[CANCER RESEARDI 39. 3479-3484. S@m@er 19791O006-5472/79/0tX19.OrSJ@_00

Analysis of Human Cancer DNA's for DNA Sequences of HumanAdenovirus Serotypes 3, 7, 11, 14, 16, and 21 in Group B'

William S. N. WOId,2 Jesse K. Mackey, Patricia Rigden, and Maurice Green3

Institute k;r A*@da W,@. &@ 1.h@@y Schoof of Me@cie. St. Louis, Missouri 63110

ABSTRACT

We have investigated whather Group B human adenoviruses(Ad) (Ad3, Ad7, Mi 1, Ad14, Mi 6, and Ad21 ), which arewidespread in the humat population and are tumOrigenic inhamsters, may play a mis in human cancer. Hybridization ofAd7-radiolabeled DNA with DNA's from an Ad7-induced primary hamster tumor aid from twe cell lines (5728 and Ad7 Pcell) established from Ad7-induced hamster tumors indicatedmultiple copies per cell of 17, 30 to 36, and 20%, respectively,of the Ml genome. Thus, cells transformedby Group B Adsresemble cells transformed by Group C and Group A Ad's inthat they retain nsjltiple copies of a variable fraction of the viralgenome. These model atutbes suggest that possible Group BAd-induced human cancer cells should contain one or morecopies of virus DNA per caL Therefore, we assayed humancancer DNA's for Ad sequences, by highly sensitive “saturation-hydndizahon r@tions with *47 or Mi 1 DNA (4 x 10@to2.1 x 1O@ cpm/pg).We concentratedon cancersoftherespiratory and @gestivesystems, because these systems arethe most common sites of E@echonby Group B Ad's. In 8independent experiments, no Ml sequences were detected inDNA's from 16 normal kmg tissues, 18 normal tissues of thedigestive system, 34 cancers of the respiratory system, 19cancers of the digestive system, 11 cancers of the urinarysystem, 5 cancers of the genital system, 3 cancers of thebreast, and 6 Hodgkin's lyn@homas. Reconstruction controlswith added Ad7 DNA indicated that about 0.05 to 0.1 copy ofAd7 DNA per cell allould be detected. Mi 1 is strongly implicated as a cause of acute hemorrhagic cystitis. In two independent experiments, no Mi 1 sequences were detected inDNA's from 9 carcinomas of bladder, 10 carcinomas of prostate, 24 caranomas of kMk@ey,3 hypernephromas, 3 Wilma'tumors, or 2 normal kiduey& Reconstruction experiments indicated that the cancer DNA @ayshad a sensitivity of 0.05 to0.1 copy of Adl 1 DNA per cal The DNA's of Group B Ad'sare >85% homologous by hytwidlzation; thus, these resufts areapplicable to all Group B serolypes. Our data provide evidence(but not formal iwoof@that none of the human cancers that weanalyzed were induced by @koupB Ad's. These tumors represent about 50% of the kunors that affect humans. Thepossible involvement of Group B Ad's in other less commonforms of human cancersis WXIerinVeStigatIOnin our laboratory.

1@ work@ a4@ndnd by@ NOl @P43359 from the V'wus Cancer

Program @mins@ Nrnunat@ b@da asl by Contract 223-74-1022 fromthe Bureai of BioIo@ca@Foud aat I@vg Adainintration.

2 To whom@@@@ tie akkessed.

3 Rec@@ieat of USPHS l@Ji @eer A@d 5k6.A14739 from the National

institute of AIer@ @atkda4@th@@Received Seg@esther15, 1978 aci@4nd May 31 . 1979.

INTRODUC@ON

The 31 recognized human Ad@serotypes (Ml to Ad31) form5 distinct groups (A to E), based upon DNA genome homologiesestablished by molecular hybridization experiments (17, ‘I8,43). The DNA's of Group B Ad's (Ad3, Ad7, Adi ‘I, *414,Adl6,and Ad21) are>85% homologousbutare c@2O%homologous with DNA's of non-Group B types. The epidemiological and pathogenic properties of Group B Ad's are reviewedinRefs.24,25,38,and 44.Ad3 and *47 aremore commonthanAdil,Ad14,Ad16,and *421.*47,and lessfrequentlyAd3, causes epidemicsof ARD among militaryrecruits in theUnited States, and more recently in Europe (see Refs. 5, 6,and 41). The respiratory tract tissues infected range from thenasopharynx to the pulmonary lobes. To alleviate ARD, entericcoated vaccines are being used that release hve unattenuatedAd7 into the intestines(3, 6). *414, *416, and *421 havefrequently caused epidemics of ARD in Europe (5, 38, 41).Although there may be serotype differences, in general, GroupB Ad's cause a variety of diseases in tissues that include theentire respiratory tract, much of the gastrointestinal tract, theurogenital tract, and the eye (2, ‘Ii,12, 24, 25, 38, 46).Serology studies indicate that the majority (50 to 80%) ofadolescents and young adults have been infected by some orall of the Group B types (5, 6, 29, 32, 41).

There is substantial evidence implicating Adi 1 as art etiological agent of acute hemorrhagic cystitis. Nwnazald at a!. (31,32) and Mufson et a!. (29, 30) isolated Adi 1 from the urineand/or demonstrated seroconversion to Adi 1 in a large fraction of patients with this disease. Using Enmunolluorescencetechniques, Beishe and Mufson (1) identified Ad's in bladderepithelial cells in urine sediments from adenoviremic patientswith acute hemorrhagic cystitis. Adi 1 hasalso been associatedwith acute hemorrhagic cystitis in immunosuppressed renaltransplant patients (8, 26). Considering that Adi 1 is a rareserotype (2, 12, 24, 38), and that Ad's are rarely isolated fromurine, these studies suppod an Adl 1 etiology of acute hemorrhagic cystitis.

