University of Southern California Dissertations and Theses

Androgens and breast cancer

(USC Thesis Other)

Androgens and breast cancer

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ANDROGENS AND BREAST CANCER
by
Elizabeth Osth Lillie
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(EPIDEMIOLOGY)
May 2004
Copyright 2004 Elizabeth Osth Lillie
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
UMI Number: 3140508
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
UMI
UMI Microform 3140508
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
A C K N O W L E D G E M E N T S
I would like to thank my dissertation chairs Drs. Giske Ursin and Theodore Krontiris for
offering their time and energy towards my future success. I could not have completed
this dissertation without their faith in my abilities. While I take pride in cutting my own
path into the forest of breast cancer literature to generate a thesis worthy of the doctoral
degree, it could not have been done without their guidance.
I would also like to acknowledge my dissertation committee members Drs. Leslie
Bernstein, Jim Gauderman and Gerry Coetzee for their additional insight and support.
In addition to my dissertation committee, various individuals have offered their expertise
and moral support that I would like to recognize. A special thanks to Dr. Garry Larson
and Mrs. Cathryn Lundberg at the City of Hope along with Dr. Alpa Patel of the
American Cancer Society.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
T A B L E O F C O N T E N T S
Acknowledgements
List of Tables
Abbreviations
Abstract
Chapter 1
Chapter 2
Chapter 3
Chapter 4
Bibliography
REVIEW: The Role o f Androgens and Polymorphisms in
the Androgen Receptor in the Epidemiology o f Breast
Cancer
GRANT PROPOSAL: The Androgen Receptor and
Mammographic Density
DATA ANALYSIS: Polymorphism in the Androgen
Receptor and Mammographic Density in Women Taking
and Not Taking Estrogen and Progestin Therapy
DATA ANALYSIS: Androgen Receptor GAG
Polymorphism and Breast Cancer Prognosis
11
IV
V ll
23
38
54
71
m
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
LIST OF TABLES
Table 1-1 The relative risk of breast cancer associated with testosterone levels
in postmenopausal women, results from prospective studies
Table 1-2 The relative risk of breast cancer associated with testosterone levels,
results from case-control studies
11
Table 1-3 Mean levels of testosterone in breast cancer cases and controls 12
Table 1-4 The relative risk of breast cancer associated with the androgen 19
receptor CAG repeat
Table 2-1 Distribution of breast cancer cases and controls according to 32
mammographic density
Table 2-2 Analysis of androgen receptor and breast cancer tumor grade in 32
breast cancer cases with a family history of breast cancer
Table 3-1 Mean percentage of mammographic density by descriptive 44
characteristics (N = 404)
Table 3-2 Least-squares mean percentage breast density by androgen receptor 46
genotype
Table 3-3 Association for a single quartile increase in percent mammographic 49
density per 'long' androgen receptor allele
Table 4-1 Distribution of cases in each prognostic indicator category stratified 61
by source of data
Table 4-2 Mean number of androgen receptor CAG repeats by breast cancer 62
prognostic indicator
Table 4-3 Association of androgen receptor CAG with breast cancer 63
prognostic indicators
Table 4-4 Mean number of androgen receptor CAG repeats by breast cancer 65
prognostic indicator in cases age > 55 at diagnosis
Table 4-5 Association of androgen receptor CAG with breast cancer 66
prognostic indicators in cases age > 55 at diagnosis
iv
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
ABBREVIATIONS
AR androgen receptor
BMI body mass index
CAG cytosine-adenine-guanine
CARE Contraceptive and Reproductive Experiences
Cl confidence interval
DHT dihydrotestosterone
E2 estradiol
EPT estrogen progestin therapy
ER estrogen receptor
ET estrogen therapy
GGC guanine-guanine-cytosine
H hours
H I hormone therapy
L long allele
ML medium long allele
MS medium short allele
NS not significant
OR odds ratio
P p-value
RR relative risk
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
s short allele
SD standard deviation
T testosterone
VL very long allele
VS very short allele
Y years
Y/N yes/no
VI
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
ABSTRACT
Testosterone binds to the androgen receptor in target tissue to mediate its effects. Both
testosterone and the androgen receptor have been exposures of interest in the etiology of
breast cancer. A triplet repeat polymorphism (CAG) in exon 1 of the gene for the
androgen receptor that encodes a polyglutamine tract has been observed to correlate with
androgen receptor activity. Paradoxical to observations that endogenous testosterone
increases breast cancer risk, results from observational studies that have examined
polymorphisms in the androgen receptor suggest that the low activity androgen receptor
increases breast cancer risk. I reviewed the quality of this evidence and based on my
findings have formulated various hypotheses to further examine the role of androgens and
the androgen receptor in breast cancer etiology.
Mammographic density is a strong breast cancer risk factor that probably reflects hormone
induced cell proliferation in the breast. I evaluated the association between the androgen
receptor CAG repeat length and mammographic density. My results suggest that the long
CAG repeat length (less active androgen receptor) is associated with increased
mammographic density in women on estrogen progestin therapy.
The short CAG repeat length (more active androgen receptor) has been observed to be
associated with increased breast cancer grade in one published study. I evaluated the
association between polymorphism in the androgen receptor and breast cancer prognosis
vii
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
using a larger study setting. I observed a significant association when the analysis was
restricted to women over age 55.
In conclusion, my review of the literature and results from my data analyses suggest that
androgens and the androgen receptor play a role in breast cancer etiology. However,
further investigation is clearly needed.
vm
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CHAPTER 1
REVIEW; The Role o f Androgens and Polymorphisms in the Androgen Receptor in the
Epidemiology o f Breast Cancer (Lillie and others 2003)
Introduction
The role of androgens in breast cancer etiology has been the subjects of both curiosity
and confusion. It is still unclear by which mechanisms testosterone (T) exerts its activity
in the female breast, and whether the effects are predominantly proliferative or anti
proliferative on breast cells at physiologic levels. In this review I evaluate the results
from epidemiologic studies on the role of circulating T and a functional polymorphism in
the androgen receptor (AR) in breast cancer. I also highlight some of the epidemiologic
challenges in addressing these questions.
Sources of Endogenous Testosterone
There are two main sources of androgens m women. T is produced directly by the ovary,
and also by conversion of the adrenal androgens dehydroepiandrosterone and
dehydroepiandrosterone-sulfate into androstenedione, and then further to T in peripheral
tissue (Norman and Litwig 1987). In premenopausal women, approximately 25% of
circulating T is secreted directly from the adrenal gland and 25% from the ovary, while
the remaining 50% is produced by peripheral conversion of androstenedione (Longcope
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
and others 1986). T levels vary over the menstrual cycle with peak levels mid-cycle, and
diumally with highest levels in the early morning (Abraham 1974).
T and androstenedione are produced by the interstitial cells of the ovarian stroma and
may continue to respond to gonadotopins and produce T after the menopause (Adashi
1994). In normal postmenopausal women the ovarian vein has been observed to have
higher concentrations of T than is found in peripheral blood (Judd and others 1974);
bilateral oophorectomy results in reductions in T levels by as much as 50% (Judd and
others 1974).
Several smaller cross-sectional studies have foimd lower T levels in postmenopausal than
premenopausal women (Bancroft and Cawood 1996; Labrie and others 1997; Zumoff and
others 1995) or lower levels in perimenopausal than premenopausal women (Longcope
and others 1986). Large longitudinal studies that have followed women through the
menopausal transition have observed either no significant change in T (Burger and others
2000; Longcope and others 1986) or a 15% decrease in both T and androstenedione at
menopause (Rannevik and others 1995). In one study of women aged 50-89 years, T
levels were lowest at the time of the menopause while women above age 70 or more than
20 years post menopause had levels approximating those of premenopausal women
(Laughlin and others 2000).
In summary, there is increasing evidence that the ovary continues to produce
androstenedione and T in healthy postmenopausal women. Levels may either remain the
2
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
same or decrease slightly at menopause. However, women with bilateral oophorectomy
may be androgen deficient.
Testosterone and Breast Cancer Risk
Prospective Studies that Examine the Association Between Testosterone and Breast
Cancer
Eight prospective cohort studies have published data on the association between
endogenous T levels and breast cancer risk using T measured from blood samples
gathered at baseline from postmenopausal women (Berrino and others 1996; Cauley and
others 1999; Dorgan and others 1996; Garland and others 1992; Hankinson and others
1998; Thomas and others 1997; Wysowski and others 1987; Zeleniuch-Jacquotte and
others 1997). Six of these studies were nested case-control studies (Berrino and others
1996; Dorgan and others 1996; Hankinson and others 1998; Thomas and others 1997;
Wysowski and others 1987; Zeleniuch-Jacquotte and others 1997); one was a case-cohort
study (Cauley and others 1999) and one a full cohort study (Garland and others 1992).
Only one of these studies published results for premenopausal women (Wysowski and
others 1987).
Six of these eight studies reported a statistically significant increase in postmenopausal
breast cancer risk with increasing levels of endogenous T (Berrino and others 1996;
Cauley and others 1999; Dorgan and others 1996; Hankinson and others 1998; Thomas
and others 1997; Zeleniuch-Jacquotte and others 1997). A recently conducted pooled
analysis (2002) of these eight prospective studies estimated that the relative risk (RR) of
3
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
breast cancer in women whose levels of T were in the top quintile compared to women in
the bottom quintile was 2.22 (95% confidence interval (Cl), 1.59-3.10). A statistically
significant dose-response relationship was also observed (P trend < 0.001)(2002). Two
of these studies also reported statistically significantly increasing breast cancer risk with
increasing levels of free T (Berrino and others 1996; Cauley and others 1999), a measure
of bioavailable T.
The study of premenopausal women (Wysowski and others 1987) found no statistically
significant differences between cases and non-cases in mean levels in either pre or
postmenopausal women, but the sample size was small (premenopausal women: 17 cases,
67 controls; postmenopausal women: 22 cases, 88 controls).
Can the Observed Association between Testosterone levels and Breast Cancer Risk be
Due to Bias?
Effects o f Measurement Biases:
One limitation to the studies reviewed is that serum T may not be the ideal measure of T.
Total T includes both free T and bound T. Further, serum levels do not take into account
the peripheral conversion of precursor androgens into T in the breast tissue itself. The
effect of this measurement error is most likely to be non-differential, therefore, biasing
results towards the null.
Measurement biases due to use of a single hormone measurement or degradation of
hormones in stored specimens over time would most likely be non-differential, resulting
4
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
in attenuated estimates of disease risk. All of the existing prospective studies (Table 1-1)
analyzed T measured from only one blood draw which might not be representative of the
cumulative exposure to T; however, given the prospective design of these studies, any
inaccuracy in measurement would likely bias risk estimates towards the null.
The consideration of time of day that blood was drawn and fasting status can help to
avoid the biases due to using a single honnone measurement. To avoid this bias, three of
the studies either matched on time of blood draw (Dorgan and others 1996; Hankinson
and others 1998) or restricted subjects to having their blood drawn in the morning
(Garland and others 1992). Since the effect of these sporadic variations would be to bias
the results towards the null, this may help to explain one of the null associations observed
(Wysowski and others 1987).
The studies that observed an association between T and breast cancer risk attempted to
reduce measurement bias due to degradation by matching cases to controls on the date of
blood draw (Berrino and others 1996; Dorgan and others 1996; Hankinson and others
1998; Thomas and others 1997; Zeleniuch-Jacquotte and others 1997) and storage
conditions such as sample location/shelf in the freezer (Berrino and others 1996). This
was not done in the studies reporting no association between T and breast cancer risk
(Garland and others 1992; Wysowski and others 1987) or in one of the positive studies
(Cauley and others 1999).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Laboratory assay variation would also most likely be non-differential since cases and
controls were analyzed concurrently in these studies. The intra and inter assay
coefficients of variation in these studies were rather good, ranging from 4 to 14 percent.
However, the coefficient of variation was not reported in one of the null studies
(Wysowski and others 1987).
Thus, although there may be some attenuation in effect estimates in all of these studies, it
is not clear whether measurement bias due to degradation can explain the discrepancies
between the two null studies and the positive studies.
Temporal Bias:
If breast cancer development inereases T levels, then studies that included subjects
diagnosed shortly after baseline hormone measurement may have artificially elevated
estimates of the risk of breast cancer associated with T levels. Two of the positive
studies and both of the null studies excluded subjects diagnosed 6-24 months after
baseline (Dorgan and others 1996; Garland and others 1992; Wysowski and others 1987;
Zeleniuch-Jacquotte and others 1997). However, the study with the most conservative
cut-point of 24 months (Dorgan and others 1996) reported a significant positive
association between T levels and breast cancer risk. Thus, although it is possible that
temporal bias played a role in the four positive studies with no exclusions, this latter
study suggests that temporal bias cannot explain the association between T and breast
cancer risk.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Effects o f Confounding:
Lack of control for body mass index (BMI) or age at menopause could result in a positive
bias away from the null. All of the studies that reported a significant association between
T levels and breast cancer risk included either BMI or height and weight as covariates in
the statistical model. All but one of the null studies (Garland and others 1992) considered
either the amount of time menopausal (Dorgan and others 1996; Thomas and others
1997; Wysowski and others 1987) or age at menopause (Berrino and others 1996; Cauley
and others 1999; Hankinson and others 1998; Zeleniuch-Jacquotte and others 1997) as a
covariate to control for the effects of menopause on T levels. Thus it is unlikely that
confounding by these variables can explain the associations observed between T levels
and breast cancer risk.
Comparabilitv/Generalizabilitv:
All of the prospective studies that examined the T-breast cancer association were
conducted using cohorts from Caucasian populations. Only two studies (Berrino and
others 1996; Thomas and others 1997) were conducted outside the United States, one in
Italy (Berrino and others 1996) and one on the island of Guernsey (Thomas and others
1997). It is therefore unlikely that differences in the populations studied can explain the
discrepant results between studies. There are, as far as we know, no prospective data
from non-white populations.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Studies o f Testosterone Measured Post-Diagnosis that Examine the Testosterone-Breast
Cancer Association:
Several case-control studies published the past 20 years have evaluated the association
between T levels and breast cancer risk. Comparisons to levels in control subjects have
shown that both postmenopausal breast cancer cases (Adlercreutz and others 1989; Hill
and others 1985; Lipworth and others 1996; Secreto and others 1991; Secreto and others
1983b) and premenopausal cases (Secreto and others 1983a; Secreto and others 1984;
Secreto and others 1989) have significantly elevated T levels. Table 1-2 presents the odds
ratios (OR) of breast cancer associated with categories of serum T, while Table 1-3
presents results of studies that compared mean levels of T. The measures of T in cases
were almost twice those of controls. Although these retrospective studies support the
prospective study results showing an association between increased T and increased
breast cancer risk, these results may not be as readily interpretable as those from the
prospective studies, given the possibility that the presence of cancer, or the treatment for
it, may have increased T levels.
Summary
Both retrospective and prospective studies have published statistically significant
associations between increased levels of T and increased breast cancer risk. These
associations are unlikely to be due to measurement biases, the influence of disease, or
lack of adjustment for the confounding effects of BMI or age at menopause.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 1-1 The relative risk of breast cancer associated with testosterone levels in postmenopausal women,
results from prospective studies
Author, year Cohort Age” Sample size Matching Adjusted for
Wysowski, 1987 Washington County,
MDU.S.
36-90 39 cases; 155
controls
Race, age, time
since last menstrual
period.
Matching variables only
Garland, 1992
Rancho Bernardo,
CAU.S.
50-79 15 cases; 409
at risk
Age, BMI (tertiles), smoking at
baseline other hormones®
Dorgan, 1996
Columbia, MO U.S. 52-73 71 cases; 133
controls
Exact age, date {+/-
1 y) and time of
blood draw (+/-2 h)
y since menopause, height, weight,
parity, family history; matched
analysis
Berrino, 1996
Study of Hormones
and Diet in the
Etiology of Breast
Tumors, Italy
40-69 24 cases; 87
controls
Recmitment center,
recruitment date,
dayli^t saving
period at time of
recruitment, location
of freezer storage
Age at menarche, age at first
childbirth, number of births, age at
menopause, weight, height, BMI
waist-to-hip ratio, other hormones;
matched anal)^is
Thomas, 1997
Guernsey, U.K. mean= 59 61 cases; 179
