488 J Clin Pathol 1990;43:488-492

Production and characterisation of a monoclonal antibody to human papillomavirus type 16 using recombinant vaccinia virus

C S McLean, M J Churcher, J Meinke, G L Smith, G Higgins, M Stanley, A C Minson

Abstract subjects. The ubiquity of viral DNA highlights A monoclonal antibody was raised the need for reagents that will detect viral against the major capsid protein LI of proteins. Several groups have recently tried to human papillomavirus type 16, using a overcome the problems associated with the lack recombinant vaccinia virus that ex- of a permissive in vitro culture system by presses the LI protein, as a target for cloning and expressing viral genes in prok- screening. This antibody, designated aryotic and eukaryotic expression systems."'2 CAMVIR-1, reacted with a 56 kilodalton In an earlier paper we reported the construc- protein in cells infected with LI-vaccinia tion of a vaccinia virus recombinant containing virus, and the protein was present in a the HPV 16 Li gene (encoding the major predominantly nuclear location. The capsid antigen), and the preparation of poly- antibody also detects the HPV-16 Li clonal antisera raised against a bacterial fusion antigen in formalin fixed, paraffin wax protein containing HPV 16 LI sequences.'3 embedded biopsy specimens and on Polyclonal antisera, however, are oflimited use routine cervical smears. The antibody in that they cannot give information about the reacts strongly and consistently with presence of type specific epitopes on the LI biopsy specimens containing HPV-16 or protein and tend to result in high background HPV-33, but very weak reactions were in immunohistochemical and immunoprecipi- occasionally observed with biopsy tation studies. There is a need for type specific specimens or smears containing HPV-6 reagents for diagnostic as well as for research or HPV- 1. purposes for viruses such as HPV 16, which has The potential advantages of using a been identified in 50-600/o of cases of cervical vaccinia recombinant are (i) the target carcinoma subjected to viral typing.5 protein is synthesised in a eukoryotic cell In this report we describe the production ofa so that its "processing" and location are monoclonal antibody to HPV 16 L1 using a normal; (ii) cells infected with vaccinia novel protocol in which viral antigen is gen- recombinants can be subjected to various erated for immunisation using a bacterial fixing procedures similar to those used expression system and hybridoma supernants for routine clinical material. This greatly are screened for specific antibody using a increases the probability that an identi- vaccinia recombinant expressing the target fied antibody will be useful in a clinical protein. setting. Methods Human papillomaviruses (HPV) are small The myeloma cell line NSO and established DNA viruses which cause a variety of pro- hybridoma cell lines were maintained in liferative lesions on epithelial or mucosal sur- Glasgow modified Eagle's medium (GMEM) Department of faces.' More than 50 HPV types have now been containing 50o fetal calf serum (FCS). African Pathology, Divisions are cells were main- of Cellular Pathology identified, and several of these associated green monkey kidney (CV-1) and Virology, with premalignant and malignant disease ofthe tained in the same medium containing 100/0 University of anogenital region. In particular, HPV 6 and 11 FCS. Baby hamster kidney 21 clone 13 cells Cambridge have been in were in GMEM C S McLean identified benign lesions; HPV 16, (BHK) maintained containing M J Churcher 18, 31 and 33 are highly significantly associated 1000 tryptose phosphate broth and 1000/ new- J Meinke with all grades of cervical intraepithelial neo- born calf serum. The construction of recom- G L Smith plasia (CIN) and cervical cancer.2 3 binant vaccinia virus (VLlRK) expressing the M Stanley A C Minson There is no suitable tissue culture system for entire HPV 16 LI open reading frame under Division of Virology, the propagation of these viruses and the iden- the control of a vaccinia late promoter has been Institute of Medical tification of viral DNA and classification of described previously.'3 The recombinant virus and Veterinary viruses is based on DNA homology using and the parent WT vaccinia virus (strain WR) Sciences, Adelaide, were in BHK cells and titrated in CV-1 South Australia hybridisation techniques.4' Viral DNA grown G Higgins sequences have been identified and typed in cells. Correspondence to: abnormal cervical lesions and in cervical Dr C S McLean, smears using blot hybridisation techniques, MOUSE