University of Southern California Dissertations and Theses

Examining the relationship between common genetic variation, type 2 diabetes and prostate cancer risk in the multiethnic cohort

(USC Thesis Other)

Examining the relationship between common genetic variation, type 2 diabetes and prostate cancer risk in the multiethnic cohort

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
EXAMINING THE RELATIONSHIP BETWEEN COMMON GENETIC VARIATION,
TYPE 2 DIABETES AND PROSTATE CANCER RISK IN THE MULTIETHNIC
COHORT

by

Kevin M. Waters

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)

August 2010

Copyright 2010 Kevin M. Waters
ii
Acknowledgements

This thesis was created with incalculable aid and guidance from my advisor and
thesis committee chair Christopher A. Haiman. The remaining thesis committee
members: Brian E. Henderson, Daniel O. Stram, Richard M. Watanabe and Gerhard A.
Coetzee were also all instrumental in helping to steer me through the process of creating
this thesis.
The analysis chapters were written in concert with many collaborators. My co-
authors on these chapters are: Chapter 2 – Loïc Le Marchand, Laurence N. Kolonel,
Kristine R. Monroe, Daniel O. Stram, Brian E. Henderson, and Christopher Haiman;
Chapter 3 – Daniel O. Stram, Loïc Le Marchand, Robert J. Klein, Camilla Valtonen-
André, Mari Peltola, Laurence N. Kolonel, Brian E. Henderson, Hans Lilja, and
Christopher A. Haiman; Chapter 4 – Brian E. Henderson, Daniel O. Stram, Peggy Wan,
Laurence N. Kolonel, and Christopher A. Haiman; Chapter 5 – Daniel O. Stram,
Mohamed T. Hassanein, Loïc Le Marchand, Lynne R. Wilkens, Gertraud Maskarince,
Kristine R. Monroe, Laurence N. Kolonel, David Altshuler, Brian E. Henderson, and
Christopher A. Haiman; Chapter 6 – Daniel O. Stram, Iona Cheng, Frederick R.
Schumacher, Lynne R. Wilkens, Kristine R. Monroe, Laurence N. Kolonel, Brian E.
Henderson, Loïc Le Marchand and Christopher A. Haiman.
Additionally, I received a great deal of technical assistance from members of the
Multiethnic Cohort lab, in particular Loreall Pooler, Grace Sheng, David Wong, Hank
Huang, Peggy Wan, Edder Duarte, Erin Carter, Enrique Zelaya, and Luke Monroe.
iii
Table of Contents

Acknowledgements ii

List of Tables v

List of Figures viii

Abstract ix

Chapter 1: Background 1
Chapter 1 References 21

Chapter 2: Generalizability of Associations from Prostate Cancer 36
Genome-Wide Association Studies in Multiple Populations
Introduction 36
Methods 37
Results and Discussion 41
Chapter 2 References 54

Chapter 3: A Common Prostate Cancer Risk Variant in the 5’ 56
Region of MSMB is a Strong Predictor of Circulating
MSP Levels in Multiple Populations: The Multiethnic
Cohort Study
Introduction 56
Methods 57
Results 62
Discussion 74
Chapter 3 References 77

Chapter 4: Association of Diabetes with Prostate Cancer Risk in 80
the Multiethnic Cohort
Introduction 80
Methods 82
Results 86
Discussion 93
Chapter 4 References 98

iv

Chapter 5: Consistent Association of Type 2 Diabetes Risk 101
Variants Found in Europeans in Diverse Racial-and Ethnic
Groups
Introduction 101
Methods 103
Results 108
Discussion 126
Chapter 5 References 133

Chapter 6: Examining Known Diabetes Risk Variants for 134
Association with Prostate Cancer in a Multiethnic
Population
Introduction 134
Methods 136
Results 141
Discussion 156
Chapter 6 References 161

Chapter 7: Summary 164
Chapter 7 References 172

References 173

v

List of Tables

Table 2-1: Genotyping Efficiency: Genotype Call Rates, 39
Hardy-Weinberg Equilibrium Testing

Table 2-2: Frequencies of Risk Alleles and Associations with 42
Prostate Cancer Risk in the MEC

Table 2-3: Association with Prostate Cancer Risk by Genotype for 44
Each Racial/Ethnic Group

Table 2-4: Estimated Power for Detecting Association Given 46
Reported ORs (α=0.05)

Table 2-5: Frequency of Risk Allele and Association with Prostate 47
Cancer by Disease Subgroup and Case-Only Testing

Table 2-6: Frequencies of Risk Alleles and Associations with 48
Advanced Prostate Cancer in the MEC

Table 2-7: Frequencies of Risk Alleles and Associations with 49
Non-Advanced Prostate Cancer in the MEC

Table 2-8: Association of Risk Allele Counts with Prostate 51
Cancer (n=4,235)

Table 3-1: Descriptive Characteristics by Race/Ethnicity 63
(n=500)

Table 3-2: Least Squares Geometric Mean MSP Levels (ng/mL) 65
and Percent Change by rs10993994 Genotype

Table 3-3: Effect of Adjustment for Global European Ancestry on 67
Association of rs10993994 with Plasma MSP Levels (n=246)

Table 3-4: The Association of MSP Levels with Age, PSA Levels 68
and Suspected Risk Factors for Prostate Cancer

Table 3-5: Least Squares Geometric Mean MSP Levels by 69
Demographic Factors (n=500)

vi
Table 3-6: Least Squares Geometric Mean MSP Levels for 70
Quintiles of PSA Levels (n=500)

Table 3-7: Least Squares Geometric Mean MSP Levels by 72
Suspected Prostate Cancer Risk Factors (n=500)

Table 4-1: Descriptive Characteristics by Race/Ethnicity and 87
Diabetes Status in the Multiethnic Cohort (n=86,303),
Los Angeles, CA and Hawaii, 1993-2005

Table 4-2: Relative Risk of Prostate Cancer Associated with 89
Diabetes Status by Age and Gleason Score in the
Multiethnic Cohort, Los Angeles, CA and Hawaii
(n=86,303), 1993-2005

Table 4-3: Geometric Mean PSA Levels by Ethnicity, Diabetes 90
Status and BMI in the Multiethnic Cohort (n=2,874),
Los Angeles, CA and Hawaii, 1993-2005

Table 4-4. PSA Screening Frequencies by Level of Education 92
and Diabetes Status in the Multiethnic Cohort (n=46,970),
Los Angeles, CA and Hawaii, 1993-2005

Table 5-1: Genotyping Efficiency: Genotype Call Rates, 107
Hardy-Weinberg Equilibrium Testing

Table 5-2: The Descriptive Characteristics of Type 2 Diabetes 110
Cases and Controls in the MEC at Baseline by Racial/Ethnic
Group and Sex

Table 5-3: Power Estimates (α=0.05) to Detect Relative Risks in 112
Previous Studies

Table 5-4: The Association of Known Risk Alleles for T2D by 113
Race/Ethnicity

Table 5-5: Association with T2D Risk by Genotype 115

Table 5-6: Effects of European Ancestry Adjustment in 118
African Americans

Table 5-7: Effects of Adjustment for Education 129

vii
Table 5-8: The Association of the Total Risk Score With T2D 122
Risk by Racial/Ethnic Population

Table 5-9: Associations with Risk Score Among Subjects with 123
and without Complete Genotype Data

Table 6-1: Genotype Assay Completion Rates and Hardy Weinberg 139
Equilibrium Testing

Table 6-2: The Association of Known T2D Risk Alleles with 143
Prostate Cancer Risk by Race/Ethnicity

Table 6-3: Association with Prostate Cancer Risk by Genotype 145

Table 6-4: Association of T2D Risk Alleles with PC Risk in 147
Subjects with No Self-Report of T2D (n=4,619)

Table 6-5: Associations by Disease Severity and Case-Only 149
Testing

Table 6-6: The Association of Known T2D Risk Alleles with 150
Advanced Prostate Cancer Risk by Race/Ethnicity

Table 6-7: The Association of Known T2D Risk Alleles with 152
Non-Advanced Prostate Cancer Risk by Race/Ethnicity

Table 6-8: The Association of the Summary T2D Risk Scores with 155
Prostate Cancer Risk by Racial/Ethnic Population

viii
List of Figures

Figure 3-1: Least Squares Mean of Log MSP (ng/mL) by 64
rs10993994 Genotype for Each Racial/Ethnic Group

Figure 5-1: Risk Allele Frequencies by Racial/Ethnic Group 113

Figure 5-2: Predicted Distribution of T2D Risk from Common 124
Variants by Racial/Ethnic Group Compared to European
Americans

Figure 6-1: Estimated Power to Detect Protective Association 158
for Alleles in Pooled Sample (α = 0.05 with Bonferroni
Correction for 17 tests)

ix
Abstract

This thesis is comprised of five studies that examine the relationship between common
genetic variation, type 2 diabetes and prostate cancer risk.

Generalizability of Associations from Prostate Cancer GWAS in Multiple Populations
Genome-wide association studies have identified multiple common alleles
associated with prostate cancer risk in populations of European ancestry. Testing these
variants in other populations is needed to assess the generalizability of the associations,
and may guide fine-mapping efforts. We examined 13 of these risk variants in a
multiethnic sample of 2,768 incident prostate cancer cases and 2,359 controls from the
Multiethnic Cohort (MEC; African Americans, European Americans, Latinos, Japanese
Americans and Native Hawaiians). We estimated ethnic-specific and pooled odds ratios
and tested for ethnic heterogeneity of effects using logistic regression. In ethnic-pooled
analyses, 12 of the 13 variants were positively associated with risk, with statistically
significant associations (p<0.05) noted with 6 variants (odds ratio, 95% confidence
interval): JAZF1, rs10486567, 1.23(1.12-1.35); Xp11.2, rs5945572, 1.31(1.13-1.51);
HNF1B, rs4430796, 1.15(1.06-1.25); MSMB, rs10993994, 1.13(1.04-1.23); 11q13.2,
rs7931342, 1.13(1.03-1.23); 3p12.1, rs2660753, 1.11(1.01-1.21); SLC22A3, rs9364554,
1.10(1.00-1.21); CTBP2, rs12769019, 1.11(0.99-1.25); HNF1B, rs11649743, 1.10(0.99-
1.22); EHBP1, rs721048, 1.08(0.94-1.25); KLK2/3, rs2735839, 1.06(0.97-1.16);
17q24.3, rs1859962, 1.04(0.96-1.13); and LMTK2, rs6465657, 0.99(0.89-1.09).
x
Significant ethnic heterogeneity of effects was noted for 4 variants (EHBP1, p
het
=
3.9x10
-3
; 11q13, p
het
= 0.023; HNF1B (rs4430796), p
het
= 0.026; and KLK2/3, p
het
=
2.0x10
-3
). Although power was limited in some ethnic/racial groups due to variation in
sample size and allele frequencies, these findings suggest that a large fraction of prostate
cancer variants identified in populations of European ancestry are global markers of risk.
For many of these regions, fine-mapping in non-European samples may help localize
causal alleles and better determine their contribution to prostate cancer risk in the
population.

Common Prostate Cancer Risk Variant in the 5’ Region of MSMB is a Strong Predictor
of Circulating MSP Levels in Multiple Populations: The Multiethnic Cohort Study
Beta-microseminoprotein (MSP) is one of the three most abundantly secreted proteins of
the prostate, and has been suggested as a biomarker for prostate cancer risk. A common variant,
rs10993994, in the 5’ region of the gene which encodes MSP (MSMB), has recently been
identified as a risk factor for prostate cancer. We examined the association between rs10993994
genotype and MSP levels in a sample of 500 prostate cancer-free men from four racial/ethnic
populations in the Multiethnic Cohort (European Americans, African Americans, Latinos, and
Japanese Americans). Generalized linear models were used to estimate the association between
rs10993994 genotype and MSP levels. We observed robust associations between rs10994994
genotype and MSP levels in each racial/ethnic population (all P<10
-8
) with carriers of the C allele
having lower geometric mean MSP levels (ng/mL) (CC/CT/TT genotypes: European Americans,
28.8/20.9/10.0; African Americans, 29.0/21.9/10.9; Latinos, 29.2/17.1/8.3; and Japanese
Americans 25.8/16.4/6.7). We estimated the variant accounts for 30-50% of the variation in MSP
xi
levels in each population. We also observed significant differences in MSP levels between
populations (P=3.5x10
-6
), with MSP levels observed to be highest in African Americans and
lowest in Japanese Americans. Rs10993994 genotype is strongly associated with plasma MSP
levels in multiple racial/ethnic populations. This supports the hypothesis that rs10993994 may be
the biologically functional allele.

Association of Diabetes with Prostate Cancer Risk in the Multiethnic Cohort
Among men of European ancestry, diabetics have a lower risk of prostate cancer
than non-diabetics. The biological basis of this association is unknown. The authors have
examined whether the association is robust across populations in a population-based
prospective study. The analysis included 5,941 prostate cancer cases identified over a
12-year period (1993-2005) among 86,303 European American, African American,
Latino, Japanese American, and Native Hawaiian men from the Multiethnic Cohort. The
association between diabetes and Prostate-Specific Antigen (PSA) levels (n=2,874), and
PSA screening frequencies (n=46,970) was also examined. Diabetics had significantly
lower risk of prostate cancer than non-diabetes (Relative Risk (RR) =0.81, 95%
Confidence Interval (CI): 0.74, 0.87, P<0.001), with relative risks ranging from 0.65
(95% CI: 0.50, 0.84, P=0.001) among European Americans to 0.89 (95% CI: 0.77, 1.03,
P=0.13) among African Americans. Mean PSA levels were significantly lower in
diabetics than in non-diabetes (mean PSA levels, 1.07 and 1.28, respectively, P=0.003) as
were PSA screening frequencies (44.7% vs. 48.6%, P<0.001), however, this difference
could explain only a small portion (~20%) of the inverse association between these
xii
diseases. Diabetes is a protective factor for prostate cancer across populations,
suggesting shared risk factors that influence a common mechanism.

Consistent Association of Type 2 Diabetes Risk Variants Found in Europeans in Diverse
Racial-and Ethnic Group
It has been recently hypothesized that many of the signals detected in genome-
wide association studies (GWAS) to T2D and other diseases, despite being observed to
common variants, might in fact result from causal mutations that are rare. One prediction
of this hypothesis is that the allelic associations should be population-specific, as the
causal mutations arose after the migrations that established different populations around
the world. We selected 19 common variants found to be reproducibly associated to T2D
risk in European populations, and studied them in a large multiethnic case-control study
(6,142 cases and 7,403 controls) among men and women from 5 racial/ethnic groups
(European Americans, African Americans, Latinos, Japanese Americans, and Native
Hawaiians). In analysis pooled across ethnic groups, the allelic associations were in the
same direction as the original report for all 19 variants, and 14 of the 19 were
significantly associated with risk. In summing the number of risk alleles for each
individual, the per allele associations were highly statistically significant (P<10
-4
) and
similar in all populations (odds ratios 1.09-1.12) except in Japanese Americans the
estimated effect per allele was larger than in the other populations (1.20; P
het
=3.8x10
-4
).
We did not observe ethnic differences in the distribution of risk that would explain the
increased prevalence of type 2 diabetes in these groups as compared to European
xiii
Americans. The consistency of allelic associations in diverse racial/ethnic groups is not
predicted under the hypothesis of Goldstein regarding “synthetic associations” of rare
mutations in T2D.

Examining Known Diabetes Risk Variants for Association with Prostate Cancer in a
Multiethnic Population
Epidemiologic studies have found evidence of an inverse association between diabetes
status and prostate cancer risk. Recently genome-wide association studies of these two
diseases have identified a single risk allele at the HNF1B locus that is associated with
both diseases along with two risk loci (JAZF1 and THADA) that have been associated
with both diseases through distinct unlinked variants. We, as part of the Population
Architecture using Genomics and Epidemiology (PAGE) Study, explored the hypothesis
that common genetic variation may explain, in part, the inverse association between type
2 diabetes and prostate cancer, by examining known diabetes risk variants for their
association with prostate cancer. Our study consists of men from a prostate cancer case-
control set of 2,746 cases and 3,317 controls from five racial-ethnic groups (African
American, European American, Latino, Japanese American, and Native Hawaiian) of the
Multiethnic Cohort. The allelic discrimination assay was used to genotype 17 diabetes
risk variants identified by genome-wide association studies. Unconditional logistic
regression was used to examine these alleles for association with prostate cancer risk. In
ethnic-pooled analysis, we did not find evidence of an association with prostate cancer
for the diabetes risk alleles in either of the two loci previously associated with prostate
xiv
cancer (JAZF1, rs864745, OR 0.98, 95% CI 0.90-1.06; THADA, rs7578597, OR 1.11,
95% CI 0.98-1.25). An allele (rs7961581) in the TSPAN8 locus, which has been shown
to be positively associated with diabetes risk had a protective association with prostate
cancer at a nominally statistically significant level (OR 0.90, 95% CI 0.83-0.99,
P=0.021). Except for the risk allele at HNF1B (allele G of rs4430796), which we
previously reported to be associated inversely with prostate cancer, we found no evidence
of an association with prostate cancer for the diabetes risk loci at JAZF1 and THADA
despite other independent alleles at these loci being associated with prostate cancer. In
summary, we found an association between the TSPAN8 allele and prostate cancer risk
that needs to be replicated in larger populations. Resequencing and fine-mapping studies
to identify causal alleles in large association studies will be important in exploring the
role of common risk variants in the inverse association between diabetes and prostate
cancer.
1
Chapter 1: Background

Prostate Cancer: General Epidemiology
According to the annual report from the American Cancer Society (ACS) which
compiles incidence data from the National Cancer Institute (NCI), Centers for Disease
Control and Prevention, and the North American Association of Central Cancer
Registries and mortality data from the National Center for Health Statistics, there were an
estimated 192,280 new cases of prostate cancer in 2009. This number accounts for 25%
of cancer cases in men in the U.S., making it the most common cancer in men (non-skin),
ranking ahead of cancers of the lung & bronchus (116,090 new cases), and colon &
rectum (75,590 new cases) (1). Prostate cancer accounted for an estimated 27,360 (9%)
cancer deaths in men in the U.S. in 2009, ranking it second in cancer deaths in men
behind lung & bronchus (88,900 deaths), and just ahead of cancers of the colon & rectum
(25,240 deaths) (1). The age-adjusted incidence rate of prostate cancer peaked at ~240
cases per 100,000 in 1991 just after PSA-screening became a common clinical practice,
and has since leveled out to between 150 to 180 cases per 100,000 from 1995-2005 (1).
The annual estimated age-adjusted death rate for prostate cancer also peaked around 1991
at nearly 40 deaths per 100,000 probably due to increased detection but has since leveled
out at ~25 deaths per 100,000 in 2005 (1). Age is strongly associated with prostate
cancer as the estimated probability of developing prostate cancer rises from 1-in-41
between the ages of 40-59 to 1-in-16 between the ages of 60-69 and 1-in-8 after the age
of 70 (1). Typically, prostate cancer is an indolent disease with a 5-year survival rate of
2
~99% (1). However, the 5-year survival rate is only 32% in the prostate cancers that
were reported to have a distant stage at the time of diagnosis (1).
The incidence of prostate cancer also varies widely across racial and ethnic
populations. In the ACS 2009 report African Americans had the highest incidence rate
(248.5 per 100,000), followed by Whites (156.7), Hispanics/Latinos (138.0), Asian
Americans and Pacific Islanders (93.8), and American Indians and Alaskan Natives
(73.3) (1). Disparities in mortality rates are also observed with African Americans
having the highest (59.4 deaths per 100,000), followed by Whites (24.6), American
Indian and Alaskan Natives (21.1), Hispanics/Latinos (20.6), and Asian American and
Pacific Islanders (11.0) (1). Racial disparities in prostate cancer incidence have also been
reported in the population-based Multiethnic Cohort (MEC) in Hawaii and Los Angeles
with cancer incidence rates in African Americans being more than twice that of other
racial/ethnic groups while Japanese subjects had slightly lower incidence than Whites,
Latinos and Native Hawaiians (2).
Increased incidence of prostate cancer in Latinos and Japanese immigrants in the
U.S. (3), and Japanese immigrants in Hawaii (4) have been reported in comparison to
prostate cancer incidence in their country of birth. These differences are likely to be
partially due to differences in prostate cancer screening processes between the countries.
Another explanation of these disparities in incidence may be due to different levels of
exposure to environmental risk factors between the countries.
Many putative environmental risk factors for prostate cancer have been studied,
but none have had a consistent reproducible association with risk. Vitamin D was
3
hypothesized to be potential protective factors for prostate cancer because of ecologic
correlation studies which have shown higher rates of prostate cancer at higher geographic
latitudes where there is less exposure to the sun (5). A laboratory study in 3 prostate
cancer cell lines has also reported that prostate cancer cells have vitamin D receptors and
that Vitamin D inhibited growth and induced differentiation (6). Growth was also found
to be inhibited in human prostate epithelial cells in culture (7), and in-vivo in mouse
xenografts of human prostate cancers (8). However, meta-analysis of ten original articles
examining the association between serum Vitamin D levels and prostate cancer risk (and
an additional study with prostate cancer mortality) did not find evidence of an association
(OR 1.03; 95% CI 0.96-1.11) (9). Lycopene/tomato product intake, has also been
reported as having a potential protective effect on prostate cancer risk (10-14), but this
association has not been consistently replicated (15, 16). A potential mechanism of
prostate cancer protection for lycopene may lie in its ability to quench singlet oxygen
species (17). Likewise, selenium and Vitamin E also have antioxidative properties and
have been proposed as potential protective factor for prostate cancer. However, a recent
multi-site randomized double-blind placebo-controlled clinical trial with 35,533 men
(Selenium and Vitamin E Cancer Prevention Trial (SELECT)) found no association
between the arms assigned supplementation of Vitamin E (HR=1.13; 95% CI 0.95-1.35),
selenium (HR=1.04; 95% CI 0.87-1.24), or the combination of the two (HR=1.05; 95%
CI 0.88-1.25) with prostate cancer as compared to the placebo arm (18). Early
epidemiologic studies of dietary intake of red meat and fat reported that these dietary
4
factors may increase prostate cancer risk (19). However, recent prospective cohort
studies have not replicated these findings (20, 21).
Androgens have been proposed as potential risk factors for prostate cancer
because 1) androgens are required for the growth and differentiation of prostate epithelial
cells (22), 2) androgen deprivation therapy (medical or surgical castration) is a common
treatment to inhibit tumor growth and has been shown to have survival benefit in high-
risk cancer patients (23), and 3) a clinical trial (Prostate Cancer Prevention Trial) with
18,882 male subjects aged 55 or older studied the effects of finasteride, a 5α-reducatase
inhibitor which inhibits the conversion of testosterone to dihydrotestosterone, on prostate
cancer risk and reported that subjects in the finasteride arm had a 24.8% (95%CI 18.6%-
30.6%; P<0.001) reduction in prostate cancer prevalence over the seven year study period
(24). This study did, however, note an increase in high-grade tumors in the finasteride
group (6.4%) compared to the placebo arm (5.1%) (P=0.005) (24). Pooled analysis from
18 prospective studies of the association between endogenous sex hormones and prostate
cancer risk found no association between prostate cancer risk and serum concentrations
of testosterone, calculated free testosterone, dihydrotestosterone dehydroepiandrosterone
sulfate, androstenedione, androstanediol glucuronide, estradiol, or calculated free
estradiol (25). They did however find that serum sex hormone–binding globulin (SHBG)
was modestly protective (RR for highest quintile versus lowest = 0.86(95%CI 0.75-0.98);
P
trend
= 0.01) for prostate cancer risk (25). The biologic link between steroid hormones
with prostate cancer was also the basis for many of the early candidate gene studies (see
below) that focused on genes involved with androgens and their metabolism.
5

Genetic Associations with Prostate Cancer
Genetic susceptibility factors play a role in prostate cancer pathogenesis. A meta-
analysis of 13 case-control and cohort studies reported that men with a first-degree family
history have a relative risk of 2.5 (95%CI 2.2-2.8) for developing prostate cancer
compared to men with no family history (26). However, in addition to shared genetic
factors, this finding could also be explained by increased screening in individuals with an
affected family member. Twin studies comparing disease concordance between
monozygotic and dizygotic twins have been used to estimate the heritability of prostate
cancer. One study found a concordance rate of 21.1% among 723 pairs of monozygotic
twins and 6.4% among 13,769 pairs of dizygotic twins (27), and a second study found a
concordance rate of 27.1% among 5,933 pairs of monozygotic twins and 7.1% pairs of
dizygotic twins (28), suggesting that genetic susceptibility factors play a large role in
prostate cancer pathogenesis with estimates suggesting that between 40 and 60 percent of
prostate cancer may have a genetic component (27, 28).
Family based segregation analysis, generally of “heavily loaded pedigrees” has
been conducted in attempt to determine the most likely genetic model of prostate cancer
transmission (29). Multiple studies have found that the best fit was an autosomal
dominant model with a rare allele of high penetrance (30-35). Another study has found
that the best model included two or three rare autosomal dominant alleles with high
penetrance (36). The above studies did have differences in their estimates of the allele
frequencies (range 0.0003-0.017) and penetrance (range 63%-99%) for the risk variants
6
(30-37). Two of the studies found that there was some residual correlations not explained
by a simple autosomal dominant model due to stronger relative risk for brothers
compared to fathers (31, 33), and the population-based study in Australia found that
either an autosomal recessive or an X-linked model with more common alleles was the
best fit at older ages (35). Lastly, a population-based study with European Americans,
African Americans, and Asian Americans from the United States suggested another
possible mode of genetic risk transmission, reporting that a multifactorial model with
more common variants of low penetrance fit just as well as an autosomal dominant
model (37).

Familial Linkage Studies
The familial linkage study is an approach designed to screen the entire genome to
detect regions that are shared more than expected among affected family members. This
study design, suited to detect highly penetrant rare risk alleles, was notably successful in
detecting the BRCA1 and BRCA2 regions in familial breast cancer (38, 39). Linkage
studies have reported a number of candidate regions for hereditary prostate cancer (40),
with associations reported in the 1q23-25 (41), 1q41-43 (42), Xq27-28 (43), 20q11-13
(44), 1p36-35 (45), 4q24-25 (41), 8p22-23 (41), 16p13 (46, 47), 7q11-21 (48), 17p11
(49), 17q22 (50), and 19q13 (51) chromosomal regions. Many of the significant linkage
found in the intial reports has not been consistently reproduced (52, 53). The difficulty in
identifying prostate cancer susceptibility loci through linkage analysis may lie in genetic
heterogeneity with the pedigrees in linkage analysis being caused by many different
7
susceptibility loci (54), and/or that it can be difficult to differentiate sporadic forms of
prostate cancer from hereditary forms (40). Positional cloning, sequencing, and
subsequent association studies identified RNASEL (HPC1/1q23-25 region) (55, 56),
ELAC2 (HPC2/17p11 region) (49, 57), and MSR1 (8p23-22) (58, 59) as possible
susceptibility genes responsible for the observed linkage, but some of these associations
have not been consistently reproduced (60-63).

The Study of Common Genetic Variation: Candidate Gene Studies
The common disease common variants hypothesis (CDCV) suggests that genetic
susceptibility to common diseases may be determined by multiple variants common
within the population with relatively small penetrance (64, 65). The detection of such
modest (RR<1.5) effects requires study designs with large numbers of cases and controls.
Candidate gene studies take advantage of previous biologic knowledge, and study genes
known to be associated with prostate growth and development. Early studies focused on
the AR (Androgen Receptor) gene, with an initial report of a strong association between
prostate cancer and a CAG repeat in the AR locus (66). However, multiple subsequent
replication studies have been inconsistent (67). Additional candidate gene studies within
the MEC and elsewhere have found suggestive evidence of association of variation
within the IGF-1 locus with prostate cancer (68), but did not find evidence of association
for polymorphisms within the IGFBP1 and IGFBP3 (69), and PTEN (70) genes. More
recently, the National Cancer Institute Breast and Prostate Cancer Cohort Consortium
(BPC3), a pooled group of cohort studies with more than 8,000 prostate cancer cases and
8
controls, was designed to conduct well-powered analyses of common genetic variation in
gene regions involved in the steroid-hormone metabolism and insulin-like growth factor
pathways for association with prostate and breast cancer (71). However, the BPC3 found
little evidence of association with prostate cancer in polymorphisms in HSD17B1 (72),
CYP17 (73), IL6 and PTGS2 (74), ESR2 (75), and CYP19A1 (76). The lack of detection
of reproducible prostate cancer risk markers from candidate gene studies suggested that
common genetic susceptibility factors may lie in loci that are not yet associated with
prostate biology.

The Study of Common Genetic Variation: GWAS in Prostate Cancer
Large-scale multistage genome-wide association studies recently became feasible
because 1) technological advances in genotyping made it possible to efficiently genotype
up to 1 million SNPs in large numbers of subjects; and 2) the HapMap project provided a
public database of allele frequencies and linkage disequilibrium patterns in multiple
racial/ethnic populations that made it possible to efficiently tag common variants in the
genome without selecting redundant markers (77). These studies are designed to scan the
entire genome for outcome associations in an agnostic manner that does not consider
previous biologic knowledge of gene pathways and their association with the outcome.
GWAS utilize a case-control design and are suited to detect modest associations with
common variation. To date, they have reproducibly identified at least 30 risk loci that
explain >20% of the familial risk of prostate cancer in whites (78). GWAS in prostate
cancer have been more successful than those of other cancers such as cancer of the breast,
9
where 17 loci have been identified that explain ~8% of familial risk (79), and colon,
where 10 loci explain ~6% of the familial risk (78).
The first reports from prostate cancer genome-wide association studies (GWAS)
identified the 8q24 region as a risk locus for prostate cancer (80, 81). One of these scans
was an admixture-based genome-wide scan of African Americans which identified a 3.8
Mb region on 8q24 as having a higher percentage of African ancestry in the case subjects
(and not the controls) suggesting that there may be risk variants in this region that are
strongly associated with African ancestry which may help explain the higher prostate
cancer incidence in people of African ancestry (81). Since its initial discovery multiple
independent associations have been reported (82-87), with a recent report that there are 8
SNPs in five LD blocks that are independently associated with prostate cancer risk (82).
Additionally, risk variants specific to African Americans have also been identified (86,
87). The 8q24 region is also associated with other cancers as variants in 8q24 have been
reported in association with colorectal, breast, and bladder cancer risk (88-91). The
biological action and the mechanism(s) of the association between genetic variation and
cancer risk at 8q24 are not fully understood. The region has no annotated genes and one
study found no evidence of miRNA expression in the region or of an association between
the MYC oncogene (one of the closest genes to the region) expression and risk allele
genotypes (92). Recently, two functional enhancers occupied by androgen receptors
were detected, one of which contained a FoxA1 binding site and a prostate cancer risk
variant (rs11986220) (93). The risk allele of this SNP was associated with stronger
FoxA1 binding and androgen responsiveness (93). Another study has found that a
10
colorectal and prostate cancer variant in 8q24, rs6983267, differentially binds to
transcription factor 7-like 2 (TCF7L2) and that this region physically interacts with the
MYC oncogene (94). Additionally, it has been reported that this risk allele was associated
with enhanced responsiveness to Wnt signaling (95).
Including 8q24, >30 variants at 23 loci have been associated with prostate cancer
risk through the use of multi-stage GWAS in populations of European ancestry (84, 96-
102). These scans have been conducted by three main groups: deCode Genetics, Cancer
Research United Kingdom (CRUK), and the National Cancer Institute’s Cancer Genetics
Markers of Susceptibility project (CGEMS).
Multiple alleles have been identified by deCode Genetics which is based out of
Iceland. In 2007, a deCode GWAS reported two risk variants on chromosome 17 (96).
The study included 1,501 cases and 11,290 controls from Iceland with follow-up in 3
case-control studies from the Netherlands, Spain, and Chicago totaling 1,989 cases and
3,056 controls (96). The risk variant on 17q12 (rs4430796) was also associated with
diabetes risk. It lies in intron 2 of HNF1B (Hepatocyte-Nuclear Factor 1 Beta), a locus
previously associated with T2D (103), and transcription factor which had been previously
reported to be involved in nephrogenesis (104) and to contain rare mutations associated
with MODY (maturity-onset diabetes of the young) (105). A second variant in intron 5
of HNF1B (rs11649743) was found to be independently associated with prostate cancer
risk, but has thus far not been associated with T2D (106). The second risk variant
reported in the article was in the 17q24 (rs1859962) region and was not near any
annotated genes. A second report in 2008, this time with 1,833 cases and 21,372 controls
11
from Iceland and 7 case-control replication studies totaling 8,211 cases and 7,506
controls from Europe and the United States, reported risk alleles on chromosomes 2p15
(rs721048) and Xp11 (rs5945572) (97). The allele on 2p15 lies in an intron of the
EHBP1 gene, which has been reported to alter insulin-regulated GLUT4 recycling in
adipocytes in culture (107), but has no known association to prostate cancer. The Xp11
SNP is in the region of the NUDT10 and NUDT11 genes however a precise location of
the functional variant within the region is difficult to determine because the region has a
large LD block (~700 kb in European populations). The finding of a risk variant on the X
chromosome supports the hypothesis of an X-linked genetic component to prostate
cancer risk that was based on an observed higher correlation of risk between a case
subject and his brother than with his father (108). In 2009, a deCode report, this time
with 1,968 cases and 35,227 controls from Iceland with replication in 11 studies totaling
11,806 cases and 12,387 controls, revealed one variant in an intron of the EEFSEC gene
on chromosome 3q21 (rs10934853) and another on 19q13 (rs8102476) that is ~6 kb away
from the closest gene (PPP1R14A) (84). Lastly, a SNP (rs401681) in an intron of the
CLPTM1L gene on chromosome 5p15 was identified in a GWAS of basal cell carcinoma
and was also associated with cancer of the prostate, lung, bladder, and cervix (98); a
second locus that has been associated with multiple cancer types.
The CRUK scan, followed by replication in the PRACTICAL consortium which
includes >16,000 cases and >14,000 controls, has also revealed multiple prostate cancer
risk variants. The initial report from the UK group was based on an initial genome-wide
scan of 541,129 SNPs in 1,854 cases and 1,894 controls from the UK with replication of
12
11 SNPs in 3,268 cases and 3,366 controls from the UK and Australia (99). In their first
report, they identified risk variants at 3p12, 6q25, 7q21, 10q11, 11q13, and 19q13 and
simultaneously identified a proxy of the SNP in Xp11 that was identified by the deCode
group (99). The 10q11 variant (rs10993994) is biologically related to the prostate and is
a good candidate gene for association with prostate cancer as it is located in the 5’ region
of the MSMB (beta-microseminoprotein) gene, which encodes beta-microseminoprotein
(MSP). This protein is one of the 3 most abundantly secreted proteins of the prostate
(109). Low levels of MSP have been reported as a marker of increased prostate cancer
risk and worse prognosis after radical prostatectomy (110, 111). This along with a report
that MSP may regulate growth of the prostate through apoptosis (112), suggest that MSP
may be a tumor suppressor. There is some support that rs10993994 is the functional
variant as fine-mapping studies of the region have found it to be the best marker of
disease risk (113, 114), and functional studies have found that the non-risk allele
preferentially binds the CREB transcription factor and is associated with higher MSMB
transcription in tumor cell lines (114). The risk allele was recently associated with lower
serum MSP levels in 60 Chinese men with prostate cancer (115), and is the subject of one
of the chapters of this thesis (Chapter 3). The 19q13 region also has ties to the prostate as
the risk allele (rs2735839) lies in proximity to the KLK3 (kallikrein-related peptidase 3)
gene which encodes Prostate-specific Antigen (PSA). There is some controversy over
whether the SNP is an actual marker of prostate cancer risk as it is associated with PSA
levels which are widely used for prostate cancer screening (116). The variant in the 6q25
region (rs9364554) lies in the region of the SLC22A3 (solute carrier family 22 member 3)
13
gene which encodes a cation transporter that removes endogenous organic cations. The
variant on chromosome 7q21 (rs6465657) lies in the region of the LMTK2 (lemur
tyrosine kinase 2) gene which has been reported to be involved in endocytic and exocytic
membrane trafficking pathways and regulation of tubule formation (117). The 3p12
(rs2660753) and 11q13 (rs7931342) variants lie in gene deserts. Subsequently, a second
independent prostate cancer risk variant (rs12418451) has been detected in the 11q13
region (118).
In a second report from the UK group, which was based on the same initial
genome-wide scan of 541,129 SNPs in 1,854 cases and 1,894 controls from the UK, but
this time included stage 2 follow-up of 43,671 SNPs in 3,650 cases and 3,940 controls
from the UK and Australia, and stage 3 follow-up of 12 SNPs in 16,229 cases and 14,821
controls from the PRACTICAL Consortium (100). This study identified seven novel risk
loci in the 2p21, 2q31, 4q22, 4q24, 8p21, 11p15, and 22q13 chromosomal regions (100).
They also reported that a number of these loci are located near genes which are good
biological candidates for being associated with prostate cancer risk (Supplemental
Material; (100)). For example, the risk variant (rs1465618) in the 2p21 region is located
in an intron of THADA (Thyroid Adenoma Associated). Translocation mutations in this
gene have been associated with thyroid adenomas (119), and a missense mutation
(rs7578597) in this gene has been associated with T2D (120). The variant on 2q31
(rs12621278) is located in an intron of ITGA6 which encodes the integrin alpha 6 protein.
It has been reported that blocking production of its receptor in human prostate cancer
cells in a mouse xenograft model altered the ability of the tumor cells to migrate (121).
14
The variants on 4q22 (rs12500426 and rs17021918) are located in introns of a LIM-
domain-containing protein, PDLIM5, that is expressed higher in prostate, skeletal, brain,
muscle, colon, and leukocyte tissue than in other tissue (122). The TET2 gene is in the
proximity of the risk variant located in 4q24 (rs7679673). Mutations in this gene have
been associated with myelodysplastic/myeloproliferative neoplasms, and this gene is
highly expressed in the prostate as well as the bone marrow (123). NKX3.1 is in the
vicinity of the risk variants (rs2928679 and rs1512268) at 8p21. NKX3.1 is a homeobox
gene involved in normal differentiation of prostate epithelium, and loss of its function in
both mouse models and human patients is involved in the initiation of prostate cancer
(124). There are multiple possible candidate genes (IGF2, IGF2AS, INS, and TH) in the
vicinity of the 11p15 variant (rs7127900), with the first three being part of the insulin
family of polypeptide growth factors (100). Lastly, they report the variant (5759167) in a
non-coding region of 22q13 is ~7Mb away from a region previously associated with
prostate cancer through familial linkage studies (125-127).
In 2008, the CGEMS group reported risk variants at chromosomes 7p15 and
10q26 while also simultaneously detecting the same risk marker at the MSMB locus as
the PRACTICAL group (101). This study included an initial scan of 527,869 SNPs in
1,172 cases and 1,157 controls from a nested case-control group of the Prostate, Lung,
Colorectal, and Ovarian (PLCO) Cancer screening trial with follow-up in 26,958 SNPs in
four studies totaling 3,941 cases and 3,964 controls (101). The risk allele (rs10486567)
in the 7p15 region lies in an intron of JAZF1 (juxtaposed with another zinc finger gene
1), a gene with three zinc finger motifs that functions as a transcriptional repressor and
15
has been found to be fused to SUZ12 on chromosome 17q11 in endometrial stromal
tumors (128). A second SNP (rs864745) in an intron of JAZF1 has been reported as a risk
variant for T2D (120). Two SNPs in the region (rs849140 and rs1635852, not in LD with
rs10486567) have also been associated with variation in height (129). A recent follow-up
fine-mapping study of 106 SNPs in the JAZF1 region in a prostate cancer case-control
group of 10,286 cases and 9,135 controls of European ancestry reported that the original
prostate cancer index SNP (rs10486567) remained the strongest marker for prostate
cancer, and they did not find evidence that the index signal for type 2 diabetes (rs864745)
in the JAZF1 region was associated with prostate cancer (P=0.171) (130). The SNP in
the 10q26 region is near CTBP2 (C-terminal binding protein 2), which is a gene with two
isoforms one of which functions as a transcriptional repressor while the other is involved
in synaptic ribbons (131). Lastly, the CGEMs data has also been used in combination
with other datasets to reveal a suggestive novel marker of aggressive prostate cancer
(102). This SNP is in an intron of the DAB2IP gene, a RAS-GTPase-activating protein
that is thought to be a tumor suppressor gene for prostate cancer (102). It has recently
been reported that knocking down endogenous DAB2IP in human cancer cells of a
human prostate xenograft-mouse model led to metastases in lymph nodes and distant
organs (132).
While these studies have been extremely successful and have provided much
insight into prostate cancer biology, many questions remain as only ~20% of the familial
risk of prostate cancer is currently able to be explained by these known risk loci.
Possible explanations of the remaining familial risk include: the presence of many
16
additional common variants with even smaller effects that were missed due to a lack of
power or inadequate coverage of GWAS chips, other types of genetic variants, such as
copy number polymorphisms (CNPs) or structural rearrangements, that are not tagged by
the SNPs in the genome-wide scans, rare variants with large effects that are not suitable
to be detected by GWAS, and/or gene x gene or gene x environment interactions have yet
to be explored. It is also possible that heritability has been overestimated by the twin
studies and the monozygotic prostate cancer concordance is due to shared environmental
risk factors that are unaccounted for (78). In addition, the 30 risk loci for prostate cancer
have primarily been studied in populations of European descent. Chapter 2 of this thesis
studies whether these associations are applicable to other racial/ethnic groups as
differences in linkage disequilibrium, allele frequencies, and exposure to environmental
factors across racial/ethnic groups may alter these associations.

The Inverse Association between Prostate Cancer and Type 2 Diabetes
Epidemiologic studies have suggested a protective relationship between Type 2
diabetes (T2D) and prostate cancer risk. This is contrary to other cancers in which meta-
analysis has reported diabetes to increase the risk of breast, colorectal, bladder, and
endometrial cancer (133-136). A meta-analysis of 19 publications examining an
association between prostate cancer and T2D reported an inverse relationship, with
diabetics having a ~20% lower risk of developing prostate cancer than non-diabetics
(137). Twelve of the 19 publications reported an inverse association between the two
diseases (7 of which were statistically significant), and 7 reported a positive association
17
though none were statistically significant (137). Significant heterogeneity between
studies was observed with stronger inverse associations reported in studies that adjusted
for 3 or more covariates and BMI (137). They also found a stronger association in
studies carried out during the post-PSA era (RR 0.73; 95%CI 0.64-0.83) than those in the
pre-PSA era (RR 0.94; 95%CI 0.85-1.03) (137). Unlike prostate cancer, many
environmental factors have been consistently associated with T2D risk. Increased age,
BMI, physical inactivity, gestational diabetes, family history of T2D, and ethnicity are all
considered risk factors of T2D (138). The majority of these studies have been comprised
of subjects of European ancestry (137, 139-145), and to our knowledge this observation
had not been previously replicated in a large prospective multi-ethnic population.
Chapter 4 of this thesis focuses on whether the association between these diseases is
observed in multiple racial/ethnic populations and explores possible explanations of this
inverse association.

Genetics of Type 2 Diabetes
Like prostate cancer, T2D susceptibility also has a genetic component, as
heritability estimates of 26% and 61% have been reported for type 2 diabetes and
abnormal glucose tolerance, respectively (146). Genetic association studies have
identified many common genetic susceptibility variants (96, 120, 147-157). The
rs1801282 (Pro12Ala) mutation was reported as a risk factor in a candidate gene study of
PPARG (147), which was under study because of its involvement in adipogenesis
regulation and because it is a target of the thiazoladinedione drug class (158, 159). A risk
18
allele in the KCNJ11 gene was identified in a candidate gene study of genes encoding the
beta-cell KATP channel subunits (148). Common variants in the WFS1 locus were
examined for T2D risk because of its association with Beta-cell function and Wolfram’s
syndrome (diabetes mellitus is part of the syndrome) (149). A common risk variant in
HNF1B was identified as part of a candidate study of MODY (maturity-onset diabetes of
the young) genes (103). The TCF7L2 locus, the region most strongly and consistently
associated with T2D, was confirmed as a risk locus (150) after being suggested in a
linkage analysis (160, 161). Large-scale, multi-stage GWAS of T2D have been
conducted in subjects of European descent, with a few in subjects of Asian descent, and
have identified many additional risk loci and confirmed many of the risk loci from the
above candidate gene studies (96, 120, 151-157). Additional loci have been associated
with T2D through GWAS using quantitative traits of T2D (including fasting plasma
glucose, insulin response to glucose, fasting insulin levels, adiposity traits, glycated
hemoglobin levels, and 28-year average fasting plasma glucose) as the dependent
outcome variables (156, 157, 162-173). Chapter 5 of this thesis explores whether these
associations are consistent in multiple populations and whether they account for observed
differences in T2D prevalence across the populations of the MEC.
As previously mentioned, the two diseases have three genetic risk loci in
common; HNF1B, JAZF1, and THADA. An allele (rs757210) in intron 2 of HNF1B
(TCF2) was associated with diabetes in a candidate gene study of MODY genes (103).
Later, the deCode group identified the ‘A’ allele of rs4430796 (associated with
rs757210), also in intron 2 of the HNF1B gene, as a risk factor for prostate cancer and
19
also as a protective factor for T2D (96). To date, this is the only single SNP that is
common to both diseases. The JAZF1 and THADA loci, albeit with distinct and
uncorrelated alleles, have also been associated with both diabetes and prostate cancer
through GWAS of each disease (100, 101, 120). Chapter 6 of this thesis focuses on the
potential role of T2D risk variants as underlying the inverse association of T2D with
prostate cancer risk.

The Multiethnic Cohort
Examining allele frequencies and diseases associations in multiple racial-ethnic
populations may provide insight into the role that genetic susceptibility plays in the large
differences in incidence rates of both prostate cancer and T2D among racial-ethnic
populations. All of the studies included in this thesis were conducted within the
Multiethnic Cohort (MEC), a large population-based prospective cohort study with five
main racial-ethnic groups from Los Angeles and Hawaii (African Americans and Latinos
from Los Angeles, and European Americans, Japanese Americans, and Native Hawaiians
from Hawaii) (174). The MEC collected detailed information on dietary factors, non-
dietary factors (e.g. smoking, physical activity), past medical history, and family history
of disease, with the goal of studying the relationship between these factors and cancer
and also studying the contribution of these factors to ethnic differences in cancer
incidence (174). The addition of the MEC biorepository, which began in 2002, made it
possible to also study germline genetic variation and biomarker levels in the blood. The
MEC has the advantage of providing a large prospective multiethnic study population for
20
cohort analysis, and providing a defined base population from which to select cases and
controls for nested genetic association studies. As established, the MEC provides an
excellent opportunity to compare exposure frequencies, disease rates, and effect estimates
between racial-ethnic groups. Lastly, the prospective nature of the MEC gives the benefit
of describing temporal relationships between exposure and disease occurrence.

Objectives
This thesis has two main focuses. First, we will explore prostate cancer genetic
susceptibility factors in multiple racial/ethnic populations 1) Chapter 2 describes the
testing of 13 established prostate cancer risk alleles for association with prostate cancer in
the MEC to examine the generalizability of risk to multiple racial-ethnic populations, and
2) Chapter 3 presents results of an investigation of the MSMB allele in association with
plasma MSP levels in the racial/ethnic groups in the MEC, providing a functional and
physiologic basis for its association with prostate cancer risk. As a second focus of this
thesis we examined the strength and potential causes of the association between prostate
cancer and type 2 diabetes in a multiethnic population: 1) Chapter 4 is a prospective
cohort analysis in the MEC comparing prostate cancer risk among diabetics to non-
diabetics, 2) Chapter 5 is a study of established T2D genetic risk variants in the MEC
populations in order to assess whether these SNPs are markers of genetic susceptibility to
T2D in multiple ethnic populations; and 3) in Chapter 6 we examine these T2D markers
individually and in aggregate for association with prostate cancer risk..

21
Chapter 1 References

1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2009. CA Cancer J Clin
2009;594:225-49.

2. Kolonel LN, Altshuler D, Henderson BE. The multiethnic cohort study: exploring
genes, lifestyle and cancer risk. Nat Rev Cancer 2004;47:519-27.

3. Shimizu H, Ross RK, Bernstein L, et al. Cancers of the prostate and breast among
Japanese and white immigrants in Los Angeles County. Br J Cancer
1991;636:963-6.

4. Kolonel LN. Racial and Geographic variations in prostate cancer and the effect of
migration. In: Fortner JG, Sharp P.A. , ed. Accomplishments in Cancer Research.
Philadelphia (PA): Lippincott-Raven, 1996:221-30.

5. Schwartz GG, Hulka BS. Is vitamin D deficiency a risk factor for prostate cancer?
(Hypothesis). Anticancer Res 1990;105A:1307-11.

6. Skowronski RJ, Peehl DM, Feldman D. Vitamin D and prostate cancer: 1,25
dihydroxyvitamin D3 receptors and actions in human prostate cancer cell lines.
Endocrinology 1993;1325:1952-60.

7. Peehl DM, Skowronski RJ, Leung GK, et al. Antiproliferative effects of 1,25-
dihydroxyvitamin D3 on primary cultures of human prostatic cells. Cancer Res
1994;543:805-10.

8. Oades GM, Dredge K, Kirby RS, et al. Vitamin D receptor-dependent antitumour
effects of 1,25-dihydroxyvitamin D3 and two synthetic analogues in three in vivo
models of prostate cancer. BJU Int 2002;906:607-16.

9. Yin L, Raum E, Haug U, et al. Meta-analysis of longitudinal studies: Serum
vitamin D and prostate cancer risk. Cancer Epidemiol 2009.

10. Giovannucci E, Ascherio A, Rimm EB, et al. Intake of carotenoids and retinol in
relation to risk of prostate cancer. J Natl Cancer Inst 1995;8723:1767-76.

