Jpn J Clin Oncol 2002;32(10)398–402
Significant Reduction in Breast Cancer Risk for Japanese Women
with Interleukin 1B –31 CT/TT Relative to CC Genotype
Lucy Sayuri Ito1, Hiroji Iwata2, Nobuyuki Hamajima3, Toshiko Saito3, Keitaro Matsuo3, Mitsuhiro Mizutani2, Takuji
Iwase2, Shigeto Miura2, Katashi Okuma4, Manami Inoue3, Kaoru Hirose3 and Kazuo Tajima3
1Department of Neurology, School of Medicine, São Paulo University, São Paulo, Brazil and 2Department of Breast
Surgery, 3Division of Epidemiology and Prevention and 4Department of Clinical Laboratories, Aichi Cancer Center,
Nagoya, Japan
Received May 7, 2002; accepted June 12, 2002
Objective: The present case-control study aimed to examine the associations between breast
cancer risk and three functional polymorphisms (Interleukin (IL) –1A C-889T, IL-1B C-31T and
IL-1RN 86-bp variable number tandem repeat) related to expression of IL-1b, which combines
estrogen receptor.
Methods: Cases were 231 patients with breast cancer who had been diagnosed 1 month to 6
years before their enrollment in 1999–2000 at Aichi Cancer Center Hospital. Controls were 186
non-cancer outpatients recruited during the same period at the digestive tract, breast surgery
and gynecology clinics.
Results: There were no differences in the genotype distributions of the IL-1A and IL-1RN polymorphisms, but individuals harboring a IL-1B C-31T T allele (high expression allele) were less
frequent among cases (74.3%) than among controls (84.9%). The age-adjusted odds ratio
(OR) relative to CC genotype was 0.52 (95% confidence interval, 0.30–0.88) for CT genotype,
0.58 (0.32–1.02) for TT genotype and 0.54 (0.33–0.90) for CT/TT genotype. Subgroup analysis
showed that the preventive effect was significantly stronger for postmenopausal women than
for premenopausal women (interaction 0.30, 0.11–0.84).
Conclusions: Although this is the first report on the association between breast cancer risk
and IL-1B C-31T, the observed association seems plausible in a biological sense.
Key words: breast cancer – interleukin 1B – polymorphisms – PCR-CTPP
INTRODUCTION
The interleukin (IL)-1 family includes three members, IL-1a,
IL-1b and IL-1 receptor antagonist (IL-1Ra), which play
important roles in inflammatory and immunological responses
(1). IL-1a is a membrane-bound IL-1, whose precursor, proIL1a, is fully active in the cytosol. Some investigators consider
that the intracellular form regulates normal cellular differentiation, particularly in epithelial cells (1). It is encoded by IL-1A,
which is mapped to chromosome 2q13, along with the other
two member genes. IL-1b is a molecule that triggers inflammation, by binding to IL-1 receptor type I. It is encoded by IL-1B
with several promoter elements, including a TATA box, a
typical motif of inducible genes. IL-1Ra is an antagonist,
For reprints and all correspondence: Nobuyuki Hamajima, Division of
Epidemiology and Prevention, Aichi Cancer Center Research Institute, 1–1
Kanokoden, Chikusa-ku, Nagoya 464-8681, Japan.
E-mail: [email protected]
which combines with IL-1 receptors (IL-1RI and IL-1RII)
competitively. Its encoding gene is called IL-1RN.
Recently, IL-1a was reported to contribute to metastasis of
breast cancer cells by enhancing their expression of IL-6 and
IL-8 and to induce in fibroblasts these two interleukins and
matrix metalloproteinase 3 (2). In breast cancer cells, IL-1b
was found to combine with the estrogen receptor (ER) a,
resulting in transcriptional activation (3). The concentration of
IL-1b is reported to be higher in invasive breast cancer tissue
than in benign lesions (4). In breast cancer tissue homogenates,
IL-1a concentration correlates inversely with ER, a wellknown prognostic factor for breast cancer, while the IL-1Ra
concentration correlates with both ER and IL-1b. However,
there does not appear to be any difference in IL-1a and IL-1b
levels between ER-negative and ER-positive breast cancer
tissues (5). Some of the above findings suggest that interindividual variations in IL-1 production may impact on the
prognosis with breast cancers. Although prognostic factors are
not necessarily risk factors, the functional polymorphisms of
IL-1s may influence breast cancer risk.
