BACTERIOLOGICAL REVIEWS, Dec., 1965Copyright © 1965 American Society for Microbiology

Vol. 29, No. 4Printed in U.S.A.

Recovery and Identification of Adenovirus inInfections of Infants and Children'

A. J. VARGOSKO, H. W. KIM, R. H. PARROTT, B. C. JEFFRIES, D. WONG,AND R. M. CHANOCK

Children's Hospital Research Foundation, Washington, D.C., and Laboratory of InfectiousDiseases, National Institutes of Health, Bethesda, Maryland

INTRODUCTION.................................................................. 487

MATERIALS AND METHODS ..................................................... 487Population Studied................................................ 487

Specimen Collection, Tissue Inoculation, and Virus Recovery................... 487

Identification................................................................. 488

Selection and preparation of erythrocytes..................................... 489

Treatment of hyperimmune sera............................................. 489

HI test................................................ 490

Neutralization test................................................ 490

Complement-Fixation Test................................................ 490

RESULTS........................................................... 491

Virus Recovery........................ 491

CF Antibody Response........................ 492DISCUSSION AND CONCLUSIONS........................ 494

SUMMARY...................................................................... 495

LITERATURE CITED........................ 495

INTRODUCTIONDuring the past 6 years, the major concern of

our laboratory has been to define the etiologicalrole, ecology, and importance of new or newlyrecovered viruses and Mycoplasma in childhoodrespiratory-tract illness. Between October, 1957,and May, 1961, we studied 7,509 children, andwere able to recover adenoviruses from 504 in-dividuals. This experience made it possible todefine certain facets of behavior of adenovirusesin the laboratory as well as in natural infection.In this paper, we will describe the frequency ofrecovery of different adenoviruses from throatand anal swab specimens, the sensitivity of pri-mary simian and continuous-passage humanheteroploid tissue culture for recovery of adeno-viruses, and the complement-fixing (CF) anti-body response of individuals from whom adeno-viruses were recovered.

MATERIALS AND METHODS

Population StudiedSpecimens for virus isolation were obtained

from 7,509 infants and children from the Wash-ington, D.C., metropolitan area during theperiod from October, 1957, through May, 1961.Of these, 1,879 individuals were admitted to

' Presented at the Annual Meeting of theAmerican Society for Microbiology, Washington,D.C., 6 May 1964.

the hospital for croup, bronchopneumonia, se-vere bronchitis, pharyngitis, or bronchiolitis,and 2,726 patients were seen in the OutpatientDepartment of the hospital for mild bronchitisor pharyngitis. In addition, control patientsfree from respiratory-tract disease were se-lected to match the first two groups as closely aspossible in age, week of study, and other at-tributes. The control group consisted of 1,974ambulatory patients and 930 individuals ad-mitted to the hospital. Almost all individualsin this study population were from a low socio-economic group and were under 6 years of age.

Specimen Collection, Tissue Inoculation, andVirus Recovery

Serum specimens were collected during theacute phase of illness and during convalescencefrom the infants and children who were hospital-ized. Throat and anal swab specimens werecollected by use of dry cotton swabs which wereimmersed in Hank's saline containing 0.5% gela-tin, 500 units of penicillin, 500 ,ug of strepto-mycin, 50 ,ug of tetracycline, and 125 yg ofeither nystatin or amphotericin B per ml. Speci-mens obtained in the clinics or at bedside werekept refrigerated from time of collection untilthey were inoculated into tissue culture approxi-mately 2 to 3 hr later. During periods whentissue cultures were not available, the specimenswere immediately frozen at -65 C and storeduntil tested. Over 90% of the specimens were

487

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VARGOSKO ET AL.

inoculated into tissue culture immediately,whereas those specimens stored frozen wereusually inoculated into tissue culture within 14days after collection.

Adenovirus isolation was attempted in mono-layer tissue cultures of continuously passaged(CP) human heteroploid epithelial cells (KB orHEp-2 cells) and primary monkey kidney (MK)epithelial cells. Two CP and two MK tubeswere each inoculated with 0.2 ml of throat oranal swab specimen. Usually one tube of rhesusMK (Macaca mulatta) and one tube of vervetMK (Cercopithecus aethiops) were inoculatedconcurrently during the latter half of the study.There were periods, however, when only onekind of MK tissue was available. Rhesus MKwas used exclusively during the first 2 years.From October, 1957, to March, 1958, KB cellswere utilized, whereas from March, 1958, throughMay, 1961, HEp-2 cells were used. CP cultureswere either purchased or prepared in the labora-tory from KB or HEp-2 cells subcultivatedtwice weekly in bottles. Throat swabs weretested in MK and CP cultures during the entirestudy period. Anal swabs were also routinelytested in CP cultures only after August, 1959.