All the human Ad types (except *44) have been shown toinduce tumors in experimental animals, and/or to transformcultured cells. Group B Ad's induce tumors in a low fraction ofnewborn hamsters within 4 to 18 months (15-1 7, 23, 43).Hamsters bearing virus-induced tumors produce antibodiesthat react specifically with ‘T-antigens―produced in cellsinfected with or transformed by Group B Ad's (23). The Tantigens consist of several viral-coded proteins, including oneor more whose function(s) is thought to maintain the bansformed phenotype of the tumor cell. Cells transformed by

4 The @t,reviations used ae: Ad. nden@ius A@, @de .a@t1iI@,,er@

re@ dinease.

SEPTEMBER 1979 3479

on June 30, 2020. © 1979 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

w. S. M. Wo!d et a!.

Group B Ad's have not been extensively studied. This is incontrast to cells transformed by Group C and A Ad's.

Considering that Group B Ad's have oncogenic propertiesand are widespread in the human population, an importantquestion is whether these viruses cause human cancer. Severalscenarios of Ad-induced human cancer are conceivable (20,44). As one example, an Ad could infect a cell, and the viralgenome could become stably associated with the cell, eitherby integrating into the cell DNA or by forming a latent infection.As the individual ages, the transforming information inherent inthe viral genome could become activated to produce mRNAand functional protein and transform the cell to produce atumor. Studies with tumors induced by and/or cells transformed by Group A and C Ad's, and as shown in this communication by Group B Ad's, indicate that all tumors and transformed cells retain viral DNA sequences. Therefore, if Group BAd's cause human cancer, cancer DNA's should contain GroupB Ad DNA sequences. This communication presents results ofour analyses of 125 distinct human cancer DNA's and 37normal tissue DNA's for Group B Ad sequences, by molecularhybridization with highly radioactive Ad7 or Adi 1 DNA.

MATERIALS AND METHODS

Human tumors were obtained from Dr. J. Gruber and Dr. I.Sekely (Office of Program Resources and Logistics, NationalCancer Institute), Dr. E. Harrison (deceased) (Mayo Clinic), Dr.M. Gardner (University of Southern California), and Dr. H.Pinkerton (deceased) (St. Louis University). Human and hamster tumor DNA's and Ad7-transformed cell DNA's were extracted as described elsewhere (22, 44). Two different Ad7tumor cell lines, designated ‘5728' ‘(13) and ‘‘Ad7P-cell―(derived by R. Huebner, obtained from Flow Laboratories),established from Ad7-induced hamster tumors, were propagated in Eagle's minimal essential medium containing 5% fetalcalf serum.

Ad7 and Adi 1 DNA's were used as radiolabeled probes(32). These DNA's are characteristic of Group B DNA's, because (a) their buoyant density is typical of Group B Ad's (1,34), (b) they hybridized >85% to Group B Ad DNA's but <20%to non-Group B Ad DNA's (2), and (c) the Ad7 DNA had thesame restriction endonuclease patterns5 as have been published for Ad7 DNA (35, 39).

Standard procedures were used for growth of virus in suspension culture, plaque assay of virus, and protease-assistedextraction of viral DNA (21 , 44). Ad DNA was labeled in vitrowith [3H]dTTP or [32P]dCTP,using the ‘‘nicktranslation' ‘procedure, as described in detail earlier (28, 44). The specificradioactivities of the various probes are given in ‘‘Results.―

All hybridizations were conducted in duplicate. Reactionmixtures contained 6 mg of DNA per ml (tumor, transformedcell, or calf thymus), 500 or 1000 cpm of probe DNA, 0.72 MNaCI, 10 mM piperazine-N,N'-bis(2-ethanesulfonic acid) (pH6.7), 1 mM EDTA, and 0.05% sodium dodecyl sulfate. Hybridization was quantitated by batchwise chromatography on hydroxyapatite (44). In all tumor DNA analyses, reconstructionexperiments were done to estimate the sensitivity of the experiments. The reconstruction reactions contained 0.2 copy

5 W. S. M. Wold, J. K. Mackey, P. Rigden, and M. Green, unpublished data.

(equivalent) or 1 copy of unlabeled Ad DNA per cell, and 6 mgof calf thymus DNA per ml in place of tumor DNA. A molecularweight of 4 X 1012was used for diploid cell DNA. The Ad7-transformed cell DNA was also assayed at concentration of 6mg/mI.

RESULTS

Hybridization Properties of in Vitro-labeled Ad7 DNAProbes Used In Tumor Analyses. To test whether humantumor DNA's contain Group B Ad-specific DNA sequences, weperformed ‘‘saturation-hybridization'‘experiments with in vitro3H- or 32P-Iabeled Ad7 or Adi 1 DNA. As detailed elsewhere(28, 44), these in vitro-labeled probes are uniformly labeledand reassociate with expected kinetics. The specific activitiesof the probes used were high, ranging from 4.3 x 106 to 2.1x 108 cpm/@g. With such probes, it is possible to performhighly sensitive saturation-hybridization experiments, in whichreassociation of the probe is low, and in which most of theprobe will be driven into a hybrid form by viral DNA present ata concentration of 1 copy per diploid quantity of cell DNA.Chart 1 shows typical hybridization kinetics of 3H-labeled Ad7probe with 1 copy of Ad7 DNA per cell (in the presence of 6mg of calf thymus DNA per ml). After a 48-hr hybridization (theusual time for our tumor studies), 85% of the probe had formeda hybrid, whereas only 15% had self-annealed. Thus, the probecould easily detect less than 1 copy of Ad7 DNA per cell. Thissensitivity generally cannot be achieved by other procedures(e.g. , reassociation kinetics) often used to assay tissue DNA'sfor viral sequences.

