JOURNAL OF BACrERIOLOGY, Mar., 1966Copyright © 1966 American Society for Microbiology

Vol. 91, No. 3Printed in U.S.A.

Immunofluorescence in Cells Derived from Burkitt'sLymphoma

GERTRUDE HENLE AND WERNER HENLE1Virus Laboratories, Children's Hospital of Philadelphia, and School of Medicine, University of Pennsylvania,

Philadelphia, Pennsylvania

Received for publication 18 October 1965

ABSTRACT

HENLE, GERTRUDE (Children's Hospital of Philadelphia, Philadelphia, Pa.), ANDWERNER HENLE. Immunofluorescence in cells derived from Burkitt's lymphoma. J.Bacteriol. 91:1248-1256. 1966.-Indirect immunofluorescence tests led to thebrilliant staining of a small proportion of the cells in five different cultures derivedfrom Burkitt's (African) lymphomas. The reaction was not restricted to the 17 serafrom cases of this disease but extended to many sera from American individuals,whether healthy donors or patients suffering from a variety of illnesses. The inci-dence of positive sera increased with age from about 30% in childhood to >90% inadults. Fluorescein-isothiocyanate-conjugated human 7-globulins were suitable fordirect staining of the same proportion of cells. The stained cells appeared to be invarying stages of degeneration, but cultural conditions leading to an increase in thecellular death rates failed to result in a rise in fluorescent cells. Several observationssuggest that the stainable cells might be those which are seen to harbor virus par-ticles under the electron microscope. Two cell lines derived from leukemic patientsin this country also contained a small fraction of stainable cells but two others, andnumerous primary human leukocyte cultures, gave consistently negative results.Attempts to relate the staining to known viral antigens have failed to implicateherpes simplex, varicella, cytomegalo, and reo viruses types 1, 2, and 3. The natureof the virus carried by the lymphoma cells as well as of the staining reactions remainsto be determined.

A number of cell lines have been isolated fromBurkitt's (African) lymphomas which are lympho-blastic in character and grow freely in culturemedia without attachment to the surface of theglass vessel (6-9, 18, 22; Epstein, Barr, andAchong, Wistar Inst. Symp. Monograph No. 4,in press; Rabson et al., Intern. J. Cancer, inpress). Five of these six lines have been examinedelectron microscopically, and the presence ofvirus particles in a small proportion of the cellshas been reported (5, 8, 22; Epstein et al., inpress; Rabson et al., in press). These particlesresemble herpes simplex virus in shape but seemto be somewhat smaller in size. They were seenin the nuclei and also in the cytoplasm of cells.Infected cells showed as a rule, some degree ofdegeneration (10). There is an indication that thevirus particles may disappear, or at least decreasein number, on prolonged cultivation of the cells(Epstein, personal communication; Rabson et al.,

1Career Investigator, 1-K6-AI-22,683, NationalInstitutes of Health, U.S. Public Health Service.

in press). Efforts to transmit the virus to othertypes of cultures, experimental animals, or chickembryos by routine procedures thus far havefailed (10). Several of the lines revealed suggestiveevidence of a viral infection by means other thanelectron microscopy in that they were noted tobehave like certain viral carrier cultures (12;Henle and Henle, Wistar Inst. Symp. MonographNo. 4, in press); i.e., they showed a considerabledegree of resistance to the usually highly cyto-pathic vesicular stomatitis virus (VSV); they werecapable of transmitting resistance to certainhuman and simian cells on mixed cultivation;and they released into the culture media a factorpossessing all the attributes of an interferon.With the availability of sera from patients withBurkitt's lymphoma, it became possible to studythe various cell lines by the indirect immuno-fluorescence and complement-fixation techniques.Positive results were obtained in both tests, andthey were not restricted to sera from Africanpatients with this disease but extended to many

1248

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 91, 1966 IMMUNOFLUORESCENCE OF BURKIrT LYMPHOMA CELLS

sera from individuals residing in the UnitedStates. Neither were positive reactions obtainedsolely with the cell lines derived from cases ofBurkitt's lymphoma, but also with certain linesisolated from leukemic patients in the UnitedStates. The results of immunofluorescence will bepresented in this conmnunication; those obtainedin complement-fixation tests are reported sep-arately (1).

