- Email: [email protected]
The value of HPV DNA typing in the distinction between adenocarcinoma of endocervical and endometrial origin
Pathology (October 2003) 35(5), pp. 397–401
ANATOMICAL PATHOLOGY
The value of HPV DNA typing in the distinction between adenocarcinoma of endocervical and endometrial origin MYFANWY K. PLUNKETT*, BRIAN BRESTOVAC{, JANE E. THOMPSON{, GREGORY F. STERRETT*, PIERRE R. FILION*, DAVID W. SMITH{ AND FELICITY A. FROST* Departments of *Tissue Pathology and {Division of Microbiology and Infectious Diseases, The Western Australian Centre for Pathology and Medical Research (PathCentre) Nedlands, Western Australia and {Tissue Pathology, King Edward Memorial Hospital, Subiaco, Western Australia
Summary Aims: Distinguishing between adenocarcinomas of endocervical and endometrial origin histologically can be difficult, particularly in small biopsies. Most endocervical adenocarcinomas contain human papillomavirus (HPV) deoxyribonucleic acid (DNA) of ‘high-risk’ (HR) types, whereas this has not been consistently demonstrated in endometrial adenocarcinomas. The aim of this study was to determine whether HPV DNA testing could aid in this differential diagnosis. Methods: The frequency of HPV DNA in paraffin-embedded tissue samples from 50 endocervical and 50 endometrial adenocarcinomas was investigated using polymerase chain reaction (PCR) amplification techniques involving (i) a screening HPV test followed by HPV DNA sequencing, and (ii) a test designed to detect HR genotypes 16, 18, 31, 33, 35, 45 and 58. Control specimens included cervical intraepithelial neoplasia (CIN) III lesions, squamous cell carcinomas (SCCs) of the cervix and lung, and colonic adenocarcinomas. Measures to minimise cross-contamination were implemented. Results: The screening test followed by HPV DNA sequencing had the highest sensitivity. By this test HR HPV DNA was detected in 11 of 11 (100%) cervical intraepithelial neoplasia (CIN III) lesions, nine of 10 (90%) cervical SCCs, none of 10 (0%) colorectal adenocarcinomas and none of 10 (0%) SCCs of the lung. Thirty-nine (78%) endocervical adenocarcinomas contained HR HPV DNA, compared to one (2.0%) endometrial adenocarcinoma. Conclusions: The results suggest that HPV DNA testing could be a useful adjunct in distinguishing between endocervical and endometrial adenocarcinomas in curettings or small biopsy specimens. Key words: Endocervical adenocarcinoma, endometrial adenocarcinoma, human papillomavirus, polymerase chain reaction, HPV DNA sequencing, high-risk HPV DNA. Abbreviations: CIN, cervical intraepithelial neoplasia; DNA, deoxyribonucleic acid; HR, high risk; HPV, human papillomavirus; PCR, polymerase chain reaction; SCC, squamous cell carcinoma. Received 26 February, revised 23 June, accepted 1 July 2003
INTRODUCTION The histological distinction between adenocarcinomas of endocervical and endometrial origin can be difficult, particularly in small biopsy and curettage specimens where minimal material is available for examination. Adenocarcinomas in both sites show considerable overlap in terms of histological appearance, with up to 30% of endocervical adenocarcinomas showing an endometrioid morphology. In addition, endometrial adenocarcinoma can secondarily involve the cervix providing further diagnostic difficulty. Mucin histochemistry and immunohistochemistry may be helpful, but are not always conclusive. The distinction between these adenocarcinomas may be clinically significant in terms of prognosis and subsequent patient management.1–3 Invasive carcinomas and high-grade intraepithelial precursor lesions of the cervix have been shown to contain human papillomavirus (HPV) DNA and HPV is an accepted cause of cervical squamous carcinoma.4 Endocervical adenocarcinomas have also been shown to contain human papilloma virus (HPV) DNA of certain ‘high-risk’ (HR) genotypes, with in situ studies confirming the presence of viral DNA in tumour cells.5,6 The reported frequency of HPV detection has varied greatly depending on the molecular techniques used.7 However, in the larger and more recent series performed using the sensitive polymerase chain reaction (PCR) amplification technique, the majority of endocervical adenocarcinomas have been shown to be HR HPV DNA positive.3,8–14 In contrast, the presence of HR HPV DNA has not been consistently demonstrated in endometrial adenocarcinomas. Although occasional studies have shown a relatively high rate of positivity,15–18 the majority of recent studies have reported significantly lower rates of HPV infection in endometrial adenocarcinomas compared to that reported in endocervical adenocarcinomas.3,19–23 Although some studies have used frozen tumour tissue for PCR analysis,12,14,15,23 most have been performed on sections cut from archival paraffin-embedded tissue.3,8–11,13,17,19,20,24–30 The ability of the PCR technique to allow highly specific and sensitive HPV DNA detection in archival paraffin-embedded material has previously been demonstrated31 and PCR techniques for the identification
ISSN 0031-3025 printed/ISSN 1465–3931 online/03/050397–05 # 2003 Royal College of Pathologists of Australasia DOI: 10.1080/00313020310001602611
398
Pathology (2003), 35(5), October
PLUNKETT et al.
of HPV DNA in cell samples have previously been validated in our laboratory.32 The aims of the current study were to examine the frequency of HR HPV DNA positivity in a series of endocervical and endometrial adenocarcinomas and adenosquamous carcinomas using PCR amplification techniques on paraffin-embedded material from small biopsy and curettage specimens, and to determine whether the differences in frequency were great enough to allow HPV DNA testing by PCR to be a useful adjunct in this differential diagnosis.
METHODS Selection of test cases Fifty endocervical adenocarcinomas and adenosquamous carcinomas and 50 endometrial adenocarcinomas and adenosquamous carcinomas were selected from the files of the histopathology departments of PathCentre (Nedlands, WA) and King Edward Memorial Hospital (Subiaco, WA), and included material reported between January 1992 and May 2001. Patient demographic data were obtained from the original pathology report. All of the endometrial adenocarcinomas were curettage specimens from cases with subsequent definitive hysterectomy diagnoses. The endocervical adenocarcinomas included seven endocervical curetting specimens (all with hysterectomy diagnoses), one endometrial curettage specimen, 13 cervical biopsies (eight with hysterectomy diagnoses and one with cone biopsy diagnosis), 15 large loop excisions of the transformation zone (LLETZ) or cone biopsy specimens (10 with hysterectomy diagnoses) and 14 hysterectomy specimens for which no previous small biopsy or curettage specimen was available. All biopsy specimens and hysterectomies were reviewed by a gynaecological pathologist. Tumours were graded according to World Health Organisation (WHO) guidelines,33 on the definitive surgical specimen, where these were available. Histological subtyping was also performed according to WHO guidelines.33 A tumour was considered ‘mixed’ if the second component comprised more than 10% of the total tumour. Cases were excluded if there was minimal material in tissue blocks, if cervical squamous dysplasia was present in the sections together with the adenocarcinoma, or if there were only microscopic foci of adenocarcinoma. Cases were also excluded if definitive endocervical or endometrial location had not been discerned macroscopically or microscopically. One representative block was selected from each case for the PCR studies. Selection of control cases The positive control cases were cervical biopsy and LLETZ or cone biopsy specimens from 10 CIN III lesions and 10 invasive squamous cell carcinomas (SCCs) of the cervix. The negative control cases were 10 SCCs of the lung and 10 colorectal adenocarcinomas. Specimen processing and PCR technique Three 20-mm sections were cut from each paraffin block and placed directly into Eppendorf tubes. The material was submitted for PCR analysis without knowledge of the tissue diagnosis. An initial HPV screening test was performed and used a nested PCR technique with primers targeting the L1 region of HPV. The outer MYO9 and MY11 primers were designed by Manos et al.34 using magnesium chloride at a concentration of 2.0 nM and an annealing temperature at 50‡C resulting in a product size of 450 base pairs. The inner primers, GP5z and GP6z, were designed by deRhoda Husman et al.,35 using magnesium chloride at a concentration of 2.0 nM and an annealing temperature at 45‡C, resulting in a product of 150 base pairs. The test detects episomal DNA and cannot detect integrated HPV DNA because the L1 region is lost on integration.