Analysis of Ad7-induced Tumor DNA's and Tumor CellLines for Group B Ad-specific DNA Sequences. Chart 1 showsthe hybridization of an Ad7 probe with DNA extracted from anAd7-induced hamster tumor and with DNA from 2 Ad7-tumorcell lines, 5728 and Ad7 P-cell, established from 2 differenthamster tumors. The hamster tumor DNA drove 17% (above

z0

0

Chart 1. Analysis of DNA's from an Ad7-induced hamster tumor and from 2Ad7 hamster tumor cell lines for Ad7 sequences. Ad7 3H-labeled DNA (11.1 x1o@cpm/@g)was annealed with the following: •,1 copy equivalent of Ad7 DNAper diploid quantity of cell DNA plus 6 mg of thymus DNA per ml; S. 2.4 mg ofuncloned 5728 hamster cell DNA per ml plus 3.6 mg of calf thymus DNA per ml;A. 2.4 mg of 5728 (clone 6) hamster cell DNA per ml plus 3.6 mg of calf thymusDNA per ml; & 2.4 mg of Ad7 P-cell hamster tumor cell DNA per ml plus 3.6 mgof calf thymus DNA per ml; 0, 2.4 mg of Ad7-induced hamster tumor DNA per mlplus 3.6 mg of calf thymus DNA per ml and 0, 6 mg of calf thymus DNA per ml.Hybridization conditions and assays are described in “Materialsand Methods.―Each time point represents duplicate 50-id aliquots containing 500 cpm (4.6 x105@tg)of probe DNA.

TI@I ( hours

3480 CANCER RESEARCH VOL. 39

on June 30, 2020. © 1979 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Analysis of human cancer DNA@ for Group B adenovirus-specificDNAsequencesHybridizations

were conducted to an equivalent Cotof 2 x 1o@.EquivalentC,trefersto the product of the initial DNA concentration, as mol of nucleotideperliter,

and the hybridization time in eec, with the hybridization rate correctedforanequivalent Na@concentration of 0.1 8 M. The specific activity of the@H-iabeledAd7

probewas4.3 x lO@ @%ofhybridizationBackground

notLung cancer DNA SubtraCted― CorrectedvaIues@'Carcinoma

(1134f 23.1 , 23.5 0.9,1.3Carcinoma(1235) 25.8, 25.8 3.6,3.6Carcinoma(1236) 17.8, 20.9 —4.4,-‘3.1Carcinoma(1339)

24.8,22.8 2.6,0.6Carcinoma(1341)25.9,24.7 3.7,2.5Carcinoma

(1366) 27.7, 26.4 5.0,4.6Carcinoma(1518)24.7,23.9 2.5,1.7Carcinoma

(1519) 26.3, 24.8 4.1 ,2.6Carcinoma(1556)16.7,20.7 —5.5,—5.1Carcinoma

(1555) 17.9, 21.0 —4.4.—1.2Mean23.2Squamous

cell carcinoma(1293) 21.5, 16.3 —0.7,—5.9Squamouscell carcinoma (1295) 19.7, 20.0 —2.5,—2.2Squamouscell carcinoma (1307) 22.0, 21.2 —0.2,—1.05quamouscell carcinoma (1669) 18.2, 21 .3 —4.0,—0.9Squamouscell carcinoma (1703) 17.8, 17.8 —4.4,—4.4Squamouscell carcinoma (1167) 25.0, 23.1 2.8,0.9Mean

20.3Hybridization

controlsCallthymus DNA 22.200.2copyAd7DNApercell 34.715.11

copy Ad7 DNA per cell 80.774.7267copies Ad7 DNA per cell 98.70Ad7tumorcell(5728,clone6) 52.137.2DNA

Group B Adenoviruses and Human Cancer

background) of the probe into a hybrid form. DNA from uncloned and clone 6 line 5728 cells gave 30 and 36% hybndization, respectively, and DNA from Ad7 P-cells gave 20%hybridization. Reassociation kinetic analyses suggested that5728 cells, Ad7 P-cells, and the hamster tumor, respectively,contain 70 to 90 copies, 70 copies, and 6 or 7 copies of Ad7DNA per cell (data not shown). Thus, these tumor cell DNA'scontain multiple copies of only a fraction of the Ad7 genome,which our probe could readily detect after only about 7 hr ofhybridization. These studies suggest that DNA's from possibleGroup B Ad-induced human cancers should contain detectableviral DNA sequences. Further, if Ad7-induced human tumorcells resemble Ad7-induced hamster tumor cells, we shouldeasily detect Ad7 sequences under our standard conditions of48- or 72-hr hybridization.

Analysis of Human Lung Tumor DNA's for Group B Adspecific DNA Sequences. Table 1 presents the results of atypical analysis of human cancer DNA's for Group B Ad-spacific sequences. Both the raw data (duplicate hybridizationvalues without background subtraction or normalization) andthe corrected values are given. Reconstruction experimentswith added Ad7 DNA are also shown, to provide an estimate ofthe sensitivity of the tumor assays. The probe self-annealed22.2% in the presence of 6 mg of calf thymus DNA per ml.None of the cancer DNA's caused an increase in the reassociation of the probe, as would be expected from DNA samplescontaining Group B Ad-specific DNA sequences. Under thesame conditions as the tumor DNA assays, 0.2 copy of Ad7DNA per cell gave 15.1% hybridization, and 1 copy gave74.7% (corrected values). DNA from 5728 (clone 6) cells gave37.2% hybridization. Because there is little scatter to the tumordata, 5 to 10% hybridization above background would besignificant. We estimate that our probe could detect about 0.1copy of the complete Ad7 genome per cell. Therefore, weconclude that these tumor DNA's do not contain detectable (ata sensitivity of 0.1 copy per cell) Group B Ad DNA sequences.

Summary of Analyses of Human Cancer DNA's for GroupB Ad-specific DNA Sequences. Table 2 summarizesthe resultsof 8 experiments, using 8 distinct radiolabeled Ad7 DNAprobes, in which DNA's from different types of human cancersand normal tissues were assayed for Group B Ad-specific DNAsequences. Only the mean corrected hybridization values aregiven (the data are typical of that in Table 1). No viral sequenceswere detected in DNA's from 16 normal lung tissues, 18 normaltissues of the digestive system, 34 cancers of the respiratorysystem, 19 cancers of the digestive system, 11 cancers of theurinary system, 5 cancers of the genital system, 3 cancers ofthe breast, and 6 Hodgkin's lymphomas.

