J Clin Pathol 1990;43:806-809

In situ hybridisation with digoxigenin-labelledDNA probes for detection of viral genomes

Y Furuta, T Shinohara, K Sano, M Meguro, K Nagashima

AbstractThe applicability of a recently developednon-radioactive DNA labelling anddetection method, which uses the digox-igenin (DIG) enzyme linked immuno-sorbent assay (ELISA) system, for thedetection of viral infections in pathologyspecimens by in situ hybridisation, wasexamined. Its efficacy was comparedwith that of biotin and radioisotopelabelling methods. Three cases ofprogressive multifocal leucoence-phalopathy, two of verruca vulgaris, andseven cases of laryngeal papilloma werestudied. The sensitivity of the DIGlabelled probe was almost the same asthat of a "S-labelled probe in the dot-blot hybridisation test. Using in situhybridisation with "S-labelled and DIGlabelled probes, the levels of thehybridised signals detected were similar.The biotin labelled probe was less sen-sitive, particularly in the cases of laryn-geal papilloma. The DIG labelling anddetection method was highly sensitiveand applicable to the detection of viralinfection by ISH, and is preferable to aradiolabelled probe, especially when insitu hybridisation is done in the patho-logy laboratory.

Department ofPathology II,Hokkaido UniversitySchool ofMedicine,Sapporo, JapanY FurutaT ShinoharaK SanoM MeguroK NagashimaCorrespondence to: DrYasushi Furuta, Departmentof Pathology II, HokkaidoUniversity School ofMedicine, Kita 15 Nishi 7,Kits-ku, Sapporo 060, JapanAccepted for publication15 March 1990

Electron microscopical examination andimmunohistochemistry have been used toshow the presence of viral particles and viralproteins, respectively, in pathologyspecimens. In situ hybridisation (ISH) meth-ods for detecting viral genomes have recentlybeen introduced, becoming widespread. DNAor RNA probes specific for viral genomes are

used, and these probes are labelled either withradioisotopes or with non-radioactiveproducts. For routine in situ hybridisation inthe pathology laboratory, non radioactiveprobes are better than radioactive ones,because radiolabelled probes require cumber-some procedures for handling theradiosotopes.We examined the applicability of a recently

developed non-radioactive DNA labelling anddetection method, in which the digoxigenin(DIG) enzyme linked immunosorbent assay(ELISA) system is used, for the demonstra-tion of viral genomes by in situ hybridisation,and we compared its efficacy with that of thebiotin and radioisotope methods.

MethodsThree cases of progressive multifocal leucoen-cephalopathy, which is induced by humanpolyoma virus (JCV); two cases of verrucavulgaris, caused by human papillomavirus(HPV); and seven cases of laryngealpapilloma, also predominantly caused byHPV infection were studied. Tissue sectionswere fixed with 10% formalin and embeddedin paraffin wax.Human polyomavirus DNA' and whole

genomes of HPV DNA types, 22 6,3 and 1 14were used. The DNA probes were labelledwith digoxigenin-dUTP (Boehringer, Mann-heim, West Germany), 3S-dCTP (4&8 TBq/mmol; NEN Research Products), and biotin-dUTP (Boehringer, Mannheim) by the Ran-dom Primed method.5 The radiolabelledprobes were purified by a G-50 Sephadexcolumn to specific activities of 2 to 4 x 108cpm/ig. Digoxigenin- and biotin labelledprobes were purified by ethanol precipitation.To test for the sensitivity of each probe, 0 1

pg-lng of HPV 11 DNA was spotted on to anylon filter (Biodyne; Pall), and dot-blothybridisation was carried out with DIGlabelled, biotin labelled, or 35S-labelled HPV11 DNA probes.

IN SITU HYBRIDISATIONFive micrometre thick sections were cut fromparaffin wax embedded tissues and mountedon slides coated with 3-aminopropyltri-etoxysilan.6 The tissues were waxed andrehydrated by sequential immersion in xyleneand graded ethanols. When biotin labelledprobes were used, the slides were immersed in0 3% hydrogen peroxide diluted with meth-anol for 15 minutes, followed by rinsing indeionised water so that the endogenous perox-idase reaction was blocked. The sections wereimmersed in 0-2 N HC1 for 20 minutes andwashed in 2 x SSC (1 x SSC: 015 Msodium chloride, 15 mM sodium citrate) forsix minutes at room temperature. They werethen treated with 100 1ig/ml proteinase K(Boehringer, Mannheim) for 15 minutes at37°C. For detection of viral DNA or mRNA,the sections were also digested with 100 gg/mlribonuclease A (RNAse A, Boehringer, Man-nheim) or with 100 Mg/ml deoxyribonuclease I(DNAse I, Boehringer, Mannheim), respec-tively, for one hour at room temperature. Thesections were then refixed with 4% parafor-maldehyde in 01 M phosphate buffer for 10minutes at 4°C and immersed in 0-2% glycine

806

on March 25, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.43.10.806 on 1 October 1990. D

ownloaded from

In situ hybridisation with digoxigenin-labelled DNA probes

Figure 1 A case ofprogressive multifocalleucoencephalopathy. Theprotein antigen and viralgenomes were detected byboth immunohistochemistry(A); and in situ hybridisation(B); "S-labelled probe (C);digoxigenin-labelled probe(D); biotin-labelled probein infected glial cells.(Counterstaining withhaematoxylin).

