50 Genitourin Med 1992;68:50-54 Laboratory techniques in the investigation of human papillomavirus infection

Ethel-Michele de Villiers

Introduction The papillomavirus genome Although the existence of a human The genome of a papillomavirus consists of a papillomavirus was demonstrated by electron- number of open reading frames (ORF), each microscopy as early as 1949,' the plurality of capable of coding for a messenger RNA which this group of viruses only became evident in turn will be translated into a protein. These about 30 years later.23 With the advent of gene ORFs are grouped into those coding for technology, methods to clone and characterise proteins active in the nucleus and cytoplasm the viral DNA became available, the result of of the host cell (early proteins), and those which is a list to date totalling 68 different coding for viral capsid proteins (late proteins), genotypes of known human papillomaviruses, the L1 and L2. The L1 ORF contains with probably a number of yet to be identified sequences conserved in all papillomaviruses types still to come. and codes for a group specific antigen.78 The L2 polypeptide has been suggested to be the Plurality of the human papillomaviruses type specific antigen, but such sequences have After the recognition of the existence of dif- not yet been identified unequivocally amongst ferent types of human papillomaviruses those HPV DNAs sequenced. The El has, at (HPV), Orth et a13 demonstrated the presence least in the bovine papillomavirus, been iden- of different HPVs in lesions of patients suffer- tified as the gene responsible for the replica- ing from epidermodysplasia verruciformis. tion and maintenance of the episomal form of Here a number of different HPV types can the DNA molecule within the host cell.910 occur in lesions from the same patient, The E2 controls the transcription rate of the sometimes even more than one type within other early genes,"1 12 such as the E6 and the one lesion. Table 1 presents a very generalised E7, both of which are known as the genes summary of papillomaviruses detected in skin playing a major role in the malignant transfor- lesions (benign and malignant). mation of a host cell.'3 The function of the E4 The association of a papillomavirus infec- gene has not been fully understood. The E4 tion with the development of genital carcin- gene product can be detected within the cyto- oma was postulated by zur Hausen.4 The plasmic inclusion granules of papillomatous rapid progress in papillomavirus research, as lesions'4 and probably acts by binding and well as the development of methods to detect disrupting certain cytokeratins.`5 The func- an HPV infection, have largely been influen- tions of the other ORFs in the HPV genome ced by this association. Not only have 27 have not yet been clarified. different HPV types been isolated from benign The transcription patterns of the ORFs and malignant genital lesions (table 2), but the have been studied in great detail.617 The various aspects of the hypothesis been inves- localisation of the single mRNAs has been tigated and verified on a molecular [zur demonstrated in genital lesions varying from Hausen, in press] and epidemiological5 level. condyloma acuminata, low grade CIN to high Although the available data can only serve as grade CIN to invasive carcinomas.'8 The an indication (table 3),6 papillomaviruses will detection of such transcripts could be of most probably be shown to play an important inportance in the diagnosis of a lesion, that is, etiological role in the development of the which grade in the development of malig- majority of benign and malignant tumours of as well as in head and neck the oral mucosa, Table 2 HPV types present in genital lesions tumours. Condylomata acuminata 6, 11, 42, 44, 51, 55, (53), 67 Intraepithelial neoplasia 6, 11, 16, 18, 30, 31, 33, 34, 35,39,40,42,43,45,51,52, Table 1 HPV types present in skin lesions 56,57,59,61,62,64 Referenzzentrum fur Carcinoma 6,11, 16,18, 31, 33, 35, 39, Humanpathogene Verrucae 1, 2, 3, 4, 10, 26, 27, 28, 29, 45, 51, 52, 54, 56,66 Papillomviren, 38,41,49,57,63,65 Deutsches Epidermoid cyst 60 Butcher's wart 2, 7 Table 3 HPV types present in tumours of the head and Krebsforschungszent- Epidermodysplasia rum, Im Neuenheimer verruciformis 2, 3, 10 neck Feld 506, 6900 5,8,9,12,14,15,17,19,20 Heidelberg, Federal 21, 22, 23, 24, 25, 37, 47, 50 Papillomatoses 6,11, 32 Republic of Germany Bowenoid changes 16, 34, 35 7, 57 (2) E-M de Villiers Squamous cell Focal epithelial carcinoma 5, 8, 14, 17, 20, 47, hyperplasia 13, 32 Accepted for publication 41 Carcinoma 2,6,11, 16,18, 30 30 September 1991 Laboratory techniques in the investigation ofhuman papillomavirus infection 51