Table 3 presents the results of reconstruction experimentsconducted for each of the 8 cancer analysis experiments,giving both uncorrected and corrected hybridization values.Depending upon the experiment, 0.2 copy of Ad7 DNA per cellgave from 13 to 31% hybridization (above background), and 1copy gave 59 to 70%. Therefore, these probes should havedetected viral DNA sequences in tumor DNA's containing about0.05 to 0.1 copy of Ad7 DNA per cell.

Analysis of Human Urogenital Cancer DNA's for Adi 1-specific DNA Sequences. As noted above, Adi 1 has beenshown to infect bladder epithelial cells and is thought to be acause of acute hemorrhagic cystitis. Therefore, Adi 1 may bea cause of human urogenital cancer. Using the Ad7 probe, no

Table 1

aThevaluespresentedrepresentduplicatehybridizationvalueswithoutsubtraction of the probe self-annealing background (22.2%; i.e., the expected value).

b iercentage of hybridization with the self-annealing background (22.2%)subtracted and the resulting values normalized to maximum hybridization In thepresence of excess Ad7 DNA (98.7%).

C Numbers in parentheses, tumor inventory numbers in the tumor storage

facility at the Institute for Molecular Virology.

viral sequences were detected in DNA's from 16 urogenitalcancers (Table 2). The Ad7 probe hybridizes about 88% withAdi 1 DNA (18, 43) and therefore probably would have dotected Adi 1 DNA sequences if present. However, we cannotexclude the possibility that the DNA's might contain a portionof the Adi 1 genome that is not homologous with Ad7 DNA.Therefore, we also assayed urogenital cancer DNA's using invitro-labeled Adi 1 DNA as probe. Table 4 shows typical rawdata obtained by saturation-hybridization of an Adi 1-radiolabeled DNA probe with DNA's from carcinomas of the bladderand the prostate. The hybridization values obtained with alltumor DNA's cluster closely about the mean values of 21 .9 and22.1 % for carcinomas of the bladder and prostate, respectively. Further, the probe self-annealed to the same extent inthe presence of calf thymus DNA or the tumor DNA's. Weconclude that none of the tumor DNA's contained detectableAdi 1 DNA sequences. In reconstruction experiments, 0.2 copyof added Adi 1 DNA per cell gave 40.7% hybridization, 1 copyper cell gave 60.2%, and 267 copies of Adi 1 DNA per cellgave 90%. After subtraction of background (22. 1%) and normalization to maximum possible hybridization (90.5%), 0.2copy of Adi 1 DNA gave 20.6%, and 1 copy gave 42.9%.Based upon these reconstruction values, we estimate that wewould have detected, at about 10% hybridization above back

SEPTEMBER1979 3481

on June 30, 2020. © 1979 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Summary of 8 experiments in which DNA s from normal tissues andhumancancerswere analyzed for Group B Ad-specific DNAsequencesThe

values given represent data compiled from 8 separate experiments.Table3gives reconstruction hybridization values, probe-specific activities, and equiv

alent C0tvalues for each experiment.%

ofhybridization(correctedDNA

sample No.values)8Normal

tissuesLung16 0.0±Ileum

1‘-3.9Colon17 —0.8±2.1Rectum

1—2.1Cancer

tissuesRespiratorysystemCarcinoma

of lung 15 0.8 ±2.8Squamouscell carcinoma lung 9 —1.0 ±2.5Adenocarcinoma

of lung 4 —1.7 ±3.0Oatcell carcinoma of lung 3 —0.1±2.7Metastatic

hemangiopericytoma 13.6Metastaticembryonal carcinoma of lung 10.2Epidermoidcarcinoma of lung 11.4Digestive

systemCarcinomaof stomach 5 —0.3 ±2.7Leiomyosarcoma

of intestine 1—3.8Carcinomaofcolon7 —2.1±1.7Adenocarcinoma

of colon 3 —1.1 ±2.0Carcinomaof cecum 2—0.3Adenocarcinoma

of cecum 10.5UrinarysystemCarcinoma

of bladder 7 —0.4±1.0Carcinomaof kidney 10.3Hypernephroma

20.9Sarcomahypernephroma 12.7Genital

systemCarcinomaof prostate 4 0.6 ±1.2Carcinomaof ovary 12.2Breast

carcinoma 3 —0.9±2.1Hodgkin'slymphomaLymph

node 3 1.7 ±0.6Spleen11.5Lung2 2.5

Reconstruction hybridization controls for the human tumor analyses summarized in Table2Allhybridizations were done at a final DNA concentration of 6 mg/mI and were carried to an equivalent C0tof 2or3

)( 10@.Calf thymus DNA was used in the reconstruction experiments. With the 3H-labeledprobes, hybridizationsweredonein duplicate 5O-pl aliquots, each with 500 cpm of probe. With the 32P-Iabeledprobes, hybridizations wereinduplicate

10-@tlaliquots containing 1000 cpm ofprobe..