© 2002 Foundation for Promotion of Cancer Research
Jpn J Clin Oncol 2002;32(10)
399
Table 1. Characteristics of cases and controls
Characteristic
Cases
n = 231
Age at diagnosis for cases or at interview for controls
Age at first birth (years)
Breast cancer family history‡
%
9.5
19
10.2
40–49
87
37.7
40
21.5
50–59
76
32.9
70
37.6
46
19.9
57
30.6
109
47.4
93
50.0
14–15
94
40.9
65
34.9
16³
27
11.7
28
15.1
<13
<24
No birth
BMI† (kg/m2)
n = 186
22
25–38
Menopause
%
24–39
60–70
Age at menarche (years)*
Controls
94
40.7
67
36
109
47.2
98
52.7
28
12.1
21
11.3
Premenopausal
128
55.4
71
38.2
Postmenopausal
103
44.6
115
61.8
<20
51
22.1
46
24.7
20–22
70
30.3
53
28.5
22–24
58
25.1
40
21.5
³24
52
22.5
47
25.3
No
202
87.4
173
93.0
Yes
29
12.6
13
7
*Unknown for one case. †BMI, body mass index. ‡Breast cancer history of the mother and/or sister (s).
Several polymorphisms have been reported for the three IL1 members (6,7). Among those, IL-1A C-889T, IL-1B C-31T
and IL-1RN 86-bp variable number tandem repeat (VNTR) at
intron 2 are reported to be functional (8,9). The IL-1A polymorphism has been reported to have an association with
Alzheimer’s disease (10). Associations with IL-1B polymorphisms have been reported for stomach cancer risk (7),
Helicobacter pylori infection (11), periodontitis (12) and
inflammatory bowel diseases (13). The inflammatory bowel
diseases were also associated with the IL-1RN polymorphism
(13–15). The present case-control study was conducted to
investigate possible associations with the IL-1A C-889T, IL-1B
C-31T and IL-1RN 86-bp VNTR polymorphisms among
Japanese women, using the same approach as in previous
reports (16–18).
PATIENTS AND METHODS
patients to participate in this study. Between March 1999 and
April 2000, 243 breast cancer patients were enrolled. Twelve
cases diagnosed 6 years before were excluded and the remaining 231 cases were used for analysis. Controls were 186
women without a history of cancer; 123 outpatients mainly of
the gastroenterology clinic and 63 outpatients mainly of the
breast surgery or gynecology clinics, aged 24–69 years. It was
not rare that outpatients visited two or more clinics, because of
anxiety about whether they might be suffering from cancers. In
our hospital, approximately 70% of first-visit non-cancer
patients were found to be disease-free, presenting for the purpose of an annual checkup or further examination after a positive result at a cancer screening facility (19).
Those who signed an informed consent form were asked to
complete a self-administered questionnaire and to provide a
7 ml peripheral blood sample. This study was approved by the
Ethical Committee at Aichi Cancer Center in 1999 (Ethical
Committee Approval Numbers 12-20 and 12-23).
STUDY SUBJECTS
GENOTYPING
Cases were female breast cancer patients aged 26–70 years
histologically diagnosed at Aichi Cancer Center Hospital.
Controls were female patients without cancer who visited
outpatient clinics at the same hospital. This study was openly
announced to female outpatients at the reception desk for firstvisit patients and in waiting rooms of breast surgery clinic. In
addition, doctors at the breast surgery clinic invited eligible
DNA was extracted from 200 ml of buffy coat preserved at
–40°C with a QIAamp DNA Blood Mini Kit (Qiagen,
Valencia, CA). PCR-CTPP (polymerase chain reaction with
confronting two-pair primers) was applied for IL-1A C-889T
and IL-1B C-31T polymorphisms as described in our previous
papers (11,20) and IL-1RN genotyping was conducted by the
same method as described by Mansfield et al. (14).