nancy. Another factor of increasing impor- Antigen detection tance in the diagnosis is whether the viral The papillomavirus group-specific antigen7 can DNA exists as an episomal molecule or be detected using commercially available tests. whether it is integrated into the cellular gen- Only highly differentiated cells containing large ome, in which case the E2 ORF is usually numbers of viral particles will stain positive, disrupted.'6 The HPV genome is in an indicating that only lesions in which viral episomal state in benign and premalignant capsids are being produced, will be seen as genital lesions, but integrated in the majority HPV positive. This is misleading in so far as the of invasive carcinomas.' 2' majority of lesions induced by, for example, HPV 1 will be seen as positive, whereas lesions Definition of an HPV type containing, for example, HPV 16, will hardly Papillomaviruses need a differentiating cell be detected. The different types apparently layer for the circular double-stranded DNA differ in the frequency with which mature virus molecule to replicate and the capsid proteins particles are being produced. By detecting the to be synthesised and assembled to form group-specific antigen, no distinction can be mature virus capsids. Owing to these special made as to which type is involved in the requirements, papillomaviruses cannot be cul- infection. tivated in the laboratory on a routine basis. Therefore the methods applied to detect a Serology papillomavirus infection have had to be varied The production of individual viral proteins has from the usual serological detection methods to be executed with the help of bacterial and used in other viral infections, to nucleic acid other expression systems. This is a procedure detection. Only lately the use of serological which has been delaying progress not only in diagnostic methods can be envisaged. the development of serological tests, but in the The definition of a papillomavirus type has use ofantibodies directed against proteins from been based on the degree of DNA sequence individual HPV ORFs to detect infection on homology to other known papillomaviruses.22 histological sections. Problems still exist as to These comparisons have in the past been whether the antibodies studied are really type- conducted through hybridisation in liquid specific. Only a limited number of HPV types phase between the DNA genomes of two are at the moment being included. Although no papillomaviruses. If the homology was higher serological tests to be used routinely are at than 50%, the newly isolated clone was regar- present available, the situation is due to change ded as a subtype of the known papillomavirus. in the foreseeable future. If the homology was lower than 50%, it was classified as a new type. The degree of DNA Methods of hybridisation homology between different HPVs varies con- The methods most commonly used at present, siderably. Some types are very closely related all involve the principle of nucleic acid such as HPV 6, HPV 11, HPV 13 and HPV hybridisation. Hybridisation means, in very 55, many of the viruses isolated from epider- simple terms, that two single stranded frag- modysplasia verruciformis lesions, HPV 5, ments of nucleic acid (either DNA or RNA or HPV 8, HPV 12, HPV 19-23, HPV 25, HPV both) will, under the experimental conditions 47 and HPV 50. Others show almost no provided, attach to each other like a perfect degree of homology to any of the known zipper if they originate from a mutual parent types, for example HPV 41. It has become molecule (high stringency), or imperfect (that quite evident that the present form of clas- is, with intermittant loops of discordance) if sification is very unsatisfactory. With the homology exists only in certain segments ofthe increasing number of complete DNA sequen- fragment in question. In the latter, hybridisa- ces of different HPV types becoming available tion conditions can be varied according to the and more unknown HPV types being isolated purpose involved. The DNA sequences of and characterised, the possibility of a reclas- HPV 6 and HPV 11 share such a high degree of sification of the HPVs is in sight. The present homology that they will cross-hybridise even definition of a new type is based on less than under conditions of high stringency. In con- 90% sequence homology in