J Clin Pathol 1991;44:33-38 33 Interphase cytogenetics using and

digoxigenin labelled probes: III Increased J Clin Pathol: first published as 10.1136/jcp.44.1.33 on 1 January 1991. Downloaded from sensitivity and flexibility for detecting HPV in cervical biopsy specimens and cell lines

C S Herrington, A K Graham, J O'D McGee

Abstract background to the same extent as biotin. A monoclonal to en- Digoxigenin labelled probes, however, are less abled sandwich techniques to be used for sensitive than biotinylated probes using single the detection of hybridised digoxigenin step detection methods.6 The sensitivity of labelled probes in cultured cells and detection procedures can be enhanced using paraffin wax sections. This system has antibody sandwich techniques using a linker greater flexibility than alkaline phos- antibody to bridge between antibody directed phatase conjugated polyclonal anti- against reporter molecule and an enzyme con- digoxigenin antibody and permits the jugate.9 A refinement of this concept is to use use of alternative detector enzymes, such biotinylated , which can be used to as horseradish peroxidase and fluores- bridge between the first antibody, and an cence labels. The APAAP detection avidin enzyme conjugate.'0 This gives system that does not require the use of increased sensitivity due to the high affinity of biotin can also be used in situations avidin for biotin (dissociation constant = I0-` where endogenous biotin is a problem. M) and improved specificity when the first The low level of background staining antibody is of monoclonal type. combined with precise substrate deposi- Digoxigenin, the aglycone derivative of the tion of the amplified peroxidase system cardiac digoxin, has high affinity for gives higher sensitivity and resolution. antibodies to digoxin."'3 This suggested to us This permits localisation of closely that the antibodies which have been adjacent chromosomal loci in interphase developed for the monitoring and treatment of nuclei. The most sensitive peroxidase clinical digoxin toxicity would increase the http://jcp.bmj.com/ based digoxigenin detection system sensitivity of detection of digoxigenin labelled visualises two and a half to 12 copies of probes in NISH. human papillomavirus (HPV) per The aims of this investigation were: (1) to nucleus. This system is also suitable for develop systems of high sensitivity and the analysis of low copy number HPV resolution for digoxigenin labelled probes infection of cervical tissues. which are applicable to interphase cyto- genetics-for the latter, CaSki and HeLa cells on September 30, 2021 by guest. Protected copyright. were used as a model system; and (2) to adapt Interphase cytogenetics encompasses, among detection systems for digoxigenin labelled other things, visual discrimination of probes so that they are more sensitive and individual chromosomes or genes in inter- flexible for detection of genes in archival phase nuclei.'-3 This has previously been biopsy specimens. performed by fluorescence45 and more recently by enzyme based chromogenic procedures using biotin and digoxigenin Methods labelled probes.' During the course of the All chemicals were obtained from Sigma (UK) latter work, it became apparent that the detec- or BDH (UK) unless otherwise stated. tion systems for biotin or digoxigenin labelled The HPV6 probe was the amp2 fragment probes using alkaline phosphatase based cloned into pBR322,'4 the HPV16 probe the detection systems did not give sufficient whole viral genome in pAT153,'5 and HPV18 resolution to discriminate between closely 16 University of Oxford the whole viral genome in pBR322. Biotin and Nuffield Department adjacent chromosomal loci. digoxigenin were incorporated into plasmid of Pathology and Digoxigenin labelled probes have been used DNAs by nick translation, and the degree of Bacteriology, John both as alternatives and in addition to biotin Radcliffe Hospital, labelling and size of fragments were as Headington, Oxford labelled probes in non-isotopic in situ previously described.6 17 OX3 9DU hybridisation (NISH).67 The main advantage Paraffin wax sections were prepared and C S Herrington of digoxigenin over biotin labelled probes for nucleic acids unmasked as previously A K Graham NISH is that digoxigenin (a plant alkaloid J O'D McGee described.6 analogue), unlike biotin, is not present con- Correspondence to: HeLa 229 cells (ATCC, USA) were grown to Professor McGee stitutively in mammalian cells. Hence, in confluence in 80 cm2 flasks in RPMI 1640 Accepted for publication immunohistochemical detection systems, medium (Gibco, UK) supplemented with L- 16 August 1990 digoxigenin does not generate endogenous glutamine (1 x; Gibco, UK); penicillin/ 34 Herrington, Graham, McGee

streptomycin (1 x; Gibco, UK); and 20% fetal diluted 1 in 600 in TBT, washed in TBS, and calf serum (Gibco, UK). Detachment was the signal was developed using NBT/BCIP achieved using trypsin/EDTA solution (see below). (Gibco, UK) and the cells were pelleted at (b) Three step procedure Sections were J Clin Pathol: first published as 10.1136/jcp.44.1.33 on 1 January 1991. Downloaded from 600 x g. After two washes in PBS the cells incubated in monoclonal antidigoxin (Sigma, were fixed in 10% formalin, embedded in agar, UK) diluted 1 in 10 000 in TBT. Subsequent and processed to paraffin wax in the usual way. detection was as for the three step detection of Sections of 4 ,gm were cut and unmasked using biotinylated probes, with second incubation in proteinase K at a concentration of 1 mg/ml in biotinylated rabbit anti-mouse (F(ab')2 frag- PBS for 15 minutes at 37°C. ment) and the third incubation in avidin- CaSki cells were grown onto slides, fixed in peroxidase (dig-3-P/C) or avidin-alkaline paraformaldehyde, and nucleic acids unmasked phosphatase (dig-3-A/N). as previously described.6 (c) APAAP system Slides were incubated in monoclonal a4tidigoxin as in (a), followed by HYBRIDISATION rabbit anti-mouse immunoglobulin (Dako, Aliquots of hybridisation mix (8 u1) containing UK), diluted 1 in 50 in TBT, then APAAP 2 ng/pl ofbiotin or digoxigenin labelled HPV6, complex (Dako, UK) diluted 1 in 50 in TBT. HPV 16, or HPV18 were added to each well on Signal was developed using NBT/BCIP. multispot slides which were covered with a (d) Fluorescence detection Slides were treated 14 mm glass coverslip (Chance, UK) and the as for the three step detection method in (b) slides placed in a moist Terasaki plate. except that streptavidin- iso- Hybridisation mix consisted of 50% forma- thiocyanate (Dako, UK) was used in the final mide, 5% dextran sulphate, 2 x SSC, and 0 05 step. Optimal results were obtained with a mol/l TRIS-HCI, pH 7 3; 1 x SSC = 0 15 dilution of .1 in 20 of the fluorescent conjugate mol/l sodium chloride, 0-015 mol/l sodium in TBT containing 5% non-fat milk. citrate. Sections and probes were simul- taneously denatured at 95°C for 15 minutes on a SUBSTRATE PREPARATION solid stainless steel plate in a hot air oven and Nitrobluoe tetrazolium/5-bromo-4-chloro- then hybridised at 420C for 2 hours. 3-indolyl phosphate (NBT/BCIP) and amino- ethylcarbazole (AEC) were made as previously DETECTION OF HYBRIDISATION SIGNAL described.618 Diaminobenzidine (DAB) was All antibody/avidin incubations were carried diluted to 0-5 mg/ml in distilled water and out for 30 minutes at 22°C unless otherwise stored in 5 ml aliquots at -20°C. Hydrogen stated. Slides were washed in three changes of4 peroxide was added to DAB (final concentra- x SSC at 22°C (five minutes each change), tion 03%) immediately before use. After probe soaked in blocking agent TBT (0.05 mol/l detection slides were air dried at 42°C and mounted in TRIS-HCI, 0-10 mol/l sodium chloride (pH glycerol jelly (for enzyme/ http://jcp.bmj.com/ 7 2) containing 3% (w/v) bovine serum chromogen reactions). For fluorescence albumin and 0-05% Triton 100 (v/v)) at 22°C preparations, PBS/glycerol (1:9 v/v) contain- for 10 minutes. Subsequent visualisation of ing 2-3% diazobicyclo-octane (DABCO) signal was determined by the reporter (Sigma, UK) was used. molecule. Frequency distributions were compared using the two tailed Mann-Whitney U test with

Biotin labelled probes a correction for tied values. on September 30, 2021 by guest. Protected copyright. (a) Single step procedure Sections were incubated in avidin-alkaline phosphatase or avidin-peroxidase (Dako, UK) diluted 1 in 100 Results in TBT. Unbound conjugate was removed by The results of single step detection methods for washing for five minutes twice in 50 mM biotin and digoxigenin probes have previously TRIS-HCI, 10 mM NaCl, pH 7 2 (TBS), and been reported.67 These were less sensitive than the sections incubated in the appropriate the three step procedures described below. substrate. (b) Three step procedure Sections were COMPARATIVE SENSITIVITY AND RESOLUTION OF incubated first in monoclonal mouse anti- DIGOXIGENIN AND BIOTIN IN INTERPHASE CELLS biotin (Dako, UK) diluted 1 in 50 in TBT, The three step peroxidase AEC detection washed in TBS (twice at 10 minutes each), and method for digoxigenin labelled probes (dig-3- then in biotinylated rabbit anti-mouse (F(ab')2 P/C) produced very low background staining fragment, Dako, UK). After washing in TBS, and enhanced resolution in CaSki and HeLa the third incubation was in either avidin alka- cells (figs 1 and 2). Discrete HPV signals were line phosphatase (bio-3-A/N), or streptavidin considerably more easily resolved when com- peroxidase (Dako, UK) (bio-3-P/C). The pared with the three step alkaline phosphatase former was diluted 1 in 50 and the latter 1 in NBT/BCIP detection system for digoxigenin 100 in TBT containing 5% non-fat milk. After (dig-3-A/N) (fig 1); the latter is the counterpart washing in TBS, sections were incubated in the ofthe most sensitive system we have developed appropriate substrate (see below). hitherto." An estimate of the absolute sensitivity of the Digoxigenin labelledprobes dig-3-P/C method (fig 3) can be made by (a) Single step procedure Sections were analysis of the frequency distribution of dots incubated in alkaline phosphatase conjugated per interphase nucleus in CaSki cells. The antidigoxigenin (Boehringer, Germany) median number of dots per nucleus is nine. Up Digoxigenin probe detection 35 J Clin Pathol: first published as 10.1136/jcp.44.1.33 on 1 January 1991. Downloaded from

J *; * 4 .m

A* #I *

.0

'a 1

Si^

.* 4 *^e7 ''it

A B Figure 1 CaSki cells hybridised with digoxigenin labelled HPV16 probe and detected using (A) three step alkaline phosphatase NBT/BCIP (dig-3-A/N) and (B) three step peroxidase AEC (dig-3-P/C). The nuclei are lightly counterstained with haematoxylin in (B). Note the clear resolution of closely adjacent dots within individual nuclei in (B), but the high level of background with staining of nucleoli (small arrows) and coalescence of dots (large arrow) in (A). to four discrete dots were also detected in the Digoxigenin labelled probes nuclei of HeLa cells (which contain 10-50 The amplified procedure for the detection of copies of HPV1820), which had been fixed in. digoxigenin using alkaline phosphatase with

formalin and embedded in paraffin wax (fig 2). NBT/BCIP as substrate (dig-3-A/N) gave a http://jcp.bmj.com/ signal in a greater number of cells compared COMPARATIVE SENSITIVITY OF DIGOXIGENIN AND with the comparable three step alkaline phos- BIOTIN LABELLED PROBES ON BIOPSY SPECIMENS phatase system for biotin detection (bio-3-A/ The relative sensitivities of digoxigenin and N) (fig 4). The background produced by dig-3- biotinylated probes were assessed on routinely A/N is significantly less than that produced by processed, paraffin wax embedded archival bio-3-A/N and the signal intensity is enhanced

cervical biopsy specimens. These showed the to a greater degree using digoxigenin compared on September 30, 2021 by guest. Protected copyright. criteria of wart virus infection,21 with or with- with biotin (fig 4). Clear signal within almost all out CIN1-3. parabasal cells is present with digoxigenin labelled probes (fig 4A). The signal to noise ratio of digoxigenin labelled probes is superior to that ofbiotinylated probes-compare figs 4A and 4B. Using the three step peroxidase AEC detec- tion system for the detection of digoxigenin labelled HPV16 (dig-3-P/C) in lesions defined histopathologically as CIN3, signal can be seen in all layers of the epithelium, including the basal layer (fig 5). Notably, the morphology of this signal is different from that obtained by NISH analysis of wart virus infection with mild atypia (fig 4). The combination of the substrate AEC with haematoxylin counter- staining aids interpretation. The APAAP procedure produced a clear signal with little background (data not shown), but it was less sensitive than the procedure used in fig 4 as judged by both the number of positive cells and the fact that parabasal cells Figure 2 HeLa cellsfixed informalin and embedded in paraffin wax were hybridised were not labelled. with digoxigenin labelled HPV18 and detected using dig-3-P/C. Four discrete signals of (large arrow) can be seen within a single nucleus. Note that signals are present within Fluorescence detection digoxigenin other nuclei but that some are out of the plane offocus. labelled probes also produced satisfactory 36 Herrington, Graham, McGee

Figure 3 Frequency random priming,23 or nick translation.6 These distributions of dot number probes have been used on nitrocellulose per nucleusfor (A) dig-3- in P/C and (B) bio-3-P/C. filters2223 and for situ hybridisation.67 The Dig-3-P/C is significantly simplest way to detect digoxigenin labelled J Clin Pathol: first published as 10.1136/jcp.44.1.33 on 1 January 1991. Downloaded from more sensitive than bio-3- probes is by using an alkaline phosphatase P/C (p < 0 001), with conjugated polyclonal antidigoxigenin median dot numbers of nine and seven, antibody. This system is less sensitive than the respectively. 0~ corresponding avidin based single step detec- tion system for biotin,6 a difference likely to be due to the higher affinity of avidin for biotin (KD 10-1') compared with the affinity of antibody for antigen (KD 10-12 at best). To improve the sensitivity of digoxigenin Number of dots labelled probes we have used a monoclonal antibody to digoxin.- As shown here, this produces considerably greater sensitivity. The signal also develops within minutes with NBT/ BCIP and AEC compared with an hour or more for single step digoxigenin detection and up to 20 16 hour incubations in NBT/BCIP in some 0 a) detection systems for biotin.24 These methods 0) increase the sensitivity of these probes as Cc a)10410 determined by the number of HPV signals per cell in cultured cells and by the number ofHPV positive cells in clinical samples. The detection systems for biotin produce more background staining than the corresponding systems for digoxigenin. This is likely to be due to Number of dots amplification of signal from endogenous biotin. By increasing the number of rounds of amplification of digoxigenin, the signal can be results, with low background staining of enhanced still further but, as this requires the surrounding connective tissue (data not use of an anti-biotin antibody and an avidin shown). This procedure, however, was less conjugate, the background staining is corres- sensitive as signal was not detected within pondingly increased. Nevertheless, amplified parabasal cells. detection systems for digoxigenin amplify endogenous biotin by one round fewer than the

corresponding system for biotin. http://jcp.bmj.com/ Discussion Advantages of the peroxidase-AEC system Digoxigenin has been used to label probes are both its high degree of resolution and the by terminal labelling of oligonucleotides,22 ease of interpretation of preparations counter-

o * 0 ap 0

. 4 on September 30, 2021 by guest. Protected copyright. '9#*.X Je, 4. * X* 0 *e

* i0 V:.:t 9.*

b .4 41ni~~~~~~~~~~~~~~~~~~_ ,,U i .,

.l0.'*. * *

B 9w Figure 4 Sections of a condyloma acuminatum were 4.. probed with (A) digoxigenin labelled and (B) biotin labelled HPV6, detected using the three step methods dig-3-A/N and bio-3-A/N, respectively. Parabasal cells are positive only with the digoxigenin labelled probe (arrows). Note the low level ofbackground staining in the subjacent stroma in (A). The biotinylated probe produces greater background staining in the region of the basement membrane rendering interpretation impossible. Digoxigenin probe detection 37

In routine cervical biopsy specimens, the three step digoxigenin system is clearly the most sensitive developed for clinical practice.

An HPV signal is evident in almost all J Clin Pathol: first published as 10.1136/jcp.44.1.33 on 1 January 1991. Downloaded from parabasal cells. It is of note that the mor- phology of the signal (single dots) obtained in the biopsy specimen shown in fig 5 resembles that seen in CaSki and HeLa cells, in which the viral genomes are known to be integrated.2526 It is tempting to speculate that this NISH signal morphology is indicative of viral integration and this is under investigation by routine molecular biological techniques in this laboratory. APAAP and fluorescence detection of digoxi- genin labelled probes can be performed by minor modification to the three step detection system. These may be useful in clinical laboratories for routine screening by NISH of HPV in cervical biopsy specimens but they are Figure 5 A biopsy specimen of CIN 3 probed with digoxigenin labelled HPV16 ani d less sensitive than the methods described in cktected using the peroxidase AEC system (dig-3-P/C). Note that signal is present i'n detail above. It should be noted, however, that basal cells (large arrow). the APAAP system has the advantage that it does not require biotin. This may be useful in tissues where endogenous biotin creates high stained with haematoxylin. Thus cle)sely background. adjacent dots can be readily distinguished in Overall, the preferred methods are therefore: CaSki and HeLa cell nuclei and in arc}hival (i) three step alkaline phosphatase NBT/BCIP biopsy specimens. Both these cell lines conitain detection of digoxigenin labelled probes where their respective HPV types integrated intc) the high sensitivity and low background but not host genome at multiple sites.25 26 i\EC high signal resolution are required; (ii) three produces better resolution and colour contrast step peroxidase AEC detection of digoxigenin than DAB, making it more suitable for siImul- where high sensitivity, high resolution, and low taneous double labelling of genes.7 Isigh background are required in addition to nuclear resolution is not possible using NBT/B(CIP, counterstaining. because diffusion of substrate leads to loEss of In conclusion, HPV probes labelled with definition and hence loss of spatial resoluition. digoxigenin can be detected using long estab-

Thus closely adjacent gene loci can be dis- lished monoclonal antibodies to digoxin. This http://jcp.bmj.com/ criminated with AEC: this is not possible with permits amplification ofthe signal produced by alkaline phosphatase based systems. these probes to a degree that detects low copy The three step peroxidase AEC detec-tion number viral infection in cell lines and routine system (dig-3-P/C) is more sensitive than1 the clinical samples. Flexibility is also enhanced in comparable biotin based system (bio-3-J P/C) that digoxigenin detection can be achieved with (fig 3). Analysis ofthe frequency distributican of horseradish peroxidase, fluorescence and

dot numbers per nucleus for these two metlhods APAAP systems. The latter has the advantage on September 30, 2021 by guest. Protected copyright. shows that many more cells containing rnore that it does not require biotin. The enhanced than 10 dots per nucleus are detectable uIsing sensitivity and resolution also make digoxi- digoxigenin. Given that these systems iwere genin an attractive label for interphase applied to similar cell populations, this dif- cytogenetics and for the detection of HPV in ference is likely to have been due to the grccater cervical disease. resolution of signal using digoxigenin as p,robe label (fig 1). Calculation of sensitivity fronra the CSH is a CRC Clinical Research Fellow and holds a Junior median values of these distributions gives an Research Fellowship at Green College, Oxford. This work was supported by grants to J O'D McG from the Cancer Research absolute sensitivity of 40 viral genomes foror Campaign(UK). biotin and 30 for digoxigenin detection u sing the amplified peroxidase methods, given that the total number of copies of HPV per CaSki cell nucleus is about 270.6 This measure is, 1 Burns J, Chan VT-W, Jonasson JA, Fleming KA, Taylor S, McGee JO'D. Sensitive system for visualising bio- however, an underestimate as it asse-sses tinylated DNA probes hybridised in situ: rapid sex relative sensitivity more accurately than determination in intact cells. J Clin Pathol 1985;38: ~ 1085-92. absolute sensitivity,6 and as can be seen frorm figm fig 2 Cremer T, Landegent J, Bruckner A, et al. Detection of 2, the presence of multiple signals in HeLa1 cell chromosome aberrations in the human interphase nucleus by visualisation of specific target DNAs with radioactive nuclei using this system suggests that the and non-radioactive in situ hybridisation techniques: sensitivity is greater. Indeed, it has 1been diagnosis of trisomy 18 with probe L1.84. Hum Genet 1986;74:346-52. estimated that HeLa cells contain 10-50 cc )pies 3 Bhatt B, McGee JO'D. Chromosomal assignment of genes. ofHPV18.20 As four NISH signals were visiual- In: Polak J, McGee JO'D, eds. In situ hybridisation: principles and practice. Oxford: OUP, 1990:149-64. ised in paraffin wax embedded HeLa celh;s an 4 Hopman AHN, Ramaekers FCS, Raap AK, et al. In situ estimate of the sensitivity of this detec:tion'.ion hybridisation as a tool to study numerical chromosome aberrations in solid bladder tumours. Histochemistry 1988; system is two and a half to 12 viral copic-s in 89:307-16. routinely processed material. 5 Pinkel D, Landegent J, Collins C, et al. Fluorescence in situ 38 Herrington, Graham, McGee

hybridisation with human chromosome specific libraries: papillomavirus DNA, its presence in genital cancer Detection oftrisomy 21 and translocations ofchromosome biopsies and in cell lines derived from cervical cancer. 4. Proc Natl Acad Sci USA 1988;85:9138-47. EMBO J 1984;3:1151-7. 6 Herrington CS, Burns J, Graham AK, Bhatt B, McGee 17 Chan VT-W, Fleming KA, McGee JO'D. Detection of JO'D. Interphase cytogenetics using biotin and digoxi- subpicogram quantities of specific DNA sequences on blot genin labelled probes I: relative sensitivity of both repor- hybridisation with biotinylated probes. Nucleic Acids Res J Clin Pathol: first published as 10.1136/jcp.44.1.33 on 1 January 1991. Downloaded from ter molecules for the detection of HPV16 in CaSki cells. J 1985;13:8083-91. Cell Pathol 1989;42:591-600. 18 Burns J, Graham AK, Frank C, Fleming KA, Evans MF, 7 Herrington CS, Burns J, Graham AK, Bhatt B, McGee McGee JO'D. Detection of low copy human papilloma JO'D. Interphase cytogenetics using biotin and digoxi- virus DNA and mRNA in routine paraffin sections of genin labelled probes II: simultaneous detection of viral cervix by non-isotopic in situ hybridisation. J Clin Pathol and mammalian nucleic acids in individual nuclei. J Clin 1987;40:858-64. Pathol 1989;42:601-6. 19 Hcrrington CS, Burns J, Graham AK, McGee JO'D. 8 Herrington CS, McGee JO'D. Interphase cytogenetics. Discrimination of closely homologous HPV types by Neurochem Res 1990;15:467-74. nonisotopic in situ hybridisation: definition and deriva- 9 Guesdon J-L, Ternynck T, Avrameas S. The use of avidin- tion of tissue lIms. Histochem J 1990;22:545-54. biotin interaction in immunoenzymatic techniques. 20 Schwarz E, Freese UK, Gissmann L, et al. Structure and J Histochem Cytochem 1979,27:1 131-9. transcription of human papillomavirus sequences in 10 Hsu S-M, Raine L, Fanger H. Use of avidin-biotin-per- cervical carcinoma cells. Nature 1985;314:111-14. oxidase complex (ABC) in immunoperoxidase techniques. 21 Coleman DV, Evans DMD. Biopsy pathology and cytology of J Histochem Cytochem 198 1;29:577-80. the cervix. London: Chapman and Hall, 1988. 11 Smith JW, Butler VP, Haber E. Characterisation of 22 Schafer R, Zischler H, Epplen JT. DNA fingerprinting antibodies of high affinity and specificity for the using non-radioactive oligonucleotide probes specific for glycoside digoxin. Biochemistry 1970;9:331-7. simple repeats. Nucleic Acids Res 1988;16:9344. 12 Monji N, Ali H, Castro A. Quantification of digoxin by 23 Dooley S, Radtke J, Blin N, Unteregger G. Rapid detection enzyme immunoassay: synthesis ofa maleimide derivative of DNA-binding factors using protein-blotting and of digoxigenin succinate for enzyme coupling. Experientia digoxygenine-dUTP marked probes. Nucleic Acids Res 1980;36:1141-3. 1988;16: 11839. 13 Valdes R, Brown B, Graves SW. Variable cross-reactivity of 24 Mougin C, Guitteny AF, Fouque B, Viennet G, Teoule R, digoxin metabolites in digoxin immunoassay. Am J Clin Bloch B. Histochemical detection of the messenger RNAs Pathol 1984;82:210-13. coding for calcitonin and calcitonin gene-related peptide 14 De Villiers E-M, Gissmann L, Zur Hausen H. Molecular in medullary carcinomas of the thyroid with radioactive cloning of viral DNA from human genital warts. J Virol and biotinylated oligonucleotide probes. J Pathol 1990; 1981;40:932-5. 160:187-94. 15 Durst M, Gissmann L, Ikenberg H, Zur Hausen H. A 25 Popcscu NC, DiPaolo JA, Amsbaugh SC. Integration sites papillomavirus DNA from a cervical carcinoma and its ofhuman papillomavirus 18 DNA sequences on HeLa cell prevalence in cancer biopsy samples from different chromosomes. Cytogenet Cell Genet 1987;44:58-62. geographical regions. Proc Natl Acad Sci USA 1983; 26 Mincheva A, Gissmann L, Zur Hausen H. Chromosomal 80:3812-15. integration sites of human papillomavirus DNA in three 16 Boshart ML, Gissmann L, Ikenberg H, Kleinheinz A, cervical cancer cell lines mapped by in situ hybridisation. Scheurlen W, Zur Hausen H. A new type of Med Microbiol Immunol 1987;1 76:245-56. http://jcp.bmj.com/ on September 30, 2021 by guest. Protected copyright.