HR HPV testing was performed using a protocol designed to detect HPV genotypes 16, 18, 31, 33, 35, 45 and 58. For this test the primers targeted the E1–7 border regions, and were designed by Pizzighella et al.36 The outer primers were 3s and 1as (product size 670 base pairs), with inner primers 2s and 2as (product size 161 base pairs). Both outer and inner PCRs used a magnesium chloride concentration of 2.0 nM and an annealing temperature of 55‡C. The test should detect both episomal and integrated HPV DNA because the early region is kept on integration. For all PCR tests, 45 cycles were performed with the hot start Taq Polymerase, Taq Gold (Applied Biosystems). A positive control (HPV 18 extract) was used for each PCR ‘run’, with water controls performed after each five cases to ensure no contamination had occurred. PCR products were run on 2.5% ethidium bromide gel and photographed. Duplicate tests were performed for all cases and PCR photographs were read by two scientists independently with no disagreement.
Initial validation of PCR techniques DNA integrity was tested by amplification of endogenous retroviral PCR, which was amplified in all cases. For both nested PCR reactions the sensitivity was approximately 50 copies of HPV 18, as detected by serial dilutions of HeLa cell lines. The sensitivity of the PCR technique for clinical use has previously been validated against a cohort of 304 routine Pap smears with HR HPV detected in 15 of 16 (93.8%) ThinPrep1 (Cytyc Corporation, Boxborough, Massachusetts) samples from biopsy-proven high-grade cervical squamous intra-epithelial lesions. HR HPV was detected in 12 of 227 (5.3%) negative smears.32 HPV-DNA sequencing In control and test cases where HPV DNA was detected on the initial screening PCR, subsequent HPV DNA genotyping was performed. DNA sequencing was performed on the first round product of the screening PCR using the MYO 9 primer as the sequencing primer. However, if the PCR product was weak (low amount of sequencing target), the second round product was used, and in these cases GP5z was used as the sequencing primer. The DNA sequencing was performed using the Applied Biosystems BigDye Terminator V 1.0 in accordance with manufacturer’s instructions. Procedures to minimise cross-contamination during section cutting The material was cut in batches of 10–20 by several laboratory scientists. Each block was cut at a different point along the cutting edge of the microtome blade, with the blade and forceps wiped clean after each case. During cutting of the test cases, extra cleaning of the microtome blades using xylene was performed. Statistical analysis The significance of differences was tested using the x2 goodness of fit test. All tests were for two-tailed difference, with a type one error set at a level of 95% (P~0.05).
RESULTS For the endocervical adenocarcinomas and adenosquamous carcinomas, the patient’s ages ranged from 24 to 80 years (mean 46.6), and for the endometrial adenocarcinomas and adenosquamous carcinomas the ages ranged from 37 to 86 years (mean 68.2). Of the endocervical adenocarcinomas and adenosquamous carcinomas, 24 (48%) were well differentiated, 17 (34%) were moderately differentiated and nine (18%) were poorly differentiated. Of the endometrial adenocarcinomas, 13
HPV DNA IN FEMALE GENITAL ADENOCARCINOMAS
DISCUSSION
(26%) were well differentiated (grade I), 23 (46%) were moderately differentiated (grade II) and 14 (28%) were poorly differentiated (grade III). The results of the screening HPV DNA and HR HPV DNA tests and subsequent HPV DNA sequencing in the control cases are shown in Table 1. HR HPV DNA was detected in 100% of CIN III lesions and 90% of invasive squamous cell carcinomas of the cervix. Only one of the 20 negative controls, a SCC of lung, was found to contain HPV DNA, of a non-HR type. The results of the screening HPV DNA and HR HPV DNA tests and subsequent sequencing in the test cases are shown in Table 2, together with tumour typing. The screening HPV test was positive in 78% of endocervical adenocarcinomas, whereas HR HPV DNA testing was positive in only 54% of cases. All of the screening positive, HR-negative cases were found to contain HR HPV DNA by sequencing. The HPV DNA types detected by HPV DNA sequencing included genotype 18 (20 cases), genotype 16 (14 cases), and genotype 45 (five cases). The rates of HR HPV DNA positivity for well, moderately and poorly differentiated endocervical adenocarcinomas were 83.3, 76.5 and 66.7%, respectively (P~0.5785). The mean age of patients with HR HPV-positive endocervical carcinomas was 45.3 years compared to 50.8 years in patients with HPV DNA-negative tumours (P~0.5785). Only one (2.0%) endometrial adenocarcinoma was shown to be HPV-positive (Grade II endometrioid adenocarcinoma), which was HR HPV subtype 16 on HPV DNA subtyping.
TABLE 1
399
The consistent results in the positive and negative control cases provide further validation for the PCR techniques used in our laboratory. The finding of a single non-HR HPV DNA-positive squamous cell carcinoma of the lung is consistent with previous reports of a low rate of positivity in lung cancer.37 There was a significant difference between the rate of HR HPV DNA positivity in endocervical and endometrial carcinomas. Whilst 78.0% of endocervical carcinomas were shown to contain HR HPV DNA, only one (2.0%) endometrial adenocarcinoma was positive (P~v0.001). A positive HPV DNA result therefore makes endometrial adenocarcinoma or adenosquamous carcinoma very unlikely. A negative result by our methods, however, has less significance since HR HPV DNA was not detected in 22.0% of endocervical adenocarcinomas and adenosquamous carcinomas. The detection rate in endocervical adenocarcinomas is similar to that seen in the larger and more recent studies using the PCR technique, with reported frequencies of HPV DNA positivity ranging from 64 to 95%.3,8–14 A number of reasons have been suggested for these differences. PCR techniques may have variable sensitivities. Some PCR tests have been designed to detect only highrisk HPV types 16 and 18,8,13 whereas other studies, including the current study, detected a wider range of HR HPV DNA types.3,9–12,14,38 Geographical variation in levels of HPV infection has also been suggested.16 Earlier studies using various molecular techniques showed a much wider range of HPV positivity, reflecting the variable
Results of HPV DNA testing in control cases.
Diagnosis
HPV DNA zve by HPV screening test
Total
CIN III Invasive SCC cervix Colorectal adenocarcinoma SCC lung
11 10 10 10
11 9 0 1
(100%) (90%) (0%) (10%)
HR* HPV DNA zve by HR* HPV test 11 7 0 0
(100%) (70%) (0%) (0%)
HR* HPV DNA zve by sequencing of screening test zve cases 11 9 0 0
(100%) (90%) (0%) (0%)
*HR, high risk.
TABLE 2
Results of HPV DNA testing in endocervical and endometrial adenocarcinomas and adenosquamous carcinomas by histological subtype.
Diagnosis
Total
Endocervical carcinomas Mucinous (endocervical and intestinal) Well-differentiated villoglandular Endometrioid Adenosquamous Total
32 8 5 5 50
Endometrial carcinomas Endometrioid Villoglandular Serous papillary Mixed adenocarcinomas** Adenosquamous Total
38 3 1 7 1 50
HPV DNA zve by HPV screening test
24 7 4 4 39
(75.0%) (87.5%) (80.0%) (80.0%) (78.0%)
1 (2.6%) 0 0 0 0 1 (2.0%)
HR* HPV DNA zve by HR HPV test
18 5 3 1 27
(56.3%) (62.5%) (60.0%) (20.0%) (54.0%)
0 (0%) 0 0 0 0 0 (0.0%)
HR* HPV DNA zve by sequencing of screening test zve cases
24 7 4 4 39
(75.0%) (78.5%) (80.0%) (80.0%) (78.0%)
1 (2.6%) 0 0 0 0 1 (2.0%)
HR*, high risk; **the mixed adenocarcinomas included endometrioid/clear cell (two), serous papillary/clear cell (one), and endometrioid/clear cell (four).
