By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
There is strong epidemiologic evidence indicating that human papillomavirus (HPV) plays a central role in the etiology of cervical cancer.1, 2, 3, 4 The women positive for HPV DNA have a risk of developing cervical cancer 15–50 times higher than those without HPV DNA.1, 2, 4, 5 Although HPV infection is common among young women, only a small minority go on to develop cervical cancer. This situation was reviewed by a group of researchers, who concluded that viral persistence of oncogenic HPV appears to be crucial for the development of cervical cancer.6 Indeed, HPV-16, -18, -31 and -33 have been officially declared to be oncogenic by the World Health Organization (WHO).7
The International Biological Study on Cervical Cancer (IBSCC) study group revealed that on average 92.9% of the tumors contained HPV DNA, with a range of 75–100%.8 Although this international survey included 10 of the 18 regions of the world and provided the most extensive global view of HPV in cervical cancer, there were no data from China, 1 of the largest populations in the world. Furthermore, there are few published data from a well-designed study using a quality-controlled method on the prevalence of either HPV infection or HPV genotypes in cervical cancers in China.
We conducted a multicenter study of HPV infection in cervical cancer by selecting 5 geographic regions of China: Shanghai (Eastern China), Guangzhou (Southern China), Sichuan (Western China), Beijing (Northern China) and Hong Kong (Specific Administration Region).
MATERIAL AND METHODS
Cervical cancer specimens were collected from each of the 5 regions in China, including Shanghai, Guangzhou, Sichuan, Beijing and Hong Kong. Each of the 5 settings (Tumor Hospital, Shanghai Medical University in Shanghai, Tumor Hospital, Sun Yat-sen University of Medical Sciences in Guangzhou, Second Hospital, West China Medical University in Sichuan, Tumor Hospital, Chinese Academy of Medical Sciences in Beijing, and Prince of Wales Hospital, The Chinese University of Hong Kong in Hong Kong) was responsible for collecting 150 or more tumor specimens consecutively from the year 1997 to 1999. Patients from Hong Kong were prospectively recruited with informed consent, and all specimens were freshly frozen. Those specimens collected in other parts of China were paraffin-embedded and retrieved retrospectively from the pathology files in those reference centers. Ethical approvals were obtained from each of the ethical review boards of the respective institutes.
All histologic slides submitted were reviewed by 1 of the investigators (M.K.M.C.) to reestablish the histologic type and grade of differentiation according to the criteria of the WHO. Equivocal and unusual specimens were examined jointly with a second investigator (Y.F.W.).
Preparation of DNA
DNA from paraffin-embedded tissue sections was extracted as previously described.9 Briefly, 5–15 10 μm thick sections were deparaffinized with organic solvents, and high molecular weight DNA was prepared from tumor specimens by proteinase K digestion and phenol/chloroform extraction. Microdissection was performed to enrich the tumor cells if the proportion of cancer cells in tissue section was less than 50%. Procedures to prevent specimen contamination and PCR carryover were observed at every step.10, 11 Specimens from Hong Kong were freshly frozen and confirmed histologically to contain no less than 50% tumor cells. Tumor tissue was also digested by proteinase K, and DNA was isolated after phenol/chloroform extraction and ethanol precipitation.
Detection and Typing of HPV
All HPV assays were done at the Gynecologic Cancer Research Laboratory, The Chinese University of Hong Kong. The following steps were used to ensure the highest probability of detecting HPV DNA (Fig. 1).
Use of a pair of degenerate MY09/11 consensus primers in the highly conserved L1 open reading frame of HPV yielded a 450 bp fragment.12 The PCR conditions were optimized. In brief, a 50 μl amplification mixture consisted of 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 100 μM of each dNTP, 0.25 μM of each primer, 2 units of Taq polymerase (GIBCO BRL, Life Technologies, Paisley, Scotland) and 1 μl of cervical cancer sample DNA (containing about 25 ng of DNA). Thirty-five cycles of amplification were carried out as follows: denaturation at 94°C for 30 sec, annealing at 55°C for 30 sec and extension at 72°C for 1 min. The initial denaturation was for 3 min at 94°C, and the final elongation step was extended to 10 min. The products of the reaction were electrophoresed on 2% agarose gel. The positive PCR products were subjected to direct DNA sequencing for HPV genotyping.