the ORFs E6, E7 trast, HPV 41 DNA will hybridise to any other and L1 to any other known HPV type. HPV DNA only if the conditions are such that duplexes will form at all inspite of numerous mismatches. Such a form ofhybridisation is, in contrast to high-stringency hybridisation, very Methods to detect an HPV infection unstable and can easily be reversed. In order to At present, each of the techniques generally detect any of these hybridisation products, one used to detect any papillomavirus infection has of the single strands involved ("probe") is its advantages, as well as disadvantages. They either radio-labelled or labelled with a product vary greatly in sensitivity and in specificity. which can be made visible by additional None of the tests can be labelled as the ideal enzymatic reactions. The non-labelled single- method to be used as a diagnostic tool, but, as stranded nucleic acid molecule to be tested, is long as the question to be answered is well usually fixed to a solid phase. defined, and seen within the limitations of the test used, any of the following methods can be Probes used. A crucial necessity in the interpretation To avoid false positive results, the HPV DNA of results obtained with any test, is the used as labelled probe, should always be inclusion of adequate controls throughout. separated from the vector used in its produc- 52 de Villiers

tion. The latter could hybridise to sequences this can at present be done with, for example, present in bacteria possibly contaminating the digoxigenin- versus biotin-labelled probes, lesion under investigation. The DNA or RNA large differences in sensitivity (the latter 10-fold probe can be labelled using one ofthe following less sensitive), as well as undesired background methods: staining should be taken into account.33 Other (1) Nick-translation:23 with the help ofDNase, newly recognised advantages of this method is "gaps" are introduced into the double-stran- the distinction between integrated molecules ded DNA into which new DNA strands are (signal localised within the nucleus) versus synthesised with DNA polymerase I incor- episomalDNAmolecules (diffusehybridization porating labelled nucleotides. The specific signal over the entire cell).3"36 Another advan- activity of the resulting probe depends on the tage is the detection of the RNA transcripts of number and activity of the labelled nucleotides the individual ORFs of an HPV genome. The being incorporated. exact localisation ofthese individual transcripts (2) Random-primer synthesis:24 a random pool can help in the diagnosis of the grade of a ofhexanucleotides, being annealed to the DNA lesion.8 molecule, act as primers for the enzymatic synthesis of new DNA strands containing the Filter in situ hybridisation37 labelled nucleotides. As this can be repeated Cells obtained from a scraping or lavage, are several times, the resulting DNA probe not filtered onto a membrane, denatured in situ and only spans the length of the input DNA hybridised to a labelled probe. Although molecule, but the specific activity of the probe thousands of samples can be tested with can be higher than obtained when using nick- relative ease, this advantage is overruled by the translation. many disadvantages of this method. These are: In both these methods the DNA molecules -very low sensitivity (only 50 HPV genome are melted into single strands before being copies per cell or more can be detected), added to the hybridisation solution. -high background combined withnon-specific (3) Synthesis of single-stranded RNA hybridisation, in many cases due to contamin- probes:25 HPV DNA, cloned into a plasmid ating bacteria or blood and mucous present in vector carrying bacteriophage RNA the sample, polymerase binding sites, is transcribed into -the extremely strong dependence of HPV labelled virus-specific single-stranded RNA detection on the manner in which the clinical molecules after addition of RNA polymerase sample was taken (surface area, number and and the appropriate labelled nucleotides. Such nature of cells, blood and mucus present), a probe is free of contaminating bacterial -the limitation ofnumber ofHPV types which sequences and the specific activity obtained is can be applied as probes, very high. -owing to resulting background, non-radioac- (4) Polymerase chain reaction (PCR):2627 The tive probes cannot be used, PCR can also be used to synthesise labelled -adequate controls are neither available nor probes consisting of short nucleotide sequen- can these be imitated in