IMMUNISATION AND CELL FUSION Department of Pathology, University of Cambridge, and more recently by the polymerase chain Mice were immunised subcutaneously with 20 Cambridge CB2 IQP reaction.7 The exquisite sensitivity of the latter ig of a # galactosidase-LI fusion protein (con- Accepted for publication has resulted in the detection of low copy viral taining amino acids 198-C terminus ofHPV-16 15 February 1990 DNA in histologically and cytologically normal L113) in complete Freund's adjuvant. After Monoclonal antibody to HPV 16 489 four weeks a series ofsubsequent doses of 15 Mg routine colposcopy clinic. Sections were protein in Freund's incomplete adjuvant were dewaxed and rehydrated through a xylene given intraperitoneally at two weekly intervals. alcohol series. Before immunostaining with the A final intravenous immunisation of 15 Mg was Vectastain ABC complex (Vector Labora- given three days before the fusion. The spleen tories, Peterborough, England) sections were was removed aseptically and fused to the mouse incubated in 0-10% trypsin at 37°C for five hybridoma NSO, using the original proto- minutes and quenched in 0-3% H202 for 15 col.'41 5 Cells were plated at a density of2 x 106/ minutes. well in 24-well Linbro plates, and the growth Routine cervical Papanicolau smears were medium changed every two to three days. fixed in 500 formaldehyde in PBS for five Positive hybrids were cloned twice by limiting minutes, washed in PBS, permeabilised in 1% dilution, and clones secreting antibody to L1 Triton and subsequently treated with trypsin, were grown in 75 cm2 flasks and stored in liquid quenched, and stained in the same way as the nitrogen. cervical sections.

SCREENING OF HYBRIDOMA SUPERNATANTS IN SITU HYBRIDISATION USING ELISA Sections of cervical lesions embedded in paraf- The supernatants of wells containing hybrids fin wax were dewaxed in xylene and alcohol and were tested in an ELISA as follows: CV-1 cells hybridised with probes containing DNA of were grown in 5 mm wells to confluency and HPV 16/18, HPV 31/33, or HPV 6/1 1, using infected with 30 plaque forming units/cell WT the Vira Type in situ hybridisation kit (Life or VL1RK vaccinia virus. After 18 hours at Technologies Inc, Gaithersburg, Maryland, 37°C cells were fixed in 5% formaldehyde, USA) according to the manufacturer's instruc- 10% sucrose, in phosphate buffered saline tions. In certain cases in situ hybridisation was (PBS), and air dried. Before use cells were carried out using 35S-labelled DNA, according permeabilised with 1% Triton X-100, 10% to the method described by Gowans et al.'6 sucrose in PBS. After blocking with 5% FCS, Briefly, slides were rehydrated, digested with monolayers were reacted with hybridoma proteinase K and pronase, and treated with supernatants, followed by incubation with acetic anhydride before hybridisation with 35S- peroxidase-linked protein A, each for 30 min- labelled HPV-16 DNA. After coating with utes at room temperature. The wells were then radiographic emulsion slides were stored for reacted with 0-phenylene diamine and the eight days at 4°C and developed, fixed, and absorption of the supernatants at 492 nm was stained with haematoxylin. measured in an ELISA reader.

IMMUNOFLUORESCENCE Results CV-1 cells were grown on coverslips and Mice repeatedly immunised with the f galac- infected, fixed, and permeabilised as described tosidase-HPV 16 Li fusion protein developed for the ELISA method. Cells were reacted with strong responses to the fusion protein, but the hybridoma supernatant and subsequently response to the vaccinia recombinant express- incubated with an anti-mouse IgG fluorescein ing HPV-16 Li was variable. Some mice also conjugate. gave high "background" responses to cells infected with wild type vaccinia virus. Spleen WESTERN BLOT ANALYSIS cells from mice showing a good differential BHK-cell monolayers, infected with 30 pfu/ response in ELISAs to the Li recombinant cell WT or VL1RK vaccinia virus, were har- vaccinia virus relative to the wild type virus vested in PBS 20 hours after infection, lysed in were used in fusions with cell line NSO. The sample buffer (24 mm TRIS-HCI, pH 6-18; results of a series of three fusions were dis- 100 mm dithiothreitol; 2% sodium dodceyl appointing in that only two positive hybridoma sulphate (SDS); 20% glycerol; 0-02% bromo- supernatants were identified by ELISAs on phenol blue), and sonicated. Samples were cells infected with VL1RK. Of these, only one denatured at 100°C for three minutes, subjec- retained activity following two rounds of clon- ted to SDS-polyacrylamide gel electrophoresis