Gynecologic Oncology 93 (2004) 499 – 505

www.elsevier.com/locate/ygyno

p53 and p21 genetic polymorphisms and susceptibility to

$
endometrial cancer
Ju Won Roh, a Jae Weon Kim, b,c,d,* Noh Hyun Park, b,c Yong Sang Song, b,c In Ae Park, e
Sang-Yoon Park, a Soon Beom Kang, b,c and Hyo Pyo Lee b,c
a
Research Institute and Hospital, National Cancer Center, Goyang, Gyeonggi, South Korea
b
Department of Obstetrics and Gynecology, Seoul National University, Seoul, 110-744, South Korea
c
Cancer Research Institute, Seoul National University, Seoul, South Korea
d
Human Genome Research Institute, Seoul National University, Seoul, South Korea
e
Department of Pathology, Seoul National University, Seoul, South Korea

Received 11 November 2003

Abstract

Objective. Recently, there has been considerable interest in the association of specific cancers with single nucleotide polymorphisms
(SNPs). In this regard, genetic polymorphism at codon 72 (CCC/proline to CGC/arginine [Pro72Arg]) of the p53 gene is one of the most
frequently studied subjects. An association between endometrial cancer and the polymorphism at codon 31 (AGC/serine to AGA/arginine
[Ser31Arg]) of the p21 gene, which is known to be a downstream mediator of p53, has also been reported.
Methods. The authors designed a hospital-based case-control study of 95 endometrial cancer patients and 285 non-cancer controls. For the
determination of p53 and p21 polymorphism, allele-specific polymerase chain reaction (PCR) and PCR-restriction fragment length
polymorphism assay was applied, respectively.
Results. We found statistically significant differences in the frequency of the p53 and p21 genotypes between these two groups ( P <
0.001), respectively. The p53 genotypes containing the Pro allele were significantly associated with endometrial cancer with an odds ratio
(OR) of 3.56 (95% confidence interval [CI] 2.10 – 6.04). Also, homozygous carriers of the p21 Ser allele showed a substantially increased
risk of developing endometrial cancer (OR 2.68, 95% CI 1.59 – 4.51) as compared to homozygous and heterozygous carriers of the Arg allele.
In addition, the combination of the pro allele containing genotypes of p53 and the Ser homozygous genotype of p21 posed a remarkably
increased risk (OR 9.55, 95% CI 4.30 – 21.24) of endometrial cancer development. These significant differences were maintained throughout
the groups after they were stratified by menopausal status.
Conclusions. These data suggest that there is a significant association between the genetic polymorphisms of p53, p21, and specific
combinations of the at-risk genotypes of these genes and the risk of developing endometrial cancer in Korean women.
D 2004 Elsevier Inc. All rights reserved.

Keywords: p21; p53; Endometrial cancer; Polymorphism

Introduction cancer can be traced to reproductive events, which influ-

ence the lifetime levels of estrogen. A large proportion of
Endometrial cancer is the most common female tract endometrial cancer cases cannot, however, be explained
malignancy in many developed countries [1], and its purely by the hormonal risk factors. The identification of
prevalence has also been increasing in Korea during the susceptibility factors, that is, genetic factors that predispose
last decade [2]. The major risk factor for endometrial individuals to endometrial cancer would provide further
insight into the etiology of this malignancy. Recently, there
$
Presented as a poster at the 33rd Annual Meeting of the Society of has been a great deal of interest in the association of
Gynecologic Oncologists in Miami, FL. specific diseases with single nucleotide polymorphisms
* Corresponding author. Department of Obstetrics and Gynecology,
Cancer Research Institute, Seoul National University, 28 Yungun-Dong,
(SNPs), which might explain the individual differences
Chongno-Ku, Seoul, 110-744, South Korea. Fax: +82-2-762-3599. of susceptibility to specific malignant neoplasms. Because
E-mail address: [email protected] (J.W. Kim). of their higher frequencies in the general population, SNPs