. Self-annealProbe specific ac- in ‘@k- 0.2 copy Ad7 1.0 copy Ad7 267 cop,es 5728 clone 6

tivity DNA DNA Ad7 DNA DNAgroundExperi

ment Label cpm/g@g Ub C U C U C U C UC1

3H 4.3 x 106 22.2 0 34.7 15.1 80.7 74.7 98.7 100 52.137.223H 4.3 x 106 20.2 0 30.4 12.7 77.1 70.2 95.8 100 48.432.133H 4.3 x 10@ 19.9 0 39.1 24.9 66.4 60.3 97.0 100 47.636.043H 1.0 x i0@ 19.9 0 ND ND 62.2 59.9 90.7 100 48.236.953H 1.0 x 10' 20.2 0 45.1 31.2 67.0 58.5 97.6 100 49.236.1632P 2.1 x 10° 11.4 0 ND ND 69.3 72.5 91.6 100 NDND732P 1.5 x i0@ 15.3 0 ND ND 77.0 76.3 95.7 100 NDND832P 1.3 x 108 16.4 0 ND ND 81.5 78.5 93.3 100 48.7 37.5

w. S. M. Wo!d et a!.

ground, Adi 1 DNA sequences in a tumor DNA sample containing 0.1 copy of Adi 1 DNA per cell.

Table 5 summarizes the results of another set of tumorassays, using a different probe, and different tumor samplesfrom that in Table 4. The hybridization values are presented

Table 2

without subtraction of background (15.6%) or normalization tomaximum possible hybridization (90.1 %). No Adi 1 DNA sequences were detected in DNA's from 9 carcinomas of bladder,10 carcinomas of prostate, 24 carcinomas of kidney, 3 hypernephromas, 3 Wilms' tumors, or 2 normal kidneys. In reconstruction experiments, 0.2 copy of added Adl 1 DNA per cellgave 13% hybridization above the self-annealing backgroundof 16%, and 1 copy per cell gave 41 % hybridization abovebackground. Thus, as in the experiment presented in Table 4,we would have detected Adi 1 sequences in a tumor containing1 copy of 5 to 10% of the viral genome per cell.

DISCUSSION

The objective of this investigation was to examine whetherGroup B (Ad3, Ad7, Adll, Ad14, Ad16, and Ad21) humanAd's might be a cause of human cancer. Our approach was toassay cancer tissue DNA's for Ad7 or Adi 1 DNA sequences,based on the assumption that DNA's from Group B Ad-inducedcancers would contain Group B Ad DNA sequences. Thisapproach is based upon a very large number of animal modeland cultured cell studies with Group C and Group A Ad's andon a limited number of studies with Group B Ad's. With GroupC Ad's, all viral-transformed cells contain the leftmost 12 to14% of the viral genome, usually in more than one copy (10,14). Multiple copies of variable portions of the remainder of thegenome may also be present. With Group A Ad's, viral-transformed cells and viral-induced tumors usually contain multiplecopies of most or all of the complete genome (7, 21, 22, 27).The genetic information responsible for transformation is located within the extreme molecular left (7 to 8%) of the genomes of Groups C and A Ad's, because only restrictionendonuclease DNA fragments that contain these sequencescan transform cells in transfection experiments (37, 40, 41,45). DNA's from Adl 2-induced hamster tumors and Adi 2-transformed hamster cells contain similar patterns of integratedAdi 2 DNA sequences, suggesting that transformed cells arerealistic models of Ad tumorigenesis (22). All Group C and AAd-transformed and tumor cells synthesize virus-specific RNA(9, 19, 33), and all cells also produce virus-specific immunologically reactive proteins. It is highly probable that the transformed phenotype of viral-transformed or tumor cells is maintamed by the constant functioning of a viral-coded protein. It

Table 3

8The self-annealing background of the probe has been subtracted, and thevalues were normalized to maximum hybridization in the presence of 267 copiesof Ad7 DNA per cell (see Table 3).

b Mean ±S.D.

a u. uncorrected hybridization values; C, corrected hybridization values; i.e. , the self-annealing background hasbeen subtracted and the values normalized to maximum hybridization with 267 copies per cell of Ad7 DNA; ND, notdone.

3482 CANCERRESEARCHVOL. 39

on June 30, 2020. © 1979 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

DNAsample%

of hybridization (backgroundnot subtracted)Normali@edmeanCarcinoma

of bladder (201 1)cCarcinoma of bladder (1574)Carcinoma of bladder (1976)Carcinoma of bladder (2002)Carcinoma of bladder (2601 )Carcinoma of bladder (2595)Carcinoma of bladder (1532)

Mean ±S.D.22.1

. 22.519.5, 20.319.3, 20.324.9, 21.523.0, 21 .818.8,24.421.5,26.121.9 ±2.50.3

—1.5—1.6

1.70.5

—0.7—0.4Carcinoma

of prostate (1971)Carcinoma of prostate (1972)Carcinoma of prostate (1864)Carcinoma of prostate (1969)

Mean ±S.D.23.9,

23.418.9, 22.219.9, 20.724.5, 23. 122.1 ±1.92.2

—2.2—2.6

2.5Hybridization

controlsSelf-annealing in the presence of calf

thymus DNA0.2 copyAdi 1 DNApercell1 copy Adl 1 DNA per cell267 copies(10 @g/ml)Adi 1 DNAper

cell22.1

40.760.290.50

20.642.9100

DNANo.of sampIes%

of hybridization(background not sub

tracted)Tissue

DNACarcinomaof bladder91 4.0 ±13aCarcinomaof prostate1 01 2.6 ±1.3Carcinomaof kidney241 2.5 ±1.9Hypernephroma