vitro, because the ces. Although the resulting labelling can be constitution of smears differs between each extremely high and very specific, oligonu- individual patient. cleotides are usually 17 to 20 nucleotides in length requiring a modification of regularly Southern blot hybridisation38 used hybridisation conditions to assure specific For routine diagnostic purposes this method is annealing.' too time consuming and labour intensive, al- New methods for labelling probes, using though it can be regarded as the test from which non-radioactive substances, are constantly the most information can be obtained. Cellular being investigated and developed: kits for HPV DNA is cut with selected restriction enzyme(s) detection via chemiluminiscense will probably and electrophoretically separated on an agarose be available in the near future. gel. After denaturation, the DNA is transferred and fixed to a membrane. The latter is then In situ hydridisation hybridised with a labelled HPV probe. The Tissue sections are treated to obtain single- sensitivity ranges between 0,1 and 0,01 HPV stranded DNA molecules. Although the mor- genome copies per cell, depending on the phology of the tissue can be distinguished after amount ofcellular DNA and specific activity of hybridisation and the signal can be located the probe. Questions such as episomall within one cell, the sensitivity of this method is integrated, deletions, HPV type involved in the rather low. Using a radio-labelled probe, one single or even double infections, relatedness of cell has to contain at least 20 to 50 DNA one HPV type to another, detection of un- genome copies to induce a visible signal upon known HPV types and presence of bacterial hybridisation,29 whereas the sensitivity infection can all be answered in the minimum decreases to 350 genome copies or more per cell number of experiments through the critical using non-radioactive probes.30'3 choice of hybridisation conditions and probes. The question whether a distinction can be made between two closely related HPV types Reverse blot hybridisation39 present in the same lesion if the probes are Essentially this method is similar to the usual identically labelled, remains controversial.32 Southern blothybridisation, with the difference Ways to detect double or multiple infections that here the individual HPV types are digested will in future be possible with the use of from the Vector sequences and these samples different labelling/detection systems. Although then electrophoresed on an agarose gel. After Laboratory techniques in the investigation of human papillomavirus infection 53

denaturation and transfer to a membrane, ple is separated by gelectrophoresis. Important hybridisation follows using radio-labelled total in avoiding a false interpretation of the end- cellular DNA (1 gg). The number and range of product, is to separate the amplified DNA by HPV types present on such a blot, that is, tested gelectrophoresis, with subsequent hybridisa- for in one experiment, can be varied according tion to type specific HPV probes, or that the to the problem posed. If one of these HPV amplification samples are digested with restric- types is present in the sample tested, a positive tion enzymes to result in type specific fragment signal will be seen in the ca. 8 kb fragment of sizes. Problems can occur if the L1 ORF is the specific type. If the HPV type is represen- deleted in a tumour in which the HPV DNA tative of a group of closely related HPVs, this has been integrated. This is often the case in result could be confirmed with a subsequent carcinomas.`6 For this reason, the use of Southern blot using the HPV type in question primers in the E6 ORF4` or E7 ORF4' might be as probe. The advantage of this method is that, advisable, even though the range ofHPV types with an input ofthe minimum amount ofDNA, detectable might be restricted. Irrespective of only one experiment is needed to test for as which primers are being used, a percentage of large a number of HPV types as wished. The the probes could still be uninterpretable45 or sensitivity is 5 HPV genome copies per cell designated as "new HPV types".4' ' The latter using nylon membrane and a radio-labelled could only be accepted if these "HPV types" probe. An additional advantage is the detection would be isolated and characterised as a full of contaminating bacterial sequences in the length DNA genome. Regarding the short time sample-the vector DNA present on the blot interval in which this method has been availa- will