Tissue culture maintenance media were variedonly slightly during the study. During the first2 years, MK cell cultures were maintained withEagles' basal medium and 0.2% SV-5 antiserum.During the last 2 years of the study, equal partsof Eagles' basal medium and medium 199 with0.2% SV-5 antiserum were used. Maintenancemedium of the primary cell cultures was ex-changed every 5 to 7 days. Eagles' basal mediumwith 3% inactivated chick serum was used tomaintain CP cell cultures during the study period;this medium was exchanged every 3rd or 4th day.Antibiotics were present in all maintenancemedium in the following concentration per milli-liter: penicillin, 100 units; streptomycin, 100Mig; tetracycline, 10 Mg; and nystatin or ampho-tericin B, 25 Mg.Both MK and CF cells (KB and HEp-2)

were incubated for the first 3 years of the studyin stationary racks at 35 C for 2 to 3 weeks.During the last year, inoculated HEp-2 tissuecultures were placed on a drum, rotating 10rev/hr, for 7 days and were then transferredto stationary racks for the remaining 2 weeks.Tissue cultures inoculated with throat and analswab specimens were examined every 2nd or3rd day for cytopathic effect (CPE). During thefirst half of the study, when KB tissue culturewas used, each specimen was subpassaged rou-tinely. The subpassaged KB cultures wereobserved for 10 to 14 days. During the last halfof the study when HEp-2 tissue was used, cul-

S COMPLEMENT FIXATION (CR) SEROLOGIC (CFIO-WITH GROUP REACTIVE CONFIRETATION USINGL HUMAN SERUM PAIRED SERA FROMA INDIVIDUALS INFECTEDE WITH ADENOVIRUS

TEST FORHEMAGGLUT INAT ION

_HA)RHESUS MONKEY ERYTHROCYTE RAT ERYTHROCYTE POS IT IVE

POS IT IVENO HA

TISSUE CULTURE NEUTRALIZATIONTYPES: 12 AND 18

HEMAGGLUTINATION-INHIBITION OlATH HMAGGCUTINATION-INHIBITION WITHSPECIFIC RABBIT SERA, TYPES: SPECIFIC RABBIT SERA, TYPES: 1.2.3.7. 11 14 16.23.25,28 4 .5, 6 8 9.10. 13.15, 17, 19, 22. 23 24 26.

270 "R29'.'30''

FIG. 1. Laboratory procedures for identificationof adenoviruses.

tures were observed for 21 days but were notroutinely subpassaged.

IdentificationIsolates which produced a CPE consistent

with adenovirus and isolates which showed otherCPE but which could not be otherwise identifiedwere tested by complement fixation (Fig. 1).Preliminary identification was accomplishedwith a human serum pool containing adenovirusgroup-reactive CF antibody. Confirmatory identi-fication was accomplished by use of paired serafrom adenovirus-infected patients who developeda rise in group-reactive CF antibody duringconvalescence.Adenovirus isolates thus identified were then

placed in one of the four categories originallydescribed by Rosen; these categories are basedupon the hemagglutination pattern displayedwith rat or rhesus erythrocytes (5). Viruses ofgroup 1 agglutinate only rhesus erythrocytes andserotypes in this group include types 3, 7, 11,14, 16, 20, 21, 25, and 28. Agents in group 2produce complete agglutination of rat erythro-cytes as well as rhesus erythrocytes, but titersare lower with the latter cells. This group includesadenovirus types 8, 9, 10, 13, 15, 17, 22, 23, 24,26, 27, "29," and "30" (6). Viruses in group 3produce incomplete agglutination of rat erythro-cytes, but fail to agglutinate rhesus erythrocytes.Included in this group are types 1, 2, 4, 5, and 6.Viruses of group 4 do not agglutinate either rator rhesus red cells. Thus far, this group includesonly types 12 and 18. Agents falling into thiscategory were tested by tissue culture neutrali-zation with rabbit antisera for type 12 and type18 adenoviruses. All the other isolates weretyped by hemagglutination inhibition (HI) withthe use of type-specific rabbit sera for thoseadenoviruses which produce the type of hemag-glutination pattern observed. All dilutions forthe HI tests were performed with isotonic saline,

488 BACTERIOL. REV.

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VOL. 29, 1965 ADENOVIRUS INFECTION OF INFANTS AND CHILDREN

except for viruses in group 3. The hemaggluti-nation reaction of agents in this group is en-hanced when a heterotypic antiserum from group3 is added to the reaction mixture (4, 5). Sincetype 6 adenovirus was one of the least frequentlyencountered group 3 viruses in our studies, a1:100 dilution of type 6 rabbit antiserum wasused as a diluent for isolates of group 3. If anisolate in this group could not be identified astype 1, 2, 4, or 5, the test was repeated withanother heterotype group 3 antiserum for en-hancement.

