p53 Codon 72 Polymorphism (C/G) and the Riskof Human Papillomavirus-AssociatedCarcinomas in China

Tao Li1

Zhe-Ming Lu1

Mei Guo, M.S.1

Qin-Jiao Wu, M.S.1

Ke-Neng Chen, M.D.2

Hai-Ping Xing, B.S.3

Qiang Mei, B.S.3

Yang Ke, M.D., Ph.D.1

1 Laboratory of Genetics, Beijing Institute for Can-cer Research, School of Oncology, Peking Univer-sity, Beijing, People’s Republic of China.

2 Department of Surgery, Beijing Institute for Can-cer Research, School of Oncology, Peking Univer-sity, Beijing, People’s Republic of China.

3 Anyang Tumor Hospital, Anyang City, People’sRepublic of China.

Supported by the National Nature and ScienceFoundation (grant 30070834). This study was akey project of the Beijing Metropolitan Science andTechnology Committee and Human Disease GeneCenter at Peking University, Peking, and the Chi-nese National Human Genome Center, Beijing.

Address for reprints: Yang Ke, M.D., Ph.D., Labo-ratory of Genetics, Beijing Institute for Cancer Re-search, School of Oncology, Peking University, No.1 Da Hong Luo Chang Street, Beijing, People’sRepublic of China 100034; Fax: 011-8610-66175832; E-mail: keyang@mx.cei.gov.cn

Received February 11, 2002; revision receivedJuly 7, 2002; accepted July 15, 2002.

BACKGROUND. Human papillomavirus (HPV) plays an important role in the devel-

opment of carcinomas at various body sites. It was found previously that the p53

codon 72 polymorphism (C/G) is a high-risk factor for the development HPV-

associated cervical carcinoma. However, it still was considered controversial in

several studies of cervical and esophageal carcinoma.

METHODS. In the current study, the authors used an allele specific polymerase

chain reaction (PCR) method to analyze correlation between the p53 codon 72

(C/G) polymorphism and HPV-associated, noncancerous esophageal epithelium as

well as esophageal, ovarian, and breast carcinoma in the Chinese population.

Esophageal balloon cytology examination samples were obtained from high-inci-

dence and low-incidence populations for esophageal carcinoma in Anyang (Henan

Province).

RESULTS. Thirty-six of 48 esophageal balloon samples from the high-incidence

population were HPV positive, and 13 of 33 esophageal balloon samples from the

low-incidence population were HPV positive. Thirty-nine of 62 esophageal carci-

noma samples from Anyang Tumor Hospital were HPV positive. Twenty-six of 39

ovarian carcinoma samples from the Second Affiliated Hospital of Inner Mongolia

Medical College were HPV positive. Nineteen of 82 breast carcinoma samples from

Beijing Cancer Hospital were HPV positive. It is noteworthy that the distribution of

the p53 codon 72 Arg homozygous genotype in HPV positive samples of esophageal

epithelium, ovarian carcinoma, and breast carcinoma was significantly higher

compared with HPV negative tumor samples. (P � 0.05).

CONCLUSIONS. The current results suggest that the p53 codon 72 Arg homozygous

genotype is one of the high-risk genetic factors for HPV-associated malignancies

among the Chinese population. Cancer 2002;95:2571– 6.

© 2002 American Cancer Society.

DOI 10.1002/cncr.11008

KEYWORDS: p53, polymorphism, human papillomavirus, allele specific polymerasechain reaction analysis.

Human papillomavirus (HPV) plays an important role in the devel-opment of malignant disease at various body sites, including the

anogenital area, the upper respiratory tract, the digestive tract, andthe breast.1,2 However, not all HPV-containing lesions proceed to thefull malignant phenotype. It has been shown that HPV infection alonemay not be sufficient for the malignant transformation of a host celland requires additional genetic lesions. Germline polymorphisms ofsome genes involved in multiple steps of carcinogenesis may accountfor the genetic susceptibility to HPV-associated carcinoma.