ing by limiting dilution. This hybridoma was (PAGE) and the proteins transferred to nitro- stable to subsequent passage and subcloning cellulose by electroblotting as described and was designated CAMVIR-1. previously.9 After blocking in 5% dried milk powder in PBS for 30 minutes at room PROPERTIES OF CAMVIR-1 temperature, filters were incubated with 1/2 Tissue culture supernatant from CAMVIR-1 diluted supernatant for one hour at room reacted with a 56 kilodalton protein in Western temperature, washed three times in PBS-1% blots of cell lysates infected with VL1RK (fig NP40 and reacted with affinity purified 125I1 la). A minor species of 64 kilodaltons was also protein A (Amersham International, UK) at a detected. This probably represents a glycosy- concentration of 0 I pCi-ml. The blots were lated form ofthe protein as it is absent in lysates washed again in PBS-i% NP40, dried, and ofcells infected in the presence oftunicamycin, autoradiographed. an inhibitor ofN-glycosylation. Similar results were reported using a polyclonal rabbit serum IMMUNOCYTOCHEMICAL STUDIES raised against an HPV-16 LI fusion protein." Formalin fixed, paraffin wax embedded sec- No reactivity was observed against cells infec- tions were obtained from diagnostic biopsy ted with wild type vaccinia. The antibody also specimens taken from patients attending a reacted with the fusion protein used for im- 490 McLean, Churcher, Meinke, Smith, Higgins, Stanley, Minson

Figure I Western blot a b c d lesion staining positive with CAMVIR-1 con- analysis of reactivity of CAMVIR-1 with LI tained HPV31/33DNA. In most cases lesions fusion protein and LI containing HPV 6 or HPV 11 did not stain with vaccinia recombinant. (a) CAMVIR-1, but occasionally a few cells in the CVI cells were infected with Li-vaccinia (lane a) - pH X2 superficial layers of these sections would show or wild type vaccinia (lane nuclear staining. We also compared the sen- b) at 30 pfu/cell, and lysed sitivity of CAMVIR-1 in detecting LI protein after 20 hours. Samples to that ofa commercially available anti-BPV-l- were subjected to SDS- PAGE electrophoresis, antibody (Dako). On sections positive for and after transfer to HPV-16 or -33 little or no staining was nitrocellulose the blots were obtained with the Dako antibody, although incubated with CAMVIR-I supernatant CAMVIR-1 gave a strongly positive nuclear diluted at l/I0,followed reaction (figs 2c and d). by affinity purified 125I- protein A at a concentration of 0 1 DETECTION OF LI PROTEIN IN CERVICAL SMEARS pCi/ml. After washing, the 56kD-- The ability of CAMVIR-1 to detect L1 protein blots were dried and in routine cervical smears was investigated in a autoradiographed. (b) 01 pg of the purified blind trial. Samples were taken from 50 general f-galactosidase LI fusion practice patients. DNA extractions were per- protein (pHX2, 1A lB formed on exfoliated cells which were sub- containing amino acids to 221-C terminus (lane d) jected dot blot hybridisation with probes and 0)1 pg or an irrelevant specific for HPV 6, 11, 16 or 18. The smears fusion protein (containing munisation but not with an irrelevant fusion were examined cytologically and -stained with amino acids 1-220 of the protein containing the N-terminal p*ortion of CAMVIR-1 tissue culture supernatant at a HPV-16 Ll protein, (lane the HPV-16 Li protein (ammno acids c) were subjected to 1-197). 1/100 dilution using the biotin-streptavidin- electrophoresis on a 125% To determine the cellular localisatiion of the* peroxidase system. The cells were fixed briefly SDS-acrylamide gel and antigen detected by ChAMVIR 1, ce.lls were in 5% formaldehyde in PBS, and trypsin was transferred to infected with recombinant or WT virus and nitrocellulose. The blots added before reaction. In 14 of the 50 cases an were reacted with after formaldehyde fixation readt:ed with insufficient number of cells was present on the CAMVIR-1 at a dilution CAMVIR-1 and fluorescein-labeliled anti- smear to determine if positively stained cells of 1/20,followed by mouse immunoglobullin. In cells mnfelcted with peroxidase-linked anti- were present. Of the remaining 36 cases, nine mouse IgG (Amersham Li vaccinia strong nuclear flulorescq:ence was contained cells showing nuclear staining and International, UK) at a observed, while no reaction coulld be bdetected were of koilocytic morphology (figs 2e and f). dilution of 1/200. Bands in cell monolayers infected with WT fvirus (figs Decoding showed that two of these contained were visualised using 2g and h). An diaminobenzidine irrelevant monoclonal antibody HPV-16 DNA by dot blot