11. Gann PH, Ma J, Giovannucci E, et al. Lower prostate cancer risk in men with
elevated plasma lycopene levels: results of a prospective analysis. Cancer Res
1999;596:1225-30.

12. Giovannucci E, Rimm EB, Liu Y, et al. A prospective study of tomato products,
lycopene, and prostate cancer risk. J Natl Cancer Inst 2002;945:391-8.
22

13. Norrish AE, Jackson RT, Sharpe SJ, et al. Prostate cancer and dietary carotenoids.
Am J Epidemiol 2000;1512:119-23.

14. Lu QY, Hung JC, Heber D, et al. Inverse associations between plasma lycopene
and other carotenoids and prostate cancer. Cancer Epidemiol Biomarkers Prev
2001;107:749-56.

15. Kirsh VA, Mayne ST, Peters U, et al. A prospective study of lycopene and tomato
product intake and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev
2006;151:92-8.

16. Stram DO, Hankin JH, Wilkens LR, et al. Prostate cancer incidence and intake of
fruits, vegetables and related micronutrients: the multiethnic cohort study*
(United States). Cancer Causes Control 2006;179:1193-207.

17. Di Mascio P, Kaiser S, Sies H. Lycopene as the most efficient biological
carotenoid singlet oxygen quencher. Arch Biochem Biophys 1989;2742:532-8.

18. Lippman SM, Klein EA, Goodman PJ, et al. Effect of selenium and vitamin E on
risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer
Prevention Trial (SELECT). JAMA 2009;3011:39-51.

19. Kolonel LN. Fat, meat, and prostate cancer. Epidemiol Rev 2001;231:72-81.

20. Crowe FL, Key TJ, Appleby PN, et al. Dietary fat intake and risk of prostate
cancer in the European Prospective Investigation into Cancer and Nutrition. Am J
Clin Nutr 2008;875:1405-13.

21. Park SY, Murphy SP, Wilkens LR, et al. Fat and meat intake and prostate cancer
risk: the multiethnic cohort study. Int J Cancer 2007;1216:1339-45.

22. Platz EA, Giovannucci E. The epidemiology of sex steroid hormones and their
signaling and metabolic pathways in the etiology of prostate cancer. J Steroid
Biochem Mol Biol 2004;924:237-53.

23. Sharifi N, Gulley JL, Dahut WL. Androgen deprivation therapy for prostate
cancer. JAMA 2005;2942:238-44.

24. Thompson IM, Goodman PJ, Tangen CM, et al. The influence of finasteride on
the development of prostate cancer. N Engl J Med 2003;3493:215-24.

23
25. Roddam AW, Allen NE, Appleby P, et al. Endogenous sex hormones and prostate
cancer: a collaborative analysis of 18 prospective studies. J Natl Cancer Inst
2008;1003:170-83.
26. Johns LE, Houlston RS. A systematic review and meta-analysis of familial
prostate cancer risk. BJU Int 2003;919:789-94.

27. Lichtenstein P, Holm NV, Verkasalo PK, et al. Environmental and heritable
factors in the causation of cancer--analyses of cohorts of twins from Sweden,
Denmark, and Finland. N Engl J Med 2000;3432:78-85.

28. Page WF, Braun MM, Partin AW, et al. Heredity and prostate cancer: a study of
World War II veteran twins. Prostate 1997;334:240-5.

29. Schaid DJ. The complex genetic epidemiology of prostate cancer. Hum Mol
Genet 2004;13 Spec No 1:R103-21.

30. Carter BS, Beaty TH, Steinberg GD, et al. Mendelian inheritance of familial
prostate cancer. Proc Natl Acad Sci U S A 1992;898:3367-71.

31. Schaid DJ, McDonnell SK, Blute ML, et al. Evidence for autosomal dominant
inheritance of prostate cancer. Am J Hum Genet 1998;626:1425-38.

32. Verhage BA, Baffoe-Bonnie AB, Baglietto L, et al. Autosomal dominant
inheritance of prostate cancer: a confirmatory study. Urology 2001;571:97-101.

33. Valeri A, Briollais L, Azzouzi R, et al. Segregation analysis of prostate cancer in
France: evidence for autosomal dominant inheritance and residual brother-brother
dependence. Ann Hum Genet 2003;67Pt 2:125-37.

34. Gronberg H, Damber L, Damber JE, et al. Segregation analysis of prostate cancer
in Sweden: support for dominant inheritance. Am J Epidemiol 1997;1467:552-7.

35. Cui J, Staples MP, Hopper JL, et al. Segregation analyses of 1,476 population-
based Australian families affected by prostate cancer. Am J Hum Genet
2001;685:1207-18.

36. Conlon EM, Goode EL, Gibbs M, et al. Oligogenic segregation analysis of
hereditary prostate cancer pedigrees: evidence for multiple loci affecting age at
onset. Int J Cancer 2003;1055:630-5.

37. Gong G, Oakley-Girvan I, Wu AH, et al. Segregation analysis of prostate cancer
in 1,719 white, African-American and Asian-American families in the United
States and Canada. Cancer Causes Control 2002;135:471-82.

24
38. Hall JM, Lee MK, Newman B, et al. Linkage of early-onset familial breast cancer
to chromosome 17q21. Science 1990;2504988:1684-9.

39. Wooster R, Bignell G, Lancaster J, et al. Identification of the breast cancer
susceptibility gene BRCA2. Nature 1995;3786559:789-92.

40. Simard J, Dumont M, Labuda D, et al. Prostate cancer susceptibility genes:
lessons learned and challenges posed. Endocr Relat Cancer 2003;102:225-59.

41. Smith JR, Freije D, Carpten JD, et al. Major susceptibility locus for prostate
cancer on chromosome 1 suggested by a genome-wide search. Science
1996;2745291:1371-4.

42. Berthon P, Valeri A, Cohen-Akenine A, et al. Predisposing gene for early-onset
prostate cancer, localized on chromosome 1q42.2-43. Am J Hum Genet
1998;626:1416-24.

43. Xu J, Meyers D, Freije D, et al. Evidence for a prostate cancer susceptibility locus
on the X chromosome. Nat Genet 1998;202:175-9.

44. Berry R, Schroeder JJ, French AJ, et al. Evidence for a prostate cancer-
susceptibility locus on chromosome 20. Am J Hum Genet 2000;671:82-91.

45. Gibbs M, Stanford JL, McIndoe RA, et al. Evidence for a rare prostate cancer-
susceptibility locus at chromosome 1p36. Am J Hum Genet 1999;643:776-87.

46. Suarez BK, Lin J, Burmester JK, et al. A genome screen of multiplex sibships
with prostate cancer. Am J Hum Genet 2000;663:933-44.

47. Gibbs M, Stanford JL, Jarvik GP, et al. A genomic scan of families with prostate
cancer identifies multiple regions of interest. Am J Hum Genet 2000;671:100-9.

48. Friedrichsen DM, Stanford JL, Isaacs SD, et al. Identification of a prostate cancer
susceptibility locus on chromosome 7q11-21 in Jewish families. Proc Natl Acad
Sci U S A 2004;1017:1939-44.

49. Tavtigian SV, Simard J, Teng DH, et al. A candidate prostate cancer susceptibility
gene at chromosome 17p. Nat Genet 2001;272:172-80.

50. Gillanders EM, Xu J, Chang BL, et al. Combined genome-wide scan for prostate
cancer susceptibility genes. J Natl Cancer Inst 2004;9616:1240-7.

51. Witte JS, Goddard KA, Conti DV, et al. Genomewide scan for prostate cancer-
aggressiveness loci. Am J Hum Genet 2000;671:92-9.
25
52. Xu J, Dimitrov L, Chang BL, et al. A combined genomewide linkage scan of
1,233 families for prostate cancer-susceptibility genes conducted by the
international consortium for prostate cancer genetics. Am J Hum Genet
2005;772:219-29.

53. Christensen GB, Baffoe-Bonnie AB, George A, et al. Genome-wide linkage
analysis of 1,233 prostate cancer pedigrees from the International Consortium for
Prostate Cancer Genetics using novel sumLINK and sumLOD analyses. Prostate
2010;707:735-44.

54. Ostrander EA, Stanford JL. Genetics of prostate cancer: too many loci, too few
genes. Am J Hum Genet 2000;676:1367-75.

55. Carpten J, Nupponen N, Isaacs S, et al. Germline mutations in the ribonuclease L
gene in families showing linkage with HPC1. Nat Genet 2002;302:181-4.

56. Rokman A, Ikonen T, Seppala EH, et al. Germline alterations of the RNASEL
gene, a candidate HPC1 gene at 1q25, in patients and families with prostate
cancer. Am J Hum Genet 2002;705:1299-304.

57. Rebbeck TR, Walker AH, Zeigler-Johnson C, et al. Association of HPC2/ELAC2
genotypes and prostate cancer. Am J Hum Genet 2000;674:1014-9.

58. Xu J, Zheng SL, Komiya A, et al. Germline mutations and sequence variants of
the macrophage scavenger receptor 1 gene are associated with prostate cancer
risk. Nat Genet 2002;322:321-5.

59. Xu J, Zheng SL, Komiya A, et al. Common sequence variants of the macrophage
scavenger receptor 1 gene are associated with prostate cancer risk. Am J Hum
Genet 2003;721:208-12.

60. Severi G, Giles GG, Southey MC, et al. ELAC2/HPC2 polymorphisms, prostate-
specific antigen levels, and prostate cancer. J Natl Cancer Inst 2003;9511:818-24.

61. Camp NJ, Tavtigian SV. Meta-analysis of associations of the Ser217Leu and
Ala541Thr variants in ELAC2 (HPC2) and prostate cancer. Am J Hum Genet
2002;716:1475-8.

62. Rokman A, Ikonen T, Mononen N, et al. ELAC2/HPC2 involvement in hereditary
and sporadic prostate cancer. Cancer Res 2001;6116:6038-41.

63. Wang L, McDonnell SK, Cunningham JM, et al. No association of germline
alteration of MSR1 with prostate cancer risk. Nat Genet 2003;352:128-9.

26
64. Schork NJ, Murray SS, Frazer KA, et al. Common vs. rare allele hypotheses for
complex diseases. Curr Opin Genet Dev 2009;193:212-9.

65. Reich DE, Lander ES. On the allelic spectrum of human disease. Trends Genet
2001;179:502-10.

66. Giovannucci E, Stampfer MJ, Krithivas K, et al. The CAG repeat within the
androgen receptor gene and its relationship to prostate cancer. Proc Natl Acad Sci
U S A 1997;947:3320-3.

67. Freedman ML, Pearce CL, Penney KL, et al. Systematic evaluation of genetic
variation at the androgen receptor locus and risk of prostate cancer in a
multiethnic cohort study. Am J Hum Genet 2005;761:82-90.

68. Cheng I, Stram DO, Penney KL, et al. Common genetic variation in IGF1 and
prostate cancer risk in the Multiethnic Cohort. J Natl Cancer Inst 2006;982:123-
34.

69. Cheng I, Penney KL, Stram DO, et al. Haplotype-based association studies of
IGFBP1 and IGFBP3 with prostate and breast cancer risk: the multiethnic cohort.
Cancer Epidemiol Biomarkers Prev 2006;1510:1993-7.

70. Haiman CA, Stram DO, Cheng I, et al. Common genetic variation at PTEN and
risk of sporadic breast and prostate cancer. Cancer Epidemiol Biomarkers Prev
2006;155:1021-5.

71. Hunter DJ, Riboli E, Haiman CA, et al. A candidate gene approach to searching
for low-penetrance breast and prostate cancer genes. Nat Rev Cancer
2005;512:977-85.

72. Kraft P, Pharoah P, Chanock SJ, et al. Genetic variation in the HSD17B1 gene
and risk of prostate cancer. PLoS Genet 2005;15:e68.

73. Setiawan VW, Schumacher FR, Haiman CA, et al. CYP17 genetic variation and
risk of breast and prostate cancer from the National Cancer Institute Breast and
Prostate Cancer Cohort Consortium (BPC3). Cancer Epidemiol Biomarkers Prev
2007;1611:2237-46.

74. Dossus L, Kaaks R, Canzian F, et al. PTGS2 and IL6 genetic variation and risk of
breast and prostate cancer: results from the Breast and Prostate Cancer Cohort
Consortium (BPC3). Carcinogenesis;313:455-61.

27
75. Chen YC, Kraft P, Bretsky P, et al. Sequence variants of estrogen receptor beta
and risk of prostate cancer in the National Cancer Institute Breast and Prostate
Cancer Cohort Consortium. Cancer Epidemiol Biomarkers Prev 2007;1610:1973-
81.

76. Travis RC, Schumacher F, Hirschhorn JN, et al. CYP19A1 genetic variation in
relation to prostate cancer risk and circulating sex hormone concentrations in men
from the Breast and Prostate Cancer Cohort Consortium. Cancer Epidemiol
Biomarkers Prev 2009;1810:2734-44.

77. The International HapMap Project. Nature 2003;4266968:789-96.

78. Varghese JS, Easton DF. Genome-wide association studies in common cancers-
what have we learnt? Curr Opin Genet Dev.

79. Turnbull C, Ahmed S, Morrison J, et al. Genome-wide association study identifies
five new breast cancer susceptibility loci. Nat Genet;426:504-7.

80. Amundadottir LT, Sulem P, Gudmundsson J, et al. A common variant associated
with prostate cancer in European and African populations. Nat Genet
2006;386:652-8.

81. Freedman ML, Haiman CA, Patterson N, et al. Admixture mapping identifies
8q24 as a prostate cancer risk locus in African-American men. Proc Natl Acad Sci
U S A 2006;10338:14068-73.

82. Al Olama AA, Kote-Jarai Z, Giles GG, et al. Multiple loci on 8q24 associated
with prostate cancer susceptibility. Nat Genet 2009;4110:1058-60.

83. Yeager M, Chatterjee N, Ciampa J, et al. Identification of a new prostate cancer
susceptibility locus on chromosome 8q24. Nat Genet 2009;4110:1055-7.

84. Gudmundsson J, Sulem P, Gudbjartsson DF, et al. Genome-wide association and
replication studies identify four variants associated with prostate cancer
susceptibility. Nat Genet 2009;4110:1122-6.

85. Gudmundsson J, Sulem P, Manolescu A, et al. Genome-wide association study
identifies a second prostate cancer susceptibility variant at 8q24. Nat Genet
2007;395:631-7.

86. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24
independently affect risk for prostate cancer. Nat Genet 2007;395:638-44.

28
87. Robbins C, Torres JB, Hooker S, et al. Confirmation study of prostate cancer risk
variants at 8q24 in African Americans identifies a novel risk locus. Genome Res
2007;1712:1717-22.

88. Zanke BW, Greenwood CM, Rangrej J, et al. Genome-wide association scan
identifies a colorectal cancer susceptibility locus on chromosome 8q24. Nat Genet
2007;398:989-94.

89. Haiman CA, Le Marchand L, Yamamato J, et al. A common genetic risk factor
for colorectal and prostate cancer. Nat Genet 2007;398:954-6.

90. Easton DF, Pooley KA, Dunning AM, et al. Genome-wide association study
identifies novel breast cancer susceptibility loci. Nature 2007;4477148:1087-93.

91. Kiemeney LA, Thorlacius S, Sulem P, et al. Sequence variant on 8q24 confers
susceptibility to urinary bladder cancer. Nat Genet 2008;4011:1307-12.

92. Pomerantz MM, Beckwith CA, Regan MM, et al. Evaluation of the 8q24 prostate
cancer risk locus and MYC expression. Cancer Res 2009;6913:5568-74.

93. Jia L, Landan G, Pomerantz M, et al. Functional enhancers at the gene-poor 8q24
cancer-linked locus. PLoS Genet 2009;58:e1000597.

94. Pomerantz MM, Ahmadiyeh N, Jia L, et al. The 8q24 cancer risk variant
rs6983267 shows long-range interaction with MYC in colorectal cancer. Nat
Genet 2009;418:882-4.

95. Tuupanen S, Turunen M, Lehtonen R, et al. The common colorectal cancer
predisposition SNP rs6983267 at chromosome 8q24 confers potential to enhanced
Wnt signaling. Nat Genet 2009;418:885-90.

96. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome
17 confer prostate cancer risk, and the one in TCF2 protects against type 2
diabetes. Nat Genet 2007;398:977-83.

97. Gudmundsson J, Sulem P, Rafnar T, et al. Common sequence variants on 2p15
and Xp11.22 confer susceptibility to prostate cancer. Nat Genet 2008;403:281-3.

98. Rafnar T, Sulem P, Stacey SN, et al. Sequence variants at the TERT-CLPTM1L
locus associate with many cancer types. Nat Genet 2009;412:221-7.

99. Eeles RA, Kote-Jarai Z, Giles GG, et al. Multiple newly identified loci associated
with prostate cancer susceptibility. Nat Genet 2008;403:316-21.

29
100. Eeles RA, Kote-Jarai Z, Al Olama AA, et al. Identification of seven new prostate
cancer susceptibility loci through a genome-wide association study. Nat Genet
2009;4110:1116-21.

101. Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified in a genome-wide
association study of prostate cancer. Nat Genet 2008;403:310-5.

102. Duggan D, Zheng SL, Knowlton M, et al. Two genome-wide association studies
of aggressive prostate cancer implicate putative prostate tumor suppressor gene
DAB2IP. J Natl Cancer Inst 2007;9924:1836-44.

103. Winckler W, Weedon MN, Graham RR, et al. Evaluation of common variants in
the six known maturity-onset diabetes of the young (MODY) genes for
association with type 2 diabetes. Diabetes 2007;563:685-93.

104. Dudziak K, Mottalebi N, Senkel S, et al. Transcription factor HNF1beta and novel
partners affect nephrogenesis. Kidney Int 2008;742:210-7.

105. Fajans SS, Bell GI, Polonsky KS. Molecular mechanisms and clinical
pathophysiology of maturity-onset diabetes of the young. N Engl J Med
2001;34513:971-80.
106. Sun J, Zheng SL, Wiklund F, et al. Evidence for two independent prostate cancer
risk-associated loci in the HNF1B gene at 17q12. Nat Genet 2008;4010:1153-5.

107. Guilherme A, Soriano NA, Furcinitti PS, et al. Role of EHD1 and EHBP1 in
perinuclear sorting and insulin-regulated GLUT4 recycling in 3T3-L1 adipocytes.
J Biol Chem 2004;27938:40062-75.

108. Monroe KR, Yu MC, Kolonel LN, et al. Evidence of an X-linked or recessive
genetic component to prostate cancer risk. Nat Med 1995;18:827-9.

109. Lilja H, Abrahamsson PA. Three predominant proteins secreted by the human
prostate gland. Prostate 1988;121:29-38.

110. Nam RK, Reeves JR, Toi A, et al. A novel serum marker, total prostate secretory
protein of 94 amino acids, improves prostate cancer detection and helps identify
high grade cancers at diagnosis. J Urol 2006;1754:1291-7.

111. Reeves JR, Dulude H, Panchal C, et al. Prognostic value of prostate secretory
protein of 94 amino acids and its binding protein after radical prostatectomy. Clin
Cancer Res 2006;1220 Pt 1:6018-22.

30
112. Garde SV, Basrur VS, Li L, et al. Prostate secretory protein (PSP94) suppresses
the growth of androgen-independent prostate cancer cell line (PC3) and
xenografts by inducing apoptosis. Prostate 1999;382:118-25.

113. Chang BL, Cramer SD, Wiklund F, et al. Fine mapping association study and
functional analysis implicate a SNP in MSMB at 10q11 as a causal variant for
prostate cancer risk. Hum Mol Genet 2009;187:1368-75.

114. Lou H, Yeager M, Li H, et al. Fine mapping and functional analysis of a common
variant in MSMB on chromosome 10q11.2 associated with prostate cancer
susceptibility. Proc Natl Acad Sci U S A 2009;10619:7933-8.

115. Xu B, Wang J, Tong N, et al. A functional polymorphism in MSMB gene
promoter is associated with prostate cancer risk and serum MSMB expression.
Prostate 2010;7010:146-52.

116. Ahn J, Berndt SI, Wacholder S, et al. Variation in KLK genes, prostate-specific
antigen and risk of prostate cancer. Nat Genet 2008;409:1032-4; author reply 5-6.

117. Chibalina MV, Seaman MN, Miller CC, et al. Myosin VI and its interacting
protein LMTK2 regulate tubule formation and transport to the endocytic recycling
compartment. J Cell Sci 2007;120Pt 24:4278-88.

118. Zheng SL, Stevens VL, Wiklund F, et al. Two independent prostate cancer risk-
associated Loci at 11q13. Cancer Epidemiol Biomarkers Prev 2009;186:1815-20.

119. Rippe V, Drieschner N, Meiboom M, et al. Identification of a gene rearranged by
2p21 aberrations in thyroid adenomas. Oncogene 2003;2238:6111-4.

120. Zeggini E, Scott LJ, Saxena R, et al. Meta-analysis of genome-wide association
data and large-scale replication identifies additional susceptibility loci for type 2
diabetes. Nat Genet 2008;405:638-45.

121. King TE, Pawar SC, Majuta L, et al. The role of alpha 6 integrin in prostate
cancer migration and bone pain in a novel xenograft model. PLoS One
2008;310:e3535.

122. Wu M, Li Y, Ji C, et al. Cloning and identification of a novel human gene
PDLIM5, a homolog of AD-associated neuronal thread protein (AD7c-NTP).
DNA Seq 2004;152:144-7.

123. Jankowska AM, Szpurka H, Tiu RV, et al. Loss of heterozygosity 4q24 and TET2
mutations associated with myelodysplastic/myeloproliferative neoplasms. Blood
2009;11325:6403-10.
31
124. Abate-Shen C, Shen MM, Gelmann E. Integrating differentiation and cancer: the
Nkx3.1 homeobox gene in prostate organogenesis and carcinogenesis.
Differentiation 2008;766:717-27.

125. Camp NJ, Farnham JM, Cannon-Albright LA. Localization of a prostate cancer
predisposition gene to an 880-kb region on chromosome 22q12.3 in Utah high-
risk pedigrees. Cancer Res 2006;6620:10205-12.

126. Johanneson B, McDonnell SK, Karyadi DM, et al. Fine mapping of familial
prostate cancer families narrows the interval for a susceptibility locus on
chromosome 22q12.3 to 1.36 Mb. Hum Genet 2008;1231:65-75.

127. Camp NJ, Cannon-Albright LA, Farnham JM, et al. Compelling evidence for a
prostate cancer gene at 22q12.3 by the International Consortium for Prostate
Cancer Genetics. Hum Mol Genet 2007;1611:1271-8.

128. Koontz JI, Soreng AL, Nucci M, et al. Frequent fusion of the JAZF1 and JJAZ1
genes in endometrial stromal tumors. Proc Natl Acad Sci U S A 2001;9811:6348-
53.

129. Johansson A, Marroni F, Hayward C, et al. Common variants in the JAZF1 gene
associated with height identified by linkage and genome-wide association
analysis. Hum Mol Genet 2009;182:373-80.

130. Prokunina-Olsson L, Fu YP, Tang W, et al. Refining the Prostate Cancer Genetic
Association within the JAZF1 Gene on Chromosome 7p15.2. Cancer Epidemiol
Biomarkers Prev.

131. Zhao LJ, Subramanian T, Chinnadurai G. Inhibition of transcriptional activation
and cell proliferation activities of adenovirus E1A by the unique N-terminal
domain of CtBP2. Oncogene 2008;2739:5214-22.

132. Xie D, Gore C, Liu J, et al. Role of DAB2IP in modulating epithelial-to-
mesenchymal transition and prostate cancer metastasis. Proc Natl Acad Sci U S
A;1076:2485-90.

133. Larsson SC, Mantzoros CS, Wolk A. Diabetes mellitus and risk of breast cancer:
a meta-analysis. Int J Cancer 2007;1214:856-62.

134. Larsson SC, Orsini N, Wolk A. Diabetes mellitus and risk of colorectal cancer: a
meta-analysis. J Natl Cancer Inst 2005;9722:1679-87.

135. Larsson SC, Orsini N, Brismar K, et al. Diabetes mellitus and risk of bladder
cancer: a meta-analysis. Diabetologia 2006;4912:2819-23.
32

136. Friberg E, Orsini N, Mantzoros CS, et al. Diabetes mellitus and risk of
endometrial cancer: a meta-analysis. Diabetologia 2007;507:1365-74.

137. Kasper JS, Giovannucci E. A meta-analysis of diabetes mellitus and the risk of
prostate cancer. Cancer Epidemiol Biomarkers Prev 2006;1511:2056-62.

138. MedlinePlus [Internet]. Bethesda (MD): National Library of Medicine (US);
[updated 2009 Dec 7]. Type II Diabetes - risk factors; [updated 2009 May 20;
cited 2009 Dec 7. Available from:
http://www.nlm.nih.gov/medlineplus/ency/article/002072.htm.

139. Calton BA, Chang SC, Wright ME, et al. History of diabetes mellitus and
subsequent prostate cancer risk in the NIH-AARP Diet and Health Study. Cancer
Causes Control 2007;185:493-503.

140. Giovannucci E, Rimm EB, Stampfer MJ, et al. Diabetes mellitus and risk of
prostate cancer (United States). Cancer Causes Control 1998;91:3-9.

141. Gong Z, Neuhouser ML, Goodman PJ, et al. Obesity, diabetes, and risk of
prostate cancer: results from the prostate cancer prevention trial. Cancer
Epidemiol Biomarkers Prev 2006;1510:1977-83.

142. Velicer CM, Dublin S, White E. Diabetes and the risk of prostate cancer: the role
of diabetes treatment and complications. Prostate Cancer Prostatic Dis
2007;101:46-51.

143. Rodriguez C, Patel AV, Mondul AM, et al. Diabetes and risk of prostate cancer in
a prospective cohort of US men. Am J Epidemiol 2005;1612:147-52.
144. Pierce BL, Plymate S, Ostrander EA, et al. Diabetes mellitus and prostate cancer
risk. Prostate 2008;6810:1126-32.

145. Tavani A, Gallus S, Bosetti C, et al. Diabetes and the risk of prostate cancer. Eur J
Cancer Prev 2002;112:125-8.

146. Poulsen P, Kyvik KO, Vaag A, et al. Heritability of type II (non-insulin-
dependent) diabetes mellitus and abnormal glucose tolerance--a population-based
twin study. Diabetologia 1999;422:139-45.

147. Altshuler D, Hirschhorn JN, Klannemark M, et al. The common PPARgamma
Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes. Nat
Genet 2000;261:76-80.

33
148. Gloyn AL, Weedon MN, Owen KR, et al. Large-scale association studies of
variants in genes encoding the pancreatic beta-cell KATP channel subunits Kir6.2
(KCNJ11) and SUR1 (ABCC8) confirm that the KCNJ11 E23K variant is
associated with type 2 diabetes. Diabetes 2003;522:568-72.

149. Sandhu MS, Weedon MN, Fawcett KA, et al. Common variants in WFS1 confer
risk of type 2 diabetes. Nat Genet 2007;398:951-3.

150. Grant SF, Thorleifsson G, Reynisdottir I, et al. Variant of transcription factor 7-
like 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet 2006;383:320-3.

151. Sladek R, Rocheleau G, Rung J, et al. A genome-wide association study identifies
novel risk loci for type 2 diabetes. Nature 2007;4457130:881-5.

152. Saxena R, Voight BF, Lyssenko V, et al. Genome-wide association analysis
identifies loci for type 2 diabetes and triglyceride levels. Science
2007;3165829:1331-6.

153. Zeggini E, Weedon MN, Lindgren CM, et al. Replication of genome-wide
association signals in UK samples reveals risk loci for type 2 diabetes. Science
2007;3165829:1336-41.

154. Unoki H, Takahashi A, Kawaguchi T, et al. SNPs in KCNQ1 are associated with
susceptibility to type 2 diabetes in East Asian and European populations. Nat
Genet 2008;409:1098-102.

155. Yasuda K, Miyake K, Horikawa Y, et al. Variants in KCNQ1 are associated with
susceptibility to type 2 diabetes mellitus. Nat Genet 2008;409:1092-7.

156. Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al. A variant in CDKAL1
influences insulin response and risk of type 2 diabetes. Nat Genet 2007;396:770-
5.
157. Rung J, Cauchi S, Albrechtsen A, et al. Genetic variant near IRS1 is associated
with type 2 diabetes, insulin resistance and hyperinsulinemia. Nat Genet
2009;4110:1110-5.

158. Tontonoz P, Hu E, Spiegelman BM. Stimulation of adipogenesis in fibroblasts by
PPAR gamma 2, a lipid-activated transcription factor. Cell 1994;797:1147-56.

159. Lehmann JM, Moore LB, Smith-Oliver TA, et al. An antidiabetic
thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated
receptor gamma (PPAR gamma). J Biol Chem 1995;27022:12953-6.

34
160. Duggirala R, Blangero J, Almasy L, et al. Linkage of type 2 diabetes mellitus and
of age at onset to a genetic location on chromosome 10q in Mexican Americans.
Am J Hum Genet 1999;644:1127-40.

161. Reynisdottir I, Thorleifsson G, Benediktsson R, et al. Localization of a
susceptibility gene for type 2 diabetes to chromosome 5q34-q35.2. Am J Hum
Genet 2003;732:323-35.

162. Meigs JB, Manning AK, Fox CS, et al. Genome-wide association with diabetes-
related traits in the Framingham Heart Study. BMC Med Genet 2007;8 Suppl
1:S16.

163. Chen WM, Erdos MR, Jackson AU, et al. Variations in the G6PC2/ABCB11
genomic region are associated with fasting glucose levels. J Clin Invest
2008;1187:2620-8.

164. Prokopenko I, Langenberg C, Florez JC, et al. Variants in MTNR1B influence
fasting glucose levels. Nat Genet 2009;411:77-81.

165. Bouatia-Naji N, Bonnefond A, Cavalcanti-Proenca C, et al. A variant near
MTNR1B is associated with increased fasting plasma glucose levels and type 2
diabetes risk. Nat Genet 2009;411:89-94.

166. Bouatia-Naji N, Rocheleau G, Van Lommel L, et al. A polymorphism within the
G6PC2 gene is associated with fasting plasma glucose levels. Science
2008;3205879:1085-8.

167. Palmer ND, Langefeld CD, Ziegler JT, et al. Candidate loci for insulin sensitivity
and disposition index from a genome-wide association analysis of Hispanic
participants in the Insulin Resistance Atherosclerosis (IRAS) Family Study.
Diabetologia 2009.

168. Rich SS, Goodarzi MO, Palmer ND, et al. A genome-wide association scan for
acute insulin response to glucose in Hispanic-Americans: the Insulin Resistance
Atherosclerosis Family Study (IRAS FS). Diabetologia 2009;527:1326-33.

169. Chambers JC, Elliott P, Zabaneh D, et al. Common genetic variation near MC4R
is associated with waist circumference and insulin resistance. Nat Genet
2008;406:716-8.

170. Pare G, Chasman DI, Parker AN, et al. Novel association of HK1 with glycated
hemoglobin in a non-diabetic population: a genome-wide evaluation of 14,618
participants in the Women's Genome Health Study. PLoS Genet
2008;412:e1000312.
35

171. Lindgren CM, Heid IM, Randall JC, et al. Genome-wide association scan meta-
analysis identifies three Loci influencing adiposity and fat distribution. PLoS
Genet 2009;56:e1000508.

172. Norris JM, Langefeld CD, Talbert ME, et al. Genome-wide association study and
follow-up analysis of adiposity traits in Hispanic Americans: the IRAS Family
Study. Obesity (Silver Spring) 2009;1710:1932-41.

173. Timpson NJ, Lindgren CM, Weedon MN, et al. Adiposity-related heterogeneity in
patterns of type 2 diabetes susceptibility observed in genome-wide association
data. Diabetes 2009;582:505-10.

174. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and
Los Angeles: baseline characteristics. Am J Epidemiol 2000;1514:346-57.

36
Chapter 2: Generalizability of Associations from Prostate Cancer
Genome-Wide Association Studies in Multiple Populations

Published in Cancer Epidemiology Biomarkers Prevention 2009;18(4). April 2009.

Introduction

Genome-wide association studies (GWAS) in men of European ancestry have
revealed multiple variants consistently associated with prostate cancer risk (1-5). Testing
of these risk alleles across populations is an important first step to address the pan-ethnic
nature of their associations, as differences in linkage disequilibrium (LD) and minor
allele frequencies (MAF) may make it difficult to generalize the associations to
populations of non-European descent. We have recently demonstrated the power that
multiethnic genetic studies of common complex diseases possess, having revealed a
number of common variants for prostate cancer at 8q24 that were not identified in larger
comprehensive studies in populations of European ancestry (6). In the present study, we
have evaluated 13 variants considered to be established risk variants for prostate cancer
among men of European ancestry in association with prostate cancer risk in a large
multiethnic case-control study.

37
Methods

Study Population
The Multiethnic Cohort Study (MEC) is a population-based prospective cohort
study that was initiated between 1993 and 1996 and includes subjects from various ethnic
groups – African-Americans and Latinos primarily from California (mainly Los Angeles)
and Native Hawaiians, Japanese-Americans, and European Americans primarily from
Hawaii (7). State driver’s license files were the primary sources used to identify study
subjects in Hawaii and California. Additionally, in Hawaii, state voter’s registration files
were used, and, in California, Health Care Financing Administration (HCFA) files were
used to identify additional African American men.
All participants (n=215,251) returned a 26-page self-administered baseline
questionnaire that obtained general demographic, medical and risk factor information. In
the cohort, incident cancer cases are identified annually through cohort linkage to
population-based cancer Surveillance, Epidemiology, and End Results (SEER) registries
in Hawaii and Los Angeles County as well as to the California State cancer registry.
Information on stage and grade of disease are also obtained through the SEER registries.
Blood sample collection in the MEC began in 1994 and targeted incident prostate
cancer cases and a random sample of study participants to serve as controls for genetic
analyses. This nested prostate cancer case-control study in the MEC consists of 2,768
invasive prostate cancer cases and 2,359 controls. This study was approved by the
38
Institutional Review Boards at the University of Southern California and at the University
of Hawaii and informed consent was obtained from all study participants.

Laboratory Assays
Genotyping for this study was performed using genomic DNA samples, and the
allelic discrimination assay (8). The assay for rs4962416, which was previously reported
as a risk allele near the CTBP2 gene (4), failed in genotyping so it was replaced by
rs12769029 for association testing of this risk allele (pairwise r
2
= 1.0 in HapMap CEU
population (9). We included ~5% duplicate samples to assess genotyping reproducibility.
In total, the concordance was 99.9% among the replication sets. For the 13 variants, the
overall genotyping call rate was 98.7%. Call rates were also similar between cases and
controls for each population (largest difference was 5.5% for rs5945572 in African
Americans). For each variant, we examined Hardy-Weinberg equilibrium (HWE) using a
chi squared test (1 df) among the controls for each ethnic group. Three variants were
nominally statistically significantly (rs2660753, European Americans, p=0.035;
rs9364554, Japanese Americans, p=0.047; and rs2735839, European Americans,
p=0.035). Based on the number of tests we would have expected ~3 of these tests to be
significant by chance alone. Details regarding genotyping efficiency and HWE are
provided in Table 2-1.

Table 2-1: Genotyping Efficiency: Genotype Call Rates, Hardy-Weinberg Equilibrium Testing

Genotype Call Rates Cases/Controls
HWE p value (1 df) Cases/Controls
SNP
Chr.
Gene
Allele
Tested
African Americans

860 ca / 575 co
European
Americans
468 ca / 419 co
Latinos

603 ca / 572 co
Japanese
Americans
725 ca / 684 co
Native Hawaiians
112 ca / 109 co
rs721048

2p15
EHBP1
A

99.2%/99.5%
0.56/0.20
99.6%/99.0%
0.21/0.16
97.0%/96.7%
0.49/0.32
98.5%/98.2%
0.92/0.31
99.1%/100%
0.51/0.16
rs2660753

3p12.1

T

95.9%/99.8%
0.012/0.17
99.8%/99.8%
0.18/0.035
99.5%/98.6%
0.23/0.92
98.6%/99.7%
0.61/0.16
100%/100%
0.97/0.33
rs9364554

6q25.3
SLC22A3
T

95.9%/99.7%
0.56/0.24
99.6%/99.0%
0.18/0.41
99.3%/98.3%
0.33/0.13
98.3%/98.8%
0.67/0.047
100%/99.1%
0.70/0.85
rs10486567

7p15.2
JAZF1
G

95.3%/99.5%
0.94/0.26
99.1%/99.5%
0.87/0.23
98.3%/97.9%
0.23/0.35
98.8%/99.4%
0.92/0.51
100%/99.1%
0.96/0.76
rs6465657

7q21.3
LMTK2
C

99.5%/99.3%
0.90/0.20
99.4%/99.3%
0.96/0.75
98.8%/98.4%
0.55/0.072
98.8%/99.6%
0.44/0.29
100%/100%
0.34/0.34
rs10993994

10q11.23
MSMB
T

96.0%/99.1%
0.069/0.57
98.9%/99.0%
0.54/0.052
98.0%/98.1%
0.44/0.16
99.0%/98.2%
0.30/0.44
100%/100%
0.25/0.66
rs12769019

10q26.13
CTBP2
G

99.5%/99.5%
0.13/0.14
99.6%/98.8%
0.60/0.19
99.3%/98.4%
0.27/0.84
99.4%/100%
1.6x10
-6
/0.79
100%/100%
0.48/0.44
rs7931342

11q13.2

G

98.4%/99.8%
0.94/0.89
99.1%/99.0%
0.97/0.26
98.8%/98.4%
0.81/0.59
98.6%/99.4%
0.68/0.86
99.1%/99.1%
0.27/0.43
rs11649743

17q12
HNF1B
G

98.4%/99.7%
0.75/0.38
100%/99.0%
0.16/0.37
99.0%/98.1%
0.16/0.84
99.3%/98.8%
0.34/0.45
99.1%/100%
0.21/0.94
rs4430796

17q12
HNF1B
A

98.3%/99.1%
0.52/0.41
97.9%/98.3%
0.35/0.73
99.3%/98.1%
0.32/0.88
99.0%/98.7%
0.069/0.63
99.1%/97.2%
0.12/0.87
rs1859962

17q24.3

G

96.7%/98.8%
0.29/0.14
98.7%/99.3%
0.043/0.69
99.0%/98.6%
0.57/0.60
98.5%/99.3%
0.32/0.47
99.1%/100%
0.37/0.37
rs2735839

19q13
KLK2/3
G

95.7%/99.1%
5.4x10
-5
/0.86
99.6%/99.5%
0.35/0.035
98.7%/98.3%
0.54/0.64
99.0%/99.4%
0.42/0.90
99.1%/100%
0.30/0.92
rs5945572

Xp11.22
NUDT10/11
A

93.8%/99.3%
NA
99.4%/99.8%
NA
98.0%/98.3%
NA
99.2%/99.4%
NA
100%/100%
NA

39
40
Statistical Analysis
Odds ratios (OR) and 95% confidence intervals (CI) of the effect of each variant
on prostate cancer risk were computed using logistic regression in ethnic-specific and
ethnic-pooled analyses (SAS version 9.1, SAS Institute Inc., Cary, North Carolina). We
estimated the log-additive effect of each risk allele as well as the OR for heterozygotes
and homozygotes separately. All estimates are adjusted for age (quintiles) and race (in
pooled analysis). In the admixed populations (African Americans, Latinos, and Native
Hawaiians), we adjusted for their global proportion of European ancestry as previously
described (6). First degree family history of prostate cancer (father or full brother) was
also examined as a potential confounding variable, but was not included in the model
because it had no effect (<2% change) on the pooled risk associations. We tested for
allelic heterogeneity of effects by including an interaction term between variant and
racial/ethnic group in the regression model (4 df test). We also tested for gene x gene
interaction by including an interaction term for every combination of two risk alleles.
We also examined genetic associations with prostate cancer risk among disease
subgroups using the standard case-control approach, limiting the cases to those with a
specific phenotype (‘advanced disease’) and all controls, and a case-only analysis to test
for differences by disease subgroup. We defined the cancer as ‘advanced’ if high stage
(regional by direct extension, regional by lymph nodes, regional by both direct extension
and lymph nodes, regional NOS, or distant metastases/systemic disease), and/or high
grade (low level of cell differentiation; Gleason score > 7). Non-advanced disease was
defined as having both a localized stage and low grade (Gleason Score ≤ 7). We were
41
unable to define cases as either advanced or localized if both stage and grade data were
missing or if either the stage was localized or grade was low (Gleason Score ≤ 7), and
information for the other variable was missing (n=207).
We also estimated the per allele effects for the sum of 28 established risk alleles
(13 risk variants from this study, 8 established risk variants from MEC data as part of the
PRACTICAL consortium (10), and 7 independent risk variants from a previous MEC
study of the 8q24 region (6). This analysis was performed both with the seven 8q24
SNPs separate from the 21 other risk variants and with all 28 alleles summed together.
This analysis included 2,214 cases and 2,021 controls that were included in all 3 studies.

Results and Discussion

Cases in this study ranged in age of entry into the cohort from 44 to 78 with a
mean of 64.3 (age at diagnosis ranged from 46 to 87). Controls ranged in age from 45 to
77 with a mean of 62.5. The Japanese-Americans were slightly older (mean age of entry,
64.6 years) while the Native Hawaiians were slightly younger than the other three groups
(mean age of entry, 62.7 years).
Six of the variants were nominally statistically significant (p<0.05) in pooled
analyses (JAZF1, rs10486567, OR= 1.23; (95% CI, 1.12-1.35); Xp11.2, rs5945572,
1.31(1.13-1.51); HNF1B, rs4430796, 1.15(1.06-1.25); MSMB, rs10993994, 1.13(1.04-
1.23); 11q13.2, rs7931342, 1.13(1.03-1.23), and 3p12.1, rs2660753, 1.11(1.01-1.21);
Table 2-2). These associations were similar in magnitude (RR>1.10) and in the same

Table 2-2: Frequencies of Risk Alleles and Associations with Prostate Cancer Risk in the MEC

OR(95% CI)
a

Risk Allele Frequency
SNP
Risk Allele
Chr
Gene
African
Americans
860 ca / 575 co
European
Americans
468 ca / 419 co
Latinos

603 ca / 572 co
Japanese
Americans
725 ca / 684 co
Native
Hawaiians
112 ca / 109 co
Pooled
2768 ca
/2,359 co P value P
het
b
rs721048
A
2p15
EHBP1
0.86(0.59-1.26)
0.05
0.87(0.67-1.12)
0.19
1.49(1.19-1.87)
0.14
1.05(0.71-1.56)
0.04
0.52(0.24-1.12)
0.09
1.08(0.94-1.25)
0.09
0.26

3.9x10
-3

rs2660753
T
3p12.1

0.97(0.83-1.14)
0.46
1.06(0.81-1.39)
0.13
1.15(0.94-1.40)
0.20
1.30(1.09-1.55)
0.24
0.94(0.56-1.57)
0.18
1.11(1.01-1.21)
0.26
0.034

0.16

rs9364554
T
6q25.3
SLC22A3
1.10(0.82-1.48)
0.07
1.06(0.86-1.30)
0.27
1.15(0.95-1.39)
0.21
1.09(0.93-1.29)
0.34
1.07(0.69-1.68)
0.22
1.10(1.00-1.21)
0.22
0.062

0.99

rs10486567
G
7p15.2
JAZF1
1.18(1.00-1.40)
0.70
1.50(1.19-1.89)
0.74
1.19(1.00-1.40)
0.53
1.14(0.88-1.48)
0.09
1.25(0.83-1.89)
0.36
1.23(1.12-1.35)
0.47
2.1x10
-5

0.48

rs6465657
C
7q21.3
LMTK2
0.91(0.72-1.14)
0.85
1.08(0.89-1.31)
0.45
0.94(0.78-1.12)
0.70
1.04(0.82-1.33)
0.90
0.98(0.65-1.49)
0.67
0.99(0.89-1.09)
0.75
0.80

0.77

rs10993994
T
10q11.23
MSMB
1.05(0.90-1.24)
0.59
1.15(0.96-1.39)
0.42
1.06(0.90-1.25)
0.37
1.26(1.08-1.46)
0.45
1.10(0.75-1.61)
0.64
1.13(1.04-1.23)
0.47
3.1x10
-3

0.52

rs12769019
G
10q26.13
CTBP2
1.20(0.98-1.47)
0.16
1.12(0.90-1.39)
0.26
1.00(0.83-1.21)
0.24
1.43(0.75-2.76)
0.01
1.42 (0.68-2.95)
0.07
1.11(0.99-1.25)
0.15
0.062

0.60

rs7931342
G
11q13.2

1.12(0.93-1.35)
0.76
1.28(1.05-1.55)
0.51
1.27(1.07-1.51)
0.37
0.87(0.73-1.05)
0.23
1.19(0.79-1.80)
0.48
1.13(1.03-1.23)
0.45
8.4x10
-3

0.023

rs11649743
G
17q12
HNF1B
1.04(0.79-1.38)
0.91
1.05(0.82-1.35)
0.82
1.29(1.04-1.61)
0.82
1.08(0.91-1.27)
0.70
0.96(0.65-1.41)
0.62
1.10(0.99-1.22)
0.80
0.067

0.58

rs4430796
A
17q12
HNF1B
0.99(0.84-1.16)
0.35
1.44(1.18-1.74)
0.48
1.26(1.07-1.50)
0.57
1.04(0.89-1.22)
0.64
1.23(0.79-1.90)
0.70
1.15(1.06-1.25)
0.53
9.1x10
-4

0.026

rs1859962
G
17q24.3

1.01(0.86-1.19)
0.32
1.00(0.83-1.20)
0.51
1.10(0.93-1.30)
0.60
1.06(0.89-1.25)
0.26
1.03(0.69-1.52)
0.56
1.04(0.96-1.13)
0.42
0.35

0.95

rs2735839
G
19q13
KLK2/3
0.80(0.67-0.95)
0.71
1.33(1.02-1.75)
0.84
1.15(0.94-1.40)
0.77
1.21(1.03-1.41)
0.58
0.91(0.61-1.35)
0.51
1.06(0.97-1.16)
0.70
0.20

2.0x10
-3

rs5945572
A
Xp11.22
NUDT10/11
1.34(1.05-1.71)
0.26
1.25(0.95-1.66)
0.35
1.32(0.98-1.77)
0.17
1.25(0.86-1.82)
0.08
1.65(0.61-4.46)
0.06
1.31(1.13-1.51)
0.19
2.6x10
-4
0.98

a
ORs adjusted for age (quintiles), genome-wide European ancestry (African Americans, Latinos and Native Hawaiians) and age-ethnicity strata
(pooled analysis).
b
P
het
= p value for heterogeneity of allelic effects across ethnic groups (4 df test).
42
43
direction as reported in previous GWAS among men of European ancestry (1-5). The
associations for each genotype class are provided in Tables 2-3. These 6 risk variants
were common in all populations with frequencies ranging from 0.06-0.76, and
frequencies ≥0.19 in the combined sample. For two variants we detected significant
heterogeneity of the effect across populations (HNF1B, rs4430796, p
het
= 0.026; 11q3.2,
rs7931342, p
het
= 0.023). For these variants, positive associations were noted in all
populations except African Americans and Japanese, respectively, the two largest groups,
suggesting that these variants are poorly linked to the causal alleles in these populations.
Non-significant positive associations were also observed in the expected direction
for 6 other variants (SLC22A3, rs9364554, 1.10(1.00-1.21); CTBP2, rs12769019,
1.11(0.99-1.25); HNF1B, rs11649743, 1.10(0.99-1.22); EHBP1, rs721048, 1.08(0.94-
1.25); KLK2/3, rs2735839, 1.06(0.97-1.16); and 17q24.3, rs1859962, 1.04(0.96-1.13))
and for most of these variants, positive associations were observed consistently across
population (Table 2-2). Two of these variants had frequencies <0.20 in the combined
sample with ethnic-specific frequencies <0.05 in some populations. We noted significant
ethnic heterogeneity in the associations for EHBP1 (rs721048, p
het
= 3.9 x10
-3
) and
KLK2/3 (rs2735839, p
het
= 2.0x10
-3
), (Table 2-2) and no evidence of an association with
variant rs6465657 in LMTK2, (OR=0.99; 95% CI: 0.89-1.09). Interestingly, the KLK2/3
variant was inversely associated with risk in African Americans.