kidney31 1.1 ±0.3Wilms'tumor310.7 ±1.6Normalkidney21 4. 1 ±0.6Reconstruction

controlsCalfthymusDNA15.60.2

copy Adl 1 DNA percell29.01.0 copy Adi 1 DNA percell56.9Copies

Adl 1 DNA per cell90.1

Group B Adenoviruses and Human Cancer

Table4Analysis of human bladder arid prostate tumors for Adi 1-specific DNA

sequences

of the genomes of Ad3 and Ad7 is very similar to that of GroupA and B Ad's (39). Cell lines established from hamster tumorsinduced by Group B Ad's are known to contain Viral-specificDNA and RNA (19). As shown in this communication, DNA froman Ad7-induced hamster tumor contained 17% of the Ad7genome, and DNA's from 5728 and Ad7 P-cell hamster tumorcell lines contained 30 to 36 and 20% of the Ad7 genome,repectively. Very probably, the molecular left end of the Ad7genome is present in these DNA's. We estimate that, per cell,the tumor DNA contains 6 or 7 copies of the Ad7 sequences,5728 cells contain 70 to 90 copies, and Ad7 P-cells contain70 copies. Earlier hybridization-kinetic studies, using 32P-labeled Ad7 DNA restriction fragments, suggested that uncloned5728 cells contained 200 to 400 copies of 10 to 20% of theAd7 genome (13), in reasonable agreement with our currentestimate of 70 to 90 copies of 30% of the genome. Mostimportant, these results indicate that hamster tumors inducedby Group B Ad's contain multiple copies of a variable fractionof Ad7 DNA sequences and that these sequences are easilydetected after as little as 1 hr under our hybridization conditions. If possible Group B Ad-induced human tumors resemblehamster tumors, we should readily detect viral sequences inthe human tumors DNA's.

We did not detect Ad7 DNA sequences in any of the DNA'sfrom 31 cancers of the respiratory system, 19 cancers of thedigestive system, 16 cancers of the urogenital system, 3 cancers of breast, and 6 Hodgkin's lymphomas. Eight experimentswere conducted, using 8 different probes. In all experiments,controls were done with DNA's from Ad7-transformed hamstercells, and with Ad7 DNA added to calf thymus DNA. Thereconstructions indicated that we should have detected Ad7sequences in a tissue DNA containing about 0.1 copy of Ad7DNA per cell (or 1 copy of 10% of the Ad7 genome per cell).Most of the tumor samples were obtained from the NationalCancer Institute; all have been diagnosed by histological examination to be malignant, and some have been certified tocomprise mainly (>50%) tumor cells. Therefore, the sensitivityof our assays is within 0.1 to 0.2 copy per cell. Consideringthis, and that most Ad-induced tumors and -transformed cellscontain multiple copies of from about 12 to possibly 100% ofthe viral genome, these data are strong evidence that none ofthe human cancers examined were induced by Ad7.

Adi 1 is probably a cause of acute hemorrhagic cystitis. In 2experiments, we did not detect Adi 1 sequences in the DNA'sfrom 17 cancers of bladder, 15 cancers of prostate, 24 cancersof kidney, 3 hypernephromas, 3 Wilms' tumors, or 2 normalkidneys. These studies were done at a sensitivity of about 0.1copy of Adi 1 DNA per cell and therefore are strong evidencethat none of these urogenital cancers were induced by Adl 1.

Ad7 and Adi 1 are >85% homologous with other Group BAd's. Therefore, these data are probably also applicable toAd3, Adl 4, Adi 6, and Ad21 . However, these data cannot beextended to Ad types not in Group B, because the Ad7 andAdi 1 probes hybridize <20% with DNA's from non-Group Btypes. We note that it is theoretically possible that Ad-inducedcancers could contain 1 copy of about 3 to 6% of the viralgenome per cell (the estimated size of the Group C Ad-transforming region), so that these data are not formal proof that thecancers tested did not have a Group B Ad etiology.

In conclusion, although Group B Ad's have oncogenic properties, cause respiratory and probably bladder disease, and

aActualhybridizationvaluesnotcorrectedforself-annealingoftheprobeinthe absenceof homologousDNAor normalizedto maximumhybridizationobserved (90.5%) with excess Adl 1 DNA. Hybridizations were carried to anequivalent C0t of 20,000. The probe was 3H-labeled Adi 1 DNA with specificactivity of 8 x lO@cpm/ig.

b Background due to self-annealing ofthe probe (22.1 %) was subtracted, andresulting values were normalized to the maximum percentage of the probehybridized with excess Adi 1 DNA (90.5%). S.D. of all values is ±1.2%.

C Numbers in parentheses, tumor inventory numbers in the human tumor

storage facility at the Institute for Molecular Virology.

Table 5Summary of analyses of human bladder, prostate, and kidney cancers, and

normal kidney tissues, for Adi 1-specific DNA sequences

aMeanvaluesforeachtypeoftissueDNA,withoutsubtractionofselfannealing background (15.6%) or normalization to maximum hybridization withAdi 1 DNA(90.1%).Theprobewas @P-IabeledAdl 1 DNAwithspecificactivityof 1.6 x 108cpm/@g.Hybridizations were carried to an equivalent C,,tof 25,000.

therefore follows that all viral-induced human cancers shouldcontain viral DNA, RNA, and protein sequences. We cannotexclude the possibility that Ad's may play a transient role inhuman cancer that does not necessitate the persistence ofviral-transforming sequences, such as potentiating the actionof chemical carcinogens. However, there is no evidence forthis possibility from animal model studies.

Although considerably less is known about Group B Ad's, itis very likely that the studies with Group C and A Ad's can beextended to include Group B Ad's. As with Group C and AAd's, restriction fragments from the left end of the Ad3 or Ad7genome are able to transform cells (36). Also, the organization

3483SEPTEMBER1979

on June 30, 2020. © 1979 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

19. Green,U., Psona, J. T.,Piña,It, Fujinaga,IC,chattier,H.,andLaidgrafLaura, I. Transcription di adenovirus genes in prodectivoly Infected end intransformed cells. Cold Spring Harbor Symp. Quant. 9101.35: 803—818,1970.

20. Green, It, and Wold, W. S. N. Oncogenic DNA viruses—replication, tumorgene expression, and role In humancancer. Semin. Oncol., 3: 64-79, 1976.

21. Green,U. R., chinnadurai,G., Mackey,J. K., and Green,M. A uniqueI*ttem of E@egr@adviral genes in tuenster cells transformed by highly0(100_c human adeno*us 12. Cal, 7: 419-428, 1976.

22. Green, M. R., Mackey, J. K., and Green, M. Multiple copies of humanadenovirus 12 genomes are Integrated In virus-Induced hamster tumors. J.V'wol.,22: 238—242,1977.