400
PLUNKETT et al.
sensitivities of the different techniques as well as variation in sample size.7,8,38 Our low rate of detection of HR HPV DNA in endometrial adenocarcinomas is comparable to the majority of recent reports in the literature.3,19–23,25 In particular the findings are very similar to those found by Hording et al.3 in 1997, a study with a very similar methodology to our own, with similar numbers of endometrial and endocervical adenocarcinomas examined. However, the endometrial adenocarcinomas used in the study by Hording et al.3 were of surgical stage 2 (with cervical extension), specifically selected to simulate the situation where differential diagnostic problems arise. They showed none of the endometrial adenocarcinomas studied to contain HPV DNA, whilst detecting the virus in 70% of endocervical adenocarcinomas, similar to that seen in the present study. Nevertheless, occasional studies have shown higher rates of HPV DNA positivity.15–18 One of the highest levels of detection was recorded in a 1994 study by Lai et al.15 who found high-risk HPV DNA in 61.1% of endometrial adenocarcinomas by reverse transcriptase PCR, but found the DNA to be transcriptionally inactive. In some of these studies, HPV DNA was also detected in benign endometrial lesions18 and in benign and malignant ovarian tissue.15,17 These authors did comment that the amount of HPV-DNA detected was always low compared to most cervical and vulvar lesions.18 Differences in sample size and geographic variation have again been suggested as reasons for the varying results.16 Contamination from cervical or laboratory sources is a potentially important cause of false-positive results in PCR-based studies, particularly those using archival tissue. Contamination from adjacent squamous epithelium and especially CIN III lesions has been suggested as a possible mechanism for false-positive results in endocervical adenocarcinomas. In our study we ensured that no adjacent squamous intraepithelial lesion was present in the slides together with the adenocarcinoma. Some researchers have tried to prevent contamination by trimming away nontumorous tissue on sections cut from paraffin blocks.16 Other workers have used sterile scalpel blades to microdissect the cervical adenocarcinoma from adjacent squamous epithelium and stroma either on glass slides or directly from the paraffin block.7,9,10,38 The benign cervical stroma from the same slide has also been used as a negative control.9 The presence of the same HPV type in paraffin tissue blocks from histologically normal cervical tissue and endometrial adenocarcinoma tissue from the same patient has been noted, but might be used to argue either for contamination, or a consistent finding validating the study.19,20 Minimising cross-contamination of specimens during section cutting has been addressed in different ways. Some workers cleaned the microtome thoroughly with alcohol or xylene after cutting of each block and used new surgical gloves for each case.11,27 Others thoroughly cleaned17,24 or changed blades between cases.9 Some adopted even stricter anticontamination protocols with ultraviolet irradiation of all work surfaces and microtome blades between cases.19 We achieved good results despite relatively limited measures to prevent cross-contamination. Although sections from the positive and negative controls were cut at the same time, using the same microtome, HPV DNA was
Pathology (2003), 35(5), October
detected in only one negative control. Similarly, the endocervical and endometrial adenocarcinomas were cut in mixed batches, with only one endometrial adenocarcinoma found to contain HPV DNA. These findings suggest the technique used could be applicable to the routine diagnostic histopathology laboratory. We attempted to find a relationship between the HPV DNA status of tumours and patient age, tumour grade and tumour subtype. In our study, HPV-positive tumours were more common in younger patients (mean age 45.3 versus 50.8 years) but this was not statistically significant (P~0.5785). Some studies have demonstrated a significant correlation between younger age of the patient and HPV DNA positivity in endocervical adenocarcinoma;3,10,13 however, this has not been demonstrated in other papers.7,9,37 We found that the frequency of HPV DNA positivity in endocervical adenocarcinomas and adenosquamous carcinomas decreased with increasing tumour grade, but this did not reach statistical significance (P~0.5785). Other studies have demonstrated a similar correlation with more advanced clinical stage.10,13 However, the majority of studies have found no such correlation between tumour grade or stage and HPV DNA positivity.7,9,16 No significant correlation with histological subtype was demonstrated in our study, although only very small numbers of serous papillary and clear cell adenocarcinomas were included. Several studies have found a significantly higher proportion of mucinous tumours to be HPV DNA-positive compared to non-mucinous tumours, including clear cell tumours.8–10,13 High rates of positivity have been previously demonstrated in adenosquamous carcinomas.9,25,27 Some previous reports have suggested HPV 18 to be the most commonly identified HPV type in endocervical adenocarcinomas, leading to the assumption that HPV 18 may be more capable of infecting and transforming glandular epithelial cells than HPV 16, which is more commonly associated with squamous cell carcinoma.6,7 However, not all studies have found this correlation.8,10 We reported HPV 18 in 54.5% of all HPV DNA types detected in adenocarcinomas. In this study we used two separate HPV DNA PCR tests, one a ‘screening’ test and the other designed to detect a panel of HR HPV DNA. The HR HPV DNA test failed to detect HR HPV DNA in a number of the test cases that were positive by the screening HPV test. By sequencing these cases, HPV types 16, 18 or 45 were detected in all screening positive, HR-negative cases. This represents a significant failure of the HR HPV test. This may be due to degradation of the DNA during formalin fixation, possibly making PCR, particularly of greater than 200 base pairs, unreliable. Low levels of PCR inhibitors, related to DNA extraction from paraffin-embedded tissue, may have affected the HR HPV assay more than the PCR screening assay. However, testing for DNA integrity did show all of the test and control cases to amplify endogenous retroviral DNA. Improving the sensitivity of the HR HPV test in paraffin-embedded tissue would significantly improve the efficiency of the testing process, eliminating the need for DNA sequencing and possibly for the first HPV DNA PCR screening test. HPV DNA typing by the PCR tests and DNA sequencing methods used here, may be useful in the distinction between adenocarcinomas of endocervical and endometrial origin in
HPV DNA IN FEMALE GENITAL ADENOCARCINOMAS
small biopsy and curetting specimens. A positive result for HR HPV DNA makes endometrial adenocarcinoma very unlikely. Good results were achieved despite limited measures to prevent cross-contamination, suggesting that the methods would be applicable in the routine diagnostic histopathology laboratory. Further efforts to improve the sensitivity of PCR-based detection of HR HPV DNA in paraffin-embedded tissue from endocervical carcinomas will be worthwhile. ACKNOWLEDGEMENTS This research was funded by a grant from the Education and Research fund of The Western Australian Centre for Pathology and Medical Research (PathCentre), Nedlands, Western Australia. The authors gratefully acknowledge the assistance of the PathCentre and King Edward Memorial Hospital laboratory staff, and Pierre Filion, Senior Scientist, Electron Microscopy Unit, for help with the statistical analysis. Address for correspondence: Dr M. Plunkett, Histopathology Department, PathCentre, Hospital Avenue, Nedlands, 6009 WA, Australia. E-mail: littlefi[email protected]
17.
18.
19.
20.
21.
22.
23.
24.
25.