Selection and preparation of erythrocytes. Theconcentration of erythrocytes used in hemad-sorption (HA) and HI tests was determinedvisually. This procedure differed from the originalmethod of Rosen, who used spectrophotometryfor this purpose (5). Erythrocytes were dilutedserially in saline to give successive concentrationsranging from 0.1 to 1.0%. A 0.4-ml amount ofsaline was placed in each tube, followed by 0.2ml of erythrocyte suspension. The mixture wasallowed to settle at 37 C for 1 to 2 hr. The highestdilution of erythrocytes that produced a com-pact button in the bottom of a tube was takenas the dilution to be used in the HA test. Thisusually corresponded to a 0.9% suspension forrat erythrocytes and a 0.6% suspension forrhesus erythrocytes.Monkeys which yielded red cells suitable for

adenovirus HA continued to yield suitableerythrocytes when such animals were testedover a 2-year interval. Blood was collected inAlsevers' solution (4 parts Alsevers' to 1 partblood) and stored at 4 C for a few days beforeit was washed three times in dextrose-gelatin-Veronal solution (DGV).Those erythrocytes which had a satisfactory

HA titer with type 3 adenovirus were also suit-able for all other serotypes belonging to group 1.Antigen was prepared by growing the prototypestrain of type 3 virus in CP cells; the inoculumwas diluted so that CPE was maximal notearlier than 3 days. A single pool of type 3 viruswas used throughout the study period to assessthe agglutinability of different lots of rhesuserythrocytes. Rhesus erythrocytes were con-sidered satisfactory if they could be aggluti-nated by type 3 virus diluted at least 1:32.However, antigen titration end points of 1:64to 1:128 were obtained in most tests. Of themonkeys tested, 30 to 40% yielded satisfactoryerythrocytes. Those erythrocytes which didnot agglutinate type 3 adenovirus were usedfor adsorption of typing sera before such serawere used in the HI test.Blood obtained by cardiac puncture from rats

(Sprague-Dawley strain) was collected in Al-

severs' solution (4 parts Alsevers' and 1 partblood) and kept at 4 C overnight. This materialwas then washed three times in DGV, andstored as a 10% suspension in DGV no longerthan 7 days. Because of the unusual sensitivityof rat red cells to washing compounds, best re-sults were achieved when disposable tubes wereused. Although rhesus red cells were not asprone to spontaneous agglutination as rat redcells, disposable tubes were nevertheless usedwith rhesus erythrocytes as well.Rat erythrocytes were satisfactory if a standard

type 4 virus preparation produced agglutinationat a dilution of 1:32 or greater. Rat erythrocyteswhich agglutinated satisfactorily with type 4virus were also suitable for use with all other ratcell reactive adenoviruses. Approximately 75to 80% of the rats tested were found acceptable.Rats yielding nonhemagglutinating erythrocyteswere exsanguinated, and the red cells were usedfor adsorption of rat cell agglutinins from typingsera.

Treatment of hyperimmune sera. Hyperimmunerabbit sera were prepared by use of prototypestrains of adenovirus grown in KB cells. Thesesera were then absorbed with kaolin and rator rhesus monkey red blood cells (5). Equalamounts of 25% acid-washed kaolin suspended insaline and heat-inactivated (56 C for 30 min)rabbit serum diluted 1:5 were mixed and incu-bated for 20 min at room temperature withoccasional agitation. The kaolin was sedimentedby centrifugation at 325 X g, after which theserum was decanted. To each 1 ml of dilutedserum, 0.1 ml of a 50% suspension of either rator rhesus red cells was added, and the mixturewas incubated for 1 hr at 4 C. The type of redcell used for adsorption was determined by thesubgroup in which the serum belonged. Erythro-cytes were removed by centrifugation and thetreated serum was tested for red-cell agglutina-tion. This adsorption procedure was repeatedif agglutinins were not removed by the firsttreatment.Each serum was titered against the homologous

antigen as well as against antigens within itsown erythrocyte-reactive group. The virus (0.2ml)-serum (0.2 ml) mixtures were shaken andwere then incubated for 1 hr at room tempera-ture before the addition of 0.2 ml of erythrocytesuspension. The erythrocytes were allowed tosediment for 2 to 3 hr at room temperature.The lowest dilution of each serum was tested forits ability to agglutinate the appropriate erythro-cytes in the absence of antigen. The highestdilution of serum which completely inhibitedrhesus or rat erythrocyte agglutination wasconsidered to contain 1 unit of antibody. Those