400
Breast cancer and IL-1B polymorphism
Table 2. Genotype frequencies for IL-1A, IL-1B and IL-1RN polymorphisms
Genotype
Cases
Controls
c2 test
n = 231
%
n = 186
%
CC
194
85.1
152
82.2
c2 = 1.70
CT
34
14.9
32
17.3
d.f. = 2
TT
0
0.0
1
Not genotyped
3
IL-1A C-889T
–
1
0.5
P = 0.427
–
IL-1B C-31T
CC
58
25.7
28
15.1
c2 = 6.85
CT
103
44.7
99
53.5
d.f. = 2
TT
66
29.5
58
31.4
P = 0.033
Not genotyped
4
–
1
–
IL-1RN 86-bp VNTR
22
2
0.9
0
0.0
c2 = 5.04
24
13
5.7
14
7.6
d.f. = 5
25
0
0.0
1
0.5
P = 0.411
34
2
0.8
0
0.0
44
209
92.1
169
91.4
45
1
0.4
1
0.5
Not genotyped
4
–
1
–
STATISTICAL ANALYSIS
An unconditional logistic model was applied for estimating
odds ratios (ORs) and 95% confidence intervals (CIs), as well
as interaction terms, with the computer program STATA Version 7 (STATA Corp., College Station, TX). All the ORs were
adjusted for age by the logistic model including a continuous
variable of age at diagnosis for cases and age at enrollment for
controls. Women less than 50 years of age with hysterectomy
before menopause were classified as being premenopausal.
Those younger than 50 years whose menstruation stopped after
chemotherapy were also classified into the premenopausal
group.
RESULTS
Table 1 summarizes data for 231 cases and 186 controls. The
controls were slightly older, with a mean age of 52.9 as compared with 50.4 years for the cases. There were no significant
differences in age at menarche, age at first birth and body mass
index (BMI, kg/m2) between cases and controls. The number
of postmenopausal subjects was greater in controls than in
cases, which reflected the age distribution of non-cancer
patients of our hospital. The cases had a higher percentage for
mother and/or sister family history of breast cancer than the
controls (P = 0.072 by Fisher’s exact test).
Genotype frequencies for IL-1A C-889T, IL-1B C-31T and
IL-1RN 86-bp VNTR are shown in Table 2. No significant differences in genotype frequency between the cases and controls
were observed for IL-1A and IL-1RN polymorphisms. However, the frequency of CC genotype of IL-1B C-31T was 25.7%
for cases and 15.1% for controls; the difference was statistically significant (P < 0.05). Further analysis was focused on
the IL-1B C-31T polymorphism.
The age-adjusted ORs estimated by a logistic model are
shown in Table 3. The estimates relative to the CC genotype
were 0.52 (95% CI, 0.30–0.88) for the CT genotype and 0.58
(0.32–1.02) for the TT genotype. When CT and TT were
combined, the OR was 0.54 (0.33–0.90). They did not differ
substantially according to the interval between diagnosis and
enrollment. No significant differences in the OR were found
with reference to age at menarche, age at first childbirth and
BMI, although the difference in the point estimate was substantial for the age at menarche and BMI.
With regard to menopause status, the difference was significant; the interaction term between menopausal status (0 for
premenopausal and 1 for postmenopausal) and genotype (0 for
CC and 1 for CT/TT) was 0.31 (0.11–0.84). The OR among
premenopausal women was 0.91 (0.43–1.96) for CT genotype,
0.74 (0.32–1.75) for TT genotype and 0.85 (0.42–1.75) for
CT/TT genotype, whereas that for postmenopausal women
was 0.24 (0.11–0.53), 0.37 (0.16–0.85) and 0.29 (0.14–0.61),
respectively. There was no difference in the OR between the
postmenopausal women with a BMI < 22 and those with a
BMI ³ 22; OR for CT/TT was 0.27 (0.09–0.80) and 0.27
(0.09–0.79), respectively.
As shown in Table 3, none of 12 controls with breast cancer
history of mother and/or sister(s) had CC genotype, whereas
nine out of 29 cases had CC genotype (P = 0.039 by Fisher’s
exact test).
DISCUSSION
The present study showed that women with CT/TT genotype of
IL-1B C-31T had a reduced risk of breast cancer. Subgroup
analysis revealed that the preventive effect was marked for
postmenopausal women. The observed association needs to be
discussed in terms of (1) studies with prevalent cases, (2) possible bias for genotype frequency, (3) statistical power and (4)
possible biological explanation.