Table 2-3: Association With Prostate Cancer Risk by Genotype for Each Racial/Ethnic Group
African Americans European Americans Latinos Japanese Americans Native Hawaiians Ethnically Pooled
SNP
Risk Allele Het Hom Het Hom Het Hom Het Hom Het Hom Het Hom
rs721048
A
0.81
(.55-1.20) NA
0.87
(0.65-1.16)
0.77
(0.30-1.94)
1.64
(1.25-2.14)
1.54
(0.76-3.11)
1.02
(0.68-1.52) NA
0.64
(0.28-1.48) NA
1.09
(0.93-1.28)
1.13
(0.66-1.92)
rs2660753
T
1.28
(0.99-1.65)
0.90
(0.65-1.23)
1.11
(0.80-1.55)
0.95
(0.41-2.16)
1.09
(0.85-1.40)
1.51
(0.87-2.61)
1.17
(0.93-1.46)
2.11
(1.34-3.33)
0.85
(0.47-1.53)
1.49
(0.24-9.39)
1.16
(1.03-1.32)
1.14
(0.93-1.42)
rs9364554
T
1.13
(0.81-1.57)
1.00
(0.30-3.40)
1.03
(0.78-1.38)
1.14
(0.69-1.89)
1.18
(0.92-1.52)
1.24
(0.75-2.06)
1.00
(0.80-1.26)
1.31
(0.91-1.87)
1.01
(0.57-1.78)
1.34
(0.40-4.50)
1.07
(0.94-1.22)
1.26
(0.99-1.60)
rs10486567
G
1.31
(0.87-1.98)
1.48
(0.99-2.21)
1.09
(0.57-2.10)
1.79
(0.94-3.39)
0.99
(0.73-1.34)
1.37
(0.98-1.92)
1.11
(0.84-1.47)
1.75
(0.50-6.12)
1.20
(0.66-2.18)
1.60
(0.67-3.81)
1.11
(0.93-1.31)
1.48
(1.22-1.79)
rs6465657
C
1.12
(0.54-2.32)
0.97
(0.47-2.01)
1.09
(0.80-1.49)
1.16
(0.79-1.71)
1.27
(0.85-1.91)
1.03
(0.69-1.54)
1.24
(0.48-3.18)
1.26
(0.50-3.15)
0.56
(0.23-1.35)
0.74
(0.29-1.86)
1.11
(0.89-1.38)
1.04
(0.83-1.30)
rs10993994
T
1.19
(0.86-1.64)
1.16
(0.83-1.62)
1.33
(0.98-1.81)
1.28
(0.88-1.87)
1.11
(0.86-1.43)
1.10
(0.77-1.56)
1.10
(0.86-1.42)
1.60
(1.18-2.17)
0.85
(0.37-1.93)
1.08
(0.48-2.47)
1.16
(1.01-1.33)
1.28
(1.08-1.50)
rs12769019
G
1.22
(0.95-1.57)
1.36
(0.75-2.46)
0.99
(0.74-1.31)
1.57
(0.90-2.75)
0.93
(0.73-1.19)
1.16
(0.70-1.91)
1.16
(0.56-2.40) NA
1.22
(0.56-2.66) NA
1.05
(0.91-1.22)
1.42
(1.04-1.94)
rs7931342
G
1.24
(0.74-2.06)
1.34
(0.81-2.21)
1.48
(1.05-2.10)
1.66
(1.13-2.44)
1.31
(1.01-1.70)
1.60
(1.12-2.27)
0.86
(0.68-1.08)
0.79
(0.48-1.30)
1.25
(0.63-2.50)
1.42
(0.63-3.23)
1.12
(0.97-1.29)
1.27
(1.06-1.52)
rs11649743
G
0.76
(0.17-3.37)
0.81
(0.19-3.54)
0.61
(0.27-1.34)
0.72
(0.33-1.57)
0.96
(0.48-1.92)
1.32
(0.67-2.59)
1.41
(0.94-2.11)
1.36
(0.91-2.03)
0.73
(0.33-1.62)
0.83
(0.37-1.87)
1.03
(0.78-1.38)
1.16
(0.88-1.54)
rs4430796
A
1.02
(0.80-1.28)
0.96
(0.68-1.37)
1.30
(0.92-1.83)
2.05
(1.39-3.02)
1.42
(1.02-1.99)
1.66
(1.17-2.36)
1.32
(0.93-1.86)
1.21
(0.86-1.72)
0.72
(0.27-1.92)
1.13
(0.42-3.06)
1.18
(1.02-1.37)
1.33
(1.13-1.57)
rs1859962
G
1.06
(0.84-1.34)
0.98
(0.68-1.41)
0.76
(0.55-1.07)
0.98
(0.67-1.43)
0.95
(0.67-1.34)
1.15
(0.80-1.64)
1.04
(0.83-1.31)
1.14
(0.75-1.72)
1.51
(0.73-3.14)
1.15
(0.52-2.56)
0.99
(0.87-1.13)
1.11
(0.93-1.30)
rs2735839
G
1.31
(0.87-1.98)
0.86
(0.57-1.30)
3.51
(1.22-10.10)
3.95
(1.41-11.11)
1.11
(0.62-1.98)
1.28
(0.73-2.26)
1.06
(0.78-1.45)
1.39
(1.01-1.92)
0.74
(0.38-1.46)
0.82
(0.37-1.83)
1.17
(0.95-1.43)
1.19
(0.96-1.46)
rs5945572
A
NA
(X chrom.)
1.34
(1.05-1.71)
1.25
(0.95-1.66)
1.32
(0.98-1.77)
1.25
(0.86-1.82)
1.65
(0.61-4.46)
1.31
(1.13-1.51)
Het = heterozygous for risk allele; Hom = homozygous for risk allele; NA = sample size too small to test
a
ORs adjusted for age (quintiles), genome-wide European ancestry (African Americans, Latinos and Native Hawaiians).
44
45
The original GWAS found effect sizes of 1.10-1.25 per allele with frequencies of
the risk alleles ranging from 0.10-0.85 (1-5). A replication study of seven of these alleles
by the PRACTICAL Consortium found per allele effect sizes
ranging from 1.08 to 1.30 (11). In our study, a lack of power due to smaller sample size
and/or low MAFs in some populations (and thus in the combined sample) was likely to
contribute to some of the variants not reaching statistical significance (Table 2-4). We
had relatively limited power (50-65%) to detect statistically significant pooled effects of
1.10-1.12 for variants with frequencies as low as 0.20. Power was improved (≥ 81%) for
effects ≥1.20 and risk alleles with frequencies ≥0.10 in the combined sample.
We detected 6 significant gene x gene interactions; however, it is difficult to
determine whether any of these are true effects, since this analysis included 78 tests and
we would have expected ~4 significant interactions by chance alone. The most
significant interaction (p=9.0x10
-3
) was between rs4430796 (HNF1B) and rs1859962
(17q24) which are both located on the same chromosome, albeit in distant areas.
We also examined allelic associations by disease subgroup (advanced vs. non-
advanced; Tables 2-5, 2-6, and 2-7) and tested for differences in case-case analyses.
None of the differences in prostate cancer risk between advanced and non-advanced
subgroups were statistically significant.

Table 2-4: Estimated Power for Detecting Association Given Reported ORs (α=0.05)

African
Americans
(860ca/575co)
European
Americans
(468ca/419co)
Latinos
(603ca/572co)
Japanese
Americans
(725ca/684co)
Native
Hawaiians
(112ca/109co)
Pooled
(2768ca/2359co)
SNP
Reported
ORs
a
RAF
Est
Power RAF
Est
Power RAF
Est
Power RAF
Est
Power RAF
Est
Power RAF
Est
Power
rs721048 1.15 0.05 13% 0.19 22% 0.14 23% 0.04 12% 0.09 7% 0.09 55%
rs2660753 1.18 0.46 58% 0.13 23% 0.20 38% 0.24 48% 0.18 11% 0.26 96%
rs9364554 1.17 0.07 19% 0.27 32% 0.21 35% 0.34 51% 0.22 11% 0.22 92%
rs10486567 1.30 0.70 87% 0.74 65% 0.53 88% 0.09 55% 0.36 27% 0.47 99%
rs6465657 1.12 0.85 18% 0.45 22% 0.70 24% 0.90 14% 0.67 9% 0.75 68%
rs10993994 1.25 0.59 81% 0.42 64% 0.37 75% 0.45 84% 0.64 20% 0.47 99%
rs12769019 1.17 0.16 34% 0.26 31% 0.24 38% 0.01 7% 0.07 7% 0.15 83%
rs7931342 1.19 0.76 48% 0.51 45% 0.37 54% 0.23 51% 0.48 15% 0.45 99%
rs11649743 1.28 0.91 42% 0.82 48% 0.82 60% 0.70 83% 0.62 24% 0.80 99%
rs4430796 1.22 0.35 71% 0.48 55% 0.57 66% 0.64 70% 0.70 16% 0.53 99%
rs1859962 1.20 0.32 62% 0.51 48% 0.60 57% 0.26 58% 0.56 16% 0.42 99%
rs2735839 1.20 0.71 60% 0.84 28% 0.77 44% 0.58 66% 0.51 16% 0.70 99%
rs5945572 1.19 0.26 30% 0.35 24% 0.17 21% 0.08 15% 0.06 6% 0.19 70%
a
As previously reported (1-4, 10, 12-14).
46

Table 2-5: Frequency of Risk Allele and Association with Prostate Cancer by Disease Subgroup and Case-Only Testing

OR(95% CI)
a

Risk Allele Frequency
OR(95% CI)
a

Risk Allele Frequency
SNP
Chr.
Gene
Allele
Tested
Advanced Cases
961 ca / 2359 co
Non-Advanced Cases
1600 ca / 2359 co P
het
b
rs721048

2p15
EHBP1
A

1.00(0.82-1.21)
0.09
1.11(0.94-1.30)
0.09
0.33

rs2660753

3p12.1

T

1.15(1.01-1.30)
0.26
1.06(0.96-1.19)
0.26
0.26

rs9364554

6q25.3
SLC22A3
T

1.06(0.93-1.21)
0.22
1.12(0.99-1.25)
0.22
0.53

rs10486567

7p15.2
JAZF1
G

1.13(0.99-1.29)
0.47
1.31(1.17-1.46)
0.47
0.088

rs6465657

7q21.3
LMTK2
C

0.95(0.83-1.08)
0.75
1.04(0.93-1.17)
0.75
0.17

rs10993994

10q11.23
MSMB
T

1.07(0.96-1.20)
0.47
1.16(1.06-1.27)
0.47
0.19

rs12769019

10q26.13
CTBP2
G

1.16(1.00-1.36)
0.15
1.08(0.94-1.23)
0.15
0.36

rs7931342

11q13.2

G

1.17(1.04-1.32)
0.45
1.10(1.00-1.22)
0.45
0.37

rs11649743

17q12
HNF1B
G

1.14(0.99-1.32)
0.80
1.10(0.98-1.25)
0.80
0.69

rs4430796

17q12
HNF1B
A

1.16(1.03-1.30)
0.53
1.15(1.05-1.27)
0.53
0.88

rs1859962

17q24.3

G

1.00(0.89-1.12)
0.42
1.07(0.97-1.18)
0.42
0.27

rs2735839

19q13
KLK2/3
G

1.03(0.91-1.17)
0.70
1.09(0.98-1.21)
0.70
0.33

rs5945572

Xp11.22
NUDT10/11
A

1.27(1.04-1.55)
0.19
1.31(1.11-1.54)
0.19
0.84

a
ORs adjusted for age(quintiles)-ethnicity strata, and genome-wide European ancestry.
(African Americans, Latinos and Native Hawaiians).
b
P
het
= p value for heterogeneity (advanced vs. non-advanced).
47

Table 2-6: Frequencies of Risk Alleles and Associations with Advanced Prostate Cancer in the MEC

OR(95% CI)
a

Risk Allele Frequency
SNP
African Americans

241 ca / 575 co
European
Americans
152 ca / 419 co
Latinos

217 ca / 572 co
Japanese
Americans
306 ca / 684 co
Native
Hawaiians
45 ca / 109 co
Pooled

961 ca / 2359 co P value P
het
b
rs721048

1.16(0.69-1.93)
0.05
0.62(0.42-0.92)
0.19
1.41(1.05-1.89)
0.14
0.87(0.51-1.50)
0.04
0.55(0.20-1.57)
0.09
1.00(0.82-1.21)
0.09
0.99

0.013

rs2660753

1.01(0.81-1.27)
0.46
0.98(0.67-1.44)
0.13
1.31(1.01-1.70)
0.20
1.24(0.99-1.55)
0.24
1.16(0.60-2.25)
0.18
1.15(1.01-1.30)
0.26
0.034

0.53

rs9364554

1.17(0.78-1.76)
0.07
0.96(0.71-1.29)
0.27
1.17(0.90-1.51)
0.21
1.05(0.86-1.30)
0.34
0.87(0.46-1.62)
0.22
1.06(0.93-1.21)
0.22
0.36

0.81

rs10486567

1.01(0.80-1.29)
0.70
1.59(1.13-2.24)
0.74
1.03(0.82-1.30)
0.53
1.21(0.88-1.67)
0.09
1.12(0.61-2.03)
0.36
1.13(0.99-1.29)
0.47
0.066

0.25

rs6465657

0.94(0.68-1.31)
0.85
1.21(0.92-1.58)
0.45
0.87(0.68-1.10)
0.70
0.87(0.65-1.17)
0.90
0.78(0.45-1.37)
0.67
0.95(0.83-1.08)
0.75
0.43

0.36

rs10993994

1.04(0.82-1.30)
0.59
1.14(0.88-1.47)
0.42
1.05(0.83-1.32)
0.37
1.10(0.90-1.33)
0.45
1.01(0.59-1.71)
0.64
1.07(0.96-1.20)
0.47
0.21

0.98

rs12769019

1.18(0.89-1.56)
0.16
1.05(0.77-1.43)
0.26
1.19(0.92-1.53)
0.24
1.38(0.61-3.14)
0.01
1.59(0.62-4.09)
0.07
1.16(1.00-1.36)
0.15
0.056

0.90

rs7931342

1.24(0.94-1.63)
0.76
1.64(1.25-2.15)
0.51
1.22(0.97-1.53)
0.37
0.86(0.68-1.09)
0.23
1.03(0.60-1.78)
0.48
1.17(1.04-1.32)
0.45
9.0x10
-3

0.013

rs11649743

1.05(0.70-1.58)
0.91
0.92(0.65-1.31)
0.82
1.39(1.01-1.91)
0.82
1.16(0.94-1.45)
0.70
1.04(0.61-1.76)
0.62
1.14(0.99-1.32)
0.80
0.072

0.52

rs4430796

0.97(0.77-1.23)
0.35
1.41(1.07-1.85)
0.48
1.29(1.02-1.62)
0.57
1.09(0.89-1.34)
0.64
1.15(0.65-2.02)
0.70
1.16(1.03-1.30)
0.53
0.011

0.26

rs1859962

0.96(0.75-1.22)
0.32
1.01(0.77-1.33)
0.51
0.98(0.78-1.23)
0.60
1.02(0.82-1.26)
0.26
1.10(0.65-1.87)
0.56
1.00(0.89-1.12)
0.42
0.94

0.99

rs2735839

0.78(0.61-0.99)
0.71
1.09(0.76-1.56)
0.84
1.19(0.91-1.57)
0.77
1.14(0.94-1.39)
0.58
0.99(0.57-1.71)
0.51
1.03(0.91-1.17)
0.70
0.60

0.12

rs5945572

1.51(1.08-2.12)
0.26
1.24(0.84-1.83)
0.35
1.15(0.77-1.73)
0.17
1.13(0.69-1.84)
0.08
0.69(0.13-3.58)
0.06
1.27(1.04-1.55)
0.19
0.017

0.74

a
ORs adjusted for age (quintiles), genome-wide European ancestry (African Americans, Latinos and Native Hawaiians) and age-ethnicity strata
(pooled analysis).
b
P
het
= p value for heterogeneity of allelic effects across ethnic groups (4 df test).
48

Table 2-7: Frequencies of Risk Alleles and Associations with Non-Advanced Prostate Cancer in the MEC

OR(95% CI)
a

Risk Allele Frequency
SNP
African Americans
545 ca / 575 co
European
Americans
288 ca / 419 co
Latinos
342 ca / 572 co
Japanese
Americans
363 ca / 684 co
Native
Hawaiians
62 ca / 109 co
Pooled
1600 ca/ 2359 co P value P
het
b
rs721048

0.76(0.49-1.17)
0.05
1.02(0.76-1.36)
0.19
1.50(1.16-1.95)
0.14
1.02(0.63-1.64)
0.04
0.44(0.17-1.14)
0.09
1.11(0.94-1.30)
0.09
0.23

0.016

rs2660753

0.94(0.79-1.12)
0.46
1.13(0.84-1.54)
0.13
1.06(0.84-1.35)
0.20
1.29(1.04-1.59)
0.24
0.78(0.41-1.47)
0.18
1.06(0.96-1.19)
0.26
0.26

0.19

rs9364554

1.04(0.74-1.47)
0.07
1.10(0.87-1.40)
0.27
1.15(0.92-1.44)
0.21
1.12(0.92-1.37)
0.34
1.14(0.67-1.92)
0.22
1.12(0.99-1.25)
0.22
0.066

0.99

rs10486567

1.28(1.05-1.54)
0.70
1.56(1.19-2.05)
0.74
1.29(1.06-1.58)
0.53
1.18(0.86-1.62)
0.09
1.20(0.74-1.95)
0.36
1.31(1.17-1.46)
0.47
2.2x10
-6

0.70

rs6465657

0.92(0.71-1.18)
0.85
1.06(0.86-1.32)
0.45
1.05(0.85-1.29)
0.70
1.17(0.86-1.60)
0.90
1.11(0.67-1.82)
0.67
1.04(0.93-1.17)
0.75
0.49

0.80

rs10993994

1.04(0.87-1.24)
0.59
1.13(0.91-1.40)
0.42
1.08(0.89-1.31)
0.37
1.42(1.18-1.71)
0.45
1.16(0.73-1.85)
0.64
1.16(1.06-1.27)
0.47
2.1x10
-3

0.17

rs12769019

1.24(0.99-1.54)
0.16
1.14(0.89-1.46)
0.26
0.86(0.69-1.09)
0.24
1.19(0.51-2.80)
0.01
1.31 (0.55-3.13)
0.07
1.08(0.94-1.23)
0.15
0.28

0.24

rs7931342

1.05(0.85-1.29)
0.76
1.08(0.87-1.34)
0.51
1.35(1.10-1.64)
0.37
0.91(0.73-1.14)
0.23
1.22(0.74-2.01)
0.48
1.10(1.00-1.22)
0.45
0.063

0.13

rs11649743

1.18(0.86-1.62)
0.91
1.04(0.78-1.37)
0.82
1.23(0.95-1.59)
0.82
1.10(0.90-1.34)
0.70
0.83(0.52-1.33)
0.62
1.10(0.98-1.25)
0.80
0.11

0.66

rs4430796

1.02(0.86-1.22)
0.35
1.46(1.17-1.82)
0.48
1.30(1.06-1.58)
0.57
0.97(0.80-1.17)
0.64
1.30(0.75-2.27)
0.70
1.15(1.05-1.27)
0.53
4.1x10
-3

0.026

rs1859962

1.02(0.85-1.22)
0.32
0.97 (0.79-1.21)
0.51
1.23(1.00-1.50)
0.60
1.11(0.90-1.36)
0.26
1.00(0.62-1.59)
0.56
1.07(0.97-1.18)
0.42
0.16

0.57

rs2735839

0.81(0.67-0.98)
0.71
1.48(1.08-2.04)
0.84
1.23(0.96-1.57)
0.77
1.26(1.04-1.52)
0.58
0.79(0.49-1.29)
0.51
1.09(0.98-1.21)
0.70
0.13

1.4x10
-3

rs5945572

1.20(0.92-1.58)
0.26
1.23(0.89-1.68)
0.35
1.50(1.07-2.10)
0.19
1.35(0.87-2.11)
0.08
2.10(0.71-6.17)
0.06
1.31(1.11-1.54)
0.19
1.2x10
-3

0.75

a
ORs adjusted for age (quintiles), genome-wide European ancestry (African Americans, Latinos and Native Hawaiians) and age-ethnicity strata for
pooled analysis.
b
P
het
= p value for heterogeneity of allelic effects across ethnic groups (4 df test).
49
50
The per allele effects of the total risk alleles were significantly associated with
prostate cancer in all racial/ethnic groups for both the sum count of the 7 8q24 SNPs from
Haiman et al. (6) as well as the for the sum count of the combination of the 13 established
prostate cancer risk variants from this study and 8 from Eeles et al (10) (Table 2-8).
There was significant heterogeneity of association between racial-ethnic groups
(p=0.029) among the 21 non 8q24 SNPs with African Americans having the weakest per
allele association (OR=1.05) and Native Hawaiians having the strongest (1.24) per allele
association. The mean number of risk alleles ranged from 2.7 (European Americans and
Native Hawaiians) to 5.5 (African Americans; p<0.0001 vs. each ethnic group) for the
8q24 SNPs and from 16.8 (Japanese Americans) to 20.9 (African Americans) for the 21
non-8q24 SNPs. The high number of 8q24 risk alleles among African Americans
compared to the other racial/ethnic groups (and European Americans in particular)
corresponds with the previous findings of an admixture scan which reported that this
region was overrepresented by DNA of African ancestry among cases (15).
While the majority of the risk variants examined in this study were positively
associated with risk in the pooled analysis, the lack of consistent effects in all populations
for some markers suggests that, for these associations, the underlying causal variant may
not be of appreciable frequency in all populations and/or differences in LD may be
obscuring effects in some populations. One example where this is likely to be case is
rs4430796 (HNF1B) where we noted no evidence of an association in African Americans
(OR = 0.99, 95% CI 0.84-1.16). The ethnic heterogeneity observed for some markers
may also be due to interactions with other genetic risk factors and environmental

Table 2-8: Association of Risk Allele Counts with Prostate Cancer (n=4,235)
European
Americans
446 cases
398 controls
African
Americans
721 cases
954 controls
Latinos
590 cases
562 controls
Japanese
Americans
443 cases
445 controls
Native
Hawaiians
69 cases
62 controls
Pooled
2,214 cases
2,021 controls
7 Independent Risk Alleles for 8q24 genotyped in Haiman et al. 2007 (6)
Avg. # of risk alleles
(range)

2.7
(0-6)
5.5
a
(0-11)
3.1
(0-7)
3.2
(0-8)
2.7
(0-7)
3.7
(0-11)
Per allele OR
b
P value

1.17(1.07-1.29)
1.9x10
-3
1.20(1.12-1.28)
3.4x10
-8

1.23(1.13-1.34)
1.6x10
-6

1.30(1.20-1.40)
3.9x10
-11
1.50(1.22-1.85)
1.3x10
-4
1.24(1.19-1.28)
2.2x10
-27
P
het
c
0.14
21 Established Risk Alleles 13 genotyped for Waters et al. 2009 and 8 for Eeles et al. 2010 (10)
Avg. # of risk alleles
(range)
19.5
(11-27)
20.9
(14-29)
19.8
(11-30)
16.8
(9-25)
18.0
(11-26)
19.3
(9-30)
Per allele OR
b
P value

1.15(1.09-1.20)
1.3x10
-8

1.05(1.01-1.10)
0.023
1.14(1.09-1.18)
1.0x10
-9
1.10(1.05-1.16)
1.8x10
-4
1.24(1.08-1.42)
2.5x10
-3
1.11(1.09-1.14)
3.4x10
-21
P
het
c
0.029
All 28 risk alleles combined
Avg. # of risk alleles
(range)
22.2
(12-32)
26.4
(17-36)
22.9
(14-34)
20.1
(10-29)
20.7
(12-27)
23.1
(10-36)
Per allele OR
b
P value

1.15(1.10-1.20)
2.0x10
-10
1.10(1.06-1.14)
4.3x10
-7
1.15(1.11-1.20)
8.5x10
-14

1.16(1.11-1.21)
1.9x10
-11
1.33(1.17-1.51)
1.5x10
-5

1.14(1.12-1.17)
5.0x10
-41
P
het
c
0.039
a
P<0.0001 for mean value compared to each ethnic group (t-test).
b
ORs were adjusted for age (quintiles), genome-wide European ancestry (African Americans, Latinos, and Native Hawaiians) and age-ethnicity strata
(pooled analysis).
c
P
het
= P value for heterogeneity of allelic effects across ethnic groups (4 df test).
51
52
exposures that vary in frequency across populations, which we plan to explore in future
analyses.
Our previous studies on 8q24 and prostate cancer provide strong support for the
hypothesis that the higher incidence of prostate cancer in African American men,
compared to men in other racial and ethnic populations, is due to common risk variants
that are more common in men of African descent (6). It is interesting to note that only
three of the variants examined in this study were more common in African Americans
than in the other racial/ethnic groups (Table 2-2). Fine-mapping of these candidate
prostate cancer risk loci in a multiethnic sample will be important in order to identify the
strongest markers of risk in each population and hopefully will help us to better
understand the excess risk of prostate cancer in African Americans.
Among the alleles examined, very little is known about the genes involved and/or
the potential biological mechanisms underlying their association with prostate cancer
risk. None of the risk variants examined in this study are located in exons. Decreased
serum levels of the protein product of MSMB, Prostate Secretory Protein of 94 amino
acids has been associated with increased prostate cancer risk (16). MSMB may be a
tumor suppressor and altering its expression could play an important role in
tumorigenesis. HNF1B, a transcription factor, is involved in nephrogenesis, and
heterozygous mutations in HNF1B are known to cause maturity-onset diabetes of the
young (MODY5) (17, 18). HNF1B is located on chromosome 17q12 where two
independent risk alleles for prostate cancer have been detected in non-coding sequence
(1, 5). KLK3 encodes the PSA protein, which raises the question as to whether the
53
previously reported relationship is causal or an artifact of differential selection of cases
and controls based on PSA levels (19). Controls were unselected with regard to PSA
level in the present study. A statistically significant positive association was found with
the KLK3 SNP for subjects of European and Japanese ancestry, whereas a significant
inverse association was found in African Americans. This may suggest that either the
variant is not causal, and/or that distinct mechanisms are at play in these populations. In
addition to the already established 8q24 region, the variants at both 3p12 and 11q13 lie in
gene deserts, with the closest annotated genes being ~70 kb and ~67 kb away,
respectively.
In conclusion, we have confirmed that the majority of associations noted with
prostate cancer risk variants from GWAS in European populations can be generalized to
other populations. Moreover, they appear to act independently. Deep resequencing and
fine-mapping of these regions in samples from multiple populations is now
recommended, specifically for loci that display significant ethnic heterogeneity, to both
define the full spectrum of risk alleles in the population, as well as further localize the
causal alleles.
54
Chapter 2 References
1. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome
17 confer prostate cancer risk, and the one in TCF2 protects against type 2
diabetes. Nat Genet 2007;398:977-83.

2. Eeles RA, Kote-Jarai Z, Giles GG, et al. Multiple newly identified loci associated
with prostate cancer susceptibility. Nat Genet 2008;403:316-21.

3. Gudmundsson J, Sulem P, Rafnar T, et al. Common sequence variants on 2p15
and Xp11.22 confer susceptibility to prostate cancer. Nat Genet 2008;403:281-3.

4. Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified in a genome-wide
association study of prostate cancer. Nat Genet 2008;403:310-5.

5. Sun J, Zheng SL, Wiklund F, et al. Evidence for two independent prostate cancer
risk-associated loci in the HNF1B gene at 17q12. Nat Genet 2008;4010:1153-5.

6. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24
independently affect risk for prostate cancer. Nat Genet 2007;395:638-44.

7. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and
Los Angeles: baseline characteristics. Am J Epidemiol 2000;1514:346-57.

8. Lee LG, Connell CR, Bloch W. Allelic discrimination by nick-translation PCR
with fluorogenic probes. Nucleic Acids Res 1993;2116:3761-6.

9. The International HapMap Project. Nature 2003;4266968:789-96.

10. Eeles RA, Kote-Jarai Z, Al Olama AA, et al. Identification of seven new prostate
cancer susceptibility loci through a genome-wide association study. Nat Genet
2009;4110:1116-21.

11. Kote-Jarai Z, Easton DF, Stanford JL, et al. Multiple novel prostate cancer
predisposition loci confirmed by an international study: the PRACTICAL
Consortium. Cancer Epidemiol Biomarkers Prev 2008;178:2052-61.

12. Gudmundsson J, Sulem P, Gudbjartsson DF, et al. Genome-wide association and
replication studies identify four variants associated with prostate cancer
susceptibility. Nat Genet 2009;4110:1122-6.

13. Rafnar T, Sulem P, Stacey SN, et al. Sequence variants at the TERT-CLPTM1L
locus associate with many cancer types. Nat Genet 2009;412:221-7.
55
14. Duggan D, Zheng SL, Knowlton M, et al. Two genome-wide association studies
of aggressive prostate cancer implicate putative prostate tumor suppressor gene
DAB2IP. J Natl Cancer Inst 2007;9924:1836-44.

15. Freedman ML, Haiman CA, Patterson N, et al. Admixture mapping identifies
8q24 as a prostate cancer risk locus in African-American men. Proc Natl Acad Sci
U S A 2006;10338:14068-73.

16. Nam RK, Reeves JR, Toi A, et al. A novel serum marker, total prostate secretory
protein of 94 amino acids, improves prostate cancer detection and helps identify
high grade cancers at diagnosis. J Urol 2006;1754:1291-7.

17. Dudziak K, Mottalebi N, Senkel S, et al. Transcription factor HNF1beta and novel
partners affect nephrogenesis. Kidney Int 2008;742:210-7.

18. Fajans SS, Bell GI, Polonsky KS. Molecular mechanisms and clinical
pathophysiology of maturity-onset diabetes of the young. N Engl J Med
2001;34513:971-80.

19. Ahn J, Berndt SI, Wacholder S, et al. Variation in KLK genes, prostate-specific
antigen and risk of prostate cancer. Nat Genet 2008;409:1032-4; author reply 5-6.

56
Chapter 3: A Common Prostate Cancer Risk Variant in the 5’ Region of
MSMB is a Strong Predictor of Circulating MSP Levels in Multiple
Populations: The Multiethnic Cohort Study

Introduction

Beta-microseminoprotein (MSP) has been implicated as a biomarker of prostate
cancer risk, detection and prognosis (1, 2). MSP, also called prostate secretory protein of
94 amino acids (PSP94), is one of the most highly secreted proteins by the prostate gland
(3). MSP may have a physiologic role in fertility as it has been shown to bind sperm and
inhibit the acrosome reaction (4). In the prostate, MSP expression has been observed to
be higher in benign than in cancerous tissue (5). Serum MSP levels have been found to be
lower in men with aggressive prostate cancer (1) as well as men with a prostate cancer
recurrence after radical prostatectomy (2, 6). Studies in vivo and in vitro have also
implicated MSP as having a functional role in regulating cellular growth in the prostate
through apoptosis (7).
Genome-wide association studies (GWAS) of prostate cancer have identified a
common risk variant (rs10993994) in the 5’ region of MSMB on chromosome 10q11.2
(the gene that encodes MSP) (8, 9). We recently replicated the association of this variant
with prostate cancer risk in a multiethnic population (10). The variant, rs10993994, is
located 57 bp upstream of the first exon of MSMB (11, 12) with the C allele (low risk
allele) found to preferentially bind the CREB transcription factor and to be associated
57
with higher MSMB gene expression in tumor cell lines (12). A recent study among Chinese
men, in which 60 of 251 cases and 30 of 258 controls were evaluated for the potential association
between MSP levels and MSMB genotype, they reported serum MSP levels to be lower in those
who carried the CT or TT genotype than in those with the CC genotype among the 60 cases (13).
These genetic data, together with the clinical and laboratory observations, are consistent
with the hypothesis that lower MSMB expression is associated with prostate cancer
development.
In an attempt to clarify the biological mechanism underlying the association of
genetic variation at the MSMB locus and prostate cancer risk, we examined the
association between rs10993994 genotype and MSP levels in the blood plasma from
prostate cancer-free men from four racial/ethnic populations (European Americans,
African Americans, Latinos, and Japanese Americans). We also examined known and
putative risk factors for prostate cancer as potential contributors to inter-individual
variability of circulating levels of MSP in the study population.

Methods

Study Subjects
This study consisted of male participants of the Multiethnic Cohort (MEC). The
MEC includes 215,251 men and women and is comprised mainly of African Americans,
Japanese Americans, Latinos, Native Hawaiians and European Americans (14). Between
1993 and 1996, adults between 45 and 75 years old were enrolled by completing a 26-
58
page, self-administered questionnaire asking detailed information about demographic
factors, personal behaviors, dietary habits and history of prior medical conditions.
Potential cohort members were identified through Department of Motor Vehicles drivers’
license files, voter registration files and Health Care Financing Administration data files.
Between 1995 and 2006, blood specimens were collected from ~67,000 MEC
participants for genetic and biomarker analyses.
Blood plasma MSP and PSA levels and rs10993994 genotype were measured in
500 men from four racial/ethnic groups (n=125 of each European Americans, African
Americans, Latinos, and Japanese Americans). Men aged 50-69 years old at the time of
blood draw and with a self-reported BMI between 19 and 35 were randomly selected
from the control group of the nested prostate cancer case-control study we previously
reported in the MEC (10). Four hundred and twelve (84%) of the men had fasted 8 or
more hours before blood collection. The Institutional Review Boards at the University of
Southern California and University of Hawaii approved the study protocol.

Lab Assays
Archived EDTA anti-coagulated blood plasma samples (stored frozen for 3-14
years at -80
o
C after initial processing within 4 hours from venipuncture) were retrieved
and shipped frozen on dry ice to Malmö, Sweden in fall of 2009. Analyses of free and
total PSA, and MSP were performed blinded to ethnicity or genotyping data in Dr. Lilja’s
laboratory at the Wallenberg Research Laboratories, Department of Laboratory Medicine,
Lund University, Skåne University Hospital in Malmö, Sweden during 2009 - 2010.
59
Production and purification of the polyclonal rabbit anti-MSP antibody and seminal
plasma MSP have been described elsewhere (15). Biotinylation and Europium labeling of
the anti-MSP antibody were done as previously described (16). Streptavidin coated
microtiter plates were from Kaivogen Oy, Turku, Finland. Assay buffer and wash
solution have been described previously (15).
The MSP immunoassay was done with AutoDelfia 1235 automatic immunoassay
system (Perkin-Elmer Life Sciences, Wallac, Turku, Finland) and all incubations were
done at room temperature with shaking. First, 150 ng of the biotinylated MSP-antibody
was attached to the streptavidin coated wells in 100 μl of assay buffer during a 30 min
incubation. After two washes, 100 μl of assay buffer and 10 μl of standard or sample
material were added to the wells, incubated 1 h and washed twice. The Eu-labelled anti-
MSP antibody (50 ng) was added in 200 μl of assay buffer. After 1 h incubation, the
wells were washed six times before adding 200 μl of the Delfia enhancement solution
(Perkin-Elmer Life Sciences, Wallac, Turku, Finland) which was then incubated for 5
min at room temperature with shaking before the time-resolved measurement of the Eu-
fluorescence. To measure free and total PSA, we used the dual-label DELFIA
Prostatus® total/free PSA-Assay (Perkin-Elmer, Turku, Finland) (17), which is calibrated
against the WHO 96/670 (PSA-WHO) and WHO 68/668 (free PSA-WHO) standards.
Samples were run in 8 batches. Mean coefficients of variation of 7.4%, 11.8%, and 7.8%
were observed for MSP, total PSA, and free PSA respectively among 20 blinded
duplicate samples (5 per ethnic group) in this study.

60
Genotyping
Genotyping of rs10993994 and other GWAS-validated risk alleles for prostate
cancer was performed using the Taqman allelic discrimination, as previously described
(10, 18, 19).

Statistical analysis
Demographic variables are presented as the mean ± the standard deviation and
ANOVA and ANCOVA were used to test for differences in the mean levels of
demographic variables between racial/ethnic groups. MSP and PSA levels were log
transformed to better normalize the distributions and mean levels are presented as least
squares geometric means. Generalized linear models were used to examine the
association between rs10993994 and MSP, adjusted for age (continuous), BMI
(continuous), laboratory batch, and race/ethnicity (pooled analysis). Laboratory batch
was not significantly associated with MSP levels. We also adjusted for PSA levels, as
PSA was found to be modestly correlated with MSP levels. Adjustment for either total,
free PSA (ng/mL) or percent free PSA provided similar results. The least squares
geometric mean MSP levels and the percent change relative to the wild-type (i.e. low
risk) rs10993994 genotype (CC) was calculated in ethnic-specific and pooled analysis.
The effect of global European ancestry on the association between MSP levels and
rs10993994 was examined in the admixed racial/ethnic populations (African Americans
and Latinos) using ancestry estimates from a previous study (18). Fasting time before
61
blood collection was not significantly associated with MSP levels and had no effect on
the association between rs10993994 and MSP.
As the determinants of MSP levels are largely unknown, we also examined other
known and suspected risk factors for prostate cancer in relation to MSP levels including
demographic, dietary, and established common risk variants from GWAS. More
specifically we examined age at the time of blood draw, BMI (kg/m
2
), weight (kg), height
(cm), total PSA (ng/mL), free PSA (ng/mL), percent free PSA, physical activity (hours of
vigorous work and exercise), and intake of saturated fat (percentage of calories from
saturated fat), red meat (calorie adjusted intake per day; g/kcal/day), lycopene (calorie
adjusted intake per day; 100mcg/kcal/day), calcium (calorie adjusted intake per day;
mg/kcal/day), vitamin D (calorie adjusted intake per day; IU/kcal/day), and alcohol
(grams per day). We tested the association between an increase of 1 unit of each factor
and the percent change in geometric mean MSP levels. Age, BMI, weight, and height
were coded as continuous variables. PSA levels and the dietary variables were
categorized into quintiles and assigned the mean value of the quintile. Physical activity
(0, >0-1.5, and >1.5 hours/week) and alcohol intake (0, >0-10, >10-30, and >30
grams/day) were assigned the mean value of the category. We also examined 26 other
validated genetic risk variants for prostate cancer in association with MSP levels (10, 18,
19). Coefficients of determination (r
2
) were estimated in multivariable models and used
to assess the variance of MSP levels explained by rs10993994 genotype and other
covariates.

62
Results

The mean ages at the time of blood draw were not significantly different between
racial/ethnic groups (p=0.21), with African Americans being younger on average (mean
age, 60.2 years) and Japanese Americans being slightly older (mean age, 61.5 years;
Table 3-1). There were significant differences in mean BMI among racial/ethnic groups
(p=1.9x10
-3
), with Japanese Americans having the lowest mean BMI (25.0 kg/m
2
) and
Latinos the highest (26.5 kg/m
2
). There were also significant differences in both weight
and height among the groups. We observed significant differences in mean MSP levels
across racial/ethnic groups (P=2.0x10
-4
). On average, European Americans had the
highest levels (20.4 ng/mL), followed by Latinos (19.7 ng/mL) and African Americans
(19.0 ng/mL) with the lowest levels observed for Japanese Americans (14.8 ng/mL). No
significant differences were noted with mean PSA levels (total or free) levels across
populations (P=0.40 and 0.28 respectively).
63
Table 3-1: Descriptive Characteristics by Race/Ethnicity (n=500)

European
Americans
African
Americans
Latinos

Japanese
Americans
p value

N 125 125 125 125
Age (yr) 61.1 ± 5.2 60.2 ± 5.5 60.6 ± 4.7 61.5 ± 5.4 0.21
a

BMI (kg/m
2
) 26.1 ± 3.3 26.3 ± 3.5 26.5 ± 3.4 25.0 ± 3.0 1.9x10
-3 a
Weight (kg) 84 ± 12 86 ± 14 82 ± 12 72 ± 10 1.0x10
-16 a
Height (cm) 177 ± 7 179 ± 7 175 ± 7 169 ± 6 3.5x10
-29 a
MSP (ng/mL)
c
20.4 19.0 19.7 14.8 2.0x10
-4 b
MSP (ng/mL)
d
17.8 19.3 16.4 14.1 3.5x10
-6 b
PSA – total
(ng/mL)
e
1.09 0.99 1.02 0.92 0.40
b

PSA – free
(ng/mL)
e
0.33 0.29 0.28 0.28 0.28
b

Rs10993994
Frequency of
T allele 0.41 0.55 0.36 0.48
a
Global test, ANOVA.
b
Global test, ANCOVA.
c
Least squares geometric means adjusted for laboratory batch, free PSA, age, and BMI.
d
Least squares geometric means adjusted for laboratory batch, free PSA, age, BMI, and
rs10993994 genotype.
e
Least squares geometric means adjusted for laboratory batch, MSP, age, and BMI.

The risk allele at rs10993994 (T allele) was common in all populations and
ranged in frequency from 0.36 in Latinos to 0.55 in African Americans. The variant,
rs10993994, was strongly and highly statistically associated with MSP levels in each
racial/ethnic group (all global P values < 4 x 10
-9
; Figure 1). Compared to men with the
CC genotype, men with the CT genotype had a percent change in geometric mean PSA
levels between -24.3% (African Americans) and -41.5% (Latinos) (all P<0.05), and men
with the TT genotype had changes ranging from -62.3% in African Americans to -74.2%
in Japanese Americans (all P< 10
-8
) (Table 3-2). We observed little evidence of
heterogeneity of the association of genotype and MSP levels between racial/ethnic groups

Figure 3-1: Least Squares Mean of Log MSP (ng/mL) by rs10993994 Genotype for Each Racial/Ethnic Group

Figure 3-1. The least squares geometric mean of log MSP levels (ng/mL) for each racial/ethnic group (European Americans, African
Americans, Latinos, and Japanese Americans) by the genotype of rs10993994 (n=500). The error bars represent the upper and lower 95%
confidence limits of the mean. The values are adjusted for age (continuous), BMI (continuous), free PSA level (ng/mL), laboratory batch,
and ethnicity (pooled). P values are from a global ANCOVA test (Type III sum of squares F-test)
64

Table 3-2: Least Squares Geometric Mean MSP Levels (ng/mL) and Percent Change by rs10993994 Genotype
rs10993994
Genotype
European Americans (n=125)

African Americans
(n=125)

Latinos
(n=125)

Japanese Americans
(n=125)

n Mean
a
%
change
P value n Mean
a
%
change
P value N Mean
a
%
change
P value n Mean
a
%
change
P
value
CC 44 28.8
ref
ref 23 29.0
ref
Ref 52 29.2
Ref
ref 32 25.8
ref
ref
CT 60 20.9
-27.6%
1.0 x 10
-3
66 21.9
-24.3%
0.041 55 17.1
-41.5%
1.4 x 10
-9
66 16.4
-36.4%
1.3 x 10
-5
TT 21 10.0
-65.2%
5.0 x 10
-12
36 10.9
-62.3%
6.6 x 10
-9
18 8.3
-71.5%
9.2 x 10
-19
27 6.7
-74.2%
8.8 x 10
-20
r
2b
0.32 0.30 0.49 0.50
r
2c
0.44 0.39 0.60 0.61
a
Values adjusted for age, BMI, free PSA, laboratory batch, and race/ethnicity (for pooled sample).
b
Univariate model with rs10993994 (categorized as genotype class) as independent variable.
c
Multivariate model with rs10993994 (categorized as genotype class), age, BMI, laboratory batch, and race/ethnicity (for pooled sample) as independent
variables.
65
66
(P=0.37; Table 3-2). The associations were similar when adjusting for global European
ancestry in the admixed African American and Latino populations (Table 3-3).
Differences in the frequency of rs10993994 genotype could not explain the ethnic
differences in MSP levels (global P = 3.5x10
-6
; Table 3-1). Following adjustment for
genotype, MSP levels were highest in African Americans (19.3 ng/mL) and lowest in
Japanese Americans (14.1 ng/mL; Table 3-1).
We found no significant association between age and MSP levels (P=0.10; Table
3-4), however the youngest 5-year age group (50-54 years) was noted as having lower
least squares geometric mean levels than the older groups (55-70 years; Table 3-5). We
observed significant inverse associations of MSP levels with BMI (P=7.4x10
-3
) and
weight (P=3.6x10
-3
) and a suggestive association with height (P=0.072). Family history
of prostate cancer was associated with MSP levels in African Americans (P=0.023), but
not in the pooled population (P=0.34). We observed modest, yet, statistically significant
positive correlations between MSP levels and free (P=3.2x10
-9
), and total (P=4.2x10
-5
)
PSA (Table 3-4 and 3-6), with Pearson correlation coefficients of 0.15 (0.07-0.21 among
racial/ethnic groups) and 0.13 (0.08-0.19 among racial/ethnic groups), respectively.
Aside from age and height in Latinos the direction of the associations with these factors
was consistent across populations.
We found no strong associations between suspected risk factors for prostate
cancer and MSP levels (Table 3-4). Of the dietary variables, we observed a -0.44%
decrease in MSP levels per 1 unit (100mcg/kcal/day) increase in lycopene intake

Table 3-3: Effect of Adjustment for Global European Ancestry on Association of rs10993994 with Plasma MSP Levels
(n=246)
a

rs10993994
Genotype
African Americans
Unadjusted
(n=123)

African Americans
Adjusted
(n=123)

Latinos
Unadjusted
(n=123)

Latinos
Adjusted
(n=123)

N

%
change
b
P value N %
change
c
P value n %
change
b
P value n %
change
c
P value
CC 22 ref ref 22 Ref Ref 50 Ref ref 50 Ref Ref
CT 66 -0.22 0.059 66 -0.24 0.038 55 -0.41 2.0x10
-9
55 -0.41 2.5x10
-9

TT 35 -0.63 1.0x10
-9
35 -0.64 3.8x10
-10
18 -0.72 1.4x10
-18
18 -0.72 2.0x10
-18

r
2
0.44 0.45 0.60 0.60
a
4 subjects (2 African American, 2 Latino) not included because of missing ancestry information.
b
Values adjusted for age, BMI, free PSA, and laboratory batch.
c
Values adjusted for age, BMI, free PSA, laboratory batch, and percent global European ancestry.
67

Table 3-4: The Association of MSP Levels with Age, PSA Levels and Suspected Risk Factors for Prostate Cancer
European Americans
(n=125)
African Americans
(n=125)
Latinos
(n=125)
Japanese Americans
(n=125)
Pooled
(n=500)

%
Change
P value % Change P value %
Change
P value %
Change
P value %
Change
P value P
het
a
Age (years)
b
0.79% 0.35 1.52% 0.11 -0.93% 0.26 1.06% 0.19
0.69% 0.10 0.40
BMI (kg/m)
c
-0.62% 0.64 -3.13% 0.032 -2.11% 0.057 -1.11% 0.42
-1.73% 7.4x10
-3
0.71
Weight (kg)
c
-0.18% 0.26 -0.46% 8.2x10
-3
-0.19% 0.18 -0.21% 0.26
-0.23% 3.6x10
-3
0.57
Height (cm)
c
-2.87% 0.076 -3.13% 0.089 0.47% 0.71 -1.52% 0.39
-1.43% 0.072 0.59
1
st
degree Family History
of Prostate Cancer
d
-26.4% 0.068 37.1% 0.023 8.8% 0.54 16.3% 0.52 7.6% 0.34 0.065
Total PSA (ng/mL)
e
11.76% 2.2x10
-3
7.06% 0.12 6.66% 0.054 5.22% 0.17
7.70% 4.2x10
-5
0.41
Free PSA (ng/mL)
e
79.3% 7.1x10
-5
34.6% 0.11 55.5% 1.5x10
-3
51.3% 7.7x10
-3

56.5% 3.2x10
-9
0.37
% Free PSA
e
0.48% 0.24 0.10% 0.84 1.18% 2.7x10
-3
0.26% 0.52
0.47% 0.022 0.33
Physical Activity
(Hrs/Wk)
d
-13.25 0.051 -1.71% 0.83 1.21% 0.84 0.73% 0.93
-4.64% 0.18 0.68
Saturated Fat (% of
calories)
d
2.05% 0.51 6.03% 0.13 -2.03% 0.47
-0.29% 0.94 2.24% 0.18 0.50
Red Meat (g/kcal/day)
d
-0.15% 0.55 -0.08% 0.81 0.04% 0.87
-0.03% 0.92 -0.04% 0.76 0.98
Lycopene
(100mcg/kcal/day)
d
-0.07% 0.86 -0.48% 0.37 -0.31% 0.40
-0.95% 0.048 -0.44% 0.047 0.70
Calcium (mg/kcal/day)
d
0.01% 0.79 0.00% 0.98 -0.04% 0.31
0.08% 0.062 0.01% 0.55 0.27
Vitamin D (IU/kcal/day)
d
0.03% 0.76 -0.03% 0.83 -0.07% 0.51
0.25% 0.033 0.04% 0.54 0.31
Alcohol (g/day)
d
-0.21% 0.14 -0.39% 0.034 0.22% 0.16
-0.07% 0.64 -0.11% 0.15 0.14
a
P value for heterogeneity of effect across race/ethnicity (Type III sum of squares F-Test).
b
Adjusted for BMI, free PSA, laboratory batch rs10993994, and race/ethnicity (for pooled sample).
c
Adjusted for age, free PSA, laboratory batch, rs10993994, and race/ethnicity (for pooled sample).
d
Adjusted for age, BMI, Free PSA, laboratory batch, rs10993994, and race/ethnicity (for pooled sample).
e
Adjusted for age, BMI, laboratory batch, rs10993994, and race/ethnicity (for pooled sample).
68

Table 3-5: Least Squares Geometric Mean MSP Levels by Demographic Factors (n=500)
European
Americans
African
Americans
Latinos

Japanese
Americans
Pooled

n MSP
(ng/mL)
n MSP

(ng/mL)
n MSP
(ng/mL)
N MSP
(ng/mL)
n MSP
(ng/mL)
Age (yrs)
a
50-54 18 14.8 25 16.5 11 17.4 15 12.3 69 14.9
55-59 26 20.7 28 18.4 39 17.0 27 13.8 120 17.2
60-64 39 18.1 44 19.9 47 15.6 35 14.2 165 16.8
65-69 42 18.5 28 21.0 28 15.3 48 14.9 146 17.3
P value
b
0.16 0.45 0.65 0.58 0.18
BMI (kg/m
2
)
c
Quintile 1 21 17.9 21 24.7 22 20.7 36 14.2 100 18.2
Quintile 2 26 18.7 29 20.0 16 15.4 29 14.4 100 17.0
Quintile 3 31 19.2 20 17.4 24 16.9 25 15.9 100 17.6
Quintile 4 18 15.7 30 17.2 33 14.4 19 13.4 100 15.5
Quintile 5 29 18.3 25 18.1 30 14.9 16 11.8 100 15.6
P value
b
0.72 0.25 0.026 0.36 0.076
Weight (kg)
c
Quintile 1 17 22.6 14 29.8 22 18.1 53 14.7 106 19.6
Quintile 2 26 16.3 27 20.7 20 17.0 33 13.9 106 16.5
Quintile 3 26 21.5 26 19.6 30 15.4 21 15.1 103 17.6
Quintile 4 31 14.5 17 13.8 30 15.1 14 11.7 92 13.9
Quintile 5 25 18.6 41 17.7 23 15.3 4 12.6 93 16.0
P value
b
5.8x10
-3
6.8x10
-3
0.48 0.49 5.1x10
-5

Height
(cm)
c
Quintile 1 18 22.4 13 24.4 27 16.0 66 15.2 124 18.7
Quintile 2 13 19.2 24 21.0 33 16.5 37 12.9 107 17.0
Quintile 3 25 16.6 13 19.7 21 14.8 9 13.2 68 15.4
Quintile 4 36 19.0 31 17.1 25 17.0 12 13.6 104 16.4
Quintile 5 33 16.3 44 17.6 19 15.2 1 14.9 97 15.2
P value
b
0.16 0.35 0.81 0.49 0.036
a
Least squares geometric mean of MSP adjusted for BMI, free PSA, laboratory batch, rs10993994, and race/ethnicity (pooled).
b
Global test, ANCOVA.
c
Least squares geometric mean of MSP adjusted for age, free PSA, laboratory batch, rs10993994, and race/ethnicity (pooled).
69

Table 3-6: Least Squares Geometric Mean MSP Levels for Quintiles of PSA Levels (n=500)

European
Americans
(n=125)

African
Americans
(n=125)
Latinos

(n=125)
Japanese
Americans
(n=125)
Pooled

(n=500)

n MSP
a

(ng/mL)

n MSP
a

(ng/mL)
N MSP
a

(ng/mL)
n MSP
a
(ng/mL)
n MSP
a

(ng/mL)
Total PSA
(ng/mL)
Quintile 1 19 19.1 28 17.1 26 15.1 29 13.5 102 15.6
Quintile 2 27 15.7 26 17.5 26 14.9 19 9.9 98 14.4
Quintile 3 22 14.6 23 18.5 26 15.0 29 14.5 100 16.2
Quintile 4 31 19.2 24 20.5 18 16.7 27 18.6 100 18.4
Quintile 5 26 23.9 24 21.8 29 18.5 21 15.0 100 19.5
P value
b
6.2x10
-3
0.56 0.4 4.3x10
-4
7.0x10
-5

Free PSA
(ng/mL)
Quintile 1 23 14.1 33 15.8 30 12.9 27 12.6 113 13.8
Quintile 2 18 19.2 18 16.7 29 15.0 32 12.3 97 15.5
Quintile 3 27 17.1 24 23.4 23 17.2 20 14.4 94 17.6
Quintile 4 28 16.4 26 19.2 20 16.3 26 16.3 100 16.9
Quintile 5 29 25.4 24 21.5 23 20.3 20 17.5 96 21.2
P value
b
5.0x10
-4
0.074 0.012 0.042 3.2x10
-8

Percent Free
PSA
Quintile 1 22 18.2 25 20.5 28 15.0 27 15.0 102 17.2
Quintile 2 33 17.1 19 17.3 25 15.5 23 11.9 100 15.4
Quintile 3 24 18.9 27 19.5 25 14.2 23 14.5 99 16.1
Quintile 4 15 16.8 33 17.2 24 16.8 27 13.5 99 16.0
Quintile 5 31 20.7 21 21.7 23 21.8 25 15.3 100 19.5
P value
b
0.58 0.55 0.012 0.38 6.8x10
-3

a
Least squares geometric mean of MSP adjusted for age, BMI, laboratory batch, rs10993994, and race/ethnicity (pooled).
b
Global test, ANCOVA.
70
71
(p=0.047; Table 3-4, and 3-7). Since the biological function of most genetic risk variants
for prostate cancer are not known, we examined their association
with MSP levels. In testing 26 validated risk variants (8, 9, 18-22), we observed a
nominally significant inverse association with only two variants, rs721048 on
chromosome 2p15 (EHBP1 gene locus; -11.5% change in MSP (ng/mL) per allele
P=0.013), and rs7931342 on chromosome 11q13 (7.2% change in MSP per allele
P=0.041).
In univariable models, rs10993994 genotype explained between 30% (African
Americans) and 50% (Japanese Americans) of the variation in log MSP levels. Together
with age, BMI, free PSA, and laboratory batch, between 39% (African Americans) and
61% (Japanese Americans) of the variation in log MSP levels could be explained. In
ethnic-pooled analysis, rs10993994 explained 38% of the variance in log MSP levels,
while 48% of the variance could be explained when including age, BMI, free PSA,
ethnicity, and laboratory batch. Lycopene, rs721048, and rs7931342 explained little
additional variation (<0.02 increase in r
2
).