489

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VARGOSKO ET AL.

sera from different rabbits having similar ho-mologous HI titers when tested against prototypeantigen were pooled. After the titer of the pooledsera was rechecked, the serum was diluted tocontain 8 antibody units per 0.2 ml and thenwas stored at -20 C.HI test. Disposable flint-glass tubes with hem-

ispherical bottoms were used in all tests.A preliminary virus titration was performed

in which each isolate was diluted 1:8 to 1:512 intwofold steps. To 0.4 ml of each dilution, 0.2ml of the appropriate erythrocyte suspensionwas added. Erythrocytes were allowed to sedi-ment at room temperature. The end point of an

antigen titration was read as the last tube show-ing complete agglutination of rhesus or raterythrocytes, or, in the case of partially aggluti-nating types, the last tube showing approxi-mately 50% agglutination of rat red cells.

Duplicate tests were performed with eachserum and 8 HA units of virus. The virus-serummixtures were incubated for 1 hr at room tempera-ture. Then 0.2 ml of erythrocytes was added toeach tube. The test was read after it had been in-cubated for 2 to 3 hr at room temperature. Thetest was not considered satisfactory unless repli-cate HI readings with a given serum agreed. Asimultaneous antigen titration of each isolate wasincluded in the test. If the isolate could not beidentified and if more than 8 units of antigen were

used, the test was repeated.Adenoviruses that did not hemagglutinate

initially were passaged in CP cells until all or

almost all culture cells were destroyed after 3to 5 days. Then the isolate was retested forhemagglutination with rat and rhesus erythro-cytes.

Neutralization test. In the neutralization testfor type 12 and 18 adenoviruses, an equal volumeof undiluted infected tissue culture fluid anda 1:10 dilution of inactivated hyperimmunerabbit serum was incubated for 1 hr at room

temperature and was then inoculated into rhesusmonkey kidney, human embryonic kidney, or

HEp-2 tissue culture. The test was read whenthe virus control cultures exhibited 50 to 75%tissue destruction while CPE was not detectablein cultures containing either adenovirus 12 or18 serum.

Complement-Fixation Test

In the complement-fixation test, 4 to 8 unitsof adenovirus type 2 antigen were used, andthe technique was a modification of that de-scribed by Rowe (7). Type 2 was the adenovirusserotype most often recovered during the studyperiod.

TABLE 1. Serotypes of 506 adenovirus strainsrecovered from 7,509 infants and children*

Virus recoveryType

No. of children Per cent of isolates

1 125 252 166 333 73 144 2 0.45 56 116 6 17 25 59 7 110 5 112 13 316 6 118 4 0.822 1 0.223 1 0.225 1 0.226 2 0.427 2 0.429 3 0.630 2 0.4

Untyped 6 1.2

* A total of 506 virus strains were recoveredfrom 504 infants and children. Type 3 and type22 were recovered from one patient and type 5and an untyped strain from another.

TABLE 2. Comparison of recovery of adenoviruses from throat and anal swab specimens of3,230 infants and children (August, 1959-May, 1961)

No. of individuals from whom adenovirus recovered

Type Respiratory disease patients Control individuals

Throat only Throat and Anal only Throat only Throat and Anal onlyThroaonly anal naoy raoy anal n

1, 2, 3, 4, 5, 6, or 7 32 77 89 17 12 459, 10, 12, 16, 18, 22, 23, 25, 26, 0 0 34 0 0 16

27, 29, 30, or untyped

490 BACTERIOL. REV.

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VOL. 29, 1965 ADENOVIRUS INFECTION OF INFANTS AND CHILDREN

TABLE 3. Sensitivity of primary monkey kidneyand continuously passaged heteroploid humancell cultures for recovery of 411 adenoviruses(August, 1959-May, 1961)

Per cent of isolates recoveredin indicated culture system

Type No. of Hetero-isolates Monkey loiPkidney human Bothonly cell only(HEp-2

or KB)

1, 2, 3, 4, 5, 6, 361 0.8 68 31or 7

9, 10, 12, 16, 18, 50 2 86 1222, 23, 25, 26,27, 29, 30, oruntyped

Adeno Type 1

23 TS ISOLATESIn1..I AS ISOLATE S

75

S0 / ,' TS55 -,-

/ ---- AS

25

e I I I0 5 15 15 20

Day of Det

Adeno Type 2

22 rS ISOLATE$.1 I tInsl c

505 /

25 /

0 5 10 15 20

FIG. 2. Time required for detection of adenovirusin monkey kidney tissue culture (October, 1955-May, 1961).