give a positive signal. Only stringent ble, as well as the many influencing factors, hybridisation conditions can be used, owing to even, for example type of fixative used or the fact that the genomes of several ofthe HPV fixation time,47 it is not surprising that the types cross-hybridise to cellular DNA sequen- present state of the HPV detection rate using ces under relaxed experimental conditions. the PCR is fairly confusing. Time and The specific genome sections/ORFs of the experience is needed to sort out the facts. The HPVs which share these homologous sequen- use ofthe PCR will probably expand in future. ces, vary from type to type, rendering a general For example methods are being developed to rule impossible [de Villiers, Kyono, Silver- amplify the DNA in the intact cells of a stein, Delius and zur Hausen, unpublished histological section (Nuovo, personal com- results]. munication). By doing this the localisation of the amplified DNA could be determined within Dot/slot hybridisation the context of intact cellular morphology. This method differs from the two preceeding in that the DNA to be fixed onto the membrane, is not electrophoretically separated, but fixed as a Comparisons "dot" (drop) or "slot" (manifold used). Many Many investigators have recently used two of the commercial kits available make use of or more of the above mentioned methods on this method. Upon evaluating the results, no the same samples to compare sensitivity and estimation can be made on the degree of non- specificity. As mentioned previously, each test specific hybridisation, therefore adequate con- can be used, provided the user is aware of its trols should be run parallel together with limitations, positive or negative. A mistake probes chosen which would not contain con- with confusing consequences is the division of taminating vector sequences. clinical samples before comparing different methods. If, for example, a biopsy is divided Polymerase chain reaction6 27 and the separate sections tested individually, The PCR is rapidly becoming the most completely different results can be obtained. frequently used technique in demonstrating an One area ofthe lesion could contain HPV DNA HPV infection. With the use of selected and the other not. Similarly, if the same primers, a certain genome segment is amplified instruments are applied to take more than one through ca. 30-40 cycles of denaturation/ biopsy, a carry-over of viral DNA/particles annealing/amplification with a temperature could occur between biopsies. resistant DNA polymerase. Owing to the extremely high sensitivity of this method (one HPV genome copy in 100 000 cells), The value ofHPV detection experimental conditions have to be very tightly Several studies have been conducted to deter- controlled to rule out contamination and mine the prevalence rate of genital HPV infec- therefore false positive results. This contamin- tion in healthy individuals. A negative result ation could occur at any stage, even as early as obtained from one sample is not necessarily the moment the sample is being taken from the indicative of the absence of an HPV infection. patient. Several possibilities to avoid the pos- Again, apart from the many varying handling sibility ofcontamination have been introduced, and experimental conditions, a physiological such as the boiling of samples to exclude a step variation can be detected. Multiple sampling of of DNA extraction.' The most generally used one women within a time span, results in the primers are chosen in the Li ORF of the HPV fluctuation between positivity and negativity genome.41`4 These primers are seen as consen- [de Villiers et al, in press]. Although the sus primers with which a large spectrum of ultimate proof has not yet been obtained, HPV types can be identified. After the indications are that the HPV infections persist amplification has been done, the resulting sam- as latent infections, with intermittent cycles of 54 54 de Villiers