Table 3-7: Least Squares Geometric Mean MSP Levels by Suspected Prostate Cancer Risk Factors (n=500)

European
Americans
African
Americans
Latinos

Japanese
Americans
Pooled

n
a
MSP
b

(ng/mL)

n
a
MSP
b

(ng/mL)
n
a
MSP
b

(ng/mL)
n
a
MSP
†
(ng/mL)
n
a
MSP
b

(ng/mL)
Saturated Fat
(Percent of
calories)
Quintile 1 15 17.8 16 17.0 17 17.5 50 14.2 98 16.3
Quintile 2 24 17.0 26 17.3 19 15.1 29 14.7 98 16.0
Quintile 3 26 20.1 19 18.0 26 17.7 27 13.2 98 17.2
Quintile 4 26 16.7 29 20.9 29 15.1 14 14.0 98 16.5
Quintile 5 28 19.6 31 20.5 34 15.6 4 16.9 97 17.9
P value
c
0.48 0.61 0.58 0.83 0.49
Lycopene
Density
(mcg/kcal/
day)
Quintile 1 21 19.7 33 23.7 13 16.3 31 17.2 98 19.7
Quintile 2 15 20.7 24 14.5 27 16.7 32 13.6 98 16.3
Quintile 3 24 16.9 24 16.9 25 16.9 25 14.4 98 16.1
Quintile 4 31 17.5 20 20.9 26 15.4 21 12.2 98 16.0
Quintile 5 28 19.0 20 17.5 34 15.4 15 12.8 97 16.1
P value
c
0.58 0.020 0.87 0.070 7.0x10
-3

Calcium
Density
(mg/kcal/
day)
Quintile 1 13 19.2 29 19.9 6 18.8 50 13.8 98 17.1
Quintile 2 21 19.0 30 18.1 17 17.3 30 13.9 98 16.7
Quintile 3 20 17.1 27 20.2 31 15.5 20 15.0 98 16.7
Quintile 4 26 17.5 19 15.6 38 16.3 15 12.3 98 15.7
Quintile 5 39 19.1 16 20.8 33 15.3 9 21.7 97 17.8
P value
c
0.87 0.49 0.73 0.050 0.48
72

Table 3-7: Continued

European
Americans
African
Americans
Latinos

Japanese
Americans
Pooled

n
a
MSP
b

(ng/mL)

n
a
MSP
b

(ng/mL)
n
a
MSP
b

(ng/mL)
n
a
MSP
†
(ng/mL)
n
a
MSP
b

(ng/mL)
Vitamin D
Density
(IU/kcal/day)
Quintile 1 15 18.3 28 18.6 22 16.9 33 13.5 98 16.8
Quintile 2 19 19.6 16 17.9 32 16.8 31 13.2 98 16.8
Quintile 3 24 17.2 29 21.0 27 15.2 18 13.5 98 16.6
Quintile 4 28 17.9 25 18.1 22 15.8 23 13.9 98 16.1
Quintile 5 33 19.1 23 18.4 22 15.6 19 17.6 97 17.5
P value
c
0.88 0.86 0.87 0.24 0.78
Red Meat
Density
(g/kcal/day)
Quintile 1 29 19.2 22 20.2 22 16.2 25 14.0 98 17.2
Quintile 2 26 19.0 17 16.4 28 14.9 27 14.8 98 16.2
Quintile 3 23 17.8 32 18.5 17 18.0 25 13.5 97 16.4
Quintile 4 17 17.4 29 20.4 23 16.1 29 14.2 98 17.5
Quintile 5 24 18.1 21 17.4 35 16.0 18 14.0 98 16.3
P value
c
0.94 0.70 0.71 0.97 0.75
Alcohol
(g/day)
0 45 19.2 49 22.2 39 16.0 42 14.7 175 17.7
>0-10 32 18.5 39 18.6 48 16.2 36 15.0 155 16.9
>10-30 20 17.9 13 15.2 24 14.3 29 12.9 86 15.4
>30 22 16.0 20 15.2 14 19.6 17 14.1 73 16.0
P value
c
0.51 0.040 0.18 0.61 0.10
Physical
Activity
(Hrs/Wk)

0 37 20.3 41 18.7 31 15.4 29 13.7 138 16.8
>0-1.5 76 17.8 72 19.6 72 16.6 90 14.3 310 16.6
>1.5 12 15.1 12 16.9 22 15.7 6 13.0 52 15.2
P value
c
0.15 0.69 0.71 0.86 0.41
a
Subjects do not add up to 125 in each group and 500 total due to missing values.
b
Least squares geometric mean of MSP adjusted for age, BMI, Free PSA, laboratory batch, rs10993994, and race/ethnicity (pooled).
c
Global test, ANCOVA.
73
74
Discussion

In this study, we found that rs10993994 genotype is an important and highly
statistically significant predictor of circulating MSP levels. The association of genotype
and MSP levels was observed in all racial/ethnic groups which supports fine-mapping
studies pinpointing rs10993994 as the most strongly associated prostate cancer risk
variant in the region (11, 12). We also found MSP levels to vary significantly among
racial/ethnic groups, and after conditioning on genotype, African Americans were
observed to have the highest levels and Japanese the lowest. Low levels of MSP, as a
potential risk factor for prostate cancer, are unlikely to contribute to ethnic differences in
prostate cancer incidence, as we have previously found Japanese Americans to have the
lowest rates of prostate cancer and African Americans the highest in the MEC (23).
The variant is located in a putative CREB transcription factor binding site and the
C to T change at this position has been shown to influence CREB binding, with
preferential binding observed with the low risk C allele. In cell lines, the C allele has also
been associated with increased MSMB expression (12). A recent study in a Chinese Han
population found significantly lower circulating MSP levels in men with prostate cancer
and the CT or TT genotype (16.32 μg/L) compared to men with the CC genotype (19.33
μg/L) (P=0.022) (24). A similar but non-significant association was also noted in 30
unaffected men. In our lone population of Asian ancestry (Japanese Americans), we
observed geometric mean levels of 6.7, 16.4, and 25.8 (ng/mL) in men with the TT, CT,
and CC genotypes respectively. While our findings are supportive, it is difficult to
75
compare mean MSP values between studies since we measured MSP in plasma and these
are different populations. Our observation among unaffected men from multiple
populations provides further support for the hypothesis that variant rs10993994 is
biologically functional, and that the risk for prostate cancer may be conferred by altering
expression of MSP.
Numerous studies have investigated the genetic basis underlying inter-individual
variation in circulating levels of biomarkers with relevance to prostate cancer including
PSA, SHBG, testosterone, 3α-androstanediol-glucoronide, and 17β-estradiol (25, 26).
The genetic variants identified to date in association with these biomarkers however
explain only a small fraction of the variation in circulating levels in any population. In
contrast, the variant at the MSMB locus alone may account for 30% (African Americans)
to 50% (Japanese Americans) of the variation in levels in any one population. This
finding supports those of Xu et al. in a population of 18-21 year old Swedish men (paper
co-submitted). While rs10993994 does explain a large portion of the variance in MSP
levels, a portion of the remaining variance may be explained by day-to-day inter-
individual variation, environmental risk factors, and other genetic variants. We observed
a nominally significant inverse association between MSP levels and lycopene intake. This
would not be expected if MSP is protective for prostate cancer as both high lycopene
intake and lycopene measured in the blood have been associated with lower risk of
prostate cancer (27, 28). However, this finding was not replicated in the MEC (29). The
geometric mean MSP levels were lower in men with more risk alleles for rs721048 which
lies in an intron of the EHBP1 gene on chromosome 2, and MSP levels were higher in
76
men with more risk alleles of rs7931342 located in a gene desert on chromosome 11.
Both of these modest associations (P>0.01) should be interpreted cautiously as neither
was significant after adjusting for multiple tests. Sequencing efforts in the MSMB region
(30) may also reveal additional variants that independently influence MSP levels. The
associations with these genetic variants and lycopene should be interpreted cautiously as neither
was significant after adjusting for multiple tests.
In summary, we observed a strong association between rs10993994 genotype and
plasma levels of MSP in a multiethnic sample of men without prostate cancer. The
association was robust and statistically significant in all four racial/ethnic groups. The
variant explained a large proportion of the variability in plasma MSP levels in all groups.
Examining MSP in a prospective study will be necessary to determine whether the
prostate cancer risk association with rs10993994 is mediated through influence on MSP
levels, as well as the population risk associated with MSP levels.
77
Chapter 3 References
1. Nam RK, Reeves JR, Toi A, et al. A novel serum marker, total prostate secretory
protein of 94 amino acids, improves prostate cancer detection and helps identify
high grade cancers at diagnosis. J Urol 2006;1754:1291-7.

2. Reeves JR, Dulude H, Panchal C, et al. Prognostic value of prostate secretory
protein of 94 amino acids and its binding protein after radical prostatectomy. Clin
Cancer Res 2006;1220 Pt 1:6018-22.

3. Lilja H, Abrahamsson PA. Three predominant proteins secreted by the human
prostate gland. Prostate 1988;121:29-38.

4. Anahi Franchi N, Avendano C, Molina RI, et al. beta-Microseminoprotein in
human spermatozoa and its potential role in male fertility. Reproduction
2008;1362:157-66.

5. Imasato Y, Xuan JW, Sakai H, et al. PSP94 expression after androgen deprivation
therapy: a comparative study with prostate specific antigen in benign prostate and
prostate cancer. J Urol 2000;1645:1819-24.

6. Bjartell AS, Al-Ahmadie H, Serio AM, et al. Association of cysteine-rich
secretory protein 3 and beta-microseminoprotein with outcome after radical
prostatectomy. Clin Cancer Res 2007;1314:4130-8.

7. Garde SV, Basrur VS, Li L, et al. Prostate secretory protein (PSP94) suppresses
the growth of androgen-independent prostate cancer cell line (PC3) and
xenografts by inducing apoptosis. Prostate 1999;382:118-25.

8. Eeles RA, Kote-Jarai Z, Giles GG, et al. Multiple newly identified loci associated
with prostate cancer susceptibility. Nat Genet 2008;403:316-21.

9. Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified in a genome-wide
association study of prostate cancer. Nat Genet 2008;403:310-5.

10. Waters KM, Le Marchand L, Kolonel LN, et al. Generalizability of associations
from prostate cancer genome-wide association studies in multiple populations.
Cancer Epidemiol Biomarkers Prev 2009;184:1285-9.

11. Chang BL, Cramer SD, Wiklund F, et al. Fine mapping association study and
functional analysis implicate a SNP in MSMB at 10q11 as a causal variant for
prostate cancer risk. Hum Mol Genet 2009;187:1368-75.
78
12. Lou H, Yeager M, Li H, et al. Fine mapping and functional analysis of a common
variant in MSMB on chromosome 10q11.2 associated with prostate cancer
susceptibility. Proc Natl Acad Sci U S A 2009;10619:7933-8.

13. Xu B, Wang J, Tong N, et al. A functional polymorphism in MSMB gene
promoter is associated with prostate cancer risk and serum MSMB expression.
Prostate 2010;7010:146-52.

14. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and
Los Angeles: baseline characteristics. Am J Epidemiol 2000;1514:346-57.

15. Valtonen-Andre C, Savblom C, Fernlund P, et al. Beta-microseminoprotein in
serum correlates with the levels in seminal plasma of young, healthy males. J
Androl 2008;293:330-7.

16. Vaisanen V, Peltola MT, Lilja H, et al. Intact free prostate-specific antigen and
free and total human glandular kallikrein 2. Elimination of assay interference by
enzymatic digestion of antibodies to F(ab')2 fragments. Anal Chem
2006;7822:7809-15.

17. Mitrunen K, Pettersson K, Piironen T, et al. Dual-label one-step immunoassay for
simultaneous measurement of free and total prostate-specific antigen
concentrations and ratios in serum. Clin Chem 1995;418 Pt 1:1115-20.

18. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24
independently affect risk for prostate cancer. Nat Genet 2007;395:638-44.

19. Eeles RA, Kote-Jarai Z, Al Olama AA, et al. Identification of seven new prostate
cancer susceptibility loci through a genome-wide association study. Nat Genet
2009;4110:1116-21.

20. Gudmundsson J, Sulem P, Rafnar T, et al. Common sequence variants on 2p15
and Xp11.22 confer susceptibility to prostate cancer. Nat Genet 2008;403:281-3.

21. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome
17 confer prostate cancer risk, and the one in TCF2 protects against type 2
diabetes. Nat Genet 2007;398:977-83.

22. Sun J, Zheng SL, Wiklund F, et al. Evidence for two independent prostate cancer
risk-associated loci in the HNF1B gene at 17q12. Nat Genet 2008;4010:1153-5.

23. Kolonel LN, Altshuler D, Henderson BE. The multiethnic cohort study: exploring
genes, lifestyle and cancer risk. Nat Rev Cancer 2004;47:519-27.
79
24. Xu B, Wang J, Tong N, et al. A functional polymorphism in MSMB gene
promoter is associated with prostate cancer risk and serum MSMB expression.
Prostate.

25. Ahn J, Berndt SI, Wacholder S, et al. Variation in KLK genes, prostate-specific
antigen and risk of prostate cancer. Nat Genet 2008;409:1032-4; author reply 5-6.

26. Ahn J, Schumacher FR, Berndt SI, et al. Quantitative trait loci predicting
circulating sex steroid hormones in men from the NCI-Breast and Prostate Cancer
Cohort Consortium (BPC3). Hum Mol Genet 2009;1819:3749-57.

27. Giovannucci E, Ascherio A, Rimm EB, et al. Intake of carotenoids and retinol in
relation to risk of prostate cancer. J Natl Cancer Inst 1995;8723:1767-76.

28. Gann PH, Ma J, Giovannucci E, et al. Lower prostate cancer risk in men with
elevated plasma lycopene levels: results of a prospective analysis. Cancer Res
1999;596:1225-30.

29. Gill JK, Franke AA, Steven Morris J, et al. Association of selenium, tocopherols,
carotenoids, retinol, and 15-isoprostane F(2t) in serum or urine with prostate
cancer risk: the multiethnic cohort. Cancer Causes Control 2009;207:1161-71.

30. Kote-Jarai Z, Leongamornlert D, Tymrakiewicz M, et al. Mutation analysis of the
MSMB gene in familial prostate cancer. Br J Cancer;1022:414-8.

80
Chapter 4: Association of Diabetes with Prostate Cancer Risk in the
Multiethnic Cohort

Published in: AM J Epidemiol 2009; 169: 937-945

Introduction

Prostate cancer and type 2 diabetes (T2D) are two of the most common chronic
diseases that afflict the aging male population. Epidemiologic studies conducted
primarily in populations of European ancestry, have provided evidence of an inverse
relationship between these diseases, with diabetics having a ~20% lower risk of
developing prostate cancer than non-diabetics (1-8). However, considerable effect
heterogeneity has been noted among studies, highlighting the need for additional
prospective analyses of these endpoints in large representative population-based studies.
The biological basis of this suspected relationship is currently unknown, and aside from
age, and perhaps obesity (4, 9, 10), these two diseases share no known non-genetic risk
factors.
The inverse relationship between these endpoints may be due to direct effects on
prostate cancer growth and development as men with T2D have been found to have lower
PSA levels, on average, than men without T2D (11, 12). The reported protective effects
of T2D may also be attributed to differences in prostate cancer screening in diabetic and
non-diabetic patients. Differences in health maintenance, access to medical care, and the
81
presence of serious medical conditions may result in more (or less) medical attention and
preventive measures (13). Thus, examining the association between T2D status and
prostate cancer screening frequencies is important to quantify the degree to which
detection bias may explain the apparent relationship.
The incidence rates of T2D and prostate cancer vary widely across populations
however the ethnic disparities for these common diseases are not correlated (i.e. not all
populations with high rates of diabetes are at low risk of prostate cancer) (14-20). The
extent to which these diseases are linked in non-European populations is not clear. To
confirm the previously reported association in a large, representative population-based
prospective study, as well as to examine the consistency of the association across
racial/ethnic populations with differing rates of prostate cancer, we evaluated prostate
cancer incidence by T2D status in a multiethnic sample of 86,303 men from the
Multiethnic Cohort (MEC) (21). We also assessed the presumed effect of diabetes status
in the etiology of prostate cancer by examining PSA levels in a multiethnic sample of
men with and without T2D. We also evaluated PSA screening frequencies in diabetics
and non-diabetics to define the role of screening bias in explaining the observed
association between these common diseases.

82
Methods

Study population
The MEC is a prospective cohort study that includes 215,251 men and women,
the majority from five racial-ethnic groups in Hawaii and Los Angeles (European
Americans, Native Hawaiians, Japanese Americans, and Latinos) (21). Between 1993
and 1996, participants entered the cohort by completing a 26-page, self-administered
questionnaire that asked about diet and demographic factors, personal behaviors (e.g.
physical activity), history of prior medical conditions (e.g. diabetes) and family history of
common cancers. Potential cohort members were identified primarily through
Department of Motor Vehicles drivers’ license files and additionally for African-
Americans, Health Care Financing Administration data files. Participants were between
the ages of 45 and 75 at the time of recruitment.
In the cohort, incident prostate cancer cases are identified annually through cohort
linkage to population-based cancer Surveillance, Epidemiology, and End Results (SEER)
registries in Hawaii and Los Angeles County as well as the California State cancer
registry. Information on stage and grade of disease is also obtained through these
registries. Linkage with these registries is complete through December 31, 2004 in
Hawaii and December 31, 2005 in California. Over this period, 5,941 incident cases of
invasive prostate cancer were identified. Deaths within the cohort are determined from
linkages to the death certificate files in Hawaii and California, supplemented with
linkages to the National Death Index. In the MEC, diabetes status is defined based on
83
self-report on the baseline questionnaire. This question did not differentiate between type
1 diabetes (T1D) and T2D, and thus we expect a small fraction (<10%) of the
respondents to have type 1 diabetes and be potentially misclassified (22).
In addition to self-reported race/ethnicity, the following risk factors were included
in the analysis: body mass index (BMI, weight in kg/height in m
2
), education level (≤12
years, some college or vocational, and college graduate), first degree family history of
prostate cancer, and amount of vigorous physical activity (0, 0-1.5, 1.5-5 and >5
hours/week). Vigorous activity includes both vigorous sports and vigorous work.
We limited our analysis to 91,018 men in the MEC from the 5 major racial/ethnic
groups. We excluded men with a prevalent report of prostate cancer (n=3,004) based on
self-report or from the SEER registries. We excluded men with missing information for
BMI (n=838), education level (n=872) and diabetes status (n=1). The prospective
analysis of the association between diabetes status and prostate cancer incidence in this
study includes 86,303 men.
PSA levels were previously measured on 4,623 men in the MEC (23). These men
were randomly selected from the cohort to evaluate the distribution of PSA levels across
ethnic groups. We excluded 194 men with prevalent prostate cancer at baseline. We also
excluded 1,527 men with incident prostate cancer during the follow-up period, to ensure
that elevated PSA levels among undiagnosed cases did not influence the results. Another
28 men with missing BMI data were excluded from the analysis, leaving 2,874 men who
are included in the final analysis of the effect of T2D on PSA levels.
84
In 2001, we sent a short follow-up questionnaire to cohort members. On this
questionnaire we also asked about PSA screening prior to 1999. Of the 86,303 men
included in the primary analysis of T2D and prostate cancer, 23,768 (27.5%) did not
complete the follow-up questionnaire. We also excluded 4,649 men with incident prostate
cancer. Lastly, we excluded men under the age of 50 (n=10,916) because annual PSA
screening is recommended to begin at the age of 50 (24). This leaves 46,970 men
included in the analysis of the association between T2D and PSA screening.
The informed consent and study protocol were approved by the institutional
review boards at the University of Southern California and the University of Hawaii.

Statistical analysis
Cox regression was used to estimate hazard ratios (reported as relative risks, RR)
for the effect of T2D on prostate cancer incidence (Stata Version 8; Stata Corporation,
College Station, TX). We adjusted for age, BMI, education level, and race/ethnicity (in
pooled analyses). Neither BMI or education level were associated with prostate cancer
risk, but both remained in the model as BMI was found to be associated with PSA levels
and education level was found to be a highly significant predictor of PSA screening.
Physical activity and family history of prostate cancer were left out of the final model
because neither had an effect on the association between T2D and prostate cancer.
Stratified analyses were performed in older age groups to assess whether T2D duration
and long-term exposure to declining insulin levels may be important in prostate cancer
development. Because men may be at increased risk of prostate cancer within the first
85
few years following a diabetes diagnosis as a result of higher insulin levels, and since the
date of T2D diagnosis is unknown for cohort members, we also performed a sensitivity
analysis to examine whether the association might be attenuated by recently diagnosed
diabetics. In this analysis we censored follow-up of incident prostate cancer cases
incrementally by year from one to five years after cohort entry. We also examined the
association in analyses stratified by BMI (≥25 kg/m
2
and <25 kg/m
2
). Analyses stratified
by Gleason score to determine the effect of T2D status on prostate cancer severity were
also conducted. This latter analysis excludes 370 prostate cancer cases with missing
information on Gleason score.
In the analysis of PSA levels, generalized linear models were used to estimate
least-squared mean PSA levels by T2D status (SAS version 9.1, SAS Institute Inc., Cary,
North Carolina).. Models were adjusted for the putative confounders of age, BMI and
race/ethnicity. We calculated PSA screening frequencies adjusted for both age and
education level by T2D status and we tested for a difference using logistic regression;
BMI was not found to influence the effect of T2D on PSA screening. The fraction of the
association between T2D and prostate cancer that may attributable to PSA screening was
estimated. Assuming prostate cancer incidence roughly doubled since the initiation of
PSA screening (25), with about 50% of men being screened, we estimate that incidence
rates have increased by 0.02 per 1% of the population screened. We then used this slope
to estimate the relative impact of screening on prostate cancer incidence in diabetic and
non-diabetic men as follows: RR
PSA
=(1+0.02*screening frequency in non-
diabetics)/(1+0.02*screening frequency in diabetics), with (1-RR
PSA
/1-RR
T2D
) being an
86
estimate of the fraction of the association between T2D and prostate cancer incidence
attributable to PSA screening.

Results

The mean age of the men (n=86,303) in this study was 59.9 years (SD 8.8), and
ranged from 56.6 for Native Hawaiians to 61.3 for African Americans (Table 4-1). Age-
standardized prostate cancer incidence rates (per 100,000) were nearly two times greater
in African Americans (830.2) than in the other populations (Table 4-1). The age-adjusted
prevalence of T2D also varied widely across populations, from 6.9% in European
Americans to 18.0% in Native Hawaiians. The mean age of the diabetic men in our study
at baseline was slightly higher than non-diabetic men for each racial-ethnic group,
ranging from 59.2 (versus 56.1 in non-diabetics) in Native Hawaiians to 63.6 (versus 60.6
in non-diabetics) in Japanese Americans. The age-standardized prostate cancer incidence
rates were lower in diabetic men than in non-diabetic men for each racial-ethnic group,
ranging from 242.8 (per 100,000) in European American diabetics (versus 413.9 in non-
diabetics) to 686.2 in African American diabetics (versus 845.2 in non-diabetics). First
degree family history of prostate cancer was also less common in diabetic men than in
non-diabetic men for each racial-ethnic group, ranging from 5.1% in Latino diabetic men
(5.8% in non-diabetics) to 7.9% in African-American diabetic men (8.8% in non-
diabetics). As expected, in each population, diabetic men were more likely to be over-
weight and less physically active than men without T2D (Table 4-1).

Table 4-1: Descriptive Characteristics by Race/Ethnicity and Diabetes Status in the Multiethnic Cohort (n=86,303), Los
Angeles, CA and Hawaii, 1993-2005
European
Americans
African Americans Native Hawaiians Japanese
Americans
Latinos Total
Diabetes (Yes/No) Y N Y N Y N Y N Y N
No. of Men 1,440 20,606 1,791 9,475 988 5,088 3,160 22,927 3,446 17,382 86,303
Age, mean (SD) 62.8
(8.1)
58.5
(9.0)
63.4
(8.0)
60.9
(8.9)
59.2
(8.0)
56.1
(8.6)
63.6
(8.3)
60.6
(9.2)
61.7
(7.0)
59.7
(7.8)

No. of Prostate Cancer Cases 60 1,193 215 1,295 35 213 160 1,302 198 1,270 5,941
Prostate Cancer Incidence Rates
a
242.8 413.9 686.2 845.2 257.9 373.8 270.1 354.9 304.2 433.7
Family History of Prostate
Cancer(%)
b

7.0 7.9 7.9 8.8 5.6 5.7 6.1 6.2 5.1 5.8
BMI (kg/m
2
), %
b

<23
13.1 21.3 11.1 17.5 7.6 13.8 21.0 26.3 12.3 16.7

23-24.99
8.1 17.9 9.6 14.6 4.7 11.1 17.3 25.1 8.1 10.8

25-29.99
42.7 45.7 44.5 47.8 40.0 45.6 45.1 41.4 49.1 53.8

30-34.99
23.7 11.9 24.4 15.9 28.1 20.3 12.9 6.2 22.8 15.1

≥35
12.3 3.1 10.0 4.3 19.6 9.1 3.7 1.0 7.7 3.5

Education Level, %
b

≤12 years
34.6 23.2 41.8 40.0 63.0 53.8 36.6 34.8 67.4 64.1

Some College or Vocational
29.0 29.1 36.9 37.0 25.8 28.5 34.3 30.5 22.5 23.5

College Graduate
36.4 47.7 20.8 22.9 11.2 17.7 29.1 34.6 10.2 12.3

Physical Activity (hrs/wk), %
b,c

0
39.5 27.9 43.7 34.1 29.7 22.1 39.3 33.0 37.7 28.9

>0-1.5
16.0 14.4 17.5 15.6 13.9 13.4 19.6 18.2 14.3 14.3

>1.5-5
20.8 23.5 15.7 22.0 23.4 25.7 20.9 23.4 18.0 20.9

>5
20.1 31.6 17.2 23.9 29.3 35.7 17.3 23.0 24.4 31.1

a
Adjusted to 1970 US standard population.
b
Age-standardized (5-year age groups) to the total population included in the study.
c
Percentages do not add up to 100% due to missing values.
87
88
In multivariate analyses, men with T2D had significantly lower risk of prostate cancer
than men without T2D (RR=0.81; 95% CI: 0.74, 0.87; P<0.001; Table 4-2). The inverse
association was observed consistently in all five populations, and ranged from 0.65 (95%
CI: 0.50, 0.84) in European Americans to 0.89 (95% CI: 0.77, 1.03) in African
Americans (P for heterogeneity =0.32). We also examined the effect of T2D status on
prostate cancer incidence by age at entry into the cohort as a surrogate for duration of
T2D, as the progressive decline of insulin levels with age among type 2 diabetics has
been suggested to be protective for prostate cancer (1) (3) (6, 8). We found no evidence
that the inverse association was strengthened among older men (Table 4-2). However,
we did observe a slight, yet consistent, decrease in the relative risk when censoring
follow-up of incident cases, by year, within the first five years of follow-up(Appendix
Figure 1). We observed no significant difference in the association when stratified by
BMI (≥25 kg/m
2
: RR=0.78; 95% CI 0.71, 0.86; <25kg/m
2
: RR=0.86; 95% CI 0.74, 1.00;
p for interaction = 0.24). We also observed consistent effects by disease severity
(Gleason score ≤7, n=3,853: RR=0.81; 95% CI: 0.73, 0.90; Gleason score >7, n=1,703:
RR=0.76; 95% CI: 0.65, 0.89).

The association of PSA levels with T2D status and BMI
In the subset of 2,874 men with PSA measurements, diabetic men (n=344) were found to
have significantly lower least square geometric mean PSA levels than non-diabetic men
(n=2,530; 1.04 vs. 1.29 ng/ml; P<0.001). Adjusting for BMI had little effect on this
association (1.07 vs. 1.28 ng/ml; P=0.003; Table 4-3). This association was noted in all

Table 4-2: Relative Risk of Prostate Cancer Associated with Diabetes Status by Age and Gleason Score in the
Multiethnic Cohort, Los Angeles, CA and Hawaii (n=86,303), 1993-2005
European Americans African Americans Native Hawaiians Japanese Americans Latinos All
No. of
cases
RR
a
(95% CI)
No. of
cases
RR
a
(95% CI)
No. of
Cases
RR
a
(95% CI)
No. of
cases
RR
a
(95% CI)
No. of
cases
RR
a
(95% CI)
No. of
cases
RR
a
(95% CI)
All Men 1,253 0.65
(0.50-0.84)
1,510 0.89
(0.77-1.03)
248 0.73
(0.51-1.05)
1,462 0.81
(0.69-0.96)
1,468 0.78
(0.67-0.91)
5,941 0.81
(0.74-0.87)
P value 0.001 0.13 0.09 0.01 0.001 <0.001
P
het
, 0.32

Age≥50 1,192 0.66
(0.51-0.86)
1,436 0.87
(0.75-1.01)
233 0.72
(0.49-1.04)
1,422 0.80
(0.68-0.95)
1,434 0.79
(0.68-0.91)
5,717 0.80
(0.74-0.87)
P value 0.002 0.06 0.08 0.01 0.002 <0.001
P
het
, 0.47

Age≥60 932 0.66
(0.49-0.87)
1,109 0.90
(0.77-1.06)
170 0.76
(0.50-1.14)
1,216 0.81
(0.68-0.97)
1,069 0.77
(0.65-0.92)
4,496 0.81
(0.74-0.89)
P value 0.004 0.22 0.19 0.02 0.003 <0.001
P
het
, 0.25

Age≥70 346 1.03
(0.70-1.49)
371 0.92
(0.70-1.21)
47 0.71
(0.31-1.60)
521 0.80
(0.62-1.05)
293 0.69
(0.49-0.97)
1,578 0.84
(0.72-97)
P value 0.90 0.55 0.40 0.11 0.03 0.02
P
het
, 0.57

Gleason
≤7
753 0.66
(0.47-0.92)
1,075 0.87
(0.73-1.03)
139 0.86
(0.54-1.35)
830 0.78
(0.63-0.98)
1,056 0.80
(0.67-0.95)
3,853 0.81
(0.73-0.90)
P value 0.02 0.11 0.51 0.03 0.01 <0.001
P
het
, 0.60

Gleason
>7
384 0.68
(0.43-1.07)
340 0.95
(0.71-1.29)
95 0.61
(0.32-1.14)
558 0.76
(0.58-1.00)
341 0.68
(0.49-0.94)
1,718 0.76
(0.65-0.89)
P value 0.09 0.76 0.12 0.05 0.02 <0.001
P
het
, 0.54
RR, relative risk; CI, confidence interval:
a
Adjusted for age, BMI, and education level. Adjusted for race in pooled analysis.
89

Table 4-3: Geometric Mean PSA Levels by Ethnicity, Diabetes Status and BMI in the Multiethnic Cohort (n=2,874),
Los Angeles, CA and Hawaii, 1993-2005
Geometric Mean PSA Levels (ng/ml), (n)

European
Americans
African
Americans
Native
Hawaiians
Japanese
Americans Latinos All
e

PSA
(n)
P
value
PSA
(n) P value
PSA
(n)
P
value
P
value
PSA
(n)
P
value
PSA
(n)
P
value P
het

All
Men
a

1.19
(446)
1.50
(916)
1.14
(313)
1.22
(485)
1.26
(714)
1.26
(2,874) <0.001

Diabetes
Status
b

Yes
0.62
(25)
1.46
(126)
1.00
(44)
1.15
(56)
0.99
(93)
1.07
(344)
No
1.21
(421) 0.003
1.64
(790) 0.29
0.97
(269) 0.86
1.26
(429) 0.53
1.27
(621) 0.02
1.28
(2,530) 0.003 0.11

BMI
(kg/m
2
)
c

<25
1.09
(174) Ref.
1.66
(266) Ref.
1.12
(59) Ref.
1.37
(240) Ref.
1.27
(196) Ref.
1.30
(935) Ref.

25-29.9
1.27
(198) 0.18
1.69
(457) 0.84
0.96
(148) 0.26
1.14
(216) 0.03
1.27
(365) 0.97
1.28
(1,384) 0.70

≥30
1.12
(74) 0.84
1.38
(193) 0.07
0.92
(106) 0.18
1.07
(29) 0.18
1.09
(153) 0.16
1.12
(555) 0.009 0.49

BMI
(kg/m
2
)
d

%
Change -0.6% 0.65 -1.8% 0.03 -1.9% 0.06 -2.4% 0.06 -1.6% 0.06 -1.6% <0.001 0.93
a
Adjusted for diabetes status, BMI, and age at blood draw. AA vs each ethnic group, P<.001.
b
Adjusted for BMI and age at blood draw.
c
Adjusted for diabetes status and age at blood draw.
d
Percent change in geometric PSA levels with increase of 1 BMI unit adjusted for diabetes status and age at blood draw.
e
Adjusted for BMI, diabetes status, ethnicity and age at blood draw.
90
91
populations except Native Hawaiians, and was statistically significant in European
Americans (0.62 vs. 1.21, P=0.003) and Latinos (0.99 vs. 1.27, P=0.02). Consistent with
previous reports (11, 12), we also observed an inverse relationship between BMI and
PSA levels (Table 4-3). In ethnic-pooled analyses, compared with men with a BMI < 30
kg/m
2
, men with a BMI ≥ 30 had 13.8% lower mean PSA levels (P=0.009). In
multivariate generalized linear models, adjusted for T2D status and age at blood draw, we
estimated a 1 unit increase in BMI to be associated with a 1.6% decrease in mean PSA
level (P<0.001).

The association of T2D status and education with PSA screening frequencies
In the sample of 46,970 men over 50 with information on PSA screening, 48.2% reported
a PSA screening test prior to 1999. European American men were more likely to report
having been screened (55.8%), while Native Hawaiians (34.3%) were the least likely to
have had PSA testing (P for heterogeneity <0.001). We observed a modest yet highly
statistically significant difference in age and education level standardized PSA screening
frequencies between diabetics (44.7%) and non-diabetics (48.6%, P<0.001; Table 4-4).
The lower PSA screening frequencies among diabetics were noted in all populations
except in African Americans and were statistically significant in Japanese Americans
(P<0.001) and Latinos (P=0.02). Men with higher education levels were much more
likely to have had a PSA test than men with ≤ 12 years of schooling (Table 4-4). We
adjusted for education level as a surrogate for PSA screening in the primary cohort
analyses discussed above, however, it had little impact on the association.

Table 4-4. PSA Screening Frequencies by Level of Education and Diabetes Status in the Multiethnic Cohort (n=46,970),
Los Angeles, CA and Hawaii, 1993-2005

Number of Men with PSA Screening Data
% Screened
European Americans African Americans Native Hawaiians Japanese Americans Latinos All
P value P value P value P value P value P value P
het

All Men
a

12,035
55.8%
4,946
52.1%
2,995
34.3%
16,021
46.3%
10,973
44.7%
46,970
48.2% <0.001
Education Level
a,b

≤12 years
2,747
44.5% Ref.
1,954
46.3% Ref.
1,549
30.1% Ref.
6,247
37.1% Ref.
6,856
41.1% Ref.
19,353
40.0% Ref.
College or
Vocational
3,347
53.6% <0.001
1,819
53.3% <0.001
874
35.6% 0.001
4,846
46.0% <0.001
2,702
50.8% <0.001
13,588
49.1% <0.001
College Graduate
5,941
63.4% <0.001
1,173
59.8% <0.001
572
45.7% <0.001
4,928
56.4% <0.001
1,415
50.5% <0.001
14,029
58.6% <0.001 <0.001
Diabetes Status
c,d

Yes
796
51.4%
742
52.5%
487
35.1%
1,947
42.1%
1,793
44.3%
5,765
44.7%
No
11,239
53.0% 0.46
4,204
52.1% 0.84
2,508
36.4% 0.37
14,074
46.0% <0.001
9,180
47.4% 0.02
41,205
48.6% <0.001 0.45
a
Percentages age-standardized (5-year age groups) to the total population included in the study.
b
P-values are from a logistic regression model and are adjusted for age, education level, and race/ethnicity (pooled analysis).
c
Percentages age (5-year age groups) and education standardized to the total population in the study.
d
P-values are from a logistic regression model and are adjusted for age and race/ethnicity (pooled analysis).
92
93
Next, we examined the potential impact of detection bias on the observed
association between T2D and prostate cancer. Based on the 3.9% difference in PSA
screening frequencies observed between diabetics and non-diabetics we estimated that
detection bias is likely to account for only ~20% of the inverse association between T2D
and prostate cancer risk.

Discussion

In the prospective analysis of five racial/ethnic populations we found a highly
significant association between T2D status and prostate cancer incidence, with diabetics
having a ~20% lower risk of developing prostate cancer. This inverse association was
observed in all populations with the magnitude of the effect being consistent with the
majority of other studies conducted in men of European ancestry (1-8).
In this study, T2D status was based on self-report which may have lead to
misclassification. Previous studies however have shown that self-reported responses for
many common chronic diseases such as diabetes are reliable when compared to medical
records (26-28). The analysis does not account for incident cases of T2D over the 8-year
follow-up period. However, incident cases of diabetes in the non-diabetes group would
make the two groups more similar and create an underestimation of the inverse
association. Another limitation of our study is that we cannot differentiate between cases
of T1D and T2D. While they have similar phenotypes, T1D and T2D have distinct
mechanisms of pathogenesis and may have dissimilar associations with prostate cancer
94
incidence. However, we expect this differential misclassification to be minimal as the
prevalence of T1D is comparatively low in these populations (22).
Diabetes and prostate cancer are both traditionally underdiagnosed diseases. In
this study, undiagnosed T2D would result in prostate cancer incidence rates being more
similar between the diabetic and non-diabetic groups. Undiagnosed cases of prostate
cancer would result in lower rates of prostate cancer among both diabetics and non-
diabetics. As a result, we would expect these simultaneous events of disease
misclassification to counter the inverse association that we noted in this study towards the
null. It is also possible that men who do not receive frequent medical care would be
underdiagnosed and misclassified for both diseases. As a result, the underdiagnosis of
prostate cancer and the lower risk of prostate cancer among the misclassified diabetics
would also result in a bias towards the null of the underlying association. While we
expect that the underdiagnosis of these diseases are unable to explain their inverse
association, future studies demanding regular blood glucose and PSA screening will be
needed to quantify the impact of this bias.
In this study, we also found that men with diabetes are less likely to report PSA
screening than men without diabetes. These findings are contrary to a previous study that
reported that men with diabetes are more likely to undergo screening for prostate cancer
(3). PSA screening frequencies were lower among diabetics in all populations except in
African Americans. Education, which is a surrogate for socioeconomic status and access
to health care, was significantly associated with both PSA screening frequencies and
diabetes status. However, further adjustment for education in the main cohort analyses
95
did not change the results. We estimated that the potential bias incurred by differential
PSA screening (~4%) in diabetics and non-diabetics to explain only ~20% of the
protective effect of T2D on prostate cancer risk. In addition, if the association between
these conditions was influenced by detection bias, then one would expect the inverse
association to diminish among severe cases of prostate cancer since they are likely to
have been diagnosed without the use of PSA screening. However, we observed only a
minimal change in the association between diabetes status and prostate cancer incidence
when stratified by disease severity. Thus, detection bias associated with lower PSA levels
and/or lower PSA screening frequencies in diabetics is unlikely to explain the strong and
highly significant inverse association between T2D and prostate cancer in this study.
Studies have suggested that the protective effect of diabetes on prostate cancer
incidence may be greater among men with long-standing T2D (3, 6, 8). One theory is
that hyperinsulinemia, which is observed at onset, is associated with increased levels of
growth factors (e.g. IGF-I) that may induce prostate cancer during the first few years of
T2D. Subsequently, prostate cancer rates would decrease in the later stages of T2D when
insulin levels decrease and men become hypoinsulinemic. We do not have data on the
date of diagnosis for T2D, but we did analyze the association between T2D status and
prostate cancer incidence by age at entry to the cohort as a surrogate for long-standing
T2D. With both of these theories, one would expect the magnitude of the inverse
association between diabetes and prostate cancer to be greater among older men.
However, we found no difference in the association in older men. We did however notice
a modest increase in the magnitude of inverse association when removing incident cases
96
within the first 5 years of follow-up, which supports the hypothesis that men with newly
diagnosed diabetes may have an increased risk of prostate cancer.
Most, but not all studies have shown that on average, men with diabetes have
lower PSA levels than those without diabetes (11, 12, 29). In our multiethnic sample,
PSA levels were lower in diabetic men. However, what this is an indication of isn’t clear.
Lower PSA levels in diabetics may signal a lower prevalence of prostate cancer and an
indication of a biological effect of T2D status on prostate growth and development. At
the same time, the effect of diabetes status on PSA levels could result in decreased
follow-up for prostate cancer diagnosis among diabetics which may partially account for
the inverse relationship between T2D and prostate cancer risk. Additional work will be
needed to understand whether T2D status influences the accuracy of PSA screening or
directly contributes to prostate cancer risk. Consistent with previous studies (11, 12, 30),
our analysis also suggests an inverse relationship between BMI and PSA levels. Further
studies of this association are necessary to determine if obese men should have lower
PSA thresholds to indicate further work-up for prostate cancer.
Only a small number of studies have investigated the relationship between T2D
and prostate cancer risk in non-European populations (2, 31-33). Most of these studies
have observed non-significant inverse associations, however small sample sizes have
limited interpretation of the findings. Our results, from a population-based prospective
study of over 5,900 prostate cancer cases from five racial/ethnic populations provides
strong support for the pan-ethnic nature of the association between these common
diseases.
97
Recently, a common variant in the HNF1β gene (hepatocyte nuclear factor-1 β)
was found to be associated with an increased risk of prostate cancer. This same variant
was also found to be associated with a decreased risk of T2D (34). Common genetic
variation in another gene, JAZF1, has also been associated with risks of both prostate
cancer and T2D (35, 36). These findings along with other studies that have shown that
diabetes is inversely associated with a family history of prostate cancer (5, 37), which we
also noted, point to both shared genetic risk and common molecular and/or metabolic
pathways that are important in the etiology of these diseases.
In summary, in this large multiethnic prospective study, we observed consistent
inverse associations between T2D and prostate cancer risk across multiple racial/ethnic
populations. These findings provide strong support for the hypothesis that T2D is a
protective factor for prostate cancer. We also confirmed previous studies, showing that
PSA levels are decreased in diabetic men. Our findings that diabetic men are less likely to
be screened for prostate cancer could not account for these results. Future studies aimed
at determining the biological link between diabetes and prostate cancer are warranted and
should focus on common environmental and genetic factors that are shared across
populations.
98
Chapter 4 References
1. Kasper JS, Giovannucci E. A meta-analysis of diabetes mellitus and the risk of
prostate cancer. Cancer Epidemiol Biomarkers Prev 2006;1511:2056-62.

2. Calton BA, Chang SC, Wright ME, et al. History of diabetes mellitus and
subsequent prostate cancer risk in the NIH-AARP Diet and Health Study. Cancer
Causes Control 2007;185:493-503.

3. Giovannucci E, Rimm EB, Stampfer MJ, et al. Diabetes mellitus and risk of
prostate cancer (United States). Cancer Causes Control 1998;91:3-9.

4. Gong Z, Neuhouser ML, Goodman PJ, et al. Obesity, diabetes, and risk of
prostate cancer: results from the prostate cancer prevention trial. Cancer
Epidemiol Biomarkers Prev 2006;1510:1977-83.

5. Velicer CM, Dublin S, White E. Diabetes and the risk of prostate cancer: the role
of diabetes treatment and complications. Prostate Cancer Prostatic Dis
2007;101:46-51.

6. Rodriguez C, Patel AV, Mondul AM, et al. Diabetes and risk of prostate cancer in
a prospective cohort of US men. Am J Epidemiol 2005;1612:147-52.

7. Pierce BL, Plymate S, Ostrander EA, et al. Diabetes mellitus and prostate cancer
risk. Prostate 2008;6810:1126-32.

8. Tavani A, Gallus S, Bosetti C, et al. Diabetes and the risk of prostate cancer. Eur J
Cancer Prev 2002;112:125-8.

9. Giovannucci E, Rimm EB, Stampfer MJ, et al. Height, body weight, and risk of
prostate cancer. Cancer Epidemiol Biomarkers Prev 1997;68:557-63.

10. Huxley R, James WP, Barzi F, et al. Ethnic comparisons of the cross-sectional
relationships between measures of body size with diabetes and hypertension.
Obes Rev 2008;9 Suppl 1:53-61.

11. Fukui M, Tanaka M, Kadono M, et al. Serum prostate-specific antigen levels in
men with type 2 diabetes. Diabetes Care 2008;315:930-1.

12. Werny DM, Saraiya M, Gregg EW. Prostate-specific antigen values in diabetic
and nondiabetic US men, 2001-2002. Am J Epidemiol 2006;16410:978-83.

99
13. Ross LE, Berkowitz Z, Ekwueme DU. Use of the prostate-specific antigen test
among U.S. men: findings from the 2005 National Health Interview Survey.
Cancer Epidemiol Biomarkers Prev 2008;173:636-44.

14. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2009. CA Cancer J Clin
2009;594:225-49.

15. Abate N, Chandalia M. The impact of ethnicity on type 2 diabetes. J Diabetes
Complications 2003;171:39-58.

16. Borrell LN, Crawford ND, Dailo FJ. Race/ethnicity and self-reported diabetes
among adults in the National Health Interview Survey: 2000-2003. Public Health
Rep 2007;1225:616-25.