RESULTS

Virus RecoveryOf the isolates, 89% were adenoviruses of

types 1 to 7, and most of these were type 1, 2,3, or 5 (Table 1). The remainder of the isolateswere adenoviruses of the higher numbered desig-nations, including three of proposed type 29 andtwo of proposed type 30 (6). In addition, sixstrains were recovered which have not yet beenidentified and may represent new serotypes.Thirteen of the adenoviruses recovered weretype 12, and four were type 18.

Adenoviruses were recovered approximatelytwo to three times more frequently from analswab specimens than from throat swab speci-mens, as indicated in Table 2. This was truenot only for patients with respiratory-tractillness but also for individuals in the controlgroup. Adenoviruses of type 9 and above wererecovered exclusively from anal swab speci-mens, although some patients with respiratory-tract illness yielded these viruses, and the throat

swab specimens from these patients were ob-tained in the early phase of the illness.The sensitivity of primary MK and CP

heteroploid human cell cultures for recovery ofadenoviruses is compared in Table 3 for thatperiod during which comparable testing wasperformed. The CP HEp-2 or KB cells werefar more sensitive than were MK tissue cultures.Lower numbered adenoviruses, i.e., type 1, 2, 3,4, 5, 6, or 7, were frequently recovered on bothtissues, but those of type 9 or above were muchless frequently recovered in MK tissue culture.In fact, most adenoviruses above type 18 wererecovered only in the heteroploid human cellcultures. Had we not used monkey kidney tissuecultures, we would have missed less than 1%of the adenoviruses of types 1 through 7 andonly 2% of the adenoviruses above type 7.

Extensive experience with certain adenovirusserotypes permits comparisons to be made con-cerning (i) the time required for detection ofviruses by use of different types of tissue culture,(ii) the time required for detection of virusfrom throat swab and anal swab specimens, and(iii) the time required for detection of CPE pro-duced by different serotypes. These data aredepicted in Fig. 2 to 5.

Several conclusions can be drawn from a re-view of these data.

(i) With most of the adenoviruses of lower nu-merical designation, 50% of the virus-positivethroat swab specimens produced a CPE inmonkey kidney tissue culture by the 6th or 7thday after inoculation, and 90% did so by the 11thor 12th day. Although continuous human cellcultures were more sensitive for detection of theseadenoviruses, the time required for detection ofCPE was longer than in MK. Thus, 50% of thevirus-positive throat swab specimens containinga lower numbered adenovirus produced a CPEby the 9th day in heteroploid cell cultures, and90% did so by the 15th day.

100

75z

E:

CL

8e

50

25

Adeno Type 3

-, T. 23ISC L TE _-_* 10

_TS 75

-

/ -AS

5 10 215-

e I I I o5 10 1 5 20

Adeno Type 5

TS ISOLATES e

/ASISO E ----AS

5 10 15 20Day of Detection

FIG. 3. Time required for detection of adenovirusin monkey kidney tissue culture (October 1957-May, 1961).

491

I

z

7E

1.

C-

_3 AS 1J _ I)0 L A T F S 0

0

0

1

* II

0o

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VARGOSKO ET AL.

Adeno Type 1

>100 82 TS ISOLATES. 90 59 AS ISOLATES ,X90_

ff 75 _OiE 740, 60- 45 _ I

_ 30 / T S* 15 _Ad ---AS

i 0 3 6 9 12 15 18

Adeno Type 2

7 5 - /0_60

45 - 0,

30 -/ 1 515 /., --- AS

0 3 6 9 12 15 18Day of Detect ion

Adeno Type 3

1 00 53 TS ISOLATESF32 AS ISOLATES,490 X75 - etS'60 - *45 I-/

30 I/ TS15 / , --- AS

s3 I I I I I0 3 6 9 1 2 1 5 1 8

FIG. 4. Time required for detection of adenovirus in continuous human cell cultures (KB or HEp-2),October, 1957-May, 1961.

Adeno Type 5

33 TS ISOLATES

100 32 AS ISOLATES 0

X 9075 _ 0UEee60 -

45 /-: 30 _ --S

115

0 3 6 9 12 15 18

Adeno Type 7

1 1 TS ISOLATES

loo 19 AS ISOLATES 0

90 0-0 ,

75_,60 _- -* -.