References 1. Zaino RJ. The fruits of our labours: distinguishing endometrial from endocervical adenocarcinoma. Int J Gynecol Pathol 2002; 21(1): 1–3. 2. Dabbs DJ, Sturtz K, Zaino RJ. The immunohistochemical discrimination of endometrioid adenocarcinomas. Hum Pathol 1996; 27(2): 172–7. 3. Hording U, Daugaard S, Visfeldt J. Adenocarcinoma of the cervix and adenocarcinoma of the endometrium: distinction with PCRmediated detection of HPV DNA. APMIS 1997; 105: 313–6. 4. zur Hausen H. Human papillomaviruses in the pathogenesis of anogenital cancer. Virology 1991; 184(1): 9–13. 5. King LA, Tase T, Twiggs LB, et al. Prognostic significance of the presence of human papillomavirus DNA in patients with invasive carcinoma of the cervix. Cancer 1989; 63: 897–900. 6. Tase T, Okagaki T, Clark BA, et al. Human papillomavirus types and localization in adenocarcinoma and adenosquamous carcinoma of the uterine cervix: a study by in situ DNA hybridisation. Cancer Res 1988; 48: 993–8. 7. Duggan MA, Benoit JL, McGregor SE, Nation JG, Inoue M, Stuart GCE. The human papillomavirus status of 114 endocervical adenocarcinoma cases by dot blot hybridization. Hum Pathol 1993; 24(2): 121–5. 8. Tenti P, Romagnoli S, Silini E, et al. Human Papillomavirus types 16 and 18 infection in infiltrating adenocarcinoma of the cervix. PCR analysis of 138 cases and correlation with histological type and grade. Am J Clin Pathol 1996; 106(1): 52–6. 9. Pirog EC, Kleter B, Olgnac S, et al. Prevalence of human papillomavirus DNA in different histological subtypes of cervical adenocarcinoma. Am J Pathol 2000; 157: 1055–62. 10. Milde-Langosch K, Schreiber C, Becker G, Loning T, Stegner HE. Human papillomavirus detection in cervical adenocarcinoma by polymerase chain reaction. Hum Pathol 1993; 24: 590–4. 11. van Muyden RCPA, ter Harmsel BWA, Smedts FMM, et al. Detection and typing of human papillomavirus in cervical carcinomas in Russian women. Cancer 1999; 85(9): 2011–6. 12. Bosch FX, Manos MM, Munoz N, et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. J Natl Cancer Inst 1995; 87(11): 796–802. 13. Tenti P, Romagnoli S, Silini E, et al. Analysis and clinical implications of K-ras gene mutations and infection with human papillomavirus types 16 and 18 in primary adenocarcinomas of the uterine cervix. Int J Cancer 1995; 64: 9–13. 14. Chen S, O’Sullivan H, Tabrizi SN, Fairley CK, Quinn MA, Garland SM. Prevalence and genotyping of HPV in cervical cancer among Australian women. Int J Gynecol Obstet 1999; 67: 163–8. 15. Lai C, Wang C, Lin C, Pao CC. Detection of human papillomavirus RNA in ovarian and endometrial carcinomas by reverse transcription/ polymerase chain reaction. Gynecol Obstet Invest 1994; 38: 276–80. 16. Fujita M, Shroyer KR, Markham NE, et al. Association of human
26.
27.
28. 29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
401
papillomavirus with malignant and premalignant lesions of the uterine endometrium. Hum Pathol 1995; 26(6): 650–8. Lai C-H, Hsueh S, Lin C-Y, et al. Human papillomavirus in benign and malignant ovarian and endometrial tissues. Int J Gynecol Pathol 1992; 11: 210–5. Milde-Langosch K, Becker G, Loning T. Human papillomavirus and c-myc/c-erbB2 in uterine and vulvar lesions. Virchow’s Arch A Pathol Anat 1991; 419: 479–85. O’Leary JJ, Landers RJ, Crowley M, et al. Human papillomavirus and mixed epithelial tumours of the endometrium. Hum Pathol 1998; 29(4): 383–9. Wong WSF, Wong YF, Tam OSP, Tam JSL. Detection of human papilloma virus (HPV) infection in paraffin-embedded tissues of endometrial carcinoma. Aust NZ J Obstet Gynaecol 1993; 33(2): 180–2. Bergeron C, Shah K, Daniel R, Ferenczy A. Search for human papillomaviruses in normal, hyperplastic, and neoplastic endometria. Obstet Gynecol 1988; 72:3: 383–7. Czerwenka K, Lu Y, Heuss F, Manavi M, Kubista E. Human papillomavirus detection in endometriod carcinoma with squamous differentiation of the uterine corpus. Gynecol Oncol 1996; 61: 210–4. Brewster WR, Monk BJ, Burger RA, Bergen S, Wilczynski SP. Does human papillomavirus have a role in cancers of the uterine corpus? Gynecol Oncol 1999; 75: 51–4. Anciaux D, Lawrence WD, Gregoire L. Glandular lesions of the uterine cervix: prognostic implications of human papillomavirus status. Int J Gynecol Pathol 1997; 16: 103–10. Hachisuga T, Matsuto N, Iwasaka T, Sugimori H, Tsuneyoshi M. Human papilloma virus and P53 overexpression in carcinomas of the uterine cervix, lower uterine segment and endometrium. Pathology 1996; 28: 28–31. Koulos JP, Wright TC, Follen Mitchell M, et al. Relationships between c-Ki-ras Mutations, HPV types and prognostic indicators in invasive endocervical adenocarcinomas. Gynecol Oncol 1993; 48: 364–9. Yamakawa Y, Forslund O, Teshima H, Hasumi K, Kitagawa T, Hansson BG. Human papillomavirus DNA in adenocarcinoma and adenosquamous carcinoma of the uterine cervix detected by polymerase chain reaction (PCR). Gynecol Oncol 1994; 53: 190–5. Parker MF, Arroyo GF, Geradts J, et al. Molecular characterization of adenocarcinoma of the cervix. Gynecol Oncol 1997; 64: 242–51. Lee MF, Chang MC, Wu CH. Detection of human papillomavirus types in cervical adenocarcinoma by the polymerase chain reaction. Int J Gynecol Obstet 1998; 63: 265–70. Claas ECJ, Melcher WJG, van der Linden HC, Linderman J, Quint WGV. Human papillomavirus detection in paraffin-embedded cervical carcinomas and metastases of the carcinomas by the polymerase chain reaction. Am J Pathol 1989; 135(4): 703–9. Resnick RM, Cornelissen MTE, Wright DK, et al. Detection and typing of human papillomavirus in archival cervical cancer specimens by DNA amplification with consensus primers. J Natl Cancer Inst 1990; 82: 1477–84. Choo C, Rankine C, Williams V, Brestovac B, Smith D, Frost FA. Correlation of high-risk HPV DNA testing with cervical cytology in 304 ThinPrep split samples. Poster Display, Annual Scientific Meeting, Australian Society of Cytology, Broome, June 2000. Scully RE, Banfiglio TA, Kurman RJ, et al. Histological Typing of Female Genital tract Tumours. In: WHO International Histological Classification of Tumours. 2nd edn. Berlin: Springer, 1994; 13–5. Manos MM, Wright DK, Lewis AJ, Broker TR, Wolinsky SM. The use of polymerase chain reaction amplification for the detection of genital papillomavirus. In: Furth M, Greaves M, editors. Molecular Diagnostics of Human Cancer. Cold Spring Harbor, NY: Cold Spring Harbor Press, 1989; 7: 209–14. deRhoda Husman AM, Walboomers JM, van den Brule AJ, Meijer CJ, Snijders PJ. The use of general primers GP5 and GP6 elongated at their 3’ ends with adjacent highly conserved sequences improves human papillomavirus detection by PCR. J Gen Virol 1995; 76: 1057–62. Pizzighella S. Simultaneous polymerase chain reaction detection and restriction typing for the diagnosis of HPV. J Virol Methods 1995; 55: 245–56. Clavel CE, Nawrocki B, Bosseaux B. Detection of human papillomavirus DNA in bronchopulmonary carcinomas by hybrid capture II. Cancer 2000; 88: 1347–52. Duggan MA, McGregor SE, Benoit JL, Inoue M, Nation JF, Stuart GCE. The human papillomavirus status of invasive cervical adenocarcinoma: a clinicopathological and outcome analysis. Hum Pathol 1995; 26(3): 319–25.