17. Mau MK, Glanz K, Severino R, et al. Mediators of lifestyle behavior change in
Native Hawaiians: initial findings from the Native Hawaiian Diabetes
Intervention Program. Diabetes Care 2001;2410:1770-5.

18. Aluli NE. Prevalence of obesity in a Native Hawaiian population. Am J Clin Nutr
1991;536 Suppl:1556S-60S.

19. Brawley OW, Jani AB, Master V. Prostate cancer and race. Curr Probl Cancer
2007;313:211-25.

20. Kolonel LN, Altshuler D, Henderson BE. The multiethnic cohort study: exploring
genes, lifestyle and cancer risk. Nat Rev Cancer 2004;47:519-27.

21. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and
Los Angeles: baseline characteristics. Am J Epidemiol 2000;1514:346-57.

22. National Center for Chronic Disease Prevention and Health Promotion. National
Diabetes Fact Sheet. Atlanta GA: Centers for Disease Control and Prevention;
2005. Available at: http://www.cdc.gov/diabetes/pubs/general.htm#what.

23. Cheng I, Yu MC, Koh WP, et al. Comparison of prostate-specific antigen and
hormone levels among men in Singapore and the United States. Cancer Epidemiol
Biomarkers Prev 2005;147:1692-6.

24. Smith RA, Cokkinides V, Eyre HJ. American Cancer Society guidelines for the
early detection of cancer, 2003. CA Cancer J Clin 2003;531:27-43.

25. Collin SM, Martin RM, Metcalfe C, et al. Prostate-cancer mortality in the USA
and UK in 1975-2004: an ecological study. Lancet Oncol 2008;95:445-52.

100
26. Midthjell K, Holmen J, Bjorndal A, et al. Is questionnaire information valid in the
study of a chronic disease such as diabetes? The Nord-Trondelag diabetes study. J
Epidemiol Community Health 1992;465:537-42.

27. Okura Y, Urban LH, Mahoney DW, et al. Agreement between self-report
questionnaires and medical record data was substantial for diabetes, hypertension,
myocardial infarction and stroke but not for heart failure. J Clin Epidemiol
2004;5710:1096-103.

28. Walitt BT, Constantinescu F, Katz JD, et al. Validation of self-report of
rheumatoid arthritis and systemic lupus erythematosus: The Women's Health
Initiative. J Rheumatol 2008;355:811-8.

29. Chan JM, Latini DM, Cowan J, et al. History of diabetes, clinical features of
prostate cancer, and prostate cancer recurrence-data from CaPSURE (United
States). Cancer Causes Control 2005;167:789-97.

30. Fowke JH, Signorello LB, Chang SS, et al. Effects of obesity and height on
prostate-specific antigen (PSA) and percentage of free PSA levels among African-
American and Caucasian men. Cancer 2006;10710:2361-7.

31. Coker AL, Sanderson M, Zheng W, et al. Diabetes mellitus and prostate cancer
risk among older men: population-based case-control study. Br J Cancer
2004;9011:2171-5.

32. Mishina T, Watanabe H, Araki H, et al. Epidemiological study of prostatic cancer
by matched-pair analysis. Prostate 1985;64:423-36.

33. Rosenberg DJ, Neugut AI, Ahsan H, et al. Diabetes mellitus and the risk of
prostate cancer. Cancer Invest 2002;202:157-65.

34. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome
17 confer prostate cancer risk, and the one in TCF2 protects against type 2
diabetes. Nat Genet 2007;398:977-83.

35. Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified in a genome-wide
association study of prostate cancer. Nat Genet 2008;403:310-5.

36. Zeggini E, Scott LJ, Saxena R, et al. Meta-analysis of genome-wide association
data and large-scale replication identifies additional susceptibility loci for type 2
diabetes. Nat Genet 2008;405:638-45.

37. Meyer P, Zuern C, Hermanns N, et al. The association between paternal prostate
cancer and type 2 diabetes. J Carcinog 2007;6:14.
101
Chapter 5: Consistent Association of Type 2 Diabetes Risk Variants
Found in Europeans in Diverse Racial-and Ethnic Groups

Introduction

Multiple common risk alleles have been identified as reproducibly associated with
risk of type 2 diabetes (T2D) (1-13). With the exception of the KCNQ1 locus which was
identified in the Japanese population (12, 13), all of the well-replicated risk variants were
first identified in populations of Northern European ancestry (1-11). T2D morbidity
varies widely across racial/ethnic groups; the prevalence is more than twice as high
among African Americans, Japanese Americans, Latinos and Native Hawaiians as
European Americans (14, 15). It is important to evaluate whether and how genetic
variation may contribute to health disparities between populations. For example, genetic
variation at 8q24 may contribute to population differences in risk of prostate cancer (16,
17), and genetic variation at MYH9 contributes substantially to the higher rates of kidney
disease in African Americans (18).
More generally, it has recently been argued that single rare causal variants and/or
collections of multiple different rare variants on unrelated haplotypes may create
“synthetic associations” of common variants with disease risk (19, 20). One prediction of
this model is that the associations with common variants will not be consistent across
populations (since many of the mutations will be young in age, and post-date the
migrations that led to the founding of modern continental populations). In these papers,
102
type 2 diabetes was specifically called out as a possible case in which synthetic
associations might be operative, based on the lack of statistical significance in very small
studies that examined allelic associations for T2D in multi-ethnic samples.
Testing the association of each validated risk allele for T2D in multiple
populations is an important step to determine (a) whether these genetic markers can be
used to better understand population risk in non-European populations, (b) to measure
their association with racial/ethnic variation in disease risk, and (c) to test a prediction of
the Goldstein “common SNP, rare mutation” hypothesis (19, 20).
To allow for comparability of estimates of genetic risk among racial/ethnic groups
requires large studies comprised of cases and controls defined using identical criteria and
sampled ideally from the same study population. In the present study, we as part of the
Population Architecture using Genomics and Epidemiology (PAGE) Study, examined
genetic associations with 19 validated risk alleles for T2D in African American,
European American, Japanese American, Latino and Native Hawaiian T2D cases
(n=6,142) and controls (n=7,403) from the population-based Multiethnic Cohort Study
(MEC). We also evaluated whether these variants can be utilized to model the genetic
risk of T2D in each population and their association to disparities in risk.

103
Methods

Study Population:

The MEC consists of 215,251 men and women, and comprises mainly five self-
reported racial/ethnic populations: African Americans, Japanese Americans, Latinos,
Native Hawaiians and European Americans (21). Between 1993 and 1996, adults
between 45 and 75 years old were enrolled by completing a 26-page, self-administered
questionnaire asking detailed information about dietary habits, demographic factors, level
of education, personal behaviors, and history of prior medical conditions (e.g. diabetes).
Potential cohort members were identified through Department of Motor Vehicles drivers’
license files, voter registration files and Health Care Financing Administration data files.
In 2001, a short follow-up questionnaire was sent to update information on dietary habits,
as well as to obtain information about new diagnoses of medical conditions since
recruitment. Between 2003 and 2007, we re-administered a modified version of the
baseline questionnaire. All questionnaires inquired about history of diabetes, without
specification as to type (1 vs. 2). Between 1995 and 2004, blood specimens were
collected from ~67,000 MEC participants at which time a short questionnaire was
administered to update certain exposures, and collect current information about
medication use.
Cohort members in California are linked each year to the California Office of
Statewide Health Planning and Development (OSHPD) hospitalization discharge
104
database which consists of mandatory records of all in-patient hospitalizations at most
acute-care facilities in California. Records include information on the principal diagnosis
plus up to 24 other diagnoses (coded according to ICD-9), including T1D and T2D. In
Hawaii cohort members have been linked with the diabetes care registries for subjects
with Hawaii Medical Service Association (HMSA) and Kaiser Permanente Hawaii
(KPH) health plans (~90% of the Hawaiian population has one of these two plans) (15).
Information from these additional databases have been utilized to assess the percentage of
T2D controls (as defined below) with undiagnosed T2D, as well as the percentage of
identified diabetes cases with T1D rather than T2D. Based on the OSHPD database <3%
of T2D cases had a previous diagnosis of T1D. We did not use these sources to identify
T2D cases because they did not include information on diabetes medications, one of our
inclusion criteria for cases (see below).
In this study, diabetic cases were defined using the following criteria: (a) a self-
report of diabetes on the baseline questionnaire, 2
nd
questionnaire or 3
rd
questionnaire;
and (b) self-report of taking medication for T2D at the time of blood draw; and (c) no
diagnosis of T1D in the absence of a T2D diagnosis from the OSHPD (California
Residents). Controls were defined as: (a) no self-report of diabetes on any of the
questionnaires while having completed a minimum of 2 of the 3 (79% of controls
returned all 3 questionnaires); and (b) no use of medications for T2D at the time of blood
draw; and (c) no diabetes diagnosis (type 1 or 2) from the OSHPD, HMSA or KPH
registries. To preserve DNA for genetic studies of cancer in the MEC, subjects with an
incident cancer diagnosis at time of selection for this study were excluded. Controls were
105
frequency matched to cases on age at entry into the cohort (5-year age groups) and for
Latinos, place of birth (U.S. vs. Mexico, South or Central America), oversampling
African American, Native Hawaiian and European American controls to increase
statistical power.
Fasting glucose (FG) and HbA
1C
measurements were used to validate the case-
control selection criteria. Among 185 T2D cases and 1,048 controls who met the T2D
case-control definitions above and with FG measurements available from ongoing studies
in the MEC, 57% of cases (ranging from 43% in European Americans to 63% in Japanese
Americans) and 3% of controls (ranging from 1% in African Americans to 6% in
Latinos) had a FG value >125 mg/dl. We also measured HbA
1C
(ARUP Laboratories, Salt
Lake City, Utah) in 50 cases and 50 controls per each sex-ethnic group. Just over 1%
(6/500) of controls were likely to have unreported T2D (HbA
1C
value ≥7%). In contrast,
~47% (234/500) of T2D cases had HbA
1C
≥7% (ranging from 41% in European
Americans to 57% of Native Hawaiians). Since hypoglycemic medication use was part of
the case selection criteria, some cases were expected to have FG and HbA
1C
levels in the
normal range.
Altogether, this study included 6,142 T2D cases and 7,403 controls (African
American (1,077/1,469), European American (533/1,006), Latino (2,220/2,184), Japanese
American (1,736/1,761) and Native Hawaiian (576/983)). Genotyping was conducted by
the TaqMan allelic discrimination assay (Applied Biosystems, Foster City, CA) (22). For
all SNPs, genotype call rates were >95% among case and control groups in each
population and HWE p-values were >0.05 among controls in at least 4 of the 5 ethnic
106
groups and none of the values were <0.01 (Table 5-1). Subjects missing data for >5 SNPs
(n=82) were removed from the analysis.

Statistical Analysis
Odds ratios and 95% confidence intervals were calculated for each allele in
unconditional

logistic regression models while adjusting for age at cohort entry
(quartiles), body mass index (BMI, kg/m
2
, quartiles), and sex in ethnic-stratified analyses.
Associations with the two variants at KCNQ1 were examined adjusting for the other
allele. Potential confounding factors including, smoking history, education, physical
activity, and history of hypertension were evaluated but did not influence the results.
Potential confounding by percent European ancestry was examined in a subset of African
American men (336 cases, 397 controls) with available genetic ancestry information (16,
23, 24).
We also modeled the cumulative genetic risk of T2D using these markers. We
summed the number of risk alleles for each individual and estimated the odds ratio per
allele for this aggregate unweighted allele count variable as an approximate risk score
appropriate for unlinked variants with independent effects of approximately the same
magnitude for each allele. We also examined a second model where each allele was
weighted and multiplied by the log of the published odds ratio prior to summing the total
alleles. The two risk scores were highly correlated in each ethnic group (Pearson r≥0.92)
and the results for the more parsimonious unweighted risk score are presented. For
individuals missing genotypes for a given SNP, we assigned the average number of risk
107
Table 5-1: Genotyping Efficiency: Genotype Call Rates, Hardy-Weinberg
Equilibrium Testing

Genotype Call Rates (Cases/Controls)
HWE p value (1 df) (Cases/Controls)
SNP
European
Americans
533 cases
1,006 controls
African
Americans
1,077 cases
1,469 controls
Latinos

2,220 cases
2,184 controls
Japanese
Americans
1,736 cases
1,761 controls
Native
Hawaiians
576 cases
983 controls
rs10923931 100%/99.3%
0.92/0.72
98.8%/98.8%
0.40/0.034
99.7%/99.6%
0.64/0.83
99.8%/99.8%
0.87/0.40
99.5%/99.5%
0.59/0.78
rs7578597 99,2%/100%
0.47/0.15
99.7%/99.5%
0.39/0.55
99.8%/99.5%
0.58/0.18
99.8%/99.4%
0.048/0.066
99.8%/99.7%
0.18/0.27
rs1801282 100%/99.9%
0.084/0.20
99.7%/99.5%
0.20/0.70
99.5%/99.5%
0.41/0.99
99.9%/99.6%
0.43/0.99
99.1%/99.6%
0.084/0.091
rs4607103 99.2%/99.7%
0.82/0.68
99.7%/99.5%
0.40/0.48
99.5%/99.5%
0.10/0.44
99.8%/99.3%
0.71/0.72
99.7%/99.4%
0.14/0.92
rs4402960 99.1%/98.8%
0.80/0.20
99.4%/98.9%
0.16/0.79
99.7%/99.4%
0.35/0.10
99.4%/99.0%
0.33/0.39
99.3%/98.6%
0.23/0.39
rs10010131 98.9%/99.5%
0.87/0.41
99.8%/99.3%
0.60/0.45
99.7%/99.7%
0.61/0.97
99.8%/99.7%
0.013/0.40
99.8%/99.8%
0.28/0.71
rs7754840 97.9%/97.8%
0.30/0.24
99.7%/99.4%
0.15/0.99
99.8%/99.4%
0.62/0.31
96.9%/98.2%
0.074/0.30
98.1%/98.3%
0.60/0.15
rs864745 97.6%/99.1%
0.89/0.98
98.4%/99.0%
0.62/0.024
98.5%/98.9%
0.88/0.13
98.6%/98.4%
0.39/0.24
98.4%/97.5%
0.59/0.71
rs13266634 100%/99.1%
0.97/0.045
99.8%/99.5%
0.88/0.87
99.8%/99.5%
0.64/0.61
99.7%/99.5%
0.66/0.37
99.8%/98.8%
0.49/0.082
rs2383208 96.2%/95.2%
0.96/0.76
99.4%/98.5%
0.81/0.97
99.6%/99.4%
0.45/0.64
98.0%/97.1%
0.58/0.53
97.9%/98.3%
0.54/0.67
rs1111875 98.9%/99.2%
0.13/0.19
99.0%/98.6%
0.48/0.94
99.0%/99.7%
0.66/0.79
99.4%/99.5%
0.51/0.74
99.7%/99.5%
0.94/0.90
rs7903146 96.6%/99.3%
0.44/0.62
99.3%/98.4%
0.19/0.28
99.7%/98.9%
0.10/0.19
99.2%/98.8%
0.87/0.25
98.8%/99.0%
0.16/0.14
rs12779790 96.8%/98.3%
0.87/0.87
99.0%/99.3%
0.094/0.25
98.8%/99.3%
0.73/0.75
97.7%/97.2%
0.19/0.46
99.0%/98.8%
0.31/0.19
rs2237895 99.2%/99.4%
0.86/0.67
99.4%/98.8%
0.18/0.80
99.7%/99.8%
0.26/0.76
98.4%/99.4%
0.23/0.71
98.1%/99.4%
0.19/0.86
rs2237897 99.4%/98.8%
0.81/0.050
99.6%/99.0%
0.26/0.91
99.3%/99.4%
0.73/0.26
98.4%/98.8%
0.67/0.62
98.4%/97.5%
0.56/0.75
rs5219 98.3%/98.2%
0.062/0.84
99.1%/99.0%
0.79/0.27
98.6%/99.0%
0.99/0.81
98.6%/98.6%
0.27/0.65
99.7%/99.0%
0.19/0.70
rs7961581 98.5%/98.5%
0.30/0.29
99.1%/99.3%
0.58/0.35
99.5%/99.5%
0.71/0.89
98.7%/98.8%
0.35/0.24
99.7%/99.2%
0.91/0.86
rs8050136 97.9%/99.4%
0.32/0.99
99.7%/99.4%
0.12/0.59
99.7%/99.6%
0.92/0.24
99.3%/99.2%
0.30/0.38
98.8%/98.7%
0.60/0.16
rs4430796 97.7%/99.1%
0.27/0.64
99.5%/98.4%
0.53/0.95
99.3%/98.9%
0.13/0.026
99.5%/99.1%
0.53/0.66
99.8%/99.1%
0.93/0.084

108
alleles within each ethnic group (2 x risk allele frequency) to replace the missing value
for that SNP. We used these ethnic-specific per allele summary odds ratios and the total
number of risk alleles among control subjects to estimate the distribution of relative risks
conveyed by all risk alleles. To avoid making the reference group carriers of zero risk
alleles (a group which does not exist) we centered the distribution on the mean number of
risk alleles observed in the control population (18.5). The log relative risk for each
subject was calculated as logRR = (RA – 18.5) x log(OR
i
) (where RA is equal to the
subject’s total risk alleles and log(OR
i
) is the log of the ethnic specific per allele odds
ratio. A spline function was used to capture the shape of the distributions of log OR for
display purposes. Two variants in KCNQ1 were included in the risk modeling because
both were significantly associated with T2D when co-modeled (results were similar when
only the most significant of the two, rs2237897, was included). The variant in FTO was
excluded from risk modeling procedures, as we found (as have others) that it is not a risk
factor for diabetes independent of its effect on obesity.

Results

The age of the cases and controls ranged from 45 to 77 at cohort entry, with cases
(mean 59.0 years) being essentially the same as the controls (mean 58.8 years), and
African-Americans being on average the oldest (mean 60.2 years) and Native Hawaiians
the youngest (mean 55.6 years). Compared to controls, cases were heavier, more likely to
be a current or former smoker, less physically active and had fewer years of education
109
(Table 5-2). Compared to the other groups, the Japanese were leaner (for cases and
controls, men and women).
The established T2D risk SNPs were polymorphic (frequency>0.05) in all
racial/ethnic groups, except for rs10923931 (NOTCH2) in Japanese and Native
Hawaiians and rs7903146 (TCF7L2) in Japanese (Figure 3-1). In European populations
these 19 SNPs have very modest odds ratios (1.1-1.3 per copy of the risk allele), and
required studies of more than ten thousand cases and controls to reach genome-wide
significance (1-13). Our sample sizes, although substantial, provided limited power to
detect these modest effects (Table 5-3; power to achieve nominal significance (P=0.05)
of 34%, 47%, 67%, 54%, and 33%, in European Americans, African Americans, Latinos,
Japanese Americans, and Native Hawaiians, respectively).
We first assessed whether the “risk allele” of each SNP was associated in the
same direction (odds ratios>1) in each ethnic group. Whereas the null hypothesis is that
50% of “risk” alleles would trend in the same direction, we observed from 12 (63%;
P=0.18; binomial probability) in European Americans to 19 (100%; P=1.9x10
-6
) in
Japanese. The number of these associations that reached nominal significance (P<0.05)
ranged from 3 (P=0.067; binomial probability) in Native Hawaiians to 10 (P=5.9x10
-9
) in
Japanese (Table 5-4). For the majority of alleles with positive associations, odds ratios
for homozygotes carriers were greater than for heterozygous carriers in each population,
which provides support for their associations and allele dosage effects (Table 5-5). In
African Americans, results were similar after adjustment for percent European

Table 5-2: The Descriptive Characteristics of Type 2 Diabetes Cases and Controls in the MEC at Baseline by
Racial/Ethnic Group and Sex
Characteristic European Americans African Americans Latinos Japanese Americans Native Hawaiians
Cases Controls Cases Controls Cases Controls Cases Controls Cases Controls
Men
N 288 504 408 634 1,063 1,051 978 986 258 493
Age (yr, mean ± SD) 57.6 ± 8.1 57.5 ± 8.0 62.6 ± 7.9 61.9 ± 7.9 59.7 ± 6.9 59.7 ± 6.9 59.0 ± 8.2 59.1 ± 8.2 55.5 ± 7.0 55.6 ± 7.0
BMI (kg/m
2
, mean ± SD) 29.4 ± 4.7 25.1 ± 3.1 28.9 ± 4.5 26.2 ± 3.8 28.4 ± 4.2 26.2 ± 3.3 26.9 ± 4.0 24.5 ± 3.1 31.0 ± 4.8 28.0 ± 5.2
Weight (lbs, mean ± SD) 209 ± 38 180 ± 25 206 ± 37 187 ± 30 188 ± 31 174 ± 25 172 ± 30 156 ± 23 215 ± 38 194 ± 40
Former/current smoker (%) 77.7 60.1 75.7 73.5 72.4 63.3 72.3 67.8 70.9 66.1
Education (≤ 12 years, %)

27.9 11.1 39.6 39.7 59.3 55.9 29.1 26.0 48.8 40.6
Physical activity (0 hrs / wk,
%)
a
30.9 21.0 45.8 38.3 34.6 29.5 35.3 30.9 24.4 19.3
Women
N 245 502 669 835 1,157 1,133 758 775 318 490
Age (yr, mean ± SD) 58.1 ± 8.3 58.3 ± 8.3 59.1 ± 8.6 58.6 ± 8.6 59.0 ± 7.1 59.1 ± 7.0 59.4 ± 8.5 59.5 ± 8.4 55.7 ± 7.6 55.7 ± 7.6
BMI (kg/m
2
, mean ± SD) 30.2 ± 6.2 24.4 ± 4.5 31.8 ± 6.3 27.6 ± 5.7 30.0 ± 5.9 26.2 ± 4.4 26.6 ± 4.6 22.8 ± 3.5 31.4 ± 6.7 26.4 ± 5.2
Weight (lbs, mean ± SD) 179 ± 39 147 ± 28 191 ± 39 166 ± 35 170 ± 34 148 ± 26 145 ± 27 124 ± 21 186 ± 42 155 ± 33
Former or current smoker
(%)
45.7 48.6 51.9 51.3 36.3 31.0 35.6 30.2 51.3 47.1
Education (≤ 12 years, %)

30.2 20.9 37.1 30.5 74.9 68.5 34.7 31.9 51.4 46.0
Physical activity (0 hrs / wk,
%)
a
55.1 44.2 64.6 58.6 66.8 59.7 64.3 61.9 50.3 37.1
a
Hours per week of vigorous activity of both work and sports.
110

Figure 5-1: Risk Allele Frequencies by Racial/Ethnic Group

Figure 1. Risk allele frequencies for each variant in European Americans (Blue), African Americans (Yellow), Latinos (Purple), Japanese Americans (Red), and Native
Hawaiians (Green). The order of the variants is based on the frequency of the risk allele in European Americans (high to low).
111

Table 5-3: Power Estimates (α=0.05) to Detect Relative Risks in Previous Studies

European
Americans
533 cases
1,006 controls
African Americans
1,077 cases
1,469 controls
Latinos
2,220 cases
2,184 controls
Japanese
Americans
1,736 cases
1,761 controls
Native Hawaiians
576 cases
983 controls
Pooled
6,142 cases
7,403 controls
SNP
Published
odds
ratios
a
RAF
Estimated
Power
RAF
Estimated
Power
RAF
Estimated
Power
RAF
Estimated
Power
RAF
Estimated
Power
RAF
Estimated
Power
rs10923931 1.13 0.12 19% 0.29 51% 0.09 39% 0.02 11% 0.05 11% 0.11 89%
rs7578597 1.15 0.90 19% 0.75 55% 0.94 33% 0.99 9% 0.97 10% 0.91 89%
rs1801282 1.14 0.89 18% 0.97 12% 0.90 44% 0.96 18% 0.93 14% 0.93 76%
rs4607103 1.09 0.73 17% 0.70 28% 0.69 46% 0.61 42% 0.72 18% 0.68 90%
rs4402960 1.14 0.31 37% 0.49 64% 0.27 79% 0.30 72% 0.27 36% 0.33 99%
rs10010131 1.11 0.59 27% 0.66 41% 0.71 59% 0.98 9% 0.81 19% 0.76 95%
rs7754840 1.12 0.28 28% 0.55 51% 0.31 70% 0.40 65% 0.52 33% 0.40 99%
rs864745 1.10 0.51 24% 0.73 31% 0.61 58% 0.77 38% 0.75 20% 0.68 95%
rs13266634 1.12 0.68 28% 0.89 23% 0.75 62% 0.60 64% 0.62 31% 0.72 98%
rs2383208 1.20 0.81 45% 0.81 69% 0.85 84% 0.56 96% 0.74 56% 0.75 99%
rs1111875 1.13 0.61 35% 0.74 46% 0.63 78% 0.28 64% 0.28 32% 0.52 99%
rs7903146 1.37 0.27 97% 0.28 99% 0.23 99% 0.04 79% 0.14 88% 0.19 99%
rs12779790 1.11 0.17 18% 0.14 26% 0.17 46% 0.17 38% 0.18 20% 0.17 90%
rs2237895 1.23 0.42 77% 0.20 85% 0.40 99% 0.35 99% 0.33 76% 0.34 99%
rs2237897
b
1.19 0.95 16% 0.92 36% 0.76 93% 0.62 94% 0.78 47% 0.79 99%
rs5219 1.14 0.35 38% 0.09 27% 0.37 85% 0.35 75% 0.37 40% 0.31 99%
rs7961581 1.09 0.28 18% 0.23 25% 0.21 39% 0.21 32% 0.29 19% 0.23 85%
rs8050136 1.17 0.41 54% 0.43 79% 0.27 91% 0.20 77% 0.23 44% 0.30 99%
rs4430796 1.10 0.50 24% 0.65 36% 0.42 60% 0.36 49% 0.31 22% 0.45 97%
a
Odds ratio estimates from previous scans in European populations.(2, 3, 5, 10, 11, 25).
b
Odds ratio estimate for rs2237897 from personal communication (Ben Voight, DIAGRAM consortium).
112

Table 5-4: The Association of Known Risk Alleles for T2D by Race/Ethnicity
ab
SNP /
Allele
Tested
b

Gene
Chr
European
Americans
533 cases
1,006 controls
African
Americans
1,077 cases
1,469 controls
Latinos
2,220 cases
2,184 controls
Japanese
Americans
1,736 cases
1,761 controls
Native
Hawaiians
576 cases
983 controls
Pooled
6,142 cases
7,403 controls P value P
het
c
rs10923931
T
NOTCH2
1
0.84 (0.64-1.10)
0.12
1.10 (0.97-1.25)
0.29
1.17 (1.01-1.35)
0.09
1.00 (0.70-1.44)
0.02
0.76 (0.50-1.15)
0.05
1.06 (0.98-1.16)
0.11 0.16 0.086
rs7578597
T
THADA
2
1.42 (1.06-1.91)
0.90
1.04 (0.90-1.19)
0.75
1.12 (0.93-1.35)
0.94
1.10 (0.68-1.78)
0.99
1.65 (1.01-2.70)
0.97
1.13 (1.02-1.25)
0.91 0.016 0.21
rs1801282
C
PPARG
3
1.26 (0.95-1.68)
0.89
1.95 (1.30-2.92)
0.97
1.03 (0.89-1.19)
0.90
1.05 (0.82-1.36)
0.96
1.14 (0.84-1.55)
0.93
1.13 (1.02-1.26)
0.93 0.018 0.048
rs4607103

C
ADAMTS9
3
1.11 (0.92-1.34)
0.73
1.03 (0.91-1.18)
0.70
0.99 (0.89-1.08)
0.69
1.05 (0.94-1.17)
0.61
0.98 (0.82-1.16)
0.72
1.02 (0.97-1.08)
0.68 0.45 0.78
rs4402960
T
IGF2BP2
3
1.02 (0.85-1.23)
0.31
1.14 (1.01-1.28)
0.49
1.05 (0.95-1.16)
0.27
1.24 (1.11-1.38)
0.30
1.17 (0.99-1.39)
0.27
1.13 (1.07-1.19)
0.33 2.2 x 10
-5
0.26
rs10010131
G
WFS1
4
1.18 (0.99-1.40)
0.59
0.94 (0.83-1.07)
0.66
1.15 (1.04-1.27)
0.71
1.45 (0.98-2.13)
0.98
1.27 (1.03-1.56)
0.81
1.11 (1.04-1.19)
0.76 1.8 x 10
-3
0.032
rs7754840
C
CDKAL1
6
1.26 (1.05-1.50)
0.29
1.03 (0.92-1.16)
0.55
1.09 (0.99 -1.19)
0.31
1.37 (1.24-1.52)
0.40
1.39 (1.19-1.62)
0.52
1.20 (1.13-1.26)
0.40
4.1 x 10
-
11
6.2 x 10
-4
rs864745
T
JAZF1
7
0.98 (0.83-1.16)
0.51
1.16 (1.01-1.32)
0.73
1.30 (1.19-1.43)
0.61
1.19 (1.05-1.35)
0.77
1.13 (0.93-1.36)
0.75
1.20 (1.13-1.27)
0.68 1.3 x 10
-9
0.054
rs13266634
C
SLC30A8
8
1.28 (1.06-1.54)
0.68
1.21 (0.99-1.48)
0.89
1.11 (1.00-1.23)
0.75
1.18 (1.06-1.31)
0.60
1.04 (0.88-1.22)
0.62
1.15 (1.08-1.22)
0.72 9.7 x 10
-6
0.47
rs2383208

A
CDKN2B
9
1.35 (1.07-1.69)
0.81
1.14 (0.98-1.33)
0.81
1.15 (1.01-1.31)
0.85
1.26 (1.13-1.40)
0.56
1.02 (0.85-1.22)
0.74
1.18 (1.11-1.26)
0.75 2.1 x 10
-7
0.27
rs1111875
C
HHEX
10
0.93 (0.78-1.11)
0.61
1.10 (0.95-1.26)
0.74
1.03 (0.94-1.13)
0.63
1.21 (1.08-1.36)
0.28
0.93 (0.78-1.11)
0.28
1.07 (1.01-1.13)
0.52 0.028 0.037

113

Table 5-4: Continued

SNP /
Allele
Tested
b

Chr
Gene
European
Americans
533 cases
1,006 controls
African
Americans
1,077 cases
1,469 controls
Latinos
2,220 cases
2,184 controls
Japanese
Americans
1,736 cases
1,761 controls
Native
Hawaiians
576 cases
983 controls
Pooled
6,142 cases
7,403 controls P value P
het
c
rs7903146
T
TCF7L2
10
1.55 (1.29-1.87)
0.27
1.32 (1.16-1.51)
0.28
1.31 (1.19-1.45)
0.23
1.74 (1.38-2.20)
0.04
1.12 (0.90-1.40)
0.14
1.36 (1.27-1.45)
0.19 1.1 x 10
-19
0.068
rs12779790
G
CDC123
10
1.02 (0.82-1.28)
0.17
1.09 (0.92-1.29)
0.14
1.19 (1.06-1.33)
0.17
1.01 (0.88-1.15)
0.17
1.16 (0.95-1.41)
0.18
1.10 (1.03-1.18)
0.17 5.9 x 10
-3
0.36
rs2237895
d
C

KCNQ1
11
0.98 (0.82-1.16)
0.42
1.04 (0.90-1.21)
0.20
1.15 (1.05-1.28)
0.40
1.12 (0.98-1.27)
0.35
1.16 (0.97-1.39)
0.33
1.11 (1.04-1.18)
0.34 7.8 x 10
-4
0.17
rs2237897
d
C
KCNQ1
11
0.86 (0.59-1.26)
0.95
1.13 (0.90-1.42)
0.92
1.23 (1.09-1.39)
0.76
1.26 (1.11-1.44)
0.62
1.06 (0.86-1.31)
0.78
1.21 (1.13-1.30)
0.79 3.2 x 10
-7
0.10
rs5219
T
KCNJ11
11
1.24 (1.05-1.47)
0.35
1.03 (0.84-1.26)
0.09
1.09 (1.00-1.20)
0.37
1.26 (1.13-1.40)
0.35
1.03 (0.88-1.21)
0.37
1.15 (1.08-1.21)
0.31 3.3 x 10
-6
0.13
rs7961581
C
TSPAN8
12
1.03 (0.86-1.24)
0.29
0.92 (0.80-1.06)
0.23
1.03 (0.93-1.15)
0.21
1.00 (0.88-1.14)
0.21
1.12 (0.94-1.33)
0.29
1.01 (0.95-1.08)
0.23 0.71 0.51
rs8050136
A
FTO
16
0.89 (0.75-1.06)
0.41
1.07 (0.95-1.20)
0.43
1.02 (0.92-1.12)
0.27
1.04 (0.92-1.18)
0.20
1.01 (0.84-1.21)
0.23
1.02 (0.96-1.08)
0.30 0.48 0.68
rs4430796
G
HNF1B
17
0.96 (0.82-1.14)
0.50
1.10 (0.97-1.25)
0.65
0.96 (0.88-1.05)
0.42
1.18 (1.06-1.32)
0.36
1.09 (0.92-1.30)
0.31
1.05 (1.00-1.11)
0.45 0.058 0.043
a
Odds ratios (and 95% confidence intervals) for allele dosage effects are adjusted for age (quartiles), BMI (quartiles), sex, and ethnicity (in pooled analysis).
b
Ethnic specific risk allele frequencies (RAF) calculated for each SNP.
c
NCBI build 36 (forward strand).
d
P
het
= P value for heterogeneity of allelic effects across ethnic groups (4 df test).
e
rs2237895 and rs2237897 adjusted for each other.
114

Table 5-5: Association with T2D Risk by Genotype
a
European Americans African Americans Latinos Japanese Americans Native Hawaiians Pooled
SNP
Risk Allele Het Hom Het Hom Het Hom Het Hom Het Hom Het Hom
rs10923931
T
0.79
(0.59-1.07)
1.03
(0.37-2.91)
1.24
(1.04-1.48)
1.05
(0.78-1.40)
1.20
(1.02-1.41)
1.12
(0.59-2.11)
0.98
(0.68-1.41) NA
0.76
(0.49-1.16)
0.64
(0.05-8.45)
1.10
(1.00-1.22)
1.00
(0.78-1.29)
rs7578597
T
1.94
(0.58-6.53)
2.64
(0.82-8.58)
1.06
(0.72-1.58)
1.09
(0.75-1.58)
0.52
(0.16-1.66)
0.60
(0.19-1.91)
1.05
(0.06-17.7)
1.16
(0.07-18.7) NA NA
1.00
(0.72-1.40)
1.16
(0.83-1.61)
rs1801282
C
4.84
(0.60-39.0)
5.61
(0.71-44.6)
0.79
(0.06-10.8)
1.60
(0.12-21.1)
0.86
(0.46-1.59)
0.90
(0.49-1.65)
0.50
(0.10-2.44)
0.54
(0.11-2.63) NA NA
1.20
(0.74-1.96)
1.36
(0.84-2.19)
rs4607103

C
1.32
(0.80-2.19)
1.38
(0.84-2.25)
1.12
(0.81-1.54)
1.12
(0.81-1.54)
1.03
(0.82-1.28)
0.99
(0.79-1.24)
1.03
(0.83-1.28)
1.09
(0.87-1.36)
0.78
(0.51-1.19)
0.83
(0.55-1.26)
1.03
(0.91-1.18)
1.05
(0.92-1.20)
rs4402960
T
0.93
(0.72-1.20)
1.17
(0.77-1.77)
1.25
(1.01-1.53)
1.29
(1.02-1.64)
1.08
(0.94-1.23)
1.07
(0.85-1.36)
1.26
(1.08-1.47)
1.49
(1.16-1.92)
1.16
(0.92-1.47)
1.39
(0.93-2.06)
1.15
(1.06-1.24)
1.26
(1.11-1.42)
rs10010131
G
1.44
(1.01-2.07)
1.50
(1.04-2.16)
0.98
(0.75-1.28)
0.90
(0.69-1.19)
1.05
(0.82-1.34)
1.24
(0.98-1.58) NA NA
0.79
(0.41-1.52)
1.11
(0.59-2.09)
1.05
(0.90-1.22)
1.19
(1.02-1.39)
rs7754840
C
1.20
(0.93-1.55)
1.66
(1.11-2.48)
0.88
(0.70-1.10)
1.03
(0.81-1.31)
1.11
(0.97-1.26)
1.16
(0.93-1.43)
1.33
(1.12-1.56)
1.91
(1.55-2.35)
1.56
(1.15-2.12)
1.98
(1.44-2.74)
1.17
(1.07-1.27)
1.44
(1.30-1.61)
rs864745
T
0.99
(0.74-1.34)
0.97
(0.69-1.36)
1.29
(0.92-1.82)
1.43
(1.03-2.00)
1.18
(0.97-1.43)
1.62
(1.33-1.98)
0.91
(0.63-1.30)
1.16
(0.82-1.65)
1.06
(0.63-1.78)
1.21
(0.73-2.01)
1.11
(0.98-1.27)
1.38
(1.21-1.57)
rs13266634
C
1.08
(0.68-1.72)
1.47
(0.93-2.32)
1.08
(0.44-2.69)
1.33
(0.54-3.25)
1.07
(0.81-1.42)
1.20
(0.91-1.59)
1.11
(0.89-1.38)
1.36
(1.09-1.70)
1.33
(0.95-1.86)
1.18
(0.84-1.66)
1.12
(0.97-1.29)
1.29
(1.12-1.49)
rs2383208

A
1.65
(0.78-3.49)
2.13
(1.03-4.40)
1.15
(0.71-1.86)
1.31
(0.82-2.09)
1.30
(0.80-2.11)
1.47
(0.91-2.35)
1.26
(1.03-1.54)
1.58
(1.28-1.96)
1.14
(0.71-1.83)
1.11
(0.70-1.77)
1.26
(1.08-1.48)
1.46
(1.25-1.70)
rs1111875
C
1.07
(0.75-1.52)
0.91
(0.63-1.31)
1.17
(0.80-1.70)
1.25
(0.87-1.81)
0.98
(0.81-1.19)
1.04
(0.86-1.27)
1.25
(1.07-1.45)
1.41
(1.08-1.83)
0.96
(0.76-1.21)
0.84
(0.54-1.30)
1.11
(1.00-1.22)
1.14
(1.02-1.27)

115

Table 5-5: Continued
European Americans African Americans Latinos Japanese Americans Native Hawaiians Pooled
SNP
Risk Allele Het Hom Het Hom Het Hom Het Hom Het Hom Het Hom
rs7903146
T
1.55
(1.20-2.00)
2.43
(1.61-3.67)
1.33
(1.11-1.58)
1.73
(1.27-2.36)
1.30
(1.14-1.48)
1.76
(1.37-2.27)
1.79
(1.39-2.29)
2.11
(0.60-7.40)
1.10
(0.85-1.43)
1.36
(0.67-2.74)
1.36
(1.25-1.48)
1.84
(1.56-2.18)
rs12779790
G
1.06
(0.81-1.39)
0.90
(0.44-1.86)
0.97
(0.80-1.17)
2.07
(1.16-3.72)
1.19
(1.04-1.36)
1.40
(0.98-2.01)
0.98
(0.83-1.15)
1.15
(0.74-1.77)
1.18
(0.92-1.51)
1.29
(0.74-2.25)
1.07
(0.99-1.17)
1.33
(1.07-1.66)
rs2237895
b
C
1.11
(0.85-1.45)
0.90
(0.63-1.29)
1.01
(0.84-1.22)
1.18
(0.78-1.78)
1.22
(1.05-1.42)
1.30
(1.06-1.60)
1.09
(0.91-1.30)
1.27
(0.97-1.66)
1.12
(0.87-1.44)
1.39
(0.94-2.04)
1.11
(1.02-1.21)
1.22
(1.08-1.39)
rs2237897
b
C
1.89
(0.28-12.7)
1.52
(0.23-9.85)
1.79
(0.44-7.19)
1.97
(0.50-7.81)
1.38
(1.01-1.89)
1.65
(1.20-2.27)
1.19
(0.92-1.54)
1.55
(1.17-2.05)
1.05
(0.60-1.83)
1.12
(0.65-1.95)
1.28
(1.07-1.53)
1.52
(1.27-1.83)
rs5219
T
1.11
(0.86-1.44)
1.64
(1.15-2.36)
1.06
(0.85-1.34)
0.80
(0.32-1.96)
1.10
(0.96-1.26)
1.20
(0.99-1.45)
1.38
(1.18-1.62)
1.49
(1.18-1.87)
0.94
(0.74-1.19)
1.12
(0.80-1.57)
1.15
(1.06-1.24)
1.31
(1.16-1.48)
rs7961581
C
1.15
(0.89-1.48)
0.94
(0.61-1.44)
0.93
(0.78-1.12)
0.81
(0.55-1.19)
1.01
(0.88-1.15)
1.15
(0.84-1.56)
0.93
(0.80-1.09)
1.24
(0.86-1.78)
1.09
(0.86-1.38)
1.31
(0.86-1.98)
0.99
(0.92-1.07)
1.07
(0.91-1.26)
rs8050136
A
0.98
(0.75-1.28)
0.75
(0.52-1.09)
0.97
(0.80-1.18)
1.16
(0.91-1.47)
1.07
(0.93-1.21)
0.96
(0.76-1.22)
1.02
(0.87-1.19)
1.14
(0.80-1.63)
1.01
(0.79-1.28)
1.02
(0.64-1.63)
1.03
(0.95-1.11)
1.04
(0.91-1.18)
rs4430796
G
0.83
(0.62-1.11)
0.93
(0.67-1.29)
1.26
(0.95-1.66)
1.29
(0.98-1.71)
0.99
(0.86-1.13)
0.92
(0.77-1.10)
1.21
(1.03-1.41)
1.38
(1.10-1.73)
1.01
(0.80-1.27)
1.32
(0.90-1.96)
1.06
(0.98-1.16)
1.11
(0.99-1.23)
Het = heterozygous for risk allele; Hom = homozygous for risk allele; NA = small number of individuals in a cell and can not be estimated.
a
ORs adjusted for age (quartiles), BMI (quartiles), sex, and ethnicity (pooled analysis).
b
rs2237895 and rs2237897 adjusted for one another.
116
117
ancestry (Table 5-6). Adjustment for education, a proxy for socio-economic status (SES)
and European ancestry, did not influence the results (Table 5-7) (26).
We next performed analyses that combined evidence for association across the
five ethnic groups. In this analysis the power to achieve nominal significance for the
allelic effects reported previously was >80% for 18 out of 19 alleles (average 94%; Table
5-4). In this analysis all 19 (100%; P=1.9x10
-6
, binomial probability) variants were
associated with risk in the same direction as the initial report (odds ratios>1) and 14
(P=5.7x10
-15
; binomial probability) with nominal statistical significance (P<0.05). All 19
associations remained in the same direction as previous reports (OR>1) and 13 of the
variants were significantly associated with T2D risk when the European American
subjects were excluded from the analysis. The association of rs8050136 in FTO was
attenuated by adjustment for BMI (odds ratio (95% confidence interval), 1.06(1.00-1.11)
prior to adjustment; 1.02(0.96-1.08) after adjustment). Only 5 of the 19 risk variants
showed nominal evidence for heterogeneity in the odds ratio across ethnic groups, and
only one of these (CDKAL1) was significant after correction for having performed 19
tests (PPARG, rs1801282, P
het
= 0.048; WFS1, rs10010131, P
het
= 0.032; CDKAL1,
rs7754840, P
het
= 6.2x10
-4
; HHEX, rs1111875, P
het
= 0.037; and, HNF1B, rs4430796, P
het

= 0.043; Table 5-4).