Prognostic factors influence ORs derived from prevalent
case-control studies seriously for cancers with a poor outcome,
such as pancreatic cancer, for which the T allele at 3954 of IL1B was reportedly associated with poor prognosis (21). However, the impact may be limited for cancers with a relatively
good prognosis, such as breast cancer (22). In this study, the
OR for survivors 3 years or longer after diagnosis was not different from that for patients diagnosed in the past 2 years, indicating that no adjustment was required for prognostic effects.
Since recall bias is not relevant for genotyping, there seem no
logical problems in using case-control studies as long as the
OR for the long survivors is the same for the cases enrolled
more recently (22). Even in case-control studies with incident
cases using a cancer registry, it is not rare that it takes 6 months
Jpn J Clin Oncol 2002;32(10)
401
Table 3. Age-adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) for IL1B C-31T genotype
Subjects
Cases
Controls
OR (95% CI)
CC
CT
TT
CT/TT
All subjects
227
185
1
0.52 (0.30–0.88)
0.58 (0.32–1.02)
0.54 (0.33–0.90)
<3 years
111
185
1
0.46 (0.25–0.87)
0.52 (0.26–1.02)
0.48 (0.27–0.87)
³3 years
116
185
1
0.57 (0.30–1.06)
0.66 (0.34–1.29)
0.60 (0.33–1.08)
<14
106
92
1
0.82 (0.38–1.77)
0.85 (0.37–1.95)
0.83 (0.40–1.72)
³14
120
93
1
0.31 (0.14–0.68)
0.39 (0.17–0.91)
0.34 (0.16–0.71)
94
67
1
0.42 (0.17–1.06)
0.63 (0.24–1.69)
0.49 (0.21–1.18)
133
118
1
0.55 (0.28–1.07)
0.50 (0.24–1.04)
0.53 (0.28–1.00)
Interval after diagnosis
Age at menarche (years)*
Age at first birth (years)
<25
³25 or no birth
Menopause
Premenopausal
126
71
1
0.91 (0.43–1.96)
0.74 (0.32–1.75)
0.85 (0.42–1.75)
Postmenopausal
101
114
1
0.24 (0.11–0.53)
0.37 (0.16–0.85)
0.29 (0.14–0.61)
<22
119
99
1
0.58 (0.29–1.17)
1.10 (0.50–2.41)
0.73 (0.38–1.42)
³22
108
86
1
0.41 (0.18–0.97)
0.28 (0.11–0.68)
0.35 (0.16–0.80)
1
0.61 (0.35–1.06)
0.61 (0.34–1.11)
0.61 (0.36–1.03)
9/0†
10/9†
10/3†
20/12†
Body mass index
(kg/m2)
Family history of breast cancer (mother or sisters)
No
198
173
Yes
29
12
*Unknown for one case. †Cases/controls.
or so after diagnosis for research staff to contact them, with the
result that some patients die before the first contact.
The IL-1B C-31T polymorphism was genotyped by the same
method for patients with esophageal or colorectal cancer in our
laboratory. The CC genotype was found in 17.2% (16/93) and
17.3% (24/139), respectively (unpublished data). The percentages were similar to those in the present controls. Only breast
cancer series showed a higher percentage. In addition, the genotype frequency for controls was calculated according to the
clinic that they visited; 14.6% for 123 outpatients of the digestive tract clinic and 16.1% for the remaining 62 outpatients,
mainly from breast surgery and gynecology clinics. It seems
unlikely that the genotype frequency for the controls was
biased. There was no way for patients or research staff to have
access to genotype information before genotyping. Most of the
breast cancer cases were invited to participate in the study by
doctors in charge. Only six patients with breast cancer refused
to provide blood because of their anxiety about genotyping.
Generally, associations with rare alleles are difficult to
examine in case-control studies for a moderately elevated risk.
The present study was conducted within a project to detect
genotypes related to breast cancer risk, so that the population
was not large. However, the statistical power to detect OR = 2
for 231 cases and 186 controls under P < 0.05 (two-sided test)
is 0.75 when the high risk genotype is 15% among controls,
e.g. for CC genotype of IL-1B C-31T.