45 - / 1*'30 / 15T

15 J ' -- AS

0 3 6 9 12 15 18Day of Detprt ijrn

Adeno Type 12

0 TS ISOLATES

100 11 AS ISOLATES

90 _

60_w145 0I30 - I15 --- AS0 36 125l I1

0 3 6 9 1 2 1 5 18

FIG. 5. Time required for detection of adenovirus in continuous human cell culture (KB or HEp-2),October, 1957-May, 1961.

TABLE 4. CF antibody response by age of infantsand children from whom adenovirus wasrecovered (October, 1957-August, 1963)

CF antibody rise*Age individuals

No. Per cent

months

0-6 54 8 157-12 52 16 3113-24 64 12 1825-36 24 4 1737-48 14 0 _49-> 17 3 18

Total 225 43 19

* For cross-reactive antigen of type 2 virus.

(ii) Throat swab specimens contained slightlygreater amounts of virus than did anal swabspecimens, or possibly some other factor operatedto result in an earlier detection of pharyngealadenovirus in both MK and continuous humancell culture.

(iii) Adenoviruses of types 1, 2, 3, and 5 seemedto produce CPE at similar rates. Adenovirustype 7 throat swab isolates produced CPE more

rapidly in continuous cell cultures than the othertypes.

(iv) Those adenoviruses of a lower numericaldesignation were detected much sooner thanthose of types 9 and above. In general, the highertypes not only exhibited slower growth in tissueculture but were often confirmed only on sub-passage, a presumptive CPE having appeared atthe time of last reading.

CF Antibody ResponseTable 4 shows the CF antibody response by

age of infants and children from whom adeno-viruses were recovered during the period from1957 through 1963. Of 225 infants and children,19% developed a group-reactive antibody re-sponse to the type 2 CF antigen. The highest per-centage of responses occurred in infants betweenthe ages of 7 and 12 months.

In Table 5 the CF antibody response of pa-tients from whom an adenovirus was recoveredis analyzed to compare patients with or withoutillness, to show the response according to variousadenovirus types, and to indicate the responseaccording to whether adenovirus was recoveredfrom a throat swab specimen or anal swab speci-men, or both. CF antibody rises occurred almost

492 BACTE:RIOL. REV.

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VOL. 29, 1965 ADENOVIRUS INFECTION OF INFANTS AND CHILDREN

TABLE 5. CF antibody response of patients from whom adenovirus was recovered from throator anal swab specimen, or both, (October, 1957-August, 1963)

Adenovirus recovered from

Category Type Throat swab only Throat and anal swab Anal swab only

No. CF anti- No. CF antibody No. CF anti-body rise rise body rise

Respiratory 1, 2, 3, 5, or 7 41 39 43 37 69 9illness 9, 10, or 12 0 0 11 0

Other types* 0 1 (100) 7 14

Control 1, 2, 3, 5 or 7 8 0 5 0 29 39, 10, or 12 0 - 0 3 (33)Other types* 0 0 8 13

* Includes types 4, 6, 16, 18, 22, 23, 25, 27, 29, and untyped.

TABLE 6. CF antibody response of adenovirus-positive patients with or without priorCross-reactive antibody (October, 1957-August 1963)

Adenovirus CF Antibody in acute-phase serum

Category Type < 1:8 1:8 or >

No. of patients CF antibody No. of pa- CF anti-rise tients body rise

Respiratory illness 1, 2, 3, 5, or 7 91 36 62 89, 10, or 12 9 0 2 0Other types* 4 25 4 25Total 104 33 68 9

Control 1, 2, 3, 5, or 7 20 0 22 59, 10, or 12 2 (50) 1 0Other types* 3 (33) 5 0Total 25 8 28 3

* Includes types 4, 6, 16, 18, 22, 23, 25, 29, and untyped.

exclusively in infants and children from whomadenovirus 1, 2, 3, 5, or 7 was recovered and whowere ill. CF antibody responses were infrequentin infants and children in whom virus was re-covered only from the anal specimen.

