Comparative Analysis of Human Papillomavirus Detection by Polymerase Chain Reaction and Virapap/Viratype Kits

Transcription

1 Comparative Analysis of Human Papillomavirus Detection by Polymerase Chain Reaction and Virapap/Viratype Kits GLENNA C. BURMER, M.D., PH.D., JAY D. PARKER, B.S., JEANNIE BATES, B.S., KATRINA EAST, B.S., AND BRUCE G. KULANDER, M.D. Cervical samples from 7 women referred by area physicians were analyzed for the presence of human papillomavirus (HPV) types,,, and 8 by the polymerase chain reaction (PCR) followed by Southern hybridization. Samples from S patients were concurrently analyzed by a commercial filter hybridization technique (Virapap and Viratype Kits, Life Technologies, Bethesda Research Labs, Gaithersburg, MD). The sensitivity of the Southern blot procedure combined with PCR was significantly higher than that of the Virapap and Viratype methods. HPV was detected in 7% of women who had positive results for dysplasia by PCR and in 7% by the Virapap method. HPV types /8 were found more commonly than types / in every diagnostic category. More than one HPV type was detected in % of HPV-positive patients. The prevalence of HPV in cytologically negative or indefinite patients as measured by PCR was % and %; in contrast, by the Virapap method, these values were 7% and %. These results demonstrate that PCR combined with Southern hybridization provides a higher level of sensitivity than methods that use hybridization without amplification of HPV DNA and also show that the prevalence of HPV is highest in cytologically positive smears. (Key words: Human papillomavirus; Polymerase catalyzed chain reaction; Virapap Kit; Cervical dysplasia) Am J Clin Pathol 99;9: 55-5 ALTHOUGH CERTAIN TYPES of human papillomaviruses (HPVs) have been strongly implicated in the genesis of cervical carcinoma, the exact prevalence of HPV and its role in the progressive development of neoplasia have not been clearly established. ' Prevalence data have depended on methods of detection such as in situ or filter hybridization with HPV-specific probes to detect the viruses. ' 9-5,8 Papillomaviruses have been detected in up to 9% of cervical carcinomas and between % and 8% of specimens exhibiting negative cytologic results. 8 ' 9 " 79 ' ' 7 ' ' Recent studies using the polymerase chain reaction (PCR) to amplify the virus have suggested that these prevalence rates are underestimates, and detection rates as high as 8% have been reported in cytologically negative patients. 5, ' Such high prevalence Received November 9, 989; received revised manuscript and accepted for publication March 9, 99. Supported in part by the NH NIDDK. grant DK97 and National Institute on Aging grant AG759 to Dr. Glenna Burmer. Address reprint requests to Dr. Burmer: SM-, Department of Pathology, School of Medicine, University of Washington, Seattle, Washington Department of Pathology, University of Washington School of Medicine and the Laboratory of Pathology of Seattle, Seattle, Washington rates in cytologically negative patients bring into question the clinical utility of screening for HPV. We have examined the prevalence of HPV,,, and 8 in 7 Seattle area patients referred to the Laboratory of Pathology of Seattle and the Swedish Hospital Medical Center, using the PCR combined with Southern hybridization to detect viral DNA sequences. In 5 cases, we have compared the sensitivity of detection by PCR and Southern hybridization with filter hybridization alone, using a commercially available assay, the Virapap and Viratype Kits (Life Technologies, Bethesda Research Labs, Gaithersburg, MD). Results obtained with the PCR and Virapap/Viratype assays were then correlated with cytologic classification and histopathologic diagnosis. Our data demonstrate the significantly increased sensitivity of detection when PCR is used to amplify the presence of the virus in patient samples. In contrast to recent reports, however, we detect significantly fewer positive samples by PCR in patients with negative or indefinite cytologic results. Our results therefore support the clinical utility of this assay for screening high-risk patients for papillomavirus. Study Design Experimental Design Two hundred seventy cervical samples were collected by Seattle area physicians between January and October 989, evaluated for the presence of HPV DNA at the University of Washington, and evaluated cytologically at the Laboratory of Pathology of Seattle. This referral-based patient population consisted of women attending clinics of area gynecologists and family practitioners. For cytologic evaluation, Papanicolaou (Pap) -stained cervical smears that had been obtained at the same time or within three months of HPV DNA analysis were evaluated by a cytopathologist (B.G.K.) without his knowledge of the results of the molecular studies. Smears were cat- 55 Downloaded from on 7 July 8