The integrity of DNA in negative specimens tested by PCR using the MY09/11 consensus primer was examined by a β-actin PCR assay, yielding a human β-actin product of 303 bp using a primer pair of 5′-GAAACTACCTTCAACTCCATC-3′ and 5′-CTAGAAGCATTTGCGGTGGACGATGGAGGGGCC-3′. This was also used to confirm that nonspecific inhibitors were absent.
Negative specimens of the β-actin PCR assay were excluded from the study, and positive specimens were further screened for HPV infection using HPV-16, -18, -31, -33, -52 and -58 specific PCR. For PCRs of HPV-16, -18, -31 and -33, we adopted the methods previously described by Baay and colleagues;13 for PCRs of HPV-52 and -58, we used the primer pair: 5′-GCATTCATAGCACTGCCAC-3′; and 5′-GCCTCTACTTCAAACCAGCC-3′, corresponding to positions of the sense strand 761–779 and antisense strand 909–928 and yielding a 168 bp amplimer, as well as the primer pair: 5′-CATGTACCATTGTGTGCCC-3′ and 5′-ACCGCTTCTACCTCAAACC-3′, resulting in a 118 bp amplimer corresponding to positions of the sense strand 833–851 and antisense strand 932–950, yielding a 118–bp amplimer. A 50 μl amplification mixture consisted of 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 100 μM of each dNTP, 0.25 μM of each primer, 2 units of Taq polymerase (GIBCO BRL, Life Technologies) and 1 μl of cervical cancer sample DNA. Thirty-five cycles of amplification were carried out as follows: denaturation at 94°C for 30 sec, annealing at 59°C for 30 sec and extension at 72°C for 1 min. The initial denaturation was 3 min at 94°C, and the final elongation step was extended to 10 min.
In negative specimens tested by the 6 genotypes of HPV type-specific PCR described above, HPV DNA was further examined by PCR using another pair of consensus primers, GP5+/6+, from the HPV L1 region, yielding a 110 bp fragment.14
The positive specimens found by using GP5+/6+ PCR were subjected to direct DNA sequencing again. If the direct sequencing was unsuccessful, the specimens were defined as unknown type HPV infection. Nevertheless, the negative specimens by GP5+/6+ PCR were defined as HPV infection negative.
In specimens in which direct DNA sequencing failed after the detection of HPV infection positivity by MY09/11 PCR, HPV typing using restriction fragment length polymorphism (RFLP) was performed.15 If the HPV typing was still unsuccessful, then steps 3–5 were repeated. Each PCR included a positive and negative control. Purified DNA obtained from plasmids or known type-specific positive cervical cancer DNA samples was used as positive controls, and HPV-negative DNA was from the K-562 cell line.
Log-linear models were fitted to the contingency table cross-classified by 4 factors: geographic regions, HPV type, histology and grade. The adequacy of the independence model was first addressed by a likelihood ratio test. The highest order of associations to be included in the model was also found. Finally, a parsimonious model was constructed by using backward elimination. Goodness-of-fit of the final model was assessed by the likelihood ratio test. Chi-squared tests were used to detect any geographic differences in HPV prevalence, geographic differences in cell differentiation and differences in prevalence of HPV types by histologic type. Two-sided p-values were calculated, and p < 0.05 was considered significant to take into account the multiple comparisons. Statistical analysis was performed using the Statistical Analysis System software package (SPSS 10.0, SPSS, Chicago, IL).
After exclusion of 54 cases with no detectable DNA (cases with negative β-actin PCR) of which 33 cases were from Shanghai, 20 from Sichuan and 1 from Beijing, a total of 809 cervical cancer specimens were collected from the 5 regions of China (Table I). The diagnosis of invasive cervical carcinoma was confirmed in all specimens, which consisted of 731 (90.4%) squamous cell carcinomas (SCCs), 74 (9.1%) adenocarcinomas (ACs) and 4 (0.5%) adenosquamous carcinomas (ASs). Based on the degree of nuclear and architectural abnormality and the proportion of cells exhibiting cytoplasmic differentiation, tumors were graded as well differentiated in 145 (17.9%), moderately differentiated in 427 (52.8%) and poorly differentiated in 237 (29.3%) cases.
Table I. HPV Prevalence by Geographic Regions Studied1
HPV, case nos. (% within regions)
HPV, human papillomavirus.