hybridisation (be- tetrahydrochloride of the same subclass (IgG2a) did notxreact with tween 2 x 105'106 genomic copies). In four (DAB) at a concentration the recombinant infected cells. cases no HPV DNA could be detected (less of I mg/ml. than 105 genomic copies). In one case HPV REACTIVITY OF CAMVIR-1 ON TISSUE SE( CTIONS DNA was present at low copy number but it The ability of CAMVIR-l to detec:t the LI was not possible to determine which type. In protein in formalin fixed, paraffin wax em- two cases where positive staining with bedded tissue sections was examinled using CAMVIR-1 was observed a very high copy a biotin-streptavidin-peroxidase staining number ofHPV-11 was present (more than 106 method. Sections were obtained Ifrom the genomic copies a cell). In three cases hybridisa- cervical biopsy specimens of patiebnts with tion showed the presence ofHPV-16 DNA, but abnormal cervical smears, attending a routine no staining with CAMVIR-1 was observed clinic for colposcopy and examination. These because the smear was inadequate. lesions were histologically graded as III. The sections were tested for the presence of Li protein using CAMVIR-1 tissuie culture Discussion supernatant at a concentration of 1/ 1010. Figure The absence ofa suitable in vitro system for the 2a shows a lesion in which the nuclei of several propagation of human papillomaviruses, to- cells in the superficial layers are strongly gether with the fact that in clinical lesions only stained. These cells have the morp}hology of very low levels of virus are present, has neces- koilocytes. A parallel section of this 14esion was sitated the use of alternative strategies for the subjected to in situ hybridisation to establish production of immunological reagents to these the presence and type of viral DNA (fig 2b). viruses. In the production of monoclonal anti- The cells with the maximum copy number of bodies suitable targets for the screening of HPV 16 DNA are found in the same region of hybridoma fluids are an important require- the section as the cells containing LI Iprotein- ment. To date, the monoclonal antibodies that is, in the superficial layers ofthe 1Iesion. We reported have been raised and screened against examined a substantial number of lesiions (n = bacterial derived proteins.'6"8 We used cells 50), and in those cases where stainiing with infected with a vaccinia virus recombinant CAMVIR-1 was observed, in situ hLybridisa- expressing the HPV 16 LI open reading frame tion showed high copy numbers of HPV-16 as a target for the selection of monoclonal DNA in that region of the section. The op- antibodies. This system provided a very con- posite was not true, however-that iis, a min- venient and plentiful source of viral protein, ority of lesions were positive for HPV!-16 by in free of potentially crossreactive bacterial situ hybridisation but showed no staiining with elements. CAMVIR-1. In one instance we foumd that a The antibody against the HPV-16 LI pro- Monoclonal antibody to HPV 16 491

Figure 2 Immunocytochemical staining with CAMVIR-1 of sections crf cervical biopsy specimens containing HPV-16 or HPV-33, routine cervical smears, and cells infected ,.AO with the Li vaccinia recombinants. (a) Staining with CAMVIR- 1 at 1/100 dilution on formolfixed, paraffin wax embedded section of a cervical lesion containing A B HPV-16. Sections were dewaxed and rehydrated through a xylene alcohol series, incubated in 0 100 'I trypsin at 37°Cforfive .' 4 minutes, quenched in 033% '. H202for 15 minutes, and Z * s~ stained using the 4 Vectastain ABC complex. (b) In situ hybridisation 4 using .t * te ~4 * 8 * 35S-labelled total HPV-16 DNA on the section >%'i.:S* ^ s * :.s~~~ e :... adjacent to that infig 2a. (c) Staining with CAMVIR-1 at 1/100 on a cervical lesion containing HPV-33 using the same C D protocol as infig 2a. (d) Staining of the same lesion as infig 2c, using a commercially available polyclonal anti-BPV- antibody (Dako) at 1/1 00,followed by the Vectastain ABC complex. (e andf) Staining with CAMVIR-1 at 1/100 of routine cervical smears fixed in 500 formaldehyde-PBSfor five minutes, permeabilised in l °O Triton, and subsequently treated with trypsin, quenched, and stained in the same way as E F the cervical sections. (g and h) Reactivity of CAMVIR-1 at 1/100 dilution on CV-1 cells infected with 30 pfu/cell Li vaccinia (g) or wild type vaccinia (h). Cells werefixed in 50o formaldehyde-PBS for five minutes, washed in PBS, and after incubation with CAMVIR-1 for 30 minutes at room temperature, washed again in PBS, and reacted with FITC-labelled anti-mouse IgG at 1/70 (Dako).