2 Vol. 9. No. 5 HUMAN PAPILLOMAVIRUS DETECTION BY PCR 555 egorized as negative, indefinite for dysplasia (I), cervical intraepithelial neoplasia (CIN) (mild dysplasia or condyloma), CIN (moderate dysplasia), or CIN (severe dysplasia). Patients with a negative diagnosis had a history of negative Pap smears and/or currently have negative findings for dysplasia; those with an indefinite for dysplasia diagnosis exhibited minimal cytologic alterations (e.g., inflammatory atypia) not diagnostic for condyloma or dysplasia. For HPV detection, cervical samples were collected by a Dacron swab using the Virapap Collection Kits according to the manufacturer's instructions (Life Technologies, Inc., Bethesda Research Labs, Gaithersburg, MD). Ten percent of the sample was immediately removed from the Virapap collection buffer, DNA was extracted, and ng of DNA was used as a template in PCR reactions using primers specific for HPV types,,, and 8 as detailed below. Ten percent of the PCR product was analyzed by Southern blotting using probes specific for each viral type. Controls for the PCR reactions consisted of primers amplified without template DNA or in the presence of human breast DNA (negative controls); and for positive controls, DNA extracted from HeLa cells (HPV 8), SiHa cells (HPV ), 9 and a patient with a condyloma known to be positive for HPV type /. Twenty-five percent of the remainder of the original sample was analyzed by the Virapap method, and viruspositive samples were further analyzed by the Viratype Kit according to the manufacturer's directions. Molecular analyses were all conducted without the examiner's prior knowledge of the corresponding cytologic results. DNA Extraction Genomic DNA was extracted from patient samples by standard methods. One hundred microliters from each Virapap Collection Kit was incubated overnight at 7 C in the presence of 5 fig/ml proteinase K and.5% (w/ v) sodium dodecyl sulfate in buffer containing mmol/ L TRIS-HC (ph 8.), 5 mmol/l NaCl, and mmol/ L EDTA, followed by serial phenol-chloroform (:) and chloroform extractions, and precipitated with. vol of mol/l sodium acetate (ph 5.) and.5 vol of ethanol at 7 C overnight. The DNA was resuspended in /il of water after centrifugation, and the concentration was determined spectrophotometrically. When less than ng of total DNA was present, 5 nl of the sample was used as a template in the PCR reaction; when larger amounts were recovered, ng was routinely used as a template for PCR. Polymerase Chain Reaction Genomic DNA was incubated in a total volume of nl in solution containing the following: 5 mmol/l KC; mmol/l TRIS-HC (ph 8.); 5 mmol/l MgCl ;.% (w/v) gelatin; Mmol/L each datp, dctp, dgtp, and dttp; ng/jul of oligonucleotide primers; and.5 units of Taq polymerase (Perkin-Elmer Cetus, Norwalk, CT). The mixture was overlaid with mineral oil and underwent cycles of PCR with denaturation for minute at 9 C, followed by -minute incubations at C (annealing) and minutes at 7 C (polymerization). 8 Samples were separately amplified for HPV types / and or 8 with the use of primers specific for each viral type (Table ). The oligonucleotide primers and probes were synthesized by our laboratory with the use of the beta-cyanoethyl phosphoramidite method on a Cruachem synthesizer (Cruachem Corporation, Herndon, VA). Southern and Dot Blot Hybridizations of PCR Products Ten microliters of the PCR product was electrophoresed on % (w/v) agarose, denatured for minutes by incubation in buffer containing.5 mol/l NaCl and.5 mol/ L NaOH, neutralized for minutes in mol/l NaCl with.5 mol/l TRIS-HC (ph 7.), and transferred to Nytran membranes (Schleicher and Schuell, Keene, NH). For dot blot hybridization, the DNA was denatured by addition of. vol of N NaOH and boiled for Table J. Oligonucleotide Primers and Probes Used in PCR Assay Primer Region Base Pair Reference Sequence /-U / l-d / l-p -U -D -P 8-U 8-D 8-P El El El , 5, 5, 5'-AAA GGT AAA GCG ACG GCT GTT-' 5'-TTT AAA TGG CCT AAT TAA ATC-' 5'-GGG AAC TAA CGG ACA GTG GAT ATG GCT ATT CTG AAG TGG AAG CT-' 5'-ATT AGT GAG TAT AGA CAT TA-' 5'-GGC TTT TGA CAG TTA ATA CA-' 5'-ATG GAA CAA CAT TAG AAC AGC AAT ACA ACA AAC CGT TCT G-' 5'-ACT ATG GCG CGC TTT GAG GAT CCA-' 5'-GGT TTC TGG CAC CGC AGG CA-' 5'-ATG GAG ACA CAT TGG AAA AAC TAA CTA ACA CTG GGT TAT-' U = upstream PCR primer: D = downstream PCR primer; P = probe. Downloaded from on 7 July 8

3 55 BURMER ET AL. A.J.C.P. November 99 minutes at C, neutralized by addition of.5 vol of 5 mol/l ammonium acetate (ph 5.) at C, and spotted onto Nytran membranes with the use of a Vacudot-VS apparatus (American Bionetics, Inc., Hayward, CA). The membranes were baked at 8 C for one hour; prehybridized for four hours in 5X SSC, X Denhardt's solution, Mg/mL salmon sperm DNA, and.% (w/v) sodium dodecyl sulfate (SDS) and hybridized overnight at 8 C with [7- P]-end-labeled oligonucleotide probes specific for HPV types /,, or 8. The membranes were washed in 5X SSC/.5% (w/v) SDS at 7 C for 5 minutes, followed by two 5-minute washes in.5x SSC/.5% (w/v) SDS at 5 C, and exposed for two to three days to x-ray film at 7 C with two intensifying screens. Experimental Results SOUTHERN BLOT OF PCR AMPLIFIED HPV DNA 5789 A m PCR Analysis of Cervical Specimens Validation of the PCR Assay for Detection of HPV Types /,, and 8. Initial experiments were conducted with tissues and cell lines known to contain specific HPV types to establish the sensitivity and specificity of the PCR assay used in these studies. PCR primers and probes specific for types /,, and 8 were synthesized after a Genebank dot plot analysis of regions exhibiting the least homology between viral types and within the region for HPV types and 8, and El region for HPV types / (Table ). The choice of sequences with least homology ensured the greatest specificity for viral subtypes. The region was chosen to maximize the sensitivity of detection of integrated, low copy number viral sequences in dysplastic lesions, which may have lost viral sequences less necessary for maintenance of the transformed state.' Figure shows a representative Southern blot of /il of the amplification product after cycles of PCR of DNA from HeLa cells (B, lane ), which contain -5 integrated copies of HPV 8 per genome ; SiHa cells (C, lane ), which contain - copies of HPV per genome 9 ; a patient with a urethral condyloma known to be positive for HPV / (A, lane ); and seven patient samples. The patient samples in lanes 7 and 8 are positive for HPV /, and the patient in lane 5 had positive results for HPV 8 and faintly positive results for HPV. The hybridization is specific for each viral type, and crosshybridization was not observed between the type-specific probes and the products of these PCR reactions, except in cases in which dual infection was apparent (e.g., Fig., patient sample 5). Hybridization also was not seen when samples of cultures from herpes simplex types I and II were amplified with HPV-specific primers and hybridized with probes for HPV types /,, or 8 (unpublished data). B c m~ ~^K ^sk! FIG.. Autoradiogram of Southern blot of PCR products after amplification with primers specific for the El region of HPV / (Panel A and Table, /-U and /-D), the region of HPV 8 (Panel B and Table, 8-U and 8-D), and HPV (Panel C and Table, - U and -D). DNA from HeLa cells (Lane ), SiHa cells (Lane ), a condyloma containing HPV type / (Lane ), and seven patient samples was used as a template in a -cycle PCR reaction. Ten percent of the PCR product was electrophoresed on a % (weight/volume) ethidium bromide-stained agarose gel, transferred to Nytran membranes, and each blot separately hybridized with end-labeled probes specific for HPV types / (/ -P), 8 (8-P) or (-P). Thirty percent of the PCR product from the type / positive control patient, SiHa cells, and HeLa cells (Fig. I, lanes,, and, respectively) was then directly sequenced with the use of an internally nested, kinased primer. The sequencing autoradiograms verified the identity of the amplified fragment as specific for types / II,, and 8 (unpublished data). The sensitivity of the PCR method was then tested by performance of serial dilutions of HeLa DNA before PCR and performance of dot blots with nl (%) of the reaction mixture after cycles of PCR (Fig. ). Adjacent wells were loaded with diluted HeLa DNA without am-,-a Downloaded from on 7 July 8