In Table 6 the CF antibody response of pa-tients from whom an adenovirus was recoveredis analyzed according to the presence or absenceof group-reactive CF antibody in the acute-phasesera. A CF antibody response occurred in 36% ofill patients from whom adenovirus 1, 2, 3, 5, or7 was recovered and who did not have antibodyin their acute-phase serum. Only 8% of patientsfrom whom these agents were recovered but whohad prior group-reactive antibody showed a re-sponse. This difference was not seen in the fewpatients from whom other types of adenoviruseswere recovered. Antibody responses occurred in-

frequently in control patients; however, thenumber of subjects was too few to draw anyconclusions.An inhibition of human CF antibody was ob-

served when an excess of antigen was employedin the CF test. This phenomenon, however, wasnot responsible for the limited CF antibody re-sponse of adenovirus-positive infants and childrendescribed in this report. Chess-board titrationsinvolving both antigen and serum dilutions indi-cated that the antigen dilution used in ourroutine tests was in the zone of optimal reactivity.

Thus, it appears that several factors are corre-lated with the group-reactive CF antibody re-sponse of patients from whom adenovirus isrecovered. Factors which favor a CF antibodyresponse include an age of 7 to 12 months, theoccurrence of illness, adenovirus serotype (types

493

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VARGOSKO ET AL.

1, 2, 3, 5, or 7), the absence of prior group-reactiveantibody, and recovery of adenovirus from theoropharyngeal area rather than the lower gastro-intestinal tract.

DISCUSSION AND CONCLUSIONS

Although adenoviruses were among the earliestreported of the "newly uncovered" viral agentsproducing illness in man, the multiplicity of typesand the expense and tedium of neutralizationtests for identification of those viruses has limitedthe extent of epidemiological studies. Thus, therehas been a need for sensitive, simple, accurate,and rapid isolation and identification procedures.The procedures reported in this paper, modifiedslightly from those originally described by Rosen,were found to meet these requirements.The key to these procedures is that all but

types 12 and 18 of the 28 previously recognizedserotypes of human adenoviruses and apparentlycertain other newly defined serotypes hemagglu-tinate either rat or rhesus erythrocytes (5). Asdescribed in Materials and Methods, theseadenoviruses can be divided into three hemagglu-tinating subgroups and one nonhemagglutinatinggroup. The accuracy of this "differential hemag-glutination" screening procedure is indicated bythe fact that each of the adenovirus serotypesidentified by this procedure exhibited the hemag-glutination pattern characteristic of the proto-type strain of its serotype. Selection of rat orrhesus erythrocytes which are sensitive in thisprocedure is important, and failure to screenerythrocytes might account for the difficulty inusing the hemagglutination technique. Rhesuserythrocytes were used throughout our study,but grivet erythrocytes are also suitable (8).

In this study, a majority of the serotypes re-covered from children were those previouslythought to be prevalent in children, namely,types 1, 2, 3, and 5. Of all the serotypes, only11% were of the higher numerical designation andall but 1% (six isolates) could be identified.Probably the presently untyped adenoviruses areindeed new adenovirus serotypes.Although many of the adenovirus isolates in

this study were recovered in both MK and hetero-ploid human tissue culture cells, 68% of theadenovirus types 1, 2, 3, 4, 5, 6, and 7, and 85%of the so-called "higher" types of adenovirus,were recovered only in heteroploid human tissueculture cells. Only 0.8 to 2% of the isolates wouldhave been missed had heteroploid tissue culturealone been used. These findings confirm thosereported by Grayston et al. (1) and Pal et al. (3).Direct comparison of the HEp-2 and KB typesof heteroploid cells was not made, but the re-

covery rate of adenoviruses seemed to be aboutthe same in both of these tissue lines.

In most instances in which an adequate com-parison could be made, it appeared that CPE wasevident sooner in tissue cultures inoculated withthroat swab specimens than with anal swabspecimens. This suggests that the amount of viruspresent in throat swab specimens may have beenslightly greater, although it also is possible thata slight delay in production of cytopathogenicitycould have been occasioned by the presence ofsome viral inhibitor in anal swab material. None-theless, in all types of subjects and with all sero-types, more adenoviruses were recovered fromanal swab specimens than from throat swabspecimens. During the period of this study, noadenoviruses of the higher numerical designationwere recovered from throat swab specimens.(Three were recovered during a subsequentperiod.) More than likely, these findings resultfrom the fact that these adenoviruses are presentin the nasopharynx or oropharynx for a briefperiod early during human infection and thatthey then persist for a longer interval either ingastrointestinal mucosa or lymphoid tissue. Alongitudinal study in which specimens are takenfrom the same subjects over a period of timewould help to clarify some of these questions.Only 19% of 225 individuals from whom an