MATERIAL AND METODS

Cells. The various Burkitt lymphoma as well asother cell lnes used are lsted in Table 1 together withthe media employed in this laboratory for their culti-vation and optimal growth. (lhe lymphoma lines hadbeen in culture for 9 to 14 months at the start oftheseexperiments. Since they are maintained mainly bydilution in fresh medium, at most preceded by sedi-mentation of the cells, it seemed inappropriate tosignify each feeding by serial passage numbers.) Theywere kindly furnished by M. A. Epstein, Bland SuttonInstitute, Middlesex Hospital, London, England

(EB1, EB2, and EB3); Sarah Stewart (SL1) and A. S.Rabson (ALh), National Cancer Institute, Bethesda,Md.; J. T. Grace, Roswell Park Memorial Institute,Buffalo, N.Y. (64-10); and B. Clarkson, Sloan Ketter-ing lstitute, New York, N.Y. (SKL-1 and SKL-2).The LK-1D line was isolated in this laboratory.Pimary cultures of peripheral leukocytes from

healthy donors were prepared according to the methodemployed for chromosome studies (16). The method,involving stimulation by phytohemagglutinin, hasbeen slightly modified (Due-Nguyen and Henle, to bepublished). Human diploid cells (WI-38) were ob-tained from the Wistar Institute, Phadelphia, Pa.,and human embryonic kidney (HEK) cultures fromFlow Laboratories, Rockville, Md., or Microbio-logical Associates, Inc., Bethesda, Md. HeLa and

Earle's strain L cell cultures were prepared from stocklines maintained in this laboratory. Primary rabbitkidney cultures were kindly supplied by T. Tokumaruhere at the hospital.

Sera. For sera from patients with Burkitt's lym-

phoma, we are indebted to W. Ray Bryan, National

Cancer Institute, D. H. Wright, Makerere UniversityCollege Medical School, Kampala, Uganda, andM. A. Epstein. Sera from children with leukemia,solid tumors, and various metabolic and infectiousdiseases were furnished by staff members of TheChildren's Hospital of Philadelphia. Sera from infantswith cytomegalic inclusion disease and from motherswith or without complement-fixing antibodies to thisvirus were kindly supplied by R. M. McAllister,The Children's Hospital of Los Angeles, Los Angeles,Calif., and J. B. Hanshaw, School of Medicine, Uni-versity of Rochester, Rochester, N.Y. Sera from adultpatients with leukemias and cancers were provided bythe Virus Research Resources Branch of the NationalInstitutes of Health, Bethesda, Md. Other adult ser

were obtained from blood donors and medical stu-dents.

Fluorescein isothiocyanate (FITC)-conjugated 'y-

globulins. For the indirect staining technique (23),rabbit anti-human and anti-guinea pig globulin con-

jugates were prepared by conventional means as de-scribed elsewhere (13) or were purchased from Balti-more Biological Laboratories, as were labeled goatanti-rabbit and anti-rhesus monkey globulins. Fordirect staining (3), preparations ofcommercial humanev-globulin or of human mumps hyperimmune -y-glob-ulin (Philadelphia Serum Exchange) were coupledwith fluorescein isothiocyanate (19). The latter y-

globulin was obtained from donors who had beenimmunized against mumps by two doses ofa fonnalin-inactivated vaccine. The donors had been subjectedthereafter to plasmapheresis every 2 weeks for severalmonths, and the plasmas so obtained revealed acontinued high titer of antibodies to mumps (21).The FTC-conjugated 'y-globulins prepared from thepooled plasmas had been stored at 4 C since 1959.Rabbit anti-herpes simplex virus conjugates wereprovided by K. Munk, Virologisches Institut derUniversitit Heidelberg, and monkey anti-murineleukemia (Rauscher) virus conjugate was kindlysupplied by M. A. Fink, National Cancer Institute,Bethesda, Md.

Preparation of cell smears. Suspended cell cultureswere subjected to centrifugation at 600 X g for 10 to

TABLE 1. Cell lines and culture media

Cell line Source Authors Medium

EB1 Burkitt's lymphoma Epstein and Barr (6) BME (45%), MEM (45%),FCS (10%)

EB2 Burkitt's lymphoma Epstein et al. (5) BME (90%/0), FCS (10%)EB3 Burkitt's lymphoma Epstein et al (in press) No. 1629 (90%), FCS (10%O)SL1 Burkitt's lymphoma Stewart et al. (22) BME (90%/), FCS (10%)AL, Burkitt's lymphoma Rabson et al. (in press) No. 199 (80%), FCS (20%O)64-10 Myeloblastic leukemia Ikawata and Grace (14) No. 1629 (90%O), FCS (10%)SKL-1 Acute monocystic leukemia Clarkson (to be published) No. 1629 (90%), FCS (10%)SKL-2 Acute lymphoblastic leuke- Clarkson (to be published) No. 1629 (90%), FCS (10%)

miaLK-1D Acute lymphocytic leukemia Armstrong (to be pub- No. 1629 (90%), FCS (10%)

lished)

* BME, Basal Medium Eagle; MEM, minimal essential medium (Eagle); FCS, fetal calf serum.

1249

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

HENLE AND HENLE

15 min. The supematant fluids were drained off, andthe cells were resuspended in the remaining drops withthe aid of a finely drawn-out Pasteur pipette. A smalldrop of this suspension was placed on a cover slip(6 by 30 mm), evenly distributed over the surface,and allowed to dry thoroughly at 37 C. The coverslips were fixed at room temperature with acetone for10 niin, drained, and dried. The fixed preparationswere kept at -20 C until staining was performed.

Other cells (WI-38, HEK, and HeLa) were grownon cover slips in test tubes and infected with herpessimplex, varicella, or cytomegalo (Davis strain) vir-uses. When lesions became evident, the cover slipswere removed, washed in phosphate-buffered salinesolution (PBS), dried, and fixed as described.

Staining of cover slips. For the direct technique,the cover slips were overlayed with FITC conjugatesand incubated in a moist chamber for 1 hr at 37 C.Subsequently, they were washed twice for 10 min inpetri dishes containing PBS, agitated by a magneticstirrer, and then dipped several times into two con-secutive beakers containing distilled water. Afterdrying, the cover slips were fixed onto slides by meansof a semipermanent mounting medium (20) and ex-amined under ultraviolet illumination (Osram HBO200-w lamp) with a Zeiss microscope.

For indirect staining, the cover-slip preparationswere first exposed to serum at a dilution of 1:8 orgreater for 1 hr at 37 C. After thorough washing, theywere overlaid with FITC-conjugated antibodies toglobulins of the appropriate species and processedfurther as described.

Vital staining, by both the direct and indirectmethods, differed from the above techniques only in-sofar as sedimented unfixed cells were taken up insmall volumes of the reagents, incubated for 1-hrperiods at 37 C and washed by repeated sedimenta-tion and resuspension in PBS. The final suspensionswere spread on cover slips, which, after drying, weremounted on slides with or without prior fixation inacetone.

Other techniques are described below.

RESULTS

Indirect immunofluorescence. In the first test,plasma obtained from a case of Burkitt's lym-phoma, treated at the Clinical Center of theNational Institutes of Health, and a serum from aleukemic American child as control were used in a1:8 dilution to stain EB1 and EB2 cells by theindirect method. With the patient's plasma, apple-green fluorescence was noted in about 1% of thecells. As shown in Fig. 1A and B, the staining wassomewhat granular, often revealing larger ag-gregates within the confines of the cells. Many ofthe cells showed irregular outlines and appearedto be in varying states of disintegration. Thecontrol serum failed to mediate any fluorescence;neither did the rabbit anti-human globulinconjugate per se stain any cells except for oc-casional faint stippling of the cells of the type seenin Fig. 1A.

The indirect immunofluorescence techniquewas subsequently applied to sera from additionalpatients with Burkitt's tumor as well as to nu-merous sera from American individuals of varyingages, and also to other Burkitt lymphoma celllines. The data concerning the reactivity of thevarious sera in a 1 :8 dilution are summarized inTable 2. All of the 17 sera from Burkitt lymphomapatients gave positive results, but indirectimmunofluorescence was elicited also with manysera from American individuals. About 30% ofthe sera from children reacted regardless ofwhether they were taken from patients withleukemia, metabolic diseases, or infectiousdiseases. The incidence of positive staining reac-tions was distinctly higher (up to >90%) withsera of adult persons which were obtained eitherfrom patients with malignancies or from healthydonors.A number of sera were titrated by the indirect

immunofluorescence technique as well as incomplement-fixation tests (1). The results (Table3) revealed that sera which were strongly positivein a 1: 8 dilution could be diluted as much as 64-fold and still yield definite, though weak, fluores-cence. The intensity of staining at low serumdilutions (despite the subjective nature of thegrading) seemed to relate well to the titersmeasured by the staining technique but not tothose obtained in complement-fixation tests.Yet, some degree of correlation between the twoassay methods seems to be apparent from thesummary presented in Table 4, based merelyupon positive or negative results at a serumdilution of 1:8. Of the 96 sera studied by bothmethods, the results agreed in 72% of the cases;i.e., either both tests were positive or both werenegative. However, 24% of the sera yieldedpositive fluorescence in the absence of detectablecomplement-fixing activity and 4% failed to stainbut reacted in the complement-fixation test.

In control experiments, no staining wasobtained when the EB cells were exposed first tosera from several adult rabbits, guinea pigs, orrhesus monkeys and then to FITC-conjugatedantibodies to homologous or human globulins.

Direct immunofluorescence tests. The frequencyof positive reactions with sera of adult personsin the indirect test rendered it likely that FITC-conjugated pooled human y-globulin would besuitable for the direct method of staining. Threesuch preparations were available, and all threefulfilled the expectation. Two derived from com-merically produced normal y-globulin gavedistinct staining of moderate intensity. The thirdwas prepared from pooled plasmas obtained bybiweekly plasmapheresis from a few donors whohad been vaccinated against mumps. Since fre-

1250 J. BACTEURIOL.

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 91, 1966 IMMUNOFLUORESCENCE OF BURKITIT LYMPHOMA CELLS

FIG. 1. Immunofluorescence elicited in cells derived from Burkitt's lymphoma. (A) Indirect stain ofEB2 cellsby serum from a patient with Burkitt's lymphoma and FITC-conjugated rabbit antibodies to human globulin. (B)The same except with serum from a healthy donor for the primary reaction. (C) Direct staining ofEBI cells byFITC-conjugated human mumps hyperimmune y-globulin. (D) Direct stain ofEB2 cells by FITC-conjugated nor-mal human y-globulin. X 1,250.

TABLE 2. Incidence ofpositive reactions in indirect immunofluorescence tests with EBI or EB3 cells

Primary serum FluorescenceNo. of individuals

Age group Diagnosis + A - Per cent +

Children Burkitt's lymphoma 17 17 0 0 100Leukemias 16 5 0 11 31Solid tumors 7 3 0 4 43Other diseases 32 9 1 22 31

Adults Leukemias 10 5 1 4 60Carcinomas 15 14 0 1 93Normal 33 26 2 5 85

quent plasmapheresis is known to maintain highlevels of viral antibodies (21), this Py-globulinconjugate was suitable for staining of mumps aswell as several other viral antigens (see below),and elicited brilliant fluorescence also in a frac-

tion of the EB cells (Fig. IC and D). The reactionwas blocked by prior exposure of the fixed cellsmears to human sera which gave positive reac-tions in the indirect test, but not by negative sera(Table 5). Sera from rhesus monkeys, rabbits,

1251

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

TABLE 3. Titration of sera in indirect immunofluorescence tests with EB3 cells

Serum | Fluorescence with primary serum diluted

fixing titerNo. Diagnosis 1:8 1:16 1:32 1:64 1:128 1:256

P-91 Burkitt's lymphoma 1:128 ++ + + - -

B-2239 Burkitt's lymphoma 1:32 +++ ++ + (+) (±)B-220 Burkitt's lymphoma 1:8 ++ + ± (:1) -B-1251 Burkitt's lymphoma 1:8 ++ ++ + (+) (±B-18l19 Burkitt's lymphoma 1:32 +++ ++ + (+) ±E-2 Burkitt's lymphoma act ++ ++ + (+) ±E-4 Burkitt's lymphoma ac + ± (±-) - _R000503 Erythroleukemia 1:256 ++ ++ + (+) ±R000222 Adenocarcinoma 1:64 ++ + ± - _SGD55 Healthy donor 1:16 + + ± - -R000197 Reticulum cell sarcoma <1:8 (4±)64-680 Adeno infection <1:8 _LK-13 Acute lymphocytic leukemia <1:8 _LK-14 Acute lymphocytic leukemia <1:8 - _

* Symbols: +++ to (±) = decreasing degrees of brilliance.t Anticomplementary serum.

guinea pigs, calves, and horses failed to blockstaiing by the human y-globulin conjugates.The labeled mumps hyper mune y-globulinwas employed predominantly in further analysisof the staining reaction.

Relation of degenerating ana dead cels to thestaining reaction. As pointed out earlier, theinimunofluorescent cells often appeared to be in astate of degeneration, and therefore the possibilityhad to be ruled out that dead cells present in thecultures adsorbed the labeled proteins non-specifically. Although the percentage of fluores-cent cells was generally smaler than that of deadcells as determined by staining with trypan blue,a more definitive analysis was needed. Conse-quently EB cells were subjected to variousprocedures which increase their death rates: (i)reduction of the serum concentration in themedia to 1% or nil; (ii) crowding of the cellsduring cultivation in the tips of pointed centri-fuge tubes; and (iii) incubation of the cultureswithout refeeding for up to 2 weeks. Even thoughsuch preparations contained from 30 to 80%trypan blue-staining ceLs, immunofluorescencewas elicited only in the usual low percentage ofthe cells.

In several experiments, cells were stainedvitally; i.e., without prior fixation. The resultswere similar to those obtained by the routinetechnique which are shown in Fig. 1. The stainedcells again showed, as a rule, evidence of degen-eration. It was impossible, therefore, to decidewhether this fluorescence was due to surfacesta g or to penetration of the labeled proteinsinto the degenerating cells. In addition, the faintstippling of some of the cells, noted on occasionafter staining of fixed preparation (Fig. 1A) was

somewhat more pronounced on vital staining,especially by the indirect technique. Indeed, thestippling (but not the intense staining of the smallproportion of the cells) was seen after vitalstaining solely by the anti-human globulinconjugate. It clearly outlined the circumferenceof the cells and seemed to cover also the cellsurface.

Comparison of various cell lines. The fiveBurkitt lymphoma cell lines as well as variousother cells were compared in direct as well as in-direct immunofluorescence tests. The results aresummarized in Table 6. In the direct tests, allBurkitt lymphoma lines contained a variableproportion of cells which stained with the labeledhuman mumps hyperimmune and normal y-globulins. The percentage of staiing cells washighest in EB3 preparations, followed in decreas-ing order by the EBI, EB2, SL1, and AL1 cells.This order appears to match the reported easewith which virus particles have been detected insectioned cells under the electron microscope (5,8; Epstein et al., in press). The AL1 line originallycontained sufficient virus particles to be readilyfound, but by the time the line was received nonewas detected (Rabson et al., in press). Two of theother lines yielded positive iunofluorescencein a very low percentage of the cells; i.e., the SKL-1 and SKL-2 lines which were isolated fromAmerican patients with leukemias (Clarkson, tobe published). It has not been reported whetherthese cultures contain virus particles. The stain-ing was blocked by prior exposure of the SKLcells to human sera which gave positive resultsin the indirect tests with EB cells, but not bynegative sera (Table 5). The 64-10 line, which atone time harbored a virus (14), and the LK-1D

1252 HIENLE AND HENLE J. BAcTERioL.

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 91, 1966 IMMUNOFLUORESCENCE OF BURKIlT LYMPHOMA CELLS

TABLE 4. Comparison ofresults obtained in comple-ment-fixation and indirect immunofluorescence

tests

Complment Indirect Per centComplnt fluor No. of sera of total

+ + 34 35.4+ _ 4 4.2_ + 23 24.0_ _ 35 36.4

Total 96 100.0

TABLE 5. Blocking by various sera ofdirect staining ofEB3, SKL-1, and SKL-2 cells by human mumps

hyperimmune -y-globulin conjugate

Staining by labeledBlocking serum human mumps hyper-

immune 7y-globulin

Species No. Indirect EB3 SKL-l SKL-2staining cells cells cells

Human B-2239 ++ (=It) -B-1145 ++ (t) - (i)B-1819 ++ (41:) - (d)PK ++ (±) - (:)LK-6 - ++ + +LK-15 - ++ + +LK-33 - ++ + +LK-43 - ++ + +

Rhesus monkey 3* NTt ++Rabbit 5 NT ++Guinea pig 4 NT ++Horse 4 NT ++Calf 4 NT ++

_ ,~~ ++ +_ +

* Number of animal sera tested.t Not tested.

line gave consistently negative results, and so didnumerous cultures of normal peripheral leuko-cytes, stimulated by phytohemagglutinin. Thelast type of cultures was examined within 1 to 16days of incubation. Various routine cell cultures(HeLa, WI-38, HEK, primary rabbit kidney,and Earle's strain L) likewise failed to revealpositive cells.The results of direct staining were matched by

those obtained with the indirect technique, withthe use of a positive serum for the primaryphase of the reaction. The controls, omitting theprimary serum, revealed that the rabbit anti-human globulin conjugate alone elicited fluores-cence in some of the cell lines. As mentionedearlier, the three types of EB cells showed at mosta fine stippling of moderate intensity under theseconditions, especially on vital staining. Similarresults were obtained with the SL, and the AL1

TABLE 6. Direct and indirect staining of various celllines

Cells

EB1.......EB2.......EB3.......SL1 .....AL1.......64-10......LK-.D....SKL-1 ....SKL-2 ....Primaryleuko-cyte cul-tures....

HeLa.....WI-38.....HEK .....Rabbitkidney..

L.........

Direct Indirecttet() test0tet() (b)

++++

+

+

+N+d+

NDND

Per centof cellsrositive(a or b),range

1-30.5-2

5-10ca. 1ca. 0.100

0.1-0.50.1-0.5

0

000

00

Controlof

indiotest° (c)

-e0

+

Per centof celspositive

(c),range

000005-10

ca. 1010-505-20

<1

000

00

- Positive human serum plus FITC-conju-gated rabbit antibodies to human globulin.

b Only FITC-conjugated rabbit antibodies tohuman globulin.

¢ Except for occasional fine surface stipplingof some of the lymphoma cells.

d Not done.

TABLE 7. Staining ofvarious viral antigens by FITC-conjugated human mumps

y-globulinhyperimmune

Virus Cells Stainig

Mumps HeLa ++Herpes simplex HEK ++Varicella HEK (±)Cytomegalo WI-38ReoType 1 FLType 2 FLType 3 FL

Simian virus 40 HEKInfluenza A HeLa + (S and V antigen)

lines. The stippling appeared to be a surfacereaction. In contrast, many of the 64-10, SKL-1,SKL-2, and LK-1D cells, and a few of the cellsin peripheral leukocyte cultures, revealed granularcytoplasmic staining of moderate to markedintensity when exposed to rabbit anti-humanglobulin conjugates. Rabbit anti-guinea pigglobulin conjugates gave very weak staining, anot unexpected cross-reaction (13). The per-

1253

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

HENLE AND HENLE

centage of stained cells in the four lines mentionedvaried to some extent from test to test. Thenature of this type of reaction is still obscure,especially since it is blocked significantly, if nottotally, by prior exposure of the cells to humansera or globulin fractions.

Attempts to relate the immunofluorescence oflymphoma cells to viral antigens. For the indirectmethod of staining of EB cells, sera were usedwhich had been collected from patients infectedwith herpes simplex, varicella or herpes zoster,and cytomegalic inclusion disease viruses, aswell as from individuals with known antibodylevels to these agents, as determined by comple-ment fixation or virus-specific immunofluores-cence. In each category, only a proportion of thesera gave positive results in line with the expectedpercentages for given age ranges.The labeled mumps hyperimmune y-globulin

was tested against a number of viral antigens(Table 7). It stained herpes simplex antigen(s)brilliantly, whether present in human embryonicor rabbit kidney cells. The conjugates failed tostain varicella virus-infected HEK cells, exceptfor a barely detectable dull-green coating, in-sufficient to account for the brilliant fluorescenceelicited in the lymphoma cells. No staining wasobtained of appropriate cells infected by cyto-megalovirus (Davis strain), simian virus 40, orreovirus types 1, 2, and 3. In addition, attemptswere made to stain lymphoma and some of theother cell lines with FITC-conjugated rabbitantibodies to herpes simplex virus or the reo-viruses with negative results. The tests concern-ing reovirus were kindly performed by R.Spendlove, Viral and Rickettsial Disease Lab-oratory, California State Department of PublicHealth, Berkeley. In addition, no specific stain-ing was noted with labeled monkey antibodiesversus murine leukemia (Rauscher) virus (11).None of the above tests appeared to link any

of the viruses studied to the immunofluorescenceelicited in the lymphoma cells. Nevertheless, sincethe FITC-conjugated human y-globulins stainedherpes simplex antigens; additional tests werecarried out for a final exclusion of this agent. Themumps hyperimmune conjugate was absorbedwith concentrated herpes simplex virus growneither in rabbit or human embryonic kidneycultures. Volumes of 1 ml of the conjugate weremixed with an equal volume of either centrif-ugally concentrated herpes simplex virus orHanks' solution, incubated at 37 C for 1 hr andovernight at 4 C, and centrifuged thereafter at100,000 X g for 1 hr. The supematant fluids,representing a twofold dilution of the originalconjugate, were used for staining of EBI, EB3,herpes simplex virus-infected and normal HEK

TABLE 8. Absorption of FITC-conjugated humanmumps hyperimmune y-globulin by herpes sim-

plex virus (HSV)

Staining of cells

FITC-conjugate Absorbent HEKEB1 EB3

Infected ControlHSVCotl

Human None + + +mumps HSV (rabbit + + -

hyperim- kidneymune HSV (HEK) + + -

Rabbit None _ - +anti-HSV

or rabbit kidney cells. The results (Table 8)showed that the absorbed conjugates no longerstained HSV antigens, but the reaction with theexpected proportion ofEB cells was unaffected.

DIscussIoNThe immunofluorescence obtained with the

various cell lines tested appears to fall into at leastthree categories.The first type, the brilliant fluorescence ob-

served in a low percentage of the lymphoma cells(Fig. 1), was obtained by the indirect methodonly with certain human sera and anti-human7y-globulin conjugates. Direct staining by FITC-coupled pooled human y-globulins was blockedby human sera which gave positive reactions inthe indirect test but not by negative human oranimal sera. It would seem that this type ofreaction is due to specific antibodies to an as yetunidentified antigen(s).The second type, the cytoplasmic fluorescence

of moderate to marked intensity elicited byFITC-conjugated rabbit antibodies to humanglobulins only in cells of the lines derived fromvarious leukemias and of normal leukocytecultures, has not been analyzed, and its nature isat present obscure. Aside from the fact that thepercentage of staining cells in given lines wasvariable from test to test, this type of staining wasunexplainably reduced or totally blocked byprior exposure of the cells to human sera orglobulin, and therefore is assumed to be due tononspecific factors.The third type of fluorescence likewise has not

been clarified; i.e., the fine surface stippling ofmoderate intensity (Fig.IA) observed occasionallyespecially on vital staining of lymphoma cells inthe indirect tests with human sera, or in directtests with human y-globulin or rabbit anti-human globulin conjugates. It might be due to

1254 J. BACTERIOL.

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 91, 1966 IMMUNOFLUORESCENCE OF BURKIIT LYMPHOMA CELLS

nonspecific adsorption of labeled proteins or toisoantibodies. In any event, it is readily dif-ferentiated from the first type of staining reactionwhich has been the main subject of this report.

It is tempting to ascribe the brilliant immuno-fluorescence elicited in a small proportion of thelymphoma cells to the virus carried by a fractionof the cells in these cultures. Indeed, severalobservations support this possibility. (i) Thereseems to be a rough correlation between the per-centage of stainable cells in the five cell linesderived from Burkitt's lymphoma and the per-centage of cells which harbor virus particles asdetermined by electron microscopy. Although theelectron microscopic studies do not lend them-selves readily to accurate quantitation, about 1to 2% of the EBI and EB2 cells were estimatedto harbor virus particles (5, 8; Epstein et al., inpress). The percentage of immunofluorescentcells in these two lines has fluctuated in numeroustests between 1 and 3 and between 0.5 and 2%,respectively. In the EB3 line, which contains morevirus-infected cells than the first two (Epstein etal., in press), the stainable cells have rangedbetween 3 and 10%. The percentage of stainablecells in the AL1 line was in the order of 0.1 whencells containing virus particles could no longer befound (Rabson et al., in press). (ii) The cellsrevealing virus particles, especially when foundin the cytoplasm, exhibited marked degrees ofdegeneration (10), as did the cells showing im-munofluorescence. Finally (iii), exposure of EBIand EB3 cultures to 5-methylamino-2'-deoxy-uridine (MADU), kindly supplied by M. R.Hilleman, Merck Sharp & Dohme, Inc., West-point, Pa., reduced significantly the percentageof stainable cells within 7 to 10 days (Henle andHenle, to be published). MADU was shownrecently to be of low toxicity for cultured cellsand effective in the treatment of cultures infectedwith herpes simplex virus (15). These pointsprovide only indirect evidence, and final proofof the identity of the stainable and virus-carry-ing cells must still be furnished.

If the above interpretation should turn out to becorrect, the virus present in the lymphomacultures would not be restricted to the regions ofprevalence of Burkitt's tumor but would be wide-spread also in the United States on the basis of theindirect immunofluorescence test. The increasein the incidence of reactive American sera fromabout 30% in children to more than 90% withadvancing age of the serum donors resembles thegradual acquisition by populations of antibodiesto common viruses. The significance of the factthat positive immunofluorescence was obtainedwith all of 17 sera from African children sufferingfrom Burkitt's lymphoma cannot be assessed

because no sera have been available from otherAfrican children for comparison.

Needless to say, it is still unknown whether thevirus present in the lymphoma cell cultures isetiologically related to Burkitt's tumor orrepresents merely a passenger. The finding of thesame kind of virus particles in five cell linesisolated from four patients in three differentlaboratories tends to favor the first alternative,but, obviously, further cell lines need to beisolated and examined for the presence of virusparticles and inumunofluorescence. The widedistribution of antibodies, presumably to thisvirus, would not necessarily preclude an etio-logical relationship, since some oncogenic animalviruses are widespread among their respectivehosts, evoking immune responses but rarelytumor formation. Furthermore, cases resemblingBurkitt's lymphoma have been reported recentlyto occur in the United States (4, 17). Finally, twocell lines isolated from American leukemicpatients (SKL-1 and SKL-2) have revealed avery small percentage of cells, which were stain-able by the same procedures as employed for thelymphoma cultures. As in the case of EB cells,staining by FITC-conjugated human y-globulinwas blocked by prior exposure of the SKL cellsto sera from patients with Burkitt's lymphomaand other sera which gave positive reactions inthe indirect immunofluorescence tests, but not bynegative sera. The presence of cell-associatedvirus particles in these two lines has not beenreported. They would be difficult to find under theelectronmicroscope if the assumption is correctthat mainly the few staining cells harbor theagent.

All attempts to relate the staining to knownviral antigens have failed. Morphologically, thevirus resembles the herpes group, and conse-quently efforts were concentrated upon herpessimplex, varicella, and cytomegaloviruses. Inaddition, reovirus was considered merely be-cause of the reported isolation of a member ofthis group from a Burkitt tumor (2). All theseagents seem to be excluded by the tests recorded.Only a proportion of sera from patients infectedwith these agents or from individuals with knownantibody levels to them elicited fluorescence inindirect tests with EB cells. The FITC-conjugatedhuman y-globulins failed to detect varicellavirus, cytomegalovirus (Davis strain), or reo-virus types 1, 2, and 3 antigens in appropri-ate, infected host cells. Although these conjugatesstained herpes simplex antigen(s) brilliantly, thisreactivity could be removed completely byadsorption with herpes simplex virus withoutaffecting the staining of EB cells. Efforts toelicit fluorescence in the lymphoma cells by

1255

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

HENLE AND HENLE

labeled rabbit antibodies to herpes simplex virusor the three types of reoviruses, likewise, gavenegative results. It would seem that the cellscarry an as yet unidentifiable virus which re-sembles in size and shape members of the herpesgroup.

AcKNowLEDGomNTSThis investigation was supported by Public Health

Service grant CA-04568 from the National CancerInstitute.The competent technical assistance of Marianne

Checkoff and Marie Adams is gratefully acknowl-edged.

ADDENDUM IN PROOF

Four additional Burkitt lymphoma cell lines, iso-lated by Drs. Pulvertaft and Osunkoya, were suppliedby Dr. P. Manaker of the National Cancer Instituteand subjected to direct and indirect immuno-fluores-cence tests without knowledge ofthe results ofelectronmicroscopic examinations. Two of these (P 2002 andP 2003) revealed from 1-2 and about 5% stainingcells, respectively; the other two, designated P 2004and P 2028 (Reji), were negative. These results again

corresponded roughly to the number of cells shown tocontain virus particles. The two negative cell lineshad been in culture for more than 3 years; the positivelines were 10 and 3 months old.

LruATu Crrw

1. ARmnRoNo, D., G. HE, AND W. HENLE. 1966.Complement-fixation tests with cell lines de-rived from Burkitt's lymphoma and acuteleukemias. J. Bacteriol. 91:1257-1262.

2. BELL, T. M., A. MASE, M. G. R. Ross, ANDM. C. WA±L4MS. 1964. Isolation of a reovirusfrom a case of Burkdtt's lymphoma. Brit. Med.J. 1:1212-1216.

3. COONs, A. H., AN M. H. KAPLAN. 1950. Localiza-tion of antigen in tissue cells. II. Improvementsin a method for the detection of antigens bymeans of fluorescent antibody. J. Exptl. Med.91:1-13.

4. DoamnsN, R. F. 1965. Childhood lymphosarcomain St. Louis, Missouri, clinically and histo-logically resembling Burkitt's tumor. Cancer18:418-430.

5. EPsmy, M. A., B. G. ACHONG, AND Y. M. BARR.1964. Virus particles in cultured lymphoblastsfrom Buridtt's lymphoma. Lancet 1:702-703.

6. EPmiN, M. A., AND Y. M. BARR. 1964. Cultiva-tion in vitro of human lymphoblasts fromBurkitt's malignant lymphoma. Lancet 1:252253.

7. EPrmN, M. A., AND Y. M. BARR. 1965. Charac-teristics and mode of growth of a tissue culturestrain (EB1) of human lymphoblasts fromBurkitt's lymphoma. J. Natl. Cancer Inst. 34:231-240.

8. EPmIN, M. A., Y. M. BARR, AND B. G. ACHONG.

1964. A second virus-carrying tissue culturestrain (EB2) of lymphoblasts from Burkitt'slymphoma. Pathol. Biol. Semaine Hop. 12:1233-1234.

9. EPsTEIN, M. A., Y. M. BARR, AND B. G. ACHONG.1965. The behaviour and morphology of asecond tissue culture strain (EB2) of lympho-blasts from Burkitt's lymphoma. Brit. J. Cancer19:108-115.

10. EPSEIN, M. A., G. HENLE, B. G. ACHONG, ANDY. M. BARR. 1965. Morphological and bio-logical studies on a virus in cultured lympho-blasts from Burkitt's lymphoma. J. Exptl. Med.121:761-770.

11. FINK, M. A., R. A. MALMGREN, F. J. RAUSCHER,H. C. Ona, AND M. KARON. 1964. Applicationof imunofluorescence to the study of humanleukemia. J. Natl. Cancer Inst. 33:581-588.

12. HENLE, G., AND W. HENLE. 1965. Evidence for apersistent viral infection in a cell line derivedfrom Burkitt's lymphoma. J. Bacteriol. 89:252-258.

13. HENLE, G., AND W. HasLE. 1965. Cross-reac-tions among -y globulins of various species inindirect unofluorescence. J. Immunol. 95:118-124.

14. IWAKATA, S., AND J. T. GRACE, JR. 1964. Culti-vation in vitro of myeloblasts from humanleukemia. N.Y. State J. Med. 64-.2279-2282.

15. NEMEs, M. M., AND M. R. HwLEmAN. 1965.Effective treatment of experimental herpessimplex keratitis with new derivative, 5-methyl-amino-2'-deoxyuridine (MADU). Proc. Soc.Exptl. Biol. Med. 119:515-520.

16. NowELL, P. C. 1961. Phytohemagglutinin: aninitiator of mitosis in cultures of human leuko-cytes. Cancer Res. 21:1518-1521.

17. O'CoNoR, G. T., H. RAPPAPORT, AND E. B. SMITH.1965. Childhood lymphoma resembling "Bur-ldtt Tumor" in the United States. Cancer 18:411-417.

18. PuLvERTAr, R. J. V. 1964. Cytology of Burkitt'stumour (African lymphoma). Lancet 1:238-240.

19. RIoGs, J. L., R. J. SEiwALD, J. H. BURCKHALTER,C. M. DowNs, AND T. G. METcALF. 1958.Isothiocyanate compounds as fluorescentlabeling agents for immune serum. Am. J.Pathol. 34:1081-1097.

20. RODRIGUEZ, J., AND F. DENHARDT. 1960. Prepa-ration of a semipermanent mounting mediumfor fluorescent antibody studies. Virology 12:316-317.

21. SMOLENS, J., J. STOKES, JR., AND A. B. VOOT. 1957.Human plasmapheresis and its effect on anti-bodies. 3. Immunol. 79:434-439.

22. STEWART, S. E., E. LOvEACE, J. J. WHANG, ANDV. A. NGu. 1965. Burkitt tumor: tissue culture,cytogenetic, and virus studies. J. Natl. CancerInst. 34:319-327.

23. WELLER, T. H., AND A. H. CooNs. 1954. Fluores-cent antibody studies with agents of varicellaand herpeszoster propagated in vitro. Proc.Soc. Exptl. Biol. Med. 86:789-794.

1256 J. BACTRIOL.

on August 23, 2019 by guest

http://jb.asm.org/

Dow

nloaded from