ANATOMICAL PATHOLOGY
The value of HPV DNA typing in the distinction between adenocarcinoma of endocervical and endometrial origin MYFANWY K. PLUNKETT*, BRIAN BRESTOVAC{, JANE E. THOMPSON{, GREGORY F. STERRETT*, PIERRE R. FILION*, DAVID W. SMITH{ AND FELICITY A. FROST* Departments of *Tissue Pathology and {Division of Microbiology and Infectious Diseases, The Western Australian Centre for Pathology and Medical Research (PathCentre) Nedlands, Western Australia and {Tissue Pathology, King Edward Memorial Hospital, Subiaco, Western Australia
Summary Aims: Distinguishing between adenocarcinomas of endocervical and endometrial origin histologically can be difficult, particularly in small biopsies. Most endocervical adenocarcinomas contain human papillomavirus (HPV) deoxyribonucleic acid (DNA) of ‘high-risk’ (HR) types, whereas this has not been consistently demonstrated in endometrial adenocarcinomas. The aim of this study was to determine whether HPV DNA testing could aid in this differential diagnosis. Methods: The frequency of HPV DNA in paraffin-embedded tissue samples from 50 endocervical and 50 endometrial adenocarcinomas was investigated using polymerase chain reaction (PCR) amplification techniques involving (i) a screening HPV test followed by HPV DNA sequencing, and (ii) a test designed to detect HR genotypes 16, 18, 31, 33, 35, 45 and 58. Control specimens included cervical intraepithelial neoplasia (CIN) III lesions, squamous cell carcinomas (SCCs) of the cervix and lung, and colonic adenocarcinomas. Measures to minimise cross-contamination were implemented. Results: The screening test followed by HPV DNA sequencing had the highest sensitivity. By this test HR HPV DNA was detected in 11 of 11 (100%) cervical intraepithelial neoplasia (CIN III) lesions, nine of 10 (90%) cervical SCCs, none of 10 (0%) colorectal adenocarcinomas and none of 10 (0%) SCCs of the lung. Thirty-nine (78%) endocervical adenocarcinomas contained HR HPV DNA, compared to one (2.0%) endometrial adenocarcinoma. Conclusions: The results suggest that HPV DNA testing could be a useful adjunct in distinguishing between endocervical and endometrial adenocarcinomas in curettings or small biopsy specimens. Key words: Endocervical adenocarcinoma, endometrial adenocarcinoma, human papillomavirus, polymerase chain reaction, HPV DNA sequencing, high-risk HPV DNA. Abbreviations: CIN, cervical intraepithelial neoplasia; DNA, deoxyribonucleic acid; HR, high risk; HPV, human papillomavirus; PCR, polymerase chain reaction; SCC, squamous cell carcinoma. Received 26 February, revised 23 June, accepted 1 July 2003
INTRODUCTION The histological distinction between adenocarcinomas of endocervical and endometrial origin can be difficult, particularly in small biopsy and curettage specimens where minimal material is available for examination. Adenocarcinomas in both sites show considerable overlap in terms of histological appearance, with up to 30% of endocervical adenocarcinomas showing an endometrioid morphology. In addition, endometrial adenocarcinoma can secondarily involve the cervix providing further diagnostic difficulty. Mucin histochemistry and immunohistochemistry may be helpful, but are not always conclusive. The distinction between these adenocarcinomas may be clinically significant in terms of prognosis and subsequent patient management.1–3 Invasive carcinomas and high-grade intraepithelial precursor lesions of the cervix have been shown to contain human papillomavirus (HPV) DNA and HPV is an accepted cause of cervical squamous carcinoma.4 Endocervical adenocarcinomas have also been shown to contain human papilloma virus (HPV) DNA of certain ‘high-risk’ (HR) genotypes, with in situ studies confirming the presence of viral DNA in tumour cells.5,6 The reported frequency of HPV detection has varied greatly depending on the molecular techniques used.7 However, in the larger and more recent series performed using the sensitive polymerase chain reaction (PCR) amplification technique, the majority of endocervical adenocarcinomas have been shown to be HR HPV DNA positive.3,8–14 In contrast, the presence of HR HPV DNA has not been consistently demonstrated in endometrial adenocarcinomas. Although occasional studies have shown a relatively high rate of positivity,15–18 the majority of recent studies have reported significantly lower rates of HPV infection in endometrial adenocarcinomas compared to that reported in endocervical adenocarcinomas.3,19–23 Although some studies have used frozen tumour tissue for PCR analysis,12,14,15,23 most have been performed on sections cut from archival paraffin-embedded tissue.3,8–11,13,17,19,20,24–30 The ability of the PCR technique to allow highly specific and sensitive HPV DNA detection in archival paraffin-embedded material has previously been demonstrated31 and PCR techniques for the identification
ISSN 0031-3025 printed/ISSN 1465–3931 online/03/050397–05 # 2003 Royal College of Pathologists of Australasia DOI: 10.1080/00313020310001602611
398
Pathology (2003), 35(5), October
PLUNKETT et al.
of HPV DNA in cell samples have previously been validated in our laboratory.32 The aims of the current study were to examine the frequency of HR HPV DNA positivity in a series of endocervical and endometrial adenocarcinomas and adenosquamous carcinomas using PCR amplification techniques on paraffin-embedded material from small biopsy and curettage specimens, and to determine whether the differences in frequency were great enough to allow HPV DNA testing by PCR to be a useful adjunct in this differential diagnosis.
METHODS Selection of test cases Fifty endocervical adenocarcinomas and adenosquamous carcinomas and 50 endometrial adenocarcinomas and adenosquamous carcinomas were selected from the files of the histopathology departments of PathCentre (Nedlands, WA) and King Edward Memorial Hospital (Subiaco, WA), and included material reported between January 1992 and May 2001. Patient demographic data were obtained from the original pathology report. All of the endometrial adenocarcinomas were curettage specimens from cases with subsequent definitive hysterectomy diagnoses. The endocervical adenocarcinomas included seven endocervical curetting specimens (all with hysterectomy diagnoses), one endometrial curettage specimen, 13 cervical biopsies (eight with hysterectomy diagnoses and one with cone biopsy diagnosis), 15 large loop excisions of the transformation zone (LLETZ) or cone biopsy specimens (10 with hysterectomy diagnoses) and 14 hysterectomy specimens for which no previous small biopsy or curettage specimen was available. All biopsy specimens and hysterectomies were reviewed by a gynaecological pathologist. Tumours were graded according to World Health Organisation (WHO) guidelines,33 on the definitive surgical specimen, where these were available. Histological subtyping was also performed according to WHO guidelines.33 A tumour was considered ‘mixed’ if the second component comprised more than 10% of the total tumour. Cases were excluded if there was minimal material in tissue blocks, if cervical squamous dysplasia was present in the sections together with the adenocarcinoma, or if there were only microscopic foci of adenocarcinoma. Cases were also excluded if definitive endocervical or endometrial location had not been discerned macroscopically or microscopically. One representative block was selected from each case for the PCR studies. Selection of control cases The positive control cases were cervical biopsy and LLETZ or cone biopsy specimens from 10 CIN III lesions and 10 invasive squamous cell carcinomas (SCCs) of the cervix. The negative control cases were 10 SCCs of the lung and 10 colorectal adenocarcinomas. Specimen processing and PCR technique Three 20-mm sections were cut from each paraffin block and placed directly into Eppendorf tubes. The material was submitted for PCR analysis without knowledge of the tissue diagnosis. An initial HPV screening test was performed and used a nested PCR technique with primers targeting the L1 region of HPV. The outer MYO9 and MY11 primers were designed by Manos et al.34 using magnesium chloride at a concentration of 2.0 nM and an annealing temperature at 50‡C resulting in a product size of 450 base pairs. The inner primers, GP5z and GP6z, were designed by deRhoda Husman et al.,35 using magnesium chloride at a concentration of 2.0 nM and an annealing temperature at 45‡C, resulting in a product of 150 base pairs. The test detects episomal DNA and cannot detect integrated HPV DNA because the L1 region is lost on integration.