4 Vol. 9 No. 5 HUMAN PAPILLOMAVIRUS DETECTION BY PCR 557 WITH PCR NO PCR DOT BLOT OF HELA DNA A B C D E F G FIG.. Dot blot of products of PCR reaction of serially diluted HeLa DNA compared with DNA without PCR. The HeLa genomic DNA template was diluted and underwent cycles of PCR with type 8 specific primers and blotted onto Nytran (WITH PCR), or blotted at each dilution without amplification (NO PCR). DNA dilutions are as follows: Mg (A), ng (B), pg (Q, pg (D), pg ( ),. pg (F), and primers only (G). plification. The results demonstrate our ability to detect HPV 8 in pg of HeLa DNA, which contains approximately HeLa cell equivalent of DNA or -5 copies of virus. In contrast, our dot blot procedure using a labeled oligonucleotide probe was incapable of detecting HPV in jug of HeLa DNA, which represents an increase in sensitivity of, using PCR in conjunction with the dot blot, over using the dot blot alone for HPV detection. Although these results are not directly comparable to those of the Virapap/Viratype assay, which is a modified dot blot procedure that uses radiolabeled RNA probes corresponding to the full-length HPV genome, a minimum of, HeLa cells (e.g., ng of HeLa DNA) has previously been reported to be the lower limit of sensitivity of detection of HPV DNA by dot blot hybridization procedures that use full-length nick-translated HPV DNA probes. 7 ' 5, Therefore, this represents an increase in sensitivity of approximately, using PCR compared with commonly used filter hybridization methods. PCR Analysis of Clinical Samples. Two hundred seventy patients were analyzed by PCR and Southern hybridization for the prevalence of HPV types /,, and 8. Fifty percent of cases ( of 7) were positive for the presence of at least one virus by PCR; 8% ( of ) of these cases contained type or, 7% (9 of ) were positive for type or 8, and % ( of ) contained evidence of infection by viruses from both the / and /8 groups (Table ). Cases were then analyzed according to cytopathologic diagnosis. Within the group of patients without previous history and without current cytologic evidence for atypia, % (5 of 7) had positive results for HPV (Table ). Within the group of patients indefinite for dysplasia, 7% ( of 59) had positive results for HPV. A higher percentage of patients with cytologic results positive for dysplasia contained detectable HPV (%, 7%, and 7% for CIN,, and, respectively, Table ). The increased prevalence of HPV is reflected by an increase in both types / and /8 within the CIN population; however, the percentage of cases positive for HPV types / decreased in patients with moderate or severe dysplasia. In all diagnostic categories, the frequency of infection with types /8 was at least.9-fold greater than for types /. This difference was statistically significant for the diagnostic categories analyzed as a whole (P <. by a conditional binomial test). Analysis by Virapap/Viratype Kits One hundred fifty-four cases were concurrently analyzed for the presence of HPV by dot blot hybridization with the use of the Virapap and Viratype Kits. Overall, 9% (5 of 5) of cases had detectable HPV by the Virapap method (Table ). Of cases negative or indefinite for dysplasia, 8% ( of 7) contained evidence of HPV by the Virapap method. The percentage of Virapap-positive samples increased in cytologically positive cases (5% for CIN and 5% for CIN and ). Within the subset of 5 PCR amplified cases that were analyzed concurrently by the Virapap method, the percentages of positive cases in each diagnostic category did not differ significantly from the pool of total PCR cases. Cases were then analyzed according to specific viral types as distinguished by the Viratype Kit (/, /8, and //5, Table ). Because of the reduced sensitivity of the commercial Viratype Kit relative to Virapap screening, only of the 5 Virapap-positive cases could Table. PCR Positive and Cases in Each Diagnostic Class Diagnosis n / /8 Both Positive Percent Positive Indefinite CIN CIN CIN ± 5 7 ± ± ± 5+ Downloaded from on 7 July 8