controls
Age {+/- 2 y), date
of blood collection
(+/-1 y), number of
years menopausal (1-
2 y, or 3+)
Age at menarche, parity, number of
y post-menopausal, BMI, E2 and
SHBG; matched analysis
Zeleniuch-Jocquotte, NYXJ Women's 49-65 85 cases; 163 Age at enrollment BMI, age at menarche, parity, age
1997 Health Study, U.S. controls (+/-6 mos), date of
initial blood
donation (+/-3 mos),
menopausal status
at first full-term pregancy, age at
menopause, family history, history
of benign breast condition, history
of oophorectomy, lifetime mos of
lactation, smoking; matched
analysis
HanJdnson, 1998
Nurses Health
Study, U.S.
46-69 147 cases; 299
controls
Age (+/- 2 y), month
of collection, time
of day that blood
was drawn (+/- 2 h),
fasting status
BMI (at age 18, quaitiles), family
history, age at menarche (quartiles),
parity/age at 1st birth, age at
menopause (quartiles); matched
analysis
Cauley, 1999 Study of
Osteoporotic
Fractures, U.S.
65-75 97 cases; 250
controls
Age, BMI, age at menarche, first
birth, and menr^ause, surgical
menopause (Y/N), nulliparity
(Y/N), family history, past estrogen
use (Y/N)^, walking for exercise
(Y/N) and alcohol consumption
(g/d quintile)
Baseline age distribution
All units converted to pg/mL
Age adjusted model
Y/N, yes/no; NS, not significant; y, years; h,
hours
' Analysis also conducted excluding past
estrogen users
^ Adjusted for BMI and age only
* Adjustment did not change results from the
crude analysis
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 1-1 continued.
Author, year Exposure Cases Controls Categories’ ’ RR' 95% Cl P Adjusted RR 95% Cl P
Wysowski, 1987 T 39 155 N/A N/A N/A NS N/A N/A NS
Garland, 1992 T 5 132 37-176 1.0
5 137 177-284 1.1 NS
*
5 140 285-778 1.0 NS NS
Dorgan, 1996 T 9 32 <98 1.0 1.0
13 28 98-169 1.8 0.6-5.0 2.9 0.9-9.4
20 39 170-259 2.1 0.8-5.6 2.9 1.0-8.6
29 34 >259 3.7 1.4-10.0 6.2 2.0-19.0 0.02
Berrino, 1996 Iree T 24 87 <0.57 1.0
0.57-0.86 1.8 0.4-9.3
>0.86 5.7 1.5-22.2 0.005 4.6 1.1-20.0
T 24 87 <170 1.0
170-250 4.8 0.9-25.1
>250 7.0 1.4-36.4 0.026 11.5^ 1.3-99.6
Thomas, 1997 T 13 59 <210 1.00
22 61 210-360 1.83 0.82-4.12
*
26 59 >360 2.39 1.01-5.65 0.045
Zeleniuch- T 85 163 <210 1.0
Jocquotte, 1997
210-291 2.4 1.0-5.6
292-415 3.5 1.4-8.4
>415 2.7 1.1-6.8 <0.05
Hankinson, 1998 T 33 75 <160 1.00 1.0
38 79 160-220 1.12 1.12 0.60-2.10
37 78 230-310 1.10 1.07 0.57-2.00
39 67 >310 1.34 0.05 1.40 0.73-2.70 0.04
Cauley, 1999 free T 10 56 <1.6 1 1.0
23 65 1.6-2.4 1.7 0.7-4.2 2.2 0.7-7.1
33 58 2.4-3.S 3.5 1.5-8.2 6.4 2.1-19.6
31 64 >3.8 2.5 1.1-6.0 0.01 3.3 1.1-10.3 0.009
T 10 57 <121 1.0 1.0
25 61 121-176 2.1 0.9-5.0 2.2 0.7-7.1
30 66 177-276 3.0 1.3-6.7 5.5 1.8-17.0
32 60 >276 2.8 1.2-6.5 0.01 3.6 1.1-11.7 0.008
* Baseline age distribution
All units converted to pg/mL
Age adjusted model
■^y/N, yes/no; NS, not significant; y, years; h,
hours
' Analysis also conducted excluding past
estrogen users
Adjusted for BMI and age only
* Adjustment did not change results from the
crude analysis
10
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 1-2 The relative risk of breast cancer associated with testosterone levels, results from case-control
studies
A uthor, year Population Exposure Cases Controls Categories* OR 95% Cl P
Secreto, 1984*’ Prem enopausal women serum T 13 47 <590 1.0
in Milan
4 9 590-670 2.6 0.6-10.9
10 6 671 + 10.2 2.6-40.0 0.0004
urinary T 9 35 <8.2 1.0
4 7 8.2-10.5 2.3 1.2-12.9
10 5 10.6+ 8.4 2.1-33.6 0.002
Secreto, 1989*’ W omen in Milan age serum T 31 51 <309 1.0
30-49 32 17 309+ 3.4 1.6-7.3 0.05
urinary T 36 50 <7.6 1.0
24 16 7.6+ 2.1 0.9-4.8 NS
Secreto, 199 V Postmenopausal serum T 16 40 <146 1.0
women in Milan <69 y
16 40 146-212 1.2 0.5-3.0
o f age
18 38 213-275 1.5 0.6-3.7
25 32 >275 2.7 1.1-6.7 0.03
urinary T * * 11 43 <18 1.0
14 38 18-31 1.2 0.5-2.9
17 37 32-46 2.2 0.8-5.7
30 26 >46 4.7 1.8-12.1 0.001
serum DHT 15 37 <36 1.0
20 40 36-57 1.6 0.7-3.7
16 38 58-82 1.3 0 .5 -3 .I
24 35 >82 2.0 0.8-5.0 NS
Lipworth, 1996' Postmenopausal serum T * 23 35 260 1.00
women from Sweden
15 27 350 0.75 0.33-1.75
47 30 470 2.64 1.27-5.46
36 30 700 2.30 0.97-5.50 0.041
“Serum T and DHT converted to pg/ml and urinary T converted to |xg/24 h
’ ’ Age adjusted model
“ Adjusted for age, occupation, number of children
^ Units in pg/ml
“ Adjusted for age and residence
Categorized by quartile medians
11
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 1-3 Mean levels of testosterone in breast cancer cases and control
Author, year Population Exposure Group N Mean“ SD
pb
Secreto, 1983 Postmenopausal women Serum T Controls 30 310 110
Carcinoma 28 550 200 0.001
Secreto, 1983 Premenopausal women Urinary T Controls 22 6.25 3.48
Familiality 21 5.41 3.60 NS
Hyperplasia 39 6.97 4.44 NS
Carcinoma 18 11.3 6.78 0.01
Secreto, 1984 Premenopausal women Serum T Controls 55 470 160
Breast Hyperplasia 31 550 200 <0.05
Breast Cancer 23 620 220 <0.005
Hill, 1985 Postmenopausal women Serum T Healthy Caucasian 43 NS
Healthy Japanese 59 NS 0.01
Cases Japanese 33 NS" 0.01
Adlercreutz, 1989 Postmenopausal women Serum T Vegetarians 10 172.80 86.40
Omnivores 9 233.28 66.24 <0.05
Cases 8 319.68 132.48 <0.05
“ Serum T converted to pg/ml and urinary T converted to pg/24 h
’ ’ Comparisons to mean levels in controls using the t-test
' Comparison to Healthy Japanese group
12
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Androgen Receptor and the Role of a Functional Polymorphism in Androgen Receptor
and Breast Cancer Risk
The main receptor for T is the AR. A functional polymorphism in the AR gene has been
examined in female breast cancer and the literature is reviewed to shed light on the
possible mechanisms by which T may affect breast cancer risk.
Androgen Receptor Protein and Breast Cancer
The AR is expressed in the majority of breast cancers (Bayer-Gamer and Smoller 2000;
Bryan and others 1984; Brys and others 2002; Hall and others 1996; Kuenen-Boumeester
and others 1992; Lea and others 1989; Soreide and others 1992). Several studies have
been conducted to examine the effects of androgens on the growth of AR positive breast
cancer cell lines. These studies have reported both inhibitory (Ortmann and others 2002;
Poulin and others 1988) and stimulatory (Hackenberg and others 1988; Marugo and
others 1992) effects. These divergent effects have been observed to be specific to the cell
line under study (Birrell and others 1995a).
To my knowledge, the only in vivo study of the effect of T on breast cell proliferation
was conducted in rats and showed that treatment with T results in both tumor regression
and a reduction in estrogen receptor (ER) expression (Zava and McGuire 1977).
However, it is unclear whether T levels used represent physiologic doses. No in vivo or
epidemiologic studies have examined the association between serum or tissue T levels
and breast cell proliferation in tumors with varying degrees of AR expression.
13
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
In summary, the effects of androgens on breast cancer ceil growth are still unclear. In
contrast to the epidemiologic data observing a consistently reported association between
serum T levels and increasing breast cancer risk, in vivo studies report an anti
proliferative effect and in vitro studies report both proliferative and anti-proliferative
effects.
The Androgen Receptor Gene and a Polymorphic CAG Repeat
The AR is coded by a single 90 kb gene on the X chromosome (Xql I-ql2) which
encodes an 11 kb mRNA transcript composed of 8 exons (Brown and others 1989; Chang
and others 1988; Lubahn and others 1989; Lubahn and others 1988; Tilley and others
1989). Epidemiologic evidence for a role of the AR gene in breast cancer was first
suggested by studies of male breast cancer patients. A mutation in the AR in the DNA
binding domain resulting in the inability to bind androgens was first reported in a pair of
brothers with breast cancer (Wooster and others 1992). In a study of 13 male breast
cancer cases, 1 case was observed to carry a similar mutation (Lobaccaro and others
1993). Another small study of 11 male breast cancer cases, did not observe this mutation
(Hiort and others 1996). These results suggested that the mutation may play a role in the
development of breast cancer in some males.
Within the first exon of the AR lies a polymorphic CAG repeat that encodes a
polyglutamine tract of variable length. The normal size range of these repeats is between
6 and 39 repeats (Edwards and others 1992; Giovannucci and others 1997). Between 40
and 66 repeats have been observed in patients with a rare, neurodegenerative disorder
14
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
called spinal and bulbar muscular atrophy (La Spada and others 1991), a disease
characterized by androgen insensitivity with gynaecomastia, testicular atrophy,
oligospermia, azoospermia, and elevated serum gonadotropins.
AR-CAG Repeat Length and Androgen Receptor Activity
Several studies have observed an association between increasing AR-CAG repeat length
and a linear decrease in AR transaetivation activity (Chamberlain and others 1994; Irvine
and others 2000; Kazemi-Esfarjani and others 1995; Tut and others 1997). Consistent
with this, male carriers of the short AR-CAG repeat length are at increased risk of
prostate cancer (Ekman and others 1999; Giovannucci and others 1997; Hakimi and
others 1997; Hsing and others 2000; Ingles and others 1997; Irvine and others 1995;
Stanford and others 1997).
AR-CAG Repeat Length and Breast Cancer
The association between the length of the AR-CAG repeat polymorphism and hreast
cancer risk has been examined in several case-control studies (Table 1-4) (Dunning and
others 1999; Giguere and others 2001; Haiman and others 2002b; Kadouri and others
2001; Rebbeck and others 1999; Spurdle and others 1999; Suter and others 2003). The
long AR-CAG repeat, representative of the less active AR, was associated with a
statistically significant increase in breast cancer risk in a population of women jfrom
Quebec (Giguere and others 2001), and in a population of BRCAl mutation carriers
(Rebbeck and others 1999). Four additional studies (Dunning and others 1999; Kadouri
and others 2001; Spurdle and others 1999; Suter and others 2003) reported slightly
15
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
increased risk of breast cancer associated with the long allele, but none of these findings
were statistically significant. A study nested within the Nurses’ Health Study cohort
found no increased breast cancer risk associated with the long AR allele overall, but, an
increased risk was observed when analyses were limited to subjects with a first-degree
family history of breast cancer (OR = 1.70; 95% Cl, 1.2-2.4) (Haiman and others 2002b).
Another trinucleotide repeat in the AR, a GGC repeat, has been observed to be associated
with prostate cancer risk (Chang and others 2002; Hakimi and others 1997; Hsing and
others 2000; Platz and others 1998; Stanford and others 1997). One of three studies that
have examined the GGC repeat length and breast cancer (Dunning and others 1999;
Kadouri and others 2001; Suter and others 2003) found a significant association in
women diagnosed before age 45 (Suter and others 2003), but no evidence of an
interaction between the CAG and GGC repeat with breast cancer risk.
The three studies that reported a significant association between long AR-CAG repeat and
breast cancer risk (Giguere and others 2001; Haiman and others 2002b; Rebbeck and
others 1999) included both pre and postmenopausal women. One study stratified on
menopausal status and found that the significant association with the long AR-CAG was
observed only in postmenopausal (OR = 3.22; 95% Cl, 1.54-6.75) but not in
premenopausal women (0R= 1.03; 95% Cl: 0.43-2.48). If this effect modification is
true then it may explain, at least in part, the non-significant results in the studies
restricted to women under age 40 years (Spurdle and others 1999).
16
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Issues with the Studies o f AR-CAG Repeat Length and Breast Cancer Risk
The gene for the AR lies on the X chromosome; therefore, women carry two alleles while
men carry only a single allele. In general, normal women are a mosaic with one allele
randomly expressed in each cell. A recent study reports that 13% of young (27-45 years
old) breast cancer cases showed preferential activation of one of the AR alleles as
measured by genotyping of peripheral blood DNA, but there was no preference towards
the allele with the longer or shorter CAG repeat (Kristiansen and others 2002). Analyses
of the AR in women that only consider the length of the CAG repeat on one allele assume
that this is the active allele in the breast tissue. Analyses that use the average of the CAG
repeat lengths or the sum of the repeats consider the contributions of both alleles;
however, if only one allele is preferentially expressed then this would result in
misclassification. There is a high rate of heterozygosity in the AR-CAG repeat length,
therefore this is likely to be a major misclassification problem, which should bias the
results towards the null. Genotyping methods can be optimized to better detect if there is
a preferentially active AR allele by either genotyping tumor tissue or serum DNA using
methylation sensitive enzymes (Kristiansen and others 2002).
Summary
In summary, the studies conducted so far suggest that the long AR-CAG repeat (less
active AR) may be associated with increased breast cancer risk in women who are
postmenopausal, have a first-degree family history of breast cancer or a known BRCAI
mutation. The location of the AR gene on the X chromosome means that results from
17
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
epidemiological studies will be biased towards the null as long as we do not know which
allele is expressed.
18
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CD
■ D
—i
O
o .
c
o
CD
Q .
■ D
CD
C/)
( / )
Table 1-4 The relative risk o f breast cancer associated with the androgen receptor CAG repeat
o
o
■ D
c q '
CD
— i
CD
■ D
— i
O
o .
c
a
o
■ D
— i
O
CD
Q.
■ D
CD
( / )
( / )
Author, year Population AJi genotype* Long allele Cases Controls Adjusted for RR (95% Cl)
Rebbeck, 1999
Multi-insUititional study o f mutatioQ
carriers ascertained through families with a histoiy
o f breast/and or ovarian cancer between 1978 and
1997.
S/S
S/L+L/L
28+ repeats 19
146
1.81 (1.06-3.081'’
Spurdle, 1999
Early onset breast cancer (<40) and age matched
controls from Australia.
S/S
S/L
L/L
21+ repeats 78
189
101
71
138
75
Age, country o f birth, state, education,
marital status, d o f live births, height,
weight 1 year ago, age at mcnarche, OC
use, family history, ER polymorphism,
mother's cotmtry o f birth, father's country
o f birth
1.4(0.92-2.15)
1,4(0.87-2.26)
Dunning, 1999
Cases from East Anglian region o f the UK and
random controls from the EPIC cohort.
S/S
S /L
L/L
23+ repeats 209
215
84
160
212
54
0.82 (0.62-1.09)
1.31 (0.87-1.97)
Giguere, 2001
Incident cases from Quebec city and age and area of
residency matched controls. S /L + L /L
21+ repeats 17
238
61
400
Matched analysis
2.14 (1.22-3.73)
Kadouri, 2001
Affected and unaffected BRCAl/2 carriers from
two genetic clinics; one in Jerusalem, Israel and the
other in London, UK.
S/S
S /L + L /L
28+ repeats BRCAl/2 carriers: 0.80 (0.44-1.46)
non-carriers: 1.27 (0.83-1.96)
Haiman, 2002
Cases and controls from the Nurses' Health Study
and controls matched on year of birth, menopausal
status, postmenopausal hormone use, and time of
day, month, and fasting status at blood draw.
S/S
S /L + L /L
22+ repeats 179
548
247
713
Age at menarche, parity, age at first birth,
BMI at age 18, weight gain since age 18,
benign breast disease, first degree family
history, duration o f posUnenopausal
honnone use; matched analysis
1,06(0.83-1.35)
Suter, 2003
Cases (<;45 y) identified through tlte Cancer
Surveillance System o f Western Washington and
fiuquency-matched controls on 5-year age group
and reference year.
S/S
S/L
LL
22+ repeats 121
255
148
122
206
133
Age at reference and reference year
1.3(0.9-1.8)
1.2 (0.8-1.7)
^ S, short allele; L, long allele
Analyses using 29 and 30 repeats as the cut point produced progressively higher significant risk estimates and progressively earlier age of
onset, no trend test published
Published OR models the S/S genotype as the high-risk allele
Discussion
If the long AR-CAG repeat (less active AR) is associated with increased breast cancer
risk in postmenopausal women, then how do these results coincide with results showing
that increased T levels increase postmenopausal breast cancer risk?
One hypothesis to explain this apparent paradox is that the less active AR may be
involved in a physiological feedback associated with increased circulating T. However,
the only data available discount this hypothesis. Two studies have examined the
association between the ^i?-CAG repeat length and circulating T levels in normal
women. ^i?-CAG repeat length was inversely associated with T levels (Haiman and
others 2002b; Westberg and others 2001). In other words, the less active AR was
associated with lower circulating T levels and results were statistically significant in both
a study of premenopausal women (Westberg and others 2001) and a study of
postmenopausal women (Haiman and others 2002b).
If the AR is not involved in a feedback mechanism to influence T levels in
postmenopausal women then it is possible that the effect of T on the breast epithelium
does not act through binding to the AR. T may exert its effect on breast tissue through
conversion of T to estrone which is then aromatized into estradiol (E2) in adipose tissue,
and the increased E2 levels may result in increased breast cell proliferation and breast
cancer risk.
2 0
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
T may also exert an indirect effect on breast cancer proliferation by sequestering sex
hormone binding globulin, leaving more E2 in the non-protein bound state and able to act
on breast tissue (Lipworth and others 1996; Siiteri and others 1981). Approximately 66%
of total T is bound to sex hormone binding globulin, 31% is bound to albumin, and 2% is
bound to cortisol binding protein (Duim and others 1981). Two of the studies suggesting
an association between T and breast cancer reported that this association disappeared
when adjusting for E2 levels (Thomas and others 1997; Zeleniuch-Jacquotte and others
1997). But, in the pooled analysis the significant association between T and breast cancer
risk remained after adjustment for E2 (2002).
Finally, it is possible that further studies will show that AR-CAG repeat length is not
linked to breast cancer risk.
Conclusion
Prospectively conducted epidemiologic studies have found that increased levels of serum
T are associated with an increase in postmenopausal breast cancer risk. However, a
number of questions remain. Several lines of evidence suggest a role of AR in breast
cancer risk, and sparse epidemiologic data suggest that a long AR-CAG repeat yielding a
less active AR may be associated with increased risk. There still remain a number of
questions on how T increases breast cancer risk. While in vitro studies report both
proliferative and anti-proliferative effects of T on the growth of various breast cancer cell
lines, we still need to further understand under which in vivo circumstances does T exert
these effects. Finally, we do not know whether androgens affect breast cancer risk in