HR HPV testing was performed using a protocol designed to detect HPV genotypes 16, 18, 31, 33, 35, 45 and 58. For this test the primers targeted the E1–7 border regions, and were designed by Pizzighella et al.36 The outer primers were 3s and 1as (product size 670 base pairs), with inner primers 2s and 2as (product size 161 base pairs). Both outer and inner PCRs used a magnesium chloride concentration of 2.0 nM and an annealing temperature of 55‡C. The test should detect both episomal and integrated HPV DNA because the early region is kept on integration. For all PCR tests, 45 cycles were performed with the hot start Taq Polymerase, Taq Gold (Applied Biosystems). A positive control (HPV 18 extract) was used for each PCR ‘run’, with water controls performed after each five cases to ensure no contamination had occurred. PCR products were run on 2.5% ethidium bromide gel and photographed. Duplicate tests were performed for all cases and PCR photographs were read by two scientists independently with no disagreement.
Initial validation of PCR techniques DNA integrity was tested by amplification of endogenous retroviral PCR, which was amplified in all cases. For both nested PCR reactions the sensitivity was approximately 50 copies of HPV 18, as detected by serial dilutions of HeLa cell lines. The sensitivity of the PCR technique for clinical use has previously been validated against a cohort of 304 routine Pap smears with HR HPV detected in 15 of 16 (93.8%) ThinPrep1 (Cytyc Corporation, Boxborough, Massachusetts) samples from biopsy-proven high-grade cervical squamous intra-epithelial lesions. HR HPV was detected in 12 of 227 (5.3%) negative smears.32 HPV-DNA sequencing In control and test cases where HPV DNA was detected on the initial screening PCR, subsequent HPV DNA genotyping was performed. DNA sequencing was performed on the first round product of the screening PCR using the MYO 9 primer as the sequencing primer. However, if the PCR product was weak (low amount of sequencing target), the second round product was used, and in these cases GP5z was used as the sequencing primer. The DNA sequencing was performed using the Applied Biosystems BigDye Terminator V 1.0 in accordance with manufacturer’s instructions. Procedures to minimise cross-contamination during section cutting The material was cut in batches of 10–20 by several laboratory scientists. Each block was cut at a different point along the cutting edge of the microtome blade, with the blade and forceps wiped clean after each case. During cutting of the test cases, extra cleaning of the microtome blades using xylene was performed. Statistical analysis The significance of differences was tested using the x2 goodness of fit test. All tests were for two-tailed difference, with a type one error set at a level of 95% (P~0.05).
RESULTS For the endocervical adenocarcinomas and adenosquamous carcinomas, the patient’s ages ranged from 24 to 80 years (mean 46.6), and for the endometrial adenocarcinomas and adenosquamous carcinomas the ages ranged from 37 to 86 years (mean 68.2). Of the endocervical adenocarcinomas and adenosquamous carcinomas, 24 (48%) were well differentiated, 17 (34%) were moderately differentiated and nine (18%) were poorly differentiated. Of the endometrial adenocarcinomas, 13
HPV DNA IN FEMALE GENITAL ADENOCARCINOMAS
DISCUSSION
(26%) were well differentiated (grade I), 23 (46%) were moderately differentiated (grade II) and 14 (28%) were poorly differentiated (grade III). The results of the screening HPV DNA and HR HPV DNA tests and subsequent HPV DNA sequencing in the control cases are shown in Table 1. HR HPV DNA was detected in 100% of CIN III lesions and 90% of invasive squamous cell carcinomas of the cervix. Only one of the 20 negative controls, a SCC of lung, was found to contain HPV DNA, of a non-HR type. The results of the screening HPV DNA and HR HPV DNA tests and subsequent sequencing in the test cases are shown in Table 2, together with tumour typing. The screening HPV test was positive in 78% of endocervical adenocarcinomas, whereas HR HPV DNA testing was positive in only 54% of cases. All of the screening positive, HR-negative cases were found to contain HR HPV DNA by sequencing. The HPV DNA types detected by HPV DNA sequencing included genotype 18 (20 cases), genotype 16 (14 cases), and genotype 45 (five cases). The rates of HR HPV DNA positivity for well, moderately and poorly differentiated endocervical adenocarcinomas were 83.3, 76.5 and 66.7%, respectively (P~0.5785). The mean age of patients with HR HPV-positive endocervical carcinomas was 45.3 years compared to 50.8 years in patients with HPV DNA-negative tumours (P~0.5785). Only one (2.0%) endometrial adenocarcinoma was shown to be HPV-positive (Grade II endometrioid adenocarcinoma), which was HR HPV subtype 16 on HPV DNA subtyping.
TABLE 1
399
The consistent results in the positive and negative control cases provide further validation for the PCR techniques used in our laboratory. The finding of a single non-HR HPV DNA-positive squamous cell carcinoma of the lung is consistent with previous reports of a low rate of positivity in lung cancer.37 There was a significant difference between the rate of HR HPV DNA positivity in endocervical and endometrial carcinomas. Whilst 78.0% of endocervical carcinomas were shown to contain HR HPV DNA, only one (2.0%) endometrial adenocarcinoma was positive (P~v0.001). A positive HPV DNA result therefore makes endometrial adenocarcinoma or adenosquamous carcinoma very unlikely. A negative result by our methods, however, has less significance since HR HPV DNA was not detected in 22.0% of endocervical adenocarcinomas and adenosquamous carcinomas. The detection rate in endocervical adenocarcinomas is similar to that seen in the larger and more recent studies using the PCR technique, with reported frequencies of HPV DNA positivity ranging from 64 to 95%.3,8–14 A number of reasons have been suggested for these differences. PCR techniques may have variable sensitivities. Some PCR tests have been designed to detect only highrisk HPV types 16 and 18,8,13 whereas other studies, including the current study, detected a wider range of HR HPV DNA types.3,9–12,14,38 Geographical variation in levels of HPV infection has also been suggested.16 Earlier studies using various molecular techniques showed a much wider range of HPV positivity, reflecting the variable
Results of HPV DNA testing in control cases.
Diagnosis
HPV DNA zve by HPV screening test
Total
CIN III Invasive SCC cervix Colorectal adenocarcinoma SCC lung
11 10 10 10
11 9 0 1
(100%) (90%) (0%) (10%)
HR* HPV DNA zve by HR* HPV test 11 7 0 0
(100%) (70%) (0%) (0%)
HR* HPV DNA zve by sequencing of screening test zve cases 11 9 0 0
(100%) (90%) (0%) (0%)
*HR, high risk.
TABLE 2
Results of HPV DNA testing in endocervical and endometrial adenocarcinomas and adenosquamous carcinomas by histological subtype.
Diagnosis
Total
Endocervical carcinomas Mucinous (endocervical and intestinal) Well-differentiated villoglandular Endometrioid Adenosquamous Total
32 8 5 5 50
Endometrial carcinomas Endometrioid Villoglandular Serous papillary Mixed adenocarcinomas** Adenosquamous Total
38 3 1 7 1 50
HPV DNA zve by HPV screening test
24 7 4 4 39
(75.0%) (87.5%) (80.0%) (80.0%) (78.0%)
1 (2.6%) 0 0 0 0 1 (2.0%)
HR* HPV DNA zve by HR HPV test
18 5 3 1 27
(56.3%) (62.5%) (60.0%) (20.0%) (54.0%)
0 (0%) 0 0 0 0 0 (0.0%)
HR* HPV DNA zve by sequencing of screening test zve cases
24 7 4 4 39
(75.0%) (78.5%) (80.0%) (80.0%) (78.0%)
1 (2.6%) 0 0 0 0 1 (2.0%)
HR*, high risk; **the mixed adenocarcinomas included endometrioid/clear cell (two), serous papillary/clear cell (one), and endometrioid/clear cell (four).