Table 5-6: Effects of European Ancestry Adjustment in African Americans

African American Males with
% European ancestry estimates
(336 cases, 397 controls)
SNP Chr
Nearest
Gene
Risk
Allele
a
Risk Allele
Frequency,
European
Americans

Risk Allele
Frequency,
African
American
Unadjusted
OR(95% CI)
b
per allele
Adjusted
OR(95% CI)
c
per allele
rs10923931 1 NOTCH2 T 0.12 0.29 0.94(0.74-1.19) 0.93(0.73-1.18)
rs7578597 2 THADA T 0.90 0.75 1.15(0.89-1.49) 1.16(0.90-1.51)
rs1801282 3 PPARG C 0.89 0.97 2.42(1.18-4.98) 2.39(1.16-4.92)
rs4607103 3 ADAMTS9 C 0.73 0.70 1.03(0.81-1.31) 1.04(0.81-1.32)
rs4402960 3 IGF2BP2 C 0.31 0.49 0.82(0.67-1.02) 0.81(0.66-1.01)
rs10010131 4 WFS1 G 0.59 0.66 1.01(0.80-1.27) 1.01(0.80-1.27)
rs7754840 6 CDKAL1 C 0.29 0.55 1.10(0.88-1.37) 1.09(0.87-1.36)
rs864745 7 JAZF1 T 0.51 0.73 1.26(0.99-1.61) 1.26(0.98-1.61)
rs13266634 8 SLC30A8 C 0.68 0.89 1.37(0.95-1.97) 1.35(0.92-1.96)
rs2383208 9 CDKN2B T 0.81 0.81 1.05(0.79-1.39) 1.04(0.78-1.38)
rs1111875 10 HHEX C 0.61 0.74 0.96(0.74-1.25) 0.95(0.73-1.23)
rs7903146 10 TCF7L2 T 0.27 0.28 1.39(1.08-1.79) 1.39(1.08-1.79)
rs12779790 10 CDC123 G 0.17 0.14 1.06(0.78-1.43) 1.06(0.78-1.43)
rs2237895
d
11 KCNQ1 C 0.42 0.20 0.96(0.73-1.26) 0.96(0.73-1.27)
rs2237897
d
11 KCNQ1 C 0.95 0.92 1.58(0.99-2.54) 1.59(0.99-2.55)
rs5219 11 KCNJ11 T 0.35 0.09 1.32(0.90-1.92) 1.38(0.93-2.04)
rs7961581 12 TSPAN8 C 0.29 0.23 0.87(0.67-1.12) 0.88(0.68-1.13)
rs8050136 16 FTO A 0.41 0.43 1.12(0.90-1.39) 1.12(0.90-1.39)
rs4430796 17 HNF1B G 0.50 0.65 1.13(0.90-1.43) 1.12(0.89-1.42)
a
NCBI build 36 (forward strand).
b
Adjusted for age (quartiles), and BMI (quartiles).
c
Adjusted for age (quartiles), BMI (quartiles), and % European ancestry (estimated using ancestry informative markers as previously designed(16, 23,
24).
d
rs2237895 and rs2237897 adjusted for one another.
118

Table 5-7: Effects of Adjustment for Education

European Americans
529 cases
1,000 controls
a
African Americans
1,067 case
1,464 controls
a
Latinos
2,190 case
2,157 controls
a
Japanese Americans
1,727 case
1,756 controls
a
Native Hawaiians
573 cases
973 controls
a
SNP /
Risk Allele
b
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
rs10923931
T
0.83
(0.63-1.09)
0.85
(0.65-1.11)
1.10
(0.96-1.24)
1.09
(0.96-1.24)
1.15
(0.99-1.34)
1.15
(0.99-1.33)
1.00
(0.70-1.44)
1.00
(0.69-1.43)
0.76
(0.50-1.14)
0.76
(0.50-1.14)
rs7578597
T
1.41
(1.05-1.89)
1.34
(1.00-1.81)
1.04
(0.91-1.20)
1.04
(0.91-1.20)
1.11
(0.92-1.34)
1.11
(0.92-1.34)
1.10
(0.68-1.78)
1.11
(0.69-1.80)
1.66
(1.02-2.71)
1.63
(1.00-2.66)
rs1801282
C
1.25
(0.94-1.67)
1.24
(0.93-1.65)
1.92
(1.28-2.88)
1.88
(1.25-2.82)
1.02
(0.88-1.18)
1.02
(0.88-1.19)
1.05
(0.81-1.35)
1.05
(0.82-1.36)
1.16
(0.85-1.57)
1.15
(0.85-1.56)
rs4607103
C

1.11
(0.91-1.34)
1.12
(0.92-1.36)
1.04
(0.91-1.18)
1.03
(0.91-1.18)
0.98
(0.89-1.08)
0.98
(0.89-1.08)
1.05
(0.95-1.17)
1.06
(0.95-1.17)
0.99
(0.83-1.17)
0.98
(0.83-1.17)
rs4402960
T
1.01
(0.84-1.22)
0.99
(0.82-1.19)
1.14
(1.01-1.28)
1.14
(1.01-1.28)
1.06
(0.96-1.17)
1.06
(0.96-1.17)
1.24
(1.11-1.38)
1.24
(1.11-1.38)
1.18
(0.99-1.40)
1.19
(1.00-1.41)
rs10010131
G
1.18
(0.99-1.40)
1.17
(0.98-1.39)
0.94
(0.83-1.07)
0.94
(0.83-1.06)
1.14
(1.03-1.26)
1.14
(1.03-1.26)
1.45
(0.98-2.13)
1.43
(0.97-2.11)
1.25
(1.02-1.54)
1.25
(1.01-1.53)
rs7754840
C
1.24
(1.04-1.49)
1.21
(1.01-1.45)
1.03
(0.91-1.16)
1.02
(0.91-1.15)
1.07
(0.97-1.18)
1.07
(0.97-1.17)
1.38
(1.25-1.53)
1.39
(1.25-1.54)
1.38
(1.18-1.62)
1.37
(1.17-1.61)
rs864745
T
0.98
(0.83-1.16)
0.99
(0.84-1.17)
1.16
(1.01-1.32)
1.16
(1.01-1.33)
1.30
(1.18-1.42)
1.29
(1.18-1.42)
1.19
(1.05-1.35)
1.19
(1.05-1.35)
1.12
(0.93-1.35)
1.11
(0.92-1.34)
rs13266634
C
1.28
(1.06-1.55)
1.28
(1.05-1.55)
1.22
(1.00-1.49)
1.21
(0.99-1.48)
1.11
(1.00-1.24)
1.11
(1.00-1.23)
1.18
(1.06-1.32)
1.18
(1.06-1.31)
1.03
(0.88-1.21)
1.04
(0.88-1.22)
rs2383208
A

1.33
(1.06-1.67)
1.37
(1.09-1.73)
1.14
(0.98-1.33)
1.14
(0.98-1.33)
1.15
(1.01-1.31)
1.15
(1.01-1.31)
1.25
(1.13-1.39)
1.25
(1.13-1.39)
1.02
(0.85-1.22)
1.02
(0.86-1.23)
rs1111875
C
0.92
(0.78-1.10)
0.93
(0.78-1.11)
1.11
(0.96-1.27)
1.10
(0.96-1.27)
1.03
(0.94-1.13)
1.03
(0.94-1.13)
1.21
(1.08-1.36)
1.21
(1.08-1.36)
0.94
(0.78-1.12)
0.94
(0.79-1.13)

119

Table 5-7: Continued

European Americans
529 cases
1,000 controls
a
African Americans
1,067 case
1,464 controls
a
Latinos
2,190 case
2,157 controls
a
Japanese Americans
1,727 case
1,756 controls
a
Native Hawaiians
573 cases
973 controls
a
SNP /
Risk Allele
b
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
Unadj
c
OR
(95% CI)
Adj
d
OR
(95% CI)
rs7903146
T
1.54
(1.28-1.85)
1.55
(1.29-1.87)
1.32
(1.16-1.51)
1.32
(1.16-1.51)
1.31
(1.19-1.45)
1.32
(1.19-1.46)
1.78
(1.41-2.25)
1.77
(1.40-2.24)
1.12
(0.90-1.39)
1.11
(0.90-1.39)
rs12779790
G
1.01
(0.80-1.26)
1.04
(0.83-1.30)
1.09
(0.92-1.29)
1.10
(0.93-1.30)
1.20
(1.07-1.35)
1.20
(1.07-1.35)
1.03
(0.89-1.18)
1.02
(0.89-1.17)
1.17
(0.96-1.43)
1.17
(0.96-1.43)
rs2237895
C

0.98
(0.82-1.17)
0.98
(0.82-1.17)
1.05
(0.91-1.22)
1.06
(0.91-1.23)
1.15
(1.04-1.27)
1.14
(1.03-1.26)
1.12
(0.98-1.27)
1.11
(0.98-1.26)
1.16
(0.97-1.39)
1.16
(0.97-1.39)
Rs2237897

0.85
(0.58-1.24)
0.87
(0.59-1.28)
1.13
(0.90-1.42)
1.13
(0.90-1.42)
1.24
(1.10-1.40)
1.25
(1.11-1.41)
1.27
(1.12-1.45)
1.27
(1.12-1.45)
1.07
(0.87-1.32)
1.07
(0.87-1.32)
rs5219
T
1.25
(1.05-1.48)
1.26
(1.06-1.50)
1.02
(0.83-1.25)
1.03
(0.84-1.26)
1.10
(1.00-1.21)
1.10
(1.00-1.20)
1.26
(1.14-1.41)
1.27
(1.14-1.41)
1.03
(0.88-1.21)
1.03
(0.88-1.21)
rs7961581
C
1.03
(0.86-1.23)
1.04
(0.86-1.25)
0.92
(0.80-1.06)
0.92
(0.80-1.06)
1.03
(0.92-1.15)
1.03
(0.93-1.15)
1.00
(0.88-1.13)
0.99
(0.88-1.13)
1.12
(0.94-1.34)
1.12
(0.94-1.33)
rs8050136
A
0.89
(0.75-1.06)
0.91
(0.76-1.08)
1.06
(0.94-1.19)
1.06
(0.94-1.19)
1.02
(0.92-1.13)
1.02
(0.93-1.13)
1.03
(0.91-1.17)
1.04
(0.91-1.17)
1.01
(0.84-1.21)
1.02
(0.85-1.22)
rs4430796
G
0.97
(0.82-1.14)
0.95
(0.81-1.13)
1.11
(0.98-1.26)
1.11
(0.98-1.26)
0.96
(0.87-1.04)
0.96
(0.87-1.04)
1.18
(1.06-1.31)
1.18
(1.06-1.32)
1.09
(0.92-1.29)
1.10
(0.92-1.30)
a
Missing data (n cases/ n controls): European American (4/6); African American (10/5); Latino (30/27); Japanese (9/5); and Native Hawaiian (3/10).
b
NCBI build 36 (forward strand).
c
Adjusted for age (quartiles), BMI (quartiles), sex, and ethnicity (pooled).
d
Adjusted for age (quartiles), BMI (quartiles), sex, education (categorized as: high school graduate or less; some college or vocational school; and
college graduate or graduate or professional school), and ethnicity (pooled).
120
121

Summary Measures of Risk
We next calculated a summary risk score comprised of an unweighted count of the 19
risk-associated alleles. The average increment in risk per allele was generally similar in
all populations, except Japanese Americans, where the effect of each allele was nearly
double that observed in Europeans ((odds ratio, 95% confidence interval): African
Americans, 1.09, 1.05-1.12; (P=3.0x10
-6
); Native Hawaiians, 1.10, 1.06-1.15 (P=1.2x10
-
5
); European Americans, 1.11, 1.06-1.17 (P=1.2x10
-5
); Latinos, 1.12, 1.09-1.14
(P=7.5x10
-19
); and, Japanese, 1.20, 1.17-1.24; (P=7.0x10
-32
); P
het
= 3.8x10
-4
) (Table 5-8).
Results were similar when limiting the analysis to individuals with complete genotype
data for all variants and when including only those markers associated with risk (at
P<0.10) (Table 5-9). Individuals in the top quartile of the risk allele distribution were at
1.6 (African Americans, P=5.3x10
-4
) to 3.1-fold (Japanese Americans, P=7.9x10
-26
)
greater risk of diabetes compared to those in the lowest quartile (Table 5-8).
Using these ethnic-specific per allele odds ratio estimates and the aggregate risk
allele counts, we built a quantitative risk model to compare the distribution of genetic
risks between populations associated with these marker alleles. The higher average
number of risk alleles in African Americans caused their distribution to be slightly right
shifted (towards higher log ORs) compared to European Americans, however their
relatively smaller per allele odds ratio resulted in wide overlap with the European
American distribution (Figure 3-2). The Japanese Americans had a wider distribution of
risk because of the large per allele odds ratio, but the low average risk allele counts

Table 5-8: The Association of the Total Risk Score With T2D Risk by Racial/Ethnic Population
a
European
Americans
African
Americans
Latinos Japanese
Americans
Native
Hawaiians
Pooled
Total Risk alleles, Mean (range) 18.8 (10-27) 20.2 (11-27) 18.6 (10-28) 17.2 (10-26) 18.1 (10-26) 18.5 (10-28)
Per allele additive model
N (cases/controls)

533/1,006

1,077/1,469 2,220/2,184 1,736/1,761 576/983 6,142/7,403
OR(95% CI)
bc
1.11 (1.06-1.17) 1.09 (1.05-1.12) 1.12 (1.09-1.14) 1.20 (1.17-1.24) 1.10 (1.06-1.15) 1.13 (1.11-1.15)
P-value 1.2 x 10
-5
3.0 x 10
-6
7.5 x 10
-19
7.0 x 10
-32
1.2 x 10
-5
4.7 x 10
-59

P-value vs. European Americans

Ref. 0.43 0.76 0.007 0.86

Quartiles of Risk Alleles
d

Q1 n (cases/controls) 69/196 206/360 362/464 267/441 119/251 1,023/1,712
OR(95% CI)

1.00(ref) 1.00(ref) 1.00(ref) 1.00(ref) 1.00(ref) 1.00(ref)
Q2 n (cases/controls) 112/273 273/437 520/605 424/518 150/307 1,479/2,140
OR(95% CI)

1.17 (0.79-1.73) 1.10 (0.86-1.40) 1.17 (0.97-1.42) 1.39 (1.12-1.72) 1.11 (0.81-1.52) 1.21 (1.08-1.35)
P-value 0.44 0.45 0.10 3.2 x 10
-3
0.51 8.4 x 10
-4

Q3 n (cases/controls) 170/281 353/391 626/587 297/306 167/259 1,613/1,824
OR(95% CI)

1.52 (1.05-2.21) 1.59 (1.25-2.01) 1.50 (1.25-1.81) 1.85 (1.46-2.36) 1.39 (1.02-1.90) 1.61 (1.44-1.80)
P-value 0.027 1.3 x 10
-4
2.1 x 10
-5
5.4 x 10
-7
0.039 4..2 x 10
-17

Q4 n (cases/controls) 182/256 245/281 712/528 748/496 140/166 2,027/1,727
OR(95% CI)

1.88 (1.29-2.74) 1.58 (1.22-2.04) 1.99 (1.65-2.40) 3.06 (2.48-3.77) 2.09 (1.49-2.94) 2.17 (1.95-2.42)
P-value 9.3 x 10
-4
5.3 x 10
-4
7.6 x 10
-13
7.9 x 10
-26
2.0 x 10
-5
4.6 x 10
-44

a
Odds ratios (and 95% confidence intervals) adjusted for age (quartiles), BMI (quartiles), sex, and ethnicity (in pooled analysis).
b
P=3.8x10
-4
for heterogeneity of allelic effects across ethnic groups (4 df test).
c
P≤0.02 for each ethnic group for the comparison with Japanese Americans.
d
Quartiles are defined separately for each population.
122

Table 5-9: Associations with Risk Score Among Subjects with and without Complete Genotype Data
a

European
Americans
African
Americans Latinos
Japanese
Americans
Native
Hawaiians Pooled P
het
b
All subjects and all variants. Missing genotype data assigned as mean number of alleles (6,142 cases and 7,403 controls).
n (Cases/Controls) 533/1,006 1,077/1,469 2,220/2,184 1,736/1,761 576/983 6,142/7,403

OR (95% CI)

1.11 (1.06-1.17) 1.09 (1.05-1.12) 1.12 (1.09-1.14) 1.20 (1.17-1.24) 1.10 (1.06-1.15) 1.13 (1.11-1.15)
3.8 x 10
-4

P-Value 1.2 x 10
-5
3.0 x 10
-6
7.5 x 10
-19
7.0 x 10
-32
1.2 x 10
-5
4.7 x 10
-59

All subjects with complete genotype data for all variants (5,522 cases and 6,633 controls).
n (Cases/Controls) 442/890 975/1,298 2,023/1,996 1,564/1,571 518/878 5,522/6,633

OR (95% CI)

1.11 (1.06-1.17) 1.09 (1.05-1.13) 1.12 (1.09-1.14) 1.19 (1.15-1.23) 1.10 (1.06-1.16) 1.13 (1.11-1.15)
5.7 x 10
-3

P-Value 4.1 x 10
-5
2.3 x 10
-6
2.0 x 10
-17
6.0 x 10
-27
2.0 x 10
-5
2.7 x 10
-53

All subjects and all variants with P<0.10 in ethnic-pooled analysis. Missing genotype data assigned as mean number
of alleles (6,142 cases 7,403 controls).

n (Cases/Controls) 533/1,006 1,077/1,469 2,220/2,184 1,736/1,761 576/983 6,142/7,403

OR (95% CI)

1.13 (1.07-1.18) 1.10 (1.06-1.15) 1.13 (1.10-1.16) 1.23 (1.19-1.28) 1.12 (1.07-1.17) 1.15 (1.13-1.17)
2.4 x 10
-4

P-Value 5.1 x 10
-6
6.4 x 10
-7
4.9 x 10
-20
3.5 x 10
-35
4.5 x 10
-6
8.1 x 10
-65

All subjects with complete genotype data using only variants with P<0.10 in ethnic-pooled analysis (5,572 cases 6,682
controls).

n (Cases/Controls) 450/897 953/1,311 2,039/2,010 1,571/1,578 521/886 5,572/6,682

OR (95% CI)

1.12 (1.07-1.19) 1.11 (1.07-1.15) 1.13 (1.10-1.16) 1.23 (1.18-1.27) 1.12 (1.06-1.17) 1.15 (1.13-1.17) 2.1 x 10
-3

P-Value 2.7 x 10
-5
4.6 x 10
-7
4.9 x 10
-18
1.5 x 10
-30
1.1 x 10
-5
3.5 x 10
-58

a
Odds ratios adjusted for age (quartiles), BMI (quartiles), sex, and ethnicity (in pooled analysis).
b
P
het
= P value for heterogeneity of allelic effects across ethnic groups (4 df test).
123
124
Figure 5-2: Predicted Distribution of T2D Risk from Common Variants by
Racial/Ethnic Group Compared to European Americans
A. African Americans

B. Latinos

125
Figure 5-2: Continued
C. Japanese Americans

D. Native Hawaiians

Figure 2. Comparison of the predicted risk distributions conveyed by the risk alleles relative to European Americans
(Blue in Panels A-D): Panel A, African Americans (Red); Panel B, Latinos (Yellow); Panel C, Japanese Americans
(Brown); Panel D, Native Hawaiians (Green). The x-axis is the log relative risk for each population centered (at log
RR=0) around the average total risk allele count in the multiethnic sample (18.5). The y-axis is the relative frequency
of the population with that level of risk.
126
caused the Japanese distribution to be left-shifted (towards lower log ORs) compared to
European Americans. The distributions for Latinos and Native Hawaiians were very
similar to the European Americans.

Discussion

We tested 19 common genetic risk markers that were discovered in European
populations. We found that association with all 19 of these SNPs trended in the same
direction in this large multiethnic study, and the majority of these variants were
nominally significant in their association with diabetes risk. A risk score comprised of
these alleles was significantly associated with diabetes risk in all five racial/ethnic
groups, with the only significant heterogeneity being larger effect sizes in Japanese
Americans. However, in comparing the distribution of risk conferred by these alleles
between populations we found that they explain little, if any, of known differences in the
prevalence of diabetes between these populations.
These observations indicate that most, if not all, of these alleles show
directionally similar association to T2D across many populations. Such a pattern
indicates that the causal alleles at these validated risk loci (which have yet to be found)
likely predate the migrations that separated these populations now residing in Europe,
Africa, East Asia, the Pacific Islands and the Americas. We note that this pattern is
unexpected under the recently described “common SNP, rare mutation” model of
Goldstein that suggests that GWAS signals with common alleles for T2D and other
127
diseases may be “synthetic associations” created by one or more rare alleles (19, 20).
Under the Goldstein Hypothesis the consistent associations that we noted at these loci
across populations would only be observed if, in each population, one or more distinct
rare alleles arose at each locus, and they happened to arise each time on the same
haplotype background. Although possible, this scenario seems unlikely, and a more
parsimonious explanation would be the “synthetic association” hypothesis of Goldstein
does not apply to a majority of these T2D SNPs.
The modest number of cases and controls in this study (as compared to the initial
discovery studies) likely underlies the lack of statistically significant associations in some
groups. Weaker associations in some racial/ethnic groups may also be due to differences
in allele frequencies, linkage disequilibrium, and environmental and genetic modifiers. In
two cases (WFS1 and CDKAL1), significant heterogeneity by race/ethnicity reflected a
lack of association in African Americans, perhaps because of lower linkage
disequilibrium between the marker and the biologically relevant allele.
It is interesting that the odds ratios observed for these marker SNPs were larger in
Japanese Americans than in the original discovery cohorts, and in the other ethnic groups
in our study. A meta-analysis of 7 association studies in Japanese populations replicated
associations from studies in European populations for 7 loci under study (TCF7L2,
CDKAL1, CDKN2B, IGF2BP2, SLC30A8, KCNJ11, and HHEX) (27). A recent GWAS in
Japanese observed significant associations in KCNQ1 as well as these same 7 loci and,
similar to our observations, noted magnitudes of effect that were generally stronger than
128
previously observed in European populations (28). Additional studies in other Asian
populations have replicated associations with many of these loci as well (29-32).
In the Multiethnic Cohort, we have found the prevalence of T2D to be at least 2-
fold higher in African Americans, Latinos, Japanese and Native Hawaiians compared to
European Americans, with these differences being independent of body weight (14). We
examined the extent to which the known genetic risk alleles for diabetes could explain
these disparities by quantifying and comparing the relative risks distributions between
populations. Compared to European Americans, we did not observe evidence of greater
genetic risk in any population. Our findings therefore indicate that these risk markers
explain little, if any, of racial/ethnic disparities in T2D prevalence. It remains possible
that the actual causal alleles in these regions may be more common in frequency and/or
have larger effects than the index signals in non-European populations. As seen with
KCNQ1 (12, 13), GWAS in non-European populations are effective in discovering risk
loci that are important in multiple populations but difficult to identify in European
populations where the alleles are rare.
This study had a number of limitations. First, a self-report of diabetes and use of
medication for diabetes was used to define cases and controls. We observed that
approximately 1% of a random sample of the controls in this study had HbA1C levels
above 7.0%, which suggests that only a small portion of controls had undiagnosed
diabetes (see Materials and Methods). Also, our case definition did not differentiate
between T1D and T2D, however we expect this misclassification to be minor as <3% of
T2D cases had a previous diagnosis of T1D based on other sources (see Materials and
129
Methods). The highly consistent findings of this study as compared to the discovery
GWAS reports argues that our phenotypic characterization is adequate to observe the
association to T2D.
Some caution should also be given to the interpretation of the risk modeling
conducted in each ethnic group, as the genetic markers included are unlikely to be the
causal alleles. Future fine-mapping and sequencing studies to identify the functional
variants (common and/or rare) and large-scale testing of each allele will be required to
more precisely model risk as well as assess differences in the distribution of genetic risk
across populations.
Another limitation is that we did not account for the potential confounding effects
of population stratification. However, odds ratios were essentially unchanged after
adjusting for global European ancestry in a subset of African Americans (336 cases 397
controls) for whom ancestry markers were available, suggesting that effects due to
population substructure were not substantial, at least in this group. We also noted that
controlling for education, a proxy for SES which has been shown to be significantly
associated with Native American ancestry in Latinos (26), had little effect on the
associations with these risk alleles. Furthermore, the risk alleles were not generally more
frequent in Latinos than in European Americans which would be likely if these alleles
were proxies for more general ancestry differences. While population stratification is
unlikely to fully explain these findings, it remains possible that at some loci, the causal
alleles may be more correlated with ancestry than the index SNPs.
130
In summary, our data provide strong support for common genetic variation contributing
to T2D risk in multiple populations. Our findings in T2D do not support the theory that
GWAS signals are due to rare alleles. Nonetheless, GWAS and sequencing studies in
these and other racial/ethnic populations are needed to reveal a more complete spectrum
of risk alleles that are important globally as well as those that may contribute to risk
disparities.
131
Chapter 5 References
1. Sladek R, Rocheleau G, Rung J, et al. A genome-wide association study identifies
novel risk loci for type 2 diabetes. Nature 2007;4457130:881-5.

2. Saxena R, Voight BF, Lyssenko V, et al. Genome-wide association analysis
identifies loci for type 2 diabetes and triglyceride levels. Science
2007;3165829:1331-6.

3. Scott LJ, Mohlke KL, Bonnycastle LL, et al. A genome-wide association study of
type 2 diabetes in Finns detects multiple susceptibility variants. Science
2007;3165829:1341-5.

4. Zeggini E, Weedon MN, Lindgren CM, et al. Replication of genome-wide
association signals in UK samples reveals risk loci for type 2 diabetes. Science
2007;3165829:1336-41.

5. Zeggini E, Scott LJ, Saxena R, et al. Meta-analysis of genome-wide association
data and large-scale replication identifies additional susceptibility loci for type 2
diabetes. Nat Genet 2008;405:638-45.

6. Grant SF, Thorleifsson G, Reynisdottir I, et al. Variant of transcription factor 7-
like 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet 2006;383:320-3.

7. Altshuler D, Hirschhorn JN, Klannemark M, et al. The common PPARgamma
Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes. Nat
Genet 2000;261:76-80.

8. Gloyn AL, Weedon MN, Owen KR, et al. Large-scale association studies of
variants in genes encoding the pancreatic beta-cell KATP channel subunits Kir6.2
(KCNJ11) and SUR1 (ABCC8) confirm that the KCNJ11 E23K variant is
associated with type 2 diabetes. Diabetes 2003;522:568-72.

9. Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al. A variant in CDKAL1
influences insulin response and risk of type 2 diabetes. Nat Genet 2007;396:770-
5.

10. Sandhu MS, Weedon MN, Fawcett KA, et al. Common variants in WFS1 confer
risk of type 2 diabetes. Nat Genet 2007;398:951-3.

11. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome
17 confer prostate cancer risk, and the one in TCF2 protects against type 2
diabetes. Nat Genet 2007;398:977-83.
132
12. Unoki H, Takahashi A, Kawaguchi T, et al. SNPs in KCNQ1 are associated with
susceptibility to type 2 diabetes in East Asian and European populations. Nat
Genet 2008;409:1098-102.

13. Yasuda K, Miyake K, Horikawa Y, et al. Variants in KCNQ1 are associated with
susceptibility to type 2 diabetes mellitus. Nat Genet 2008;409:1092-7.

14. Maskarinec G, Grandinetti A, Matsuura G, et al. Diabetes prevalence and body
mass index differ by ethnicity: the Multiethnic Cohort. Ethn Dis 2009;191:49-55.

15. Maskarinec G, Erber E, Grandinetti A, et al. Diabetes incidence based on linkages
with health plans: the multiethnic cohort. Diabetes 2009;588:1732-8.

16. Freedman ML, Haiman CA, Patterson N, et al. Admixture mapping identifies
8q24 as a prostate cancer risk locus in African-American men. Proc Natl Acad Sci
U S A 2006;10338:14068-73.

17. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24
independently affect risk for prostate cancer. Nat Genet 2007;395:638-44.

18. Kao WH, Klag MJ, Meoni LA, et al. MYH9 is associated with nondiabetic end-
stage renal disease in African Americans. Nat Genet 2008;4010:1185-92.

19. Goldstein DB. Common genetic variation and human traits. N Engl J Med
2009;36017:1696-8.

20. Dickson SP, Wang K, Krantz I, et al. Rare Variants Create Synthetic Genome-
Wide Associations. PLoS Biol 2010;81:e1000294.

21. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and
Los Angeles: baseline characteristics. Am J Epidemiol 2000;1514:346-57.

22. Lee LG, Connell CR, Bloch W. Allelic discrimination by nick-translation PCR
with fluorogenic probes. Nucleic Acids Res 1993;2116:3761-6.

23. Cheng CY, Kao WH, Patterson N, et al. Admixture mapping of 15,280 African
Americans identifies obesity susceptibility loci on chromosomes 5 and X. PLoS
Genet 2009;55:e1000490.

24. Deo RC, Patterson N, Tandon A, et al. A high-density admixture scan in 1,670
African Americans with hypertension. PLoS Genet 2007;311:e196.

133
25. Jonsson A, Isomaa B, Tuomi T, et al. A variant in the KCNQ1 gene predicts
future type 2 diabetes and mediates impaired insulin secretion. Diabetes
2009;5810:2409-13.

26. Florez JC, Price AL, Campbell D, et al. Strong association of socioeconomic
status with genetic ancestry in Latinos: implications for admixture studies of type
2 diabetes. Diabetologia 2009;528:1528-36.

27. Tabara Y, Osawa H, Kawamoto R, et al. Replication study of candidate genes
associated with type 2 diabetes based on genome-wide screening. Diabetes
2009;582:493-8.

28. Takeuchi F, Serizawa M, Yamamoto K, et al. Confirmation of multiple risk Loci
and genetic impacts by a genome-wide association study of type 2 diabetes in the
Japanese population. Diabetes 2009;587:1690-9.

29. Lee YH, Kang ES, Kim SH, et al. Association between polymorphisms in
SLC30A8, HHEX, CDKN2A/B, IGF2BP2, FTO, WFS1, CDKAL1, KCNQ1 and
type 2 diabetes in the Korean population. J Hum Genet 2008;5311-12:991-8.

30. Ng MC, Park KS, Oh B, et al. Implication of genetic variants near TCF7L2,
SLC30A8, HHEX, CDKAL1, CDKN2A/B, IGF2BP2, and FTO in type 2 diabetes
and obesity in 6,719 Asians. Diabetes 2008;578:2226-33.

31. Sachidanandam R, Weissman D, Schmidt SC, et al. A map of human genome
sequence variation containing 1.42 million single nucleotide polymorphisms.
Nature 2001;4096822:928-33.

32. Wu Y, Li H, Loos RJ, et al. Common variants in CDKAL1, CDKN2A/B,
IGF2BP2, SLC30A8, and HHEX/IDE genes are associated with type 2 diabetes
and impaired fasting glucose in a Chinese Han population. Diabetes
2008;5710:2834-42.

134
Chapter 6: Examining Known Diabetes Risk Variants for Association
with Prostate Cancer in a Multiethnic Population

Introduction

Epidemiologic studies conducted in multiple racial/ethnic populations provide
strong support for type 2 diabetes (T2D) as a protective factor for prostate cancer (PC)
(1-8). These common chronic diseases share no known environmental risk factors with
opposite effects on risk. Genome-wide and candidate gene studies have recently revealed
a single SNP (rs4430796) at the HNF1B (hepatocyte nuclear factor 1 homeobox B) locus,
with opposing effects on T2D and prostate cancer risk (9, 10). Genome-wide association
studies (GWAS) have also identified common variants at two additional loci, JAZF1
(juxtaposed with another zinc finger gene 1) and THADA (thyroid adenoma associated),
in association with both T2D and PC risk (11-13). At these loci the risk variants
identified for each T2D and PC are not correlated which suggests that there may be
multiple distinct functional alleles in each region. Additionally, variants in JAZF1 have
been associated with height through linkage and genome-wide analysis (14).
HNF1B is a transcription factor involved in nephrogenesis (15) and missense
mutations in HNF1B have been identified in patients with MODY (maturity-onset
diabetes of the young) (16). JAZF1 (or TIP27) is a transcription factor reported to repress
NR2C2 (17) which, when knockedout in mice, results in low serum IGF1 levels, growth
retardation, and hypoglycemia (18). Chromosomal translocations resulting in a JAZF1-
135
SUZ12 fusion protein have been found in endometroid-serous tumors (19). THADA is
thought to encode a death receptor protein (20), and the T2D risk variant has been
associated with lower β-cell response to GLP-1 and arginine suggesting lower β-cell
mass due to apoptosis (21). The breakpoints of chromosomal rearrangements found in
thyroid adenomas have also been localized to THADA (20). While the role of these genes
in T2D and PC are not completely understood the genetic findings suggest that T2D and
PC might share biological processes important in their etiology.
A recent prostate cancer case-control study of 18 established T2D genetic risk
variants among 1,176 cases and 1,101 controls of European ancestry reported a summary
T2D genetic risk score (both with and without the HNF1B allele) to be associated with
decreased prostate cancer risk (OR 0.96; 95%CI 0.92-0.99; P=0.02; without HNF1B)
(22). A second smaller prospective study with 397 cases examined 13 T2D risk variants
(9 confirmed in GWAS) for association with prostate cancer and observed a significant
protective effect for the TCF7L2 allele (rs7903146) as well as suggestive associations
with alleles in T2D candidate genes UCP2 SLC2A3 and CAPN10 (23).
We, as part of the Population Architecture using Genomics and Epidemiology
(PAGE) Study, further explored the hypothesis of shared genetic risk factors underlying
these diseases. We have tested 17 established T2D risk alleles in a study of 2,746 incident
prostate cancer cases and 3,317 controls among 5 racial/ethnic populations from the
Multiethnic Cohort (MEC; European Americans, African Americans, Latinos, Japanese
Americans, and Native Hawaiians). We examined both the individual as well as
aggregate effects, as a summary risk score, in association with prostate cancer risk. We
136
also examined the potential for gene x gene interactions between validated risk variants
for T2D and PC.

Methods

Study population
The MEC consists of 215,251 men and women, mainly from five self-reported
racial/ethnic populations: African Americans, Japanese Americans, Latinos, Native
Hawaiians and European Americans (24). Between 1993 and 1996, adults between 45
and 75 years old were enrolled by completing a 26-page, self-administered questionnaire
asking detailed information about dietary habits, demographic factors, level of education,
personal behaviors, and history of prior medical conditions (e.g. diabetes). Potential
cohort members were identified through Department of Motor Vehicles drivers’ license
files, voter registration files and Health Care Financing Administration data files. In the
cohort, incident cancer cases are identified annually through cohort linkage to population-
based cancer Surveillance, Epidemiology, and End Results (SEER) registries in Hawaii
and Los Angeles County as well as to the California State cancer registry. Information on
stage and grade of disease are also obtained through the SEER registries. In the MEC
biospecimens were collected on >67,000 subjects which included targeted collection of
incident prostate cancer cases.
The prostate cancer case-control study in the MEC includes 2,746 invasive cases
with a diagnosis after cohort entry and 2,288 controls with no history of prostate cancer
137
(25). We also included 1,029 additional male controls with no history of prostate cancer
from a colorectal case-control study of the MEC which also investigated these same 17
SNPs. Altogether, this study included 2,746 prostate cancer cases and 3,317 controls
(European American (472/558), African American (721/954), Latino (668/704), Japanese
American (753/936) and Native Hawaiian (132/165)). Based on the suggestive
association with the THADA variant (rs7578597) in Japanese, we included genotype data
for this variant from a GWAS of prostate cancer in Japanese which included an additional
415 Japanese cases and 518 Japanese controls that had been genotyped using the Illumina
660W array. Genotype concordance for this variant was 99.9% for the 1,273 subjects
that overlapped between studies. The Institutional Review Boards at the University of
Southern California and University of Hawaii approved the study protocol.

Genotyping
Genotyping of 18 established T2D risk alleles was performed using the allelic
discrimination assay on whole genome amplified DNA samples (26). The allele in
SLC30A8 (rs13266634) failed genotyping due to an adjacent allele (rs16889462) and was
excluded from the study. The HNF1B allele was not included in the present study as it
was included in a previous MEC study, and found to be associated with prostate cancer
risk (27). The genotype completion rate for each SNP was >96.0% among both cases and
controls in each racial/ethnic group. As a whole the genotype completion rate was
similar between cases (99.1%) and controls (99.2%), and among racial/ethnic groups,
with a range of 98.8% in African Americans to 99.3% in Japanese Americans. Hardy-
138
Weinberg Equilibrium (HWE) was assessed for each allele in each racial/ethnic group
using a chi-square test (1df). The test was statistically significant among control subjects
for rs7578597 in Japanese Americans (P=4.5x10
-5
), rs864745 in European Americans
(P=0.011). Among 90 tests we expected 4.5 significant tests by chance, and the most
highly significant departure was with rs7578597 in Japanese Americans (minor allele
frequency, MAF,0.01) where we observed 2 rare homozygotes when 0.4 were expected.
Details about Hardy-Weinberg equilibrium test and assay completion rate are in Table
6-1.

Statistical Analysis
Odds ratios (OR) and 95% confidence intervals (95% CI) for log-additive effects
were calculated in unconditional

logistic regression models while adjusting for quartiles
of age (at diagnosis for cases and at blood draw for controls), and body mass index (BMI,
kg/m
2
, quartiles). We also examined genetic associations with prostate cancer risk
among disease subgroups using the standard case-control approach, limiting the cases to
those with a specific phenotype (‘advanced disease’) and all controls, and a case-only
analysis to test for differences by disease subgroup. We defined the cancer as ‘advanced’
if high stage (regional by direct extension, regional by lymph nodes, regional by both
direct extension and lymph nodes, regional NOS, or distant metastases/systemic disease),
and/or high grade (low level of cell differentiation; Gleason score > 7). Non-advanced
139
Table 6-1: Genotype Assay Completion Rates and Hardy Weinberg Equilibrium
Testing

Genotype Assay Completion Rate Cases/Controls
HWE P value (1 df) Cases/Controls)
SNP /
Allele
Tested
a

European
Americans
472 cases
558 controls
African Americans
721 cases
954 controls
Latinos
668 cases
704 controls
Japanese
Americans
753 cases
936 controls
Native
Hawaiians
132 cases
165 controls
rs10923931
T
99.8%/99.8%
0.32/0.27
99.3%/99.9%
0.15/0.28
99.6%/99.9%
0.84/0.52
99.7%/99.9%
0.53/0.51
100%/100%
0.52/0.44
rs7578597
T
99.6%/99.5%
0.67/0.30
99.2%/99.0%
0.97/0.16
99.7%/99.1%
0.19/0.57
99.9%/99.7%
4x10
-22
/5x10
-5
99.2%/99.4%
0.28/0.51
rs1801282
C
99.1%/97.7%
0.42/0.60
96.1%99.3%
0.45/0.82
97.0%/98.7%
0.96/0.55
98.7%/99.1%
0.98/0.82
99.2%/98.8%
0.52/0.82
rs4607103

C
99.6%/99.8%
0.046/0.62
98.8%/99.7%
0.083/0.56
98.8%/99.7%
0.65/0.69
99.9%/99.8%
0.64/0.95
99.2%/100%
0.66/0.62
rs4402960
T
100%99.3%
0.52/0.11
98.3%/99.4%
0.16/0.21
99.4%/99.6%
0.43/0.63
99.9%/99.6%
3.6x10
-3
/0.14
98.5%/99.4%
0.68/0.69
rs10010131
G
100%/99.1%
0.70/0.79
97.8%/98.1%
0.23/0.30
98.5%/99.4%
0.96/0.46
99.5%/98.3%
0.12/0.65
100%/98.8%
0.85/0.72
rs7754840
C
99.1%/99.6%
0.67/0.62
98.6%/99.3%
0.29/0.70
99.4%/99.4%
0.42/0.064
99.2%/99.6%
0.26/0.86
98.5%/98.2%
0.19/0.92
rs864745
T
99.6%/99.6%
0.69/0.011
98.9%/99.3%
0.68/0.51
99.0%/99.6%
0.26/0.98
99.6%/99.9%
0.49/0.57
100%/100%
0.90/0.39
rs2383208

A
98.3%/98.6%
0.14/0.47
98.3%/99.2%
0.034/0.96
98.7%/99.4%
0.98/0.50
99.3%/98.3%
0.93/0.84
96.2%/98.2%
0.36/0.42
rs1111875
C
99.4%/99.5%
0.27/0.17
98.3%/99.0%
0.78/0.13
99.4%/99.6%
9.8x10
-4
/0.26
99.7%/99.6%
0.068/0.49
99.2%/99.4%
0.26/0.30
rs7903146
T
99.4%/99.6%
0.99/0.64
98.6%/99.7%
0.87/0.90
99.6%/99.1%
0.76/0.83
99.6%/99.8%
0.57/0.15
99.2%/99.4%
0.38/0.38
rs12779790
G
99.1%/99.3%
0.63/0.63
98.8%/99.5%
0.36/0.85
100%/99.7%
0.41/0.64
99.3%/99.8%
0.58/0.98
97.7%/99.4%
0.22/0.90
rs2237895

C

99.6%/98.6%
0.87/0.093
98.6%/98.8%
0.89/0.46
99.3%/99.0%
0.82/0.82
99.5%/98.8%
0.15/0.30
100%/100%
0.51/0.15
rs2237897

C
100%/99.5%
1.7x10
-3
/0.75
99.2%/99.5%
0.89/0.069
99.3%/99.6%
1.5x10
-4
/0.34
98.4%/99.4%
0.53/0.97
97.0%/100%
0.37/0.33
rs5219
T
99.4%/97.1%
0.15/0.068
98.3%/96.2%
0.81/0.27
99.1%/98.3%
0.90/0.27
99.5%/99.1%
0.11/0.46
99.2%/97.6%
0.50/0.62
rs7961581
C
98.7%/99.1%
0.97/0.25
99.0%/99.2%
0.95/0.081
98.8%/99.7%
0.14/0.60
99.3%/98.1%
0.27/0.88
97.0%/100%
0.45/0.57
rs8050136
A
100%/99.1%
0.32/0.15
99.3%/99.2%
0.30/0.53
99.4%/99.1%
0.70/0.58
99.6%/98.5%
0.81/0.81
100%/98.2%
0.86/0.26
a
NCBI build 36 (forward strand).
140
disease was defined as having both a localized stage and low grade (Gleason Score ≤ 7).
We were unable to define cases as either advanced (n=714) or non-advanced (n=1,420) if
both stage and grade data were missing or if either the stage was localized or grade was
low (Gleason Score ≤ 7), and information for the other variable was missing (n=612).
We excluded 62 subjects who were missing genotype information for more than 5 of the
SNPs. To examine potential associations independent of T2D status, the analysis was
also performed while limiting the subjects to those without a self-report of T2D on the
baseline questionnaire. The potential for confounding by population stratification was
assessed in a subset of subjects (n=1,336 cases and 932 controls) from the admixed
racial/ethnic populations (African Americans, Latinos, and Native Hawaiians) by
adjusting for percent European ancestry as previously described (25). We also tested for
gene x gene interactions between the 17 T2D SNPs in this study and 28 established
prostate cancer risk variants from previous prostate cancer case-control studies in the
MEC data (13, 25, 27).
We also created aggregate T2D risk scores to test the hypothesis that overall T2D
susceptibility would be protective for prostate cancer risk. We attempted to accomplish
this in two ways. First, we created an unweighted model by summing the number of risk
alleles for each individual and estimating the odds ratio per allele for this aggregate allele
count variable as an approximate risk score appropriate for unlinked variants with
independent effects of approximately the same magnitude for each allele. Second, we
created a model where each allele in the aggregate count was weighted by the log of its
published odds ratio from GWAS in populations of European ancestry (28-32) in order to
141
account for differences in effect size among the risk variants. We presented the more
parsimonious unweighted summary risk score because the unweighted and weighted
summary scores were highly correlated in each racial/ethnic group (Pearson r > 0.92) and
we observed similar per allele results for both scores. For individuals missing genotypes
for a given SNP, we assigned the average number of risk alleles within each ethnic group
(2 x risk allele frequency) to replace the missing value for that SNP. Both KCNQ1 SNPs
were included in this analysis as they were found to have independent effects in a
previous MEC study of T2D risk (Chapter 5).

Results

Subjects in this study ranged in age (age at diagnosis for cases and age at blood
draw for controls) from 46 to 88. On average, cases (mean 69.3 years) were slightly
older than controls (68.6 years). Native Hawaiians were on average the youngest (mean
67.3 years) and Japanese Americans the oldest (mean 70.6 years). Mean BMI levels
were very similar among cases (mean 26.6 kg/m
2
) and controls (26.7 kg/m
2
) while cases
(13.2%) were more likely to have a first degree family history of prostate cancer than
controls (8.3%)
In ethnic-specific analysis, we did not observe large deviations from the null
hypothesis that 50% of the T2D risk alleles would have an inverse association with
prostate cancer, as we observed OR’s<1 for 59% (10/17; P=0.31; binomial probability) of
the alleles in Latinos, 53% (9/17; P=0.50; binomial probability) for African Americans
142
and Japanese Americans and 47% (8/17; P=0.69; binomial probability) for European
Americans and Native Hawaiians. We observed nominally statistically significant
positive associations (P<0.05) for variants in IGF2BP2 (rs4402960; odds ratio (95%CI)
1.26(1.05-1.52), P=0.013) and CDC123 (rs12779790; 1.30(1.03-1.64), P=0.030) in
European Americans, and with THADA (rs7578597; (OR 3.80; 95%CI 1.69-8.55;
P=1.3x10
-3
) in Japanese Americans. None of these associations remained significant
after adjusting for multiple tests (Table 6-2).
We observed significant heterogeneity of effect across racial/ethnic groups for the
THADA allele (P=0.017) with positive associations noted in Japanese Americans
(OR=3.80) and Latinos (OR=1.31), and relatively null associations observed in African
Americans (OR=1.04), Native Hawaiians (OR=0.99) and European Americans
(OR=0.97). Seven of the 17 alleles (P=0.83; binomial probability) were inversely
associated with prostate cancer risk (OR<1) in the pooled sample, with a nominally
significant association noted with only the TSPAN8 allele (rs7961581) (OR=0.90; 95%
CI 0. 83-0.99; P=0.021). Analysis of each allele by genotype class did not show strong
evidence of association for either the dominant or recessive model in ethnic-pooled
analysis (Table 6-3). We also observed similar associations when the subjects were
limited to those with no self-report of diabetes (n=2,073 cases and n=2,546 controls;
Table 6-4).
We also examined each allele for association in subgroups of disease severity
(advanced vs. non-advanced) (Table 6-5, Tables 6-6 and 6-7). In case-only

Table 6-2: The Association of Known T2D Risk Alleles with Prostate Cancer Risk by Race/Ethnicity
a
SNP /
Allele
Tested
b

Chr. /
Nearest
Gene
European
Americans
472 cases
558 controls
African
Americans
721 cases
954 controls
Latinos
668 cases
704 controls
Japanese
Americans
753 cases
936 controls
Native
Hawaiians
132 cases
165 controls
Pooled
2,746 cases
3,317 controls
P
value P
het
c
rs10923931
T
1
NOTCH2
0.89(0.67-1.18)
0.11
0.91(0.78-1.05)
0.32
1.02(0.79-1.31)
0.10
1.05(0.65-1.68)
0.02
1.02(0.49-2.13)
0.05
0.93(0.84-1.04)
0.14 0.22 0.90
rs7578597
d
T
2
THADA
0.97(0.73-1.29)
0.90
1.04(0.88-1.21)
0.75
1.31(0.97-1.77)
0.92
3.80(1.69-8.55)
0.99
0.99(0.47-2.08)
0.95
1.11(0.98-1.25)
0. 90 0.092 0.017
rs1801282
C
3
PPARG
1.09(0.81-1.46)
0. 89
1.04(0.69-1.59)
0.97
1.02(0.78-1.33)
0.91
0.98(0.68-1.42)
0.96
1.53(0.76-3.10)
0.93
1.05(0.90-1.23)
0.94 0.51 0.90
rs4607103

C
3
ADAMTS9
0.96(0.79-1.17)
0.74
1.06(0.91-1.23)
0.71
0.95(0.81-1.12)
0.69
1.02(0.88-1.17)
0.62
1.17(0.80-1.72)
0.73
1.01(0.93-1.09)
0.69 0.89 0.78
rs4402960
T
3
IGF2BP2
1.26(1.05-1.52)
0.29
0.97(0.85-1.11)
0.52
0.99(0.84-1.18)
0.26
1.09(0.94-1.26)
0.30
0.81(0.55-1.18)
0.29
1.05(0.97-1.13)
0.35 0.26 0.097
rs10010131
G
4
WFS1
1.02(0.85-1.22)
0.62
1.08(0.93-1.25)
0.64
1.11(0.93-1.31)
0.71
0.79(0.46-1.35)
0.99
1.46(0.93-2.31)
0. 81
1.07(0.98-1.17)
0.76 0.13 0.56
rs7754840
C
6
CDKAL1
1.02(0.85-1.23)
0.31
1.14(1.00-1.32)
0.54
0.98(0.83-1.14)
0.32
0.99(0.86-1.13)
0.44
1.05(0.74-1.48)
0.56
1.04(0.96-1.12)
0.43 0.36 0.58
rs864745
T
7
JAZF1
0.90(0.75-1.08)
0.51
0.96(0.83-1.12)
0.72
0.93(0.79-1.08)
0.62
1.16(0.98-1.38)
0.78
0.85(0.58-1.25)
0.77
0.98(0.90-1.06)
0.68 0.54 0.25
rs2383208

A
9
CDKN2B
0.95(0.75-1.20)
0. 82
1.06(0.88-1.27)
0. 81
0.93(0.76-1.14)
0. 85
1.00(0.87-1.15)
0.56
0.95(0.63-1.42)
0.77
0.99(0.91-1.08)
0.75 0.85 0.91
rs1111875
C
10
HHEX
1.11(0.92-1.32)
0.58
1.10(0.94-1.30)
0.74
1.03(0.89-1.20)
0.63
0.99(0.85-1.15)
0.28
1.03(0.72-1.48)
0.29
1.05(0.97-1.14)
0.54 0.22 0.86

143

Table 6-2: Continued
SNP /
Allele
Tested
b

Chr. /
Nearest
Gene
European
Americans
472 cases
558 controls
African
Americans
721 cases
954 controls
Latinos
668 cases
704 controls
Japanese
Americans
753 cases
936 controls
Native
Hawaiians
132 cases
165 controls
Pooled
2,746 cases
3,317 controls
P
value P
het
c
rs7903146
T
10
TCF7L2
0.98(0.81-1.18)
0.32
0.97(0.84-1.13)
0.29
1.06(0.88-1.26)
0.24
0.92(0.66-1.30)
0.05
1.02(0.64-1.64)
0.14
1.00(0.91-1.09)
0.21 0.91 0.93
rs12779790
G
10
CDC123
1.30(1.03-1.64)
0.15
1.04(0.86-1.26)
0.15
0.91(0.75-1.11)
0.18
1.00(0.83-1.20)
0.17
0.92(0.61-1.38)
0.21
1.03(0.94-1.14)
0.16 0.53 0.23
rs2237895
e
C

11
KCNQ1
0.94(0.78-1.13)
0.42
0.96(0.81-1.14)
0. 21
0.92(0.77-1.09)
0.44
1.08(0.91-1.28)
0.34
0.86(0.59-1.27)
0.35
0.98(0.90-1.06)
0.34 0.55 0.32
rs2237897
e
C
11
KCNQ1
1.03(0.70-1.50)
0.95
0.92(0.72-1.18)
0.92
0.95(0.78-1.15)
0.78
1.05(0.88-1.24)
0.62
1.25(0.79-1.97)
0.78
1.02(0.92-1.13)
0. 80 0.76 0.45
rs5219
T
11
KCNJ11
1.11(0.92-1.33)
0.36
0.97(0.77-1.22)
0.11
0.95(0.82-1.11)
0.38
0.97(0.84-1.11)
0.37
1.22(0.87-1.72)
0.36
1.00(0.93-1.09)
0.29 0.93 0.57
rs7961581
C
12
TSPAN8
0.99(0.82-1.19)
0.30
0.92(0.78-1.08)
0.23
0.89(0.74-1.06)
0.23
0.87(0.73-1.03)
0.21
0.80(0.56-1.15)
0.32
0.90(0.83-0.99)
0.24 0.021 0.79
rs8050136
A
16
FTO
1.05(0.88-1.25)
0.39
0.93(0.81-1.08)
0.44
1.07(0.90-1.26)
0.28
0. 88(0.74-1.04)
0.21
0.81(0.55-1.20)
0.24
0.97(0.89-1.05)
0.32 0.38 0.32
a
Each cell gives odds ratios (and 95% confidence intervals) for allele dosage effects along with the risk allele frequency in controls. ORs adjusted for age (quartiles), BMI
(quartiles), type 2 diabetes (self-report), and ethnicity (in pooled analysis).
b
NCBI build 36 (forward strand).
c
P
het
= P value for heterogeneity of allelic effects across ethnic groups (4 df test).
d
rs7578597 analysis includes an additional 414 cases and 517 controls for Japanese Americans.
e
rs2237895 and rs2237897 adjusted for each other.
144

Table 6-3: Association with Prostate Cancer Risk by Genotype
European Americans African Americans Latinos Japanese Americans Native Hawaiians Ethnically Pooled
SNP
Risk Allele Het Hom Het Hom Het Hom Het Hom Het Hom Het Hom
rs10923931
T
0.98
(0.72-1.34)
0.37
(0.10-1.35)
0.88
(0.72-1.09)
0.85
(0.61-1.19)
0.99
(0.75-1.31)
1.30
(0.39-4.31)
1.05
(0.65-1.68) NA
1.14
(0.53-2.47) NA
0.95
(0.83-1.09)
0.84
(0.62-1.14)
rs7578597
b
T
1.16
(0.38-3.50)
1.10
(0.37-3.21)
1.24
(0.81-1.89)
1.21
(0.80-1.82)
0.93
(0.24-3.61)
1.26
(0.33-4.74)
0.52
(0.04-6.94)
2.74
(0.25-30.4) NA NA
1.13
(0.78-1.63)
1.25
(0.87-1.79)
rs1801282
C
1.32
(0.30-5.74)
1.41
(0.33-6.00) NA NA
1.25
(0.38-4.10)
1.25
(0.39-3.98)
0.90
(0.06-14.9)
0.89
(0.06-14.4) NA NA
1.42
(0.61-3.29)
1.46
(0.63-3.36)
rs4607103
C

0.65
(0.40-1.06)
0.73
(0.45-1.17)
0.80
(0.55-1.15)
0.95
(0.67-1.36)
1.09
(0.75-1.59)
0.97
(0.67-1.41)
1.08
(0.80-1.44)
1.06
(0.78-1.43)
0.80
(0.29-2.18)
1.06
(0.40-2.83)
0.92
(0.77-1.10)
0.96
(0.81-1.15)
rs4402960
T
1.36
(1.04-1.77)
1.49
(0.99-2.26)
0.94
(0.74-1.20)
0.94
(0.72-1.23)
1.06
(0.85-1.33)
0.87
(0.56-1.34)
0.86
(0.70-1.06)
1.56
(1.11-2.19)
0.74
(0.45-1.22)
0.76
(0.30-1.92)
1.01
(0.90-1.13)
1.12
(0.95-1.32)
rs10010131
G
0.93
(0.64-1.35)
1.00
(0.68-1.46)
1.36
(0.99-1.87)
1.29
(0.94-1.79)
1.00
(0.66-1.53)
1.15
(0.76-1.73) NA NA
1.18
(0.26-5.43)
1.79
(0.40-7.96)
1.11
(0.90-1.36)
1.17
(0.95-1.44)
rs7754840
C
1.09
(0.84-1.41)
0.98
(0.63-1.50)
1.29
(0.99-1.68)
1.34
(1.01-1.78)
1.04
(0.83-1.31)
0.89
(0.63-1.27)
0.90
(0.72-1.12)
1.00
(0.76-1.32)
1.37
(0.71-2.64)
1.18
(0.58-2.39)
1.05
(0.94-1.19)
1.07
(0.92-1.24)
rs864745
T
0.76
(0.56-1.03)
0.82
(0.57-1.18)
0.98
(0.68-1.42)
0.94
(0.65-1.35)
0.83
(0.61-1.13)
0.83
(0.60-1.14)
1.26
(0.76-2.10)
1.43
(0.87-2.36)
0.57
(0.19-1.68)
0.54
(0.19-1.56)
0.86
(0.73-1.02)
0.90
(0.75-1.07)
rs2383208
A