23. Huebner, R. J. Adenovinis.dlrected tumor and T-tiigens. 1n M. Pollerd(ad.), Perspectives In Wology, Vd 5, pp. 147-166. New York: AcademicPress, 1967.

24. Huebeer,R. J., Rowe,W. P., and cha.ock, R. M. Newly recognizedresplr@ory hect viruses. Annu. Rev. Microtsol., 12: 49-76, 1958.

25. Jeckson, G. G., and Muldoon, R. L ROcMrUSeSandadenOViWSeS.In: virusesCausing Common Laboratoty Infectionsin Man, pp. 165-197. chIcago: TheLkiiversity of chicago Press, 1975.

26. Lecaleas, G.. Prozeeliy, 0. W., end van Wet, J. Adenowinis type 11 associaled with bemorvt*agh@cyalitis after renal hansp@ation 5. AIr. Med. J.,4& 1932, 1944.

27. Lee. K. C., aid Mali, 5. Adenovius type 12 DNA sequences in primaryIusn@ertumora. J. Virot, 24: 408-41 1•1977.

28. Mackey, J. K., Breclimsen, K. B., Green, M. R., and Green, M. Preparationand characterization of highly radioactive in vitro labeled adenovirus DNAarid DNA reatriction fragments. Biochemiatry, 16: 4478-4483, 1977.

29. Mufeon, N. A., Belatie. ft. B., Horrigen, T. J,, and ZoUar, L M. Canse ofecute hemorvha@ccyathtleIn chikken. Am. J. Die, QuId., 126: 605—609,1973.

30. Mutson, IL A., Zolar, L M., Ma*ad, V. N., end Manalo, 0. AdenovirusInfadlon in a:ule hemorrhagic cystitis. Am. J. Die. child., 121: 281-285,1971.

31. Numazakl, V., Kumasalia, T., Yano, N., Yamariaka, M., Miyazawa. T., Takei,5., arid Wilds, N. Fwther Mudy on ecuto hemorrhagiccystitis due toadenovinistype11. N.End. J. Med.,289:344-347, 1973.

32. Numazalil, V., Shigeta, S., kimiasalia, T., Miyazawa, T., Yananalca, N.,Yano, N., Taliel, S., arid Ishida, N. Acute hemonbagic cystitis In children. N.Engl. J. Med.. 27& 700-704, 1968.

33. OdIn, J., Scheidhliann, K. H., Greenberg, R., Westphal, N., and I@er1Ier,W. Transcription of the genome of adenovirus type 12. IL I.4r4* of stMIeRNA from producti@ Infected human cells and abortively infected andtranalurmed tusn@r coNs.J, VIraL, 20: 355-372, 1976.

34, Pills, U., arid Green, M Biochemical atudies on adenovirus multipilcalion.ix. ctiemicatandbasecompositionamiyeleof 28 humanadenoviruses.Proc. Nell. Mad. Sd. U. S. A., 54: 547-551, 1965.

35. Seldkawa, K., and Fupnoga,K Qeavagn meps of human adenovirus type 7DNA by restriction endonuclsaseIb@I and EcoRi. Virology, 82: 509-512,1977.

36. Sekikawa, K., Shlroki, IC, Shinojo, H., Ojima, S., and Fujinaga, K. Trans.tormmion of rat ceU @ieby an adenovirus 7 DNA fragnient. Virology, 88: 1—7, 1978.

37. ShirokI,K., Hanada,H.,Shimo@o,H.,Vano,S., Olina, S., aid Fujinaga,K.Establl@mentand chareciorizalion of ralcel ties Varisiormed by restrictionendonucl@e fragments of adenovirus 12 DNA. Virology, 82: 462—471.1977.

38. Sohier, 8., chardonnet, V., and Prunieras, N. Adenoviruses. Status ofcurrent knowledge. Prog. Med. V'wol.,7: 253—325,1965.

39. Tibbelts, C. Phyalcal organization of subgroqu B human adenovirus gonomes. J. Wit, 24: 564-579, 1977.

40. van der Eb. A. J., aid Houwekig, A. Transformation with specifIc fragmentsof adenawirusONA& U.Analysle of the viral DNA sequences present hi cellsfrwi@nned with a 7% fragment of adenowirus5 DNA. Gene, 2: 133-146,1977.

41. vander Eb.A. J., Mulder,C., Graham,F. L, endHouweling,A. Transfermotion with specific fragments of aienovlrus ONAs. I Isolation of specihctragmei@ with trariskiruning activity of adenovirus 2 and 5 DNA. Gene, 2:115-132, 1977.

42. van der Veeri, J., c@. K. G., and Abarbanel, N. F. W. Patterns of Infectionswith adeno*us types 4, 7 aid 21 in military recruits during a 9-year mxvey.J. Hyg.67: 255-268. 1969.

43. WoId, W. S. N., Green, N., and BUttner, W. Adeno*uses. In: D. P. Nayak(ad.). The Molecular Biology of Animal Viruses, pp. 673-768. New York:Marcel Dekker, Inc., 1978,

44. Wold, W. S. N.. Green, N., and Mackey, J. K. Methods and rationale for@sledi human tumors for nucleic acid sequences of oncogenic human

DNA viruses. Methods Cancer lies., 15: 69-161 , 1978.45. Yano, S., Olivia, 5., Fupnaga,K., Shiroki, K., aid S1*no@o,H. Transformation

ofa ratcal @iebyanadenovirustype12 DNAfragment.virology,82: 214-220, 1977.

46. Yunis, E J., Atchison. R. W., Nicheals, R. H.. and Decicco, F. A. Adenovirusand illeocecal Intussusception. 1al. Invest., 33: 347-351 •1975.