tein reacts with a protein of the expected would show nuclear staining. From the results molecular weight in recombinant vaccinia virus on cervical smears it is clear that cells contain- infected cells and detects the protein in a ing very high copy numbers of HPV-1 1 also predominantly nuclear location, corresponding give positive results with CAMVIR- 1. A sub- to earlier findings using polyclonal anti-Ll- stantial number of lesions containing HPV-6, antisera. 13 however, were totally negative when reacted We also obtained nuclear staining of with CAMVIR- 1. The antibody probably morphologically koilocytic cells on formalin binds, to a much lesser extent, to cells contain- fixed, paraffin wax embedded tissue sections ing HPV-6 or -1 1, so that only cells with a very using CAMVIR-1. It was clear from the sub- high viral DNA copy number show positive sequent typing of these sections using in situ staining. The epitope to which CAMVIR-1 is hybridisation that CAMVIR-1 binds strongly directed was identified as GFGAMDF (a.a.n- to cells containing HPV-16 or -33. Very umber 230-236) using overlapping peptides in occasionally, however, a number of cells in the an ELISA (Cason J, McCance D, personal superficial layer of an HPV-6 containing lesion communication). This region is highly con- 492 McLean, Churcher, Meinke, Smith, Higgins, Stanley, Minson

served among the different HPV types, and CAMVIR-1 may be a useful reagent in this there is one amino acid substitution between context. HPV16 and HPV-6 or -11 in this region, We are grateful to Drs Lynn Kennedy and Ian Frazer of the namely, D to N (GFGAMNF), which would Department of Medicine and Lions Immunology Laboratory, explain the weak cross reactivity with these University of Queensland, Australia, for the provision of HPV typed smears; the Histopathology Department, Addenbrooke's virus types. CAMVIR-1 identifies an epitope Hospital, Cambridge, for tissue sections of cervical biopsies; and on HPV-1 6 LI that is strongly reactive against to Mrs Mary Wright for typing the manuscript. HPV-16 and -33 and, to a much lesser extent, with HPV-6 and -11, and thus is able to identify a subgroup ofhuman papillomaviruses with malignant oncogenic potential. zur Hausen H. Human papillomaviruses and their possible role in squamous cell carcinomas. Curr Top Microbiol Expression of the L1 capsid protein, as Immunol 1977;78:1-30. identified with this monoclonal antibody, is 2 Durst M, Gissmann L, Ikenburgh H, zur Hausen H. A papillomavirus DNA from a cervical carcinoma and its strongly focal, and in any one lesion only a few prevalence in cancer biopsy samples from different cells with high copy viral DNA synthesise LI geographic regions. Proc Natl Acad Sci USA 1983;80: 3812-5. protein. This situation differs from that seen in 3 Campion MJ, McCance DJ, Cuzick J, Singer A. The some other lesions induced by papillomavirus, progressive potential of mild cervical atypia: a prospective cytological, colposcopic and virological study. Lancet particularly cutaneous warts'9 and bovine 1986;ii:237-40. papillomas20 in which many cells express viral 4 Schneider A, Kraus H, Schuhmann R, Gissmann L. Papil- lomavirus infection of the lower genital tract: detection of capsid proteins and in which viral particles are viral DNA in gynaecological swabs. Int J Cancer 1985; present in substantial numbers. The focality of 35:443-8. 5 Meanwell CA, Cox MF, Blackledge G, Maitland NJ. HPV- Li expression in the cervical lesions may be 16 DNA in normal and malignant cervical epithelium: related to the paucity of viral particles in lesions implications for the aetiology and behaviour of cervical neoplasia. Lancet 1987;i:703-7. containing HPV 16. Alternatively, it may be 6 Wickenden C, Steele A, Malcolm ADB, Coleman DV. related to the stage of infection at which the Screening for wart virus infection in normal and abnormal cervices by DNA hybridisation of cervical scrapes. Lancet cervical lesion has been biopsied. The natural 1 987;i:65-7. history of HPV- 16 related disease is virtually 7 Young LS, Bevan IS, Johnson MA, et al. The polymerase chain reaction: a new epidemiological tool for investi- unknown. It is evident from studies on other gating cervical human papillomavirus infection. Br Med J papillomaviruses, particularly CRPV and 1989;298:14-8. 