Among Japanese, the T allele of IL-1A C-889T is present in
8.5% [n = 241 (11)] and alleles other than four repeats of IL1RN in 5.4% [n = 241 (1)]. Both alleles are less frequent than
in Caucasians; for the IL-1A T allele, 33.1% [n = 400 in Finland (8)] and 28.3% [n = 242 in England (14)] and for the nonfour-repeat alleles of IL-1RN, 27.9% [n = 429 in Poland (7)],
29.9% [n = 400 in Finland (8)], 26.6% [n = 261 in England
(14)] and 27% [n = 61 in South Africa (13)]. The T allele of IL1B C-31T had a penetration of 55.0% [n = 241 (11)], less frequent than in Caucasians [70.2%, n = 429 in Poland (7)]. These
genotype frequencies indicate that the effects of IL-1B C-31T
polymorphism can be examined more efficiently among Japanese than among Caucasians. A smaller proportion of the IL1A and IL-1RN influential alleles enhances the evaluation of
IL-1B C-31T with respect to disease risk among Japanese.
As mentioned in the Introduction, the biological findings
suggest that high levels of IL-1b predispose to a poor prognosis and possibly to an increase in breast cancer risk. Since the T
allele of C-31T is thought to be the higher expression allele, if
it is correct, the protective effect observed here is paradoxical.
Although biological evidence is scarce, there are two possible
interpretations. One is that the T allele is not a high expression
allele, but linked to a low expression allele. The T allele of
C-31T is linked nearly completely to the C allele of IL-1B
C-511T (7,11). It has been reported that individuals with IL1A –889T and IL-1B –511T alleles have a higher serum IL-1b
402
Breast cancer and IL-1B polymorphism
level on average (8). This means that IL-1B –31C is a higher
expression allele among those with IL-1A –889T, who are
rare in Japanese. An alternative interpretation is that the
–31T is truly a higher expression allele and plays a role in
reducing breast cancer risk. An electrophoretic mobility shift
assay indicated that the –31T allele was highly expressed by
lipopolysaccharide (7). Since IL-1b is a multifunctional cytokine, the higher expression may induce protective mechanisms
against breast cancer cells, as a whole. Nearly every cell type
is affected by IL-1b, acting in concert with other cytokines or
small mediator molecules (1). Even though invasive breast
cancer tissues contain higher levels of IL-1b than benign
lesions (4) and IL-1b was demonstrated to enhance gene expression with an estrogen responsive element (3), the interleukin
might nevertheless be protective against breast cancer carcinogenesis. The observed protective effect is more marked for
postmenopause, a condition with a lower estrogen level which
could become more sensitive to IL-1b, may provide clues to
elucidate the biological mechanisms.
As noted above, the present study was conducted as part of a
project to detect polymorphisms potentially associated with
breast cancer risk (16–18). Some significant association may
be found by chance, while others could actually be related to
breast cancer risk. In this study, the results obtained were
opposite to the expectation we had had before the genotyping.
Since this is the first report on the association between the IL1B polymorphism and breast cancer risk, other studies are
required to confirm and extend our findings.
Acknowledgments
The authors are grateful to Ms Hiroko Fujikura, Ms Keiko
Asai, Ms Michiyo Tani, Ms Naomi Takeuchi and Ms Mayumi
Kato for their technical assistance. This work was supported in
part by a Grant-in-Aid for Scientific Research (Grant No.
12670383) from the Ministry of Education, Science, Sports,
Culture and Technology of Japan.
References
1. Dianarello CA. Biologic basis for interleukin-1 in disease. Blood
1996;87:2095–147.
2. Nozaki S, Sledge GW Jr, Nakshatris H. Cancer cell-derived interleukin 1a
contributes to autocrine and paracrine induction of pro-metastatic genes in
breast cancer. Biochem Biophys Res Commun 2000;275:60–2.
3. Speirs V, Kerin MJ, Newton CJ, Walton DS, Green AR, Desai SB, et al.
Evidence for transcriptional activation of ERa by IL-1b in breast cancer
cells. Int J Oncol 1999;15:1251–4.
4. Jin L, Yuan RQ, Fuchs A, Yao Y, Joseph A, Schwall R, et al. Expression
of interleukin-1B in human breast carcinoma. Cancer 1997;80:421–33.