3484

w. S.M. Woldeta!.

infect tissues of the digestive and lymphoid systems, there isno evidence that they cause cancer of tissues of the respiratory, urogenital, digestive, or Iymphoid systems. However,more studies are required to be certain that Group B Ad's arenot carcinogenic in humans. In particular, additional respiratoryand digestive tract tumors must be examined under assayconditions capable of detecting 1 copy of about 1% of the Adgenome per cell. The cancers that we have examined accountfor less than one-half of the types found in the United States(4). Therefore, it is important to test cancers from sites otherthan those included in our study.

ACKNOWLEDGMENTS

We thank H. Thorntonfor assistancewith virus preparationand with cellculture. We thank C. Self for technical assistance. We are grateful to J. Gruber,I. Sekely,E.Harrison(deceased),M.Gardner,andH.Pinkerton(deceased),forproviding us with human cancers and normal tissues.

REFERENCES

1. Beishe.R. B., andMufson.M. A. IdentificatiOnby knmunofiuorescenceofadenoviral antigen in exfoliated bladder epithellal cells from patients withacute hemorrhagic cystitis. Proc. Soc. Exp. Biol. Med., 146: 754—758,1974.

2. Brandt,C. D.,Kim,H. W.,Vargosko,A. J., Jefferies,B. C.,Arrobio,J. 0..Ridge, B., Parrott, R. H., and Chanock, R. M. lnfections io 18,000 infantsand children in a controlled study of respiratory tract disease. I. Adenoviruspathogenicity in relation to serologic type and Illness syndrome. Am. J.Epidemiol, 90: 484-500, 1969.

3. Couch,R. B., Q@anock,R. M., Cafe,T. R., Lang,0. J., Knight,V., andHuebner, R. J. Immunization with types 4 and 7 adenovirus by selectiveinfection of the intestinal tract. Am. Rev. Reap. Die., 88 (Suppl.): 394-403,1963.

4. Cutler, S., and Young, J. (eds.). Third National Cancer Sunrey IncidenceData, pp. 18-19. Bethesda, Md.: National Cancer Institute, 1975.

5. Dingle.J,,andLangmulr,A.D.EpidemiologyofacuterespiratorydiseaseInmilitary recruits. Am. Rev. Reap. Die., 97(Part 2): 1-65, 1968.

6. Dudding. B. A., Top, F. J., Jr., Winter, P. E.. Bueecher, E. L, Lamson, T. H..and Lelbovitz, A. Acute respiratory diseases in military trainees. The adonovirus surveillance program, 1966—1971 . Am. J. Epidemlol.. 97: 187—198, 1973.

7. Fanning, E., and Doerfter, W. lntracellular forms of adenovirus DNA. v. vIralDNA sequences in hamster cells abortively infected and transformed withhuman adenovirus type 12. J. virol., 20: 373-383, 1976.

8. AsIa,N., Payne,J. E.,Borne,T. v.. Poilsawat,5., Schieble,J., andGuze,L B. Roleof adenovirustype 11 in hemorrhagiccystitis secondarytounmunoeupi@ession.J. Urol.. 112: 595—597,1974.

9, Flint, 5. J., Galhmore, P. H., and Sharp, P. A. Camparison of viral RNAsequences in adenovirus 2-transformed and lytically infected cells. J. Mdi.Biol., 96: 47-68, 1975.

10. Flint,5. J..SambrOOk,J.,Williams,J. F.,andSharp,P.A.VIralnucleicacidsin transformed cells l@.A study of the sequences of adenovirus 5-fransformed rodent cells using specific fragments of the viral genome. VIrology,72:456—470,1976.

11. Fox,J. P.,Brandt,C.0., Wassermann,F.E.,H@.C.E. Sigland,I.,Kogon,A., and Elveback, L R. The virus watth program: a continuing surveillanceof viral infections In metropolitan New York families. Am. J. Epidemiol., 89:25—50,1969.

12. Fox, J. P., Hall,C. E., and Cooney,M. K. The Seattleviruswatch.vii.Observations of adenovirus infections. Am. J. Epidemlol., 105: 362-386,1977.

13. Fujinaga.K..Seldkawa,K.,Yamazaki,H.,andGreen,M.Analysisotmultipleviral genome fragments in adenovirus 7 transformed hamster cells. ColdSpring Harbor Symp. Quant. Blol.. 39: 633-636, 1974.

14. Gallimore,P. H., Sharp,P.A., andSambrook,J. VwalDNAin transformedcells. II. A study of the sequences of adenovirus 2 DNA in nine lines oftransformed cells using specific fragments of the viral genome. J. Mel. 9101.,98: 49—72,1974.

15. Girarcl,A. J., Hilleman,M. R., and Zwickey,R. E. Testsin hamstersforoncogenic quality of ordinary viruses including adenovirus type 7. Proc.Soc. Exp. 9101.Med., 115: 1141-1150, 1964.

16. Green,M.SearchforadenovirusmessengerRNAincancerof man.In:R.LClark, R. W. Cumley, J. E. McCoy. and M. M. Copeland (ads.), Oncology,vol.5, pp.156-165.ChIcago:YearBookMedicalPublishers,Inc.,1971.

17. Green, M. Oncogenic viruses. Annu. Rev. Biochem., 39: 701-756, 1970.18. Green,M.,Mackey,J. K.,Wold,W.5. M.,andRigden,P.Thirty-onehuman

adenovirus serotypes (*41 -3) form five groups (A—E)based upon DNAgenome homolysis. VIrology, 93: 481 -492, 1979.

CANCER RESEARCH VOL 39

on June 30, 2020. © 1979 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1979;39:3479-3484. Cancer Res   William S. M. Wold, Jesse K. Mackey, Patricia Rigden, et al.   Adenovirus Serotypes 3, 7, 11, 14, 16, and 21 in Group BAnalysis of Human Cancer DNA's for DNA Sequences of Human

  Updated version

  http://cancerres.aacrjournals.org/content/39/9/3479

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/39/9/3479To request permission to re-use all or part of this article, use this link

on June 30, 2020. © 1979 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from