21
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
premenopausal women. Further analyses of the role of^i?-CAG repeat length and breast
cancer using genotyping methods that assess which allele is the active AR allele are
clearly needed. Additional data are also needed to help understand the apparent paradox
between the AR-CAG repeat length, T levels and breast cancer risk.
2 2
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CHAPTER 2
GRANT PROPOSAL: The Androgen Receptor and Mammographic Density
(Predoctoral Fellowship awarded by California Breast Cancer Research Program 8GB-
0010)
Introduction and Hypotheses
High androgen levels, in particular high T levels, have been shown to increase breast
cancer risk in several epidemiologic studies. The AR is expressed in breast tumor tissue
and is involved in feedback mechanisms that regulate T levels. The association between
the CAG repeat length polymorphism in the gene for the AR and breast cancer risk has
been examined in a handful of epidemiologic studies. The results from these studies are
inconsistent and some even suggest that low, rather than high, androgen activity is
associated with increased breast cancer risk.
I prepare to further our understanding of the role of AR-CAG length in breast cancer
development by examining the effect of the AR-CAG repeat length on two different
outcomes. I will address the effect of AR-CAG length on mammographic density, a
strong independent breast cancer risk factor and potential marker for breast cell
proliferation. I propose to test the hypothesis that longer repeat length is associated with
higher mammographic density. A corollary of this hypothesis is that AR activity is
associated with more advanced forms of breast cancer. To test this hypothesis, I will also
23
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
examine the association between the AR-CAG repeat length and breast cancer tumor
grade. This study will use DNA and mammographic density assessments collected on a
sample of 396 African American and Caucasian cases as part of the population based
Women’s Contraceptive and Reproductive Experiences (CARE) study (PI: Leslie
Bernstein, Ph.D.), the Radiographic Densities and Breast Cancer Prevention study
(BCRP# 2RB 0058, Pis: Malcolm Pike/Giske Ursin) and the Mammographic Density and
Sex Steroid Genes study (BCRP# 6IB 0093, Pis: Sue A. Ingles/Giske Ursin).
Specific Aims
I will test the hypothesis that AR activity influences mammographic density by:
AIM I : Determining the association between mammographic density and the AR-CAG
repeat polymorphism.
AIM 2: Determining the association between mammographic density and AR-CAG
repeat polymorphism stratified by ethnicity, menopausal status, and use of hormone
replacement therapy.
I will also test the hypothesis that AR activity is associated with more advanced breast
cancers by:
AIM 3: Determining the association between polymorphism in AR and the following
prognostic indicators: tumor size, stage, grade, and nodal status.
2 4
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Background and Significance
Androgens and Breast Cancer Risk
Six of the eight published studies that examine the influence of T, measured
prospectively, on breast cancer risk report a statistically significant increase in
postmenopausal breast cancer risk with increasing levels of serum T (Berrino and others
1996; Cauley and others 1999; Dorgan and others 1996; Hankinson and others 1998;
Thomas and others 1997; Zeleniuch-Jacquotte and others 1997). While one of these
studies did not publish significant point estimates, a significant trend in risk was reported
(Hankinson and others 1998). Only two of the eight studies failed to demonstrate any
significant association (Garland and others 1992; Wysowski and others 1987).
Biological Rationale for a Role o f AR in Breast Cancer
AR expression has been observed in 50-90% of breast cancers; the more recent studies
using more modem assays show the higher levels of expression (Bryan and others 1984;
Kuenen-Boumeester and others 1992; Soreide and others 1992).
An in vitro study published by Birrell et al. (Birrell and others 1995a) observed that T
affects breast cancer cell proliferation through binding to the AR. Results are ambiguous
since both proliferative and inhibitory effects of 5a-dihydroT and a non-metabolizable
androgen, mibolerone, were observed in AR-positive cell lines. These results confirm
that T binding to AR has an influence on breast cancer tumor growth, although the
direction is still unclear.
25
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
AR Polymorphism
The AR is a member of the steroid hormone receptor subfamily of the nuclear hormone
receptor super family. Like all steroid hormone receptors, it is composed of four
domains: The N-terminal transactivation domain, the DNA binding domain, the
translocation domain, and the C-terminal ligand binding domain (Mangelsdorf and others
1995). Upon binding of the ligands, T or 5a-dihydrotestosterone (DHT), to the C-
terminal domain, the receptor directly targets sequences in DNA called androgen
response elements to regulate gene transcription (Adler and others 1992; Ho and others
1993; Roche and others 1992).
The AR is coded by a single 90 kilobase gene on the X chromosome (Xql l-ql2) which
encodes an 11 kilobase mRNA transcript composed of 8 exons (Brown and others 1989;
Chang and others 1988; Lubahn and others 1989; Lubahn and others 1988; Tilley and
others 1985). The entire N-terminal transactivation domain is encoded by the first exon.
Within the first exon lies a polymorphic CAG repeat that encodes a polyglutamine tract
of variable length. The normal size range of these repeats is between 6 and 39 repeats
(Edwards and others 1992; Giovannucci and others 1997). Expansion of repeat to
between 40 and 66 repeats has been observed to result in an androgen insensitive
phenotype of a rare, neurodegenerative disorder called spinal and bulbar muscular
atrophy (La Spada and others 1991). In addition, several epidemiologic studies have
observed an association between being a carrier of the short XR-CAG repeat length and
increased prostate cancer risk (Ekman and others 1999; Giovannucci and others 1997;
Hakimi and others 1997; Hsing and others 2000; Ingles and others 1997; Irvine and
26
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
others 2000; Stanford and others 1997). Based on these observations, functional studies
have identified a relationship between increasing yfi?-CAG repeat length and a linear
decrease in AR transactivation activity (Chamberlain and others 1994; Irvine and others
2000; Kazemi-Esfarjani and others 1995; Tut and others 1997).
AR-CAG Repeat and Breast Cancer Risk
Several epidemiological studies have been conducted to examine the association between
the length of the AR-CAG repeat polymorphism and breast cancer risk using a case-
control study design (Dunning and others 1999; Giguere and others 2001; Kadouri and
others 2001; Rebbeck and others 1999; Spurdle and others 1999). Two of the five studies
conducted demonstrated that the long ^R-CAG repeat, and therefore, the less active AR,
is associated with a significant increased breast cancer risk in a population of BRCAI
mutation carriers (Rebbeck and others 1999) and in a population of non-carriers (Giguere
and others 2001). After stratification by menopausal status, the significant association
observed remained only in postmenopausal women (Giguere and others 2001). These
results suggest that low, rather than high T activity increases breast cancer risk in
postmenopausal women and in BRCAI mutation carriers. However, two of the five
studies reported null associations. One of these was in only premenopausal women
(Spurdle and others 1999), another was of a population of Ashkenazi Jewish women
(Kadouri and others 2001). The results reported in the fifth study were supportive of the
observed association but not significant, insufficient power is an unlikely explanation
since this study was one of the largest. However, insignificant results may be due to
limitations in the analysis of the CAG repeat length as a dichotomous variable using a
27
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
cutpoint based on previous studies of prostate cancer risk (Dunning and others 1999).
Both Rebbeck and Giguere categorized the AR variable based on thorough examination
of the AR distribution in their own study populations.
AR-CAG Repeat and Breast Cancer Prognostic Indicators
The association in BRCAI mutation carriers also showed an association between
increasing CAG repeat length and progressively earlier age of breast cancer diagnosis
(Rebbeck and others 1999). Based on these results, additional case-only studies have
also been conducted to examine the association of the AR-CAG repeat length with age of
onset (Given and others 2000; Menin and others 2001) or various prognostic indicators
such as tumor size, stage, grade, nodal status, ER and progesterone receptor (PR) status
(Yu and others 2000). A significant association between short, rather than long, AR-CAG
repeats and increased breast cancer tumor grade was observed in the only study to
examine this effect (Yu and others 2000). These results contradict case-control study
results showing that the long AR-CAG was associated with increased breast cancer risk.
The study results published by Yu et al. (Yu and others 2000) also contradict results from
studies examining AR expression. AR expression has been observed to correlate with
indicators of good prognosis such as decreased tumor grade (Brentani and others 1986;
Isola 1993; Kuenen-Boumeester and others 1992), increased survival (Bryan and others
1984; Langer and others 1990), and a positive response to tamoxifen (Bryan and others
1984). In addition, it has been shown that expression of AR is correlated with ER
positive status (Brentani and others 1986; Isola 1993; Kuenen-Boumeester and others
28
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1992; Lea and others 1989; Miller and others 1985; Soreide and others 1992; Valyani and
others 1987). The study by Bryan et al. (Bryan and others 1984) confirmed that the
effect of AR on prognosis is independent of that of ER by showing that the addition of
AR significantly improved the response to tamoxifen among subjects who are ER
positive.
One of the aims of this study will be to examine the association between the AR-CAG
repeat length and breast cancer tumor size, stage, grade, and nodal status using a
population based study setting. Cases will be selected from CARE study, in Los
Angeles County (PI: Leslie Bernstein, Ph.D.) with information on possible confounders,
and sufficient sample size to stratify on menopausal status.
Mammographic Density and Breast Cancer Risk
Mammographic density, measured as either percent density or using Dr. John Wolfe’s
criteria (Wolfe 1976) has been found to be strong and independent breast cancer risk
factor in numerous epidemiological studies (Boyd and others 1995; Brisson and others
1982; Byrne and others 2001; Byrne and others 1995; Carlile and others 1985; de Stavola
and others 1990; Kxook 1978; van Gils and others 2000; Whitehead and others 1985;
Wolfe and others 1987), (Ursin et al. (in preparation)). It has been suggested that
mammographic density may represent a marker for breast cell proliferation (Spicer and
others 1994). In this study we will use a computer-assisted method of assessing density
(Ursin and others 1998). This method has shown to correlate well with the expert
outlining method (Ursin and others 1998), and can strongly predict breast cancer risk
29
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
(Ursin et al, in preparation). This method was used to measure mammographic density as
part of the Radiographic Densities and Breast Cancer Prevention Study at the University
of Southern California (BCRP #2RB 0058) and will be used in the proposed study.
Summary
The AR has been shown to have varying associations with breast cancer risk.
Contradictory to results showing that T increases breast cancer risk, two studies support
the hypothesis that the less active AR (long CAG repeat) increases breast cancer risk. On
the other hand, another study shows that the more active AR (short CAG repeat)
increases risk for more advanced types of breast cancer as measured by breast cancer
tumor grade. These initial studies support the need for further analyses of the role of the
AR in breast cancer etiology. To date, there has been no published examination of the
effects of variability in the AR on mammographic density. Mammographic density is a
strong breast cancer risk factor. It is still unclear whether AR activity may influence
development of more aggressive forms of cancers, and if so, whether this association is
through its influence on mammographic density. If^R-CAG repeat length is associated
with more aggressive forms of breast cancer and mammographic density, then these
observations support interventions to reduce mammographic density in women and,
therefore, increase the ability to detect breast cancer earlier.
30
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Preliminary Results
In this study I will use women who had their mammograms scanned as part of the
Radiographic Densities and Breast Cancer Prevention study (BCRP # 2RB 0058, Pis
Malcolm Pike/Giske Ursin). Determinations of mammographic densities were conducted
by Dr. Ursin on the craniocaudal image from the cancer free (contralateral) breast of
participating cases using a method previously described (Ursin and others 1998).
Mammograms were read in groups of 30-60, where each reading group consisted of a
proportional number from each ethnic group. Preliminary results are listed in Table 2-1
on the full sample.
The study sample that will be used in the proposed study has been used in the BCRP
funded study, Mammographic Density and Sex Steroid Genes (BCRP# 6IB 0093, Pis:
Sue A. Ingles/Giske Ursin); however, no significant associations between the hormone
metabolic genes, CYPl 7, COMT, 17HSDB1, 3HSDB1 and mammographic density were
observed.
The AR-CAG repeat and Breast Cancer Grade
I have experience examining the AR-CAG repeat in a sample of breast cancer cases at
City of Hope/Beckman Research Institute. My doctoral dissertation will use the analyses
funded by this award to supplement my initial work at City of Hope (Lillie unpublished).
Preliminary analyses (Table 2-2), although not significant, support the association
observed by Yu, that the short AR-CAG repeat length is associated with increased breast
31
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
cancer tumor grade. Failure to observe a significant association may be due to
insufficient power to observe a significant association.
Table 2-1 Distribution of breast cancer cases and controls according to mammographic density
Mammographic
Adjusted
density Cases Controls Crude OR OR" 95% Cl P
<1 23 36 1.00 1.00
1-9 70 75 1.46 1.47 0.77-2.80
10-24 109 94 1.82 1.58 0.84-2.97
25-49 266 177 2.35 2.04 1.12-3.74
50-60 97 55 2.76 2.55 1.29-5.06
60-r 110 51 3.38 3.01 1.50-6.03 0.0002
“ S, short allele; L, long allele
’ ’ Analyses using 29 and 30 repeats as the cut point produced progressively higher significant risk estimates
and progressively earlier age of onset, no trend test published
Published OR models the S/S genotype as the high-risk allele
Table 2-2 Analysis o f androgen receptor and breast cancer tumor grade in breast cancer cases with a family
history of breast cancer
AR genotype Grade <3 Grade 3+ OR 95% Cl
Average CAG 22.46 21.68 0.82 0.65-1.04
CAG categorical"
L/L 21 13 1.00
S/S+S/L 20 23 1.86 0.74-4.64
' S, <21 repeats; L, >21 repeats determined based on median of the distribution
32
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Research Design and Methods
Design
1) Genotype all study subjects for the AR-CAG repeat length polymorphism.
2) Conduct analyses to determine the associations between the AR polymorphism and
mammographic density, tumor size, grade, nodal status, and stage.
Methods
Subject Identification:
Study subjects will come from the breast cancer case-control study. Women’s CARE
study in Los Angeles County (PI: Leslie Bernstein, Ph.D.). The case patients were
diagnosed between July 1994 and April 1998, and were identified by the Los Angeles
County Cancer Surveillance Program (LACCSP), a Surveillance, Epidemiology, and End
Results (SEER) registry program. Cases also consented to let us access their
mammograms. We will use women who had their mammograms scanned as part of the
Radiographic Densities and Breast Cancer Prevention study (BCRP # 2RB 0058, Pis:
Malcolm Pike/Giske Ursin). We will use the subset of 396 women that also had a blood
sample, this is the same group of women that were used in the Mammographic Density
and Sex Steroid Genes study (BCRP# 6IB 0093, Pis: Sue A. Ingles/Giske Ursin).
33
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Questionnaire:
All women who consented to the study underwent a 75 minute interview with trained
interviewers using a questionnaire with a fairly standard set of questions on established
and suspected breast cancer risk factors (demographics, pregnancies, menstruation,
menopause, surgery, hormonal contraception, hormone replacement, medication,
infertility, medical history, mammogram history (for the 5 years preceding the
diagnosis/reference date), physical activity, smoking, body size, alcohol consumption,
family history of cancer, and prenatal exposures).
Laboratory Methods:
Using a standard protocol, genomic DNA was extracted previously from peripheral
blood. The AR-CAG repeat will be genotyped at a laboratory at the City of
Hope/Beckman Research Institute. Genotyping methods for AR-CAG repeat have been
optimized for use as a non-patemity marker in the study of affected sib pairs with cancer.
Mapping Interactive Cancer Susceptibility Loci (PI: Theodore Krontiris, M.D. Ph.D.).
All genotyping costs for the proposed study will be covered by City of Hope funds (see
memo in appendices).
The exon 1 CAG repeat of the AR will be PCR amplified using a FAM labeled forward
primer (5’-TCCAGAATCTGTTCCAGAGCGTGC-3’) and reverse primer (5’-
GCTGTGAAGGTTGCTGTTCCTCAT-3’). All microsatellite genotyping will be
performed using an ABI 377 with Genescan and Genotyper software (PE/Applied
Biosystems). Direct sequencing of a sample of homozygotes for at least three different
34
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
lengths will be conducted to determine the exact number of CAG repeats for each size
fragment. All allele calls will be made blinded to outcome data.
Statistical Analysis:
Linear regression models to test the association between the AR and mammographic
density will be used. The basic model will have the form
MD=a + pAR + Y X + e
where X is a vector of additional adjustment covariates (e.g. age, ethnicity) and the
residuals (s) are assumed to be independent and normally distributed. The parameter of
primary interest is p, which measures the association between the AR-CKG repeat length
and mammographic density. Transformations to mammographic density will be used if
necessary to satisfy the model assumptions. Other outcome variables such as tumor size,
stage, grade, and nodal status can replace mammographic density in the equation.
Categorical outcomes will be modeled using logistic regression techniques.