1.24
(0.55-2.80)
1.13
(0.51-2.51)
1.85
(0.97-3.57)
1.78
(0.94-3.36)
1.08
(0.55-2.13)
0.98
(0.50-1.89)
1.00
(0.77-1.29)
1.00
(0.75-1.32)
1.91
(0.59-6.19)
1.49
(0.47-4.70)
1.13
(0.92-1.40)
1.07
(0.86-1.32)
rs1111875
C
0.85
(0.60-1.20)
1.12
(0.78-1.62)
0.93
(0.60-1.45)
1.08
(0.70-1.67)
0.66
(0.48-0.91)
0.89
(0.65-1.24)
0.92
(0.75-1.13)
1.09
(0.76-1.54)
0.81
(0.50-1.33)
1.49
(0.63-3.55)
0.85
(0.74-0.98)
1.06
(0.91-1.24)

145

Table 6-3: Continued
European Americans African Americans Latinos Japanese Americans Native Hawaiians Ethnically Pooled
SNP
Risk Allele Het Hom Het Hom Het Hom Het Hom Het Hom Het Hom
rs7903146
T
0.95
(0.73-1.23)
0.98
(0.64-1.52)
0.97
(0.79-1.19)
0.95
(0.66-1.37)
1.03
(0.82-1.30)
1.16
(0.74-1.83)
0.86
(0.61-1.22) NA
0.85
(0.49-1.46)
2.77
(0.47-16.2)
0.97
(0.86-1.09)
1.04
(0.83-1.32)
rs12779790
G
1.23
(0.93-1.63)
2.07
(0.96-4.47)
1.09
(0.87-1.36)
0.86
(0.43-1.69)
0.97
(0.77-1.23)
0.65
(0.34-1.24)
0.97
(0.78-1.21)
1.12
(0.63-1.99)
0.75
(0.44-1.26)
1.38
(0.45-4.23)
1.04
(0.92-1.16)
1.05
(0.77-1.43)
rs2237895
c
C
0.81
(0.61-1.08)
0.93
(0.64-1.37)
1.01
(0.82-1.25)
0.81
(0.50-1.33)
0.94
(0.72-1.21)
0.84
(0.59-1.19)
1.07
(0.85-1.35)
1.18
(0.82-1.70)
1.11
(0.65-1.88)
0.62
(0.26-1.45)
0.97
(0.86-1.09)
0.96
(0.80-1.15)
rs2237897
c
C
0.35
(0.06-1.88)
0.41
(0.08-2.17)
1.71
(0.57-5.16)
1.46
(0.49-4.31)
0.59
(0.36-0.96)
0.66
(0.41-1.08)
0.98
(0.71-1.36)
1.07
(0.75-1.53)
2.25
(0.64-7.96)
2.36
(0.66-8.46)
0.94
(0.74-1.20)
0.99
(0.77-1.26)
rs5219
T
0.88
(0.67-1.15)
1.46
(0.99-2.18)
0.93
(0.72-1.19)
1.44
(0.52-4.00)
1.01
(0.80-1.27)
0.88
(0.63-1.22)
0.85
(0.69-1.04)
1.02
(0.76-1.39)
1.48
(0.89-2.47)
1.32
(0.64-2.74)
0.94
(0.83-1.05)
1.08
(0.90-1.29)
rs7961581
C
1.07
(0.82-1.39)
0.88
(0.56-1.38)
0.99
(0.80-1.22)
0.71
(0.46-1.11)
0.80
(0.64-1.01)
0.99
(0.61-1.63)
0.82
(0.66-1.01)
0.89
(0.55-1.47)
0.76
(0.46-1.26)
0.68
(0.30-1.58)
0.90
(0.80-1.00)
0.83
(0.67-1.04)
rs8050136
A
1.26
(0.96-1.65)
1.00
(0.69-1.45)
0.97
(0.78-1.21)
0.86
(0.64-1.15)
1.13
(0.90-1.41)
1.06
(0.70-1.58)
0.88
(0.71-1.08)
0.78
(0.47-1.29)
0.94
(0.57-1.56)
0.47
(0.16-1.41)
1.01
(0.91-1.13)
0.89
(0.74-1.06)
Het = heterozygous for risk allele; Hom = homozygous for risk allele; NA = sample size too small to test.
a
ORs adjusted for age (quartiles), bmi (quartiles), T2D (self-report), and age-ethnicity strata (pooled analysis).
b
rs7578597 analysis includes an additional 414 cases and 517 controls for Japanese Americans.
c
rs2237895 and rs2237897 adjusted for each other.
146

Table 6-4: Association of T2D Risk Alleles with PC Risk in Subjects with No Self-Report of T2D (n=4,619)
a
SNP /
Allele
Tested
b

Chr. /
Nearest
Gene
European
Americans
396 cases
477 controls
African
Americans
534 cases
703 controls
Latinos
476 cases
519 controls
Japanese
Americans
576 cases
732 controls
Native
Hawaiians
91 cases
115 controls
Pooled
2,073 cases
2,546 controls
P
value P
het
c
rs10923931
T
1
NOTCH2
0.86
(0.63-1.16)
0.90
(0.76-1.07)
1.09
(0.81-1.47)
0.93
(0.54-1.61)
1.31
(0.57-3.00)
0.94
(0.83-1.07) 0.32 0.70
rs7578597
d
T
2
THADA
0.89
(0.65-1.21)
1.01
(0.84-1.21)
1.26
(0.88-1.79)
4.18
(1.63-10.72)
1.17
(0.46-2.96)
1.07
(0.94-1.23) 0.31 0.025
rs1801282
C
3
PPARG
1.11
(0.81-1.52)
1.04
(0.65-1.66)
1.03
(0.75-1.42)
0.99
(0.65-1.51)
2.13
(0.89-5.10)
1.09
(0.91-1.30) 0.34 0.63
rs4607103

C
3
ADAMTS9
0.98
(0.80-1.21)
1.07
(0.90-1.27)
0.93
(0.77-1.13)
1.11
(0.94-1.30)
1.16
(0.72-1.86)
1.03
(0.94-1.13) 0.51 0.66
rs4402960
T
3
IGF2BP2
1.25
(1.02-1.52)
0.99
(0.85-1.16)
0.98
(0.80-1.21)
1.03
(0.87-1.22)
1.00
(0.64-1.58)
1.05
(0.96-1.14) 0.32 0.41
rs10010131
G
4
WFS1
1.01
(0.83-1.23)
1.10
(0.93-1.30)
1.03
(0.84-1.26)
0.74
(0.41-1.33)
1.57
(0.90-2.74)
1.05
(0.95-1.17) 0.33 0.47
rs7754840
C
6
CDKAL1
0.98
(0.80-1.20)
1.16
(0.99-1.37)
0.95
(0.79-1.14)
1.07
(0.92-1.25)
1.02
(0.67-1.56)
1.05
(0.96-1.14) 0.28 0.55
rs864745
T
7
JAZF1
0.89
(0.73-1.08)
0.93
(0.78-1.11)
0.96
(0.80-1.15)
1.10
(0.90-1.33)
0.73
(0.46-1.15)
0.96
(0.87-1.05) 0.32 0.44
rs2383208

A
9
CDKN2B
1.00
(0.78-1.29)
1.07
(0.87-1.31)
0.95
(0.75-1.21)
1.00
(0.85-1.17)
0.79
(0.49-1.29)
1.00
(0.90-1.10) 0.92 0.83
rs1111875
C
10
HHEX
1.08
(0.89-1.31)
1.05
(0.87-1.27)
1.05
(0.88-1.26)
1.03
(0.87-1.21)
0.98
(0.63-1.50)
1.05
(0.96-1.14) 0.33 0.99

147

Table 6-4: Continued

SNP /
Allele
Tested
b

Chr. /
Nearest
Gene
European
Americans
396 cases
477 controls
African
Americans
534 cases
703 controls
Latinos
476 cases
519 controls
Japanese
Americans
576 cases
732 controls
Native
Hawaiians
91 cases
115 controls
Pooled
2,073 cases
2,546 controls
P
value P
het
c
rs7903146
T
10
TCF7L2
0.98
(0.80-1.21)
0.92
(0.77-1.10)
1.03
(0.84-1.27)
1.08
(0.72-1.63)
1.15
(0.66-1.98)
0.98
(0.88-1.09) 0.75 0.87
rs12779790
G
10
CDC123
1.25
(0.97-1.61)
1.15
(0.92-1.44)
0.93
(0.73-1.17)
0.97
(0.79-1.20)
1.16
(0.69-1.94)
1.06
(0.95-1.19) 0.29 0.39
rs2237895
e
C

11
KCNQ1
0.92
(0.75-1.13)
0.98
(0.80-1.20)
0.84
(0.68-1.03)
1.05
(0.87-1.28)
1.05
(0.66-1.69)
0.96
(0.87-1.06) 0.42 0.26
rs2237897
e
C
11
KCNQ1
1.20
(0.78-1.84)
0.86
(0.65-1.15)
0.91
(0.73-1.14)
1.00
(0.83-1.21)
0.83
(0.49-1.41)
0.96
(0.85-1.08) 0.48 0.38
rs5219
T
11
KCNJ11
1.18
(0.97-1.44)
0.91
(0.69-1.19)
1.01
(0.84-1.21)
0.96
(0.82-1.13)
1.25
(0.82-1.90)
1.03
(0.93-1.13) 0.61 0.38
rs7961581
C
12
TSPAN8
0.95
(0.77-1.16)
0.94
(0.77-1.13)
0.86
(0.69-1.06)
0.83
(0.68-1.01)
0.88
(0.58-1.35)
0.89
(0.80-0.98) 0.015 0.84
rs8050136
A
16
FTO
1.08
(0.89-1.31)
0.94
(0.80-1.11)
1.09
(0.89-1.32)
0.88
(0.72-1.07)
0.81
(0.50-1.29)
0.98
(0.89-1.07) 0.64 0.36
a
Each cell gives odds ratios (and 95% confidence intervals) for allele dosage effects along with the risk allele frequency in controls. ORs adjusted for age (quartiles),
BMI (quartiles), and ethnicity (in pooled analysis).
b
NCBI build 36 (forward strand).
c
P
het
= P value for heterogeneity of allelic effects across ethnic groups (4 df test).
d
rs7578597 analysis includes an additional 362 cases and 411 controls for Japanese Americans.
e
rs2237895 and rs2237897 adjusted for each other.
148
149
Table 6-5: Associations by Disease Severity and Case-Only Testing

OR(95% CI)
a

OR(95% CI)
a

SNP
Chr.
Gene
Allele
Tested
Advanced
Cases

714 ca / 3,317 co
Non-Advanced Cases

1,420 ca / 3,317 co P
het
b
rs10923931

1
NOTCH2 T 0.94(0.79-1.13) 0.91(0.80-1.04) 0.74
rs7578597

2
THADA T 1.10(0.90-1.34) 1.09(0.95-1.26) 0.90
rs1801282

3
PPARG C 1.13(0.88-1.45) 1.11(0.91-1.35) 0.82
rs4607103

3
ADAMTS9 C 1.05(0.93-1.19) 0.95(0.86-1.04) 0.13
rs4402960

3
IGF2BP2 T 1.09(0.96-1.23) 1.01(0.92-1.12) 0.47
rs10010131

4
WFS1 G 1.12(0.97-1.30) 1.06(0.95-1.18) 0.53
rs7754840

6
CDKAL1 C 1.02(0.90-1.15) 1.02(0.93-1.12) 0.76
rs864745

7
JAZF1 T 0.97(0.85-1.10) 0.96(0.87-1.05) 0.94
rs2383208

9
CDKN2B A 0.96(0.84-1.11) 1.04(0.93-1.17) 0.30
rs1111875

10
HHEX C 1.04(0.92-1.18) 1.09(0.99-1.21) 0.62
rs7903146

10
TCF7L2 T 0.97(0.83-1.12) 1.01(0.91-1.13) 0.49
rs12779790

10
CDC123 G 1.07(0.92-1.25) 0.99(0.88-1.12) 0.34
rs2237895
d

11
KCNQ1 C 0.95(0.83-1.08) 1.00(0.90-1.11) 0.45
rs2237897
d

11
KCNQ1 C 1.10(0.93-1.30) 0.96(0.84-1.09) 0.35
Rs5219

11
KCNJ11 T 0.93(0.82-1.06) 1.05(0.95-1.17) 0.15
rs7961581

12
TSPAN8 C 0.90(0.78-1.03) 0.90(0.81-1.00) 0.86
rs8050136

16
FTO A 0.96(0.84-1.09) 0.95(0.87-1.05) 0.70
a
ORs adjusted for BMI (quartiles), age (quartiles), T2D (self-report), and ethnicity.
b
P
het
= p value for heterogeneity (advanced vs. non-advanced).
c
rs7578597 analysis includes an additional 222 advanced cases, 231 non-advanced cases, and 517 controls for
Japanese Americans.
d
rs2237895 and rs2237897 adjusted for each other.

Table 6-6: The Association of Known T2D Risk Alleles with Advanced Prostate Cancer Risk by Race/Ethnicity
a
SNP /
Allele
Tested
b

European
Americans
116 cases
558 controls
African
Americans
180 cases
954 controls
Latinos
193 cases
704 controls
Japanese
Americans
196 cases
936 controls
Native
Hawaiians
29 cases
165 controls
Pooled
714 cases
3,317 controls
P
value P
het
c
rs10923931
T
1.10(0.71-1.69)
0.11
0.90(0.71-1.15)
0.32
1.00(0.68-1.47)
0.10
0.83(0.36-1.88)
0.02
0.58(0.13-2.55)
0.05
0.94(0.79-1.13)
0.14 0.52 0.90
rs7578597
d
T
0.95(0.61-1.48)
0.90
0.98(0.76-1.27)
0.75
1.49(0.91-2.45)
0.92
2.09(0.83-5.27)
0.99
1.52(0.31-7.42)
0.95
1.10(0.90-1.34)
0. 90 0.34 0.35
rs1801282
C
1.16(0.71-1.90)
0. 89
1.43(0.64-3.18)
0.97
0.87(0.60-1.27)
0.91
1.44(0.73-2.84)
0.96
3.91(0.52-29.31)
0.93
1.13(0.88-1.45)
0.94 0.35 0.42
rs4607103

C
0.99(0.71-1.37)
0.74
1.08(0.84-1.39)
0.71
0.95(0.74-1.21)
0.69
1.14(0.91-1.44)
0.62
1.69(0.78-3.66)
0.73
1.05(0.93-1.19)
0.69 0.45 0.51
rs4402960
T
1.18(0.87-1.59)
0.29
1.15(0.92-1.44)
0.52
0.94(0.72-1.22)
0.26
1.14(0.90-1.46)
0.30
0.74(0.37-1.48)
0.29
1.09(0.96-1.23)
0.35 0.19 0.51
rs10010131
G
0.85(0.63-1.14)
0.62
1.31(1.02-1.68)
0.64
1.21(0.93-1.57)
0.71
0.60(0.27-1.29)
0.99
1.53(0.67-3.50)
0. 81
1.12(0.97-1.30)
0.76 0.13 0.087
rs7754840
C
1.05(0.77-1.43)
0.31
1.34(1.06-1.69)
0.54
0.95(0.75-1.21)
0.32
0.79(0.63-1.00)
0.44
1.26(0.68-2.34)
0.56
1.02(0.90-1.15)
0.43 0.78 0.031
rs864745
T
0.88(0.65-1.19)
0.51
1.02(0.80-1.32)
0.72
0.96(0.76-1.22)
0.62
1.06(0.81-1.38)
0.78
0.68(0.35-1.30)
0.77
0.97(0.85-1.10)
0.68 0.61 0.74
rs2383208

A
0.80(0.56-1.16)
0. 82
1.00(0.74-1.34)
0. 81
0.98(0.72-1.33)
0. 85
0.97(0.78-1.21)
0.56
1.24(0.60-2.56)
0.77
0.96(0.84-1.11)
0.75 0.59 0.84
rs1111875
C
1.09(0.81-1.47)
0.58
0.97(0.75-1.27)
0.74
1.22(0.96-1.56)
0.63
0.91(0.71-1.16)
0.28
1.04(0.54-2.00)
0.29
1.04(0.92-1.18)
0.54 0.54 0.52

150

Table 6-6: Continued

SNP /
Allele
Tested
b

European
Americans
116 cases
558 controls
African
Americans
180 cases
954 controls
Latinos
193 cases
704 controls
Japanese
Americans
196 cases
936 controls
Native
Hawaiians
29 cases
165 controls
Pooled
714 cases
3,317 controls
P
value P
het
c
rs7903146
T
0.81(0.59-1.12)
0.32
1.09(0.86-1.40)
0.29
0.95(0.73-1.25)
0.24
0.82(0.46-1.46)
0.05
1.25(0.58-2.69)
0.14
0.97(0.83-1.12)
0.21 0.66 0.55
rs12779790
G
1.63(1.13-2.36)
0.15
1.10(0.81-1.50)
0.15
0.99(0.74-1.32)
0.18
0.95(0.70-1.27)
0.17
0.58(0.25-1.34)
0.21
1.07(0.92-1.25)
0.16 0.39 0.086
rs2237895
e
C

0.87(0.64-1.19)
0.42
0.89(0.67-1.19)
0. 21
1.01(0.78-1.31)
0.44
1.06(0.81-1.39)
0.34
0.71(0.36-1.38)
0.35
0.95(0.83-1.08)
0.34 0.42 0.81
rs2237897
e
C
1.34(0.67-2.68)
0.95
1.26(0.80-1.99)
0.92
0.90(0.67-1.22)
0.78
1.07(0.81-1.42)
0.62
1.63(0.72-3.65)
0.78
1.10(0.93-1.30)
0. 80 0.25 0.66
rs5219
T
0.91(0.67-1.24)
0.36
1.11(0.76-1.60)
0.11
0.95(0.75-1.20)
0.38
0.83(0.65-1.04)
0.37
1.33(0.75-2.37)
0.36
0.93(0.82-1.06)
0.29 0.30 0.52
rs7961581
C
0.84(0.61-1.16)
0.30
0.91(0.69-1.19)
0.23
0.79(0.59-1.05)
0.23
1.09(0.83-1.43)
0.21
0.77(0.39-1.53)
0.32
0.90(0.78-1.03)
0.24 0.12 0.60
rs8050136
A
1.07(0.81-1.43)
0.39
0.97(0.77-1.23)
0.44
1.08(0.84-1.39)
0.28
0.74(0.55-0.99)
0.21
0.73(0.35-1.53)
0.24
0.96(0.84-1.09)
0.32 0.48 0.24
a
Each cell gives odds ratios (and 95% confidence intervals) for allele dosage effects along with the risk allele frequency in controls. ORs adjusted for age (quartiles), BMI
(quartiles), T2D (self-report), and ethnicity (in pooled analysis).
b
NCBI build 36 (forward strand).
c
P
het
= P value for heterogeneity of allelic effects across ethnic groups (4 df test).
d
rs7578597 analysis includes an additional 222 cases and 517 controls for Japanese Americans.
e
rs2237895 and rs2237897 adjusted for each other.
151

Table 6-7: The Association of Known T2D Risk Alleles with Non-Advanced Prostate Cancer Risk by Race/Ethnicity
a
SNP /
Allele
Tested
b

European
Americans
232 cases
558 controls
African
Americans
484 cases
954 controls
Latinos
435 cases
704 controls
Japanese
Americans
239 cases
936 controls
Native
Hawaiians
30 cases
165 controls
Pooled
1,420 cases
3,317 controls
P
value P
het
c
rs10923931
T
0.92(0.65-1.30)
0.11
0.92(0.77-1.08)
0.32
0.93(0.69-1.25)
0.10
0.47(0.18-1.21)
0.02
1.73(0.59-5.02)
0.05
0.91(0.80-1.04)
0.14 0.16 0.49
rs7578597
d
T
0.97(0.68-1.38)
0.90
1.04(0.87-1.25)
0.75
1.23(0.88-1.73)
0.92
5.21(1.28-21.21)
0.99
0.47(0.17-1.28)
0.95
1.09(0.95-1.26)
0. 90 0.24 0.077
rs1801282
C
1.18(0.80-1.72)
0. 89
0.92(0.58-1.45)
0.97
1.12(0.82-1.53)
0.91
1.05(0.60-1.82)
0.96
2.41(0.55-10.57)
0.93
1.11(0.91-1.35)
0.94 0.31 0.76
rs4607103

C
0.86(0.67-1.09)
0.74
1.05(0.89-1.25)
0.71
0.96(0.80-1.16)
0.69
0.90(0.73-1.10)
0.62
0.76(0.40-1.44)
0.73
0.95(0.86-1.04)
0.69 0.28 0.51
rs4402960
T
1.39(1.11-1.74)
0.29
0.91(0.78-1.06)
0.52
1.01(0.84-1.23)
0.26
0.97(0.77-1.21)
0.30
0.89(0.47-1.70)
0.29
1.01(0.92-1.12)
0.35 0.77 0.036
rs10010131
G
1.05(0.84-1.32)
0.62
1.05(0.89-1.24)
0.64
1.06(0.87-1.28)
0.71
1.11(0.45-2.75)
0.99
1.34(0.61-2.95)
0. 81
1.06(0.95-1.18)
0.76 0.29 0.99
rs7754840
C
1.05(0.83-1.32)
0.31
1.08(0.93-1.27)
0.54
1.00(0.83-1.19)
0.32
0.99(0.81-1.22)
0.44
0.60(0.34-1.08)
0.56
1.02(0.93-1.12)
0.43 0.66 0.44
rs864745
T
0.92(0.73-1.15)
0.51
0.95(0.80-1.12)
0.72
0.91(0.76-1.08)
0.62
1.24(0.96-1.60)
0.78
0.59(0.31-1.12)
0.77
0.96(0.87-1.05)
0.68 0.36 0.20
rs2383208

A
0.97(0.72-1.30)
0. 82
1.08(0.88-1.33)
0. 81
0.91(0.72-1.14)
0. 85
1.20(0.98-1.48)
0.56
0.92(0.48-1.76)
0.77
1.04(0.93-1.17)
0.75 0.45 0.46
rs1111875
C
1.17(0.93-1.47)
0.58
1.17(0.97-1.42)
0.74
0.96(0.81-1.14)
0.63
1.05(0.84-1.31)
0.28
2.11(1.18-3.79)
0.29
1.09(0.99-1.21)
0.54 0.076 0.090

152

Table 6-7: Continued

SNP /
Allele
Tested
b

European
Americans
232 cases
558 controls
African
Americans
484 cases
954 controls
Latinos
435 cases
704 controls
Japanese
Americans
239 cases
936 controls
Native
Hawaiians
30 cases
165 controls
Pooled
1,420 cases
3,317 controls
P
value P
het
c
rs7903146
T
1.05(0.83-1.32)
0.32
0.94(0.79-1.12)
0.29
1.10(0.90-1.34)
0.24
1.00(0.60-1.65)
0.05
1.03(0.45-2.34)
0.14
1.01(0.91-1.13)
0.21 0.82 0.80
rs12779790
G
1.12(0.83-1.51)
0.15
0.99(0.80-1.24)
0.15
0.86(0.69-1.08)
0.18
1.01(0.76-1.32)
0.17
1.55(0.80-2.99)
0.21
0.99(0.88-1.12)
0.16 0.90 0.41
rs2237895
e
C

1.10(0.87-1.39)
0.42
0.96(0.79-1.17)
0. 21
0.89(0.74-1.08)
0.44
1.13(0.88-1.45)
0.34
0.70(0.36-1.39)
0.35
1.00(0.90-1.11)
0.34 0.99 0.19
rs2237897
e
C
0.75(0.48-1.16)
0.95
0.87(0.66-1.15)
0.92
0.93(0.75-1.16)
0.78
1.07(0.83-1.38)
0.62
1.47(0.67-3.24)
0.78
0.96(0.84-1.09)
0. 80 0.51 0.26
rs5219
T
1.21(0.96-1.53)
0.36
0.90(0.69-1.17)
0.11
0.97(0.81-1.15)
0.38
1.08(0.87-1.33)
0.37
1.73(1.00-2.99)
0.36
1.05(0.95-1.17)
0.29 0.36 0.15
rs7961581
C
1.11(0.88-1.40)
0.30
0.94(0.79-1.13)
0.23
0.90(0.73-1.11)
0.23
0.64(0.48-0.85)
0.21
0.65(0.33-1.26)
0.32
0.90(0.81-1.00)
0.24 0.052 0.035
rs8050136
A
0.95(0.76-1.19)
0.39
0.91(0.78-1.07)
0.44
1.09(0.90-1.31)
0.28
0.82(0.63-1.07)
0.21
1.06(0.56-1.98)
0.24
0.95(0.87-1.05)
0.32 0.35 0.46
a
Each cell gives odds ratios (and 95% confidence intervals) for allele dosage effects along with the risk allele frequency in controls. ORs adjusted for age (quartiles),
BMI (quartiles), T2D (self-report), and ethnicity (in pooled analysis).
b
NCBI build 36 (forward strand).
c
P
het
= P value for heterogeneity of allelic effects across ethnic groups (4 df test).
d
rs7578597 analysis includes an additional 231 cases and 517 controls for Japanese Americans.
e
rs2237895 and rs2237897 adjusted for each other.
153
154
analysis we did not observe significant differences between the associations in advanced
and non-advanced case subjects.
We tested for gene x gene interactions between the 17 T2D variants from this study
and 28 established prostate cancer risk variants from previous MEC analysis (13, 25, 27).
We observed 18 statistically significant (P≤0.05) interactions. However, none of the
interactions had p values less than 0.01 and among 476 tests; we expected ~24 significant
tests due to chance.
The mean number of risk alleles in each ethnic group ranged from 15.8 in Japanese
Americans to 18.1 in African Americans (Table 6-2). The per allele odds ratios for the
unweighted sum of the total number of risk alleles ranged from 0.98 in Latinos to 1.03 in
European Americans and were not statistically significant in any population (all P-
values≥0.20). In the pooled sample, the mean risk allele count was 17.0 with a range of 8
to 25. The per allele odds ratio of the unweighted summary T2D risk score was 1.00
(95%CI 0.98-1.03; P=0.69). The weighted summary T2D risk scores were also not
significantly associated with PC risk in either the ethnic-specific or pooled analysis (OR
1.02; 95%CI 0.89-1.17; P=0.78). Similar results were observed when subjects were
limited to those with no self-report of T2D (OR 1.00; 95%CI 0.98-1.03; P=0.95).

Table 6-8: The Association of the Summary T2D Risk Scores with Prostate Cancer Risk by Racial/Ethnic Population
a
All subjects, 2,746 cases and 3,317 controls
European
Americans
472 cases
558 controls
African
Americans
721 cases
954 controls
Latinos
668 cases
704 controls
Japanese
Americans
753 cases
936 controls
Native
Hawaiians
132 cases
165 controls
Pooled
2,746 cases
3,317 controls
Average number of
risk alleles
(range)
17.3
(10-24)
18.1
(10-25)
17.0
(9-25)
15.8
(8-23)
17.2
(11-23)
17.0
(8-25)
Q1
b
Q2
Q3
Q4
P trend
1(ref)
1.19(0.83-1.68)
1.27(0.89-1.80)
1.42(0.97-2.06)
0.064
1(ref)
0.97(0.74-1.26)
0.89(0.66-1.22)
0.97(0.74-1.27)
0.76
1(ref)
0.80(0.60-1.07)
0.88(0.62-1.23)
0.95(0.71-1.26)
0.94
1(ref)
1.12(0.87-1.44)
0.83(0.62-1.11)
1.07(0.82-1.41)
0.90
1(ref)
1.58(0.83-3.00)
1.47(0.73-2.97)
0.98(0.50-1.95)
0.86
1(ref)
1.02(0.89-1.17)
0.95(0.82-1.11)
1.05(0.91-1.21)
0.69
Per allele
P value

1.03(0.98-1.08)
0.27
1.00(0.96-1.04)
0.90
0.98(0.94-1.02)
0.34
1.01(0.97-1.06)
0.52
1.00(0.91-1.11)
0.96
1.00(0.98-1.03)
0.69

Subjects with no self-report of Diabetes, 2,746 cases and 3,317 controls
European
Americans
396 cases
477 controls
African
Americans
534 cases
703 controls
Latinos
476 cases
519 controls
Japanese
Americans
576 cases
732 controls
Native
Hawaiians
91 cases
115 controls
Pooled
2,073 cases
2,546 controls
Average number of
risk alleles
(range)
17.2
(10-24)
18.1
(10-25)
17.0
(10-25)
15.6
(8-22)
17.0
(11-22)
18.4
(8-27)
Q1
b
Q2
Q3
Q4
P trend
1(ref)
1.09(0.75-1.58)
1.16(0.80-1.70)
1.45(0.97-2.18)
0.073
1(ref)
1.04(0.77-1.41)
0.95(0.67-1.35)
0.95(0.69-1.30)
0.63
1(ref)
0.75(0.54-1.04)
0.86(0.57-1.29)
0.89(0.64-1.24)
0.70
1(ref)
1.15(0.87-1.51)
0.84(0.60-1.17)
1.03(0.76-1.41)
0.73
1(ref)
1.41(0.65-3.04)
1.39(0.60-3.24)
1.27(0.55-2.93)
0.60
1(ref)
1.01(0.86-1.18)
0.94(0.79-1.13)
1.04(0.88-1.22)
0.83
Per allele
P value

1.03(0.97-1.08)
0.35
1.00(0.96-1.05)
0.99
0.97(0.93-1.02)
0.22
1.01(0.96-1.06)
0.68
1.02(0.90-1.16)
0.74
1.00(0.98-1.03)
0.95
a
Cells gives odds ratios (and 95% confidence intervals) adjusted for age (quartiles), BMI (quartiles), type 2 diabetes (self-report) and ethnicity (in pooled analysis).
b
Quartiles are defined separately for each racial/ethnic population.
155
156
Discussion

Aside for the T2D risk variant rs4430796 in HNF1B, which we and others have
confirmed as a risk factor for prostate cancer in many populations (10, 27), we observed
little evidence in the present study that known T2D risk variants are associated with
prostate cancer risk independently or in combination. These findings provide little
support for the hypothesis that the known common genetic risk variants for T2D
contribute to the inverse association observed between diabetes status and prostate cancer
risk.
Our null findings for the summary risk scores are in contrast to a previous prostate
cancer study of European Americans which found a per allele odds ratio of 0.96 (95%CI
0.92-0.99, P=0.02) for the risk allele count of 17 variants (excluding HNF1B) (22).
Aside from the SLC30A8 allele, the same variants were investigated in both studies
(though highly correlated proxies to the index SNPs were used in the previous study).
The major difference between the two studies is that the present study is comprised of 5
racial/ethnic groups, and thus, one might expect these markers to be poor proxies of the
causal alleles in non-European populations due to differences in LD and the frequency of
the causal allele between racial/ethnic groups. However, we did not observe significant
heterogeneity of effect among racial/ethnic groups in the associations of these variants
with prostate cancer risk. We also previously showed that the majority of these alleles are
associated with T2D risk in these populations and that the risk allele count of these T2D
markers was a strong marker of T2D susceptibility in all 5 populations (Chapter 5)
157
suggesting that if there was an association between T2D susceptibility from these alleles
and prostate cancer risk it would likely be consistent across these populations. This
multiethnic study was well powered to detect associations in the pooled sample, and we
had >78% power (after adjusting for 17 tests) to detect relatively strong associations
(OR~0.85) with common alleles (MAF≥0.25) (Figure 6-1). While unlikely, this does
assume that the variant is equally correlated with the functional allele in each population.
However, our previous observation that these variants are associated with T2D risk in
many of these same populations in the MEC suggests that the loss of power in this
combined multiethnic sample might not be limiting. We were underpowered; however, to
detect associations in any single racial/ethnic group and larger studies in each of these
populations will be needed to for adequate power to detect more modest effects.
We did observe a modestly significant protective association (P=0.021) for the TSPAN8
allele (rs7961581) in the ethnically-pooled analysis. The TSPAN8 gene encodes a cell
surface protein that is part of the tetraspanin superfamily and has been associated with
angiogenesis and endothelial cell activation in a rat adenocarcinoma model (33). This
finding should be interpreted with caution as this association did not remain significant
after adjustment for multiple tests and a previous study reported a small non-significant
increase in prostate cancer risk (OR=1.04) for a correlated SNP (rs1353362) in a
population of European-Americans (22). We also observed an association with the
missense variant (rs7578597, Thr1187Ala) in THADA in Japanese Americans (OR 3.80;
95%CI 1.69-8.55; P=1.3x10
-3
). If true, this association suggests the presence of a
correlated risk variant that may be exclusive to Japanese Americans, however, this

Figure 6-1: Estimated Power to Detect Protective Association for Alleles in Pooled Sample (α = 0.05 with Bonferroni
Correction for 17 tests)
0
0.2
0.4
0.6
0.8
1
0 0.1 0.2 0.3 0.4 0.5
Risk Allele Frequency
Power
OR=0.80
OR=0.85
OR=0.90
OR=0.95

Figure 6-1. Estimated power to detect significant associations (α=0.05 with Bonferroni correction for 17 tests) for alleles with odds ratios of 0.8 (red), 0.85 (yellow), 090
(blue), and (0.95 (green) over a range of minor allele frequencies from 0.0 to 0.5.
158
159
association did not remain significant after adjustment for multiple tests and the 95%
confidence interval for the OR is quite wide because rs7578597 is rare (MAF ~2%) in
Japanese American subjects. We also did not find strong evidence of any interactions
between the 17 T2D variants and 28 established prostate cancer risk alleles. Notably, we
did not find evidence of an interaction (P=0.82) between the T2D allele rs7903146 in
TCF7L2 and the prostate cancer risk allele rs6983267 in the 8q24 region which is of
interest based on evidence that the variant rs6983267 lies in a conical binding site for
TCF7L2 (TCF4) and the alleles of rs6983267 demonstrate differential binding of this
transcription factor (34).
While variants at 3 loci, HNF1B, JAZF1, and THADA, have been associated with
both risk of prostate cancer and T2D, it is not clear that the disease associations at each
locus are due to the same biological process. With the exception of the THADA T2D risk
allele (rs7578597), the SNPs are located in intronic sequence, and it is not known if the
SNPs are markers of functional variants in coding or non-coding areas. It is possible that
the variants are correlated with separate non-coding functional alleles located in
regulatory regions that effect different genes locally or at a distance. Even in HNF1B
where one allele is associated with both diseases it is possible that this marker is linked
with two different functional variants that regulate different genes in separate biological
pathways or even a single functional variant that regulates one or more genes in a tissue-
specific fashion.
In conclusion, we did not find strong evidence of a protective effect on prostate
cancer risk in a multiethnic population for validated T2D risk variants. Once identified,
160
the functional alleles at these loci as well as those at novel T2D risk loci will need to be
reexamined for a role in prostate cancer risk in these populations.

161
Chapter 6 References

1. Kasper JS, Giovannucci E. A meta-analysis of diabetes mellitus and the risk of
prostate cancer. Cancer Epidemiol Biomarkers Prev 2006;1511:2056-62.

2. Giovannucci E, Rimm EB, Stampfer MJ, et al. Diabetes mellitus and risk of
prostate cancer (United States). Cancer Causes Control 1998;91:3-9.

3. Waters KM, Henderson BE, Stram DO, et al. Association of diabetes with
prostate cancer risk in the multiethnic cohort. Am J Epidemiol 2009;1698:937-45.

4. Calton BA, Chang SC, Wright ME, et al. History of diabetes mellitus and
subsequent prostate cancer risk in the NIH-AARP Diet and Health Study. Cancer
Causes Control 2007;185:493-503.

5. Gong Z, Neuhouser ML, Goodman PJ, et al. Obesity, diabetes, and risk of
prostate cancer: results from the prostate cancer prevention trial. Cancer
Epidemiol Biomarkers Prev 2006;1510:1977-83.

6. Velicer CM, Dublin S, White E. Diabetes and the risk of prostate cancer: the role
of diabetes treatment and complications. Prostate Cancer Prostatic Dis
2007;101:46-51.

7. Pierce BL, Plymate S, Ostrander EA, et al. Diabetes mellitus and prostate cancer
risk. Prostate 2008;6810:1126-32.

8. Tavani A, Gallus S, Bosetti C, et al. Diabetes and the risk of prostate cancer. Eur J
Cancer Prev 2002;112:125-8.

9. Winckler W, Weedon MN, Graham RR, et al. Evaluation of common variants in
the six known maturity-onset diabetes of the young (MODY) genes for
association with type 2 diabetes. Diabetes 2007;563:685-93.

10. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome
17 confer prostate cancer risk, and the one in TCF2 protects against type 2
diabetes. Nat Genet 2007;398:977-83.

11. Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified in a genome-wide
association study of prostate cancer. Nat Genet 2008;403:310-5.

12. Zeggini E, Scott LJ, Saxena R, et al. Meta-analysis of genome-wide association
data and large-scale replication identifies additional susceptibility loci for type 2
diabetes. Nat Genet 2008;405:638-45.
162
13. Eeles RA, Kote-Jarai Z, Al Olama AA, et al. Identification of seven new prostate
cancer susceptibility loci through a genome-wide association study. Nat Genet
2009;4110:1116-21.

14. Johansson A, Marroni F, Hayward C, et al. Common variants in the JAZF1 gene
associated with height identified by linkage and genome-wide association
analysis. Hum Mol Genet 2009;182:373-80.

15. Dudziak K, Mottalebi N, Senkel S, et al. Transcription factor HNF1beta and novel
partners affect nephrogenesis. Kidney Int 2008;742:210-7.

16. Fajans SS, Bell GI, Polonsky KS. Molecular mechanisms and clinical
pathophysiology of maturity-onset diabetes of the young. N Engl J Med
2001;34513:971-80.

17. Nakajima T, Fujino S, Nakanishi G, et al. TIP27: a novel repressor of the nuclear
orphan receptor TAK1/TR4. Nucleic Acids Res 2004;3214:4194-204.

18. Collins LL, Lee YF, Heinlein CA, et al. Growth retardation and abnormal
maternal behavior in mice lacking testicular orphan nuclear receptor 4. Proc Natl
Acad Sci U S A 2004;10142:15058-63.

19. Li H, Ma X, Wang J, et al. Effects of rearrangement and allelic exclusion of
JJAZ1/SUZ12 on cell proliferation and survival. Proc Natl Acad Sci U S A
2007;10450:20001-6.

20. Rippe V, Drieschner N, Meiboom M, et al. Identification of a gene rearranged by
2p21 aberrations in thyroid adenomas. Oncogene 2003;2238:6111-4.

21. Simonis-Bik AM, Nijpels G, van Haeften TW, et al. Gene variants in the novel
type 2 diabetes loci CDC123/CAMK1D, THADA, ADAMTS9, BCL11A, and
MTNR1B affect different aspects of pancreatic beta-cell function.
Diabetes;591:293-301.

22. Pierce BL, Ahsan H. Genetic susceptibility to type 2 diabetes is associated with
reduced prostate cancer risk. Hum Hered;693:193-201.

23. Meyer TE, Boerwinkle E, Morrison AC, et al. Diabetes genes and prostate cancer
in the Atherosclerosis Risk in Communities study. Cancer Epidemiol Biomarkers
Prev;192:558-65.

24. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and
Los Angeles: baseline characteristics. Am J Epidemiol 2000;1514:346-57.
163
25. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24
independently affect risk for prostate cancer. Nat Genet 2007;395:638-44.

26. Lee LG, Connell CR, Bloch W. Allelic discrimination by nick-translation PCR
with fluorogenic probes. Nucleic Acids Res 1993;2116:3761-6.

27. Waters KM, Le Marchand L, Kolonel LN, et al. Generalizability of associations
from prostate cancer genome-wide association studies in multiple populations.
Cancer Epidemiol Biomarkers Prev 2009;184:1285-9.

28. Sladek R, Rocheleau G, Rung J, et al. A genome-wide association study identifies
novel risk loci for type 2 diabetes. Nature 2007;4457130:881-5.

29. Scott LJ, Mohlke KL, Bonnycastle LL, et al. A genome-wide association study of
type 2 diabetes in Finns detects multiple susceptibility variants. Science
2007;3165829:1341-5.

30. Gloyn AL, Weedon MN, Owen KR, et al. Large-scale association studies of
variants in genes encoding the pancreatic beta-cell KATP channel subunits Kir6.2
(KCNJ11) and SUR1 (ABCC8) confirm that the KCNJ11 E23K variant is
associated with type 2 diabetes. Diabetes 2003;522:568-72.

31. Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al. A variant in CDKAL1
influences insulin response and risk of type 2 diabetes. Nat Genet 2007;396:770-
5.

32. Jonsson A, Isomaa B, Tuomi T, et al. A variant in the KCNQ1 gene predicts
future type 2 diabetes and mediates impaired insulin secretion. Diabetes
2009;5810:2409-13.

33. Nazarenko I, Rana S, Baumann A, et al. Cell surface tetraspanin Tspan8
contributes to molecular pathways of exosome-induced endothelial cell activation.
Cancer Res;704:1668-78.

34. Pomerantz MM, Ahmadiyeh N, Jia L, et al. The 8q24 cancer risk variant
rs6983267 shows long-range interaction with MYC in colorectal cancer. Nat
Genet 2009;418:882-4.

164
Chapter 7: Summary
Recent breakthroughs in SNP discovery and LD pattern definition in the human
genome as well as the development of large-scale genotyping technologies have allowed
the comprehensive investigation of common variation in association with many
phenotypes. For prostate cancer (PC), this has lead to the identification of thirty loci to
date that have been reproducibly associated with the disease through multi-stage GWAS.
As with most breakthroughs, these discoveries have led to new avenues of exploration. I
examined the generalizability of 13 of these loci to five distinct racial/ethnic populations.
Many of the markers were generalizable to multiple populations, and the markers as a
whole contributed significantly to risk in each population. One limitation in this study
was that there was not adequate power for individual alleles in each racial/ethnic group.
The use of multiple racial/ethnic populations in future deep sequencing and fine-mapping
studies of each risk locus will be a powerful tool to locate alleles that are functional and
to identify additional independent functional variants that are common in non-European
populations. To date, there is no evidence of the association of these markers with
disease aggressiveness. Determining the role of genetic variation in PC, and more
specifically aggressive disease will be important as this knowledge may one day be used
to prevent disease and/or inform treatment decisions for men who are at greater risk of
dying from the disease. This is extremely important since most men diagnosed with PC
do not die of the disease. The majority of the risk markers are in non-coding areas of the
genome, and the biology behind their association with disease risk is unknown.
Collaboration with laboratories that can address functionality will be a key aspect in
165
furthering our knowledge about the disease pathway of each locus. As demonstrated at
8q24 (1), experiments can guide association testing by pointing to areas within a risk
locus likely to be of functional importance.
A genotype-biomarker relationship was investigated for a known PC risk locus.
Strong evidence emerged to the hypothesis that rs10993994 influences expression of
MSMB by a robust association between rs10993994 genotype and blood plasma levels of
MSP that was consistent in four racial/ethnic populations in elderly men unaffected by
PC. This single variant was responsible for a large proportion of the variance in MSP
levels. Future experiments will be needed to study day-to-day inter-individual variance,
dietary factors, social factors, local genetic variation around MSMB, and genome-wide
genetic variation in order to fully understand the variance in circulating levels of MSP.
At this time, the function of MSP is not known, and it is possible that the association of
rs10993994 with PC is a result of an association with the regulation of a different gene.
Measuring rs10993994 genotype and plasma MSP levels in a prospective PC case-control
study would provide the data to test the hypothesis that decreased MSP levels are a PC
risk factor independent of rs10993994 and provide valuable information regarding the
utility of plasma MSP as a screening biomarker for PC. Concurrent measurements of
circulating PSA levels would also provide an avenue to test the combined effects of these
two biomarkers and their relative predictive value. If proven to be a risk marker for PC,
the biologic function of MSP will need to be determined.
The previously observed inverse association between PC and diabetes was
consistent across multiple racial/ethnic populations. We observed that diabetes status
166
was protective for PC in our multiethnic population. Men with diabetes had, on average,
lower PSA levels than non-diabetic men in this multiethnic population as was shown
previously by others. Lastly, PSA screening rates among diabetic and non-diabetic men
explained only a small portion of the inverse association. There are numerous future
directions for this line of experimentation as the two diseases have no established
environmental risk factors in common and an explanation for this inverse association is
still unknown. One hypothesis to explain the inverse association between T2D status and
PC risk is that the decrease in risk is caused by the diabetic disease state. This hypothesis
could be tested through prospective analysis of the association of longitudinal
measurements of serum levels of androgens, insulin, glucose, IGF-1, estrogens, leptin,
and other biomarkers suspected to be associated with the diabetic state with PC risk in
both men with and without diabetes. Genetic risk factors common to both diseases are a
possible explanation for the inverse association as three loci have already been detected
that are associated with both diseases.
Common risk variants for the consistent association with T2D risk in the
racial/ethnic groups of the MEC were further examined. Despite limited power for SNPs
in individual ethnic groups significant associations with risk for the majority of alleles
(14/19) and positive associations (OR>1) for all 19 SNPs were found. A summary risk
score of the aggregate sum of risk alleles was strongly associated with risk in all 5
racial/ethnic groups. The necessary future experimentation for these T2D SNPs is similar
to those needed for PC. Many of these SNPs also lie in non-coding regions and
functional experiments will be necessary to determine their role in T2D pathogenesis.
167
Unlike PC, quantitative traits exist for T2D, and studies examining the association of
variants with fasting plasma glucose, insulin sensitivity, and insulin response to glucose
are ongoing and may aid in determining their possible mechanism of action in T2D
pathogenesis.
Based on the per allele T2D risk scores, it does not appear that this set of markers
explains ethnic differences in T2D prevalence. Improved localization of the functional
alleles may add to the ability to determine the role of genetic risk variants in ethnic
differences in T2D risk as it is unlikely that many of the variants among the current set of
risk markers are the functional variant. The identification of functional alleles will also
aid in improving the validity of the summary risk scores used in our genetic case-control
studies. Common markers weakly (or not at all) associated with the functional variants in
a particular racial/ethnic group introduce measurement error into the risk score. Until the
functional variants are identified and used in the risk scores, comparisons of mean risk
allele counts among racial/ethnic groups will be susceptible to this measurement error
and we will be unable to differentiate whether the per allele associations across
racial/ethnic groups are indicative of differences in LD between the markers and
functional variants or of differences in effect across racial/ethnic groups. Another
limitation of the risk score is that it does not fully capture risk loci with multiple risk
variants. Until functional variants are identified, the risk scores should be interpreted as a
crude global measure of the effectiveness of the risk markers in each racial/ethnic group,
and should not be interpreted as measures of differences in genetic risk conferred by the
loci.
168
The finding that many of the PC and T2D risk variants are associated with risk in
multiple racial/ethnic populations does not support the recent “common SNP rare
mutation” hypothesis which predicts that associations with common SNPs are the result
of “synthetic associations” with one or more rare causal mutations. One expectation of
this hypothesis is that risk associations would not be consistent across ethnic groups
because most rare causal mutations are relatively recent and occurred after the continental
migrations that formed the current continental racial/ethnic groups. However, this thesis
does not disprove the presence of “synthetic associations” or that rare mutations with
high penetrance may be the source of missing heritability. It was recently reported that a
model including the log-additive effects nearly 295,000 genome-wide SNPs was able to
explain 45% of the estimated of variance in height (compared to 5% explained by
established markers) (2). The authors suggested that the remaining heritability (estimated
to be 80%) in height was due to large numbers of functional variants that were relatively
rare (MAF<0.10) and thus were weakly correlated to the more common SNPs included in
the analysis. Current study designs with stringent criteria to achieve genome-wide
significance are unlikely to detect these loci. While it is possible that men and women at
high risk for PC and T2D are more likely to have traits that decrease fitness, it is unlikely
given the late onset of these diseases. This would suggest that the functional risk variants
of these two diseases would not be under evolutionary pressure and would have the same
distribution of minor allele frequencies as non-functional SNPs. Assuming no
evolutionary pressure, the minor allele frequency of risk variants for PC and T2D would
not be associated with effect size and there would likely be a very large number of rare
169
risk variants with effect sizes similar to the ones currently being reported among common
alleles. While rare alleles are likely to be important for common diseases their overall
contribution to missing heritability has yet to examined. Ongoing efforts, such as the
1000’s Genome Project as well as targeted sequencing of the known risk loci in large
numbers of cases and controls of multiple ethnicities, which will soon be followed by
genome-wide sequencing, will provide empirical data needed to address the abundance
and overall contribution of rare variants to disease susceptibility and familial risk.
Demonstrating the consistency of association of the T2D variants with T2D risk
across racial/ethnic groups provided us with a basis to examine these variants for
pleiotropic effects on PC. In chapter 6 of this thesis we examined 17 of the 19 T2D risk
variants (the HNF1B allele was included in chapter 2) for association with PC. We did
not find strong evidence of an association for either the alleles individually or as an
aggregate predictor of T2D susceptibility. We did find a modest protective effect for the
TSPAN8 allele, and a nominally significant positive association for the THADA allele in
Japanese Americans. These findings suggest that the bulk of these alleles do not play a
role in the inverse association between PC and T2D. However, we did have limited
power to detect findings in individual SNPs for each racial/ethnic group and future
studies may have more success with larger sample sizes and testing the causal alleles for
T2D once they have been identified. The observation that T2D was associated with a
decreased risk of PC in multiple populations suggests that a potential risk variant(s)
associated with both diseases will be common to multiple populations and it is unlikely
that the association in each racial/ethnic group is caused by a different rare variant. In
170
searching for risk loci with a common variant(s) that is associated with the risk of both
diseases, the known PC risk variants are good candidates as they have already been
associated with PC and should also be studied in association with T2D risk. The loci that
have already been associated with both diseases (HNF1B, JAZF1, and THADA) should be
fine-mapped to identify common pan-ethnic variants across each locus, and tag the region
for association analysis with both diseases. The HNF1B allele is the only established
genetic risk variant associated with both diseases, and it is possible that the known PC
and T2D risk variants play a very limited role in the association between T2D and PC
risk. A GWAS that compares allele frequencies between men with PC and men with
T2D, ideally sampled from the same source population, may be the most efficient
approach to detect novel common variants associated with the risk of both diseases.
Candidate gene studies comparing allele frequencies between men with diabetes and men
with PC are another possible avenue to identify risk loci associated with the risk of both
diseases. These studies could focus on genes known to be involved with the activity,
production, and metabolism of the biomarkers hypothesized to play a role in the
decreased PC risk among diabetic men (e.g. androgens, insulin, glucose, IGF-1,
estrogens, leptin).
In conclusion, deep sequencing and fine-mapping of risk loci in large multiethnic
populations will be needed to fully elucidate the genetic variation in each locus and their
associations with each disease. GWAS in multiple racial/ethnic populations will also be
necessary to provide power to identify loci that may be more common or limited to
populations of non-European ancestry. Lastly, the majority of the established risk
171
variants lie in non-coding regions, and the mechanism of disease association is not
known. While the risk variants may still be markers of yet to be identified
functional/causal alleles in coding regions, this does suggest that non-coding regions play
an important role in disease risk. Localizing the functional alleles and testing their
association in large studies will be needed to determine their contribution to ethnic
differences in risk and clinical utility as predictive markers of risk.