400
PLUNKETT et al.
sensitivities of the different techniques as well as variation in sample size.7,8,38 Our low rate of detection of HR HPV DNA in endometrial adenocarcinomas is comparable to the majority of recent reports in the literature.3,19–23,25 In particular the findings are very similar to those found by Hording et al.3 in 1997, a study with a very similar methodology to our own, with similar numbers of endometrial and endocervical adenocarcinomas examined. However, the endometrial adenocarcinomas used in the study by Hording et al.3 were of surgical stage 2 (with cervical extension), specifically selected to simulate the situation where differential diagnostic problems arise. They showed none of the endometrial adenocarcinomas studied to contain HPV DNA, whilst detecting the virus in 70% of endocervical adenocarcinomas, similar to that seen in the present study. Nevertheless, occasional studies have shown higher rates of HPV DNA positivity.15–18 One of the highest levels of detection was recorded in a 1994 study by Lai et al.15 who found high-risk HPV DNA in 61.1% of endometrial adenocarcinomas by reverse transcriptase PCR, but found the DNA to be transcriptionally inactive. In some of these studies, HPV DNA was also detected in benign endometrial lesions18 and in benign and malignant ovarian tissue.15,17 These authors did comment that the amount of HPV-DNA detected was always low compared to most cervical and vulvar lesions.18 Differences in sample size and geographic variation have again been suggested as reasons for the varying results.16 Contamination from cervical or laboratory sources is a potentially important cause of false-positive results in PCR-based studies, particularly those using archival tissue. Contamination from adjacent squamous epithelium and especially CIN III lesions has been suggested as a possible mechanism for false-positive results in endocervical adenocarcinomas. In our study we ensured that no adjacent squamous intraepithelial lesion was present in the slides together with the adenocarcinoma. Some researchers have tried to prevent contamination by trimming away nontumorous tissue on sections cut from paraffin blocks.16 Other workers have used sterile scalpel blades to microdissect the cervical adenocarcinoma from adjacent squamous epithelium and stroma either on glass slides or directly from the paraffin block.7,9,10,38 The benign cervical stroma from the same slide has also been used as a negative control.9 The presence of the same HPV type in paraffin tissue blocks from histologically normal cervical tissue and endometrial adenocarcinoma tissue from the same patient has been noted, but might be used to argue either for contamination, or a consistent finding validating the study.19,20 Minimising cross-contamination of specimens during section cutting has been addressed in different ways. Some workers cleaned the microtome thoroughly with alcohol or xylene after cutting of each block and used new surgical gloves for each case.11,27 Others thoroughly cleaned17,24 or changed blades between cases.9 Some adopted even stricter anticontamination protocols with ultraviolet irradiation of all work surfaces and microtome blades between cases.19 We achieved good results despite relatively limited measures to prevent cross-contamination. Although sections from the positive and negative controls were cut at the same time, using the same microtome, HPV DNA was
Pathology (2003), 35(5), October
detected in only one negative control. Similarly, the endocervical and endometrial adenocarcinomas were cut in mixed batches, with only one endometrial adenocarcinoma found to contain HPV DNA. These findings suggest the technique used could be applicable to the routine diagnostic histopathology laboratory. We attempted to find a relationship between the HPV DNA status of tumours and patient age, tumour grade and tumour subtype. In our study, HPV-positive tumours were more common in younger patients (mean age 45.3 versus 50.8 years) but this was not statistically significant (P~0.5785). Some studies have demonstrated a significant correlation between younger age of the patient and HPV DNA positivity in endocervical adenocarcinoma;3,10,13 however, this has not been demonstrated in other papers.7,9,37 We found that the frequency of HPV DNA positivity in endocervical adenocarcinomas and adenosquamous carcinomas decreased with increasing tumour grade, but this did not reach statistical significance (P~0.5785). Other studies have demonstrated a similar correlation with more advanced clinical stage.10,13 However, the majority of studies have found no such correlation between tumour grade or stage and HPV DNA positivity.7,9,16 No significant correlation with histological subtype was demonstrated in our study, although only very small numbers of serous papillary and clear cell adenocarcinomas were included. Several studies have found a significantly higher proportion of mucinous tumours to be HPV DNA-positive compared to non-mucinous tumours, including clear cell tumours.8–10,13 High rates of positivity have been previously demonstrated in adenosquamous carcinomas.9,25,27 Some previous reports have suggested HPV 18 to be the most commonly identified HPV type in endocervical adenocarcinomas, leading to the assumption that HPV 18 may be more capable of infecting and transforming glandular epithelial cells than HPV 16, which is more commonly associated with squamous cell carcinoma.6,7 However, not all studies have found this correlation.8,10 We reported HPV 18 in 54.5% of all HPV DNA types detected in adenocarcinomas. In this study we used two separate HPV DNA PCR tests, one a ‘screening’ test and the other designed to detect a panel of HR HPV DNA. The HR HPV DNA test failed to detect HR HPV DNA in a number of the test cases that were positive by the screening HPV test. By sequencing these cases, HPV types 16, 18 or 45 were detected in all screening positive, HR-negative cases. This represents a significant failure of the HR HPV test. This may be due to degradation of the DNA during formalin fixation, possibly making PCR, particularly of greater than 200 base pairs, unreliable. Low levels of PCR inhibitors, related to DNA extraction from paraffin-embedded tissue, may have affected the HR HPV assay more than the PCR screening assay. However, testing for DNA integrity did show all of the test and control cases to amplify endogenous retroviral DNA. Improving the sensitivity of the HR HPV test in paraffin-embedded tissue would significantly improve the efficiency of the testing process, eliminating the need for DNA sequencing and possibly for the first HPV DNA PCR screening test. HPV DNA typing by the PCR tests and DNA sequencing methods used here, may be useful in the distinction between adenocarcinomas of endocervical and endometrial origin in
HPV DNA IN FEMALE GENITAL ADENOCARCINOMAS
small biopsy and curetting specimens. A positive result for HR HPV DNA makes endometrial adenocarcinoma very unlikely. Good results were achieved despite limited measures to prevent cross-contamination, suggesting that the methods would be applicable in the routine diagnostic histopathology laboratory. Further efforts to improve the sensitivity of PCR-based detection of HR HPV DNA in paraffin-embedded tissue from endocervical carcinomas will be worthwhile. ACKNOWLEDGEMENTS This research was funded by a grant from the Education and Research fund of The Western Australian Centre for Pathology and Medical Research (PathCentre), Nedlands, Western Australia. The authors gratefully acknowledge the assistance of the PathCentre and King Edward Memorial Hospital laboratory staff, and Pierre Filion, Senior Scientist, Electron Microscopy Unit, for help with the statistical analysis. Address for correspondence: Dr M. Plunkett, Histopathology Department, PathCentre, Hospital Avenue, Nedlands, 6009 WA, Australia. E-mail: littlefi[email protected]
17.
18.
19.
20.
21.
22.
23.
24.
25.