1'^ i

ti.S1.

0a

* 0.,~4 Z-..f.I

.i.0

V.o

.

-0

,W.. 0it*

*-

0....

'1$:

4,7

A.W.v

.- Wi .;. I% ..

'r 4.0

j * 'St, _~S0~-

0 ;f4't'3.. w

0 0 171.~~~~.

VPq 0.4do, %e

ei

~Wf,

.X

wtk

¶4

4,

9

-a.6

A

a..*

4

in 0 1 M phosphate buffered saline (PBS, pH7-4) for 30 minutes at room temperature,followed by washing in 01 M PBS for 10minutes. The slides were dehydrated ingraded ethanol and dried in air.The hybridisation mixtures consisted of

500/0 deionised formamide, 5 x Denhardt'ssolution, 5 x SSPE (1 x SSPE: 0-18 MNaCl, 10 mM NaH2PO4, 1 mM EDTA, pH7-4), 0100 sodium dodecyl sulphate, 100 gg/ml of denatured salmon sperm DNA, 10%dextran sulphate, and 2 ng/gl of DIG- orbiotin labelled DNA probe, or 0 1-0-2 ng/plof 35S-labelled probe. Each 10 gl of thehybridisation mixture was pipetted on to tis-sue sections and covered with siliconised glasscoverslips. The slides were then incubated on

a hotplate at 90°C for 10 minutes andimmediately cooled on ice. For the sections tobe used for detection of viral mRNA the

denaturation step was omitted. Hybridisationwas carried out at 42°C for 16 hours. Thecoverslips were then removed and the sectionswere washed in three changes of 0-5 x SSC, 1mM EDTA at 55°C for five minutes eachonce in 5000 formamide, 2 x SSC for 10minutes at 55°C, and in two changes of 0-5 x

SSC for five minutes each at 55°C, followedby a single washing in 0-5 x SSC for fiveminutes at room temperature.

DETECTION OF HYBRIDISATION SIGNALWhen DIG labelled probes were used, theslides were blocked for 30 minutes in 0-500blocking reagent solution (Boehringer, Mann-heim) and washed in DIG buffer 1 (0- 1 MTRIS-HCl, pH 7-5, 0-15 M NaCl) for one

minute. The slides were then incubated withanti-DIG antibody (diluted 5,000 times withDIG buffer 1) conjugated to alkaline phos-

a4 *,.b

*.4,

.s ,

.,

I'47

.S* s

4

a-.v

A4: 4,I-

t:

S

1.4

.4X.8

S

0 .'S

U

40

O

4

a

I

(S.

o. F

0

C.

ak

807

2

40 C'

.tox

.0

d

on March 25, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.43.10.806 on 1 October 1990. D

ownloaded from

Furuta, Shinohara, Sano, Meguro, Nagashima

Figure 2 A case ofverruca vulgaris. (A)ISH with digoxigenin-labelled HPV 2 DNAprobe. Strong positivehybridised signals wereseen from the superficial tothe intermediate layer.Background staining,however, was present. (B)In situ hybridisation afterpretreatment with RNAse.Background staining wasdiminished. (C) In situhybridisation afterpretreatment with DNAse.Background staining wasdiminished and viralmRNA was detected in thesuperficial layer.(Counterstaining withhaematoxylin).

/ 4

r,I, -

ii I,> e1

v .

i

ItI

I

..Er~I

phatase for 30 minutes, followed by washingtwice with DIG buffer 1 for 15 minutes eachand rinsing in DIG buffer 3 (0 1 M TRIS-HCl, pH 9 5, 0 1 M NaCl, 0 05 M MgCl2) fortwo minutes. The colour reaction was perfor-med with nitroblue tetrazolium salt and 5-bromo-4-chloro-3-indolyl phosphate solutionin the dark for periods from 20 minutes toeight hours, depending on the appearance ofbackground staining. When we used biotin-labelled probes, the slides were incubatedwith the avidin-biotin-peroxidase complex(ABC Elite, Vectastain) for 30 minutes, foll-owed by washing twice in PBS for 15 minuteseach. The slides were incubated with 3,3'-diaminobenzidine for the colour reaction.When 35S-labelled probes were used, theslides were dehydrated in graded ethanol con-taining 0 3 M ammonium acetate, dried in air,and dipped in Konica NR-M2 emulsion forautoradiography. After 2 weeks' exposure theslides were developed and fixed. All slideswere counterstained with haematoxylin.