0090-8258/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.ygyno.2004.02.005
500 J.W. Roh et al. / Gynecologic Oncology 93 (2004) 499–505

may represent a greater overall risk factor as compared Korea, relatively little attention has been given to the
with mutations [3]. Candidate SNPs are involved in genetic susceptibility factors associated with this disease,
carcinogen metabolism, DNA damage repair, steroid me- as was mentioned above. In this article, we attempted to
tabolism, and steroid receptor activation pathways. Poly- identify p53 and p21 alleles associated with endometrial
morphisms in p53 and p21, which constitute a major cancer risk in Korean women.
downstream component of the p53 tumor suppressor
pathway, are considered to be candidate risk factors,
because of the crucial role played by these genes in the Materials and methods
maintenance of genomic integrity following genotoxic
insult [4,5]. The reported data which indicate that muta- Study population
tions in the p53 gene, whose incidence ranges from 4.8%
to 20%, are relatively infrequent in endometrial cancer Between March 1993 and April 2001, 95 consecutive
compared to other malignancies, supports the possible as- patients with a new diagnosis of endometrial cancer, who
sociation of these polymorphisms with endometrial cancer were managed at the Department of Obstetrics and Gy-
[6 –8]. necology in Seoul National University Hospital, were re-
Codon 72 of exon 4 was the first polymorphism cruited in this study. Clinicopathologic parameters such as
detected in p53 and it was suggested that the two alleles age, age at menarche, age at the first pregnancy, body
of codon 72 might have different oncogenic properties [9]. mass index, a history of diabetes mellitus and/or hyper-
In the field of breast cancer, the studies reported so far on tension, International Federation of Gynecology and Ob-
the polymorphism of p53 codon 72 have provided dis- stetrics (FIGO) stage, histologic grade, the presence of
crepant results [10 – 12]. Although no association between lymph node metastases, positive cytology, estrogen/proges-
the genetic polymorphisms in codon 72 of the p53 gene terone receptor positivity, and myometrial invasion were
and the risk of endometrial cancer has been found in obtained by reviewing the medical records. According to
previous studies [13,14], the significance of SNPs associ- the menopausal status, 52 cases out of a total of 95 were
ated with an increased cancer risk could be dramatically premenopausal women and the other 43 were postmeno-
different between ethnic groups [15]. Allelic differences of pausal at the time of sampling. The stage distribution of
p53 polymorphisms were observed in various ethnic the 95 cases, according to the FIGO staging system were
groups [16]. We have already published the data that the stage Ia (31 cases), stage Ib (34 cases), stage Ic (11 cases),
allele frequencies of codon 72 of the p53 gene differ stage IIa (5 cases), stage IIb (2 cases), stage IIIa (4 cases),
between Caucasian and Korean women [17]. We found stage IIIb (no cases), stage IIIc (4 cases), stage IVa (no
that p53 polymorphism was associated with the develop- cases), and stage IVb (4 cases). Histologically, 89 of the
ment of adenocarcinoma of cervix (unpublished data), 95 cases had endometrioid type carcinomas, whereas four
although the association was not found in squamous cell cases were adenosquamous carcinomas, one adenoacan-
carcinoma of cervix [17]. The authors considered the thoma, and one clear cell carcinoma. The non-cancer
possibility that p53 polymorphism might be associated with controls were composed of healthy Korean women visiting
the risk of endometrial cancer, another adenocarcinoma of our hospital to participate in a routine cancer detection
female genital tract. program for gynecologic cancer, who had no history of
p21, which is a downstream mediator of p53, functions endometrial disease and for whom there was no present
as a universal inhibitor of cyclin-dependent kinases, also evidence of endometrial pathology (n = 285). Women with
known as WAF1 or CIP1. Genetic changes in p21 are any malignant disease or other systemic problems such as
potentially important in human malignancies because they chronic liver diseases were also excluded from the control
could have an impact on the control of the cell cycle [18]. A groups. The case and control subjects were all born in
polymorphism at codon 31 of p21 resulting in a change Korea, meaning that they shared the same geographic
from serine to arginine has been reported [19]. Until now, origin and culture. The samples used for genotype analysis
only a few reports have been published on the association were obtained from peripheral nucleated cells, which were
between p21 polymorphism and the risk of developing separated from whole blood and were then immediately
gynecologic cancers, such as endometrial, cervical, and stored in liquid nitrogen freezers until DNA extraction was
ovarian cancers [20 – 22]. We have already found that the performed. The institutional review board of Seoul Na-
codon 31 Serine homozygous of p21 gene could be a risk tional University Hospital approved the protocol and
factor for the development of cervical adenocarcinoma but informed consent was obtained from all of the subjects
not for squamous cell carcinoma of cervix [21]. The authors involved in the study.
hypothesized that this polymorphism could be associated
with the risk of adenocarcinoma of uterine corpus adjacent p53 and p21 genotyping
to the uterine cervix.
Although endometrial cancer is a common female ma- DNA was extracted from peripheral blood samples
lignancy and the prevalence has been increased rapidly in according to the standard procedures. p53 polymorphism
 J.W. Roh et al. / Gynecologic Oncology 93 (2004) 499–505 501

was evaluated by polymerase chain reaction (PCR) using Statistical analysis

allele-specific primers as follows: p53 Pro sequences were
detected by PCR using the primers p53Pro+/p53-(5V- The women were stratified by menopausal status. Chi-
GCCAGAGGCTGCTCCCCC-3V, 5V-CGTGCAAGTCA- square statistics were used to evaluate the distribution of
CAGACTT-3V), and p53 Arg sequences with the primers the p53 and p21 genotypes among the cancer patients
p53+/p53Arg-(5V-TCCCCCTTGCCGTCCCAA-3V, 5V- and control subjects. Chi-square test calculations for the
CTGGTGCAGGGGCCACGC-3V). A master mix was used, purpose of determining the deviation from the Hardy–
containing 50 ng of cellular DNA, 20 pmol of each primer, Weinberg equilibrium were performed as well. Odds
1.5 mM dNTP, 67 mM EDTA, 3 mM MgCl2, and 1.25 units ratios (ORs) and 95% confidence intervals (CIs), obtained
of Taq polymerase (Bioneer, Korea). The final volume of the from unconditional logistic regression, were used to
reaction mixture was 25 Al. The PCR conditions were as measure the strength of the association between the p53
follows: 25 cycles of 30 s at 94jC, 30 s at 55jC (for or p21 polymorphism and endometrial cancer risk. The
p53Pro+/p53-) or 60jC (for p53+/p53Arg-), and 30 s at ORs were adjusted in the multiple logistic model for
72jC using a Thermal Cycler (Perkin-Elmer Inc., Wellesley, education, body-mass index (BMI, kg/m2; weight divided
MA). DNAs from patients whose p53 sequences had been by the square of the height) (V25, >25 kg/m2), history of
determined by direct sequencing was used for the positive diabetes, and family history of endometrial cancer. The
controls and we confirmed the genotype by direct sequenc- selection criteria for the variables included in the final
ing in randomly selected samples more than 10% of total multiple logistic model was the existence of a statistically
samples. The PCR-amplified products were electrophoresed significant difference in the bivariate analysis. All statis-
(Hoefer Scientific Instruments, San Francisco, CA) on a 2% tical calculations were performed using the SPSS soft-
agarose gel. ware package version 11.0 for Windows (SPSS Inc.,
p21 genotyping was performed as described previously Chicago, IL).
[23]. A 273 base pair (bp) PCR fragment was amplified with
the primer sets 5V-GGATGTCCGTCAGAACCCAT-3V, 5V-
GGTGCCAGGCCGCCTGCCTC-3V and digested with Results
BsmAI. The digested fragments were run on 8% polyacryl-
amide gel and visualized under UV light after ethidium The values of the putative risk factors for endometrial
bromide staining. The serine allele with a single BsmAI site cancer in the cases and controls are shown in Table 1.
gave two common fragments of 142 and 131 bp. The Education, obesity (BMI > 25 kg/m2), a family history of
arginine allele was detected through the presence of a endometrial cancer, and a history of diabetes showed
second BsmAI site within the 142-bp fragment, which significant differences between the cases and the controls.
generated 75 and 67 bp fragments. The results of this case-control study on the association
The DNAs from patients whose p53 and p21 sequences between p53 and p21 genetic polymorphisms and endome-
had been determined by direct sequencing were used as trial cancer are presented in Table 2. In both groups, the
positive controls. To test the reliability of the assay, 50 distribution of genotypes fits the Hardy– Weinberg equilib-
randomly selected samples were re-tested with identical rium. We found statistically significant differences in the
results being obtained. frequency of the p53 genotype between the two groups of

Table 1
Case and control differences in putative risk factors for endometrial cancer
All women Premenopausal women Postmenopausal women
Cases Controls P value Cases Controls P value Cases Controls P value
(n = 95) (n = 285) (n = 52) n = 179) (n = 43) (n = 106)
Mean age at diagnosis 49.9 (11.5) 47.9 (9.9) NS 42.5 (7.4) 41.7 (6.1) NS 59.0 (8.6) 58.5 (5.0) NS
(years) (SD)
% Education 32 48 0.003 50 47 NS 9 52 <0.001
(over high school)
Mean age at menarche 15.7 (1.7) 15.1 (1.9) NS 15.7 (1.9) 15.2 (1.9) NS 15.8 (1.4) 16.0 (2.0) NS
(years) (SD)
Mean age first pregnant 25.0 (3.7) 25.5 (3.2) NS 25.1 (3.9) 25.4 (3.3) NS 24.8 (3.6) 25.5 (3.1) NS
(years) (SD)
% Obesity 49 32 0.009 50 29 0.005 49 43 NS
(BMI > 25 kg/m2)
% Family history 4 1 0.018 8 1 0.002 0 1 NS
% Diabetes history 9 4 0.034 6 2 NS 14 8 NS
% Hypertension history 15 11 NS 4 7 NS 28 18 NS
SD, standard deviation; NS, not significant; BMI, body mass index.
502 J.W. Roh et al. / Gynecologic Oncology 93 (2004) 499–505

Table 2
Association between p53 and p21 genetic polymorphisms and endometrial cancer risk
All women Premenopausal women Postmenopausal women
Cases Controls OR Cases Controls OR Cases Controls OR
n (%) n (%) (95% CI) n (%) n (%) (95% CI) n (%) n (%) (95% CI)
p53 codon 72
Arg/Arg 29 (30.5) 166 (58.3) 1.0 17 (32.7) 98 (54.8) 1.0 12 (27.9) 68 (64.2) 1.0
(reference) (reference) (reference)
Arg/Pro or Pro/Pro 66 (69.5) 119 (41.7) 3.56 35 (67.3) 81 (45.2) 2.68 31 (72.1) 38 (35.8) 7.14
(2.10 – 6.04) (1.34 – 5.35) (2.76 – 18.45)

p21 codon 31
Arg/Arg or Arg/Ser 35 (36.8) 170 (59.7) 1.0 20 (38.5) 106 (59.2) 1.0 15 (34.9) 64 (60.4) 1.0
(reference) (reference) (reference)
Ser/Ser 60 (63.2) 115 (40.3) 2.68 32 (61.5) 73 (40.8) 2.24 28 (65.1) 42 (39.6) 3.12
(1.59 – 4.51) (1.14 – 4.42) (1.23 – 7.91)
The ORs were adjusted for education, obesity, history of diabetes, and family history of endometrial cancer and menopausal status.

cases and the non-cancer controls ( P < 0.001). The p53 Pro risk genotypes was found for all women, irrespective of
allele, including the homozygous and heterozygous catego- their menopausal status ( P for trend < 0.001); the OR was
ries, was significantly associated with endometrial cancer 2.87 (95% CI 1.36 –6.08) for women carrying only one
with an odds ratio (OR) of 3.56 (95% confidence interval risky genotype of these two genes and 9.55 (95% CI 4.30–
[CI], 2.10 – 6.04), while the risk increased to 7.14-fold (95% 21.24) for women carrying both risky genotypes compared
CI 2.76 –18.45) in postmenopausal women. The differences to those with no at-risk genotype.
in frequency of the p21 genotype between the two groups of
cases and the non-cancer controls were also statistically
significant ( P < 0.001). The homozygote of the Ser allele Discussion
at codon 31 of p21 showed a significantly increased risk
(OR = 2.68; 95% CI = 1.59– 4.51) of endometrial cancer Two forms of polymorphism can occur in exon 4 of the
as compared with the carriers of the Arg allele, while this wild-type p53 gene due to a 1-bp change, which involves
risk increased to 3.12-fold (95% CI 1.23– 7.91) in post- coding for Arginine (Arg) or Proline (Pro) residues at codon
menopausal women. However, the differences in the clin- 72. Although it was suggested that the two alleles of codon
icopathologic factors such as the stage of the disease, the 72 might have different oncogenic properties, inconsistent
histologic grade, the lymph node metastasis, positive peri- results have been reported in various type of cancers [24 –
toneal cytology, positive estrogen or progesterone receptors, 26]. In the case of breast cancer, one of the hormone-related
and the 5-year survival rate (data not shown) were not cancers and which has similar epidemiologic-risk factors
significant among the patients of either genotype group, compared with endometrial cancer, the studies reported so
irrespective of the menopausal status. far on the polymorphism of p53 codon 72 have failed to
We further investigated the potential combined effects of yield concordant results [10 – 12].
these genotypes on the risk of endometrial cancer. The As far as we know, this is the first report of a positive
respective ORs were calculated according to the number association between p53 codon 72 genetic polymorphism
of at-risk genotypes of these genes (Table 3). A trends and the risk of developing endometrial cancer. There have
toward an increased risk with an increasing number of at- been few documented case-control studies on the associa-

Table 3
The combination of p53 and p21 genotypes in endometrial cancer risk
Number Genotypes All women Premenopausal women Postmenopausal women
of at-risk p53 codon 72 p21 codon 31 Cases/ OR Cases/ OR Cases/ OR
genotypes controls (95% CI) controls (95% CI) controls (95% CI)
0 Arg/Arg Arg/Arg or 10/98 1.0 6/58 1.0 4/40 1.0
Arg/Ser (reference) (reference) (reference)
1 Arg/Arg or Ser/Ser
Arg/Pro or Arg/Arg or 43/141 2.87 25/88 2.82 18/53 2.72
Pro/Pro Arg/Ser (1.36 – 6.08) (1.06 – 7.48) (0.80 – 9.24)
2 Arg/Pro or Ser/Ser 42/46 9.55 21/33 6.28 21/13 20.79
Pro/Pro (4.30 – 21.24) (2.21 – 17.87) (5.09 – 84.83)
p for trend < 0.001 p for trend < 0.001 p for trend < 0.001
The ORs were adjusted for education, obesity, history of diabetes, and family history of endometrial cancer and menopausal status.
 J.W. Roh et al. / Gynecologic Oncology 93 (2004) 499–505 503

tion between p53 codon 72 polymorphism and the risk of human malignancies, including colorectal cancer [34], soft
developing endometrial cancer [13,14]. In these studies, tissue sarcoma [23], breast cancer [18], prostate adenocar-
however, the authors were not able to demonstrate the cinoma [35], and squamous cell carcinoma of the head and
existence of a statistically significant association with the neck [34,35]. However, most of these studies were not
risk of endometrial cancer. However, in the present study, based on molecular epidemiologic concepts and merely
we found a statistically significant association between the observed that polymorphism in p21 codon 31 was found in
presence of the Pro allele at codon 72 of p53 and the risk of a significant number of cancer patients, although several
endometrial cancer in Korean women. case-control studies designed to investigate the association
The discrepancies between the previous reports and the between p21 codon 31 polymorphism and the various
present study may stem from several factors, such as the types of cancer have been reported. Sjalander et al. [36]
sample size, the techniques used, ethnic differences, and reported an association between the Arg allele of the p21
variations in genetic susceptibility. In the previous studies, codon 31 polymorphism and lung cancer; however, Shih et
both groups of patients with cancer and the control group al. [37] reported no such significant association. Hachiya et
were relatively too small to yield acceptable statistical al. [20] reported various genotypes, including an Arg
measures, in contrast to our study. A wide variety of allele, to be somewhat more common in endometrial
different methods have been established to detect DNA cancer than in the controls among Japanese, albeit with
sequence variations. p53 genotype analysis has been per- marginal significance (OR = 2.52, 95% CI = 1.09– 5.8, P =
formed with diverse methods such as the PCR assay 0.09). Roh et al. [21] reported that the Ser/Ser homozygote
involving the amplification of specific alleles, direct se- could be a risk factor for the development of cervical
quencing, the BstUI restriction digestion of PCR products, adenocarcinoma associated with high-risk HPV (OR =
and PCR-single stranded conformational polymorphism. In 3.59, 95% CI = 1.55 – 8.31, P = 0.002). Konishi et al. [38]
the study of Esteller et al. [13] and in this present study, were not able to confirm that a specific genotype was
the PCR assay was used, involving the restriction fragment associated with an increased risk of developing skin cancer.
length polymorphism methods and amplification of spe- Milner et al. [39] even found that there was a tendency for
cific alleles, respectively. Peller et al. [14] verified the p53 p21 polymorphisms containing the Arg allele to be associ-
genotype by performing direct sequencing. Makni et al. ated with a decreased risk of ovarian cancer. Our data
[27] reported that interlaboratory variations in the identifi- provide evidence that the Arg/Arg genotype is not a risk
cation of the p53 genotypes may translate into a substantial factor for the development of endometrial cancer and
risk of misclassification and such misclassifications can suggest an elevated risk of endometrial cancer for women
lead to a bias toward the null hypothesis, that is, losing the with the Ser/Ser genotype as compared to the other two
ability to detect a significant correlation between the genotypes. There is a clear discrepancy between these
presence of the codon 72 genotype and the risk of cancer. results and those of Hachiya et al. There are several
In addition, since the loss of heterozygosity on chromo- potential reasons for this discrepancy. Our study was
some 17p has been reported in 35% of patients with conducted on a sample size almost two times larger in
endometrial cancer, the frequency of endometrial cancer terms of the number of cases and five times larger in terms
patients homozygous for Pro or Arg might be increased in of the number of controls than Hachiya’s and so gives a
the studies using samples obtained from cancer tissues more precise estimate of the association between p21
[28]. Race-specific variations in the frequency of the p53 polymorphism and endometrial cancer. In these two studies,
genotype have been documented by certain researchers different methodologies were employed, that is, dot blot
[16]. In the study of Esteller et al. [13], DNA extracted hybridization and restriction fragment length polymorphism
from Caucasian women was used, which was different in Hachiya’s study and our study, respectively. Another
from that used in this study. The same situation has been positive aspect of our study was the kind of samples used
observed in cervical cancer. Although some authors in the genotyping. All of the samples used in our study
reported a higher susceptibility of carriers of the p53 Arg consisted of DNA extracted from peripheral blood WBC,
allele at codon 72 to HPV-mediated carcinogenesis of the whereas those used in Hachiya’s study were from tumor
uterine cervix [26,29], other researchers have refuted this tissues or peripheral nucleated cells. In case-control studies,
finding [30 – 32]. In Korean women, the reported data the composition of the control population is a major factor
provided evidence that the Arg/Arg genotype was not a affecting the outcome. In Hachiya’s study, the source of the
risk factor for the development of cervical carcinoma, as samples for the control group was not described. Our study
sure as the data on endometrial cancer obtained in this was also hospital-based, and so was not completely free
study [17,33]. However, an additional large population- from the innate problems of such a study design. However,
based study is needed to confirm the finding that the Pro we attempted to obtain a representative control groups
allele of p53 codon 72 might be a general risk factor for because the factors related to inclusion, such as age, age
uterine cancer. at menarche, age at first pregnancy, body mass index, and a
Previous studies reported the existence of an association history of diabetes mellitus and/or hypertension, may be
between polymorphism in p21 codon 31 and various related to the development of endometrial cancer. Differ-
504 J.W. Roh et al. / Gynecologic Oncology 93 (2004) 499–505

ences in education, obesity, a family history of endometrial part by grants from the Cancer Research Institute Research
cancer, and a history of diabetes between case and control Fund (2003-CRI-4), Seoul National University, and in part
groups were significant in our study, as was expected by the Seoul National University Hospital Research Fund
(Table 1). So, all of these factors, which are potential (05-2003-002-0).
confounding factors for the statistical analysis of genotypes,
were adjusted to minimize the misinterpretation of the
genotype influence. We recruited consecutive patients with-
References
out selection. Mean age (49.9 years) of the cases was
compatible with nationwide statistics reported by Korea [1] Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide incidence
central cancer registry program and younger than that in of 25 major cancers in 1990. Int J Cancer 1999;80:827 – 41.
Western countries. The age of peak incidence of endome- [2] 2002 Annual report of the Korea central cancer registry (2002.1. –
trial cancer was about 50 in Korea. Genetic background 2002.12). Republic of Korea: Korea Central Cancer Registry, Minis-
might be related to the differences of prevalence and age try of Health and Welfare; 2003.
[3] Weber BL, Nathanson KL. Low penetrance genes associated with
distribution of endometrial cancer between Korea and increased risk for breast cancer. Eur J Cancer 2000;36:1193 – 9.
Western countries. Similar phenomenon could be observed [4] Lane DP. Cancer, p53, guardian of the genome. Nature 1992;358:
in the age distribution of breast cancer patients. About 60% 5 – 16.
of the patients with breast cancer in Korea are premeno- [5] Eldeiry WS, Tokini T, Velculescu VE, Levy DB, Parsons R, Trent JM,
et al. WAF1, a potential mediator of p53 tumor suppression. Cell
pausal state contrary to the Caucasians [2]. The age of peak
1993;75:817 – 25.
incidence of endometrial cancer has not been changed [6] Honda T, Kato H, Imamura T, Gima T, Nishida J, Sasaki M, et al.
during the last decade, although annual prevalence of Involvement of p53 gene mutations in human endometrial carcino-
newly diagnosed endometrial cancer of Korea was in- mas. Int J Cancer 1993;53:963 – 7.
creased 2.6-fold (825 patients) in 2002 compared with that [7] Kihana T, Hamada K, Inoue Y, Yano N, Iketani H, Murao S, et al.
in 1996 (315 patients). In this study, we could not make an Mutation and allelic loss of the p53 gene in endometrial carcinoma.
Incidence and outcome in 92 surgical patients. Cancer 1995;76:
inference that p21 and p53 polymorphisms might be related 72 – 8.
to the geographic characteristics of endometrial cancer. [8] Niwa K, Murase T, Morishita S, Hashimoto M, Itoh N, Tamaya T. p53
The rationale for the association between p21 codon 31 overexpression and mutation in endometrial carcinoma: inverted re-
polymorphism and endometrial cancer risk is unclear at lation with estrogen and progesterone receptor status. Cancer Detect
present. In a previous study, no significant functional Prev 1999;23:147 – 54.
[9] Matlashewski GJ, Tuck S, Pim D, Lamb P, Schneider J, Crawford LV.
difference about the inhibition of CDK or the inhibition of Primary structure polymorphism at amino acid residue 72 of human
tumor cell growth was demonstrated for the two forms of p53. Mol Cell Biol 1987;7:961 – 3.
p21 protein [40]. Konishi et al. [38] even suggested that [10] Sjalander A, Birgander R, Hallmans G, Cajander S, Lenner P, Athlin
codon 31 polymorphism seems to be only an innocent L, et al. p53 polymorphisms and haplotypes in breast cancer. Carci-
bystander, and that the 3’UTR site that linked firmly to nogenesis 1996;17:1313 – 6.
[11] Papadakis EN, Dokianakis DN, Spandidos DA. p53 codon 72 poly-
the codon 31 site may be involved in such an association. morphism as a risk factor in the development of breast cancer. Mol
Further experiments, taking this information into consider- Cell Biol Res Commun 2000;3:389 – 92.
ation, are required to evaluate the possible functional differ- [12] Buyru N, Tigli H, Dalay N. P53 codon 72 polymorphism in breast
ences that contribute to carcinogenesis. cancer. Oncol Rep 2003;10:711 – 4.
In the present study, we demonstrated that the poly- [13] Esteller M, Garcia A, Martinez-Palones JM, Xercavins J, Reventos J.
Susceptibility to endometrial cancer: influence of allelism at p53,
morphisms of p53 codon 72 and p21 codon 31 were glutathione S-transferase (GSTM1 and GSTT1) and cytochrome
significantly associated with the occurrence of endometrial P-450 (CYP1A1) loci. Br J Cancer 1997;75:1385 – 8.
cancer. These significant differences were maintained [14] Peller S, Halperin R, Schneider D, Kopilova Y, Rotter V. Polymor-
throughout the groups after they were stratified by meno- phisms of the p53 gene in women with ovarian or endometrial carci-
noma. Oncol Rep 1999;6:193 – 7.
pausal status, although the higher ORs were observed in
[15] Ameyaw MM, Thornton N, McLeod HL. Re: population-based, case-
postmenopausal cases. We could hypothesize that the ge- control study of HER2 genetic polymorphism and breast cancer risk.
netic influence may play a larger role in postmenopausal J Natl Cancer Inst 2000;92:1947.
women because effects of environmental factors, that is, [16] Sjalander A, Birgander R, Kivela A, Beckman G. p53 polymor-
hormonal effect, decrease in old menopausal women. These phisms and haplotypes in different ethnic groups. Hum Hered 1995;
findings still remain to be confirmed by an additional large 45:144 – 9.
[17] Kim JW, Roh JW, Park NH, Song YS, Kang SB, Lee HP. Polymor-
population-based study. phism of TP53 codon 72 and the risk of cervical cancer among
Korean women. Am J Obstet Gynecol 2001;184:55 – 8.
[18] Lukas J, Groshen S, Saffari B, Niu N, Reles A, Wen WH, et al.
Acknowledgments WAF1/Cip1 gene polymorphism and expression in carcinomas of
the breast, ovary, and endometrium. Am J Pathol 1997;150:167 – 75.
[19] Chedid M, Michieli P, Lengel C, Huppi K, Givol D. A single nucle-
We thank Professor Daehee Kang (Faculty of Preven- otide substitution at codon 31 (Ser/Arg) defines a polymorphism in a
tive Medicine, Seoul National University) for his assis- highly-conserved region of the p53-inducible gene WAF1/CIP1. On-
tance with the data analysis. This work was supported in cogene 1994;9:3021 – 4.
 J.W. Roh et al. / Gynecologic Oncology 93 (2004) 499–505 505

[20] Hachiya T, Kuriaki Y, Ueoka Y, Nishida J, Kato K, Wake N. WAF1 sen-Dale AL. p53 polymorphism and risk of cervical cancer. Nature
genotype and endometrial cancer susceptibility. Gynecol Oncol 1999; 1998;396:530 – 1.
72:187 – 92. [31] Hildesheim A, Schiffman M, Brinton LA, Fraumeni Jr JF, Herrero R,
[21] Roh JW, Kim MH, Kim JW, Park NH, Song YS, Kang SB, et al. Bratti MC, et al. p53 polymorphism and risk of cervical cancer. Na-
Polymorphism in codon 31 of p21 and cervical cancer susceptibility. ture 1998;396:531 – 2.
Cancer Lett 2001;165:59 – 62. [32] Josefsson AM, Magnusson PK, Ylitalo N, Quarforth-Tubbin P, Pon-
[22] Yair D, Ben Baruch Gk Chetrit A, Friedman T, Hirsh Yechezkei G, ten J, Adami HO, et al. P53 polymorphism and risk of cervical cancer.
Gotliev WH, et al. p53 and WAF1 polymorphisms in Jewish-Israeli Nature 1998;396:531.
women with epithelial ovarian cancer and its association with BRCA [33] Kim JW, Lee CG, Park YG, Kim KS, Kim IK, Sohn YW, et al.
mutations. Br J Obstet Gynaecol 2000;107:849 – 54. Combined analysis of germline polymorphisms of p53, GSTM1,
[23] Mousses S, Ozcelik H, Lee PD, Malkin D, Bull SB, Andrulis IL. Two GSTT1, CYP1A1, and CYP2E1. Cancer 2000;88:2082 – 91.
variants of the CIP1/WAF1 gene occur together and are associated [34] Shiohara MM, el-Deiry WS, Wada M, Nakamaki T, Takeuchi S, Yang
with human cancer. Hum Mol Genet 1995;4:1089 – 92. R, et al. Absence of WAF1 mutations in a variety of human malig-
[24] Fan R, Wu MT, Miller D, Wain JC, Kelsey KT, Wiencke JK, et al. nancies. Blood 1994;84:3781 – 4.
The p53 codon 72 polymorphism and lung cancer risk. Cancer Epi- [35] Facher EA, Becich MJ, Deka A, Law JC. Association between human
demiol Biomarkers Prev 2000;9:1037 – 42. cancer and two polymorphisms occurring together in the p21Waf1/Cip1
[25] Shepherd T, Tolbert D, Benedetti J, Macdonald J, Stemmermann G, cyclin-dependent kinase inhibitor gene. Cancer 1997;79:2424 – 9.
Wiest J, et al. Alterations in exon 4 of the p53 gene in gastric carci- [36] Sjalander A, Birgander R, Rannug A, Alexandrie AK, Tornling G,
noma. Gastroenterology 2000;118:1039 – 44. Beckman G. Association between the p21 codon 31 A1 (arg) allele
[26] Zehbe I, Voglino G, Wilander E, Delius H, Marongiu A, Edler L, et al. and lung cancer. Hum Hered 1996;46:221 – 5.
p53 codon 72 polymorphism and various human papillomavirus 16 [37] Shih CM, Lin PT, Wang HC, Huang WC, Wang YC. Lack of
E6 genotypes are risk factors for cervical cancer development. Cancer evidence of association of p21WAF1/CIP1 polymorphism with lung
Res 2001;61:608 – 11. cancer susceptibility and prognosis in Taiwan. Jpn J Cancer Res
[27] Makni H, Franco EL, Kaiano J, Villa LL, Labrecque S, Dudley R, 2000;91:9 – 15.
et al. P53 polymorphism in codon 72 and risk of human papillo- [38] Konishi R, Sakatani S, Kiyokane K, Suzuki K. Polymorphisms of p21
mavirus-induced cervical cancer: effect of inter-laboratory variation. cyclin-dependent kinase inhibitor and malignant skin tumors. J Der-
Int J Cancer 2000;87:528 – 33. matol Sci 2000;24:177 – 83.
[28] Jones MH, Koi S, Fujimoto I, Hasumi K, Kato K, Nakamura Y. [39] Milner BJ, Brown I, Gabra H, Kitchener HC, Parkin DE, Haites NE.
Allelotype of uterine cancer by analysis of RFLP and microsatellite A protective role for common p21WAF1/Cip1 polymorphisms in human
polymorphisms: frequent loss of heterozygosity on chromosome arms ovarian cancer. Int J Oncol 1999;15:117 – 9.
3p, 9q, 10q, and 17p. Genes Chromosomes Cancer 1994;9:119 – 23. [40] Sun Y, Hildesheim A, Li H, Li Y, Chen JY, Cheng YJ, et al. No point
[29] Storey A, Thomas M, Kalita A, Harwood C, Gardiol D, Mantovani F, mutation but a codon 31ser!arg polymorphism of the WAF-1/Cip-1/
et al. Role of a p53 polymorphism in the development of human p21 tumor suppressor gene in nasopharyngeal carcinoma (NPC): the
papillomavirus-associated cancer. Nature 1998;393:229 – 34. polymorphism distinguishes Caucasians from Chinese. Cancer Epi-
[30] Helland A, Langerod A, Johnsen H, Olsen AO, Skovlund E, Borre- demiol Biomarkers Prev 1995;4:261 – 7.