5. Miller LJ, Kurtzman SH, Anderson K, Wang Y, Stankus M, Renna M, et
al. Interleukin-1 family expression in human breast cancer: interleukin-1
receptor antagonist. Cancer Invest 2000;18:293–302.
6. Bailly S, di Giovine FS, Blakemore AIF, Duff GW. Genetic polymorphism of human interleukin-1a. Eur J Immunol 1993;23:1240–5.
7. El-Omar EM, Carrington M, Chow W-H, McColl KEL, Bream JH, Young
HA, et al. Interleukin-1 polymorphisms associated with increased risk of
gastric cancer. Nature 2000;404:398–402. Corrections. 2001;412:99.
8. Hulkkonen J, Laippala P, Hurme MA. Rare allele combination of the
interluekin-1 gene complex is associated with high interleukin-1b plasma
levels in healthy individuals. Eur Cytokine Netw 2000;11:251–5.
9. Santtila S, Savinainen K, Hurme M. Presence of the IL-1RA allele 2
(IL1RN*2) is associated with enhanced IL-1 beta production in vitro.
Scand J Immunol 1998;47:195–8.
10. Du Y, Dodel RC, Eastwood BJ, Bales KR, Gao F, Lohmuller F, et al.
Association of an interleukin 1a polymorphism with Alzheimer’s disease.
Neurology 2000;55:480–3.
11. Hamajima N, Matsuo K, Saito T, Tajima K, Okuma K, Yamao K, et al.
Interleukin 1 polymorphisms, lifestyle factors and Helicobacter pylori
infection. Jpn J Cancer Res 2001;92:383–9.
12. Parkhill JM, Hennig BJ, Chapple IL, Heasman PA, Taylor JJ. Association
of interleukin-1 gene polymorphisms with early-onset periodontitis. J Clin
Periodontol 2000;27:682–9.
13. Mwantembe O, Gaillard MC, Barkhuizen M, Pillay V, Berry SD, Dewar
JB, et al. Ethnic differences in allelic associations of the interleukin-1 gene
cluster in South African patients with inflammatory bowel disease (IBD)
and in control individuals. Immunogenetics 2001;52:249–54.
14. Mansfield JC, Holden H, Tarlow JK, DiGiovine FS, McDowell TJ,
Wilson AG, et al. Novel genetic association between ulcerative colitis
and the anti-inflammatory cytokine interleukin-1 receptor antagonist.
Gastroenterology 1994;106:637–42.
15. Tountas NA, Casini-Raggi V, Yang H, DiGiovine FS, Vecchi M, Kam L,
et al. Functional and ethnic association of allele 2 of the interleukin-1
receptor antagonist gene in ulcerative colitis. Gastroenterology
1999;117:806–13.
16. Hamajima N, Iwata H, Obata Y, Matsuo K, Mizutani M, Iwase T, et al.
No association of the 5¢ promoter region polymorphism of CYP17 with
breast cancer risk in Japan. Jpn J Cancer Res 2000;91:880–5.
17. Hamajima N, Matsuo K, Tajima K, Mizutani M, Iwata H, Iwase T, et al.
Catechol-O-methyltransferase and breast cancer in Japan. Int J Clin Oncol
2001;6:13–8.
18. Huang X-E, Hamajima N, Saito T, Matsuo K, Mizutani M, Iwata H, et al.
Possible association of b2 and b3 adrenergic receptor gene polymorphisms with susceptibility to breast cancer. Breast Cancer Res
2001;3:264–9.
19. Hamajima N, Hirose K, Inoue M, Takezaki T, Kuroishi T, Tajima K. Agespecific risk factors of breast cancer estimated by a case-control study in
Japan. J Epidemiol 1995;5:99–105.
20. Hamajima N, Saito T, Matsuo K, Kozaki K, Takahashi T, Tajima K.
Polymerase reaction with confronting two-pair primers for polymorphism
genotyping. Jpn J Cancer Res 2000;91:865–8.
21. Barber MD, Powell JJ, Lynch SF, Fearon KCH, Ross JA. A polymorphism of the interleukin-1b gene influences survival in pancreatic cancer.
Br J Cancer 2000;83:1443–7.
22. Hamajima N, Matsuo K, Yuasa H. Adjustment of prognostic effects in
prevalent case-control studies on genotype. J Epidemiol 2001;11:204–10.
Erratum in: 2001;11:288.