5 558 BURMER ET AL. A.J.C.P. November 99 Table. PCR or Virapap Positive Cases in Each Diagnostic Category Table 5. Comparison of Virapap and PCR/Southern Hybridization Diagnosis n PCR Positive Virapap Positive % Indefinite CIN CIN CIN ± 8 ± ± + 7 ± Standard errors are shown for the percentage of PCR and Virapap positive cases. 7± 5 ± 5 5 ± 7 5 ± 5 ± 9 ± Virapap Positive Positive 7 PCR/Southern 9 8 Shown are the number of cases exhibiting concordance and discordance. 5 be definitively subtyped. The prevalences of types /8 and //5 were equal ( of = 8%) and.7 times more prevalent than type / ( of = 8%); cases ( of = 5%) were positive for more than one subgroup of HPV. The increased frequency of /8 and //5 over the / types existed in all diagnostic categories; this difference was statistically significant as assessed by a conditional binomial test (P <.5). Comparison of Virapap/Viratype Methods with PCR/Southern Blot Complete concordance between PCR/Southern and Virapap assays (both assays positive or negative for a given case) was obtained in % of cases ( of 5, Table 5). Of the 5 discordant cases, ( of 5 = 77%) consisted of those detected as positive by PCR and were negative by the Virapap method. Five cases were positive by the Virapap method and negative by both the PCR and Viratype assays; it is uncertain whether these cases represent false positive results for the Virapap assay or false negative results for both the PCR and Viratype tests. Seven cases were positive by both the Virapap and Viratype assays, yet negative by PCR. Of these seven cases, five were positive only for HPV type //5, which was not examined by our PCR-based assay. Twelve cases were identified by the Viratype assay as positive for HPV //5, alone or in combination with types / or /8 (Table ). Although our PCR assay was not designed to detect HPV //5, seven of the Viratype-positive assays were detected as positive for HPV by PCR. In addition to the two cases that were Viratype positive for / or /8, our PCR assay detected the presence of these subtypes in the remaining five cases that were classified by the Viratype method as only positive for //5. These cases may represent instances of crosshybridization between the Viratype probes for //5 and / or /8 DNA. Alternatively, these may be cases of dual infection in which the Viratype Kit was only able to detect //5 DNA because the second viral type was present at a concentration below the threshold of detection by this method. All cases positive by the Viratype method for / or /8 were also found positive for the same virus by the PCR assay; in addition, of the cases were positive by PCR for both / and or 8. When cases were analyzed according to cytologic classification, the PCR assay was found to be more sensitive than the Virapap method in detecting the presence of HPV in every diagnostic category (Table ). This difference is highly statistically significant when comparing the PCR group as a whole with the Virapap screen (P <. by McNemar's test, Table 5). Discussion Estimates of the prevalence of HPV infection rely on the sensitivity of the technique used to detect virus in patient samples. We have compared the sensitivity and specificity of a commercially available, established filter hybridization method (Virapap and Viratype) with an assay using the PCR in conjunction with Southern hybrid- Table. Viratype Positive Cases: Classification by Diagnosis and HPV Type Diagnosis Positive / /8 //5 Two or More Indefinite CIN CIN CIN (8 ± 5) 5 (8 ± 9) (8 ±9) (5±7) Shown in parentheses is the percentage of the total contributed by each viral group, plus or minus the standard error. Downloaded from on 7 July 8

