INFECrION AND IMMUNITY, June 1978, p. 816-8260019-9567/78/0020-0816$02.00/0Copyright © 1978 American Society for Microbiology

Vol. 20, No. 3

Printed in U.S.A.

Scanning Electron Microscopy of Lymphoid Cells fromLeukemia Virus-Infected Mice

PAUL A. FARBER,' STEVEN SPECTER,2 AND HERMAN FRIEDMAN2*Department of Pathology, Temple University School of Dentistry' and Departments ofMicrobiology and

Immunology, Albert Einstein Medical Center and Temple University School ofMedicine,2Philadelphia, Pennsylvania 19122

Received for publication 21 December 1977

Spleen, lymph node, bone marrow, and thymus cells from Friend leukemiavirus (FLV)-infected mice were examined by scanning electron microscopy.Whereas splenocytes from normal, noninfected animals showed the expectedmorphological classes of lymphocytes, including cells with numerous villousprojections and smoother cell types, spleen cells from mice infected with FLVshowed a rapid alteration of surface morphology. Shortly after infection, adecrease in the number and percentage of villous cells occurred, with a concomi-tant increase in the number of cells that were larger and smoother. Within 10 to20 days after infection, the majority of splenocytes were smooth, large cellsshowing many distinct morphological charges, including surface "holes" and a"spongy" appearance. By days 25 to 35 after infection, most splenocytes wereabnormal in appearance. Similar changes occurred in the lymph nodes after FLVinfection, but the rate of change was much lower. Abnormal and larger smooth-surfaced cells did not become prominent until after week 2 or 3 of infection.Thymus and bone marrow cells showed little if any change in surface morphologyuntil late in the infectious process. However, even at that time only a few of thecells were abnormal in appearance. The changes in cell population in the spleenbut not the lymph nodes paralleled the rapid decrease in the percentage of cellswhich stained positive for surface immunoglobulin and theta antigen. Further-more, FLV antigen rapidly appeared on spleen cells after infection; fewer lymphnode cells were positive, and only low numbers of marrow and thymus cellsstained positive for FLV antigen. The marked immunosuppression induced byFLV infection paralleled and in some instances preceded the marked morpholog-ical changes.

The malignant transformation of normal cellsby leukemia viruses, both in vivo and in vitro, isthe subject of much current interest. After infec-tion of an experimental animal with an oncor-navirus such as Friend leukema virus (FLV),rapid alterations of lymphoid cell populationsbecome evident (11-13). For example, within 1to 2 weeks after infection of susceptible strainsof mice, splenomegaly occurs, followed by he-patomegaly and blood cell dyscrasia. Concomi-tant with development of these changes andoften preceding them is a marked and general-ized immunosuppression characterized by de-pressed antibody responses to a large variety ofantigens, including sheep erythrocytes (SRBC)and bacterial extracts (7, 19, 24). Cell-mediatedimmune responses to allografts and bacterialantigens also may be depressed. However, themechanisms of immunosuppression induced byleukemia virus infection, as well as the mecha-nism of virus-induced transformation of hostcells per se, is still far from clear.

Previous histological and ultrastructural stud-ies by transmission electron microscopy (TEM)revealed rapid change of lymphoid cell popula-tions in the spleens of mice infected with FLV(4, 15, 16). Those studies also suggsted thatimmunocyte precursors may be the major targetof leukemia virus infection, at least in regard toimmunosuppression. For example, a massiveproliferation of reticular and erythrocyte celltypes with atrophy of lymphoid follicles oc-curred 1 to 2 weeks after infection. Examinationby TEM showed the presence of C-type virusparticles in numerous nucleated spleen cells dur-ing the course of infection, with viruses buddingfrom many immature blast-like lymphoid cells.Virus particles were also seen within individualimmunocytes present in the centers of plaquesof hemolysis, indicating active secretion of anti-body to SRBC; thus, it was evident that virusinfection did not necessarily prevent antibodysynthesis by individual spleen cells.Scanning electron microscopy (SEM) has

816

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SEM OF FLV-INFECTED LYMPHOID CELLS 817

been widely utilized during the last few years inattempts to differentiate lymphocytes on thebasis of surface topography (17, 20, 22). Al-though it was initially believed by some thatlymphocytes with numerous surface villi mayrepresent B lymphocytes and those withsmoother surfaces may be T lymphocytes, thedifferentiation between these cell classes bySEM is controversial (1, 5, 18,21). Many lymph-oid cells have an intermediate surface morphol-ogy, and, moreover, cultural and preparativeconditions markedly affect the topography oflymphocytes. Nevertheless, marked differencesin cell types have been reported for lymphoidcell preparations from individuals with lympho-proliferative disorders as compared with cellsfrom normal individuals. Although most SEMstudies on lymphocytes have been performedwith human peripheral blood cells, the studiesby Polliack et al. indicate that similar topo-graphic criteria can be extended to murine splen-ocytes (21). It was reported that approximatelyone-third of splenic lymphocytes from normalB6 mice were highly villous, nearly one-halfwere smooth, and the remainder possessed in-termediate surface characteristics, with few villi.Although Polliack et al. felt that these surfacecharacteristics were related to cell class, espe-cially since comparative immunofluoresencestudies for lymphocyte surface markers corre-lated with SEM results, recent studies have sug-gested that villus formation is a function oflymphocyte activation and not lymphocyte class(26).

In the present study, spleen cells as well ascells from the thymus and bone marrow wereobtained from mice at various times after infec-tion with FLV and examined by SEM to deter-mine if there were discernible surface changescompatible with the changes seen in the previousTEM study. In addition, conventional immuno-fluorescent procedures were used to determinethe presence of cells with surface immunoglobu-lin molecules and theta antigen, as well as withleukemia virus-associated or -induced surfaceantigens. In all cases, lymphoid cells obtained atvarious times after infection of mice with theleukemia virus were compared with similarlymphoid cell populations from similar tissues ofnormal animals.

MATERIALS AND METHODSAnimals. Inbred male BALB/c mice, obtained

from Cumberland View Farms, Clinton, Tenn., wereused for these studies. The mice were approximately6 to 8 weeks of age at the start of an experiment. Theywere housed in groups of 6 to 10 in plastic mouse cagesand fed commercial mouse pellets and water ad libi-tum.Tumor virus. A stock preparation of FLV, ob-

tained initially from the American Type Culture Col-lection, Rockville, Md., was used for infecting mice asdescribed previously (8, 14). The virus preparation hasbeen passaged in this laboratory for approximately 10year$ and consisted of 10% clarified spleen cell homog-enates stored at -70'C in small aliquots. The viruspreparation contained both the spleen focus-formingand lymphatic leukemic virus components of theFriend complex. No detectable lactic dehydrogenaseor lymphocytic choriomeningitis virus was present.

Infection. Mice were infected by intravenous in-oculation of 0.2 ml of varying concentrations of thestock FLV homogenate. In a typical experiment, a 10-'dilution containing approximately 100 to 500 50% in-fective doses of virus was injected. Splenomegaly wasused as an indicator of infection. In addition, virustitrations were performed by determining the numberof macrofoci in the spleens of infected mice at 6 to 9days, as well as microfoci induced in vitro in DB56tissue culture cell lines using the method developed byBassin et al. (3).

Electron microscopy. At various time intervalsafter infection, mice were killed and each one's spleenbone marrow cells, thymus, and inguinal lymph nodeswere removed and placed in cold McCoy tissue culturemedium without serum or antibiotics. Thymus andlymph node cells were obtained by mincing with scis-sors. The spleen cells from individual mice were puri-fied by standard Ficoll-Hypaque gradient centrifuga-tion. The resulting cell preparations were then washedtwice with McCoy medium and fixed immediately with1% glutaraldehyde in phosphate-buffered saline. After60 min of fixation, the cells were aspirated onto Flo-tronic silver membranes (Selas Corp., Springhouse,Pa.). The membranes with cells were washed withphosphate-buffered saline, postfixed with 1% osmiumtetroxide in buffer for 15 min, washed again withbuffer, dehydrated in graded alcohols, and dried bythe critical-point method (21). Specimens were at-tached to aluminum studs, coated with carbon andgold-paladium, and examined with an Etec Autoscanmicroscope operating at 20 kV (Etec Corp., Hayward,Calif.). Cells from at least 10 mice per group wereexamined at various magnifications. In all cases, atleast several hundred sections, representing a signifi-cant sampling of an entire spleen or cell pool fromother lymphoid tissues, were examined. Since thespleens from infected mice were much larger thanthose from noninfected control mice, a proportionallylarger sample of infected spleen cells was examined.Similar sample sizes were examined for lymph node,marrow, and thymus cell suspensions, since there waslittle if any alteration in cell number between infectedand control mice.

Immunization. Representative mice, either con-trol or FLV-infected, were immunized by intravenousinjection of 0.2 ml of a 1% SRBC suspension whichhad been washed several times with isotonic saline inthe cold.Antibody determination. The numbers of indi-

vidual antibody plaque-forming cells in the spleens ofimmunized mice, both controls and FLV-infected,were determined by the standard localized hemolyticplaque assay in agar gel, as described previously (8, 15,16). In brief, 0.1-ml samples of a cell suspension wereadded to 2 ml of melted (42°C) Difco agar (1%) con-

VOL. 20, 1978

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

818 FARBER, SPECTER, AND FRIEDMAN

training an 0.1 ml inoculum of a 10% suspension offreshly washed SRBC. The agar-leukocyte-SRBCmixture was poured into the surface of a previouslyprepared 60-mm-diameter petri plate containing a

base layer of solidified agar. The plates were incubatedat 370C for 1 h and then treated with 5 ml of a 1:15dilution of guinea pig complement. After further in-cubation for 1 h, zones of hemolysis appearing on thesurface of the plates were considered to be due to 19Simmunoglobulin M hemolytic antibody. Blood was

obtained from selected mice in each group by retroor-bital venous puncture, and the resulting serum speci-mens were tested for hemolytic antibody to SRBC byserial twofold dilutions in microtiter plates.Immunofluorescent studies for surface anti-

gens. The percentate of lymphoid cells with surfaceimmunoglobulin molecules, as well as theta antigenand leukemia virus-associated antigen, was deter-mined by standard fluorescent microscopy exactly as

described previously (10, 13). In brief, rabbit anti-mouse globulin serum conjugated with fluorescein iso-thiocyanate (diluted 1:20) was incubated with an equalvolume of 5 x 106 lymphoid cells for 30 min at 4°C.The cells were then washed several times, and smears

were prepared on glass slides. At least 500 cells were

examined with an ultraviolet microscope to determinethe percentage of lymphocytes positive for surfaceimmunoglobulin. The percentage of cells with thetasurface antigen was determined by similarly incubat-ing cells with fluorescein-labeled AKR anti-C3H mousethymus serum for 30 min at 4°C. After washing, thepercentage of cells staining positive with the anti-thetaserum was determined. Anti-FLV serum was preparedby hyperimmunization of normal isogeneic BALB/cmice with formaldehyde-treated FLV homogenatesweekly for a period of at least 4 weeks. The resultingantisera, capable of completely neutralizing a 10'concentration of FLV after 30 min of incubation at25°C at a dilution of 1:50, resulted in positive patchesof surface fluorescence on splenocytes of FLV-infectedbut not normal mice. The percentage of such positivecells was determined for each cell suspension. In allcases, both positive and negative controls were used.

RESULTSEffect ofFLV infection on splenocyte cell

morphology. Spleen cells from normalBALB/c mice examined by SEM showed theexpected spectrum of lymphoid cell types. Themajority of splenocytes had numerous villousprojections. Lymphocytes with smoother sur-

faces were present in smaller numbers, as were

intermediate cell types with fewer villi. Large,irregular-shaped cells with ruffled surfaces wereevident in small numbers; these had a morphol-ogy consistent with macrophages. Within 3 to 5days after FLV infection, prior to significantsplenomegaly, marked changes in the percent-age of smooth and villous cell types becameevident (Fig. 1, Table 1). An increasing numberof large, smooth-surfaced cells appeared; thesewere larger than the smooth-surfaced lympho-cytes of normal, control spleens. This type of

cell increased during the next few days so thatby days 5 to 10 after infection they comprisednearly 30% of the splenic lymphocyte popula-tion. By days 15 to 20, 45 to 70% of the lympho-cytes had a smoother topography and werelarger than most lymphocytes present in thespleens of noninfected control mice. Few nor-mal-looking, villous lymphocytes were present.By day 25 after infection, the large smooth-sur-faced cells accounted for over 95% of the spleniclymphocytes; only occasional villous cells couldbe found (Table 1). Many of the large cellsshowed "blebs," measuring approximately 100nm, on the surface. These were first noted onabout day 10, and cells bearing these surfacemodifications increased in number as infectionsprogressed. By day 20, degenerating large cellswere also evident, with holes and blebs appear-ing on the surfaces of the smooth-surfaced-typecells (Fig. 2). These surface blebs often seemedto coalesce and perhaps to result in cell lysis.

Effect of virus infection on lymph nodecells. Lymphocytes from superficial lymphnodes were examined for changes following FLVinfection. In contrast to the spleen, wheremarked alterations in lymphocyte surfaces rap-idly occurred within 5 to 10 days, the lymphnode cells at this time appeared relatively un-changed (Fig. 3). Both normal and infectedlymph nodes had numerous cells with villousprojections, as well as smaller numbers of cellswhich seemed to be macrophages with ruffledmembranes. The number of smooth-surfacedcells, especially those with surface blebs, in-creased but did not form an appreciable per-centage until 35 days postinfection. Even at thislate date, normal-appearing villous-covered cellswere still evident in large numbers.Changes in bone marrow and thymus cell

populations after FLV infection. Both bonemarrow and thymus cell populations showed thefewest changes as a result of FLV infection (Fig.4 and 5). Few abnormal large cells with blebs orpores, which characterized the spleens andlymph nodes of FLV-infected mice, were seen ineither the bone marrow or thymus preparationsduring the first 14 to 21 days after infection.Bone marrow cells from noninfected mice weresmoother and larger than splenocytes; many ofthese cells had fewer and shorter villi than thehighly villous cells seen in the spleen and nodes.In many respects those normal cells resembledthe smoother cells that were seen in FLV-in-fected spleens. Both cell types seemed to beblast cells, one apparently occurring naturallyand the other perhaps as a result of virus infec-tion. There was little indication that the cells inthe bone marrow were changed by FLV infec-tion. Comparing bone marrow at 20 to 30 days

INFECT. IMMUN.

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SEM OF FLV-INFECTED LYMPHOID CELLS 819

I B r

FIG. 1. Scanning electron micrographs of spleen cells from (A) normal BALBIc mice showing villous andsmooth spleen cells and from FL V-infected mice injected at various times earlier with a 10-1 dilution of virus;(B) spleen cells from 5-day infected mice with moderate cellular changes, including fewer villous cells and onemoderately spongy cell; (C) spleen cells from 10-day infected mice, many of which are larger and deformed,showing altered surface topography; (D) spleen cells with smooth ruffled surfaces and several showing spongytopography from 15-day infected mice; (E) spleen cells from 20-day infected mice showing large deformedcells; and (F) splenocytes from 30-day infected mice showing marked abnormalities (x5,600).

postinfection with marrow cells from normalmice revealed few, if any, differences. Similarly,the surface morphology of thymus cells in in-fected mice was not much different from that ofnormal thymocytes. Thymocytes from normalmice were smoother than splenic or lymph node

lymphocytes. Occasionally, a villus-covered cellwas noted.Changes in lymphocyte populations with

distinctive surface antigens. Cells with dis-tinctive surface antigens reacting with anti-thetaor anti-mouse immunoglobulin sera are demon-

VOL. 20, 1978

iioa --

I'..

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

INFECT. IMMUN.820 FARBER, SPECTER, AND FRIEDMAN

TABLE 1. Changes in spleen cell populations ofBALBIc mice after infection with FLV

Avere|Antibody response Percent splenocytes staining for': SEM topography

Time after in- Avleragefiction (days) A(mg) PFC per Agglutinin Surface Theta an- FLV anti- Intermedi- V

spleen (x 103) titer globulin tgen gen ate

0 (controls) 135 ± 20 97.5 ± 7.0 1:256 38.2 34.5 <1.0 18 12 705 162 ± 39 12.5 ± 3.0 1:96 24.3 28.5 12.9 31 40 2910 520 ± 95 4.5 ± 1.1 1:24 14.3 19.5 76.2 39 50 1115 1,275 ± 210 1.1 ± 0.6 1:4 6.2 11.1 81.5 55 41 420 1,764 + 340 0.6 ± 0.1 1:2 3.1 9.9 87.5 78 19 325 2,560 ± 810 0.5 ± 0.2 1:2 2.5 6.2 92.3 99 <1 <130 2,750 ± 765 <0.1 <1:2 1.0 4.2 >95 99 <1 <135 2,810 ± 820 <0.1 <1:2 1.0 1.4 >95

a Groups of mice were infected by intraperitoneal injection of a 10-' dilution of FLV (500 50% infective doses)on day indicated before assay; 5 to 10 mice were tested at each time interval.

b Average antibody response for 5 to 10 mice 4 days after intraperitoneal immunization with 4 x 10' SRBC.PFC, Plaque-forming cells.

c Average percentage of spleen cells positive for indicated surface antigens when stained with appropriatefluorescein-labeled antisera.

d Average percentage of cells with indicated surface morphology as determined by SEM.

~~~~~~I-FIG. 2. Large smooth-surfaced spleen cells characteristic ofFLV infection. (A) Large spongy cell with large

numbers of holes and deformities. (B) Cell with several blebs or knobs and small numbers of holes, obtained15 to 20 days after infection.

strable by immunofluorescence or cytotoxicityassays in vitro. These markers are considered bymany investigators characteristic for T and Blymphocytes, respectively. Such immunoflu-orescent procedures were used to determine thepercentage of cells with these surface markers inthe spleens and lymph nodes of BALB/c miceat varying times after infection. As can be seenin Table, 1, the percentage of splenocytes flu-orescing in the presence of anti-mouse immu-noglobulin serum decreased rapidly after infec-

tion. This decrease paralleled the decrease inthe percentage of villus-covered lymphocytes.However, by day 10 to 15 after infection, spleno-megaly became quite pronounced, indicatingthat the decrease in the percentage of immuno-globulin-positive cells could be due, at least par-tially, to the rapid increase in the number ofspenocytes without surface immunoglobulinrather than an absolute decrease in positive cells.Nonetheless, it is evident that although the totalnumber of spleen cells, as determined by direct

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SEM OF FLV-INFECTED LYMPHOID CELLS 821

FIG. 3. Scanning electron micrographs of lymph node cells from (A) normal BALB/c mice and from miceinfected at various times earlier with FLV: (B) lymph node cells from 5-day infected mice; (C) lymph nodecells from 10-day infected mice, showing moderate changes with several cells having smoother surfaces andruffled or abnormal appearance; (D) lymph node cells from 15-day infected mice showing large numbers ofabnormal cells; (E) lymph node cells from 20-day infected mice showing moderate numbers of holes andsponginess; and (F) lymph node cells from 30-day infected mice, many of which are large with ruffled surfacesand abnormal appearance (x5,600).

hemacytometer count, increased during the first15 days after infection, the number of lympho-cytes with immunoglobulin increased only two-fold, representing an absolute reduction. Simi-larly, the percentate of cells with the theta sur-face antigen markers also decreased in the spleenof infected animals. This decrease also paralleled

the increase in spleen weight and cellularity. Incontrast to the decrease in the percentage ofthese cell types in the spleen, there was a mod-erate increase in the lymph nodes (unpublisheddata). No significant change in the cellularity oflymph nodes was evident after infection. Never-theless, within 5 to 10 days after FLV infection,

VOL. 20, 1978

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

INFECT. IMMUN.822 FARBER, SPECTER, AND FRIEDMANA. i~

ALd

FIG. 4. Scanning electron micrographs of bone marrow cells from (A) normal Balbic mice, and from mice

at various times after infection with a 10' dilution ofFL V; (B) bone marrow cells from mice infected 5 days

earlier; (C) bone marrow cells from mice infected 15 days earlier, most similar to control marrow cells; (E)

bone marrow cells from 20-day infected mice showing some larger cells with ruffled surfaces; and (F) bone

marrow cells from 30-day infected animals showing some cells with ruffled enlarged surfaces and some with

spongy appearance (x5,600).

the percentage of immunoglobulin-positive cellsin the nodes increased about 20 to 40%. Simi-larly, the percent of theta-positive cells also in-creased rapidly after infection.Appearance of FLV-associated antigens

on lymphoid cells. Immunofluorescent assayswere also performed to determine the numberand percentage of lymphoid cells with FLV-as-

sociated antigens on their surfaces (Table 1). Forthis purpose, anti-FLV serum conjugated withfluorescein was used to stain spleen, lymph node,bone marrow, and thumus cells at various timesafter infection. Fewer than 1.0% of these cellsfrom uninfected mice reacted with the anti-FLVserum. However, within 5 days after infectionover 10% of splenocytes reacted positively (Ta-

.. I

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SEM OF FLV-INFECTED LYMPHOID CELLS 823

11_. ~~iA.. ..

I F

%..27 e .)o. - I

A

~~~~~~~~~A..Bv~~~~~~~~..

-". .

FIG. 5. Scanning electron micrographs ofthymus cells from (A) normal BALB/c mice and from mice at (B)5 days, (C) 10 days, (D) 15 days, (E) 20 days, and (F) 30 days after FLV injection, showing very few changes incell morphology or type (x5,600).

ble 1). The number and percentage of these cellsincreased rapidly thereafter, so that by 10 to 15days postinfection about 80% of the splenocyteswere positive for FLV antigen by immunofluo-rescence (Table 1). Five days after infection,about 3 to 4% of the lymph node cells werepositive and by days 10 to 15 postinfection 20 to30% of the cells were positive (unpublisheddata). Even by week 3 to 4 after infection, only40 to 50% of the cells were positive. Fewer than1% of bone marrow or thymus cells reacted with

anti-FLV serum during the first 2 weeks afterinfection. By the end of week 3 to 4, 10 to 20% ofthe bone marrow cells and 5 to 10% of thymuscells stained positive with the FLV antiserum(unpublished data).Immune response in the spleen and

lymph nodes of FLV-infected mice. Earlierextensive studies have correlated the suppres-sion of the immune response to SRBC antigensand development of splenomegaly and othersigns of the disease. More than 95% of the anti-

VOL. 20, 1978

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

824 FARBER, SPECTER, AND FRIEDMAN

body-forming cells present following primaryimmunization with SRBC are found in thespleen, with the remainder in the lymph nodes.As can be seen in Table 1, within 5 days afterFLV infection there was a 70 to 80% depressionin the number of splenic antibody-producingcells. An even greater depression became evidentas the infection progressed and splenomegalyincreased. The plaque-forming response bylymph node cells, which accounted for less than5% of the total number of plaque-forming cellsafter immunization of mice with SRBC, was notsignificantly reduced until 10 to 20 days or longerpostinfection. Hemagglutinin titers were alsomarkedly reduced during the progression of leu-kemia (Table 1).

DISCUSSIONThe results of the present study with the SEM

are consistent with and extend results of earlierstudies in this laboratory by TEM and SEMconcerning ultrastructural changes of spleencells from FLV-infected mice (4, 15, 16). In thosestudies, a marked alteration of spleen cell mor-

phology occurred. This correlated with impair-ment of antibody formation by spleen cells,which constitute the major hemolysin-formingcell type after a single intravenous or intraperi-toneal injection of SRBC into normal mice. Theresults of the present investigation with SEMprovided further insights regarding the morpho-logical surface alterations of lymphoid cells fromthe spleen, lymph node, thymus, and bone mar-

row of mice at varying times after infection withFLV as well as changes in the proportions oflymphoid cells with distinctive cell surfacemarkers, including appearance of virus-associ-ated antigen.The earlier TEM studies had shown that

lymphoid cells rich in ribosomal particles andendoplasmic reticulum became the predominantcell type in the spleen by the end of week 1 afterinfection (15, 16). Intracellular and surface par-

ticles characteristic of C-type virus were de-tected in these cells. Budding virus particlesappeared to be associated only with lympho-blastic or erythroblastic cells and not with ma-

ture plasmacytes or lymphocytes, considered tobe the main source of antibody. In addition,examination of individual antibody-secretingcells obtained from the center of hemolyticplaques revealed that over 50% of the antibody-forming cells from infected animals had C-typevirus budding from their surfaces. The numbersof such antibody-producing cells, however, were

reduced over 90% in FLV-infected mice. It wasunlikely from such observations that depressionof antibody formation at the single-cell level was

due merely to a competition between virus andantigen for a finite number of precursor cellswhich could mature into active antibody-secret-ing cells. It was postulated earlier that the smallnumber of antibody-forming cells present in ananimal after infection with FLV and challengewith SRBC represented those immunocyteswhich had escaped virus infection (8, 9, 24). Thepresence of budding virus from individual im-munocytes appears to negate this hypothessis.SEM has been used to examine FLV-infected

cells by deHarven et al. (6). They demonstratedthat a continuous line of FLV-transformedmouse cells were homogenous in appearance,with surface blebs or knobs thought to be bud-ding virus. Similar surface modifications weredescribed for splenocytes from FLV-infectedmice in this laboratory (25). It should be noted,however, that there is considerable doubtwhether lymphocyte surface morphology as de-termined by SEM examination of cell suspen-sions can be utilized to differentiate T from Bcells. Initially, in mice, Polliack and associatesreported a close correlation between the numberof villous versus smooth cells in the spleen ofnormal mice and the percentage of theta anti-gen- versus immunoglobulin G-positive cells asdetermined by appropriate immunofluorescentassays (21). Those authors used a direct fixationprocedure in which Ficoll-prepared lymphocyteswere fixed with glutaraldehyde before stainingand preparation for SEM examination. How-ever, examination of human thymus cells thatform direct rosettes with SRBC indicated thatrosette-forming cells, though presumably T cells,developed many villi (17). Similarly, under ap-propriate cultural conditions with serum-en-riched medium, smooth-surfaced cells consid-ered to be T lymphocytes may develop villipreviously considered characteristic for B cells(5, 26). Recently Thurman et al. reported thathomozygous nude mice, which were functionalas well as histologically athymic, neverthelesshad as many smooth-surfaced cells in theirspleens as did normal mice (25). Thus, it seemsclear that current SEM criteria cannot be usedto distinguish T from B cells.A major goal of the present SEM study was

to characterize the alterations, if any, of lymph-oid cell surface morphology and surface antigensafter FLV infection and, if possible, to correlatethese alterations with development of immuno-suppression. Cell suspensions obtained from thespleens, lymph nodes, bone marrow, and thymusof FLV-infected mice were compared with thesame cell preparations from normal mice.Marked changes were evident in the spleen,including reduction in the percent of villus-cov-

INFECT. IMMUN.

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SEM OF FLV-INFECTED LYMPHOID CELLS 825

ered lymphocytes. Despite the reports of Pol-hack et al., it was not clear by SEM examinationwhich, if any, cell class these represent. It seemsunlikely that these cells are T lymphocytes,since immunofluorescent studies with anti-thetasera showed a decrease in their number also.Furthermore, fluorescent antibody studies withanti-FLV serum showed a parallel increase inthe number of cells expressing virus antigen onthe surface. Thus, these smooth-surfaced cells,especially those with blebs and knobs, could bevirus-transformed cells, either those derivedfrom hemopoietic precursor cells that infiltratedthe spleen or transformed B-lymphocyte precur-sor cells implicated earlier as major target cellsfor FLV. Many of the cells seemed similar tothose observed by deHarven et al. in vitro for acontinuous cell line transformed by FLV (26).

Alterations in the surface morphology of cellsin the lymph nodes, which are considered asecondary site for FLV replication, progressedat a slower rate. No significant enlargement oflymph nodes or increased numbers of lymphnode cells occurred until late in the disease.Furthermore, no significant change in surfacelymphocyte configuration was noted until afterweek 1 following FLV infection. By the end of 2weeks, some smooth cells with surface blebsbecame evident. These did not, however, occurin significant numbers until 25 to 35 days; evenat this late date normal-appearing villous cellswere still present, a situation which contrastedsharply with that in the spleen.The bone marrow and thymus cell population

showed little evidence of surface modificationsas a result of FLV infection, at least early aftervirus infection. The normal bone marrow con-tained mostly large, smooth cells which, in manyrespects, resembled some normal cells present inthe spleens of noninfected mice. Even thoughfluorescent staining showed that up to 30% ofthese cells were infected with FLV (i.e., positivefor FLV-associated antigen) by week 4 afterinfection, SEM examination revealed only slightchanges. Similar findings were obtained with thethymus; however, less than 10% of these cellsstained with anti-FLV serum even late in thedisease process.The presence of "spongy" cells late in the

infection, particularly in the spleen, could beinterpreted several ways. The pores in the mem-brane may represent the points at which thevirus buds have pinched off. Alternatively, theymay represent cells with membrane damage asa result of immune lysis. It is possible that thehost immune response to new cellular antigensrepresented by budding virus may not be com-pletely blocked by virus infection and that com-

plement-fixing immunoglobulins or cytolyticlymphocytes developing in other lymphoid or-gans of the mouse might interact with the virus-infected immunocytes in the spleen, resulting inimmunological damage. It is now widely ac-cepted that immune reactions against immuno-cytes infected with a variety of viruses mayresult in both functional and morphologicaldamage to those cells. Thus, it is possible thatsplenocytes infected with virus may be the targetof an immunological attack by other immuno-cytes which have not yet been infected. How-ever, it is necessary to perform additional im-munological studies, including examination ofmice for cytolytic antibody or immunocytes, be-fore this interpretation can be accepted.Changes in lymphoid cell populations in FLV-

infected mice were also evaluated by enumera-ting the percentage of cells with surface immu-noglobulin or those carrying theta antigen.Within a few days after infection, there was amarked reduction in the number of cells withsurface immunoglobulin, as well as cells withtheta antigen. This reduction only occurred inthe spleen and represented an initial decrease incell numbers and then a further dilution by theproliferation of blast cells which were negativefor these markers (presumably "null" cells). By20 to 35 days after infection, when splenomegalywas maximal due to an increase in the totalnumber of cells, there was an absolute increasein the number of surface immunoglobulin- andtheta-positive cells which, if calculated as thepercentage of splenocytes, actually representeda decrease. In contrast, lymph node cells showeda continuous increase in cells positive for bothsurface immunoglobulin and theta surface anti-gens.The changes among the lymphoid cell popu-

lations as detected by both SEM and fluorescentmicroscopy appeared to correlate directly withthe immunosuppression induced by the virus.Studies in progress with in vitro-infected splen-ocytes indicate that relatively similar changesoccur when splenocyte cultures are infected withcell-free FLV. At the time that the in vitrocultures of spleen cells lose their ability to de-velop normal immune responsiveness to SRBC,there is marked alteration in the surface mor-phology ofthe cultured cells as well as a decreasein the percentage of cells exhibiting normal sur-face immunoglobulin markers. Thus, the resultsof these studies suggest a direct relationshipbetween infection and alteration of lymphoidcell type and morphology. However, it is evidentthat further studies are necessary to establishthat the event(s) that causes transformation ofcells in the spleen, resulting in morphological

VOL. 20, 1978

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

826 FARBER, SPECTER, AND FRIEDMAN

alterations, is the same as the event which altersimmune responsiveness. SEM studies in prog-ress with enriched B- and T-lymphocyte popu-lations and the individual components of theFLV complex should provide more informationrelating cellular alterations and immunosuppres-sion.

LITERATURE CITED

1. Alexander, E. L., and B. Wetzel. 1975. Human lympho-cytes. Similarity of B and T cell surface morphology.Science 188:732-734.

2. Anderson, T. F. 1951. Techniques for preservation of 3-dimensional structure in preparing specimens for elec-tron microscopy. Trans. N.Y. Acad. Sci. 13:130.

3. Bassin, R. H., N. Tuttle, and P. J. Fischinger. 1971.Rapid cell culture assay technique for murine leukemiaviruses. Nature (London) 229:564-566.

4. Chang, G., M. W. Rancourt, W. S. Ceglowski, and H.Friedman. 1968. Leukemia virus suppression of anti-body forming cells: ultrastructure of infected spleen.Science 159:437-439.

5. Criswell, B. S., R. R. Rich, J. Dardano, and S. I.Kimzey. 1975. Scanning electron microscopy of normaland mitogen-stimulated mouse lymphoid cells. Cell.Immunol. 19:336-348.

6. DeHarven, E., N. Lampen, T. Sato, and C. Friend.1973. Scanning electron microscopy of cells infectedwith a murine leukemia virus. Virology 51:240-243.

7. Dent, P. 1972. Immunodepression by oncogenic viruses.Prog. Med. Virol. 12:1-35.

8. Friedman, H., and W. S. Ceglowski. 1971. Nature andmechanism of tumor virus induced immunosuppression.Prog. Virol. 1:815-825.

9. Friedman, H., and W. S. Ceglowski. 1973. Cellularimmunity and leukemia virus infection, p. 299-320. InW. S. Ceglowski and H. Friedman (ed.), Virus tumori-genesis and immunogenesis. Academic Press Inc., NewYork.

10. Friedman, H., and J. R. Kateley. 1975. Lymphocytesurface receptors and leukemia virus induced immuno-suppression. Am. J. Clin. Pathol. 63:735-747.

11. Friend, C. 1957. Cell-free transmission in adult Swissmice of a disease having the character of a leukemia. J.Exp. Med. 105:307-318.

12. Gross, L. 1970. Oncogenic viruses, 2nd ed. PergamonPress, London.

13. Kateley, J. R., J. Holderbach, and IL Friedman. 1974.Leukemia virus induced alteration of lymphocyte Ig

INFECT. IMMUN.

surface receptors and the "capping" response of mousespleen and lymph node cells. J. Natl. Cancer Inst.53:1135-1140.

14. Kateley, J. R., I. Kamo, G. Kaplan, and H. Friedman.1974. Suppressive effect of leukemia virus infectedlymphoid cells on in vitro immunization of normalsplenocytes. J. Natl. Cancer Inst. 53:1371-1378.

15. Koo, G. C., W. S. Ceglowski, and H. Friedman. 1971.Immunosuppression by leukemia viruses. V. Ultrastruc-tural studies of antibody forming spleen cells of miceinfected with Friend leukemia virus. J. Immunol.106:799-814.

16. Koo, G. C., W. S. Ceglowski, M. Higgins, and H.Friedman. 1971. Immunosuppression by leukemia vi-ruses. VI. Ultrastructure of individual antibody formingcells in the spleens of Friend leukemia virus infectedmice. J. Immunol. 106:815-830.

17. Lin, P. S., A. G. Cooper, and H. H. Wortis. 1973.Scanning electron microscopy of human T-cell and B-cell rosettes. N. Engl. J. Med. 289:548-551.

18. Linthicum, D. S., S. Sell, R. M. Wagner, and P. Trefts.1975. Scanning immunoelectron microscopy of mouse Band T lymphocytes. Nature (London) 253:173-175.

19. Notkins, A. L., S. E. Mergenhagen, and R. J. Howard.1970. Effect of virus infections on the function of theimmune system. Annu. Rev. Microbiol. 24:525-538.

20. Polliack, A., S. M. Fu, S. D. Douglas, Z. Bentwich, N.Lampen, and E. deHarven. 1974. Scanning electronmicroscopy of human lymphocyte sheep erythrocyterosettes. J. Exp. Med. 140:146-158.

21. Polliack, A., U. Hammerling, N. Lampen, and E.deHarven. 1975. Surface morphology of murine B andT lymphocytes: a comparative study by scanning elec-torn microscopy. Eur. J. Immunol. 5:32-39.

22. Polliack, A., N. Lampen, B. D. Clarkson, E. de-Harven, Z. Bentwich, F. P. Siegel, and H. G. Kun-kel. 1973. Identification of human B and T lymphocytesby scanning electron microscopy. J. Exp. Med.138:607-624.

23. Rauscher, F. J. 1962. A virus induced disease of micecharacterized by erythrocytopoiesis and lymphoid leu-kemia. J. Natl. Cancer Inst. 29:515-532.

24. Salaman, M. H. 1969. Immunodepression by viruses.Antibiot. Chemother. 15:393-406.

25. Thurman, G. B., P. Baur, and A. L. Goldstein. 1975.Examination of lymphocyte membranes of athymic"nude" mice by scanning electron microscopy. Ann.N.Y. Acad. Sci. 249:154-165.

26. Van Ewijk, W., N. H. C. Brons, and J. Rzoing. 1975.Scanning electron microscopy of homing and recircu-lating lymphocyte populations. Cell. Immunol.19:245-261.

on July 25, 2020 by guesthttp://iai.asm

.org/D

ownloaded from