The AR will be treated as a continuous variable that considers the contributions of both
allele lengths as an average length variable. The difference in the allele lengths will be
included in the model to adjust for the difference in effect of subjects heterozygous for
two alleles of extreme lengths versus subjects homozygous for alleles close to the mean.
This model will test for a linear trend in mammographic density with increasing number
of CAG repeats. In order to test for a threshold type effect in number of CAG repeats on
mammographic density, the AR will be categorized into S/S, S/L and L/L genotypes. The
35
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
median of the AR-CAG repeat lengths in this study sample will be used as the cutpoint of
S versus L for each allele.
Relative amount (% of breast) of mammographic densities will be modeled as the
primary outcome both continuously and categorically. Cutpoints of mammographic
density in this study sample have been previously established as part of the Radiographic
Densities study (2RB 0058). Prognostic indicators will be analyzed as dichotomous
outcomes.
All regression models will be adjusted for age, and if necessary, for family history of
breast cancer. Quality of mammograms, as well as group in which the mammogram was
read will also be evaluated to control for the possible effects of inter and intra assay
measurement errors. In addition, models will be stratified on ethnic group, menopausal
status, and hormone replacement therapy use categories. The RR of having a ‘high risk’
mammographic density or a more advanced stage prognostic indicator associated with
different AR-CAG repeat lengths will be estimated using the odds ratio. Analyses will be
performed using the SAS statistical software package (SAS Institute, Inc. Cary, NC).
Power Calculations:
Since models of the AR genotype as a continuous variable are inherently more powerful
to detect significant effects on the various outcomes being measured, power calculations
were conducted using the AR genotype as a dichotomous categorical exposure variable
(S/S+S/L vs. L/L). The data reported in the preliminary results was used to calculate the
36
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
percent difference in mammographic density and RR of a dichotomous outcome that can
be detected with 90% power at the 0.05 level of statistical significance. With a sample
size of 396 cases, a standard deviation of 21.7% mammographic density, and a
distribution of 50% of cases carrying the S/S+S/L genotype (Table 2-2), this study will be
able to detect a 7.1% difference in mammographic density between carriers of the
S/S+S/L genot3 ^ e and carriers of the L/L genotype. Using the same 50% exposure rate
{AR genotype), and an equal distribution of cases by tumor grade (Table 2-2), this study
will have 90% power to detect a RR of 1.98.
37
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CHAPTERS
DATA ANALYSIS: Polymorphism in the Androgen Receptor and Mammographic
Density in Women Taking and Not Taking Estrogen and Progestin Therapy (Lillie and
others 2004)
Introduction
Evidence suggests that circulating androgen levels in postmenopausal women are
associated with breast cancer risk (2002). In a pooled analysis of 9 prospective studies of
postmenopausal women, the trend in breast cancer risk associated with increasing level of
T was statistically significant. The main receptor for T and DHT in breast tissue is the
AR. Within the first exon of the AR gene is a polymorphic CAG repeat. Several
epidemiologic studies have examined the association between the length of the CAG
repeat polymorphism in the AR gene and breast cancer risk (Dunning and others 1999;
Giguere and others 2001; Haiman and others 2002b; Kadouri and others 2001; Liede and
others 2003; Rebbeck and others 1999; Spurdle and others 1999; Suter and others 2003).
Three studies have observed that women with the less active AR allele (long CAG repeat)
have increased breast cancer risk (Giguere and others 2001; Haiman and others 2002b;
Rebbeck and others 1999), although the evidence is not entirely consistent (Dunning and
others 1999; Kadouri and others 2001; Spurdle and others 1999; Suter and others 2003).
Only one study has examined the association between the AR genotype and mammographic
density, no association was observed (Haiman and others 2003). This apparent
38
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
contradiction, that high androgen levels, but the less active AR genotype are associated
with breast cancer risk is not well understood. To shed more light on the ^i?-breast
cancer association, we examined the association between the AR-CAG repeat length and
mammographic density, a strong and independent breast cancer risk factor (Boyd and
others 1995; Byrne and others 1995; Ursin and others 2003). We previously reported
results demonstrating that mammographic density represents a strong risk factor for
breast cancer in this study (Ursin and others 2003), and furthermore, that genes important
in estrogen metabolism do not explain the variation in mammographic density observed
in these women (Haiman and others 2002a).
Materials and Methods
Study Population
Study subjects were women newly diagnosed with a first primary invasive breast cancer
who participated in the Los Angeles part of the Women’s CARE Study (Marchbanks and
others 2002). The Women’s CARE Study was a population-based case-control study of
invasive breast cancer in United States-bom African-American and Caucasian women
aged 35-64 years. All of the patient participants were diagnosed between July 1994 and
April 1998. Each participant provided a written informed consent at the time of
interview. All of the women were interviewed in person by trained interviewers using a
standardized questionnaire to collect information on demographic factors and potential
breast cancer risk factors including histories of oral contraceptive use, hormone therapy
(HT) use, family history of breast cancer, age at menarche, complete pregnancy history,
lifetime history of participation in physical exercise activities, and mammogram history
39
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
during the five years preceding the patient’s date of diagnosis. We restricted eligibility
for this study to breast cancer patients who reported having a screening mammogram in
the 5 years prior to breast cancer diagnosis. We requested screening mammograms for
60.9% of the patients included in the Los Angeles portion of the CARE study and who
provided written permission for the release of mammograms (N = 755, 426 Caucasian-
Americans and 329 African-Americans). We received mammograms for 71.1% (N =
303) of the eligible Caucasian patients and 69.3% (N = 228) of eligible African-American
patients (Ursin and others 2003). During the conduct of the Women’s CARE Study in
Los Angeles County, we collected a blood sample from all patients willing to provide a
sample. Of the 531 cases for whom we had a screening mammogram, we had a blood
sample for 423 (79.7%) women (Haiman and others 2002a). The study protocol was
approved by the Institutional Review Board at the University of Southern California.
Genotyping
Using a standard protocol, genomic DNA was extracted from peripheral blood (Haiman
and others 2002a). DNA samples were sent to City of Hope Beckman Research Institute
for AR genotyping. The exon 1 CAG repeat of the AR was amplified by PCR using a
fluorescently labeled forward primer (5 ’ -TCCAGAATCTGTTCCAGAGCGTGC-3 ’) and
reverse primer (5’-GCTGTGAAGGTTGCTGTTCCTCAT-3’). All of the microsatellite
genotyping was performed using an ABI 377 sequencer with Genescan and Genotyper
software (PE/Applied Biosystems, Foster City, Cahfomia). The number of CAG repeats
was calculated based on direct sequencing results of 8 prostate cancer cases with different
known PCR product lengths.
40
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Mammographic Density
We measured percent mammographic density using the University of Southem California
Madena computer based threshold method of assessing density (Ursin and others 1998).
Percentage of mammographic density in the right and left breasts of both case and control
women have been shown to be highly correlated (Benichou and others 2003). Therefore,
we used screening mammograms of the contralateral (non-diseased) breast obtained
before diagnosis. The cranio-caudal mammographic images were digitized using a high-
resolution Cobrascan CX-312T scanner (Radiographic Digital Imaging, Torrance, CA)
and then viewed on a computer screen. A single reader blinded to all of the patient
characteristics evaluated all images. The reader first defined the total breast area using a
special outlining tool. The density assessments were done as follows. The reader defined
the region of interest (KOI) excluding the pectoralis muscle and other light artifacts
(prominent veins, and so forth). The reader then applied a yellow tint to gray levels
above a selected threshold using a computerized tool. The area highlighted in this
manner was considered to represent dense tissue. The software then calculated the
number of pixels within the entire breast as outlined and the number of pixels tinted
yellow within the ROI. The ratio of the number of tinted pixels within the ROI to the
total number of pixels in the breast represents the measure of percent density.
The scanned mammogram files of 15 women could not be assessed for density as the
digitized files were unusable and we were unable to obtain the films a second time. We
excluded the mammograms of 4 women who had only one mammogram and were
41
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
pregnant at the time of mammogram. Therefore, we obtained mammographic density
results for 404 breast cancer patients for whom we had a blood sample.
Menopausal Status/Hormone Therapy Use Status at Mammography
Women were assigned a menopausal status at the time of their mammogram. Women
who had menstruated and not used HT within three months before mammography were
defined as premenopausal. Women who had not menstruated within three months prior to
mammogram, who had a bilateral oophorectomy more than three months prior to
mammogram, who had a simple hysterectomy prior to mammogram with the last
menstrual period more than 6 months prior to surgery, who were 50 years of age or older
and were current or past HT users, or who were 60 years of age or older were defined as
postmenopausal. Otherwise, menopausal status was considered as unknown.
Postmenopausal women were further categorized based on their HT usage as never, past
or recent (within the past 5 years). Type of HT used was further specified as estrogen
therapy (ET) or estrogen progestin therapy (EPT). Women who had used both ET and
EPT were assigned their most recent HT regimen.
Statistical Methods
We classified each allele of the AR gene as S or L using the median number of CAG
repeats across all alleles in the study population, 21 repeats, as the cut-point. Each
woman was categorized by genotype as S/S, S/L or L/L. We used least squares linear
regression methods to model the dependence of percent mammographic density on AR
42
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
genotypes. All of the models included adjustment for the following set of potential
confounders selected a priori based on their previously reported association with density
and/or the AR genotype: age at mammogram (years); 35-39, 40-44, 45-49, 50-54, 55-59,
60-64; BMI (kg/m^): <22, 22-24.9, 25-29.9, >30; and race: African-American, Caucasian.
In addition to using linear regression, we also modeled percent density as a categorical
outcome using ordinal logistic regression methods. A four-category variable was created
representing quartiles of percent density based on the distribution of percent density
(range, 00.0-85.7%) in the study population. We calculated ORs to estimate the odds of
having a single quartile higher level in percentage of mammographic density associated
with the AR genotype. In addition to analyzing the entire sample, both linear and logistic
regression models were applied to subgroups defined by race, family history of breast
cancer (mother or sister), menopausal status, age at mammogram and HT status to
examine possible effect modification. Data were analyzed using SAS v9 software (SAS
Institute Inc., Cary, North Carolina).
Results
Our sample included 246 Caucasian and 158 African-American breast cancer patients,
with mean age at the time of mammogram 48.4 years (SD = 8.5). The mean percent
mammographic densities of African-Ameri can women (34.1%) and Caucasian women
(34.3%) did not differ (Table 3-1). Mean percent density did vary substantially by age at
mammogram, menopausal status and HT use, BMI, pregnancy history, and age at
menarche, but not by family history of breast cancer.
43
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 3-1 Mean percentage of mammographic density by descriptive characteristics (N — 404)
Characteristic N %
Mean %
density P ‘
Race”
African-American 158 39.1 34.1
Caucasians 246 60.9 34.3 0.95
Menopausal status/HT use
Premenopausal 191 47.2 35.6
Postmenopausal
Never used HT 56 13.9 30.6
Past user of EPT 25 6.2 43.2
Past user o f ET 21 5.2 29.4
Recent user of EPT 51 12.6 39.3
Recent user of ET 42 10.4 34.8
Unknown menopause 18 4.5 37.7 0.01
Age at mammogram (years)
35-39 92 22.8 45.1
40-44 84 20.8 41.2
45-49 42 10.4 37.3
50-54 76 18.8 34.0
55-59 70 17.3 27.9
60-64 40 9.9 26.7 <0.0001
BMI (kg/m^)"
<22 114 28.2 49.7
22-24.9 108 26.8 38.2
25-29.9 114 28.2 28.5
>30 68 16.8 20.5 <0.0001
Parity/age at first full tenn pregnancy (years) “
Nulliparous 85 21.0 44.3
l-2/<24 80 19.8 34.0
l-2/>24 97 24.0 40.2
>3/<24 116 28.7 27.5
>3/>24 26 6.4 32.1 <0.0001
Age at menarche (years)'
<12 108 26.7 32.1
12 108 26.7 35.4
13 99 24.5 34.0
>13 89 22.0 40.8 0.04
First-degree family history o f breast cancer ‘
No 326 80.7 35.2
Yes 57 14.1 37.3
Unknown 21 5.2 33.2 0.72
“ Analysis o f covariance
’ ’ Adjusted for age and BMI
Adjusted for age at mammogram
4 4
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
For all of the subjects, the number of Ji?-CAG repeats ranged from 10 to 35 (median:
21). Stratification by racial group indicated that the AR-CKG repeat distribution for
African-Americans (range, 10-32; median, 20) was statistically significantly lower than
that of Caucasians (range, 10-35; median, 22; P <0.0001).
Mean percent mammographic density was compared across the three AR genotype
categories S/S, S/L, and L/L (Table 3-2). We observed no statistically significant
associations between percent mammographic density and Ji? genotype in all of the
subjects. On average, subjects with the L/L genotype had higher percent mammographic
density (35.6%) than subjects with the S/S or S/L genotype (33.5%), but this difference
was not statistically significant. There were also no statistically significant differences in
mean percent mammographic density within subgroups defined by race, family history or
menopausal status.
We stratified postmenopausal women according to their histories of HT use (never HT,
ever HT, and within the group who had used HT, according to whether their most recent
use was ET or EPT). Among the EPT users, carriers of the L/L genotype had a
statistically significant higher percent mammographic density (41.4%) than carriers of the
S/S genotype (25.7%; P - 0.04). The results were comparable when this analysis was
restricted to recent (within the last year) EPT users {P = 0.03). The results were also
comparable when this analysis was restricted to Caucasians (68% of ever HT users; data
not shown). Adjustment for other covariates, such as family history of breast cancer,
parity, age at first full term pregnancy, and age at menarche did not change the results {P
45
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
= 0.01). In ET users, L/L carriers had a similar percent density (32.8%) to that of the S/S
carriers (34.5%, P = 0.83).
Table 3-2 Least-squares mean percentage breast density by androgen receptor genotype
AR Genotype
Characteristic S/S S/L L/L L/L vs S/S P P trend
All
N 74 185 145
%density“ 34.8 33.0 35.6 0.79 0.50 0.64
Caucasians
N 24 97 125
%density'’ 36.2 33.6 35.3 0.84 0.78 0.86
African-Americans
N 50 88 20
%density*’ 35.4 33.9 39.8 0.40 0.48 0.62
Premenopausal
N 35 90 66
%density“ 39.3 36.8 41.1 0.71 0.44 0.55
Postmenopausal
N 34 85 76
%density“ 34.0 33.0 34.2 0.95 0.92 0.89
Ever HT
N 22 65 52
%density° 31.4 33.5 37.3 0.29 0.50 0.25
Never HT
N 12 20 24
%density“ 38.2 33.6 26.0 0.11 0.27 0.11
ET users
N
%density“
12
34.5
27
31.6
25
32.8 0.83 0.92 0.86
EPT users
N 10 38 27
%density“ 25.7 35.7 41.4 0.04 0.11 0.04
Current EPT user‘ s
N 9 32 22
%density'‘ 24.6 37.9 41.8 0.03 0.10 0.05
“ Adjusted for race: African-American, Caucasian; age at mammogram (years): 35-39,40-44, 45-49, 50-54,
55-59, 60-54, 55-59, 60-64; BMI (kg/m^): <22, 22-24.9, 25-29.9, >30
’ ’ Adjusted for age at mammogram and BMI
Current EPT=used E P T within 1 year prior
4 6
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
In assessing the association between the AR genotype and a higher quartile in percentage
mammographic density for each subgroup of HT usage (Table 3-3), we found that among
women with a history of HT use, there was greater odds of having a single quartile
increase in percent density with each L allele; however, this was not statistically
significant in the multivariate model (adjusted OR = 1.55; 95% Cl: 0.93-2.60, p trend =
0.10). In never HT users, there was a statistically significant protective effect of the L
allele on increasing density; however, this was not observed in the generalized linear
model or in the model adjusting for family history of hreast cancer, parity, age at first full
term pregnancy, and age at menarche. After stratifying by type of HT use we found a
statistically significant higher odds for each L allele among EPT users (adjusted OR =
2.59; 95% Cl, 1.20-5.60; P trend = 0.02), but not among ET users (adjusted OR = 1.02;
95% Cl, 0.46-2.25; P trend = 0.97).
A formal test of interaction between EPT use and the AR genotype was not statistically
significant on a multiplicative scale { P - 0.10).
47
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Because no standard exists for classifying ^i?-CAG repeat lengths into S and L alleles,
we categorized the AR genotype into ten genotypes using the cut-points of 19, 21, and 24
to generate the alleles: very short (VS), medium short (MS), medium long (ML), and
very long (VL). Sample sizes were small in the ten strata; however, density was lowest
in the VS/VS group of EPT users (13.0%) and increased as number of repeats increased
up to the ML/ML group (43.2%) and then remained high in the ML/VL (39.9%) and
VL/VL (56.4%) genotypes suggesting that the findings were similar regardless of
threshold value.
48
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 3-3 Association for a single quartile increase in percent mammographic density per Tong’ androgen
receptor allele
% Density” Unadjusted Adjusted*’
AN genotype <17 18-37 37-54 55+ OR 95% Cl P trend OR 95% Cl P trend
EverHT
SS 8 7 4 3
SL 15 23 17 10
LL 10 15 14 13 1.54 1.00-2.39 0.05 1.55 0.93-2.60 0.10
Never HT
SS 3 6 0 3
SL 11 3 5 1
LL 10 8 2 4 0.85 0.46-1.59 0.61 0.38 0.15-0.95 0.04
ET user
SS 5 3 2 2
SL 8 10 7 2
LL 7 7 7 4 1.30 0.71-2.40 0.39 1.02 0.46-2.25 0.97
EPT user
SS 3 4 2 1
SL 7 13 10 8