References 1. Zaino RJ. The fruits of our labours: distinguishing endometrial from endocervical adenocarcinoma. Int J Gynecol Pathol 2002; 21(1): 1–3. 2. Dabbs DJ, Sturtz K, Zaino RJ. The immunohistochemical discrimination of endometrioid adenocarcinomas. Hum Pathol 1996; 27(2): 172–7. 3. Hording U, Daugaard S, Visfeldt J. Adenocarcinoma of the cervix and adenocarcinoma of the endometrium: distinction with PCRmediated detection of HPV DNA. APMIS 1997; 105: 313–6. 4. zur Hausen H. Human papillomaviruses in the pathogenesis of anogenital cancer. Virology 1991; 184(1): 9–13. 5. King LA, Tase T, Twiggs LB, et al. Prognostic significance of the presence of human papillomavirus DNA in patients with invasive carcinoma of the cervix. Cancer 1989; 63: 897–900. 6. Tase T, Okagaki T, Clark BA, et al. Human papillomavirus types and localization in adenocarcinoma and adenosquamous carcinoma of the uterine cervix: a study by in situ DNA hybridisation. Cancer Res 1988; 48: 993–8. 7. Duggan MA, Benoit JL, McGregor SE, Nation JG, Inoue M, Stuart GCE. The human papillomavirus status of 114 endocervical adenocarcinoma cases by dot blot hybridization. Hum Pathol 1993; 24(2): 121–5. 8. Tenti P, Romagnoli S, Silini E, et al. Human Papillomavirus types 16 and 18 infection in infiltrating adenocarcinoma of the cervix. PCR analysis of 138 cases and correlation with histological type and grade. Am J Clin Pathol 1996; 106(1): 52–6. 9. Pirog EC, Kleter B, Olgnac S, et al. Prevalence of human papillomavirus DNA in different histological subtypes of cervical adenocarcinoma. Am J Pathol 2000; 157: 1055–62. 10. Milde-Langosch K, Schreiber C, Becker G, Loning T, Stegner HE. Human papillomavirus detection in cervical adenocarcinoma by polymerase chain reaction. Hum Pathol 1993; 24: 590–4. 11. van Muyden RCPA, ter Harmsel BWA, Smedts FMM, et al. Detection and typing of human papillomavirus in cervical carcinomas in Russian women. Cancer 1999; 85(9): 2011–6. 12. Bosch FX, Manos MM, Munoz N, et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. J Natl Cancer Inst 1995; 87(11): 796–802. 13. Tenti P, Romagnoli S, Silini E, et al. Analysis and clinical implications of K-ras gene mutations and infection with human papillomavirus types 16 and 18 in primary adenocarcinomas of the uterine cervix. Int J Cancer 1995; 64: 9–13. 14. Chen S, O’Sullivan H, Tabrizi SN, Fairley CK, Quinn MA, Garland SM. Prevalence and genotyping of HPV in cervical cancer among Australian women. Int J Gynecol Obstet 1999; 67: 163–8. 15. Lai C, Wang C, Lin C, Pao CC. Detection of human papillomavirus RNA in ovarian and endometrial carcinomas by reverse transcription/ polymerase chain reaction. Gynecol Obstet Invest 1994; 38: 276–80. 16. Fujita M, Shroyer KR, Markham NE, et al. Association of human
26.
27.
28. 29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
401
papillomavirus with malignant and premalignant lesions of the uterine endometrium. Hum Pathol 1995; 26(6): 650–8. Lai C-H, Hsueh S, Lin C-Y, et al. Human papillomavirus in benign and malignant ovarian and endometrial tissues. Int J Gynecol Pathol 1992; 11: 210–5. Milde-Langosch K, Becker G, Loning T. Human papillomavirus and c-myc/c-erbB2 in uterine and vulvar lesions. Virchow’s Arch A Pathol Anat 1991; 419: 479–85. O’Leary JJ, Landers RJ, Crowley M, et al. Human papillomavirus and mixed epithelial tumours of the endometrium. Hum Pathol 1998; 29(4): 383–9. Wong WSF, Wong YF, Tam OSP, Tam JSL. Detection of human papilloma virus (HPV) infection in paraffin-embedded tissues of endometrial carcinoma. Aust NZ J Obstet Gynaecol 1993; 33(2): 180–2. Bergeron C, Shah K, Daniel R, Ferenczy A. Search for human papillomaviruses in normal, hyperplastic, and neoplastic endometria. Obstet Gynecol 1988; 72:3: 383–7. Czerwenka K, Lu Y, Heuss F, Manavi M, Kubista E. Human papillomavirus detection in endometriod carcinoma with squamous differentiation of the uterine corpus. Gynecol Oncol 1996; 61: 210–4. Brewster WR, Monk BJ, Burger RA, Bergen S, Wilczynski SP. Does human papillomavirus have a role in cancers of the uterine corpus? Gynecol Oncol 1999; 75: 51–4. Anciaux D, Lawrence WD, Gregoire L. Glandular lesions of the uterine cervix: prognostic implications of human papillomavirus status. Int J Gynecol Pathol 1997; 16: 103–10. Hachisuga T, Matsuto N, Iwasaka T, Sugimori H, Tsuneyoshi M. Human papilloma virus and P53 overexpression in carcinomas of the uterine cervix, lower uterine segment and endometrium. Pathology 1996; 28: 28–31. Koulos JP, Wright TC, Follen Mitchell M, et al. Relationships between c-Ki-ras Mutations, HPV types and prognostic indicators in invasive endocervical adenocarcinomas. Gynecol Oncol 1993; 48: 364–9. Yamakawa Y, Forslund O, Teshima H, Hasumi K, Kitagawa T, Hansson BG. Human papillomavirus DNA in adenocarcinoma and adenosquamous carcinoma of the uterine cervix detected by polymerase chain reaction (PCR). Gynecol Oncol 1994; 53: 190–5. Parker MF, Arroyo GF, Geradts J, et al. Molecular characterization of adenocarcinoma of the cervix. Gynecol Oncol 1997; 64: 242–51. Lee MF, Chang MC, Wu CH. Detection of human papillomavirus types in cervical adenocarcinoma by the polymerase chain reaction. Int J Gynecol Obstet 1998; 63: 265–70. Claas ECJ, Melcher WJG, van der Linden HC, Linderman J, Quint WGV. Human papillomavirus detection in paraffin-embedded cervical carcinomas and metastases of the carcinomas by the polymerase chain reaction. Am J Pathol 1989; 135(4): 703–9. Resnick RM, Cornelissen MTE, Wright DK, et al. Detection and typing of human papillomavirus in archival cervical cancer specimens by DNA amplification with consensus primers. J Natl Cancer Inst 1990; 82: 1477–84. Choo C, Rankine C, Williams V, Brestovac B, Smith D, Frost FA. Correlation of high-risk HPV DNA testing with cervical cytology in 304 ThinPrep split samples. Poster Display, Annual Scientific Meeting, Australian Society of Cytology, Broome, June 2000. Scully RE, Banfiglio TA, Kurman RJ, et al. Histological Typing of Female Genital tract Tumours. In: WHO International Histological Classification of Tumours. 2nd edn. Berlin: Springer, 1994; 13–5. Manos MM, Wright DK, Lewis AJ, Broker TR, Wolinsky SM. The use of polymerase chain reaction amplification for the detection of genital papillomavirus. In: Furth M, Greaves M, editors. Molecular Diagnostics of Human Cancer. Cold Spring Harbor, NY: Cold Spring Harbor Press, 1989; 7: 209–14. deRhoda Husman AM, Walboomers JM, van den Brule AJ, Meijer CJ, Snijders PJ. The use of general primers GP5 and GP6 elongated at their 3’ ends with adjacent highly conserved sequences improves human papillomavirus detection by PCR. J Gen Virol 1995; 76: 1057–62. Pizzighella S. Simultaneous polymerase chain reaction detection and restriction typing for the diagnosis of HPV. J Virol Methods 1995; 55: 245–56. Clavel CE, Nawrocki B, Bosseaux B. Detection of human papillomavirus DNA in bronchopulmonary carcinomas by hybrid capture II. Cancer 2000; 88: 1347–52. Duggan MA, McGregor SE, Benoit JL, Inoue M, Nation JF, Stuart GCE. The human papillomavirus status of invasive cervical adenocarcinoma: a clinicopathological and outcome analysis. Hum Pathol 1995; 26(3): 319–25.