adenovirus was recovered showed a rise in CFantibody when type 2 adenovirus was used as across-reactive antigen. This fact should be con-sidered in interpreting clinical or epidemiologicalstudies based on complement-fixation tests. Thefact that infants in the age group of 7 to 12months showed the highest percentage of cross-reactive CF antibody rises probably reflects thefact that this is the period during which there isthe highest incidence of infection with adeno-viruses (2). This is the time of life when primaryinfection is most likely to occur, and, therefore,virus isolation is most likely to be associated withacute infection. Virus recovery from older infantsand children is more likely to be associated withlong-term or intermittent shedding of virus as aresult of chronic or subacute infection. Group re-active CF antibody responses would be detectedless frequently during such infection than duringacute infection. The apparent influence of otherfactors, such as the greater likelihood of a CFantibody rise if the adenovirus was recoveredfrom the throat rather than the feces and if theadenovirus was of a lower numerical designation,is also probably related to the likelihood that thesubjects concerned were undergoing acute infec-tion. The apparent limitation of antibody re-sponses in subjects whose acute serum containedsome cross-reactive antibody might result from

494 BACTERIOL. REV.

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

VOL. 29, 1965 ADENOVIRUS INFECTION OF INFANTS AND CHILDREN

several factors, the most likely of which is thatCF antibody present as a result of infection witha different serotype in the recent past would maska rise in antibody due to the presently infectingserotype.

SUMMARYIn summary, the development of differential

erythrocyte hemagglutination and the HI test asa standard laboratory tool for the typing ofadenoviruses has been most effective in further-ing the ability of a virus laboratory to identifyadenoviruses simply and rapidly. Once an isolatewas placed into a rhesus or rat erythrocytehemagglutinating group, typing was readily ac-complished. Adenovirus types 1 to 7 were re-covered from both anal swab and throat swabspecimens. They were detected more often froman anal swab than from a throat swab specimenbut they were detected sooner from throat swabspecimens. Adenoviruses of the higher numericaldesignations were recovered only from anal swabspecimens and in general were detected withgreater difficulty and later than the lower adeno-virus types. CP human cell tissue cultures weremore sensitive for adenovirus recovery than pri-mary MK tissue culture. A limited number ofinfants and children from whom adenovirus wasrecovered displayed a group-reactive CF antibodyresponse, and several factors were noted to in-crease the likelihood of such a response. Thehighest percentage of responses occurred in infantsin the second half year of life, in patients fromwhom types 1, 2, 3, 4, or 7 adenoviruses hadbeen recovered, and in those from whom theadenoviruses were recovered from throat swabspecimens.

ACKNOWLEDGMENTSThis is to acknowledge gratefully the assistance

of Leon Rosen in rechecking the typing of a ma-

jority of the adenoviruses of higher numericaldesignation reported in this paper.

This investigation was supported in part byPublic Health Service research grant AI 01528from the National Institute of Allergy and In-fectious Diseases.

LITERATURE CITED

1. GRAYSTON, J. T., J. C. LASHOF, C. G. LooSLI,AND P. B. JOHNSTON. 1958. Adenoviruses.III. Their etiological role in acute respira-tory diseases in civilian adults. J. Infect.Diseases 103:93-101.

2. KIM, H. W., A. J. VARGOSKO, B. JEFFRIES,R. M. CHANOCK, AND R. H. PARROTT. Adeno-virus infection in childhood-contributionto clinical respiratory tract illness.

3. PAL, S. R., J. MCQUILLIN, AND P. S. GARDNER.1963. A comparative study of susceptibilityof primary monkey kidney cells, HEp-2cells and HeLa cells to a variety of fecalviruses. J. Hyg. 61:493-497.

4. PEREIRA, H. G., AND M. V. T. DEFIGUEIREDO.1962. Mechanism of hemagglutination byadenovirus types 1, 2, 4, 5 and 6. Virology18:1-8.

5. ROSEN, L. 1960. Typing adenoviruses by hem-agglutination inhibition. Am. J. Hyg. 71:120-128.

6. ROSEN, L., J. F. HoviS, AND J. A. BELL. 1962.Further observations on typing adenovirusesand a description of two possible additionalserotypes. Proc. Soc. Exptl. Biol. Med.110:710-713.

7. ROWE, W. P., R. J. HUEBNER, J. W. HARTLEY,T. G. WARD, AND R. H. PARROTT. 1955.Studies of adenoidal-pharyngeal-conjuncti-val (APC) group of viruses. Am. J. Hyg.61 :197-218.

8. ZUSCHEK, F. 1961. Studies of the hemaggluti-nins of type 3, 4, and 7 adenovirus. Proc.Soc. Exptl. Biol. Med. 107:27-30.

495

on May 14, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from