or viral No drug is yet 23 Rigby RWJ, Dieckmann M, Rhodes C, Berg P. Labelling replication production. deoxyribonucleic acid to high specific activity in vitro by available to succesfully treat or remove these nick translation with DNA polymerase I. J Mol Biol subclinical infections. Therefore, unless a clin- 1977;113:237-51. 24 Feinberg FP, Vogelstein B. A technique for radiolabelling ical lesion or abnormal cytology has in addition DNA restriction endonuclease fragments to high specific been the detection of an HPV activity. Anal Biochem 1983;132:6-13. diagnosed, 25 Melton DA, Krieg PA, Rebagliati MR, Maniatis T, Zinm infection should at the present time not be of K, Green MR. Efficient in vitro synthesis of biologically much Good evidence exists active RNA and RNA hybridization probes from plasmids consequence. containing a bacteriophage promotor. Nucl Acid Res especially at the molecular biological [zur 1984;12:7035-56. in and levels548 26 Mullis KB, Faloona FA. Specific synthesis of DNA in vitro Hausen, press] epidemiological via a polymerase-catalyzed chain reaction. Methods indicating that an HPV infection alone is not Enzymol 1987;155:335-50. sufficient for the development of a malignant 27 Saiki RK, Gelfand DH, Stoffel 5, et al. Primer-directed enzymatic amplification of DNA with a thermostabile lesion, but that additional host cell modifica- DNA polymerase. Science 1988;239:487-91. tions are needed for the cascade of events 28 Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning. A Laboratory Manual. Cold Spring Harbor, Cold Spring required for malignant conversion. This may Harbor Laboratory Press, 1989. be mediated as well as by 29 Schneider A, Meinhardt G, Kirchmayr R, Schneider V. by endogenous Prevalence of human papillomavirus genomes in tissues exogenous additional factors. from the lower genital tract as detected by molecular in situ hybridization.In Jt Gynecol Pathol 1991;10:1-14. 30 Beckmann AM, Myerson D. Diagnosis of human papillomavirus and human cytomegalovirus by hybridiza- 1 Strauss MJ, Shaw EW, Bunting H, Melnick JL. tion histochemistry. In: Tenover FC, ed. DNA Probes for "Crystalline" virus-like paricles from skin papillomas Infectious Diseases. Boca Raton, CRC Press, 1989;145-62. characterized by intranuclear inclusion bodies. Proc Soc 31 Crum CP, Nuovo G, Griedman D, Silverstein SV. A Exp Biol Med 1949;72:46. comparison of biotin and isotope-labeled ribonucleic acid 2 Gissmann L,zur Hausen H. Human papillomavirus: probes for in situ detection of HPV 16 ribonucleic acid in Physical mapping and genetic heterogeneity. Proc Natl genital precancers. Lab Invest 1988;58:354-9. Acad Sci USA 1976;73:1310-3. 32 Herrington CS, Graham AK, Flannery DM, Burns J, 3 Orth G, JablonskaS, Favre M, Croissant 0, Jarzabek- McGee JO. Discrimination of closely homologous HPV Chorzelska M, Rzesa G. Characterization of two types of types by nonisotopic in situ hybridization: definition and human papillomaviruses in lesions of epidermodysplasia derivation of tissue melting temperatures. Histochem J verruciformis. Proc Nat Acad Sci USA 1978;87:8170-4. 1991;22:545-54. 4 zur Hausen H. Condylomata acuminata and human genital 33 Morris RG,Arends MJ, Bishop PE, Sizer K, Duvall E, Bird cancer. Cancer Res 1976;36:794. CC. Sensitivity of digoxigenin and biotin labelled probes 5 Ley C, Bauer HM, Reingold A, Schiffnan MH, Chambers for detection of human papillomavirus by in situ JC, Tashiro CJ, Manos MM. Determinants of genital hybridiazation. J Clin Pathol 1990;43:800-5. human papillomavirus infection in young women. J Nat 34 Park JS, Kurman RJ, Kessis TD, Shah KV. Comparison of Cancer Inst 1991;83:997-1003. peroxidase-labeled DNA probes with radioactive RNA 6 de Villiers E-M. Viruses in cancers of the head and neck. In: probes for detection of human papillomaviruses by in situ Bearing of Basic Research on Clinical Otolaryngology. hybridization in paraffin sections. Mod Pathol 1991;4: Pfaltz CR, Arnold W, Kleinsasser 0, eds. Adv Otorhin- 81-5. olaryngol 1991;46:116-23. 35 Cooper K, Herrington CS, Graham AK, Evans MF, McGee 7 Jenson AB, Rosenthal JD, Olson C, Pass F, Lancaster WD, JO. In situ evidence for HPV 16, 18, 33 integration in Shah K. Human papillomavirus: frequency and distri- cervical squamous cell cancer in Britain and South Africa. bution in plantar and common warts. Lab Invest J Clin Pathol 1991;44:406-9. 1980;47:491-7. 36 Wolber RA, Clement PB. In situ DNA hybridization of 8 Kurman RJ, Jenson AB, Lancaster, WD. Papillomavirus cervical small cell carcinoma and adenocarcinoma using infection of the cervix. II. Relationship to intraepithelial biotin-labeled human papillomavirus probes. Mod Pathol neoplasia based on the presence of specific viral structural 1991;4:96-100. proteins. AmJ Surg Pathol 1983;7:39-52. 37 Wagner D, Ikenberg H, Bohm N, Gissmann L. Identifica- 9 Lusky M, Botchan MR. Characterization of the bovine tion of human papillomavirus in cervical swabs by de- papillomavirus plasmid maintenance sequences. Cell oxyribonucleic acid in situ hybridization. Obstet Gynecol 1984;36:391-401. 1984;64:767-72. 10 Lusky M, Botchan MR. Genetic analysis of bovine 38 Southern EM. Detection of specific sequences among DNA papillomavirus type 1 trans-acting replication factors. J fragments separated by gel electrophoresis. J Mol Biol Virol 1985;53:955-65. 1975;98:503-17. 11 Spalholz BA, Yang Y-C, Howley PM. Transactivation of 39 de Villiers E-M, Schneider A, Gross G, zur Hausen H. bovine papillomavirus transcriptional regulatory element Analysis of benign and malignant urogenital tumors for by the E2 gene product. Cell 1985;42:183-91. human papillomavirus infection by labelling cellular 12 Lambert PF, Spalholz BA, Howley PM. A transcriptional DNA. Med Microbiol Immunol 1986;174:281-6. repressor encoded by BPV-1 shares common carboxy- 40 van den Brule AL, Meijer CJ, Bakels V, Kenemans P, terminal domain with the E2 transactivator. Cell 1987; Walboomers J. Rapid detection of human papillomavirus 50:69-78. in cervical scrapes by combined general primer-mediated 13 Crook T, Morgenstein JP, Crawford L, Banks L. Continued and type-specific polymerase chain reaction. J Clin expression of HPV 16 E7 protein is required for mainten- Microbiol 1990;28:2739-43. ance of the transformed phenotype of cells cotransformed 41 Bauer HM, Ting Y, Greer CE, et al. Genital human by HPV 16 plus EJ-ras. EMBOJ 1989;8:513-9. papillomavirus infection in female university students as 14 Doorbar J, Coneron I, Gallimore PH. Sequence divergence determined by a PCR-based method. JAMA 1991; yet conserved physical characteristics among the E4 265:472-7. proteins of cutaneous human papillomaviruses. Virology 42 Snijders PJF, van den Brule AJC, Schrijnemakers HFJ, 1989;172:51-62. Snow G, Meijer CJLM, Walboomers JMM. The use of 15 Doorbar J, Ely S, Sterling J, McLean C, Crawford L. general primers in the polymerase chain reaction permits Specific interaction between HPV-16 E1-E4 and cyto- the detection of a broad spectrum of human keratins results in collapse of the epithelial cell inter- papillomavirus genotypes.J Gen Virol 1990;71:173-8 1. mediate filament network. Nature 1991;352:824-7. 43 Yoshikawa H, Kawana T, Kitagawa K, Mizuno M, Yoshik- 16 Schwarz E, Freese UK, Gissmann L, Mayer W, Roggen- ura H, Iwamoto A. Detection and typing of multiple buck B, Stremlau A, zur Hausen H. Structure and genital human papillomaviruses by DNA amplification transcripton of human papillomavirus sequences in with consensus primers.JpnJ Cancer Res 1991;82:524-31. cervical carcinoma cells. Nature 1985;314:111-4. 44 Cornelissen MT, van dan Tweel JG, Struyk AHB, Jebbink 17 Yee C, Krishnan-Hewlett Z, Baker CC, Schlegel R, Howley MF, Briet M, van der Noordaa J, ter Schegget J. PM. Presence and expression of human papillomavirus Localization ofhuman papillomavirus type 16 DNA using sequences in human cervical carcinoma cell lines. Am J the polymerase chain reaction in the cervix uteri of women Pathol 1985;119:361-6. with cervical intraepithelial neoplasia. J Gen Virol 1989; 18 Stoler MH, Rhodes CR, Whitbeck A, Chow LT, Broker 70:2555-62. TR. Gene expression of HPV types 16 and 18 in cervical 45 Gravitt P, Hakenworth A, Stoerker J. A direct comparison of neoplasia. In PM Howley and TR Broker, eds. methods proposed for use in widespread screening of Papillomaviruses. Wiley-Liss, Inc., 1990;1-1 1. human papillomavirus infections. Mol Cell Probes 1991; 19 Durst M, Kleinheinz A, Hotz M, Gissmann L. The physical 5:65-72. stateofhuman papillomavirus type 16 DNA in benign and 46 Fujinaga Y, Shimada M, Okazawa K, Fukushima M, Kato I, malignant genital tumors. J Gen Virol 1985;66:1515-22. Fujinaga K. Simultaneous detection and typing of genital 20 Matsukura T, Koi S, Sugase M. Both episomal and human papillomavirus DNA using polymerase chain integrated forms of human papilomavirus type 16 are reaction. J Gen Virol 1991;72:1039-44. involved in invasive cervical cancers. Virology 1989; 47 Greer CE, Peterson SL, Kiviat NB, Manos MM. PCR 172:63-72. amplification from paraffin-embedded tissues. Effects of 21 Cullen AP, Reid R, Campion M, Lorincz AT. Analysis of fixative and fixation time. Am J Clin Pathol 1991;95: thephysicalstateofdifferenthumanpapillomavirus DNAs 117-24. in intraepithelial and invasive cervical neoplasm. J Virol 48 Kjaer SK, Engholm G, Teisen C, et al. Risk factors for 1991;65:606-612. cervical human papillomavirus and herpes simplex virus 22 Coggin JR, zur Hausen H. Workshop on papillomaviruses infections in Greenland and Denmark: A population- and cancer. Cancer Res 1979;39:545-6. based study. Am JEpidermiol 1990;131:669-82.