Follicular exclusion of retroviruses in the bursa of fabricius

VIROLOGY

170, 433-441 (1989)

Follicular Exclusion of Retroviruses in the Bursa of Fabricius

D. L. EWERT, 1 N . AVDALOVIC,2 AND C. GOLDSTEIN The Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvania 19104 Received October 11 . 1988; accepted February 9, 1989 To gain insight into the regulation of retroviral infection at the cellular level, we analyzed the distribution of retroviral antigen and nucleic acid in the bursa of Fabricius of the parents and progeny of two highly inbred lines of chickens, one resistant and the other susceptible to infection . Line 151 5 chickens and line 7 2 , which are C/C and C/A, respectively, and 151 5 X 72 F1 chickens were infected with either RAV-1 or RAV-49 avian leukosis virus (ALV) . Most bursal follicles of F1 chickens infected with either virus contained a variable mixture of virus-positive and virus-negative cells and a few (1 to 20%) were void of detectable virus . However, in either parental line the respective virus was uniformly expressed among all follicles . The follicles which excluded virus in the F1 birds were indistinguishable from other infected follicles in the same bursa or in uninfected birds on the basis of histology or cellular antigen expression . It was concluded that virus susceptibility is most likely determined at the bursal stem cell level of differentiation, possibly by a process of allelic exclusion at the retroviral receptor locus . 01989 Academic areas, Inc .

INTRODUCTION A knowledge of the factors that regulate susceptibility of cell populations to retroviral infection is essential to an understanding of the pathogenesis of these viruses . The first prerequisite to infection is the adsorption of the virion to the cell followed by the penetration of the cell membrane. Studies of genetically distinct populations of animals indicate that cellular genes encode structures that mediate viral attachment and/or penetration of the cell membrane in a highly selective manner based on recognition of viral-subgroup-specific determinants of the viral envelope glycoprotein (Weiss, 1984) . In chickens host susceptibility to retroviral infection has been analyzed in the animal itself or in mixed cultures of embryonic fibroblasts by measuring viral interference, serological cross-reactivity, or host-range restrictions . These studies have led to the identification of three dominant autosomal tumor virus (tv) loci that control susceptibility to infection by five antigenically distinct subgroups of retroviruses of the avian leukosis/ sarcoma group . The loci that determine susceptibility to group A and C viruses, tva and tvc, respectively, are linked loci (Vogt and Ishizaki, 1965 ; Rubin, 1965 ; Crittenden et al., 1967 ; Duff and Vogt, 1969 ; Dren and Pani, 1977) . The tvb locus which is not linked to tva, determines susceptibility to the subgroup B, D, and E viruses (Crittenden and Motta, 1975; Pani, 1977 ; Weiss, 1984) . These genes are thought to encode cel-

lular receptors that recognize subgroup determinants of the viral envelope glycoprotein and promote viral penetration of the cell (Piraino, 1967 ; Crittenden, 1968) . Because the occurrence and dominance of each receptor gene was determined in a mixed population of cells, either in the intact animal or in cultures of fibroblasts, no conclusion was drawn in the above studies concerning the expression of retroviral receptors by individual cells . Our purpose was to obtain insight into the regulation of susceptibility to retroviral infection at the level of individual cells . The bursa of Fabricius proved useful for this investigation for several reasons. First, bursal cells are highly susceptible to retroviral infection and are the target cells of avian leukosis virus (ALV)-induced transformation leading to the formation of lymphomas of B cells in chickens (Peterson etat, 1964) . Second, bursal cells exhibit ALV subgroup restrictions to infection, as defined in fibroblasts of the same line of chicken, suggesting a common regulatory mechanism governing infection of the two cell lineages (Ewert and Goldstein, 1986) . Third, the immature B lymphocyte populations within individual follicles of the bursa have been shown to arise from a limited number (one to three) of stem cells which enterthe anlage of the bursa between Days 8 and 18 of embryonic development (Le Douarin et al., 1975 ; Houssaintetat, 1976 ; Pink etat, 1985 ; Ratcliffe et al., 1986) . The ontological organization of follicles within the bursa therefore permits analysis of clonal and pauci clonal populations of cells, a possibility not easily achieved for other cell lineages . The results presented below suggest that susceptibility or resistance

' To whom requests for reprints should be addressed . 'Present address : Beckman Instruments, Palo Alto, CA 94304 . 433

0042-6822/89 $3 .00 Copyright © 1989 by Academic Press, Inc . All rights of reproduction in any form reserved.



4 34

EWERT, AVDALOVIC, AND GOLDSTEIN

of bursal cells to retroviral infection is determined at the stem cell level of differentiation prior to or shortly after the cells enter the bursal follicle . MATERIALS AND METHODS Chickens All chickens used in this study were bred from highly inbred lines maintained under specific pathogen-free and ALV-free conditions at the Regional Poultry Research Laboratory (RPRL) (East Lansing, MI) . Fertile eggs from chicken lines 7 2 and 151 5 as well as the F, cross between line 151 5 males and line 7 2 females were obtained from RPRL . Line 151 5 is a subline of line 151, which is C/C with respect to Rous sarcoma virus (RSV) or ALV subgroup susceptibility and homozygous for the teas and tvc` viral receptor alleles (L . B . Crittenden, personal communication) . Line 7 2 is phenotypically C/ A, B, E, and homozygous for the tva' and tvc 8 alleles . Eggs were hatched and chicks reared in The Wistar Institute Animal Facility under conditions that meet the standards of the National Institutes of Health and of the U .S . Department of Agriculture . Virus infection ALV, RAV-1 (obtained from L . B . Crittenden) was subcloned by limiting dilution and grown to high titer on monolayers of fibroblasts from 9-day embryonated chf - /gs chick embryos line 11, SPAFAS (Norwich, CT) . ALV, RAV-49 was obtained from W . Mason . Both viruses originated in the laboratory of P . K . Vogt . The infectious titer of the virus stocks was determined by endpoint dilution on fibroblasts . After 5 days in culture, the titer was determined as the highest dilution that gave virus antigen-positive fibroblasts . Viral antigen was detected by fluorescence antibody analysis of fixed cytocentrifuge preparations of trypsinized fibroblasts using antisera specific for the viral envelope and core antigens as described below . Chickens were infected by injection of 0 .1 ml of culture supernatant containing 10 5 -10 5 infectious units (IU) of virus, either via the yolk sac on Day 6 of incubation, or iv, via the jugular vein, on the first or second day posthatch . Preparation and analysis of bursal tissues Individual plicae were removed from the bursa of Fabricius and processed by a modification of the paraffinembedding technique of Sainte-Marie (1962) . Tissues were incubated with constant mixing at 4°, for 18 hr in 95% ethanol, 4 hr in two changes of 100% ethanol, and 18 hr, with one change after the first 2 hr, in xylene,

and embedded in paraffin before sectioning . Deparaffinization was as described by Sainte-Marie, (1962) . Individual bursal follicles were released from fresh bursal tissue by scraping of the plicae with a scalpel in Hanks' balanced salts solution (HBSS) containing 5% bovine serum . The isolated follicles were washed several times in HBSS by 1 g sedimentation before being processed for paraffin embedding as described above . Deparaffinized sections of tissue were reacted for 20 min with antibodies specific for viral and cellular antigens in a humidified chamber and washed for 20 min in phosphate-buffered saline before being stained with a secondary fluorochrome-labeled or peroxidase-labeled antibody reactive with the primary antibody . A rabbit polyclonal antisera to the retroviral core antigen, p27, was purchased from SPAFAS, Inc . (Norwich, CT) . The rabbit antisera specific for the retroviral envelope glycoproteins has been previously described (Ewert and Halpern, 1982) . Mouse monoclonal antibodies to B lymphocyte antigens were obtained as follows : antibody to the heavy chain of the IgM immunoglobulin isotype was obtained from Southern Biotechnology Associates, Inc . (Birmingham, AL) ; antibody to the Bu-1 a Bcell alloantigen, mouse hybridoma L-22, was a gift from R . Pink (Pink and Rijnbeek, 1983) ; and antibody to monomorphic determinants of the chicken MHC class II (BL) antigens was as previously described (Ewert et al., 1984) . Tissue-reactive antibodies were detected using either fluorescein isothiocyanate (FITC)-labeled, affinity-purified goat anti-rabbit immunoglobulin (Ig) (Cooper Diagnostics, Malvern, PA), orfluorochrome-labeled or horseradish peroxidase-labeled goat antimouse Ig (Southern Biotechnology Associates, Inc .) . Peroxidase-stained tissues were exposed to diaminobenzidine-tetrahydrochloride (DAB) to develop a dark brown color in the antibody-reactive tissue . Tissue sections to be stained with peroxidase-conjugated antibody were immersed in hydrogen peroxide (30%)methanol (1/100, v/v) for 30 min before staining to quench the endogenous peroxidase . Reagent controls were obtained by substituting either normal rabbit serum or hybridoma supernatant containing antibody to an irrelevant antigen but of the same isotype for the corresponding primary antibody . In addition, sections were stained with methyl green pyronine to detect hyperplastic follicle or periodic acid-Schiff stain to detect the basement membrane separating the cortex and medulla of the follicles . In situ hybridization The method of in situ hybridization was essentially as previously described by Haase et at (1984) and



B-CELL SUSCEPTIBILITY TO RETROVIRUSES

modified by Stroopet at (1984) . Tissues were fixed and paraffin-embedded as described above . Sections, 5-6 µm thick, were mounted on poly-L-lysine-treated slides, deparaffinized with xylene, and treated with proteinase K . The 35 S-labeled env probe was denatured and diluted in hybridization mix (Stroop et al., 1984) to contain 1 ng of DNA/5 gl and approximately 10 6 cpm per tissue section . The hybridization mix containing the probe was heated for 30 sec at 100°, cooled on ice, and prehybridized for 1 hr at 45° . After 10 TIM dithiothreitol was added, 5 )AI of the hybridization mix was placed on each tissue section . The sections were then covered with baked siliconized coverslips and paraffin oil . After hybridization at 50° for 48 hr, the paraffin oil was removed with a chloroform wash . The slides were washed, dehydrated, dipped in NTB2 nuclear track emulsion (Kodak), and exposed for 1-3 days at 4° . After development with D 19 (Kodak) the sections were stained with hematoxylin and eosin . The probe used for hybridization was derived from the R-8 clone of RAV-O (Hughes, 1982) by isolating the Hindlll-EcoRl fragment containing all of the env region and a portion of pol and reinsertion into pBR322 . Before use, the pol fragment was removed from the cloned insert by digestion with Kpnl and reisolation . Prior to hybridization, this env probe was labeled with [36 S]dCTP by the primer extension method, as described by Feinberg and Vogelstein (1983) . RESULTS Bursal tissue from 151 5 x 7 2 (Fl) chickens obtained 4 to 8 weeks postneonatal infection with RAV-1 was analyzed for retroviral core antigens by indirect immunofluorescence or immunoperoxidase staining of tissues using antisera to the p27 viral core antigens . Although the majority of the bursal follicles were positive for virus as indicated by the dark brown peroxidase staining, a variable number of negative follicles were dispersed among the virus-positive follicles, (Fig . 1 B) . This phenomenon was observed in both male and female F1 chickens . Any staining of endogenous retroviral antigens expressed in these chickens was below the level of detection in the bursa, as noted in bursas of uninfected chickens . The virus-positive follicles in the F1 chickens also varied in their intensity of staining . A similar pattern of virus-positive and -negative follicles was obtained by staining a serial section of the bursa with antisera specific for the retroviral envelope glycoprotein (e .g ., Fig . 4B) . The frequency of the virus-negative follicles varied from 1 to 20% (mean = 7 .2%) in individual F1 chickens (Fig . 2) . Similar frequencies of viruspositive and virus-negative follicles were obtained by

435

infecting F1 chickens in ovo, via the yolk sac on the sixth day of incubation with 10 ° infectious units of the RAV-1 stock(Fig .2) . By comparison,bursaltissuefrom the parental line 151 6 chickens that were similarly infected and sacrificed had no detectable negative follicles and had a more uniform high level of anti-p27 associated staining (see Figs . 1A and 2) . To determine if this phenomenon could be obtained with another subgroup of retrovirus, chickens of the F1 cross the 7 2 parental line were neonatally infected with RAV-49, a group C ALV . Again, virus-negative follicles were observed in bursas of the F1 cross but not in the parental line (Fig . 2) . However, in F1 chickens neonatally infected with a mixture of RAV-1 and RAV-49, the bursas obtained at 6 weeks of age contained no detectable virus-negative follicles (Fig . 2) . Examination under high magnification of the anti-p27 immunoperoxidase-stained bursal tissues from the RAV-1 -infected 151 5 chickens revealed that essentially all cells expressed viral antigen (Fig . 3) . By contrast, individual follicles in the F1 chickens contained variable proportions of virus-positive and -negative cells . Most follicles contained a mixture of virus-positive and virusnegative cells, but a significant number contained no detectable virus-positive cells . Further analysis was made of the virus-negative follicles to determine the possible basis for the exclusion of virus . To rule out the possibility that the differential staining of follicles was an artifact of fixation, individual follicles were isolated from the bursa prior to fixation and paraffin-embedding . The frequency of negative follicles among the isolated follicles (Fig . 4) was similar to that obtained by counting 300-500 follicles in four or five plicae of the bursa, suggesting a random distribution of these virus-negative follicles among the plicae . Another possibility was that these follicles contained cells that were developmentally abnormal or contained cells, other than B lymphocytes, that were refractory to infection . To test this possibility, we analyzed sequential sections of tissue containing virus-negative follicles for expression of antigens characteristic of early B-lymphocyte development, i .e ., immunoglobulin (IgM) (Fig . 5), MHC class II (BL) antigens, and the Bu-1a alloantigen (data not shown) . In each case, the pattern of antigen-associated fluorescence of the virus-negative follicles was indistinguishable from that of uninfected chickens and indistinguishable from that of the viruspositive follicles in the same section of bursa . Also, virus-negative follicles were pyronin-negative on sequential sections stained with methyl green pyronin (not shown due to lack of photographic contrast) . The preneoplastic nodules that appear in the bursa during the early stages of ALV-induced lymphomagenesis are



4 36

EWERT, AVDALOVIC, AND GOLDSTEIN

FIG . 1 . ALV antigen expression (dark contrast) in sections of bursa of Fabricius from neonatally RAV-1-infected 151 5 (A) or 151 5 x 7 . (B) chickens . Sections were stained with rabbit antisera to the viral core, p27 antigen, followed by peroxidase-labeled goat anti-rabbit 1g . Following development with the DAB substrate the sections were stained with periodic acid-Schiff stain to highlight the cortical-medullary basement membrane (arrow) . Bar = 200 µm .

CHICKEN LINE (Receptor genotype for subroups R B C RLU)

VIRUS RLU ISubgroup)

INFECTION FEE

FREQUENCY OF VIRUS-NEGRTIUE FOLLICLES

in IN BURSRS OF INOIUIDURL CHICKENS

INFECTION 0

RPU-IIP)

1 day

0 0 0

p .11 .

0 0 0 0 0 0 I I I

I

0 0 0 0 0 0 I I I

0 0 0 0 0 0 I

I

I

I

I

I

I

o a 0 0 I

I

I

I

I

I

I

I

I

I

a

0 0 0 I

I

0 I

I

0 0 I

1

1

a 1

0 1

1

a 1 1

a I

0 0 I

f

I

I

I

I

I

I

1

I

I

I

0 a I

I

I

I

1

1

1

1

1

1

I

I

I

I

I

I

I

I

1

1

I

I

I

I

I

RflU-1IP) 151

5

R 7 C,C

ID d9-US

2

s~ RRU-49(C)

1 dog P .N .

RBU-IIRI RRU-4910)

1 dog P .0 .

~eP :r .r)r

RRU-1 (0)

1 dog .O P .

~D~a~rs ,r s )

RRU-49([)

1DA . y

I

I

I

I

I

I

I

I

00 00 00 00 I 0

72

1

0 0 I

151 5

I

0 0

I

I

I

I

I

I

00 02469101214161020 PERCENT

FIG . 2 . Frequencies of bursal follicles showing no viral antigen expression in bursal sections stained with antibody to retroviral core p27 antigen . The genotype of the chickens in each group is given for the susceptibility (s) or resistance (r) alleles of the tumor virus (tv) receptor loci encoding receptors for subgroup A and C avian leukosis viruses . Chickens were infected either 1 day posthatch (p .h .) or on the sixth day of incubation . The percentages of viral negative bursal follicles were obtained by counting 300-500 follicles in four to five plicaefrom each chicken .

B-CELL SUSCEPTIBILITY TO RETROVIRUSES

437

FIG . 3. ALV antigen expression in sections of bursa of Fabricius from neonatally RAV-1-infected 151 5 (A) or 151 5 X 7 2 (B) chickens . Staining was as in Fig . 1 . Bar = 20 µm .

FIG . 4 . ALV antigen expression (light contrast) in sections of RAV-1-infected, 151 5 x 7 2 bursal follicles that were dispersed prior to fixation . Sections were stained with antibody specific for the retroviral envelope glycoproteins followed by fluorescein isothiocyanate-labeled goat antirabbit Ig. Pictures were taken with dark field illumination (A), or uv illumination (B) . Bar = 200 Am .

4 38

EWERT, AVDALOVIC, AND GOLDSTEIN

FIG . 5. Retroviral core (A) and immunoglobulin, IgM, y chain, (B) expression in sequential sections of bursa of Fabricius from neonatally RAV1-infected 151 5 x 7 . chickens . Core antigen (dark contrast) was detected using rabbit antibody to the p27 antigen, followed by peroxidase labeled goat anti-rabbit Ig antibody . Following development with the DAB substrate, the sections were stained with periodic acid-Schiff to highlight the cortical-medullary basement membrane (arrow) . B-cell-associated IgM (light contrast) was detected by indirect immunofluorescence analysis using a monoclonal antibody specific for the µ chain, followed by FITC conjugated goat anti-mouse Ig antibody . Bar = 100 µm .

selectively stained pink by this method (Cooper et al., 1968) . Also, the basement membrane that forms the boundary between the cortex and medulla of the follicle and is disorganized in the preneoplastic follicles (Cooper etat, 1968) was intact in the virus-negative follicles as shown on sequential sections stained with periodic acid-Schiff stain (see Figs . 18 and 5A) . Finally, the level of virus-specific cell-associated nucleic acid was analyzed by in situ hybridization to determine if the lack of viral antigen expression in some follicles was a consequence of a failure to translate the viral message . Sections of the bursa were hybridized with a radiolabeled cloned DNA probe for the envelope region of a RAV-O virus which cross-hybridizes with viruses of other subgroups . As shown in Fig . 6, there was no hybridization above background levels in the anti-p27-negative follicle of the RAV-1-infected Fl chicken using an 35 S-labeled env probe, while the adjacent anti-p27-positive follicle had a significant level of hybridization . DISCUSSION A complete explanation for the observed difference in the cellular and follicular distribution of virus within

the bursas of parental and F1 chickens is not apparent . The follicles appear to be normal by criteria of cellular antigen expression and histology . Nevertheless, the analysis of retroviral core and envelope proteins and viral nucleic acids indicates that virus was excluded from bursal lymphocytes in F1 chickens and that the distribution of virus-positive and -negative cells in individual follicles resulted in the complete exclusion of virus from lymphoid cells of some follicles . The observation that the negative follicles lacked detectable envelope as well as core viral antigens indicates that a mechanism of viral interference by envelope antigens produced by a defective provirus is not a likely basis for the exclusion of virus . By contrast, in the parental lines the retrovirus was more uniformly expressed among all the follicles . The level of viral antigen expression was consistently higher in the medulla than in the follicular cortex of bursas from both the F1 and the parental lines . A possible explanation to account for virus-negative follicles is a physical exclusion of virus at the follicular level . However, it is difficult to argue for a selective follicular barrier to virus passage since the restriction appears to be at the cellular level, as evidenced by the

B-CELL SUSCEPTIBILITY TO RETROVIRUSES

439

Fro . 6 . Retroviral antigen-specific (A) and envelope (env)-specific RNA (B) expression in sections of RAV-1-infected, 151 5 X 7 2 bursal follicles . Retroviral core antigen was detected by immunoperoxidase staining with rabbit anti-p27 antibody as in Fig . 1 . Retroviral-specific RNA was detected by in situ hybridization using a 35 S-labeled probe of the env region of RAV-0 . Bar = 100 gm .

follicles containing mixtures of positive and negative cells . Also, such a physical barrier would be expected to be operative in either one or both of the parental lines . Because of the presence of retrovirus in the environment of the bursa, both in blood vessels and epithelium of the virus-negative follicles, and in follicles surrounding the negative follicles, we conclude that a subpopulation of bursal cells in the F1 bursa is refractory to infection . Another possible explanation for the lack of viral expression by these isolated follicles is that they represent a preneoplastic stage of lymphomagenesis which precedes the development of progressive bursal lymphomas in ALV-infected chickens . The preneoplastic follicles are larger than normal follicles and are characterized histologically by their pyroninaphilic staining, indicative of an elevated level of RNA synthesis and the disappearance of the basement membrane that separates the cortex and medulla of the follicle, and by their uniform blast-like cellular content . By contrast, the cellular composition and histological characteristics of the virus-negative follicles described in this study were generally indistinguishable from the normal bursal follicles . Furthermore, virus-negative follicles have been observed in chickens as early as 4 weeks after neonatal infection which is about 4 weeks priorto the appearance of the preneoplastic nodules in this line of chicken . It is possible but unlikely that these virus-negative follicles are the progenitors of the preneoplastic

follicles since the incidence of ALV induced lymphomas in the 151, parental line approximates that of the F1 cross (J . Motta, personal communication) . If such a precursor relationship existed, the frequency of such virus-negative follicles would be expected to be as high or higher in the 151 5 parental line than that of the F1 chickens . Furthermore, the observation of both viruspositive and -negative preneoplastic follicles in 151 5 and 151, X 7 2 , F1 chickens (our unpublished observation) suggests that cessation of viral synthesis is not a prerequisite of preneoplastic transformation . An explanation we favor is suggested by a striking similarity in the distribution of allotypically marked cells in bursal follicles observed in a study by Pink et al. (1985) designed to enumerate the number of lymphoid precursor cells populating bursal follicles . Chick embryos (8-11 days) were parabiosed to permit exchange of Bu-1 alloantigen-marked bursal stem cells during the time that the bursa anlage is receptive to lymphocyte precursor implantation . They observed that most of the follicles in sections of the mature bursa of Bu-1a+ chickens contained a mixture Bu-la alloantigen-positive and -negative cells but that a few, 0 to 35% (mean of 7%), contained cells exclusively of the donor (Bu-1 b) allotype . By comparison, the frequencies of virus-negative follicles of the F1 chickens in the present study ranged from 1 to 20% (mean of 7%) . Pink et al. (1985) interpreted the exclusion of cells of one allotype in individual follicles as resulting from the statistical probabil-



4 40

EWERT, AVDALOVIC, AND GOLDSTEIN

ity of a limited number of stem cells bearing either the Bu-1 a or the Bu-1b allotype entering each follicle . On the basis of the number of follicles containing cells of a single allotype, they calculated the average number of precursor cells entering each follicle to be 3 .4 (Pink et al., 1985) . If we assume that commitment to either viral susceptibility or resistance occurs at the stem cell level and make similar assumptions concerning the lack of bias toward the phenotype of the precursor cell and a binomial distribution of the precursors populating each follicle, as were made in the previous reports, we estimate the number of precursors to be 3 .8, based on an average of 7% virus-negative follicles . Therefore our findings are consistent with a model in which commitment to viral susceptibility or resistance is made prior to or shortly after the lymphoid stem cell enters the follicular anlage . An obvious difference between the biological system described by Pink et al., (1985) and that described herein is that theirs involved the mixing, by parabiosis, of stem cells of two allotypically distinct embryos whereas our observations were made in chickens derived from a single embryo . Therefore a mechanism that would account for the differentiation of bursal stem cells into viral susceptible and viral resistant populations must be invoked . Such a mechanism is known to regulate the expression of alleles of immunoglobulin genes in B lymphocytes, by the suppression of the unexpressed allele (Pernis et al., 1965) . Allelic exclusion has been shown to be operative in lymphoid stem cells of the bursa of Fabricius, restricting the precursor cells to expression of a single IgM allotype prior to or shortly after entering the follicular primoidium (Ratcliffe et al., 1986) . The resultant distribution of B cells expressing different IgM allotypes in the bursa of heterozygous chickens is also very similar to what we observed for virus infection . Ratcliffe et al. (1986) observed that 7% of the bursal follicles of heterozygous chickens contained cells expressing a single allotype, with most follicle being allotypically mixed . They also estimated the follicular stem cell pool to be 2 .6 or 3 .5 cells based on the findings of two experimental groups . Therefore a mechanism of allelic exclusion operative at the tv locus in stem cells prior to or shortly after entering the bursa can account for the observed distribution of virus-positive and virus-negative follicles as reported here . Consistent with a mechanism of allelic exclusion controlling the expression of viral receptors is the failure to observe virus-negative follicles in either parental line, indicating that the process of exclusion is operative only in the Fl chicks where both the resistance and susceptible alleles of the tv locus are present . A mechanism of differential expression based on linkage to sex

chromosomes, which are allelically excluded, is unlikely since both male and female F1 chickens demonstrated this phenomenon . In summary, we believe the best explanation for the observed follicular distribution of virus-positive and -negative cells in the bursa of 151 5 x 7 2 , F1 chickens is that commitment to expression of the resistance or susceptible alleles of particular viral receptor loci occurs at the stem cell stage of lymphocyte differentiation prior to or shortly after entering the anlage of the bursal follicle . The proportion of virus-positive and -negative cells in a mature follicle therefore reflects the number of committed precursors expressing the resistance or susceptibility allele at the tv locus . Definitive proof that a mechanism of allelic exclusion is operative for the viral receptor genes would require that a single allele of the as yet uncharacterized receptor molecule be shown to be expressed in individual cells of a heterozygous animal . ACKNOWLEDGMENTS We acknowledge Wilfried Weber, Israel Steiner, James DuHadaway, and Elsa Aglow for their excellent technical assistance and Mike Halpern for reviewing the manuscript . This investigation was supported by Grants CA-39000 and CA-10815 from the National Institutes of Health and IM-320 from the American Cancer Society .

REFERENCES COOPER, M . D ., PAYNE, L . N ., DENT, P . B ., BURMESTER, B . R ., and GOOD, R . A . (1968) . Pathogenesis of avian lymphoid leukosis . I . Hlstogenesis . ). Nat!. Cancer lnst. 41, 373-389 . CRiTTENDEN, L . B . (1968) . Observations on the nature of a genetic cellular resistance to avian tumor virusesJ. Nat?. Cancer Inst . 41, 145-153 . CRITTENDEN, L . B ., and MonA, 1 . V . (1975). The role of the tv locus in genetic resistance to RSV(RAV-0) . Virology 67, 327-334 . CRITTENDEN, L . B ., STONE, H . A ., REAMER, R . H ., and OKAIAKE, W. (1967) . Two loci controlling genetic cellular resistance to avian leukosis sarcoma viruses . J. Viro?. 1, 898-904 . DREN, Cs . N ., and PANT, P. K. (1977) . Genetic control of resistance to subgroup A and Subgroup C tumor viruses in Rhode Island Red fowl : Evidence for linkage between the tumor virus a (tva) and tumor virus c (tvc) loci . J. Gen . Viral. 35, 13-23 . DUFF, R . G ., and VOGT, P . K . (1969) . Characteristics of two new avian tumorvirus subgroups . Virology39, 18-3D . EwERT, D . L ., and GOLDSTEIN, C . (1986) . Susceptibility of avian B lymphocytes to retroviral infection parallels that of firbroblasts . Virology 152, 262-267 . EwERT, D . L ., and HALPERN, M . S . (1982) . B cell expression of endogenous retroviral envelope antigen : Biochemical and immunofluorescence characterization . .1 immune?, 129, 1724-1730 . EwERT, D . L ., MuNCHUS, M . S ., CHEN, C .-L .H ., and COOPER, M . D . (1984) . Analysis of structural properties and cellular distribution of avian la antigen by using monoclonal antibody to monomorphic determinants . J. immune? . 132, 2524-2530 . FEINBERG, A., and VOOLESTEIN, B . (1983) . A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity . Anal. Biochem . 132, 6-13 .



B-CELL SUSCEPTIBILITY TO RETROVIRUSES HAASE, A . T ., BRAHIC, M ., STOWERING, L ., and BLUM, H . (1984) . Detection of viral nucleic acids by in situ hybridization . In "Methods in Virology ." (K . Maramorsch, and H . Koprowski, Eds .), Vol . 7, pp . 189-226 . Academic Press, New York . HouSSAINT, E ., BELO, M ., and LE DouARIN, N . M . (1976) . Investiga tions on cell lineage and tissue interactions in the developing bursa of Fabricius through interspecific chimeras . Dev . Riot. 53, 250264 . HUGHES, S . H . (1982) . Sequence of the long terminal repeat and adjacent segments of the endogenous avian virus Rous-associated virus. J. Virol. 43, 191-200 . LE DOUARIN, N . M ., HoussAINT, E ., JOTEREAu, F . V., and BELO, M . (1975) . Origin of hemopoietic stem cells in embryonic bursa of Fabricius and bone marrow studied through interspecific chimeras . Proc. Nad . Aced. Sci. USA 72, 2701-2705 . PANT, P . K . (1977) . Evidence for complemetary action of Nb and ive genes that control susceptibility to subgroup E RSV tumour virus in chickens . J. Gen . Virol. 37, 639-646 . PERNI5, B ., CHIAPPINO, G ., KELUS, A . S ., and GELL, P . G . H . (1965). Cellular localization of immunoglobulins with different allotypic specificities in rabbit lymphoid tissues . J. Exp. Med. 122, 863-856 . PETERSON, R . D . A ., BURMESTER, B . R ., FREDRICKSON, T . N ., PURCHASE, H . G ., and GOOD, R . A . (1964) . Effect of bursectomy and thymectomy on the development of visceral lymphomatosis in the chicken . ). Nell Cancer Inst . 32, 1343-1354 . PINK, J . R ., and RuNBEEK, A . M . (1983). Monoclonal antibodies against chicken lymphocyte surface antigens . Hybridoma 2, 287-296 .

441

PINK, J . R ., VAINIO, 0 ., and RIINBEEK, A .-M . (1985) . Clones of B lymphocytes in individual follicles of the bursa of Fabricius . Eur. J. Immunot. 15,83-87 . PIRAINO, F . (1967) . The mechanism of genetic resistance of chick embryo cells to infection by Rous sarcoma virus-Bryan strain (BSRSV) . Virology 32, 700-707 . RATCLIFFE, M . 1 . H ., LASSILA, 0 ., PINK, J . R . L ., and VAINIO, 0 . (1986) . Avian B cell precursors : Surface immunoglobulin expression is an early, possibly bursa-independent event . Eur. J. Immunol . 16, 129133 . RUBIN, H . (1965) . Genetic control of cellular susceptibility to pseudotypes of the Rous sarcoma virus . Virology 26, 270 276 . SAINTE-MARIE, G. (1962) . A paraffin embedding technique for studies employing immunofluorescence . J. Histochem . Cytochem. 10, 250-253 . STRooP, W. G ., ROCK, D . L ., and FRASER, N . W . (1984) . Localization of herpes simplex virus in the trigeminal and olfactory systems in the mouse central nervous system during acute and latent infections by in situ hybridization . Lab. Invest. 51, 27-38 . VOGT, P . K ., and I5HIZna, R . (1965) . Reciprocal patterns of genetic resistance to avian tumor viruses in two lines of chickens . Virology 26,664-672 . WEISS, R . A. (1984) . Experimental biology and assay of RNA tumor viruses . In "RNA Tumor Viruses," (R . Weiss, N . Teich, H . Varmus, and J . Coffin, Eds .), 2nd ed . Vol . 1, Chap . 3, pp. 209-260 . Cold Spring Harbor Laboratory, Cold Spring Harbor, NY .