172
Chapter 7 References

1. Jia L, Landan G, Pomerantz M, et al. Functional enhancers at the gene-poor 8q24
cancer-linked locus. PLoS Genet 2009;58:e1000597.

2. Yang J, Benyamin B, McEvoy BP, et al. Common SNPs explain a large
proportion of the heritability for human height. Nat Genet 2010;427:565-9.

.

173
References

Abate N, Chandalia M. The impact of ethnicity on type 2 diabetes. J Diabetes
Complications 2003;171:39-58.

Abate-Shen C, Shen MM, Gelmann E. Integrating differentiation and cancer: the Nkx3.1
homeobox gene in prostate organogenesis and carcinogenesis. Differentiation
2008;766:717-27.

Ahn J, Berndt SI, Wacholder S, et al. Variation in KLK genes, prostate-specific antigen
and risk of prostate cancer. Nat Genet 2008;409:1032-4; author reply 5-6.

Ahn J, Schumacher FR, Berndt SI, et al. Quantitative trait loci predicting circulating sex
steroid hormones in men from the NCI-Breast and Prostate Cancer Cohort Consortium
(BPC3). Hum Mol Genet 2009;1819:3749-57.

Al Olama AA, Kote-Jarai Z, Giles GG, et al. Multiple loci on 8q24 associated with
prostate cancer susceptibility. Nat Genet 2009;4110:1058-60.

Altshuler D, Hirschhorn JN, Klannemark M, et al. The common PPARgamma Pro12Ala
polymorphism is associated with decreased risk of type 2 diabetes. Nat Genet
2000;261:76-80.

Aluli NE. Prevalence of obesity in a Native Hawaiian population. Am J Clin Nutr
1991;536 Suppl:1556S-60S.

Amundadottir LT, Sulem P, Gudmundsson J, et al. A common variant associated with
prostate cancer in European and African populations. Nat Genet 2006;386:652-8.

Anahi Franchi N, Avendano C, Molina RI, et al. beta-Microseminoprotein in human
spermatozoa and its potential role in male fertility. Reproduction 2008;1362:157-66.

Berry R, Schroeder JJ, French AJ, et al. Evidence for a prostate cancer-susceptibility
locus on chromosome 20. Am J Hum Genet 2000;671:82-91.

Berthon P, Valeri A, Cohen-Akenine A, et al. Predisposing gene for early-onset prostate
cancer, localized on chromosome 1q42.2-43. Am J Hum Genet 1998;626:1416-24.

Bjartell AS, Al-Ahmadie H, Serio AM, et al. Association of cysteine-rich secretory
protein 3 and beta-microseminoprotein with outcome after radical prostatectomy. Clin
Cancer Res 2007;1314:4130-8.

174
Borrell LN, Crawford ND, Dailo FJ. Race/ethnicity and self-reported diabetes among
adults in the National Health Interview Survey: 2000-2003. Public Health Rep
2007;1225:616-25.

Bouatia-Naji N, Bonnefond A, Cavalcanti-Proenca C, et al. A variant near MTNR1B is
associated with increased fasting plasma glucose levels and type 2 diabetes risk. Nat
Genet 2009;411:89-94.

Bouatia-Naji N, Rocheleau G, Van Lommel L, et al. A polymorphism within the G6PC2
gene is associated with fasting plasma glucose levels. Science 2008;3205879:1085-8.

Brawley OW, Jani AB, Master V. Prostate cancer and race. Curr Probl Cancer
2007;313:211-25.

Calton BA, Chang SC, Wright ME, et al. History of diabetes mellitus and subsequent
prostate cancer risk in the NIH-AARP Diet and Health Study. Cancer Causes Control
2007;185:493-503.

Carpten J, Nupponen N, Isaacs S, et al. Germline mutations in the ribonuclease L gene in
families showing linkage with HPC1. Nat Genet 2002;302:181-4.

Camp NJ, Cannon-Albright LA, Farnham JM, et al. Compelling evidence for a prostate
cancer gene at 22q12.3 by the International Consortium for Prostate Cancer Genetics.
Hum Mol Genet 2007;1611:1271-8.

Camp NJ, Farnham JM, Cannon-Albright LA. Localization of a prostate cancer
predisposition gene to an 880-kb region on chromosome 22q12.3 in Utah high-risk
pedigrees. Cancer Res 2006;6620:10205-12.

Camp NJ, Tavtigian SV. Meta-analysis of associations of the Ser217Leu and Ala541Thr
variants in ELAC2 (HPC2) and prostate cancer. Am J Hum Genet 2002;716:1475-8.

Carter BS, Beaty TH, Steinberg GD, et al. Mendelian inheritance of familial prostate
cancer. Proc Natl Acad Sci U S A 1992;898:3367-71.

Chambers JC, Elliott P, Zabaneh D, et al. Common genetic variation near MC4R is
associated with waist circumference and insulin resistance. Nat Genet 2008;406:716-8.

Chan JM, Latini DM, Cowan J, et al. History of diabetes, clinical features of prostate
cancer, and prostate cancer recurrence-data from CaPSURE (United States). Cancer
Causes Control 2005;167:789-97.

175
Chang BL, Cramer SD, Wiklund F, et al. Fine mapping association study and functional
analysis implicate a SNP in MSMB at 10q11 as a causal variant for prostate cancer risk.
Hum Mol Genet 2009;187:1368-75.

Chen WM, Erdos MR, Jackson AU, et al. Variations in the G6PC2/ABCB11 genomic
region are associated with fasting glucose levels. J Clin Invest 2008;1187:2620-8.

Chen YC, Kraft P, Bretsky P, et al. Sequence variants of estrogen receptor beta and risk
of prostate cancer in the National Cancer Institute Breast and Prostate Cancer Cohort
Consortium. Cancer Epidemiol Biomarkers Prev 2007;1610:1973-81.

Cheng CY, Kao WH, Patterson N, et al. Admixture mapping of 15,280 African
Americans identifies obesity susceptibility loci on chromosomes 5 and X. PLoS Genet
2009;55:e1000490.

Cheng I, Penney KL, Stram DO, et al. Haplotype-based association studies of IGFBP1
and IGFBP3 with prostate and breast cancer risk: the multiethnic cohort. Cancer
Epidemiol Biomarkers Prev 2006;1510:1993-7.

Cheng I, Stram DO, Penney KL, et al. Common genetic variation in IGF1 and prostate
cancer risk in the Multiethnic Cohort. J Natl Cancer Inst 2006;982:123-34.

Cheng I, Yu MC, Koh WP, et al. Comparison of prostate-specific antigen and hormone
levels among men in Singapore and the United States. Cancer Epidemiol Biomarkers
Prev 2005;147:1692-6.

Chibalina MV, Seaman MN, Miller CC, et al. Myosin VI and its interacting protein
LMTK2 regulate tubule formation and transport to the endocytic recycling compartment.
J Cell Sci 2007;120Pt 24:4278-88.

Christensen GB, Baffoe-Bonnie AB, George A, et al. Genome-wide linkage analysis of
1,233 prostate cancer pedigrees from the International Consortium for Prostate Cancer
Genetics using novel sumLINK and sumLOD analyses. Prostate 2010;707:735-44.

Coker AL, Sanderson M, Zheng W, et al. Diabetes mellitus and prostate cancer risk
among older men: population-based case-control study. Br J Cancer 2004;9011:2171-5.

Collin SM, Martin RM, Metcalfe C, et al. Prostate-cancer mortality in the USA and UK
in 1975-2004: an ecological study. Lancet Oncol 2008;95:445-52.

Collins LL, Lee YF, Heinlein CA, et al. Growth retardation and abnormal maternal
behavior in mice lacking testicular orphan nuclear receptor 4. Proc Natl Acad Sci U S A
2004;10142:15058-63.

176
Conlon EM, Goode EL, Gibbs M, et al. Oligogenic segregation analysis of hereditary
prostate cancer pedigrees: evidence for multiple loci affecting age at onset. Int J Cancer
2003;1055:630-5.

Crowe FL, Key TJ, Appleby PN, et al. Dietary fat intake and risk of prostate cancer in the
European Prospective Investigation into Cancer and Nutrition. Am J Clin Nutr
2008;875:1405-13.

Cui J, Staples MP, Hopper JL, et al. Segregation analyses of 1,476 population-based
Australian families affected by prostate cancer. Am J Hum Genet 2001;685:1207-18.

Deo RC, Patterson N, Tandon A, et al. A high-density admixture scan in 1,670 African
Americans with hypertension. PLoS Genet 2007;311:e196.

Dickson SP, Wang K, Krantz I, et al. Rare Variants Create Synthetic Genome-Wide
Associations. PLoS Biol 2010;81:e1000294.

Di Mascio P, Kaiser S, Sies H. Lycopene as the most efficient biological carotenoid
singlet oxygen quencher. Arch Biochem Biophys 1989;2742:532-8.

Dossus L, Kaaks R, Canzian F, et al. PTGS2 and IL6 genetic variation and risk of breast
and prostate cancer: results from the Breast and Prostate Cancer Cohort Consortium
(BPC3). Carcinogenesis;313:455-61.

Dudziak K, Mottalebi N, Senkel S, et al. Transcription factor HNF1beta and novel
partners affect nephrogenesis. Kidney Int 2008;742:210-7.

Duggan D, Zheng SL, Knowlton M, et al. Two genome-wide association studies of
aggressive prostate cancer implicate putative prostate tumor suppressor gene DAB2IP. J
Natl Cancer Inst 2007;9924:1836-44.

Duggirala R, Blangero J, Almasy L, et al. Linkage of type 2 diabetes mellitus and of age
at onset to a genetic location on chromosome 10q in Mexican Americans. Am J Hum
Genet 1999;644:1127-40.

Easton DF, Pooley KA, Dunning AM, et al. Genome-wide association study identifies
novel breast cancer susceptibility loci. Nature 2007;4477148:1087-93.

Eeles RA, Kote-Jarai Z, Al Olama AA, et al. Identification of seven new prostate cancer
susceptibility loci through a genome-wide association study. Nat Genet 2009;4110:1116-
21.

Eeles RA, Kote-Jarai Z, Giles GG, et al. Multiple newly identified loci associated with
prostate cancer susceptibility. Nat Genet 2008;403:316-21.
177

Fajans SS, Bell GI, Polonsky KS. Molecular mechanisms and clinical pathophysiology of
maturity-onset diabetes of the young. N Engl J Med 2001;34513:971-80.

Florez JC, Price AL, Campbell D, et al. Strong association of socioeconomic status with
genetic ancestry in Latinos: implications for admixture studies of type 2 diabetes.
Diabetologia 2009;528:1528-36.

Fowke JH, Signorello LB, Chang SS, et al. Effects of obesity and height on prostate-
specific antigen (PSA) and percentage of free PSA levels among African-American and
Caucasian men. Cancer 2006;10710:2361-7.

Freedman ML, Haiman CA, Patterson N, et al. Admixture mapping identifies 8q24 as a
prostate cancer risk locus in African-American men. Proc Natl Acad Sci U S A
2006;10338:14068-73.

Freedman ML, Pearce CL, Penney KL, et al. Systematic evaluation of genetic variation at
the androgen receptor locus and risk of prostate cancer in a multiethnic cohort study. Am
J Hum Genet 2005;761:82-90.

Friberg E, Orsini N, Mantzoros CS, et al. Diabetes mellitus and risk of endometrial
cancer: a meta-analysis. Diabetologia 2007;507:1365-74.

Friedrichsen DM, Stanford JL, Isaacs SD, et al. Identification of a prostate cancer
susceptibility locus on chromosome 7q11-21 in Jewish families. Proc Natl Acad Sci U S
A 2004;1017:1939-44.

Fukui M, Tanaka M, Kadono M, et al. Serum prostate-specific antigen levels in men with
type 2 diabetes. Diabetes Care 2008;315:930-1.

Gann PH, Ma J, Giovannucci E, et al. Lower prostate cancer risk in men with elevated
plasma lycopene levels: results of a prospective analysis. Cancer Res 1999;596:1225-30.

Garde SV, Basrur VS, Li L, et al. Prostate secretory protein (PSP94) suppresses the
growth of androgen-independent prostate cancer cell line (PC3) and xenografts by
inducing apoptosis. Prostate 1999;382:118-25.

Gibbs M, Stanford JL, Jarvik GP, et al. A genomic scan of families with prostate cancer
identifies multiple regions of interest. Am J Hum Genet 2000;671:100-9.

Gibbs M, Stanford JL, McIndoe RA, et al. Evidence for a rare prostate cancer-
susceptibility locus at chromosome 1p36. Am J Hum Genet 1999;643:776-87.

178
Gill JK, Franke AA, Steven Morris J, et al. Association of selenium, tocopherols,
carotenoids, retinol, and 15-isoprostane F(2t) in serum or urine with prostate cancer risk:
the multiethnic cohort. Cancer Causes Control 2009;207:1161-71.

Gillanders EM, Xu J, Chang BL, et al. Combined genome-wide scan for prostate cancer
susceptibility genes. J Natl Cancer Inst 2004;9616:1240-7.

Giovannucci E, Ascherio A, Rimm EB, et al. Intake of carotenoids and retinol in relation
to risk of prostate cancer. J Natl Cancer Inst 1995;8723:1767-76.

Giovannucci E, Rimm EB, Liu Y, et al. A prospective study of tomato products,
lycopene, and prostate cancer risk. J Natl Cancer Inst 2002;945:391-8.

Giovannucci E, Rimm EB, Stampfer MJ, et al. Diabetes mellitus and risk of prostate
cancer (United States). Cancer Causes Control 1998;91:3-9.

Giovannucci E, Rimm EB, Stampfer MJ, et al. Height, body weight, and risk of prostate
cancer. Cancer Epidemiol Biomarkers Prev 1997;68:557-63.

Giovannucci E, Stampfer MJ, Krithivas K, et al. The CAG repeat within the androgen
receptor gene and its relationship to prostate cancer. Proc Natl Acad Sci U S A
1997;947:3320-3.

Gloyn AL, Weedon MN, Owen KR, et al. Large-scale association studies of variants in
genes encoding the pancreatic beta-cell KATP channel subunits Kir6.2 (KCNJ11) and
SUR1 (ABCC8) confirm that the KCNJ11 E23K variant is associated with type 2
diabetes. Diabetes 2003;522:568-72.

Goldstein DB. Common genetic variation and human traits. N Engl J Med
2009;36017:1696-8.

Gong Z, Neuhouser ML, Goodman PJ, et al. Obesity, diabetes, and risk of prostate
cancer: results from the prostate cancer prevention trial. Cancer Epidemiol Biomarkers
Prev 2006;1510:1977-83.

Gong G, Oakley-Girvan I, Wu AH, et al. Segregation analysis of prostate cancer in 1,719
white, African-American and Asian-American families in the United States and Canada.
Cancer Causes Control 2002;135:471-82.

Grant SF, Thorleifsson G, Reynisdottir I, et al. Variant of transcription factor 7-like 2
(TCF7L2) gene confers risk of type 2 diabetes. Nat Genet 2006;383:320-3.

Gronberg H, Damber L, Damber JE, et al. Segregation analysis of prostate cancer in
Sweden: support for dominant inheritance. Am J Epidemiol 1997;1467:552-7.
179

Gudmundsson J, Sulem P, Gudbjartsson DF, et al. Genome-wide association and
replication studies identify four variants associated with prostate cancer susceptibility.
Nat Genet 2009;4110:1122-6.

Gudmundsson J, Sulem P, Manolescu A, et al. Genome-wide association study identifies
a second prostate cancer susceptibility variant at 8q24. Nat Genet 2007;395:631-7.

Gudmundsson J, Sulem P, Rafnar T, et al. Common sequence variants on 2p15 and
Xp11.22 confer susceptibility to prostate cancer. Nat Genet 2008;403:281-3.

Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome 17
confer prostate cancer risk, and the one in TCF2 protects against type 2 diabetes. Nat
Genet 2007;398:977-83.

Guilherme A, Soriano NA, Furcinitti PS, et al. Role of EHD1 and EHBP1 in perinuclear
sorting and insulin-regulated GLUT4 recycling in 3T3-L1 adipocytes. J Biol Chem
2004;27938:40062-75.

Haiman CA, Le Marchand L, Yamamato J, et al. A common genetic risk factor for
colorectal and prostate cancer. Nat Genet 2007;398:954-6.

Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24
independently affect risk for prostate cancer. Nat Genet 2007;395:638-44.

Haiman CA, Stram DO, Cheng I, et al. Common genetic variation at PTEN and risk of
sporadic breast and prostate cancer. Cancer Epidemiol Biomarkers Prev 2006;155:1021-
5.

Hall JM, Lee MK, Newman B, et al. Linkage of early-onset familial breast cancer to
chromosome 17q21. Science 1990;2504988:1684-9.

Hunter DJ, Riboli E, Haiman CA, et al. A candidate gene approach to searching for low-
penetrance breast and prostate cancer genes. Nat Rev Cancer 2005;512:977-85.

Huxley R, James WP, Barzi F, et al. Ethnic comparisons of the cross-sectional
relationships between measures of body size with diabetes and hypertension. Obes Rev
2008;9 Suppl 1:53-61.

Imasato Y, Xuan JW, Sakai H, et al. PSP94 expression after androgen deprivation
therapy: a comparative study with prostate specific antigen in benign prostate and
prostate cancer. J Urol 2000;1645:1819-24.

180
Jankowska AM, Szpurka H, Tiu RV, et al. Loss of heterozygosity 4q24 and TET2
mutations associated with myelodysplastic/myeloproliferative neoplasms. Blood
2009;11325:6403-10.

Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2009. CA Cancer J Clin
2009;594:225-49.

Jia L, Landan G, Pomerantz M, et al. Functional enhancers at the gene-poor 8q24 cancer-
linked locus. PLoS Genet 2009;58:e1000597.

Johanneson B, McDonnell SK, Karyadi DM, et al. Fine mapping of familial prostate
cancer families narrows the interval for a susceptibility locus on chromosome 22q12.3 to
1.36 Mb. Hum Genet 2008;1231:65-75.

Johansson A, Marroni F, Hayward C, et al. Common variants in the JAZF1 gene
associated with height identified by linkage and genome-wide association analysis. Hum
Mol Genet 2009;182:373-80.

Johns LE, Houlston RS. A systematic review and meta-analysis of familial prostate
cancer risk. BJU Int 2003;919:789-94.

Jonsson A, Isomaa B, Tuomi T, et al. A variant in the KCNQ1 gene predicts future type 2
diabetes and mediates impaired insulin secretion. Diabetes 2009;5810:2409-13.

Kao WH, Klag MJ, Meoni LA, et al. MYH9 is associated with nondiabetic end-stage
renal disease in African Americans. Nat Genet 2008;4010:1185-92.

Kasper JS, Giovannucci E. A meta-analysis of diabetes mellitus and the risk of prostate
cancer. Cancer Epidemiol Biomarkers Prev 2006;1511:2056-62.

Kiemeney LA, Thorlacius S, Sulem P, et al. Sequence variant on 8q24 confers
susceptibility to urinary bladder cancer. Nat Genet 2008;4011:1307-12.

King TE, Pawar SC, Majuta L, et al. The role of alpha 6 integrin in prostate cancer
migration and bone pain in a novel xenograft model. PLoS One 2008;310:e3535.

Kirsh VA, Mayne ST, Peters U, et al. A prospective study of lycopene and tomato
product intake and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev
2006;151:92-8.

Kolonel LN. Fat, meat, and prostate cancer. Epidemiol Rev 2001;231:72-81.

181
Kolonel LN. Racial and Geographic variations in prostate cancer and the effect of
migration. In: Fortner JG, Sharp P.A. , ed. Accomplishments in Cancer Research.
Philadelphia (PA): Lippincott-Raven, 1996:221-30.

Kolonel LN, Altshuler D, Henderson BE. The multiethnic cohort study: exploring genes,
lifestyle and cancer risk. Nat Rev Cancer 2004;47:519-27.

Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and Los
Angeles: baseline characteristics. Am J Epidemiol 2000;1514:346-57.

Koontz JI, Soreng AL, Nucci M, et al. Frequent fusion of the JAZF1 and JJAZ1 genes in
endometrial stromal tumors. Proc Natl Acad Sci U S A 2001;9811:6348-53.

Kote-Jarai Z, Easton DF, Stanford JL, et al. Multiple novel prostate cancer predisposition
loci confirmed by an international study: the PRACTICAL Consortium. Cancer
Epidemiol Biomarkers Prev 2008;178:2052-61.

Kote-Jarai Z, Leongamornlert D, Tymrakiewicz M, et al. Mutation analysis of the MSMB
gene in familial prostate cancer. Br J Cancer;1022:414-8.

Kraft P, Pharoah P, Chanock SJ, et al. Genetic variation in the HSD17B1 gene and risk of
prostate cancer. PLoS Genet 2005;15:e68.

Larsson SC, Mantzoros CS, Wolk A. Diabetes mellitus and risk of breast cancer: a meta-
analysis. Int J Cancer 2007;1214:856-62.

Larsson SC, Orsini N, Brismar K, et al. Diabetes mellitus and risk of bladder cancer: a
meta-analysis. Diabetologia 2006;4912:2819-23.

Larsson SC, Orsini N, Wolk A. Diabetes mellitus and risk of colorectal cancer: a meta-
analysis. J Natl Cancer Inst 2005;9722:1679-87.

Lee LG, Connell CR, Bloch W. Allelic discrimination by nick-translation PCR with
fluorogenic probes. Nucleic Acids Res 1993;2116:3761-6.

Lee YH, Kang ES, Kim SH, et al. Association between polymorphisms in SLC30A8,
HHEX, CDKN2A/B, IGF2BP2, FTO, WFS1, CDKAL1, KCNQ1 and type 2 diabetes in
the Korean population. J Hum Genet 2008;5311-12:991-8.

Lehmann JM, Moore LB, Smith-Oliver TA, et al. An antidiabetic thiazolidinedione is a
high affinity ligand for peroxisome proliferator-activated receptor gamma (PPAR
gamma). J Biol Chem 1995;27022:12953-6.

182
Li H, Ma X, Wang J, et al. Effects of rearrangement and allelic exclusion of
JJAZ1/SUZ12 on cell proliferation and survival. Proc Natl Acad Sci U S A
2007;10450:20001-6.

Lichtenstein P, Holm NV, Verkasalo PK, et al. Environmental and heritable factors in the
causation of cancer--analyses of cohorts of twins from Sweden, Denmark, and Finland. N
Engl J Med 2000;3432:78-85.

Lilja H, Abrahamsson PA. Three predominant proteins secreted by the human prostate
gland. Prostate 1988;121:29-38.

Lindgren CM, Heid IM, Randall JC, et al. Genome-wide association scan meta-analysis
identifies three Loci influencing adiposity and fat distribution. PLoS Genet
2009;56:e1000508.

Lippman SM, Klein EA, Goodman PJ, et al. Effect of selenium and vitamin E on risk of
prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial
(SELECT). JAMA 2009;3011:39-51.

Lou H, Yeager M, Li H, et al. Fine mapping and functional analysis of a common variant
in MSMB on chromosome 10q11.2 associated with prostate cancer susceptibility. Proc
Natl Acad Sci U S A 2009;10619:7933-8.

Lu QY, Hung JC, Heber D, et al. Inverse associations between plasma lycopene and other
carotenoids and prostate cancer. Cancer Epidemiol Biomarkers Prev 2001;107:749-56.

Maskarinec G, Erber E, Grandinetti A, et al. Diabetes incidence based on linkages with
health plans: the multiethnic cohort. Diabetes 2009;588:1732-8.

Maskarinec G, Grandinetti A, Matsuura G, et al. Diabetes prevalence and body mass
index differ by ethnicity: the Multiethnic Cohort. Ethn Dis 2009;191:49-55.

Mau MK, Glanz K, Severino R, et al. Mediators of lifestyle behavior change in Native
Hawaiians: initial findings from the Native Hawaiian Diabetes Intervention Program.
Diabetes Care 2001;2410:1770-5.

MedlinePlus [Internet]. Bethesda (MD): National Library of Medicine (US); [updated
2009 Dec 7]. Type II Diabetes - risk factors; [updated 2009 May 20; cited 2009 Dec 7.
Available from: http://www.nlm.nih.gov/medlineplus/ency/article/002072.htm.

Meigs JB, Manning AK, Fox CS, et al. Genome-wide association with diabetes-related
traits in the Framingham Heart Study. BMC Med Genet 2007;8 Suppl 1:S16.

183
Meyer P, Zuern C, Hermanns N, et al. The association between paternal prostate cancer
and type 2 diabetes. J Carcinog 2007;6:14.

Meyer TE, Boerwinkle E, Morrison AC, et al. Diabetes genes and prostate cancer in the
Atherosclerosis Risk in Communities study. Cancer Epidemiol Biomarkers
Prev;192:558-65.

Midthjell K, Holmen J, Bjorndal A, et al. Is questionnaire information valid in the study
of a chronic disease such as diabetes? The Nord-Trondelag diabetes study. J Epidemiol
Community Health 1992;465:537-42.

Mishina T, Watanabe H, Araki H, et al. Epidemiological study of prostatic cancer by
matched-pair analysis. Prostate 1985;64:423-36.

Mitrunen K, Pettersson K, Piironen T, et al. Dual-label one-step immunoassay for
simultaneous measurement of free and total prostate-specific antigen concentrations and
ratios in serum. Clin Chem 1995;418 Pt 1:1115-20.

Monroe KR, Yu MC, Kolonel LN, et al. Evidence of an X-linked or recessive genetic
component to prostate cancer risk. Nat Med 1995;18:827-9.

Nakajima T, Fujino S, Nakanishi G, et al. TIP27: a novel repressor of the nuclear orphan
receptor TAK1/TR4. Nucleic Acids Res 2004;3214:4194-204.

Nam RK, Reeves JR, Toi A, et al. A novel serum marker, total prostate secretory protein
of 94 amino acids, improves prostate cancer detection and helps identify high grade
cancers at diagnosis. J Urol 2006;1754:1291-7.

National Center for Chronic Disease Prevention and Health Promotion. National
Diabetes Fact Sheet. Atlanta GA: Centers for Disease Control and Prevention; 2005.
Available from: http://www.cdc.gov/diabetes/pubs/general.htm#what.

Nazarenko I, Rana S, Baumann A, et al. Cell surface tetraspanin Tspan8 contributes to
molecular pathways of exosome-induced endothelial cell activation. Cancer
Res;704:1668-78.

Ng MC, Park KS, Oh B, et al. Implication of genetic variants near TCF7L2, SLC30A8,
HHEX, CDKAL1, CDKN2A/B, IGF2BP2, and FTO in type 2 diabetes and obesity in
6,719 Asians. Diabetes 2008;578:2226-33.

Norris JM, Langefeld CD, Talbert ME, et al. Genome-wide association study and follow-
up analysis of adiposity traits in Hispanic Americans: the IRAS Family Study. Obesity
(Silver Spring) 2009;1710:1932-41.

184
Norrish AE, Jackson RT, Sharpe SJ, et al. Prostate cancer and dietary carotenoids. Am J
Epidemiol 2000;1512:119-23.

Oades GM, Dredge K, Kirby RS, et al. Vitamin D receptor-dependent antitumour effects
of 1,25-dihydroxyvitamin D3 and two synthetic analogues in three in vivo models of
prostate cancer. BJU Int 2002;906:607-16.

Okura Y, Urban LH, Mahoney DW, et al. Agreement between self-report questionnaires
and medical record data was substantial for diabetes, hypertension, myocardial infarction
and stroke but not for heart failure. J Clin Epidemiol 2004;5710:1096-103.

Ostrander EA, Stanford JL. Genetics of prostate cancer: too many loci, too few genes.
Am J Hum Genet 2000;676:1367-75.

Page WF, Braun MM, Partin AW, et al. Heredity and prostate cancer: a study of World
War II veteran twins. Prostate 1997;334:240-5.

Palmer ND, Langefeld CD, Ziegler JT, et al. Candidate loci for insulin sensitivity and
disposition index from a genome-wide association analysis of Hispanic participants in the
Insulin Resistance Atherosclerosis (IRAS) Family Study. Diabetologia 2009.

Pare G, Chasman DI, Parker AN, et al. Novel association of HK1 with glycated
hemoglobin in a non-diabetic population: a genome-wide evaluation of 14,618
participants in the Women's Genome Health Study. PLoS Genet 2008;412:e1000312.

Park SY, Murphy SP, Wilkens LR, et al. Fat and meat intake and prostate cancer risk: the
multiethnic cohort study. Int J Cancer 2007;1216:1339-45.

Peehl DM, Skowronski RJ, Leung GK, et al. Antiproliferative effects of 1,25-
dihydroxyvitamin D3 on primary cultures of human prostatic cells. Cancer Res
1994;543:805-10.

Pierce BL, Ahsan H. Genetic susceptibility to type 2 diabetes is associated with reduced
prostate cancer risk. Hum Hered;693:193-201.

Pierce BL, Plymate S, Ostrander EA, et al. Diabetes mellitus and prostate cancer risk.
Prostate 2008;6810:1126-32.
Platz EA, Giovannucci E. The epidemiology of sex steroid hormones and their signaling
and metabolic pathways in the etiology of prostate cancer. J Steroid Biochem Mol Biol
2004;924:237-53.

Pomerantz MM, Ahmadiyeh N, Jia L, et al. The 8q24 cancer risk variant rs6983267
shows long-range interaction with MYC in colorectal cancer. Nat Genet 2009;418:882-4.

185
Pomerantz MM, Beckwith CA, Regan MM, et al. Evaluation of the 8q24 prostate cancer
risk locus and MYC expression. Cancer Res 2009;6913:5568-74.

Poulsen P, Kyvik KO, Vaag A, et al. Heritability of type II (non-insulin-dependent)
diabetes mellitus and abnormal glucose tolerance--a population-based twin study.
Diabetologia 1999;422:139-45.

Prokopenko I, Langenberg C, Florez JC, et al. Variants in MTNR1B influence fasting
glucose levels. Nat Genet 2009;411:77-81.

Prokunina-Olsson L, Fu YP, Tang W, et al. Refining the Prostate Cancer Genetic
Association within the JAZF1 Gene on Chromosome 7p15.2. Cancer Epidemiol
Biomarkers Prev.

Rafnar T, Sulem P, Stacey SN, et al. Sequence variants at the TERT-CLPTM1L locus
associate with many cancer types. Nat Genet 2009;412:221-7.

Rebbeck TR, Walker AH, Zeigler-Johnson C, et al. Association of HPC2/ELAC2
genotypes and prostate cancer. Am J Hum Genet 2000;674:1014-9.

Reeves JR, Dulude H, Panchal C, et al. Prognostic value of prostate secretory protein of
94 amino acids and its binding protein after radical prostatectomy. Clin Cancer Res
2006;1220 Pt 1:6018-22.

Reich DE, Lander ES. On the allelic spectrum of human disease. Trends Genet
2001;179:502-10.

Reynisdottir I, Thorleifsson G, Benediktsson R, et al. Localization of a susceptibility
gene for type 2 diabetes to chromosome 5q34-q35.2. Am J Hum Genet 2003;732:323-35.

Rich SS, Goodarzi MO, Palmer ND, et al. A genome-wide association scan for acute
insulin response to glucose in Hispanic-Americans: the Insulin Resistance
Atherosclerosis Family Study (IRAS FS). Diabetologia 2009;527:1326-33.

Rippe V, Drieschner N, Meiboom M, et al. Identification of a gene rearranged by 2p21
aberrations in thyroid adenomas. Oncogene 2003;2238:6111-4.

Robbins C, Torres JB, Hooker S, et al. Confirmation study of prostate cancer risk variants
at 8q24 in African Americans identifies a novel risk locus. Genome Res 2007;1712:1717-
22.

Roddam AW, Allen NE, Appleby P, et al. Endogenous sex hormones and prostate
cancer: a collaborative analysis of 18 prospective studies. J Natl Cancer Inst
2008;1003:170-83.
186

Rodriguez C, Patel AV, Mondul AM, et al. Diabetes and risk of prostate cancer in a
prospective cohort of US men. Am J Epidemiol 2005;1612:147-52.

Rokman A, Ikonen T, Mononen N, et al. ELAC2/HPC2 involvement in hereditary and
sporadic prostate cancer. Cancer Res 2001;6116:6038-41.

Rokman A, Ikonen T, Seppala EH, et al. Germline alterations of the RNASEL gene, a
candidate HPC1 gene at 1q25, in patients and families with prostate cancer. Am J Hum
Genet 2002;705:1299-304.

Rosenberg DJ, Neugut AI, Ahsan H, et al. Diabetes mellitus and the risk of prostate
cancer. Cancer Invest 2002;202:157-65.

Ross LE, Berkowitz Z, Ekwueme DU. Use of the prostate-specific antigen test among
U.S. men: findings from the 2005 National Health Interview Survey. Cancer Epidemiol
Biomarkers Prev 2008;173:636-44.

Rung J, Cauchi S, Albrechtsen A, et al. Genetic variant near IRS1 is associated with type
2 diabetes, insulin resistance and hyperinsulinemia. Nat Genet 2009;4110:1110-5.

Sachidanandam R, Weissman D, Schmidt SC, et al. A map of human genome sequence
variation containing 1.42 million single nucleotide polymorphisms. Nature
2001;4096822:928-33.

Sandhu MS, Weedon MN, Fawcett KA, et al. Common variants in WFS1 confer risk of
type 2 diabetes. Nat Genet 2007;398:951-3.

Saxena R, Voight BF, Lyssenko V, et al. Genome-wide association analysis identifies
loci for type 2 diabetes and triglyceride levels. Science 2007;3165829:1331-6.

Schaid DJ. The complex genetic epidemiology of prostate cancer. Hum Mol Genet
2004;13 Spec No 1:R103-21.

Schaid DJ, McDonnell SK, Blute ML, et al. Evidence for autosomal dominant inheritance
of prostate cancer. Am J Hum Genet 1998;626:1425-38.

Schwartz GG, Hulka BS. Is vitamin D deficiency a risk factor for prostate cancer?
(Hypothesis). Anticancer Res 1990;105A:1307-11.

Schork NJ, Murray SS, Frazer KA, et al. Common vs. rare allele hypotheses for complex
diseases. Curr Opin Genet Dev 2009;193:212-9.

187
Scott LJ, Mohlke KL, Bonnycastle LL, et al. A genome-wide association study of type 2
diabetes in Finns detects multiple susceptibility variants. Science 2007;3165829:1341-5.

Setiawan VW, Schumacher FR, Haiman CA, et al. CYP17 genetic variation and risk of
breast and prostate cancer from the National Cancer Institute Breast and Prostate Cancer
Cohort Consortium (BPC3). Cancer Epidemiol Biomarkers Prev 2007;1611:2237-46.

Severi G, Giles GG, Southey MC, et al. ELAC2/HPC2 polymorphisms, prostate-specific
antigen levels, and prostate cancer. J Natl Cancer Inst 2003;9511:818-24.

Sharifi N, Gulley JL, Dahut WL. Androgen deprivation therapy for prostate cancer.
JAMA 2005;2942:238-44.

Shimizu H, Ross RK, Bernstein L, et al. Cancers of the prostate and breast among
Japanese and white immigrants in Los Angeles County. Br J Cancer 1991;636:963-6.

Simard J, Dumont M, Labuda D, et al. Prostate cancer susceptibility genes: lessons
learned and challenges posed. Endocr Relat Cancer 2003;102:225-59.

Simonis-Bik AM, Nijpels G, van Haeften TW, et al. Gene variants in the novel type 2
diabetes loci CDC123/CAMK1D, THADA, ADAMTS9, BCL11A, and MTNR1B affect
different aspects of pancreatic beta-cell function. Diabetes;591:293-301.

Skowronski RJ, Peehl DM, Feldman D. Vitamin D and prostate cancer: 1,25
dihydroxyvitamin D3 receptors and actions in human prostate cancer cell lines.
Endocrinology 1993;1325:1952-60.

Sladek R, Rocheleau G, Rung J, et al. A genome-wide association study identifies novel
risk loci for type 2 diabetes. Nature 2007;4457130:881-5.

Smith JR, Freije D, Carpten JD, et al. Major susceptibility locus for prostate cancer on
chromosome 1 suggested by a genome-wide search. Science 1996;2745291:1371-4.

Smith RA, Cokkinides V, Eyre HJ. American Cancer Society guidelines for the early
detection of cancer, 2003. CA Cancer J Clin 2003;531:27-43.

Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al. A variant in CDKAL1 influences
insulin response and risk of type 2 diabetes. Nat Genet 2007;396:770-5.

Stram DO, Hankin JH, Wilkens LR, et al. Prostate cancer incidence and intake of fruits,
vegetables and related micronutrients: the multiethnic cohort study* (United States).
Cancer Causes Control 2006;179:1193-207.

188
Suarez BK, Lin J, Burmester JK, et al. A genome screen of multiplex sibships with
prostate cancer. Am J Hum Genet 2000;663:933-44.

Sun J, Zheng SL, Wiklund F, et al. Evidence for two independent prostate cancer risk-
associated loci in the HNF1B gene at 17q12. Nat Genet 2008;4010:1153-5.

Tabara Y, Osawa H, Kawamoto R, et al. Replication study of candidate genes associated
with type 2 diabetes based on genome-wide screening. Diabetes 2009;582:493-8.

Takeuchi F, Serizawa M, Yamamoto K, et al. Confirmation of multiple risk Loci and
genetic impacts by a genome-wide association study of type 2 diabetes in the Japanese
population. Diabetes 2009;587:1690-9.

Tavani A, Gallus S, Bosetti C, et al. Diabetes and the risk of prostate cancer. Eur J
Cancer Prev 2002;112:125-8.

Tavtigian SV, Simard J, Teng DH, et al. A candidate prostate cancer susceptibility gene
at chromosome 17p. Nat Genet 2001;272:172-80.

The International HapMap Project. Nature 2003;4266968:789-96.
Thompson IM, Goodman PJ, Tangen CM, et al. The influence of finasteride on the
development of prostate cancer. N Engl J Med 2003;3493:215-24.

Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified in a genome-wide
association study of prostate cancer. Nat Genet 2008;403:310-5.

Timpson NJ, Lindgren CM, Weedon MN, et al. Adiposity-related heterogeneity in
patterns of type 2 diabetes susceptibility observed in genome-wide association data.
Diabetes 2009;582:505-10.

Tontonoz P, Hu E, Spiegelman BM. Stimulation of adipogenesis in fibroblasts by PPAR
gamma 2, a lipid-activated transcription factor. Cell 1994;797:1147-56.

Travis RC, Schumacher F, Hirschhorn JN, et al. CYP19A1 genetic variation in relation to
prostate cancer risk and circulating sex hormone concentrations in men from the Breast
and Prostate Cancer Cohort Consortium. Cancer Epidemiol Biomarkers Prev
2009;1810:2734-44.

Turnbull C, Ahmed S, Morrison J, et al. Genome-wide association study identifies five
new breast cancer susceptibility loci. Nat Genet;426:504-7.

Tuupanen S, Turunen M, Lehtonen R, et al. The common colorectal cancer predisposition
SNP rs6983267 at chromosome 8q24 confers potential to enhanced Wnt signaling. Nat
Genet 2009;418:885-90.
189

Unoki H, Takahashi A, Kawaguchi T, et al. SNPs in KCNQ1 are associated with
susceptibility to type 2 diabetes in East Asian and European populations. Nat Genet
2008;409:1098-102.

Valeri A, Briollais L, Azzouzi R, et al. Segregation analysis of prostate cancer in France:
evidence for autosomal dominant inheritance and residual brother-brother dependence.
Ann Hum Genet 2003;67Pt 2:125-37.

Valtonen-Andre C, Savblom C, Fernlund P, et al. Beta-microseminoprotein in serum
correlates with the levels in seminal plasma of young, healthy males. J Androl
2008;293:330-7.

Varghese JS, Easton DF. Genome-wide association studies in common cancers-what
have we learnt? Curr Opin Genet Dev.

Velicer CM, Dublin S, White E. Diabetes and the risk of prostate cancer: the role of
diabetes treatment and complications. Prostate Cancer Prostatic Dis 2007;101:46-51.

Verhage BA, Baffoe-Bonnie AB, Baglietto L, et al. Autosomal dominant inheritance of
prostate cancer: a confirmatory study. Urology 2001;571:97-101.

Wang L, McDonnell SK, Cunningham JM, et al. No association of germline alteration of
MSR1 with prostate cancer risk. Nat Genet 2003;352:128-9.

Walitt BT, Constantinescu F, Katz JD, et al. Validation of self-report of rheumatoid
arthritis and systemic lupus erythematosus: The Women's Health Initiative. J Rheumatol
2008;355:811-8.

Waters KM, Henderson BE, Stram DO, et al. Association of diabetes with prostate cancer
risk in the multiethnic cohort. Am J Epidemiol 2009;1698:937-45.

Waters KM, Le Marchand L, Kolonel LN, et al. Generalizability of associations from
prostate cancer genome-wide association studies in multiple populations. Cancer
Epidemiol Biomarkers Prev 2009;184:1285-9.

Werny DM, Saraiya M, Gregg EW. Prostate-specific antigen values in diabetic and
nondiabetic US men, 2001-2002. Am J Epidemiol 2006;16410:978-83.

Winckler W, Weedon MN, Graham RR, et al. Evaluation of common variants in the six
known maturity-onset diabetes of the young (MODY) genes for association with type 2
diabetes. Diabetes 2007;563:685-93.

190
Witte JS, Goddard KA, Conti DV, et al. Genomewide scan for prostate cancer-
aggressiveness loci. Am J Hum Genet 2000;671:92-9.

Wooster R, Bignell G, Lancaster J, et al. Identification of the breast cancer susceptibility
gene BRCA2. Nature 1995;3786559:789-92.

Wu M, Li Y, Ji C, et al. Cloning and identification of a novel human gene PDLIM5, a
homolog of AD-associated neuronal thread protein (AD7c-NTP). DNA Seq
2004;152:144-7.

Wu Y, Li H, Loos RJ, et al. Common variants in CDKAL1, CDKN2A/B, IGF2BP2,
SLC30A8, and HHEX/IDE genes are associated with type 2 diabetes and impaired
fasting glucose in a Chinese Han population. Diabetes 2008;5710:2834-42.

Xie D, Gore C, Liu J, et al. Role of DAB2IP in modulating epithelial-to-mesenchymal
transition and prostate cancer metastasis. Proc Natl Acad Sci U S A;1076:2485-90.

Xu B, Wang J, Tong N, et al. A functional polymorphism in MSMB gene promoter is
associated with prostate cancer risk and serum MSMB expression. Prostate
2010;7010:146-52.

Xu J, Dimitrov L, Chang BL, et al. A combined genomewide linkage scan of 1,233
families for prostate cancer-susceptibility genes conducted by the international
consortium for prostate cancer genetics. Am J Hum Genet 2005;772:219-29.

Xu J, Meyers D, Freije D, et al. Evidence for a prostate cancer susceptibility locus on the
X chromosome. Nat Genet 1998;202:175-9.

Xu J, Zheng SL, Komiya A, et al. Common sequence variants of the macrophage
scavenger receptor 1 gene are associated with prostate cancer risk. Am J Hum Genet
2003;721:208-12.

Xu J, Zheng SL, Komiya A, et al. Germline mutations and sequence variants of the
macrophage scavenger receptor 1 gene are associated with prostate cancer risk. Nat
Genet 2002;322:321-5.

Yang J, Benyamin B, McEvoy BP, et al. Common SNPs explain a large proportion of the
heritability for human height. Nat Genet 2010;427:565-9.

Yasuda K, Miyake K, Horikawa Y, et al. Variants in KCNQ1 are associated with
susceptibility to type 2 diabetes mellitus. Nat Genet 2008;409:1092-7.
Intact free prostate-specific antigen and free and total human glandular kallikrein 2.
Elimination of assay interference by enzymatic digestion of antibodies to F(ab')2
fragments. Anal Chem 2006;7822:7809-15.
191

Yeager M, Chatterjee N, Ciampa J, et al. Identification of a new prostate cancer
susceptibility locus on chromosome 8q24. Nat Genet 2009;4110:1055-7.

Yin L, Raum E, Haug U, et al. Meta-analysis of longitudinal studies: Serum vitamin D
and prostate cancer risk. Cancer Epidemiol 2009.

Zanke BW, Greenwood CM, Rangrej J, et al. Genome-wide association scan identifies a
colorectal cancer susceptibility locus on chromosome 8q24. Nat Genet 2007;398:989-94.

Zeggini E, Scott LJ, Saxena R, et al. Meta-analysis of genome-wide association data and
large-scale replication identifies additional susceptibility loci for type 2 diabetes. Nat
Genet 2008;405:638-45.

Zeggini E, Weedon MN, Lindgren CM, et al. Replication of genome-wide association
signals in UK samples reveals risk loci for type 2 diabetes. Science 2007;3165829:1336-
41.

Zhao LJ, Subramanian T, Chinnadurai G. Inhibition of transcriptional activation and cell
proliferation activities of adenovirus E1A by the unique N-terminal domain of CtBP2.
Oncogene 2008;2739:5214-22.

Zheng SL, Stevens VL, Wiklund F, et al. Two independent prostate cancer risk-
associated Loci at 11q13. Cancer Epidemiol Biomarkers Prev 2009;186:1815-20.

Abstract (if available)

Abstract This thesis is comprised of five studies that examine the relationship between common genetic variation, type 2 diabetes and prostate cancer risk.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Identification and fine-mapping of genetic susceptibility loci for prostate cancer and statistical methodology for multiethnic fine-mapping

PDF

Prostate cancer: genetic susceptibility and lifestyle risk factors

PDF

Screening and association testing of coding variation in steroid hormone coactivator and corepressor genes in relationship with breast cancer risk in multiple populations

PDF

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

PDF

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

PDF

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

PDF

Genetic risk factors in breast cancer susceptibility: The multiethnic cohort

PDF

Diet quality and pancreatic cancer incidence in the multiethnic cohort

PDF

Genetic correlates of the timing of natural menopause: the breast and prostate cancer cohort consortium

PDF

The development of type 2 diabetes mellitus (type 2 DM) in California twins

PDF

Genetic studies of cancer in populations of African ancestry and Latinos

PDF

Co-chaperone influence on androgen receptor signaling and identification of androgen receptor genes in prostate cancer

PDF

Ancestral/Ethnic variation in the epidemiology and genetic predisposition of early-onset hematologic cancers

PDF

Dietary and supplementary folate intake and prostate cancer risk

PDF

Polygenes and estimated heritability of prostate cancer in an African American sample using genome-wide association study data

PDF

Study of genetic variation in Asian Indians -- applications to complex diseases and endogamic exogamy

PDF

Missing heritability may be explained by the common household environment and its interaction with genetic variation

PDF

Understanding prostate cancer genetic susceptibility and chromatin regulation

PDF

Using genetic ancestry to improve between-population transferability of a prostate cancer polygenic risk score

PDF

Extending genome-wide association study methods in African American data

Asset Metadata

Creator Waters, Kevin M. (author)

Core Title Examining the relationship between common genetic variation, type 2 diabetes and prostate cancer risk in the multiethnic cohort

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Degree Conferral Date 2010-08

Publication Date 08/05/2010

Defense Date 06/30/2010

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

Tag common genetic variation,genetic susceptibility,MSMB,multiethnic cohort,OAI-PMH Harvest,prostate cancer,type 2 diabetes

Place Name California (states), Hawaii (states), Los Angeles (city or populated place)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Haiman, Christopher A. (committee chair), Coetzee, Gerhard A. (committee member), Henderson, Brian E. (committee member), Stram, Daniel O. (committee member), Watanabe, Richard M. (committee member)

Creator Email kmwaters@gmail.com,kwaters@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3312

Unique identifier UC1445834

Identifier etd-Waters-3931 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-373007 (legacy record id),usctheses-m3312 (legacy record id)

Legacy Identifier etd-Waters-3931.pdf

Dmrecord 373007

Document Type Dissertation

Rights Waters, Kevin M.

Type texts

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

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu