RESPONSE OF CELL AND ORGANISM TO INFECTION WITHAVIAN TUMOR VIRUSES'

HARRY RUBINDepartment of Virology and Virus Laboratory, University of California, Berkeley, California

I. Introduction ............................................................................... 1II. Rous Sarcoma Virus Infection at the Cellular Level .. 2III. Resistance-Inducing Factor and Lymphomatosis Virus Infection ... 4

A. Congenital Transmission of RIF ........................................................ 8B. The Fate of Congenitally Infected and Uninfected Chicks.10C. Experimental Establishment of Tolerance.11

IV. Discussion ............................................................................... 12V. Literature Cited ..13

I. INTRODUCTION

Fortunately, there is no need for a lengthypreamble to justify the present discussion nor theresearch which occasions it. The significance oftumor virus research has become evident to allbiologists as a result of the isolation of a numberof agents from mouse neoplasms during thepast decade. This recognition is somewhat belatedas tumor viruses have been known since 1908when Ellermann and Bang (6) isolated the agentresponsible for erythroblastosis in chickens. Therewas at first considerable reluctance to admit theimportance of this finding for cancer research,partly because the neoplastic nature of leukemiacells was questioned. This question became irrele-vant a few years later after the isolation of theRous sarcoma virus (RSV) from a naturallyoccurring connective tissue tumor in a hen (12).In subsequent years, viruses were isolated from avariety of chicken sarcomas and leukemias. Allthe viruses isolated from chicken tumors up tothe present appear to be closely related to oneanother as indicated by similarities in size,morphology, chemical constitution, and anti-genicity, and it is convenient to classify themtogether as agents of the avian leukosis complex(2). Agents of this complex cover a wide spectrumof virulence; RSV has been so effectively adaptedfor rapid growth in the laboratory that it in-variably produces highly malignant tumors withina few days, whereas visceral lymphomatosis

' Text of the Eli Lilly and Company ResearchAward Address in Bacteriology and Immunologypresented at the Annual Meeting of the AmericanSociety for Microbiology in Chicago, Ill., on April25, 1961. Based on research carried out under theU. S. Public Health Service grant, C-4-774.

virus (VLV) does not induce malignancy formany months and even then with irregularity.Other leukosis viruses such as the myeloblastosisand erythroblastosis viruses have intermediatedegrees of virulence.Under natural conditions VLV is by far

the most widely occurring of the leukosisviruses as indicated by the fact that there isnot a flock of chickens in the United States knownto be free of the associated disease conditioncalled lymphomatosis. In many respects thepathogenesis of the lymphomatosis in chickensresembles the pathogenesis of lymphocytic leu-kemias in higher organisms. Therefore, if onewere to choose an agent as a model to understandthe general patterns of viral carcinogenesis, VLVwould be the natural choice. Unfortunately,until recently the assay system for this virus hasbeen too cumbersome to permit integrated ex-perimental investigation of its behavior.

In lieu of an efficient assay system for VLV,it has been necessary for virologists to turn toother systems such as RSV as possible modelsfor investigation of the detailed interactionsbetween tumor viruses and cells. In the case ofRSV an efficient and precise assay system invitro was developed (20) and rapid progress madein understanding the interactions between virusand cell which lead to malignancy. During thecourse of the work with RSV, an assay was in-advertently discovered for VLV. Although thisassay does not have all the desirable features ofthe assay for RSV, it has proved to be a usefultool for studying certain aspects of tumor virusinfection which cannot be investigated with RSV,namely, the roles played by congenital trans-mission and immunological tolerance in per-

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

HARRYRUBI[V.N-

petuating virus in nature and causing disease.Therefore, this paper will be divided into tworelatively independent sections, one concernedwith RSV infection at the level of the cell andthe other concerned with VLV infection at thelevel of the organism.

II. Rous SARCOMA VIRUS INFECTION AT THE

CELLULAR LEVEL

In describing RSV infection at the level of thecell, I shall confine myself to recent kinetic andcytological studies carried out in vitro (14, 21, 23).When a high concentration of RSV is added tochick embryo cells growing in tissue culture thecells undergo a series of characteristic morpho-logical alterations. At about 2 days after infec-tion, the fibroblasts, which are normally fusiformin shape, become plumper and their refractilityincreases. Within the next day or two, the cellsbecome more rounded and escape from contactinhibition. Escape from the contact inhibitionwhich restricts normal fibroblasts to monolayergrowth permits the Rous sarcoma cells to movefreely over one another and over normal cells.As a result, Rous sarcoma cells are frequentlyfound in several layers. If only a small numberof RSV infectious units are added to a chickembryo culture, colonies of Rous sarcoma cells(foci) become visible against a background ofnormal cells (Fig. 1 and 2) in 5 to 6 days. Roussarcoma cells can readily be distinguished fromnormal chick embryo fibroblasts because of therounded morphology of individual cells and themultilayered growth of the colony.The morphological changes in tissue culture

occur at the same time and are of the same typeas those which occur in vivo (10). The growthof virus also follows the same pattern in vitroand in vivo (4, 11, 13). Therefore, it is likelythat the cellular events studied in tissue cultureare an accurate reflection of the events leading tomalignancy in the animal.The number of foci produced in a culture varies

linearlv with the concentration of RSV added(20). Up to 1,000 such foci can be counted on asingle 50-mm petri dish. Focus formation there-fore serves as an excellent assay for virus infec-tivity.

Experiments on the growth of Rous sarcomavirus in vitro revealed the following features(Fig. 3). There is an eclipse period of about 12hrs after infection during which very little virus

..1.~~~~~~~~~~.

FIG. 1. Low power view of a typical RSV focusat 7 days on a culture of chick embryo cells. Thisfigure first appeared in Virology 12:14-31, 1960. Seereference (14).

can be recovered from the infected cells. There-after, progeny virus particles appear and theirnumber increases exponentially until about 3 daysafter infection. At this time the cells reach aconstant rate of virus production of about 1infectious unit of virus per cell per hr (23).Since the ratio of virus particles to infectiousunits has recently been estimated by Crawford(5) as about 1,000, it would appear that an in-fected cell releases about 1,000 virus particles perhr. The volume of the virus being about 10-6 thatof the cell, the cell must produce about hoo0its own mass in virus per hr. The growth rate ofthe cells is apparently unaffected by this demandon its synthetic activity.A striking aspect of the growth of RSV is its

rapid and continuous release from the infectedcell. This was first suggested by the observationthat virus can be detected in the medium beforeit can be detected in association with washedcells when samples are taken at time intervalsas short as 4 hr, and that the level of virus in themedium usually exceeds that in the cells by afactor of about 10 during the first few days ofvirus production (21, 23). This relationshipbetween cell-associated virus and free virus canonly occur if there is a rapid and continuous re-lease from the cell of mature virus particles asthey are completed (17).

2 [VOL. 26

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

INFECTION WITH AVIAN TUMOR VIRUSES

71.

-u~~~~~~~~~--;~~~~~~~~~i'.t0e0

.4~~ ~--

var

FIG. 2. Higher magnification of RSV focus in Fig. 1

The rapid and continuous release of virus fromcells is a common feature of the multiplicationof large RNA viruses, a category which includesinfluenza, mumps, and Newcastle disease virusin addition to the avian tumor viruses (16). Avariety of findings have accrued over the yearsto indicate that there is an association betweenrapid, continuous release of a virus and itsmaturation at the cell membrane. Further supportfor the completion of RSV at the cell surfacecame from the finding that over 90% of the virusassociated with washed, intact cells was accessibleto inactivation by antiserum (23). Since theantibody could not penetrate the intact cell, itmay be assumed that almost all the cell-associatedvirus is superficial, at least during the early stagesof virus growth.

Since the surface area of the virus is about 10-4that of the cell, the release of 1,000 particles perhr is equivalent to the loss of H4o the surfacearea of the cell every hour. It is not unreasonable,therefore, to assume that virus multiplicationmay interfere with the function of the cell mem-brane. This will be discussed below at greaterlength.A major deficiency of the kinetic experiments

on infected cells is that they give no hint of theamount of viral protein which is not in maturevirus particles. Thus, the calculated ratio of

virus surface to cell surface is likely to be under-estimated. The kinetic experiments also lack theimpact of direct visualization of virus growth. Itwas with these deficiencies in mind that ananalysis was undertaken of RSV growth andlocalization with the fluorescent antibody tech-nique (23). The general plan of this work was toinfect chick embryo cultures with RSV and stainindividual cultures with fluorescent antibody atdaily intervals. The resulting observations aredescribed below.The first appearance of viral antigen can be

detected along the borders of infected cells at 2days after infection (Fig. 4). As previously notedit is at this time that the first morphologicalsigns of virus infection become apparent, and itis also at this time that the production of virusin a significant fraction of infected cells can bedetected. Proof that the viral antigen is indeedat the cell surface arises from the fact that livingcells can be stained with fluorescent antibodyjust as effectively as fixed cells (Fig. 5).At 3 to 4 days, a marked change in cell be-

havior occurs. The cells become more roundedin outline and escape contact inhibition. Theescape from contact inhibition permits the alteredcells to move over one another and over normalcells. In this respect they assume the charac-teristic behavior described by Abercrombie,

1962] 3

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

HARRY RUBIN

106

UNITSPER

PLATE

j03[

/ 4\F/FREE VIRUS

Ia - - 0 TOTAL CELLS

.1\

I,1 I

\! INFECTIVE

~~~,CENTERS

~~~~~C.A.V.

6

0

01234560 2 3 4 5 6DAYS AFTER INFECTION

FIG. 3. RSV multiplication in cultures of chickembryo cells. This figure first appeared in Virology12:1431, 1960. See reference (14). C.A.V. = Cell-associated virus.

Heaysman, and Karthauser (1) for sarcoma cells.When cells of this stage are stained with fluores-cent antibody, it is found that large amounts ofviral antigen are being shed from the cell surfaceinto a matrix-like substance around the cell(Fig. 6). It seems likely that the contemporaneousoccurrence of virus shedding at the cell membraneand the loss of contact inhibition is more thancoincidental since the cell membrane is believedto be the site at which contact inhibition ismediated (1).

Following the shedding stage, the synthesis ofviral protein can be detected both in the cyto-plasm and at the cell surface (Fig. 7). Since theappearance of cytoplasmic antigen occurs afterthe cell has escaped contact inhibition, it isunlikely that production of virus within thecytoplasm plays a central role in this primarychange of cell behavior though it might play somerole in perpetuating the malignant behavior of

the cell. For the present then, it seems mostlikely that the earliest manifestation of neoplasticbehavior in a cell infected with RSV is the resultof virus-induced alterations at the cell surface.

III. RESISTANCE-INDUCING FACTOR ANDLYMPHOMATOSIS VIRUS INFECTION

During the course of the RSV work, it wasfound that the cells obtained for tissue culturefrom certain embryos were highly resistant toRSV infection. Although the cultures and theembryos from which they were obtained appearednormal, a virus was isolated from the embryoswhich could induce resistance to RSV when addedto sensitive cultures (15). The virus was namedRIF, an acronym for "resistance-inducingfactor," but in its physical, chemical, and biologi-cal characteristics RIF proved to be indistinguish-able from VLV (7). It was also found that estab-lished strains of VLV could be detected in vitroby interference with RSV with precisely thesame technique used to assay RIF. These factsplus epidemiological observations to be discussedbelow indicate that RIF is a strain of lympho-matosis virus (3).

It was found that about 1 in 40 embryos froman ordinary flock of chickens was congenitallyinfected with RIF. However, the frequencyof infected embryos from an experimental flockwhich had been selected for a high incidence oflymphomatosis was about 1 in 4 (18). The lym-phomatosis-susceptible flock was made availableto us (courtesy of Kimber Farms, Niles, Calif.)for a study of the congenital transmission of RIF.The titer of RIF in an unknown preparation

was determined by infecting RSV-sensitive cul-tures with serial dilutions of the sample andchallenging aliquots of cells from the cultureswith RSV at each of three or four successive celltransfers. A high concentration of RIF inducedresistance to RSV at the first transfer, whereaslower concentrations induced resistance at sub-sequent transfers. By reference to standardcurves for a preparation of known infectivity,the titer of the unknown could be determined.Antibody to RIF could be determined by its

ability to eliminate the RSV-inhibitory activityof RIF. However, a large-scale study of thedistribution of virus and antibody was underconsideration, and the RIF neutralization was toocumbersome to be used on such a scale. It couldbe shown that the level of neutralizing activityof a serum against RSV was a good indicator ofthe level of its activity against RIF. Since the

4 [VOL. 26

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

INFECTION WITH AVIAN TUMOR VIRUSES

4. ... W . z W:iX. *We , * .s

.s. B It's:

.wi.

*. ..s ..

}.:'!4

FIG. 4a (top). Chick fibroblast culture, fixed and stained with fluorescent RSV antiserum on the 3rd dayafter infection. Viral antigen (arrow) appears at the cell membrane. Scale: 2.9 cm = 100 1A.

FIG. 4b (bottom). Same preparation under phase contrast. This figure first appeared in Virology 13:528-544, 1961. See reference (23). Scale: 2.9 cm = 100 1.

51962]

Hi.

VB......

OF

... .......

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

HARRY RUBIN

FIG. 5a (top), 5b (bottom). Unfixed chick fibroblast cells stained with fluorescent antibody on the 3rd (a)and 5th (b) day after infection, respectively, showing superficial localization of viral antigen. This figurefirst appeared in Virology 13: 528-544, 1931. See reference (23). Scale: (a) 5.3 cm = 100 IA; (b) 2.3 cm =

100 p.s.

6 [VOL. 26

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

INFECTION WITH AVIAN TUMOR VIRUSES

FIG. 6a (top), 6b (bottom). A focus of Rous sarcoma cells stained with fluorescent antibody and seen influorescent (a) and phase contrast (b) microscopy. Arrow points to same cell in both photographs. Viralantigen is being shed fro n cell surface. Scfile: 3 cm = 100 u.

719621

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

HARRY RUBIN

... ..|l

FIG. 7. Chick fibroblasts 6 days after infection with RSV showing cytoplasmic fluorescence. This figurefirst appeared in Virology 13: 528-544, 1961. See reference (23). Scale: 2.3 cm = 100 IL.

neutralization of RSV could be carried out withefficiency and precision, it was substituted for theRIF neutralization to assay antibodies to RIFin the population. Subsequent tests of selectedsera for ability to neutralize RIF showed thatthe RSV neutralization gave an accurate pictureof the distribution of RIF-neutralizing antibody.The plan of the study mentioned above was to

determine the status of the parental birds withregard to viremia, antibody, and ability totransmit RIF congenitally. The parental birdswere bled repeatedly, and embryos of knownparentage were obtained by trap-nesting hens.The embryos were used to prepare cultures andthe cultures were challenged with RSV to deter-mine whether they were infected with RIF.The titers of RIF and antibody in the blood

of each parental bird were determined onceduring the egg-laying period when the parentswere 12 months old and at three subsequentintervals over a period of 10 months. Discussionof the results is simplified by the fact that therewas little change in the occurrence and titers ofvirus and antibody in individual birds during

this period. The adults could be divided intotwo classes consisting of viremic birds and non-viremic birds (Table 1). Four of the 18 hens inthe initial study and 3 of the 8 roosters had hightiters of virus present in the blood throughoutthe 10-month period of study. None of the viremicbirds had antibodies to RIF and only 1 of the 8had antibodies to RSV. It is evident that per-sistent viremia and neutralizing antibodies areto a large extent mutually exclusive, suggestingthat the viremic birds were immunologicallytolerant to the virus.

A. Congenital Transmission of RIFThere was a marked distinction between

viremic and nonviremic birds in ability to trans-mit RIF to progeny. All the fertile viremic femaleswere persistent congenital transmitters of thevirus (Fig. 8). Most of the cultures from theseembryos were highly resistant to RSV whenchallenged immediately after explantation of thecells, indicating that a high proportion of cellshad been actively producing virus in ovo. Onlya few showed the delayed resistance which is

8 [VOL. 26

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

INFECTION WITH AVIAN TUMOR VIRUSES

TABLE l.a RIF-viremia and antibodyin parental birds

Viremic birdsb Nonviiemic birdsc

Antibody Antibody Antibody Antibodyto RSVd to RIFe to RSVd to RIF'

Hen Henno.: no.:1 - - 2 - +

3 - - 4 + +7 - - 5 + +9 _ 8 - +

Roosterno.:R1 + - 10 + +

11 + +R2 - - 12 + +R9 - - 13 + +

14 - +16 - +17 + +18 + +19 + +20 + +

Roos-terno.:R3 + +R4 + +R6 + +R8 + +R1O + +

a This table first appeared in Proc. Natl. Acad.Sci. U. S. 47:1058-1060, 1961. See reference (18).

bViremic birds = serum-induced resistance toRSV in the first transfer when obtained at 12, 14,17, and 22 months of age and added to RSV-sensi-tive cultures.

c Nonviremic birds = serum failed to induceresistance in cultures challenged with RSV inthree successive transfers.

d Antibody to RSV. (+) = A 1:10 dilution ofserum-reduced RSV titer > 10-fold in 40 min ofincubation at 37 C. (-) = A 1:10 dilution of se-rum-reduced RSV titers > 2-fold in 40 min of in-cubation at 37 C.

e Antibody to RIF. (+) = A 1:10 dilution ofserum eliminated the RSV-inhibitory effect of a1:10 dilution of RIF in cultures challenged withRSV after 1 transfer. (-) = A 1:10 dilution ofserum failed to eliminate the RSV-inhibitory ef-fect of RIF.

characteristic of those embryos in which only asmall proportion of cells is infected at the time ofexplantation.

Viremic Hens (No antibody)

I-

3 11EJ)

7 no eggs

9-

heavy infectiona light infection

uninfe ted

10 20Embryo number in

Non-Viremic HensI

2141 15 181

12131 l--14 1

11 + 16 no eggs17181119M2011

I10

sequence of laying20

FIG. 8. Congenital transmission of RIF. Thechart shows the embryos from viremic and non-

viremic hens in the order that the egg was laid. Thisfigure first appeared in Proc. Natl. Acad. Sci. U. S.47:1058-1060, 1961. See reference (18).

Among the nonviremic females, only one was a

persistent congenital transmitter. This one is ofspecial interest, however, since its existence showsthat virus multiplication can continue indefinitelyin an animal despite the continuous presence ofantibody. This finding has more recently beensubstantiated in studies on a larger number ofbirds of this flock in which about 1 out of every 7nonviremic hens was found to be a persistentcongenital transmitter despite the presence ofantibody. This finding brings up the questionwhether RIF continues to multiply in nonovariantissue in the remaining nonviremic birds, i.e.,those which failed to congenitally transmit thevirus. If so, the persisting antigenic stimuluswould provide a simple explanation for the con-

stancy of antibody titer over many months.There was no suggestion of congenital trans-

mission by viremic males (Table 2). Of the 4nonviremic females mated to viremic males, 3produced 37 progeny, all of which were unin-fected. The remaining hen, no. 19, which pro-

duced infected progeny, continued to do so

when mated to a nonviremic male. These results,which have been substantiated in larger numbersof birds, indicate that congenital transmission wasunder strict maternal control. Similar resultshave been reported for mouse leukemia virus (8).One possible explanation for the failure of male

transmission is the likelihood that any maturevirus which might be carried in the sperm fluidwould be inactivated upon contact with antibodyin the female. When females with low antibody

1962] 9

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

HARRY RUBIN

TABLE 2.a Failure of congenital transmissionof RIF by viremic roosters

Viremic roosters Nonviremic roosters

Infected InfectedRooster X Hen progeny/ Rooster X Hen progeny/

no. no. total no. no. totalprogeny progeny

Ri X 18 0/15 R3 X 14 0/20R2 X 17 0/7 R4 X 13 0/16R9 X 19b 7/7 R6 X 4 0/14R9 X 20 0/17 R6 X 5 0/15

R8 X 1c 16/17Total... 7/46 R8 X 2 0/20

R8 X 8 0/8R8 X 9c 14/15R8 X 10 0/12R8 X 12 1/11RIO X 3c 5/9

Total.. 36/157

a This table first appeared in Proc. Natl. Acad.Sci. U. S. 47:1058-1060, 1961. See reference (18).

b Continued to produce infected progeny whenmated with nonviremic rooster.

c Viremic hens.

titers were mated to viremic males, however,infected progeny were not produced (18). Un-fortunately, antibody-free females were notavailable, and an unequivocal result could not beobtained. The most revealing aspect of the failureof male transmission is its implication for thelocalization of the viral genome in the cell and inthis respect it seems unlikely that antibody wouldaffect the transmission of the viral genome par-ticularly if carried within spermatozoa.Another possible explanation for the failure of

male transmission is that the testicular cells ofthe viremic male had somehow escaped infection.This possibility could be explored by trypsinizing,washing, and suspending the testicular cellsand plating them as infective centers. When thisprocedure was carried out, it was found that ahigh proportion of cells obtained from the testesof viremic males were actively p)roducing virus(18).The cells obtained by this technique presuma-

bly were the larger and less differentiated ele-ments of the testes since spermatozoa were notseen. The failure of the male transmission,therefore, suggests that the viral genome is lostfrom the cell during the process of spermiogene-

sis, when RNA and cytoplasm are shed fromspermatocytes to form spermatozoa. This experi-ment, then, indicates that, unlike the temperatebacteriophages, the genome of an RNA virusis not closely associated with chromosomes of thecell.

B. The Fate of Congenitally Infected and U.nin-fected Chicks

As noted above, the high resistance to RSVof cells from congenitally infected embryos im-mediately after explantationsuggested that a highproportion of cells in the embryo was infected.This seemed remarkable in view of the fact thatthe infected embryos appeared to be normal inevery way and indeed usually hatched and ma-tured into normal adult chickens. To seek con-firmation of the impression that a high propor-tion of cells from the embryo was infected, thecells were plated as infective centers. It was foundthat the addition of as few as 2 cells from a con-genitally infected embryo could ultimately induceresistance to RSV in a culture of 106 sensitivecells (18). This finding supported the impressionthat a high proportion, and perhaps all, of thecells from congenitally infected embryos not onlywere infected but were continuously producingvirus.Having established a few parameters of con-

genital infection through study of parents andembryos, it was decided to extend the investiga-tion, and study the fate of the embryos afterhatching. The following questions were kept inmind during this portion of the study: (i) Didcongenital infection establish immunologicaltolerance to the virus and did the tolerant chicksbecome the persistently viremic adults? (ii) Didthe uninfected chicks become infected bv con-tact? If so, when did they become infected andwhen were antibodies made? (iii) Which class ofanimal was more likely to develop leukosis, thecongenitally infected or the contact infected?To get significant numbers of lymphomatosis

cases among the progeny, it was necessary toincrease the population under study to a finaltotal of 63 fertile females, 10 males, and about800 progeny. The progeny were bled in staggeredgroups at various times after birth. An attemptwas made to obtain at least two consecutiveblood samples at an interval of several monthsfrom each of the progeny birds; as many asfour samples each were obtained from about

10 [VOLJ. 26

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

INFECTION WITH AVIAN TUMOR VIRUSES

100 of the birds. Each serum was tested forvirus and antibody. In analyzing the results, theprogeny were divided into two classes consistingof those coming from hens which were regularcongenital transmitters (RIF(+) families) andthose coming from hens which were not congeni-tal transmitters (RIF(-) families).The results on the persistence of viremia in

congenitally infected birds can be described in asentence: all those birds which were congenitallyinfected had a high level of viremia over theentire 7-month period in which progeny serawere tested. Of the noninfected chicks in contactwith the congenitally infected birds, a few beganto show a slight viremia during the early weeks oflife, but involvement of a significant fraction ofthe population with viremia did not begin until9 weeks after hatching.The concentration of virus in the blood of all

but a few of the contact infections remainedseveral orders of magnitude lower than that of thecongenitally infected birds. The number of birdswith detectable viremia reached a maximum at14 weeks and thereafter decreased. The probablereason for this decrease became evident when theantibody status of these birds was determined,and is discussed below.None of the individuals known to be congeni-

tally infected developed antibodies. In theRIF(+) families only those few siblings whichhad escaped congenital infection developed anti-bodies. The pattern of antibody development inthe latter group was the same as that encounteredin the progeny of hens which were not congenitaltransmitters (RIF(-) families).

In the RIF(-) families, passively transferredantibodies were found both in the yolk of theembryo and in the sera of 2-week-old birds inabout i o the concentration present in theirparents. Passively transferred antibody couldno longer be detected when the chicks reached theage of 4 weeks, but actively produced antibodybegan to appear in some birds at 9 weeks of age.The proportion of birds with antibody then in-creased, with a particularly sharp increase occur-ring between 14 and 18 weeks. It will be recalledthat it was at 14 weeks that a downturn occurredin the number of contact birds with viremia. Itseems likely that this was related to the increaseboth in the proportion of birds with antibody andin the concentration of antibody in individualbirds.

Perhaps the most striking aspect of theseinvestigations was the clear indication that im-munological tolerance to RIF was establishedby virtue of congenital transmission. In thisrespect the RIF system resembles congenitalinfection of mice with lymphocytic choriomenin-gitis virus (LCM). In the case of LCM it hasbeen established that the congenitally infectedmouse has a persistent viremia with no antibodyproduction, thereby implying immunologicaltolerance (9, 22). As adults, tolerantly infectedmice are unaffected by an intracerebral inocula-tion of LCM which kills previously uninfectedmice. The survival of tolerant mice suggestedthat the disease produced by intracerebralinoculation of previously uninfected mice was theresult of the immunological response of the hostrather than the cytopathic interaction betweenvirus and cell. It thus became a matter of greatinterest to determine whether lymphomatosis inchickens might also be the result of the immuno-logical response of the organism to virus infection.If lymphomatosis had such an immunologicalbasis, it seemed likely that the probability for de-veloping the disease would be considerably higherin the contact birds than in the congenitally in-fected, immunologically tolerant birds.

In actual fact, the reverse of these expectationswas found. The probability for developing viscerallymphomatosis proved to be six times higherin the congenitally infected birds than in thecontact-infected birds. The level of the viremiain the congenitally infected birds did not dimin-ish shortly before their death from visceral lym-phomatosis, as might be expected if the diseasewere associated with an immunological responseto infection. Therefore, the immunological hy-pothesis, at least in its simplest form, is notsupported by the results. There remains thepossibility that the disease is the result of theexcessive proliferation of infected cells stimulatedby antigens of various kinds other than thosedirectly associated with RIF infection. If suchwere the case, the virus would have to be con-sidered as a conditioning factor which merelyincreases the probability that a cell will becomemalignant under certain stimuli.

C. Experimental Establishment of Tolerance

The experiment of nature described above hasbeen most fruitful in providing us with a pictureof the natural history of a ubiquitous tumor

1962] 11

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

HARRY RUBIN

TABLE 3. Role of early viremia in establishingtolerant infection to RIF

Viremic history during first3 weeks after hatchine Fraction developing

tolerant infection1st week 3rd week

3+ 3+ 10/102+ 3+ 6/7- 3+ 3/7_ 2+ 0/6

b 0/14

a 3+ = Heavy viremia (> 101 infective units ofRIF per ml plasma), 2+ = moderate viremia(102 to 105 infective units), and- = no viremia(<102 infective units).

b Viremia after 3 weeks.

virus, but it has also been useful in provoking anumber of questions which can be answered onlyby direct experimentation. A question which hasoccupied much of our attention recently concernsthe conditions which must be satisfied for thesuccessful establishment of immunological toler-ance. As a prelude to this work, it was necessaryto carry out a growth curve of RIF in vitro (M.Feldman, personal communication). RIF is veryslow to reach a constant level of virus production,requiring some 9 days. By way of contrast, RSVrequires about 3 days (21, 23) and a cytocidalvirus such as Newcastle disease virus (NDV)requires less than Hi day (19). Precise compari-sons of the final rates of virus production cannotbe made as yet since the experiments with RIFhave not been carried out with the same pre-cision as those with RSV and NDV. It is safe tosay, however, that the ratio of free virus tocell-associated virus has been found to be evenhigher for RIF than for RSV (M. Feldman, per-sonal communication), indicating that the releaseof the virus particle once it has attained infec-tive maturity is very rapid indeed. Therefore,it is likely that RIF, like RSV, is completed atthe cell surface.To determine the conditions required to estab-

lish tolerant infection, embryos which had beenincubated for 3 to 17 days were infected byvarious routes of inoculation. Infection was alsocarried out on chicks from 1 day to 6 weeks afterhatching. The chicks were bled every other weekfrom 1 to 17 weeks after hatching and the RIFand antibody contents of the sera determined.The results in Table 3 show that the basic

requirement for the establishment of toleranceis that the virus must multiply to a high con-centration in the infected chick within the firstfew weeks after hatching. Since this would requireat least two cycles of growth in the animal andsince the virus grows slowly, it would be neces-sary to infect the embryo at least a few daysbefore hatching to reach a high level of virusmultiplication during the first few weeks of life.Indeed, this is borne out by the results whichshow that there is great difficulty in establishingtolerant infection in chickens after hatching evenwhen high concentrations of virus are inoculatedinto 1-day-old chicks.

IV. DISCUSSION

The results of these experiments provide aclear picture of infection with RIF, an almostubiquitous tumor virus. The prevalence of thevirus would not be possible if it were as virulentfor the host as are RSV or myeloblastosis virus,because inoculation of the more virulent agentsinto the embryo results in death of the chickwithin a few days or weeks after hatching. Thislethality provides an explanation for the rarity ofthese viruses in the field. It also makes it perfectlyclear that such "hothouse" strains of virus areunsuited for studying certain crucial aspects ofthe relationship between tumor virus and host.One of these aspects is of course the role of con-genital transmission in producing immunologicaltolerance which has been adequately discussedabove. Another aspect concerns the conditionswhich shift a well-nigh perfect symbiotic relation-ship between virus and host into a lethal one. Itmay well be that the RSV model for cellular alter-ation by changing the surface is also an appropri-ate one to describe the cellular changes in RIF in-fection once the balance has been shifted. In thecase of RSV, however, multiplication of thevirus is itself sufficient to cause a malignantchange in cell behavior, but this is clearly notthe case in RIF infection. There must be otherfactors which precipitate the malignant altera-tion and at present these can best be studiedwith the naturally occurring agent itself. There-fore, it seems reasonable to conclude that thestudies of RSV infection at the level of the celland RIF infection at the level of the organismwill combine to provide a comprehensive pictureof the viral carcinogenesis in the chicken, andperhaps in other animals as well.

12 [VOL. 26

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

INFECTION WITH AVIAN TUMOR VIRUSES

V. LITERATURE CITED

1. ABERCROMBIE, M., J. HEAYSMAN, AND H.KARTHAUSER. 1957. Social behavior of cellsin tissue culture. III. Mutual influence ofsarcoma cells and fibroblasts. Exptl. CellResearch 13:276-291.

2. BEARD, J.W., 1957. Etiology of avian leukosis.Ann. N. Y. Acad. Sci. 68:473-486.

3. BURMESTER, B. R., AND N. F. WATERS. 1955.The role of the infected egg in the transmis-sion of visceral lymphomatosis. PoultrySci. 34:1415-1429.

4. CARR, J. G. 1953. The mode of multiplicationof the Rous no. 1 sarcoma virus. Proc. Roy.Soc. Edinburgh, B, 65:66-72.

5. CRAWFORD, L. V. 1960. A study of the Roussarcoma virus by density gradient centrif-ugation. Virology 12:143-153.

6. ELLERMANN, V., AND 0. BANG. 1908. Experi-mentelle Leukamie bei Huhnern. Centr.Bakteriol. Parasitenk., Abt. I (Orig.)46:595-609.

7. FRIESEN, B., AND H. RUBIN. 1961. Some physi-cochemical and immunological properties ofan avian leukosis virus (RIF). Virology15:387-396.

8. GROSS, L. 1961. Vertical transmission of pas-sage A leukemic virus from inoculated C3Hmice to their untreated offspring. Proc. Soc.Exptl. Biol. Med. 107:90-93.

9. HOTCHIN, J., AND H. WEIGAND. 1961. Studiesof lymphocytic choriomeningitis in mice. I.The relationship between age at inoculationand outcome of infection. J. Immunol.86:392-400.

10. LooMIs, L. N., AND A. W. PRATT. 1956. Thehistogenesis of Rous sarcoma. I. Induced bypartially purified virus. J. Natl. CancerInst. 17:101-123.

11. PRINCE, A. M. 1958. Quantitative studies on

Rous sarcoma virus. III. Virus multiplica-tion and cellular response following infec-tion of the chorioallantoic membrane of thechick embryo. Virology 5:435-437.

12. Rous, P. 1911. A sarcoma of the fowl transmis-

sible by an agent separable from the tumorcells. J. Exptl. Med. 13:397-411.

13. RUBIN, H. 1955. Quantitative relations be-tween causative virus and cell in the Rousno. 1 chicken sarcoma. Virology 1:445-473.

14. RUBIN, H. 1960. The suppression of morpho-logical alterations in cells infected withRous sarcoma virus. Virology 12:14-31.

15. RUBIN, H. 1960. A virus in chick embryoswhich induces resistance in vitro to infectionwith Rous sarcoma virus. Proc. Natl. Acad.Sci. U. S. 46:1105-1119.

16. RUBIN, H. 1961. Influence of tumor virus in-fection on the antigenicity and behavior ofcells. Cancer Research 21:1244-1253.

17. RUBIN, H., M. BALUDA, AND J. E. HoTcHIN.1955. The maturation of western equine en-cephalomyelitis virus and its release fromchick embryo cells in suspension. J. Exptl.Med. 101:205-212.

18. RUBIN, H., A. CORNELIUS, AND L. FANSHIER.1961. The pattern of congenital transmissionof an avian leukosis virus. Proc. Natl. Acad.Sci. U. S. 47:1058-1060.

19. RUBIN, H., R. M. FRANKLIN, AND M. BALUDA.1957. Infection and growth of Newcastle dis-ease virus (NDV) in cultures of chick em-bryo lung epithelium. Virology 3:587-600.

20. TEMIN, H. M., AND H. RUBIN. 1958. Character-istics of an assay for Rous sarcoma virus andRous sarcoma cells in tissue culture. Vir-ology 6:669-688.

21. TEMIN, H. M., AND H. RUBIN. 1959. A kineticstudy of infection of chick embryo cells invitro by Rous sarcoma virus. Virology8:209-222.

22. TRAUB, E. 1960. Observations on immunologi-cal tolerance and "immunity" in mice in-fected congenitally with the virus of lym-phocytic choriomeningitis (LCM). Arch.ges. Virusforsch. 10:303-314.

23. VOGT, P., AND H. RUBIN. 1961. Localization ofinfectious virus and viral antigen in chickfibroblasts during successive stages of in-fection with Rous sarcoma virus. Virology13:528-544.

131962]

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from