HPV DNA was detected in 677 (83.7%) of 809 specimens (Table I). The prevalence varied between 78.7% from Shanghai to 87.7% from Sichuan (p value for heterogeneity = 0.14). HPV-16 was present in 79.6% (n = 541), HPV-18 in 7.5% (n = 51), HPV-31 in 0.7% (n = 5), HPV-33 in 1.8% (n = 12), HPV-45 in 0.4% (n = 3), HPV-52 in 2.6% (n = 18) and HPV-58 in 3.8% (n = 26) of all HPV-positive specimens (n = 680, double-counted in 3 cases with double infection; Table II).Twenty-three (3.4%, 95% confidence interval [CI] = 2.3–5.2%) HPV-positive specimens harbored HPVs that could not be typed. Double infection occurred in 3 (0.8%, 95% CI = 0.2–2.2%) of 381 HPV-positive cases in which HPV DNA was not typed by direct DNA sequencing. The other 296 cases using direct sequencing were excluded from the calculation, as the process of HPV detection was stopped after 1 HPV type was found. Each of them consisted of 2 different viruses, namely, HPV-18 and -58; HPV-18 and -31; and HPV-18 and -52. HPV-16 and -18 remained the most frequently encountered and added up to a total of 87.1%. HPV-58 and -52 were found to be the third (3.8%) and fourth (2.6%) most common genotypes in our study population.
The associations among geographic region, HPV type, histologic type and grade of differentiation were explored using log-linear models. The highest order of association was found to be 2. The resulting log-linear model indicated that there were significant associations (p < 0.0005) between geographic regions and HPV types (Table II), geographic regions and grade of differentiation (Table III) and HPV type and histologic type (Table IV). The likelihood ratio test showed no evidence of poor fit to the data (p = 0.923).
Table III. Grade of Differentiation by Geographic Region Studied
The prevalence of HPV-58 and -52 was higher in southern regions (Guangzhou and Hong Kong) compared with the rest (Shanghai, Sichuan and Beijing) of China (HPV-52, 13/295 vs. 5/385, 4.4% vs. 1.3%, p = 0.02 and HPV 58, 17/295 vs. 9/385, 5.8% vs. 2.3%, p = 0.03). The prevalence rates of HPV-16 and -18 in Hong Kong were 61.7 and 14.8%, respectively, representing a lower HPV-16 (100/162 vs. 441/518, 61.7% vs. 85.1%, p < 0.00005) and a higher HPV-18 (24/162 vs. 27/518, 14.8% vs. 5.2%, p = 0.0001) proportion compared with the other regions (Table II). Among the HPV-positive cervical tumors, they were graded as well differentiated in 125 (18.5%), moderately differentiated in 356 (52.6%) and poorly differentiated in 196 (28.9%) cases (Table III). The proportion of poorly differentiated tumors was higher in Guangzhou (64/132 vs. 132/545, 48.5% vs. 24.2%, p < 0.00005) and lower in Beijing (13/122 vs. 183/555, 10.7% vs. 33.0%, p < 0.00005) compared with other regions. Among the HPV-positive cervical tumors, 623 (91.6%) were SCCs, 53 (7.8%) were ACs and 4 (0.6%) were ASs (Table IV). HPV-16 and -18 remained the most common HPV infection in both SCC and AC. The proportion of HPV-18 infection was significantly higher in AC than in SCC (15/53 vs. 34/623, 28.3% vs. 5.5%, p < 0.00005). There was no significant variation in types of HPV infection among different tumor grade (Table V).
Table V. HPV Type by Grade of Differentiation Studied
Grade of differentiation, case nos. (% within grade of differentiation)
The prevalence of HPV DNA is highly dependent on the detection method used. With sensitive PCR methods, detection of HPV genotypes in clinical specimens is usually achievable.16 These methods are capable of detecting HPV DNA at the single copy level even in an overwhelming background of human DNA. As the techniques of detecting viral DNA have improved, the reported prevalence among cervical cancer patients has increased dramatically in last decade.6, 17 The IBSCC study group revealed that on average, 92.9% of the tumors (range, 75–100%) contained HPV DNA.8 A recent reanalysis of those HPV-negative tumors, which included stringent control of histology and additional PCR-based assays, indicated that HPV is present in 99.7% of these cases,18 suggesting that HPV prevalence was previously underestimated.
In our series, we found that HPV DNA was present in 83.7% of cervical cancers of a Chinese population, similar to our previous results, which were confined to Hong Kong specimens collected before the year 1997.19 Whether this represents a true geographic variation or a detection problem is unclear. To minimize false-negative results, we centralized HPV detection in 1 research laboratory, and histologic review was conducted by 1 pathologist. We accepted an HPV-negative result only after microdissection of specimens to enrich the proportion of tumor cells.
Studies have shown that the overall prevalence of HPV is underestimated when a single method is used.13, 20 Various detection methods work in different ways and at different sites. The general consensus primers MY09/11 and GP5+/6+ target the L1 gene of the HPV genome, whereas type-specific primers target the E6 or E7 regions. Therefore, in addition to the most commonly used consensus primer MY09/11, 3 other detection methods were used in our study to ensure the highest possible detection of HPV DNA.
Among the 5 regions studied, there was no significant difference in geographic variation except in Hong Kong. HPV-16 and -18 contributed an average of 85 and 5%, respectively in the other 4 regions. However, we found HPV DNA in 83.1% patients in Hong Kong, including HPV-16 in 61.7% and HPV-18 in 14.8% patients and this was consistent with our previous result.19 The results in Hong Kong were similar to the worldwide figures reported by IBSCC study group:8: HPV-16 was by far the most frequently found virus genotype and accounted for 50–60% of all positive cases, with HPV-18 accounting for 10–20%. This may be a reflection of the fact that Hong Kong is the most international city of China and as such may have been exposed to the same worldwide viruses.
HPV-58 and -52 were found to be the third and fourth most common genotypes in our study; they are relatively uncommon in cervical cancers in the Americas, Europe, Africa and Southeast Asia.8 Their prevalence was higher in Guangzhou and Hong Kong. This is consistent with previous small studies indicating that HPV-58 and -52 were relatively more prevalent among Chinese populations in Taiwan and Shanghai.21, 22 HPV-58 and -52 were shown to be closely related to and grouped together with HPV-16, -31, -33, -35 and -67 under the same branch of an HPV phylogenetic tree, indicating similarities in their pathogenic potential.23 According to the multinational case-control study conducted by Munoz et al.,24 HPV-16 had an odds ratio (OR) for cervical cancer of 182 compared with controls. HPV-58 and -52 have ORs of 79 and 146, respectively.24 Our findings suggest that HPV-58 and -52 may also play a significant role in the etiology of cervical cancer.
We found mixed infections in only 0.8% of cases. This is lower than the rate of 4% detected in the worldwide study. As there is almost no cross-reactivity between the various genotypes, the development of vaccines and diagnostic probes may have to be genotype-specific. The finding of low prevalence of mixed infections may have some implications for both screening with combination genotype-specific diagnostic HPV probes and polyvalent vaccination program development. However, current detection methods were not optimized for the detection of double-positives.
Current data regarding the relationship between HPV-18 and histologic type are in line with those reported by others.25, 26 Although HPV-16 remained the most common HPV infection in both SCC and AC, our data demonstrate that HPV-18 has a significant association with AC. This finding may reflect the fact that different virus receptors exist in cervical cells with different morphologic potentials, or it may indicate that specific HPV infection actually plays a role in directing carcinogenesis.
The mainstay of cervical cancer prevention at present is cytologic screening. However, vaccination against HPV infections that are known to be oncogenic appears to be a viable option, although its efficacy remains to be explored in a large population. Vaccination against HPV may be particularly useful in Mainland China, where it is proving difficult to implement effective screening programs in this huge population. This is particularly so in the rural areas where most of the population lives. Vaccination may be more cost-effective in the longer term compared with the screening performed at present. However, screening, eventually at longer intervals, will remain necessary to detect those patients who do not respond to vaccination, those HPV infections caused by types not covered by available vaccines and HPV-negative cases.
There is ongoing research into the possibility of vaccination against HPV type-specific infection in the future27 and also into immunotherapeutic approachs toward disease control.28, 29 Knowledge of the HPV types that are commonly associated with cervical cancers in China allows more precise selection of HPV vaccine development. Such study is crucial for informing policy decisions pertinent to the development of genotype-specific HPV diagnostic probes for and vaccination strategies against HPV infection in Mainland China.
Profs. H. zur Hausen and E.-M. de Villiers at the German Cancer Research Institute in Heidelberg gave valuable advice to ensure optimization and validation of the methodology of our study. We are also grateful to Prof. W.D. Lancaster, Geogetown University Medical Center, Washington, DC, and Prof. T. Matsukura for kindly providing HPV-52 and -58 plasmids, respectively. We thank the Gynecologic Cancer Research Laboratory and Lee Hysan Clinical Research Laboratories, The Chinese University of Hong Kong, for providing reagents and equipment.