LL 3 8 7 9 1.85 0.98-3.46 0.06 2.59 1.20-5.60 0.02
Past EPT user'
SS 0 0 1 0
SL 0 3 1 2
LL 0 3 0 2 0.80 0.14-4.50 0.80
Current EPT user'
SS 3 4 1 1
SL 7 10 9 6
LL 3 5 7 7 2.05 1.03-4.06 0.04
“ Frequency o f subjects in each subgroup defined by AJ^ genotype and percent mammographic density
category
Adjusted for race: African-American, Caucasian; age at mammogram (years): 35-39,40-44,45-49, 50-54,
55-59, 60-64; BMI (kg/m^): <22, 22-24.9, 25-29.9, >30
Current EPT=used EPT within 1 year prior to mammogram
4 9
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Discussion
In this study the number of CAG repeats in the AR gene was not associated with
mammographic density in all women combined or in subgroups based on race,
menopausal status or family history of breast cancer. The long CAG repeat (>21 repeats)
was strongly associated with increased percent mammographic density in
postmenopausal women who were EPT users.
An increased breast cancer risk has been observed among women with the long CAG
allele in three previous studies (Giguere and others 2001; Haiman and others 2002b;
Rebbeck and others 1999), whereas only a weak, non-significant association was
observed in four studies (Dunning and others 1999; Kadouri and others 2001; Spurdle
and others 1999; Suter and others 2003). The three positive studies were conducted
among BRCAl mutation carriers (Rebbeck and others 1999), postmenopausal women
(Giguere and others 2001), and women with a first-degree family history of breast cancer
(Haiman and others 2002b). Our observed association in EPT users is compatible with
the study that found an association in postmenopausal women (Giguere and others 2001).
Mammographic density may have a strong genetic component. A study of twins reported
significant higher correlation in mammographic density between monozygotic twins than
dizygotic twins (Boyd and others 2002). No data were presented on whether this
association was stronger in twins who were both EPT users. Genes involved in estrogen
metabolism have not shown to be associated with mammographic density in this study
population (Haiman and others 2002a). Using data from the Nurses’ Health Study,
50
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Haiman et al. (Haiman and others 2003) observed no association between
mammograpbic density and the AR genotype. However, no analyses stratified by type of
HT were reported in either of these studies.
Prior epidemiological studies have found that EPT is associated with higher percent
mammographic density (Greendale and others 1999; Greendale and others 2003; Laya
and others 1995; Lundstrom and others 1999; Persson and others 1997; Stomper and
others 1990). In the placebo controlled Postmenopausal Estrogen and Progesterone
Intervention trial, women in the EPT treatment group underwent on average a five
percent increase in mammographic density (Greendale and others 2003). However, the
factors that determine the variation in the percent density response in the EPT arm of this
trial are not known. Our results suggest that AR could be one such factor.
Ligands that bind AR with high affinity include not only T and DHT, but also progestins
(Poulin and others 1989). Human clinical data suggest that the AR may mediate the
antiproliferative effects of high-dose synthetic progestins such as medroxyprogesterone
acetate on breast cancer cells. In one study, the response rate to medroxyprogesterone
acetate was significantly associated with the presence of AR {P <0.001) with a shorter
progression free interval in subjects with higher AR content (Birrell and others 1995b).
No data exist on how the much less potent EPT, known to stimulate breast cell
proliferation (Hofseth and others 1999), affects AR.
51
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Our analysis has some limitations that must be considered in interpreting results. First,
we included a subset of breast cancer patients from the Women’s CARE Study who bad
provided both a mammogram and a blood sample. This introduces the possibility of
selection bias. However for this to occur and explain the EPT findings, EPT using
patients with long AR-CAG repeats and high mammographic density would have bad to
be more likely to have participated. This seems unlikely.
Second, we assessed the relationship between germ-line genotypes and mammograpbic
density of breast cancer patients only. Although we used mammograms of the
contralateral (unaffected) breast obtained before or at the time of diagnosis, these may not
be comparable to mammograms from women without breast cancer. If there are other,
unknown cofactors that are independent risk factors for breast cancer, and that also
interact with the Hi?-CAG repeat length to increase mammographic density in EPT users,
then our estimates of the impact of the AR-CAG repeat on mammographic density may
not represent that of healthy women.
Third, a proportion of the heterozygotes for S/L may be misclassified. The AR gene is
located on the X chromosome, and, -10% of breast cancer cases between the ages of 27
and 65 show preferential X-inactivation (Kristiansen and others 2002). Because only the
S or the L allele is expressed in subjects that have one X chromosome inactivated, the
heterozygote subjects with preferential X-inactivation would be more accurately grouped
as a S/S or the L/L genotype. The effect of such misclassification would be a bias
52
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
towards the null, and this may explain why some of our results for the S/L genotype are
weaker than those of the L/L genotype alone.
A final limitation in the analysis of the AR genotype is the method of dichotomozing the
allele length into S versus L. The number of CAG repeats in the AR alleles ranged from
10 to 35 in our study. It has been observed that with increasing length, AR activity
decreases (Chamberlain and others 1994). However, to our knowledge no threshold
number of repeats has been reported. Our observation of a shorter repeat distribution in
African-Americans is concordant with observations reported previously in studies of this
polymorphism and prostate cancer (Hardy and others 1996; Stanford and others 1997).
In the absence of prior information on how to dichotomize allele length, we selected the
median of the distribution of the AR repeat lengths as our cut-point. However, our
sensitivity analysis using three successive cut-points supported the relevance of the 21
repeat cut-point.
Our results may help explain the mechanism by which EPT use increases breast cancer
risk (Rossouw and others 2002). Our results suggest that AR genotype modifies
hormone-induced cell proliferation as reflected in percent mammographic density.
Additional results are needed to determine whether knowledge of AR genotype will be
helpful to clinicians in advising patients when making decisions on whether to use EPT.
53
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
CHAPTER 4
DATA ANALYSIS: Androgen Receptor CAG Polymorphism and Breast Cancer
Prognosis
Introduction
A recent meta-analysis of six prospective studies confirms the association between
increased serum T levels and breast cancer risk in postmenopausal women (2002). In
addition, in both pre (Secreto and others 1983a; Secreto and others 1984; Secreto and
others 1989) and postmenopausal women (Adlercreutz and others 1989; Hill and others
1985; Lipworth and others 1996; Secreto and others 1991; Secreto and others 1983b), T
levels have been observed to be significantly higher in cases post-diagnosis compared to
controls.
T binds to the AR in target tissue to mediate its effects. The AR is expressed in the
majority of breast cancers (Bayer-Gamer and Smoller 2000; Bryan and others 1984; Brys
and others 2002; Hall and others 1996; Kuenen-Boumeester and others 1992; Lea and
others 1989; Soreide and others 1992). Several studies have been conducted to examine
the effects of androgens on the growth of AR positive breast cancer cell lines. These
studies have reported both inhibitory (Ortmann and others 2002; Poulin and others 1988)
and stimulatory (Hackenberg and others 1988; Marugo and others 1992) effects. These
54
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
divergent effects have been observed to be specific to the cell line under study (Birrell
and others 1995a).
The AR is encoded by a single 90 kb gene on the X chromosome (Xql l-ql2) that
specifies an 11 kb mRNA transcript composed of 8 exons (Brown and others 1989;
Chang and others 1988; Lubahn and others 1989; Lubahn and others 1988; Tilley and
others 1989). The entire N-terminal transactivation domain is encoded by the first exon.
Within the first exon lies a polymorphic CAG repeat that encodes a polyglutamine tract
of variable length. The normal size range of these repeats is between 6 and 39 repeats
(Edwards and others 1992; Giovannucci and others 1997). Several studies have
observed an association between increasing AR-CAG repeat length and a linear decrease
in AR transactivation activity (Chamberlain and others 1994; Irvine and others 2000;
Kazemi-Esfarjani and others 1995; Tut and others 1997). Consistent with this, male
carriers of the short AR-CAG repeat length are at increased risk of prostate cancer
(Ekman and others 1999; Giovannucci and others 1997; Hakimi and others 1997; Hsing
and others 2000; Ingles and others 1997; Irvine and others 1995; Stanford and others
1997).
The CAG repeat polymorphism in AR has been examined in relation to breast cancer risk
showing that longer repeat lengths are associated with an increase in risk in BRCAl
mutation carriers (Rebbeck and others 1999), women with a first degree family history of
breast cancer (Haiman and others 2002b), postmenopausal women (Giguere and others
2001) and women of the Philippines (Liede and others 2003). One study has examined
55
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
the association between this repeat length and breast cancer prognosis showing that the
short repeat length and, therefore, more active AR is associated with higher breast cancer
tumor grade (Yu and others 2000) in a group of premenopausal and postmenopausal
breast cancer patients.
We have used the ongoing Mapping Interactive Cancer Susceptibility Loci Study, a
multi-institutional study conducted as part of the Eastern Cooperative Oncology Group
(ECOG) to assess the relationship between the AR-CAG repeat length polymorphism and
breast cancer tumor grade. The parent study recruits ECOG patients diagnosed with
breast, prostate, lung, or colon cancer who have a living sibling diagnosed with the same
cancer; our study utilizes the index breast cancer case patient from each sibling pair.
Materials and Methods
Subject Identification
We included 415 breast cancer patients participating in the multi-institutional Mapping
Interactive Cancer Susceptibility Loci Study that is coordinated at the City of Hope
Beckman Research Institute. All subjects were selected for this study if they were
diagnosed with invasive breast cancer and had a living sibling with breast cancer at study
entry. Only the index patient of each sibling pair was included in this analysis. All
women included in this analysis were Caucasian and between the ages of 35 and 80 years
at diagnosis.
56
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Data on breast cancer prognostic indicators were obtained from a combination of sources.
For 189 cases, data were collected from the hospital tumor registry where we accessioned
the subject. If the subject was not initially diagnosed at the recruiting hospital, data were
extracted from the pathology report establishing the patient’s breast cancer diagnosis (N
= 226) by a single investigator (Elizabeth Lillie) who was blinded to the AR genotype.
This diagnosing pathology report was obtained for all index cases at the time of entry
onto the study to confirm the diagnosis. All data, hoth from the tumor registry and from
pathology report extraction, were coded following the methods outlined by the American
Joint Committee of Cancer (Greene and others 2002). A subset of tumor registry reports
(10% from each hospital) was randomly selected and compared with the pathology
reports to confirm consistency across data sources.
Stage was classified as I, IIA, IIB, IIIA, IllB, IV using the TNM classification system.
Tumor size is defined as the size of invasive tumor (cm) at its maximum dimension.
Histologic grade of tumor was coded as grade 1 (well differentiated), grade 2 (moderately
differentiated), grade 3 (poorly differentiated) and grade 4 (undifferentiated). Nodal
status was coded as either positive or negative for metastasis for patients who had nodes
resected. Finally, ER and PR status were recorded as positive or negative when results
were provided on the pathology report.
AR Genotyping
All subjects consented to give 10 ml of blood for DNA analysis. Blood samples were
drawn at the recruiting institution and were sent via overnight mail to the City of Hope
57
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Beckman Research Institute. Using a standard protocol genomic DNA was extracted
from peripheral blood. The exon 1 CAG repeat of the AR was PGR amplified using a
flourescently labeled forward primer (5 ’-TCCAGAATCTGTTCCAGAGCGTGC-3 ’) and
reverse primer (5’-GCTGTGAAGGTTGCTGTTCCTCAT-3’)• All microsatellite
genotyping was performed using an ABI 377 with Genescan and Genotyper software
(PE/Applied Biosystems, Foster City, California). The number of CAG repeats was
calculated based on direct sequencing results of 8 prostate cancer cases with different
known PCR product lengths.
Statistical Analyses
All outcome variables were categorized using methods comparable to those used in the
previous study of these associations (Yu and others 2000). Grades 1-2 were compared to
grade 3. Stage was grouped into three categories I, IIA and IIB combined, and IIIA, IIIB,
and IV combined. ER, PR and nodal status were all categorized as positive or negative.
The tumor sizes of our subjects were, on average smaller (mean, 2.0 cm) than that
reported by Yu et al. (mean, 2.7 cm)(Yu and others 2000); therefore, tumor size was
dichotomized using the median (1.6 cm) of the distribution in our study population as
well as the cut-point of 2.5 cm used by Yu et al. (Yu and others 2000).
The initial analyses were designed to replicate the analyses of the association between
AR-CAG repeat length and breast cancer tumor grade reported by Yu et al. (Yu and
others 2000). The AR-CAG repeat was analyzed as a continuous variable coded in three
ways; the number of CAG repeats in the short allele, the sum of the CAG repeats in both
58
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
alleles, and the absolute value of the difference in the number of CAG repeats between
alleles. CAG repeat lengths were compared among the various prognostic indicator
variables using analysis of variance.
The length of the short allele, and the sum of the allele lengths were also dichotomized
using the cut-points of 21 and 44 repeats respectively from Yu et al. (Yu and others
2000). These cut-points were the same as the means in our study population.
Associations between the categorical CAG variables and the prognostic indicator
variables were compared using the chi-square test. For dichotomous outcomes, the odds
ratio and corresponding 95% confidence limits were estimated.
Unlike the previous study, we also assessed the length of the long allele as a continuous
variable and as a categorical variable dichotomized at the mean of the CAG repeat length
of the longer of the two alleles (24 repeats) to further evaluate the association between
AR genotype and breast cancer prognosis. In addition, due to the larger study setting, all
analyses were stratified on age of diagnosis categorized into tertiles <45, 45-55, >55.
Age categories were selected a priori to assess groups that were premenopausal and
postmenopausal with reasonable certainty. Data were analyzed using SAS v9 software
(SAS Institute Inc., Cary, North Carolina).
Results
Of the 415 breast cancer patients who were genotyped for the AR-CKG repeat, 316 had
complete data on TNM stage, 339 on tumor size, 315 on histologic grade, 343 on nodal
59
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
status, 234 on ER status and 223 on PR status. Table 4-1 shows the distributions of
patients in each prognostic indicator category stratified by source of data. Patients with
data available from the tumor registry had more tumors of less than 1.6 cm than patients
whose data were extracted by pathology reports {P = 0.01).
In all subjects, we observed no statistically significant association between advanced
stage, increased tumor size, advanced histologic grade, node positive, ER positive, or PR
positive status and the length of the short AR allele, the length of the long allele, the sum
of the AR alleles, or the difference in length between the AR alleles (Table 4-2). We
observed an association of borderline statistical significance {P = 0.06) between the
difference in length of the two alleles and tumor size when using the 2.5 cm cut-point,
showing a larger difference between the repeat lengths of smaller sized tumors.
However, this was not observed when the tumor size was categorized using the 1.6 cm
cut-point. Additional analyses were conducted adjusting for the source of data, i.e. tumor
registry vs. pathology report, but this had no effect on the AR associations.
Modeling the AR genotype categorically did not alter these findings (Table 4-3). We
found no difference in the number of subjects with a higher histologic grade when using
the cutpoint of 21 repeats in at least one allele (50% for <21 repeats vs. 51% for 21+
repeats; p=0.87) or the cutpoint of 44 for the sum oiAR repeats (53% for <44 repeats
versus 51% for 44+ repeats; P = 0.67).
60
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 4-1 Distribution of cases in each prognostic indicator category stratified by source of data
Feature
All (%)
(N = 415)
Registry (%)
(N = 189)
Path (%)
(N = 226)
Stage
1
2
3+4
Unknown
155 (37)
144 (35)
17(4)
99 (24)
97 (51)
76 (40)
10(5)
6(3)
58 (26)
68 (30)
7(3)
93(41)
Tumor size (cm)
<2.5"
>2.5
Unknown
246 (59)
93 (22)
76(18)
137 (73)
38 (20)
14(7)
109 (48)
55 (24)
62 (27)
<1.6*’
>1.6
Unknown
165 (40)
174 (42)
76(18)
95 (50)
80 (42)
14(7)
70 (31)
94 (42)
62 (27)
Grade
1-2
3+
Unknown
186 (45)
129 (31)
100(24)
90 (48)
60 (32)
39 (21)
96 (42)
69 (31)
61 (27)
Nodal status
Negative
Positive
Unknown
209 (50)
134 (32)
72(17)
116(61)
66 (35)
7(4)
93 (41)
68 (30)
65 (29)
ER
Positive
Negative
Unknown
182(44)
52(13)
181 (44)
115(61)
37 (20)
37 (20)
67 (30)
15(7)
144 (64)
PR
Positive
Negative
Unknown
145 (35)
78(19)
192(46)
89 (47)
54 (29)
46 (24)
56 (25)
2 4 (11)
146 (65)
Age
<45
45-55
>55
76(18)
134 (32)
205 (49)
32(17)
65 (34)
92 (49)
44(19)
69 (31)
113(50)
“ Cutpoint used by Yu, 2000
Median cutpoint
61
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 4-2 Mean number of androgen receptor CAG repeats by breast cancer prognostic indicator
Feature Sample size
Mean CAG
repeats in
allele 1 (SD)
Mean CAG
repeats in
allele 2 (SD)
Mean CAG
repeats in both
alleles (SD)
Mean CAG
difference
between alleles
(SD)
Stage
1
2
3-4
P
155
144
17
20.40 (2.69)
20.50(2.38)
20.65 (2.37)
0.90
23.60 (2.72)
23.40(2.48)
24.23(2.17)
0.42
44.00 (4.67)
43.90 (4.23)
44.88(3.67)
0.69
3.20(2.71)
2.90 (2.42)
3.59 (2.67)
0.43
Tumor size (cm)
<2.5"
>2.5
P
246
93
20.41 (2.50)