6 Vol. 9. No. 5 HUMAN PAPILLOMAVIRUS DETECTION BY PCR 559 ization. Our results demonstrate the increased sensitivity of using PCR to detect HPV in patients with or without cytologic evidence of cervical intraepithelial neoplasia. The validity of our comparisons between the PCRbased assay and the dot blot hybridization assay has been demonstrated in several ways. The prevalence of HPV within our cytologically negative population as detected by the Virapap method is consistent with previous reports of the prevalence of HPV infection in Seattle area patients as assessed by Southern or dot blot hybridization methods. 9 Moreover, the increased prevalence of HPV types and 8 over types / within our patient population as detected by both the Viratype method and our PCR assay is consistent with a previous report. 9 The decreased prevalence of types / in higher categories of cytologic dysplasia compared with types /8 has been well established by traditional hybridization studies.' 5 ' ' 9 ' ' In addition, the frequency of dual infections within HPVpositive cases, as detected by both our Viratype and PCR assays, is similar to that reported previously. 5 The major difference between data obtained by our PCR assay and those obtained by Southern or dot blot hybridization alone resides in the increased sensitivity of detection of HPV in every diagnostic category. This difference is made particularly evident by the fact that, in many cases, only % of the patient's sample was used as a template in the PCR assay (% of original sample was used for DNA extraction, of which approximately % was used as a template in the PCR reaction), compared with the volumes used for analysis in the Virapap and Viratype Kits. Moreover, our assay only detected types /,, and 8, and our estimates of prevalence still exceed those detected by the Virapap Kit, which also includes probes for detecting HPV types //5. A practical disadvantage of the PCR procedure is that, because of its ability to amplify minute quantities of template DNA, there is an increased possibility of contamination, which may produce false positive results. We have guarded against this possibility by including negative controls (PCR reaction buffer, Taq polymerase, primers, and no template DNA) within each group of patient samples and by technical procedures that include cryopreservation of pretested, aliquoted PCR reaction mixes containing all components except for the template DNA. These procedures limit the number of manipulations to a minimum, decreasing the likelihood of contamination of patient samples. Despite the increased sensitivity of the PCR assay, the prevalence of HPV within cytologically negative patients (%) remained far short of recent estimates from other countries that report prevalence rates of 8%. 5 ' - The probability that this results from major differences in sensitivity between the PCR assays is unlikely, as demonstrated by our ability to detect HPV DNA in pg of genomic DNA. The differences between these data may reflect regional differences in prevalence and distribution of HPV types, differences in the definition of a cytologically negative or indefinite population, differences in sampling technique, or a combination of these factors. The increased prevalence of HPV types and 8 within cytologically positive cases as detected by PCR further supports the role of these HPV subtypes in the progression of cervical dysplasia. The relative ease, specificity, and sensitivity of the PCR assay will provide a powerful tool for the clinical identification of patients who may be at increased risk for the development of cervical carcinoma. References. Bedell MA, Jones KH, Laimins LA. The -E7 region of human papillomavirus type 8 is sufficient for transformation of NIH T and Rat- cells. J Virol 987;:5-.. Burmer GC, Rabinovitch PS, Loeb LA. Analysis of c-ki-ras mutations in human colon carcinoma by cell sorting, polymerase chain reaction, and DNA sequencing. Cancer Res 989;9:-.. Cole ST, Danos O. Nucleotide sequence and comparative analysis of the human papillomavirus type 8 genome. J Mol Biol 987:9: Crum CP, Mitao M, Levine RU, Silverstein S. Cervical papillomaviruses segregate within morphologically distinct precancerous lesions. J Virol 985;5: Dartmann K, Schwarz E, Gissmann L, Zur Hausen H. The nucleotide sequence and genome organization of human papilloma virus type. Virology 985;5:-.. Davis LG, Dibner MD, Battey JF. Preparation of DNA from eucaryotic cells. In: Basic methods in molecular biology. New York: Elsevier, 98:-. 