8 Li CH, Shah KV, Seth A, Gilden R. Identification of the BPV-1, that the development of a lesion in human papillomavirus type 6bLl open reading frame which viral capsid protein and viral particles protein in condylomas and corresponding antibodies in human sera. J Virol 1987;61:2684-90. are produced requires a considerable time 9 Matlashewski G, Banks L, Wu-Liao J, Spence P, Pim D, period-of the order of months.2' Anecdotal Crawford L. The expression of human papillomavirus type 18 E6 protein in bacteria and the production of anti- evidence in human disease suggests that a E6 antibodies. J Gen Virol 1986;67:1909-16. similar time period is involved in the develop- 10 Banks L, Spence P, Androphy E, et al. Identification of human papillomavirus type 18 E6 polypeptide in cells ment of genital condylomata."2 In many derived from human cervical carcinomas. J Gen Virol patients biopsy specimens may be taken at a 1987;68:1351-9. 11 Doorbar J, Gallimore PH. Identification of proteins encoded stage in the natural history of the disease when by the LI and L2 open reading frames of human papillo- viral particles are not produced, either because mavirus la. J Virol 1987;61:2793-9. 12 Doorbar J, Campbell D, Grand RJA, Gallimore PH. Iden- the lesion is at an early stage or because a tification of the human papillomavirus la E4 gene quiescent or persistent phase is another feature product. EMBO J 1984;5:355-62. 13 Browne HM, Churcher MJ, Stanley MA, Smith GL, of the disease process. Whatever the explana- Minson AC. Analysis of the LI gene product of human tion it is evident that L1 capsid protein is papillomavirus type 16 by expression in a vaccinia virus recombinant. J Gen Virol 1988;69:1263-73. synthesised in a few cells containing high 14 Koher C, Howe SC, Milstein C. Fusion between immuno- numbers ofviral DNA copies in the superficial, globulin-secreting and non-secreting myeloma cell lines. Eur J Immunol 1976;6:292-5. terminally differentiated surface layers of the 15 McLean C, Buckmaster A, Hancock D, Buchan A, epithelium. Fuller A, Minson A. Monoclonal antibodies to three non- glycosylated antigens ofherpes simplex virus type 2. J Gen Detection of viral DNA only informs about Virol 1982;63:297-305. presence of the genome and gives little infor- 16 Gowans EJ, Burrell CJ, Jilbert AR, Marmion BP. Cytoplas- mic (but not nuclear) hepatitis B virus (HBV) core antigen mation on virus replication and pathogenesis. reflects HBV DNA synthesis at the level of the infected This is particularly important in that papillo- hepatocyte. Intervirology 1985;24:220-5. 17 Oltersdorf T, Seedorf K, Rowekamp W, Gissmann L. maviruses are strictly epitheliotropic, and the Identification ofhuman papillomavirus type 16 E7 protein available information indicates that permissive by monoclonal antibodies. J Gen Virol 1987;68:2933-8. 18 Patel D, Shepherd PS, Naylor JA, McCance DJ. Reactivities viral replication only occurs in keratinocytes of polyclonal and monoclonal antibodies raised to the and specifically in the keratinocyte undergoing major capsid protein of human papillomavirus type 16. J Gen Virol 1989;70:69-77. the normal process of differentiation in the 19 Roseto A, Pothier P, Guillemin M-C, et al. Monoclonal epithelium. The intra-epithelial neoplastic antibodies to the major capsid protein of HPV1. J Gen Virol 1984;65:1319-24. lesions of the cervix are characterised by a 20 Jenson AB, Rosenthal C, Olson C, Pass F, Lancaster WD, progressive loss and distortion of the normal Shah K. Immunological relatedness of papilloma viruses from different species. JNCI 1980;64:495-500. differentiation process, and these features are 21 Olson C. Animal papillomaviruses: historical perspectives. included in the diagnostic criteria of these In: Salzman NP, Howley PM, eds. The Papovaviridae Vol. 2. The Papillomaviruses. New York: Plenum Press, lesions.2" Changes in the expression of viral 1987:39-66. proteins may be a reflection of neoplastic 22 Oriel DJ. The natural history of genital warts. British _ Journal of Venereal Diseases 1971;47:1-13. phenotype and may give useful information on 23 Buckley CH, Butler EB, Fox H. Cervical intraepithelial progression and prognosis, and the antibody neoplasia. J Clin Pathol 1982;35:1-13.