20.61 (2.36)
0.49
23.62 (2.60)
23.26(2.64)
0.25
44.03 (4.38)
43.87 (4.47)
0.77
3.22 (2.60)
2.65 (2.27)
0.06
< 1 .6 '’
>1.6
P
165
174
20.53 (2.55)
20.40 (2.38)
0.61
23.58 (2.65)
23.47 (2.57)
0.68
44.12(4.56)
43.86 (4.26)
0.60
3.05 (2.51)
3.07 (2.55)
0.94
Grade
1-2
3
P
186
129
20.56 (2.62)
20.37 (2.48)
0.52
23.59 (2.90)
23.64 (2.30)
0.89
44.15(4.91)
44.01 (3.77)
0.78
3.03 (2.54)
3.26 (2.94)
0.46
Nodal status
Negative
Positive
P
209
134
20.44 (2.55)
20.49 (2.51)
0.85
23.56 (2.60)
23.73 (2.62)
0.54
44.00 (4.34)
44.22 (4.44)
0.64
3.11 (2.77)
3.24 (2.58)
0.68
ER
Positive
Negative
P
182
52
20.40 (2.58)
20.42 (2.19)
0.94
23.42(2.13)
23.58 (2.65)
0.70
43.85 (3.70)
43.97(4.58)
0.86
3.18(2.52)
3.00 (2.22)
0.64
P R
Positive
Negative
P
145
78
20.39 (2.53)
20.44 (2.44)
0.89
23.52 (2.61)
23.59 (2.43)
0.84
43.90 (4.49)
44.03 (4.27)
0.84
3.13 (2.50)
3.15(2.34)
0.95
‘ Outpoint used by Yu, 2000
’ Median cutpoint
62
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 4-3 Association of androgen receptor CAG with breast cancer prognostic indicators
CAG repeats in allele 1 CAG repeats in allele 2 CAG repeats in both alleles
Feature Sample size <21 >21 >24 <24 <44 >44
Stage
1
2
3-4
P
155
144
17
45%
51%
53%
55%
49%
47%
0.58
32%
31%
53%
68%
69%
47%
0.18
54%
49%
71%
46%
51%
29%
0.21
Tumor size (cm)'
<2.5
>2.5
P
OR (95%CI)
246
93
47%
53%
53%
47%
0.33
0.79 (0.49-1.27)
35%
26%
65%
74%
0.11
0.65 (0.38-1.10)
52%
52%
48%
48%
0.89
1.03(0.64-1.67)
<1.6
>1.6
P
OR(95% CI)
165
174
52%
45%
48%
55%
0.86
0.96(0.63-1.47)
35%
30%
65%
70%
0.30
0.79(0.50-1.24)
47%
55%
53%
45%
0.85
1.04(0.68-1.59)
Grade
1-2
3
P
OR(95% CI)
186
129
51%
50%
49%
50%
0.87
1.04(0.66-1.63)
36%
33%
73%
67%
0.62
0.88(0.55-1.43)
51%
53%
49%
47%
0.67
0.91 (0.58-1.42)
Nodal status
Negative
Positive
P
OR(95% CI)
209
134
48%
49%
52%
51%
0.87
0.96(0.62-1.49)
31%
40%
69%
60%
0.09
1.48(0.94-2.33)
53%
55%
47%
45%
0.74
0.93 (0.60-1.43)
ER
Positive
Negative
P
OR(95% CI)
182
52
46%
44%
54%
56%
0.81
1.08(0.58-2.01)
36%
29%
64%
71%
0.32
0.71 (0.36-1.39)
52%
52%
48%
48%
0.97
1.01 (0.54-1.87)
PR
Positive
Negative
P
OR(95% CI)
145
78
48%
41%
52%
59%
0.30
1.34 (0.77-2.34)
37%
32%
63%
68%
0.50
0.82(0.46-1.47)
54%
47%
46%
53%
0.32
1.33 (0.76-2.30)
* Cutpoint used by Yu, 2000
’ ’ Median cutpoint
63
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
After stratification by age at diagnosis, no statistically significant results were observed
in either the <45 years or the 45-55 years age category. However, in women diagnosed
after the age of 55 years, higher histologic grade was associated with a smaller length of
the shorter AR allele (P = 0.04) (Table 4-4). Analyzed categorically, subjects with grade
3 tumor were more likely than those with grades 1 or 2 tumor to have a short allele (61%
for <21 repeats vs. 47% for 21+ repeats; P = 0.08) (Table 4-5). Among the older women,
analysis of the long allele also revealed an association between carrying a long AR allele
of >24 CAG repeats and having node positive breast cancer (OR = 2.26; P = 0.02).
64
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 4-4 Mean number of androgen receptor CAG repeats by breast cancer prognostic indicator in cases
age > 55 at diagnosis
Feature
Number of
Patients
Mean CAG
repeats in
allele 1 (SD)
Mean CAG
repeats in
allele 2 (SD)
Mean CAG
Mean CAG difference
repeats in both between alleles
alleles (SD) (SD)
Stage
1 82 20.18(2.61) 23.35 (2.56) 43.54 (4.48) 3.17(2.59)
2 66 20.62 (2.42) 23.02 (2.70) 43.64(4.66) 2.39(2.14)
3-4 8 20.00 (2.27) 24.00 (2.83) 44.00 (4.24) 4.00 (2.88)
P 0.53 0.52 0.96 0.07
Tumor size (cm)
<2.5“ 124 20.25 (2.37) 23.42 (2.50) 43.67 (4.18) 3.17(2.49)
>2.5 45 20.64 (2.56) 23.04 (3.11) 43.69 (5.16) 2.40 (2.42)
P 0.35 0.42 0.98 0.08
<1.6*’ 83 20.34 (2.26) 23.25 (2.52) 43.59(4.21) 2.92 (2.29)
>1.6 86 20.37 (2.57) 23.38 (2.82) 43.76 (4.70) 3.01 (2.67)
P 0.93 0.75 0.81 0.80
Grade
1-2 100 20.78 (2.51) 23.64 (3.03) 44.42 (5.01) 2.86 (2.43)
3 62 19.90 (2.70) 23.35 (2.52) 43.26 (4.06) 3.45 (3.28)
P 0.04 0.54 0.13 0.19
Nodal status
Negative 109 20.22 (2.56) 23.29 (2.58) 43.51 (4.27) 3.07 (2.87)
Positive 61 20.62 (2.62) 23.57 (2.89) 44.20 (4.93) 2.95 (2.47)
P 0.33 0.52 0.35 0.78
ER
Positive 93 20.40 (2.38) 23.43 (2.52) 43.83(4.28) 3.20 (2.46)
Negative 20 20.00 (2.08) 23.20 (2.19) 43.20(3.49) 3.03 (2.40)
P 0.49 0.71 0.54 0.78
PR
Positive 64 20.31 (2.36) 23.19(2.49) 43.50 (4.23) 2.88 (2.37)
Negative 46 20.33 (2.36) 23.65 (2.41) 43.98(4.16) 3.33 (2.34)
P 0.98 0.33 0.56 0.32
“ Cutpoint used by Yu, 2000
’ ’ Median cutpoint
65
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Table 4-5 Association of androgen receptor CAG with breast cancer prognostic indicators in cases age > 55
at diagnosis
Sample size CAG repeats in allele 1 CAG repeats in allele 2 CAG repeats in both alleles
Feature <21 >21 >24 <2A <44 >44
Stage
1
2
3-4
P
82
66
8
56%
53%
50%
44%
47%
50%
0.90
26%
29%
50%
74%
71%
50%
0.34
49%
62%
38%
51%
38%
62%
0.17
Tumor size (cm)
<2.5“
>2.5
P
OR (95% Cl)
124
45
56%
56%
44%
44%
0.99
0.99(0.50-1.98)
31%
27%
69%
73%
0.62
0.82(0.38-1.77)
49%
62%
51%
38%
0.13
1.70(0.85-3.42)
<1.6*’
>1.6
P
OR (95% Cl)
83
86
57%
55%
43%
45%
0.92(0.50-1.69)
28%
31%
72%
69%
1.19(0.62-2.31)
49%
56%
51%
44%
1.29(0.71-2.37)
Grade
1-2
3
P
O R(95% CI)
100
62
47%
61%
53%
39%
0.08
1.79(0.94-3.40)
37%
31%
63%
69%
0.41
0.75 (0.38-1.48)
50%
56%
50%
44%
0.42
1.30(0.69-2.45)
Nodal status
Negative
Positive
P
OR (95%CI)
109
6!
55%
52%
45%
48%
0.75
0.90(0.48-1.69)
25%
43%
75%
57%
0.02
2.26(1.16-4.40)
52%
51%
48%
49%
0.85
0.94(0.50-1.77)
ER
Positive
Negative
P
O R(95% CI)
93
20
55%
65%
45%
35%
0.41
1.53(0.56-4.18)
32%
35%
68%
65%
0.81
1.13 (0.41-3.12)
53%
50%
47%
50%
0.83
0.90 (0.34-2.36)
PR
Positive
Negative
P
OR(95% CI)
64
46
55%
61%
45%
39%
0.52
1.29(0.60-2.78)
31%
35%
69%
65%
0.70
1.17(0.52-2.62)
52%
54%
48%
46%
0.77
1.12(0.52-2.39)
“ Cutpoint used by Yu, 2000
’ ’ Median cutpoint
66
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Discussion
Our study shows no convincmg evidence of an overall association between the AR
genotype and breast cancer histologic grade; thus, we were not able to replicate the
findings by Yu et al. (Yu and others 2000) despite looking at the outcome in several
ways. However, we did observe an association among cases with an older age at
diagnosis. Yu et al. observed the association in a population of consecutive cases
between the ages of 25 and 93 years that had undergone either radical or modified
mastectomy at the Department of Gynecologic Oncology, University of Turin, Italy.
While the age range was similar to ours, our observed association was limited to cases
diagnosed after the age of 55. The analysis of this association by Yu et al. was smaller
(N = 92) and, therefore, analyses stratified by age were not reported.
We are the first to report an association between the long AR-CAG repeat and node
positive breast cancer. The long CAG repeat of the AR gene has been observed to be
associated with increased breast cancer risk in several studies (Giguere and others 2001;
Haiman and others 2002b; Liede and others 2003; Rebbeck and others 1999). Our
observation of an association between carrying a long allele and having node positive
breast cancer in women diagnosed after age 55 years is compatible with these results.
However, this was only observed when the long AR allele was modeled categorically and
may, given the large number of tests performed, be simply a type 1 error.
The age cut-points used in our study were determined a priori to assess groups that are
postmenopausal and premenopausal with reasonable certainty. Menopausal status was
67
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
not ascertained at the time of patient recruitment to the study. The limitation of our
observed significant associations to postmenopausal women is concordant with studies
showing that higher androgen levels are associated with breast cancer risk among
postmenopausal women (2002). It is also compatible with results suggesting that the
association between the AR genotype and breast cancer risk is limited to postmenopausal
women (Giguere and others 2001).
The AR is expressed in a large percentage of breast cancers (Bayer-Gamer and Smoller
2000; Bryan and others 1984; Brys and others 2002; Hall and others 1996; Kuenen-
Boumeester and others 1992; Lea and others 1989; Soreide and others 1992). In vitro
studies show that androgens can inhibit (Ortmann and others 2002; Poulin and others
1988) or stimulate (Hackenberg and others 1988; Marugo and others 1992) breast cancer
cell prohferation. Recent evidence of the role of androgen-regulated genes in breast
cancer prognosis supports our observations of a potential role of AR activity as measured
by the AR genotype. Kallikrein 15 expression is mainly under androgen regulation
through the AR and has been shown to be an independent marker of favorable prognosis
in breast cancer (Yousef and others 2002b). Kallikrein 14 is also under androgen
regulation through the AR and has been shown to be an independent marker of
unfavorable prognosis in breast cancer (Yousef and others 2002a).
The majority of cases in this study were initially diagnosed at major cancer centers that
are part of ECOG. The patient populations that these institutions serve have been
reported to have more severe prognosis. Further, the cases included in this study were
68
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
not all incident cases. Cases were selected because they had a sibling that was still alive
with breast cancer on the date of study entry. Therefore, the generalizability of these
results is limited and subject to prevalence bias. The magnitude and direction of this bias
is unclear.
Our study was also limited due to the inability to evaluate the role of possible
confounders in our results. The parent study was designed as a genetic linkage study and,
therefore, did not collect complete data on established breast cancer risk factors. The
observations reported by Yu et al. (Yu and others 2000) did not include adjustment for
any covariates.
There are limitations in the measure of the AR genotype that may subject this genetic
variable to measurement error. A recent study reports that 13% of young (27-45 years
old) breast cancer patients show preferential activation of one of the AR alleles as
measured by genotyping of peripheral blood DNA, but there is no preference for the
allele with the longer or shorter CAG repeat (Kristiansen and others 2002). In our study,
87% of cases were AR heterozygotes. We also analyzed the difference between the AR
alleles as a method to address this issue and observed no significant associations between
having a larger difference in the two alleles with either higher grade tumors or node
positive status. Therefore, assuming that preferential activation is random by outcome
category, the misclassification would be non-differential. The associations observed
may, therefore, be underestimated.
69
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
The length of the AR-CAG repeat correlates with AR activity (Chamberlain and others
1994; Irvine and others 2000; Kazemi-Esfarjani and others 1995; Tut and others 1997).
Activity decreases with increasing number of repeats. Therefore, the observed
association between the short CAG repeat and increased tumor grade, can be inferred as
showing that increased AR activity results in a poorer prognosis. However, the
association between the long CAG repeat and node positive cancer indicates the opposite.
Further study of the association between the AR genotype and the actual percent of AR
expression in the tumor tissue is warranted to further understand the role of the AR and
breast cancer prognosis. In addition, genotyping methods that consider which AR allele
is active in the tumor tissue are necessary. A broader imderstanding of this relationship
may prove useful to clinicians for assessing breast cancer prognosis and identifying
subgroups of patients that may respond better to certain therapeutic regimens.
70
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
BIBLIOGRAPHY
2002. Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis
of nine prospective studies. J Natl Cancer Inst 94(8):606-16.
Abraham GE. 1974. Ovarian and adrenal contribution to peripheral androgens during the
menstrual cycle. J Clin Endocrinol Metab 39(2):340-6.
Adashi EY. 1994. The climacteric ovary as a functional gonadotropin-driven androgen-
producing gland. Fertil Steril 62(l):20-7.
Adler AJ, Danielsen M, Robins DM. 1992. Androgen-specific gene activation via a
consensus glucocorticoid response element is determined by interaction with
nonreceptor factors. Proc Natl Acad Sci U S A 89(24): 11660-3.
Adlercreutz H, Hamalainen E, Gorbach SL, Goldin BR, Woods MN, Dwyer JT. 1989.
Diet and plasma androgens in postmenopausal vegetarian and omnivorous women
and postmenopausal women with breast cancer. Am J Clin Nutr 49(3):433-42.
Bancroft J, Cawood EH. 1996. Androgens and the menopause; a study of 40-60-year-old
women. Clin Endocrinol (Oxf) 45(5):577-87.
Bayer-Gamer IB, Smoller B. 2000. Androgen receptors: a marker to increase sensitivity
for identifying breast cancer in skin metastasis of unknown primary site. Mod
Pathol 13(2): 119-22.
Benichou J, Byrne C, Capece LA, Carroll LE, Hurt-Mullen K, Pee DY, Salane M,
Schairer C, Gail MH. 2003. Secular stability and reliability of measurements of
the percentage of dense tissue on mammograms. Cancer Detect Prev 27(4):266-
74.
Berrino F, Muti P, Micheli A, Bolelli G, Krogh V, Sciajno R, Pisani P, Panico S, Secreto
G. 1996. Semm sex hormone levels after menopause and subsequent breast
cancer. J Natl Cancer Inst 88(5):291-6.
71
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Birrell SN, Bentel JM, Hickey TE, Ricciardelli C, Weger MA, Horsfall DJ, Tilley WD.
1995a. Androgens induce divergent proliferative responses in human breast
cancer cell lines. J Steroid Biochem Mol Biol 52(5):459-67.
Birrell SN, Roder DM, Horsfall DJ, Bentel JM, Tilley WD. 1995b. Medroxyprogesterone
acetate therapy in advanced breast cancer: the predictive value of androgen
receptor expression. J Clin Oncol 13(7): 1572-7.
Boyd NF, Byng JW, Jong RA, Fishell EK, Little LE, Miller AB, Lockwood GA,
Tritchler DL, Yaffe MJ. 1995. Quantitative classification of mammographic
densities and breast cancer risk: results from the Canadian National Breast
Screening Study. J Natl Cancer Inst 87(9):670-5.
Boyd NF, Dite OS, Stone J, Gunasekara A, English DR, McCredie MR, Giles GG,
Tritchler D, Chiarelli A, Yaffe MJ and others. 2002. Heritability of
mammographic density, a risk factor for breast cancer. N Engl J Med
347(12):886-94.
Brentani MM, Franco EL, Oshima CT, Pacheco MM. 1986. Androgen, estrogen, and
progesterone receptor levels in malignant and benign breast tumors: a multivariate
analysis approach. Int J Cancer 38(5):637-42.
Brisson J, Merletti F, Sadowsky NL, Twaddle JA, Morrison AS, Cole P. 1982.
Mammographic features of the breast and breast cancer risk. Am J Epidemiol
115(3):428-37.
Brown CJ, Goss SJ, Lubahn DB, Joseph DR, Wilson EM, French FS, Willard HF. 1989.
Androgen receptor locus on the human X chromosome: regional localization to
Xqll-12 and description of a DNA polymorphism. Am J Hum Genet 44(2):264-
9.
Bryan RM, Mercer RJ, Bennett RC, Rennie GC, Lie TH, Morgan FJ. 1984. Androgen
receptors in breast cancer. Cancer 54(11):2436-40.
Brys M, Wojcik M, Romanowicz-Makowska H, Krajewska WM. 2002. Androgen
receptor status in female breast cancer: RT-PCR and Westem blot studies. J
Cancer Res Clin Oncol 128(2):85-90.
72
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Burger HG, Dudley EC, Cui J, Dennerstein L, Hopper JL. 2000. A prospective
longitudinal study of serum testosterone, dehydroepiandrosterone sulfate, and sex
hormone-binding globulin levels through the menopause transition. J Clin
Endocrinol Metab 85(8):2832-8.
Byrne C, Schairer C, Brinton LA, Wolfe J, Parekh N, Salane M, Carter C, Hoover R.
2001. Effects of mammographic density and benign breast disease on breast
cancer risk (United States). Cancer Causes Control 12(2): 103-10.
Byme C, Schairer C, Wolfe J, Parekh N, Salane M, Brinton LA, Hoover R, Haile R.
1995. Mammographic features and breast cancer risk: effects with time, age, and
menopause status. J Natl Cancer Inst 87(21): 1622-9.
Carlile T, Kopecky KJ, Thompson DJ, Whitehead JR, Gilbert FI, Jr., Present AJ, Threatt
BA, Krook P, Hadaway E. 1985. Breast cancer prediction and the Wolfe
classification of mammograms. Jama 254(8): 1050-3.
Cauley JA, Lucas FL, Kuller LH, Stone K, Browner W, Cummings SR. 1999. Elevated
serum estradiol and testosterone concentrations are associated with a high risk for
breast cancer. Study of Osteoporotic Fractures Research Group. Ann Intem Med
130(4 Pt l):270-7.
Chamberlain NL, Driver ED, Miesfeld RL. 1994. The length and location of CAG
trinucleotide repeats in the androgen receptor N-terminal domain affect
transactivation function. Nucleic Acids Res 22(15):3181-6.
Chang BL, Zheng SL, Hawkins GA, Isaacs SD, Wiley KE, Tumer A, Carpten JD,
Bleecker ER, Walsh PC, Trent JM and others. 2002. Polymorphic GGC repeats in
the androgen receptor gene are associated with hereditary and sporadic prostate
cancer risk. Hum Genet 110(2): 122-9.
Chang CS, Kokontis J, Liao ST. 1988. Structural analysis of complementary DNA and
amino acid sequences of human and rat androgen receptors. Proc Natl Acad Sci U
S A85(19):7211-5.
de Stavola BL, Gravelle IH, Wang DY, Allen DS, Bulbrook RD, Fentiman IS, Hayward
JL, Chaudary MC. 1990. Relationship of mammographic parenchymal pattems
with breast cancer risk factors and risk of breast cancer in a prospective study. Int
J Epidemiol 19(2):247-54.
73
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Dorgan JF, Longcope C, Stephenson HE, Jr., Falk RT, Miller R, Franz C, Kahle L,
Campbell WS, Tangrea JA, Schatzkin A. 1996. Relation of prediagnostic serum
estrogen and androgen levels to breast cancer risk. Cancer Epidemiol Biomarkers
Prev 5(7):533-9.
Dunn JF, Nisula BC, Rodbard D. 1981. Transport of steroid hormones; binding of 21
endogenous steroids to both testosterone-binding globulin and corticosteroid-
binding globulin in human plasma. J Clin Endocrinol Metab 53(l):58-68.
Dunning AM, McBride S, Gregory J, Durocher F, Foster NA, Healey CS, Smith N,
Pharoah PD, Luben RN, Easton DF and others. 1999. No association between
androgen or vitamin D receptor gene polymorphisms and risk of breast cancer.
Carcinogenesis 20(ll):2131-5.
Edwards A, Hammond HA, Jin L, Caskey CT, Chakraborty R. 1992. Genetic variation at
five trimeric and tetrameric tandem repeat loci in four human population groups.
Genomics 12(2):241-53.
Ekman P, Gronberg H, Matsuyama H, Kivineva M, Bergerheim US, Li C. 1999. Links
between genetic and environmental factors and prostate cancer risk. Prostate
39(4):262-8.
Garland CF, Friedlander NJ, Barrett-Cormor E, Khaw KT. 1992. Sex hormones and
postmenopausal breast cancer: a prospective study in an adult community. Am J
Epidemiol 135(11): 1220-30.
Giguere Y, Dewailly E, Brisson J, Ayotte P, Laflamme N, Demers A, Forest VI, Dodin S,
Robert J, Rousseau F. 2001. Short polyglutamine tracts in the androgen receptor
are protective against breast cancer in the general population. Cancer Res
61(15):5869-74.
Giovannucci E, Stampfer MJ, Krithivas K, Brown M, Dahl D, Brufsky A, Talcott J,
Hennekens CH, Kantoff PW. 1997. The CAG repeat within the androgen receptor
gene and its relationship to prostate cancer. Proc Natl Acad Sci U S A