7. Demeter T, Kulski JK, Rakoczy P, Sterrett GF, Pixley EC. Detection of human papillomavirus DNA in cell scrapes and formalin-fixed, paraffin-embedded tissue of the uterine cervix by filter in situ hybridization. J Med Virol 988;: De Villiers EM, Wagner D, Schneider A, et al. Human papillomavirus infections in women with and without abnormal cervical cytology. Lancet 987;: Kiviat NB, Koutsky LA, Paavonen JA, et al. Prevalence of genital papillomavirus infection among women attending a college student health clinic or a sexually transmitted disease clinic. J Infect Dis 989;59:9-.. Koss LG. Diagnostic cytopathology, vol. rd ed. Philadelphia: JB Lippincott, 979: Lorincz AT, Temple GF, Patterson JA, Jenson AB, Kurman RJ, Lancaster WD. Correlation of cellular atypia and human papillomavirus deoxyribonucleic acid sequences in exfoliated cells of the uterine cervix. Obstet Gynecol 98;8: Matlashewski G, Schneider J, Banks L, Jones N, Murray A, Crawford L. Human papillomavirus type DNA cooperates with activated ras in transforming primary cells. EMBO J 987;:7-7.. McCance DJ, Clarkson PK, Dyson JL, Walker PG, Singer A. Human papillomavirus types and in multifocal intraepithelial neoplasia of the female lower genital tract. Br J Obstet Gynaecol 985;9:9-.. McCance DJ, Campion MJ, Singer A. Non-invasive detection of cervical papillomavirus DNA. Lancet 98;: McNicol PJ, Guijon FB, Paraskevas M, Brunham RC. Comparison of filter in situ deoxyribonucleic acid hybridization with cytologic, colposcopic, and histopathologic examination for detection of human papillomavirus infection in women with cervical intraepithelial neoplasia. Am J Obstet Gynecol 989;:5-7. Downloaded from on 7 July 8

7 5 BURMER ET AL. A.J.C.P. November 99. Pfister H. Human papillomaviruses and genital cancer. Adv Cancer Res 987;8: Pratili MA, Le Doussal V, Harvey P, et al. Recherche de papillomavirus humains dans des cellules epitheliales du col uterin: frequence des types et 8. Resultats preliminaires d'une etude clinique, cytologique et virologique. J Gynecol Obstet Biol Reprod (Paris) 98;5: Saiki RK, Scharf S, Faloona F, et al. Enzymatic amplification of beta-globin sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 985;: Schneider A, Kraus H, Schuhmann R, Gissmann L. Papillomavirus infection of the lower genital tract: detection of viral DNA in gynecological swabs. Int J Cancer 985;5:-8.. Schneider A, Sawada E, Gissmann L, Shah K. Human papillomaviruses in women with a history of abnormal Papanicolaou smears and in their male partners. Obstet Gynecol 987;9: Schwarz E, Durst M, Demankowski C, et al. DNA sequence and genome organization of genital human papillomavirus type B. EMBOJ 98;:-8.. Schwarz E, Freese UK, Gissmann L, et al. Structure and transcription of human papillomavirus sequences in cervical carcinoma cells. Nature 985;:-.. Seedorf K, KrammerG, Durst M, Suhai S, Rowekamp WG. Human papillomavirus type DNA sequence. Virology 985; 5: Shibata DK, Arnheim N, Martin WJ. Detection of human papillomavirus in paraffin-embedded tissue using the polymerase chain reaction. J Exp Med 988;: Tidy JA, Parry GCN, Ward P, et al. High rate of human papillomavirus infection in cytologically normal cervices. Lancet 989; :.. Tidy JA, Vousden KH, Farrell PJ. Relation between infection with a subtype of HPV and cervical neoplasia. Lancet 989; : Toon PG, Arrand JR, Wilson LP, Sharp DS. Human papillomavirus infection of the uterine cervix of women without cytological signs of neoplasia. Br Med J 98;9:-. 8. Wagner D, Ikenberg H, Boehm N, Gissmann L. Identification of human papillomavirus in cervical swabs by deoxyribonucleic acid in situ hybridization. Obstet Gynecol 98;: Yee C, Krishnan-Hewlett I, Baker CC, Schlegel R, Howley PM. Presence and expression of human papillomavirus sequences in human cervical carcinoma cell lines. Am J Pathol 985; 9:-.. Young LS, Bevan IS, Johnson MA, et al. The polymerase chain reaction: a new epidemiological tool for investigating cervical human papillomavirus infection. Br Med J 989;98:-8.. Zur Hausen H, Gissmann L, Schlehofer JR. Viruses in the etiology of human genital cancer. Prog Med Virol 98;:7-8.. Zur Hausen H, Schneider A. The role of papillomaviruses in human anogenital cancer. In: Salzman NP, Howley PM, eds. The Papovaviridae. The papillomaviruses. New York: Plenum, 987: 5-. Downloaded from on 7 July 8