IMMUNOHISTOCHEMISTRYTo detect viral proteins we used an antibody

for JCV (rabbit, polyclonal) and an antibodyfor the genus specific common structuralantigen (pgs-antigen) of the papillomavirus(rabbit polyclonal, Dako). Immunohisto-chemical staining was done with the avidin-biotin-peroxidase complex method.

ResultsWith dot-blot hybridisation, we were able todetect 1 pg of HPV 11 DNA with either theDIG labelled or the "S-labelled probe. Withthe biotin labelled probe, however, we detec-ted only 100 pg of HPV 11 DNA (data notpresented). Thus the sensitivity of the DIGlabelled probe was almost the same as that ofthe "S-labelled probe.

In the cases of progressive multifocalleucoencephalopathy the protein antigen ofthe JCV was detected in scattered foci ofdemyelination in the nuclei of infected glialcells, and we could detect viral nucleic acidswith equal sensitivity by in situ hybridisationwith DIG labelled, biotin labelled, or "5S-labelled probes (fig 1).

In the cases of verruca vulgaris the pgs

..IFigure 3 A case of adultmultiple laryngealpapilloma. (A) The pgsantigen was detected in afew koilocytic cells of thesuperfical layer (arrow).(B) In situ hybridisationwith "S-labelled HPV 6DNA probe. Positivehybridisation signals wereclearly shown in thesuperficial layer. (C) Insitu hybridisation withDIG labelled probe. Thesensitivity of hybridisationwas almost equal to thatwith the "S-labelledprobe. (D) In situhybridisation with biotinlabelled probe. Positivehybridisation signals wereweak in the superficiallayer. (Counterstainingwith haematoxylin.)

C, ... .4

i..

-.. 4.

0'

808

.f

on March 25, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.43.10.806 on 1 October 1990. D

ownloaded from

In situ hybridisation with digoxigenin-labelled DNA probes

antigen was detected in the superficial layer ofthe epithelium, which showed koilocyticchanges. In situ hybridisation with the HPV 2DNA probe showed strong positive signalsfrom the superficial to the intermediate layer,but a non-specific reaction was noted in thesections with the DIG labelled probe. Wecould decrease the intensity of this non-specific reaction, however, by prior treatmentwith DNAse or RNAse. Moreover, we coulddetect viral mRNA by carrying out in situhybridisation after DNAse digestion (fig 2).We detected pgs antigen in only two of the

seven (29%) laryngeal papillomas. In bothpositive cases the reaction was confined to afew cells which showed koilocytic changes.Using in situ hybridisation with DIG- and35S-labelled probes, we detected HPV 6 orHPV 11 in five (71%) cases, at almost iden-tical sensitivity. With the biotin labelledprobe, however, we detected HPV genomes inonly three (43%) cases. In two of the sevencases we were unable to detect HPV genomesor viral protein, and in three cases infectionwith HPV was detected only by in situhybridisation and not by immunostaining (fig3).

DiscussionIt is important to detect viral infection inpathology specimens. In situ hybridisationmethods have been increasingly used for detec-ting viral genomes in tissue sections. Because insitu hybridisation is highly sensitive, we candetermine the precise localisation of infectedcells. Furthermore, in situ hybridisation can beapplied to paraffin wax embedded tissue. In ourstudy JCV infection was shown with bothimmunostaining and in situ hybridisation, butin three cases of laryngeal papilloma HPVinfection was detected only by in situ hybridis-ation, confirming that it is a more sensitivemethod for detecting infection with HPV.DNA or RNA probes specific for the viral

genome are used for in situ hybridisation.DNA probes are handled more easily in thepathology laboratory than are RNA probesbecause the latter require stringent precautionsfor avoiding contamination with RNAse.These probes were labelled with radioisotopeor non-radioactive products. To use radioac-tive probes, in situ hybridisation must beperformed in a laboratory specifically equippedto use radioactive products, and care must betaken to avoid exposure to radioactive substan-ces. Use of radioactive labelling is also moreexpensive, the product is unstable because ofits short half-life, and it is time consuming.Thus non-radioactive probes are preferable forroutine use in the pathology laboratory. Severalnon-radioactive labelling and detection meth-ods, such as the biotin7 or sulphon8 labellingmethods and the T-T dimer method,9 havebeen introduced. It has been pointed out,however, that these methods fall short of theradioactive method in terms of sensitivity,specificity, or convenience of labelling anddetection. The DIG-ELISA system, a newDNA labelling and detection methoddeveloped by Boehringer Mannheim, has been