94(7):3320-3.
Given HF, Radbourae R, Oag H, Merritt S, Barclay E, Hanby AM, Lamlum H, McGrath
J, Curran C, Tomlinson IP. 2000. The androgen receptor exon 1 trinucleotide
repeat does not act as a modifier of the age of presentation in breast cancer. Eur J
Cancer 36(4):533-4.
74
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Greendale GA, Reboussin BA, Sie A, Singh HR, Olson LK, Gatewood O, Bassett LW,
Wasilauskas C, Bush T, Barrett-Connor E. 1999. Effects of estrogen and
estrogen-progestin on mammographic parenchymal density. Postmenopausal
Estrogen/Progestin Interventions (PEPI) Investigators. Ann Intem Med 130(4 Pt
l):262-9.
Greendale GA, Reboussin BA, Slone S, Wasilauskas C, Pike MC, Ursin G. 2003.
Postmenopausal hormone therapy and change in mammographic density. J Natl
Cancer Inst 95(l):30-7.
Greene FL, Page DL, Fleming ID, Fritz A, Balch CM, Haller DG, Morrow M, editors.
2002. AJCC Cancer Staging Manual, Sixth Edition. New York: Springer-Verlag.
Hackenberg R, Hofmann J, Holzel F, Schulz KD. 1988. Stimulatory effects of androgen
and antiandrogen on the in vitro proliferation of human mammary carcinoma
cells. J Cancer Res Clin Oncol 114(6):593-601.
Haiman CA, Bernstein L, Berg D, Ingles SA, Salane M, Ursin G. 2002a. Genetic
determinants of mammographic density. Breast Cancer Res 4(3).
Haiman CA, Brown M, Hankinson SE, Spiegelman D, Colditz GA, Willett WC, Kantoff
PW, Hunter DJ. 2002b. The androgen receptor CAG repeat polymorphism and
risk of breast cancer in the Nurses' Health Study. Cancer Res 62(4): 1045-9.
Haiman CA, Hankinson SE, De Vivo I, Guillemette C, Ishibe N, Hunter DJ, Byme C.
2003. Polymorphisms in steroid hormone pathway genes and mammographic
density. Breast Cancer Res Treat 77(I):27-36.
Hakimi JM, Schoenberg MP, Rondinelli RH, Piantadosi S, Barrack ER. 1997. Androgen
receptor variants with short glutamine or glycine repeats may identify imique
subpopulations of men with prostate cancer. Clin Cancer Res 3(9): 1599-608.
Hall RE, Aspinall JO, Horsfall DJ, Birrell SN, Bentel JM, Sutherland RL, Tilley WD.
1996. Expression of the androgen receptor and an androgen-responsive protein,
apolipoprotein D, in human breast cancer. Br J Cancer 74(8): 1175-80.
Hankinson SE, Willett WC, Manson JE, Colditz GA, Hunter DJ, Spiegelman D, Barbieri
RL, Speizer FE. 1998. Plasma sex steroid hormone levels and risk of breast
cancer in postmenopausal women. J Natl Cancer Inst 90(17): 1292-9.
75
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Hardy DO, Scher HI, Bogenreider T, Sabbatini P, Zhang ZF, Nanus DM, Catterall JF.
1996. Androgen receptor CAG repeat lengths in prostate cancer: correlation with
age of onset. J Clin Endocrinol Metab 81(12):4400-5.
Hill P, Garbaczewski L, Kasumi F. 1985. Plasma testosterone and breast cancer. Eur J
Cancer Clin Oncol 21(10):1265-6.
Hiort O, Naber SP, Lehners A, Muletta-Feurer S, Sinnecker GH, Zollner A, Komminoth
P. 1996. The role of androgen receptor gene mutations in male breast carcinoma. J
Clin Endocrinol Metab 81(9):3404-7.
Ho KC, Marschke KB, Tan J, Power SG, Wilson EM, French FS. 1993. A complex
response element in intron 1 of the androgen-regulated 20-kDa protein gene
displays cell type-dependent androgen receptor specificity. J Biol Chem
268(36):27226-35.
Hofseth LJ, Raafat AM, Osuch JR, Pathak DR, Slomski CA, Haslam SZ. 1999. Hormone
replacement therapy with estrogen or estrogen plus medroxyprogesterone acetate
is associated with increased epithelial proliferation in the normal postmenopausal
breast. J Clin Endocrinol Metab 84(12):4559-65.
Hsing AW, Gao YT, Wu G, Wang X, Deng J, Chen YL, Sesterhenn lA, Mostofi FK,
Benichou J, Chang C. 2000. Polymorphic CAG and GGN repeat lengths in the
androgen receptor gene and prostate cancer risk: a population-based case-control
study in China. Cancer Res 60(18):5111-6.
Ingles SA, Ross RK, Yu MC, Irvine RA, La Pera G, Haile RW, Coetzee GA. 1997.
Association of prostate cancer risk with genetic polymorphisms in vitamin D
receptor and androgen receptor. J Natl Cancer Inst 89(2): 166-70.
Irvine RA, Ma H, Yu MC, Ross RK, Stallcup MR, Coetzee GA. 2000. Inhibition of
pl60-mediated coactivation with increasing androgen receptor polyglutamine
length. Hum Mol Genet 9(2):267-74.
Irvine RA, Yu MC, Ross RK, Coetzee GA. 1995. The CAG and GGC microsatellites of
the androgen receptor gene are in linkage disequilibrium in men with prostate
cancer. Cancer Res 55(9): 1937-40.
76
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Isola JJ. 1993. Immunohistochemical demonstration of androgen receptor in breast cancer
and its relationship to other prognostic factors. J Pathol 170(l):31-5.
Judd HL, Lucas WE, Yen SS. 1974. Effect of oophorectomy on circulating testosterone
and androstenedione levels in patients with endometrial cancer. Am J Obstet
Gynecol 118(6);793-8.
Kadouri L, Easton DF, Edwards S, Hubert A, Kote-Jarai Z, Glaser B, Durocher F,
Abeliovich D, Peretz T, Eeles RA. 2001. CAG and GGC repeat polymorphisms in
the androgen receptor gene and breast cancer susceptibility in BRCAl/2 carriers
and non-carriers. Br J Cancer 85(l):36-40.
Kazemi-Esfarjani P, Trifrro MA, Pinsky L. 1995. Evidence for a repressive function of
the long polyglutamine tract in the human androgen receptor: possible
pathogenetic relevance for the (CAG)n-expanded neuronopathies. Hum Mol
Genet 4(4):523-7.
Kristiansen M, Langerod A, Knudsen GP, Weber BL, Borresen-Dale AL, Orstavik KH.
2002. High frequency of skewed X inactivation in young breast cancer patients. J
Med Genet 39(l):30-3.
Krook PM. 1978. Mammographic parenchymal pattems as risk indicators for incident
cancer in a screening program: an extended analysis. AJR Am J Roentgenol
131(6):1031-5.
Kuenen-Boumeester V, Van der Kwast TH, van Putten WL, Claassen C, van Ooijen B,
Henzen-Logmans SC. 1992. Immunohistochemical determination of androgen
receptors in relation to oestrogen and progesterone receptors in female breast
cancer. Int J Cancer 52(4):581-4.
La Spada AR, Wilson EM, Lubahn DB, Harding AE, Fischbeck KH. 1991. Androgen
receptor gene mutations in X-linked spinal and bulbar muscular atrophy. Nature
352(6330):77-9.
Labrie F, Luu-The V, Lin SX, Labrie C, Simard J, Breton R, Belanger A. 1997. The key
role of 17 beta-hydroxysteroid dehydrogenases in sex steroid biology. Steroids
62(l):148-58.
77
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Langer M, Kubista E, Schemper M, Spona J. 1990. Androgen receptors, serum androgen
levels and survival of breast cancer patients. Arch Gynecol Obstet 247(4);203-9.
Laughlin GA, Barrett-Connor E, Kritz-Silverstein D, von Muhlen D. 2000.
Hysterectomy, oophorectomy, and endogenous sex hormone levels in older
women: the Rancho Bernardo Study. J Clin Endocrinol Metab 85(2):645-51.
Laya MB, Gallagher JC, Schreiman JS, Larson EB, Watson P, Weinstein L. 1995. Effect
of postmenopausal hormonal replacement therapy on mammographic density and
parenchymal pattern. Radiology 196(2):433-7.
Lea OA, Kvinnsland S, Thorsen T. 1989. Improved measurement of androgen receptors
in human breast cancer. Cancer Res 49(24 Pt l):7162-7.
Liede A, Zhang W, De Leon Matsuda ML, Tan A, Narod SA. 2003. Androgen receptor
gene polymorphism and breast cancer susceptibility in The Philippines. Cancer
Epidemiol Biomarkers Prev 12(9):848-52.
Lillie EO, Bernstein L, Ingles SA, Gauderman WJ, Rivas GE, Gagalang V, Krontiris T,
Ursin G. 2004. Polymorphism in the androgen receptor and mammographic
density in women taking and not taking estrogen and progestin therapy. Cancer
Res 64(4).T237-41.
Lillie EO, Bernstein L, Ursin G. 2003. The role of androgens and the androgen receptor
in the epidemiology of breast cancer. Breast Cancer Res 5(3):164-173.
Lipworth L, Adami HO, Trichopoulos D, Carlstrom K, Mantzoros C. 1996. Serum
steroid hormone levels, sex hormone-binding globulin, and body mass index in
the etiology of postmenopausal breast cancer. Epidemiology 7(1):96-100.
Lobaccaro JM, Lumbroso S, Belon C, Galtier-Dereure F, Bringer J, Lesimple T, Heron
JF, Pujol H, Sultan C. 1993. Male breast cancer and the androgen receptor gene.
Nat Genet 5(2): 109-10.
Longcope C, Franz C, Morello C, Baker R, Johnston CC, Jr. 1986. Steroid and
gonadotropin levels in women during the peri-menopausal years. Maturitas
8(3): 189-96.
78
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Ortmann J, Prifti S, Bohlmann MK, Rehberger-Schneider S, Strowitzki T, Rabe T. 2002.
Testosterone and 5 alpha-dihydrotestosterone inhibit in vitro growth of human
breast cancer cell lines. Gynecol Endocrinol 16(2): 113-20.
Persson I, Thurfjell E, Holmberg L. 1997, Effect of estrogen and estrogen-progestin
replacement regimens on mammographic breast parenchymal density. J Clin
Oncol 15(10):3201-7.
Platz EA, Giovannucci E, Dahl DM, Krithivas K, Hennekens CH, Brown M, Stampfer
MJ, Kantoff PW. 1998. The androgen receptor gene GGN microsatellite and
prostate cancer risk. Cancer Epidemiol Biomarkers Prev 7(5):379-84.
Poulin R, Baker D, Labrie F. 1988. Androgens inhibit basal and estrogen-induced cell
proliferation in the ZR-75-1 human breast cancer cell line. Breast Cancer Res
Treat 12(2):213-25.
Poulin R, Baker D, Poirier D, Labrie F. 1989. Androgen and glucocorticoid receptor-
mediated inhibition of cell proliferation by medroxyprogesterone acetate in ZR-
75-1 human breast cancer cells. Breast Cancer Res Treat 13(2):161-72.
Rannevik G, Jeppsson S, Johnell O, Bjerre B, Laurell-Borulf Y, Svanberg L. 1995. A
longitudinal study of the perimenopausal transition: altered profiles of steroid and
pituitary hormones, SHBG and bone mineral density. Maturitas 21(2): 103-13.
Rebbeck TR, Kantoff PW, Krithivas K, Neuhausen S, Blackwood MA, Godwin AK,
Daly MB, Narod SA, Garber JE, Lynch HT and others. 1999. Modification of
BRCA1 -associated breast cancer risk by the polymorphic androgen-receptor CAG
repeat. Am J Hum Genet 64(5): 1371-7.
Roche PJ, Hoare SA, Parker MG. 1992. A consensus DNA-binding site for the androgen
receptor. Mol Endocrinol 6(12):2229-35.
Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C, Stefanick ML,
Jackson RD, Beresford SA, Howard BV, Johnson KC and others. 2002. Risks and
benefits of estrogen plus progestin in healthy postmenopausal women: principal
results From the Women's Health Initiative randomized controlled trial. Jama
288(3):321-33.
80
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Lubahn DB, Brown TR, Simental JA, Higgs HN, Migeon CJ, Wilson EM, French FS.
1989. Sequence of the intron/exon junctions of the coding region of the human
androgen receptor gene and identification of a point mutation in a family with
complete androgen insensitivity. Proc Natl Acad Sci U S A 86(23);9534-8.
Lubahn DB, Joseph DR, Sullivan PM, Willard HF, French FS, Wilson EM. 1988.
Cloning of human androgen receptor complementary DNA and localization to the
X chromosome. Science 240(4850):327-30.
Lundstrom E, Wilczek B, von Palffy Z, Soderqvist G, von Schoultz B. 1999.
Mammographic breast density during hormone replacement therapy: differences
according to treatment. Am J Obstet Gynecol 181(2):348-52.
Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B,
Kastner P, Mark M, Chambon P and others. 1995. The nuclear receptor
superfamily: the second decade. Cell 83(6):835-9.
Marchbanks PA, McDonald JA, Wilson HG, Burnett NM, Daling JR, Bernstein L,
Malone KE, Strom BL, Norman SA, Weiss LK and others. 2002. The NICHD
Women's Contraceptive and Reproductive Experiences Study: methods and
operational results. Aim Epidemiol 12(4):213-21.
Marugo M, Bemasconi D, Miglietta L, Fazzuoli L, Ravera F, Cassulo S, Giordano G.
1992. Effects of dihydrotestosterone and hydroxyflutamide on androgen receptors
in cultured human breast cancer cells (EVSA-T). J Steroid Biochem Mol Biol
42(5):547-54.
Menin C, Banna GL, De Salvo G, Lazzarotto V, De Nicolo A, Agata S, Montagna M,
Sordi G, Nicoletto O, Chieco-Bianchi L and others. 2001. Lack of association
between androgen receptor CAG polymorphism and familial breast/ovarian
cancer. Cancer Lett 168(l):31-6.
Miller WR, Telford J, Dixon JM, Hawkins RA. 1985. Androgen receptor activity in
human breast cancer and its relationship with oestrogen and progestogen receptor
activity. Eur J Cancer Clin Oncol 21(4):539-42.
Norman AW, Litwig G. 1987. Hormones. Orlando: Academic Press.
79
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Secreto G, Fariselli G, Bandieramonte G, Recchione C, Dati V, Di Pietro S. 1983a.
Androgen excretion in women with a family history of breast cancer or with
epithelial hyperplasia or cancer of the breast. Eur J Cancer Clin Oncol 19(1):5-10.
Secreto G, Recchione C, Ballerini P, Callegari L, Cavalleri A, Attili A, Fariselli G,
Moglia D, Del Prato 1 . 1991. Accumulation of active androgens in breast cyst
fluids. Eur J Cancer 27(l):44-7.
Secreto G, Recchione C, Cavalleri A, Miraglia M, Dati V. 1983b. Circulating levels of
testosterone, 17 beta-oestradiol, luteinising hormone and prolactin in
postmenopausal breast cancer patients. Br J Cancer 47(2):269-75.
Secreto G, Recchione C, Fariselli G, Di Pietro S. 1984. High testosterone and low
progesterone circulating levels in premenopausal patients with hyperplasia and
cancer of the breast. Cancer Res 44(2); 841-4.
Secreto G, Toniolo P, Pisani P, Recchione C, Cavalleri A, Fariselli G, Totis A, Di Pietro
S, Berrino F. 1989. Androgens and breast cancer in premenopausal women.
Cancer Res 49(2):471-6.
Siiteri PK, Hammond GL, Nisker JA. 1981. Increased availability of serum estrogens in
breast cancer: a new hypothesis. In: Pike MC, Siiteri PK, Welsch CW, editors.
Hormones and Breast Cancer. New York: Cold Spring Harbor Laboratoiy. p 87-
106.
Soreide JA, Lea OA, Varhaug JE, Skarstein A, Kvinnsland S. 1992. Androgen receptors
in operable breast cancer: relation to other steroid hormone receptors, correlations
to prognostic factors and predictive value for effect of adjuvant tamoxifen
treatment. Eur J Surg Oncol 18(2):112-8.
Spicer DV, Ursin G, Parisky YR, Pearce JG, Shoupe D, Pike A, Pike MC. 1994. Changes
in mammographic densities induced by a hormonal contraceptive designed to
reduce breast cancer risk. J Natl Cancer Inst 86(6):431-6.
Spurdle AB, Dite GS, Chen X, Mayne CJ, Southey MC, Batten LE, Chy H, Trute L,
McCredie MR, Giles GG and others. 1999. Androgen receptor exon 1 CAG
repeat length and breast cancer in women before age forty years. J Natl Cancer
Inst 91(11):961-6.
81
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Stanford JL, Just JJ, Gibbs M, Wicklund KG, Neal CL, Blumenstein BA, Ostrander EA.
1997. Polymorphic repeats in the androgen receptor gene: molecular markers of
prostate cancer risk. Cancer Res 57(6): 1194-8.
Stomper PC, Van Voorhis BJ, Ravnikar VA, Meyer JE. 1990. Mammographic changes
associated with postmenopausal hormone replacement therapy: a longitudinal
study. Radiology 174(2):487-90.
Suter NM, Malone KE, Daling JR, Doody DR, Ostrander EA. 2003. Androgen Receptor
(CAG)n and (GGC)n Polymorphism and Breast Cancer Risk in a Population-
Based Case-Control Study of Young Women. Cancer Epidemiol Biomarkers Prev
12:127-135.
Thomas HY, Key TJ, Allen DS, Moore JW, Dowsett M, Fentiman IS, Wang DY. 1997.
A prospective study of endogenous serum hormone concentrations and breast
cancer risk in post-menopausal women on the island of Guernsey. Br J Cancer
76(3):401-5.
Tilley WD, Horsfall DJ, McGee MA, Henderson DW, Marshall YR. 1985. Distribution
of oestrogen and androgen receptors between the stroma and epithelium of the
guinea-pig prostate. J Steroid Biochem 22(6):713-9.
Tilley WD, Marcelli M, Wilson JD, McPhaul MJ. 1989. Characterization and expression
of a cDNA encoding the human androgen receptor. Proc Natl Acad Sci U S A
86(1):327-31.
Tut TG, Ghadessy FJ, Trifiro MA, Pinsky L, Yong EL. 1997. Long polyglutamine tracts
in the androgen receptor are associated with reduced trans-activation, impaired
sperm production, and male infertility. J Clin Endocrinol Metab 82(ll):3777-82.
Ursin G, Astrahan MA, Salane M, Parisky YR, Pearce JG, Daniels JR, Pike MC, Spicer
DY. 1998. The detection of changes in mammographic densities. Cancer
Epidemiol Biomarkers Prev 7(l):43-7.
Ursin G, Ma H, Wu AH, Bernstein L, Salane M, Parisky YR, Astrahan M, Siozon C, Pike
MC. 2003. Mammographic Density and Breast Cancer in Three Ethnic Groups.
Cancer Epidemiol Biomarkers Prev 12:332-338.
8 2
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Valyani SH, Roberts JV, Colletta AA, Iqbal MJ. 1987. Sex-steroid receptors in human
breast cancer. Incidence and interrelationships. Anticancer Res 7(6):1281-3.
van Gils CH, Hendriks JH, Otten JD, Holland R, Verbeek AL. 2000. Parity and
mammographic breast density in relation to breast cancer risk: indication of
interaction. European Journal of Cancer Prevention 9(2): 105-11.
Westberg L, Baghaei F, Rosmond R, Hellstrand M, Landen M, Jansson M, Holm G,
Bjomtorp P, Eriksson E. 2001. Polymorphisms of the androgen receptor gene and
the estrogen receptor beta gene are associated with androgen levels in women. J
Clin Endocrinol Metab 86(6):2562-8.
Whitehead J, Carlile T, Kopecky KJ, Thompson DJ, Gilbert FI, Jr., Present AJ, Threatt
BA, Krook P, Hadaway E. 1985. The relationship between Wolfe's classification
of mammograms, accepted breast cancer risk factors, and the incidence of breast
cancer. Am J Epidemiol 122(6):994-1006.
Wolfe JN. 1976. Breast pattems as an index of risk for developing breast cancer. Am J
Roentgenol 126(6): 1130-7.
Wolfe JN, Saftlas AF, Salane M. 1987. Mammographic parenchymal pattems and
quantitative evaluation of mammographic densities: a case-control study. AJR
Am J Roentgenol 148(6): 1087-92.
Wooster R, Mangion J, Eeles R, Smith S, Dowsett M, Averill D, Barrett-Lee P, Easton
DF, Ponder BA, Stratton MR. 1992. A germline mutation in the androgen
receptor gene in two brothers with breast cancer and Reifenstein syndrome. Nat
Genet 2(2): 132-4.
Wysowski DK, Comstock GW, Helsing KJ, Lau HL. 1987. Sex hormone levels in serum
in relation to the development of breast cancer. Am J Epidemiol 125(5):791-9.
Yousef GM, Borgono CA, Scorilas A, Ponzone R, Biglia N, Iskander L, Polymeris ME,
Roagna R, Sismondi P, Diamandis EP. 2002a. Quantitative analysis of human
kallikrein gene 14 expression in breast tumours indicates association with poor
prognosis. Br J Cancer 87(11): 1287-93.
83
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Yousef GM, Scorilas A, Magklara A, Memari N, Ponzone R, Sismondi P, Biglia N, Abd
Ellatif M, Diamandis EP. 2002b. The androgen-regulated gene human kallikrein
15 (KLK15) is an independent and favourable prognostic marker for breast
cancer. Br J Cancer 87(11): 1294-300.
Yu H, Bharaj B, Vassilikos EJ, Giai M, Diamandis EP. 2000. Shorter CAG repeat length
in the androgen receptor gene is associated with more aggressive forms of breast
cancer. Breast Cancer Res Treat 59(2): 153-61.
Zava DT, McGuire WL. 1977. Estrogen receptors in androgen-induced breast tumor
regression. Cancer Res 37(6): 1608-10.
Zeleniuch-Jacquotte A, Bruning PF, Bonfrer JM, Koenig KL, Shore RE, Kim MY,
Pasternack BS, Toniolo P. 1997. Relation of serum levels of testosterone and
dehydroepiandrosterone sulfate to risk of breast cancer in postmenopausal
women. Am J Epidemiol 145(11): 1030-8.
Zumoff B, Strain GW, Miller LK, Rosner W. 1995. Twenty-four-hour mean plasma
testosterone concentration declines with age in normal premenopausal women. J
Clin Endocrinol Metab 80(4): 1429-30.
84
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