The high-risk HPV types, such as HPV type 16 (HPV-16) and

2571

© 2002 American Cancer Society

HPV-18, encode two viral oncoproteins, E6 and E7,that are expressed in HPV-associated malignancies.1,2

It has been shown that E7 protein is capable of bind-ing and inactivating the cellular tumor suppressorprotein Rb, whereas the E6 protein interacts with anddegrades p53 through the ubiquitin pathway. Theseinteractions are responsible for the transforming ac-tivity of HPV.3–5 A common polymorphic site in thewild type p53 gene at codon 72 of exon 4 results intranslation to either an arginine residue (CGC) or aproline residue (CCC). The frequency of two allelesvaries among ethnic groups that dwell at differentlatitudes.6 The p53 codon 72 polymorphism (C/G) wasone of the first susceptible factors of HPV-associatedcarcinoma that was found in 1998 by Storey et al.,7

who reported that the p53 arginine (p53Arg) isoformincreased the susceptibility of p53 to HPV-16 andHPV-18, E6-mediated degradation either in vitro or invivo. Moreover, there is a significant over-representa-tion of the p53 Arg/Arg genotype in patients withcervical carcinoma compared with the normal popu-lation. Individuals homozygous for p53Arg were aboutseven times more susceptible to HPV-associated car-cinogenesis compared with heterozygotes.7 However,many subsequent studies in different populationshave failed to prove the results.8 –13 The correlationbetween the p53 codon 72 polymorphism (C/G) andHPV-associated carcinoma remains controversial.

We showed previously that HPV-16 plays a signifi-cant role in the pathogenesis of esophageal carcinoma inAnyang, which is recognized as a high-incidence area foresophageal carcinoma in China. Retaining of viral DNAfragments E6/E7 and stable expression of viral genesmay be associated with the malignant progression ofesophageal epithelial cells.14 Our recent investigationalso showed a correlation between HPV and breast andovarian carcinoma15 (and unpublished results). It wasfound that the HPV-16 infection rate was significantlyhigher in samples of epithelial ovarian malignancy (26 of50 samples; 52%) compared with the rate in normalovarian tissues (2 of 30 samples; 6.7%), and HPV-16 E6mRNA was found in 17 of 34 breast carcinoma samples(50%). Based on our previous studies, investigation ofgenetic cancer susceptibility in the p53 gene, which maybe involved in the steps of carcinogenesis, is importantfor the identification of high-risk individuals in an HPV-infected population.

In the current study, we examined esophagealballoon samples of different histologic grade from theAnyang area and from three types of HPV-associatedmalignancies to determine the frequency of p53 poly-morphism (C/G) in the Chinese population.

MATERIALS AND METHODSSample Collection, Preparation, and HPV DetectionEsophageal balloon cytology examination sampleswere obtained from the Anyang area, which has thehighest incidence and mortality of esophageal carci-noma in China. In this area, the esophageal carcino-ma-adjusted and age-adjusted mortality rates among13 counties vary from 141 � 105 population to 23� 105 population. We collected specimens by esoph-ageal balloon cytology examination from volunteersliving in two villages, Shangzhuang and Tiangmiao,which have age-adjusted mortality rates for esopha-geal carcinoma of 132 � 105 and 52 � 105, respec-tively. The villages are situated approximately 100 kmapart: Shangzhuang is a mountain village, and Tiang-miao is situated in the plain area. Partly because of thenatural environment and traffic, the economic condi-tions in Shangzhuang are poor compared with thosein Tiangmiao. Volunteers from Shangzhuang andTiangmiao were considered high-incidence and low-incidence populations for esophageal carcinoma, re-spectively. Sample collection and preparation proce-dures and HPV detection were described previously.14

Briefly, we had the volunteers swallow a deflated bal-loon, which was then expanded and gently removed,and their esophageal epithelial cells were then col-lected. Cells on the surface of the balloon weresmeared onto slides and sent to the Department ofPathology at Anyang Tumor Hospital for cytologicevaluation. Parallel slides were collected for in situhybridization of the HPV-16 E6 and HPV-18 E6 genes.After the slide smears were taken, the remaining cellson the balloon were washed off with saline and con-centrated by centrifugation at � 4000 g for 5 minutes.The cell precipitation was dissolved in 500 �L lysisbuffer (10 mm Tris-Cl, pH 7.8; 100 mm NaCl, 10 mmethylenediamine tetraacetic acid [EDTA], and 0.5%sodium codicil sulfate [SDS], and 10 �L of 20 mg/mLproteinase K were added to the cell lysate. After incu-bation at 55 °C overnight, the DNA was extracted withphenol/chloroform. The presence of HPV DNA wasevaluated with polymerase chain reaction (PCR) anal-ysis by using type specific primers for the HPV E6 geneof HPV-16, HPV-18, and the HPV E7 gene of HPV-16.The in situ hybridization method was used to detectHPV-16 and HPV-18 E6 mRNA. The single-strandprobe was digoxin labeled, and its sensitivity corre-sponded to the sensitivity of the isotope method.Forty-eight samples were from the high-incidencepopulation, and 36 of those samples were HPV posi-tive. Thirty-three samples were from the low-incidencepopulation, and 13 of those samples were HPV posi-tive. Cytologic evaluation revealed that 17 of 36 sam-

2572 CANCER December 15, 2002 / Volume 95 / Number 12

ples of normal epithelium were HPV positive, 23 of 34samples of mild dysplasia were HPV positive, and 8 of11 samples of severe dysplasia were HPV positive.

In addition, we studied the following samples:There were 62 esophageal carcinoma samples fromAnyang Tumor Hospital (39 were HPV positive), 39ovarian carcinoma samples from the Second AffiliatedHospital of Inner Mongolia Medical College (26 wereHPV positive), and 28 breast carcinoma samples fromBeijing Cancer Hospital (19 were HPV positive). All ofthese samples were paraffin embedded archival sam-ples that were collected between 1999 and 2000. HPVdetection (HPV-16 and HPV-18) was performed withboth PCR and in situ hybridization methods in ourprevious study.14,16 We selected 50 healthy blood do-nors among the Beijing population as a control group.Although our samples were from three different geo-graphic areas, we studied the distribution of the threep53 genotypes in age-matched, healthy control popu-lations in different geographic areas and found nosignificant difference among these populations (un-published data). Informed consent was obtained fromall individuals.

DNA ExtractionDNA extraction from esophageal balloon cytologysamples was described previously.14 DNA from paraf-fin embedded carcinoma samples was extractedthrough the following procedures: Five-micrometer,formalin fixed, paraffin sections of carcinoma sampleswere deparaffinized in xylene; washed with 100%,95%, and 75% ethanol; pelleted; air dried; and incu-bated overnight at 55 °C with 10 �L of 20 mg/mLproteinase K in 500 �L lysis buffer (10 mM Tris-Cl, pH7.8; 100 mM NaCl, 10 mM EDTA, and 0.5% SDS). TheDNA was extracted with phenol/chloroform and pre-cipitated with cold ethanol. We extracted DNA from 50healthy blood donors as follows: First, white bloodcells were separated from red blood cells by washingthree times in lysis buffer. Then, the DNA was ex-tracted with phenol/chloroform and precipitated withcold ethanol. All DNA samples were dissolved in waterand stored at �20 °C. The quality of DNA was con-firmed by amplification of each sample with microsat-ellite primers (D3S1561; GIBCO BRL, Gaithersburg,MD).

Allele Specific PCR Analysis of the p53 PolymorphismGenotyping of p53 at codon 72 was carried out with anallele specific PCR amplification procedure. Two pairs ofprimers with different 3� terminal bases were used toamplify p53 sequences separately (Fig. 1): p53-p1 (GC-CAGAGGCTGCTCCCC) and p53-p2 (CGTGCAAGTCA-CAGACTT) for proline (178 base pairs [bp]) and p53-a1

(TCCCCCTTGCCGTCCCAA) and p53-a2 (CTGGTG-CAGGGGCCACG) for arginine (142 bp). The PCR reac-tion was performed in a total volume of 25 �L. The PCRmixture contained 10 mM Tris-Cl, pH 8.3; 50 mM KCl;1.5 mM MgCl; 0.01% gelatin; 200 pmol of each primer; 2U Taq DNA polymerase; and 100 ng DNA. Amplificationwas performed with denaturation at 94 °C for 2 minutesfollowed by 30 seconds at 94 °C, 50 seconds at 62 °C, and30 seconds at 72 °C for 35 cycles, and a final extension for5 minutes at 72 °C. Then, the PCR products of eachsample were mixed and electrophoresed on 2% agarosegels and were visualized with ethidium bromide stain-ing. The samples were renumbered before the p53 PCRassay. Experimenters were blinded to the cytopathologicgrade and HPV status of all samples. The PCR assay wasreiterated in a different laboratory (located on a differentfloor of the building) to verify our results. Amplified PCRproducts were sequenced using both forward and re-verse primers (ABI PRISM model 377, version 3.3).

Statistical AnalysisStatistical analysis was performed by using chi-squaretests with SPSS software (version 10.0; SPSS, Inc., Chi-cago, IL) to determine the differences of p53 codon 72genotypes among the study groups. P � 0.05 wasconsidered statistically significant.

RESULTSDNA amplification with microsatellite primers showedthat DNA from all specimens was suitable for PCRanalysis (data not shown). p53Arg homozygotes andp53 proline (p53Pro) homozygotes were observed asspecific bands with the expected sizes of 178 bp and142 bp, respectively. The heterozygotes were revealedby both bands (Fig. 2). Sequencing results showed thatthe PCR products were genuine (Fig. 3).

The distribution of p53 codon 72 genotypes insamples from 50 healthy blood donors from Beijingwere as follows: Ten of 50 samples (20%) were p53Arg

homozygotes, 14 of 50 samples (28%) were p53Pro ho-

FIGURE 1. Detection of p53 codon 72 polymorphism (C/G) by allele specific

polymerase chain reaction (PCR) analysis. The structure of the polymorphic site

in exon 4 and the primers used for PCR analysis are indicated by arrows. The

sequences are described in Materials and Methods.

p53 Polymorphisms and HPV in China/Li et al. 2573

mozygotes, and 26 of 50 samples (52%) were heterozy-gotes. The distribution of p53 codon 72 genotypes insamples from high-incidence and low-incidence pop-ulations for esophageal carcinoma as well as samplesof esophageal, ovarian, and breast carcinoma is sum-marized in Table 1. The distribution of the p53 codon72 polymorphism (C/G) showed no significant differ-ences between the Beijing population and the Anyangpopulation (chi-square test, 0.685; degrees of freedom

[df] � 2; P � 0.710). However, there were significantdifferences in the frequency of the Arg genotype be-tween HPV positive and HPV negative carcinoma sam-ples and between HPV positive carcinoma samplesand the control samples. The p53Arg homozygoteswere significantly higher in HPV positive carcinomasamples. The p53 codon 72 genotypes in HPV positivesamples of different histologic grades of esophagealmucosa and in HPV positive carcinoma samples areshown in Table 2. The distribution of the p53 codon72 polymorphism (C/G) among different histologicgrades of HPV positive esophageal epithelium showedno significant differences (chi-square test, 0.041; df� 2; P � 0.980). Similar results were found in HPVnegative samples. However, the frequency of p53Arg

homozygotes in HPV positive carcinoma samples wassignificantly higher compared with HPV positive sam-ples from normal esophageal epithelium (P � 0.003, P� 0.035, and P � 0.049, respectively).

DISCUSSIONThe p53 gene is one of the most intensely studiedhuman genes because of its role as a tumor suppressorgene.17 Mutations in p53 are found in over 50% of allhuman malignancies, indicating that there is a pow-erful selection for loss of p53 activity during tumordevelopment.17,18 Normally, p53 arrests the cell cycleat the G1/S check point when it detects genotoxicdamage of the DNA strand integrity, thus allowingDNA repair or apoptosis and avoiding replication of adamaged template. Mutation of p53 results in the lossof tumor suppressor function, which has been associ-ated with cell carcinogenesis.17 Wild type p53 is poly-morphic at codon 72 in human populations, and it hasbeen found that this polymorphism is a high-risk fac-tor for HPV-associated cervical carcinoma.7 In the cur-rent study, we used an allele specific PCR method toanalyze the p53 codon 72 (C/G) polymorphism inHPV-associated malignancies among the Chinesepopulation.

Table 1 shows that there were no significant dif-ferences in the distribution of p53 polymorphism be-tween the Beijing population and the Anyang popula-tion (chi-square test, 0.685; df � 2; P � 0.710). Becauseour previous studies indicated a high HPV prevalencerate in Anyang among the high-incidence populationfor esophageal carcinoma (HPV-16, 72%; HPV-18,17%),14 the results suggest that p53 codon 72 polymor-phism has no effect on HPV infection. However, thefrequency of p53Arg homozygotes in samples of esoph-ageal, ovarian, and breast carcinoma were signifi-cantly higher in HPV-positive samples (53%, 42%, and47%, respectively), whereas, in HPV negative samples,the p53Arg homozygotes comprised 26%, 23% and

FIGURE 2. Allele specific polymerase chain reaction (PCR) amplification of

p53 codon 72 polymorphism. Lanes 1 and 9: Marker (100 base-pair DNA

ladder; GIBCO BRL). Lanes 3 and 7: p53Pro homozygotes. Lanes 4 and 8: p53Arg

homozygotes. Lanes 2, 5, and 6: p53 heterozygotes.

FIGURE 3. Sequencing results of p53 polymerase chain reaction products.

(A) p53Pro sequenced by the reverse primer. (B) p53Arg sequenced by the

forward primer.

2574 CANCER December 15, 2002 / Volume 95 / Number 12

22%, respectively, similar the control group (P � 0.559,P � 1.000, and P � 1.000, respectively). Our resultsstrongly suggest that the p53Arg homozygote genotypeis one of the high-risk genetic factors for HPV-associ-ated malignancies among the Chinese population. Ithas been reported that these two polymorphic vari-ants of wild type p53 differ biochemically and biolog-ically.19 p53Arg is significantly more susceptible com-pared with p53Pro to the degradation induced by high-risk HPV E6 protein and the resulting loss of function.Normally, p53 acts as a genome guard. Loss of p53function may increase the probability of carcinogen-esis in HPV-infected cells. Accordingly, individualswho are homozygous for p53Arg are more susceptibleto carcinogenesis after HPV infection. Our data alsomay explain the controversial reports of the correla-tion of p53 codon 72 polymorphism with a suscepti-bility to carcinoma. Most of the previous studies did

not separate the HPV positive and negative sampleswhen p53 codon 72 polymorphism was checked andanalyzed.

In addition, our previous studies showed that HPVwas present in different histologic grades of esopha-geal epithelium from normal mucosa, to mild dyspla-sia, to severe dysplasia; and, with the progression ofhistologic grade, the HPV infection rate and viral on-cogene expression increased.14 According to Table 2,the distribution of p53 codon 72 polymorphism (C/G)among different histologic grades of HPV positiveesophageal epithelium showed no significant differ-ences (chi-square test, 0.041; df � 2; P � 0.980). Sim-ilar results were found in HPV negative samples. Theseresults suggest that p53 polymorphism may not berelated to an increased HPV infection rate or to theprogression of cytopathologic grade in esophageal ep-ithelium, implying that p53 polymorphism may have

TABLE 1Distribution of Three Genotypes of the p53 Gene at Codon 72 in Esophageal Carcinoma among High-Incidence and Low-Incidence Populationsand Patients with Squamous Cell Carcinoma of the Esophagus, Ovarian Carcinoma and Breast Carcinoma

Group No.

No. with genotype (%)

Chi-square P value 95%CIArg/Arg Arg/Pro Pro/Pro

High-incidence population 48 10 (21) 24 (50) 14 (29) — — 0.10–0.34Low-incidence population 33 9 (27) 17 (52) 7 (21) — — 0.15–0.49Beijing population 50 10 (20) 26 (52) 14 (28) 0.685 0.710 0.10–0.34Esophageal carcinoma 62 27 (43) 21 (34) 14 (23) 6.938 0.008 —

HPV positive 39 21 (53) 10 (26) 8 (21) 11.057 0.001 —HPV negative 23 6 (26) 11 (48) 6 (26) 0.341 0.559 —

Ovarian carcinoma 39 14 (36) 20 (51) 5 (13) 2.812 0.094 —HPV positive 26 11 (42) 12 (46) 3 (12) 4.257 0.039 —HPV negative 13 3 (23) 8 (62) 2 (15) 0.060 1.000 —

Breast carcinoma 28 11 (39) 10 (36) 7 (25) 3.393 0.065 —HPV positive 19 9 (47) 6 (32) 4 (21) 5.168 0.023 —HPV negative 9 2 (22) 5 (56) 2 (22) 0.023 1.000 —

95%CI: 95% confidence interval; HPV: human papilloma virus.

TABLE 2p53 Codon 72 Genotypes in Human Papillomavirus Positive Samples of Different Pathologic Types of Esophageal Mucosaand Ovarian and Breast Carcinoma

Pathologic types No.

No. with genotype (%)

Chi-square P value 95%CIArg/Arg Arg/Pro Pro/Pro

Normal mucosa 17 4 (24) 9 (52) 4 (24) 0.041 0.980 0.06–0.44Mild dysplasia 23 5 (22) 11 (48) 7 (30) — — 0.09–0.49Severe dysplasia 8 2 (25) 4 (50) 2 (25) — — 0.03–0.56Total 48 11 (23) 24 (50) 13 (27) — — —Esophageal carcinoma 39 21 (53) 10 (26) 8 (21) 8.852 0.003 —Ovarian carcinoma 26 11 (42) 12 (46) 3 (12) 2.005 0.035 —Breast carcinoma 19 9 (47) 6 (32) 4 (21) 3.887 0.049 —

95%CI: 95% confidence interval.

p53 Polymorphisms and HPV in China/Li et al. 2575

less importance in the susceptibility to carcinoma un-til the last step in carcinogenesis. In other words, thep53 codon 72 polymorphism (C/G) may be a factor forsusceptibility only in the very late stage of carcinogen-esis.

In summary, the current study suggests that thep53 codon 72 Arg homozygous genotype is one of thehigh-risk genetic factors for HPV-associated malig-nancies among the Chinese population. However, it isaccepted widely that many environmental factors andgenes are involved in the multiple steps of carcino-genesis, and the p53 gene may be only one of them.Thus, it will be important in future studies to investi-gate the germline polymorphism of other relatedgenes, for example, the genes in the p53 pathway.

REFERENCES1. Zur Hausen H. Papillomavirus infection—a major cause of

human cancers. Biochim Biophys Acta. 1996;1288:F55–F78.2. Zur Hausen H. Human papillomaviruses in the pathogene-

sis of human anogenital cancer. Virology. 1991;184:9 –13.3. Vousden KH, Doniger J, DiPaolo JA, Lowy DR. The E7 open

reading frame of human papillomavirus type 16 encodes atransforming gene. Oncogene Res. 1988;3:167–175.

4. Barbosa MS, Schlegel R. The E6 and E7 genes of HPV-18 aresufficient for inducing two-stage in vitro transforming ofhuman keratinocytes. Oncogene. 1989;4:1529 –1532.

5. Watanabe S, Kanda T, Yoshiike K. Human papillomavirustype 16 transformation of primary human embryonic fibro-blasts requires expression of open reading frames E6 and E7.J Virol. 1989;63:965–969.

6. Beckman G, Birgander R, Sjalander A, et al. Is p53 polymor-phism maintained by natural selection? Hum Hered. 1994;44:266 –270.

7. Storey A, Thomas M, Kalita A, et al. Role of a p53 polymor-phism in the development of human papillomavirus-asso-ciated cancer. Nature (London). 1998;393:229 –234.

8. Yamashita T, Yaginuma Y, Saitoh Y, et al. Codon 72 poly-morphism of p53 as a risk factor for patients with humanpapillomavirus-associated squamous intraepithelial lesions

and invasive cancer of the uterine cervix. Carcinogenesis.1999;20:1733–1736.

9. Szarka K, Veress G, Konya J, Gergely L. Frequency of p53codon 72 genotypes in human papillomavirus associatedsquamous intraepithelial lesions and cervical cancer. Anti-cancer Res. 1999;19:2377–2379.

10. Ngan HYS, Liu VWS, Liu SS. Risk of cervical cancer is notincreased in Chinese carrying homozygous arginine atcodon 72 of p53. Br J Cancer. 1999;80:1828 –1829.

11. Minaguchi T, Kanamori Y, Matsushima M, Yoshikawa H,Taketanl Y, Nakamura Y. No evidence of correlation be-tween polymorphism at codon 72 of p53 and risk of cervicalcancer in Japanese patients with human papillomavirus16/18 infection. Cancer Res. 1998;58:4585– 4586.

12. Kawaguchi H, Ohno S, Araki K, et al. p53 polymorphism inhuman papillomavirus-associated esophageal cancer. Can-cer Res. 2000;60:2753–2755.

13. Peixoto GD, Hsin LS, Snijders P, et al. Absence of associationbetween HPV DNA, TP53 codon 72 polymorphism, and riskof oesophageal cancer in a high-risk area of China. CancerLett. 2001;162:231–235.

14. Li T, Lu ZM, Chen KN, et al. Human papillomavirus type 16is an important infectious factor in the high incidence ofesophageal cancer in Anyang area of China. Carcinogenesis.2001;22:929 –934.

15. Guo M, Lu Z-M, Yao L-H, et al. HPV infection and itsrelationship with breast cancer with or without erb�-2 geneover expression. Chinese Med J. [url: www.cmj.org/16wl/guomei62.htm]

16. Lu ZM, Chen KN, Guo M, et al. Detection of HPV and EBV inhuman esophageal cancer in high-incident area and corre-lation with p53 expression. Chinese J Oncol. 2001;23:220 –223.

17. Levine A. p53, the cellular gatekeeper for growth and divi-sion. Cell. 1997;88:323–331.

18. Hollstein M, Rice K, Greenblatt M, et al. Database of p53gene somatic mutations in human tumours and cell lines.Nucleic Acids Res. 1994;22:3551–3555.

19. Thomas M, Kalita A, Labrecque S, Pim D, Banks L, Mat-lashewski G. Two polymorphic variants of wild-type p53differ biochemically and biologically. Mol Cell Biol. 1999;19:1092–1110.

2576 CANCER December 15, 2002 / Volume 95 / Number 12