widely used for various areas of research'>3because ofits h-igh sensitivity and convenienceof labelling. Heiles et al first applied a DIGlabelled probe in an effort to detect HPV incultured cells by in situ hybridisation, and theyfound that the DIG labelled probe was moresensitive than a biotin labelled probe.'0 Theydid not make comparisons with radioautogra-phic methods, however, and they did not applyDIG-ELISA to pathology specimens. Her-rington et al reported that the sensitivity of aDIG labelled probe was equivalent to that ofbiotin by in situ hybridisation using culturedcells, and applied the simultaneous differentialdetection of two nuclei acid sequences in thesame nucleus.'2 13

In this study the sensitivity of the DIGlabelled probe was almost the same as that ofthe 35S-labelled probe in dot-blot hybridisa-tion. Similar levels of positive signals weredetected by in situ hybridisation with the 35S-and the DIG labelled probe; the biotin labelledprobe was less sensitive, particularly in thecases of laryngeal papilloma. A non-specificreaction, however was occasionally noted withthe DIG labelled probe. We decreased theintensity of the non-specific reaction bypretreatment with DNAse or with RNAse,because the non-specific binding of the DIGlabelled probe to nucleic acids in tissue wasdecreased.

The plasmids of human papillomavirus types 2, 6, and 11 werekindly provided by Drs PeterM Howley and Harald zur Hausenthrough the auspices ofthe Japanese Cancer Research ResourcesBank (JCRB)-Gene.

I Frisque RJ, Bream GL, Cannella MT. Humanpolyomavirus JC virus genome. J Virol 1984;51:458-69.

2 Heilman CA, Law MF, Israel MA, Howley PM. Cloning ofhuman papilloma virus genomic DNAs and analysis ofhomologous polynucleotide sequences. J Virol 1980;36:395-407.

3 Villiers EM, Gissmann L, zur Hausen H. Molecular cloningof viral DNA from human genital warts. i Virol1981;40:932-5.

4 Gissmann L, Diehl V, Schultz-Coulon HJ, zur Hausen H.Molecular cloning and characterization of humanpapilloma virusDNA derived from a laryngeal papilloma.J Virol 1982;44:393-400.

5 Feinberg AP, Vogelstein B. A technique for radiolabelingDNA restriction endonuclease fragments to high specificactivity. Analyt Biochem 1983;132:6-13.

6 Rentrop M, Knapp B, Winter H, Schweizer J. Aminoalkyl-silane-treated glass slides as support for in situ hybridiza-tion of keratin cDNAs to frozen tissue sections undervarying fixation and pretreatment conditions. Histochem J1986;18:271-6.

7 Brigati DJ, Myerson D, Leary JJ, et al. Detection of viralgenomes in cultured cells and paraffin-embedded tissuesections using biotin-labeled hybridization probes.Virology 1983;126:32-50.

8 Poverenny AM, Podgorodnichenko VK, Bryksina LE,Monastyrskaya GS, Sverdlov ED. Immunochemicalapproaches to DNA structure investigation-I. Immuno-chemical identification of the product of cytosinemodification with bisulphite and o-methylhydroxylaminemixture. Molec Immunol 1979;16:313-6.

9 Nakane PK, Moriuchi T, Koji T, Tanno M, Abe K. In situlocalization ofmRNA using thymine-thymine dimerizedcDNA. Acta Histochem Cytochem 1987;20:229-43.

10 Heiles HBJ, Genersch E, Kessler C, Neumann R, EggersHJ. In situ hybridization with digoxigenin-labeled DNAof human papillomaviruses (HPV16/18) in HeLa andSiHa cells. BioTechniques 1988;6:978-81.

11 Schafer R, Zischler H, Epplen JT. DNA fingerprintingusing non-radioactive oligonucleotide probes specific forsimple repeats. Nucleic Acids Res 1988;16:9344.

12 Herrington CS, Bums J, Graham AK, Evans M,McGee JO'D. Interphase cytogenetics using biotin anddigoxigenin labelled probes I: relative sensitivity of bothreporter molecules for detection ofHPV16 in CaSki cells.J Clin Pathol 1989;42:592-600.

13 Herrington CS, Bums J, Graham AK, Evans M,McGee JO'D. Interphase cytogenetics using biotin anddigoxigenin labelled probes II: simultaneous differentialdetection of human and papilloma virus nucleic acids inindividual nuclei. J Clin Pathol 1989;42:601-6.

809

on March 25, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.43.10.806 on 1 October 1990. D

ownloaded from