BRCA1 mutations and polymorphisms in African American women with a family history of breast cancer identified through high throughput sequencing

PDF

beta3-adrenergic receptor gene Trp64Arg polymorphism and obesity-related characteristics among African American women with breast cancer: An analysis of USC HEAL Study

PDF

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

PDF

CYP17 polymorphism and risk for colorectal adenomas

PDF

Androgen receptor gene and prostate -specific antigen gene in breast cancer

PDF

A large-scale genetic association study of prostate cancer in a multi-ethnic population

PDF

Association between body mass and benign prostatic hyperplasia in Hispanics: Role of steroid 5-alpha reductase type 2 (SRD5A2) gene

PDF

Determinants of mammographic density in African-American, non-Hispanic white and Hispanic white women before and after the diagnosis with breast cancer

PDF

Genetic risk factors in breast cancer susceptibility: The multiethnic cohort

PDF

Descriptive epidemiology of thyroid cancer in Los Angeles County, 1972-1995

PDF

Dietary fats, fat metabolizing genes, and the risk of breast cancer

PDF

Assessment of fatigue as a late effect of therapy among survivors of childhood leukemia

PDF

Evaluation of the accuracy and reliability of self-reported breast, cervical, and ovarian cancer incidence in a large population-based cohort of native California twins

PDF

Dietary fiber intake and atherosclerosis progression: The Los Angeles Atherosclerosis Study

PDF

Family history, hormone replacement therapy and breast cancer risk on Hispanic and non-Hispanic women, The New Mexico Women's Health Study

PDF

Biopsychosocial factors in major depressive disorder

PDF

Do reproductive factors have the same effect on breast cancer risk in African-American and White women?

PDF

Extent, prevalence and progression of coronary calcium in four ethnic groups

PDF

A case/parental/sibling control study of Ewing's sarcoma/peripheral primitive neuroectodermal tumor (pPNET)

PDF

Association of vitamin D receptor gene polymorphisms with colorectal adenoma

Asset Metadata

Creator Lillie, Elizabeth Osth (author)

Core Title Androgens and breast cancer

School Graduate School

Degree Doctor of Philosophy

Degree Program Epidemiology

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

Tag biology, biostatistics,health sciences, public health,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Krontiris, Theodore (committee chair), Ursin, Giske (committee chair), Bernstein, Leslie (committee member), Coetzee, Gerhard (committee member), Gauderman, William James (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-518240

Unique identifier UC11340888

Identifier 3140508.pdf (filename),usctheses-c16-518240 (legacy record id)

Legacy Identifier 3140508.pdf

Dmrecord 518240

Document Type Dissertation

Rights Lillie, Elizabeth Osth

Type texts

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

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

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA