Genitourin Med 1986;62:240-6

Morphological studies of the cytotoxicity ofTrichomonas vaginalis to normal human vaginalepithelial cells in vitro

S E RASMUSSEN,* M H NIELSEN,t I LIND,4t AND J M RHODES*

From the *International Escherichia and Klebsiella Centre, Statens Seruminstitut, the tInstitute ofPathological Anatomy, Copenhagen University, and the $Neisseria Department, Statens Seruminstitut,Copenhagen, Denmark

SUMMARY In vitro cultured monolayers of normal human vaginal epithelial cells were incubatedwith axenic cultures of Trichomonas vaginalis. Two strains freshly isolated from patients withtrichomoniasis and one strain that had been maintained in axenic culture for several years werestudied.

Freshly isolated trichomonads showed amoeboid movements, adherence to epithelial cell surfaces,and were cytotoxic to epithelial cells in vitro. In contrast, the laboratory strain maintained for years inaxenic culture did not adhere to the epithelial cell monolayer and was only cytotoxic at a concentra-tion 100 times that of freshly isolated trichomonads.

Electron microscopy of monolayers exposed to T vaginalis for 24 hours showed that all epithelialcells in intimate contact with trichomonads had more or less disintegrated, whereas in monolayersexposed for six hours some of the epithelial cells in contact with T vaginalis were normal.T vaginalis organisms with amoeboid morphology contained a dense network ofmicrofilaments in

the part of the trichomonad that was in contact with an epithelial cell. Occasionally a pseudopodiumwas projected into the cytoplasm of disintegrated epithelial cells.

Introduction

Trichomoniasis, a disease of the urogenital tractcaused by Trichomonas vaginalis, affects both menand women and is the most prevalent protozoan infec-tion in man. The symptoms of trichomoniasis mayvary, either because different strains of T vaginalisappear to have different pathogenic capabilities orbecause of variation in the host response. ' Forexample, though trichomoniasis in women is usuallycharacterised by a copious, evil smelling discharge,about 25% of women harbouring trichomonads have

Address for reprints: Dr J M Rhodes, Intemational Escherichia andKlebsiella Centre, Statens Seruminstitut, Artillerivej 5, DK-2300Copenhagen S, Denmark

Accepted for publication 24 February 1986

no symptoms at all.2Despite the high incidence oftrichomoniasis, little is

known about the mechanism of the onset of the severeinflammation and cytopathogenic changes that can becharacteristic of the disease.' Several reports havebeen published, however, on the cytopathogenic effectof freshly isolated T vaginalis on various types ofmammalian cells or human cancer cells in tissueculture. 3-6To investigate the cytotoxic effect of different

strains of freshly isolated T vaginalis in vitro, wedecided to explore the use of normal human vaginalepithelial cells as target cells. In this report we describemorphological studies on the cytotoxicity of freshlyisolated T vaginalis compared with T vaginalis strain171 1, which was previously used in extensive studieson the ultrastructure of T vaginalis,7 towards primarycultures of human vaginal epithelial cells.

240

Cytoxicity ofT vaginalis to normal human vaginal epithelial cells in vitro

Materials and methods

STRAINS OF PARASITET vaginalis strain 1711, originally described byNielsen,7 was obtained from a cryopreserved stock inthe Neisseria Department, Statens Seruminstitut.T vaginalis strain 130354 was freshly isolated fromthe genital tract of a patient with severe symptoms oftrichomoniasis, and Tvaginalis strain 10 was obtainedas a fresh isolate from the Neisseria Department. Allstrains of T vaginalis were maintained in Diamond'smedium8 containing agar and were transferred dailyinto fresh Diamond's medium with agar. At least oneweek before the experiments the trichomonads weretransferred to Diamond's medium without agar. Thegrowth characteristics of Tvaginalis strains 171 1 and130354 in this medium were assessed by counting thenumber of trichomonads in a haemocytometer atdifferent times within 50 hours after inoculation ofthe medium.

EPITHELIAL CELLSVaginal epithelial tissue was obtained from the vaginalcuff that is removed with the uterus at operation onpatients requiring a hysterectomy for non-malignantdisease. Immediately after surgery the tissue wasplaced in Eagle's minimum essential medium(EMEM) supplemented with 10% fetal calf serum,0-01% streptomycin, and 0-002% neomycin(EMEM+). On arrival at the laboratory, the epithelialtissue was dissected from the connective tissue. Theepithelial sheet was cut into cubes (8 mm3), which wereplaced in chicken embryo extract (Gibco, Europe).Small drops of chicken plasma (Gibco, Europe) wereplaced on the bottom of each well of a Linbro tissueculture plate (24 wells). The epithelial explants werethen transferred to the drops ofplasma, and the plasmawas allowed to clot. When the clot had formed 2 mlEMEM+ was added to each well, and the plate wasincubated at 37°C in an atmosphere of 5% carbondioxide. The medium was changed after one week andthereafter twice a week. A confluent sheet of epithelialcells was seen after four weeks.

MORPHOLOGICAL STUDIES OF MONOLAYERSAFTER INOCULATION WITH T VAGINALISConfluent monolayers were washed twice withEMEM, and 2ml fresh EMEM+ was added.Trichomonads were harvested by centrifugation at 400x g for five minutes, resuspended in Diamond'smedium without agar, and counted in ahaemocytometer.

Light microscopyThe following concentrations of T vaginalis wereused: 104, 105, 106, and 10' per ml Diamond's mediumwithout agar, 100 ,u ofthese suspensions was added to

the epithelial monolayers. All experiments wererepeated three times, and triplicate estimations werecarried out within each experiment. The Linbro plateswere incubated at 37°C for 24 hours in an atmosphereof 5% carbon dioxide. The condition of the cellsthroughout the incubation period was observed byphase contrast microscopy. After incubation theepithelial monolayers were washed with phosphatebuffered saline (PBS), pH 7-0, and fixed and stainedwith Wright's stain. Cytotoxicity was assessed asfollows: 0 = no damage to the epithelial cells; 1 = 10-25% dettruction; 2 = 25-50% destruction; 3 = 50-75% destruction; and 4 = 75-100% destruction.

Electron microscopyMonolayers of epithelial cells were inoculated with100 lul of 106/ml T vaginalis organisms 6-24 hoursbefore they were processed for electron microscopy.Each culture had 0-5 ml Karnovsky's fixative9 added,and five minutes later this was replaced by Kamovsky'sfixative dilutred 1:2 with distilled water. After onehour the cultures were washed in 0-1 molI cacodylatebuffer, pH 7 3, post fixed in 1% osmium oxide in 0-1mol/l cacodylate, pH 7-3, for three hours, washedbriefly, and dehydrated in graded ethanol followed by100% propylene dioxide. After this treatment theepithelial cell monolayer became loose. The cells weretransferred to a mixture of equal parts of propylenedioxide and epon. The cells were rolled up using anorange stick and embedded in Beam capsules with theroll standing upright.

In some experiments all fixatives, buffer solutions,and ethanol solutions up to 70% alcohol contained0 1% ruthenium red.

Ultrathin sections were post stained with uranylacetate and lead citrate.

Results

LIGHT MICROSCOPYGrowth characteristicsT vaginalis strain 130354 grew more slowly thanstrain 1711 in Diamond's medium without agar anddid not attain the same total number of trichomonads.The peak numbers reached 30 hours after inoculationwith 5 x 10 T vaginalis were about 4 x 106 for strain130354 and 10' for strain 1711. Thereafter strain1711 died rapidly and there were no viable cells 40hours after inoculation, whereas 60 hours wererequired before strain 130354 was dead.

Interaction of T vaginalis with monolayersTvaginalis 171 1 organisms retained their ovoid shapeafter addition to epithelial cell cultures, and movedabout actively, either singly or in clumps of up toseveral hundreds of trichomonads. Adherence to the

241

S E Rasmussen, MH Nielsen, I Lind, and JMRhodes

monolayers was rarely seen even after incubation for24 hours (fig 1(a), (b), and (c)). Only a few Tvaginalis1711 organisms were visible in the fixed stainedpreparations, most having been washed away.T vaginalis 130354 and 10 organisms also moved

about actively for the first half hour after inoculation,but after one hour these trichomonads settled on theepithelial cells singly and then showed amoeboidmovements with projecting pseudopodia. After 48hours they often grew into sheets partly covering themonolayer (fig 2(a), (b), and (c)). These two strainsadhered to the epithelial monolayer, and most of thetrichomonads still adhered to the epithelial cells afterbeing washed and fixed.The adherence of T vaginalis to the monolayers

seemed to be associated with their cytotoxic activity.Cell destruction caused by the freshly isolated strainswas thus initially confined to areas with adheringtrichomonads. The table shows that Tvaginalis strain171 1, which did not adhere to the monolayers, onlycaused a demonstrable cytotoxic effect towardsepithelial cells when 1 05 trichomonads were added perwell. In contrast, T vaginalis strain 130354 showed a

TABLE Cytotoxic effect of two different strains ofTrichomonas vaginalis on normal human vaginal epithelialcells

No of T vaginalis Cytotoxic effect * of strains:micro-organismsadded per well 1711 (laboratory) 130354 (fresh)

103 0 1l04, 0 3-4105 1 4106 3 4

*0 = no damage to the epithelial cells; I = 10-25% destruction; 2 =25-50% destruction; 3 = 50-75% destruction; 4 = 75-100%destruction.Experiments were carried out in triplicate. Figures are averages of twoidentical experiments carried out in triplicate.

visible cytotoxic effect at a concentration of 103trichomonads per well, and more than 103 organismsper well totally destroyed the monolayers. Thecytotoxic effect of strain 10 (data not shown) wassimilar to that of strain 130354.

FIG I Phase contrast light micrographs ofconfluent mono-layer of normal human vaginal epithelial cells(a) immediately, (b) 3'/ hours, and (c) 22 hours afterinoculation with 104 T vaginalis strain 1711. At all timesmost trichomonads (T) were aggregated and non-adherentand epithelial cells (EP) were undamaged.

242

Cytoxicity ofT vaginalis to normal human vaginal epithelial cells in vitro

FIG 2 Phase contrast light micrographs ofconfluent mono-layer of normal human vaginal epithelial cells(a) immediately, (b) 3% hours, and (c) 22 hours afterinoculation with J0" T vaginalis strain 130354.Immediately after inoculation some trichomonads adheredto epithelial cells and became amoeboid (A). After three anda hafhours mostparasites had adhered and were amoeboid(A) but epithelial cells (EP) were undamaged. After22 hourstrichomonads lay in sheets over large lesions (L) in theepithelial monolayer. Epithelial cells (EP), which seemed tobe normal, were present in unparasitised areas.

ELECTRON MICROSCOPYElectron miscroscopy confirmed the light microscopyfinding that strain 1711, in contrast with strains130354 and 10, was not retained on the monolayers ofepithelial cells after fixation. T vaginalis strain 17 11was never observed in intimate contact with epithelialcells, which as a rule showed normal ultrastructuralmorphology and intact cell membranes.T vaginalis 130354 and 10 organisms acquired an

amoeboid shape and were often flattened out inintimate contact with the surface of epithelial cells,with neighbouring T vaginalis, or with the plastic onthe bottom of the culture well (fig 3 and 4). They werealso observed on, between, or under the epithelial cells(figs 3 and 4).The gap between the cell membranes of

trichomonads and epithelial cells that were in closeproximity varied, but was never less than 20 nm. At thesite of contact the cytoplasm of the trichomonads con-sisted almost exclusively of a network of microfila-ments (fig 5). Sometimes endocytic vesicles withengulfed material were noted. Ruthenium red stainedthe surface of T vaginalis irrespective of whether thetrichomonad was in intimate contact with other cells orthe plastic on the bottom ofthe culture well or was freein the medium (fig 4).

Vaginal epithelial cells incubated with freshlyisolated T vaginalis were affected to different degreesdepending on the number ofT vaginalis organisms andthe period of incubation. Six hours after inoculationmany epithelial cells in intimate contact with Tvaginalis still had a normal cell membrane andcytoplasmic ultrastructure (fig 4), whereas after 24hours the outer membrane ofthe epithelial cells in con-tact with T vaginalis was broken and the cytoplasmonly contained a few organelles (figs 3 and 5).

Discussion

Several reports have indicated that T vaginalis can becytotoxic towards cells in tissure cultures.' 3-6 Thesetarget cells originated either from different animalspecies or from human tumours. Our studies using

243

S E Rasmussen, MH Nielsen, I Lin4 and JM Rhodes

FIG3 Electron micrograph ofvaginal epithelial cell culture broken and cytoplasmic organelles were present in24 hours after inoculation with T vaginalis strain 130354. intracellular space (arrows). CG denotes chromaticNoteflattened trichomonads (7) on top ofand intercalated granules (hydrogenosomes) and V denotes large foodbetween epithelial cells (EP). Epithelial cell membrane was vacuoles in trichomonads.

normal human vaginal epithelial cells showed that astrain of T vaginalis (130354) freshly isolated from apatient with severe symptoms and a fresh isolate(strain 10) caused extensive disruption ofthe epithelialcell monolayer, whereas a strain (171 1) maintained inaxenic cultures for several years did not. This may bedue to the fact that strain 1711 had been passagedfrequently in Diamond's medium, which is known tocause a loss in pathogenicity.'The cytopathogenicity of strains 130354 and 10,

which was confined to focal areas, was clearlyassociated with the fact that the trichomonads settledon the epithelial layer and became attached to the cells.This result agrees with those of other authors,' 3-6 whoobserved that physical contact between the parasiteand target cells was necessary for cell destruction invitro. It is still not known, however, whether thecytotoxic mechanism is "mechanical" or is due tomembrane bound lytic factors. A recent report on thehaemolytic activity oftrichomonads might support thelatter hypothesis.'0

It has also been suggested that chemical factorsreleased into the culture medium by trichomonads maybe responsible for or augment cell destruction. "I In ourexperiments, however, the cell destruction was focaland some normal epithelial cells always seemed to bepresent after exposure to the pathogenic strains. Inaddition, supernatants from strains 1711 and 130354

of trichomonads used in the present studies had noeffect on the human vaginal epithelial cells (ownunpublished observations). Similar observations havebeen made by others.3 6 This implies that it is unlikelythat release of a toxin into the culture medium isresponsible for the destruction of the epithelial cellmonolayers. Furthermore, ifa toxin is produced by theprotozoa, then it is either present at a very low level,acts locally, or is inhibited by the medium.The electron microscopic observation that freshly

isolated T vaginalis organisms with amoeboidmorphology attach to monolayers of human vaginalepithelial cells in vitro is similar in most respects to theresults of earlier electron microscopic studies onvaginal biopsy specimens from patients withtrichomoniasis. 12 The only difference noted was that invaginal biopsy specimens trichomonads wereexclusively attached to necrotic superficially locatedcells, whereas in cultures inoculated in vitro with Tvaginalis intimate contact between trichomonadsand apparently unaltered epithelial cells was observedafter six hours. This may imply that the cytotoxicity ofTvaginalis towards epithelial cells is a "slow" processrequiring several hours of contact with the targetcell.

Intimate contact in vitro between T vaginalis andepithelial cells was apparently required for cytotox-icity to occur, but the fine structure of the zone of

244

Cytoxicity ofT vaginalis to normal human vaginal epithelial cells in vitro

FIG 4 Electron micrograph of cell culture six hours qfter FIG 5 Electron micrograph of cell culture 24 hours afterinfection with T vaginalis strain 10. Trichomonad cells (7) inoculation with T vaginalis strain 130354. T vaginalisin intimate contact with a detached undamaged epithelial pseudopodia (P) project into cytoplasm (EP) ofpartiallycell (EP) on bottom of well (B). Uniform ruthenium red disintegrated epithelial cell. MI denotes microfilaments, Nstaining (arrows) ofall cell surfaces implied that there was nucleus, GO golgi zone, and CG chromatic granulesno tight junction contact between closely apposed cell (hydrogenosomes) of T vaginalis.membranes.

contact gave no indication of how this might takeplace.

Inoculation of monolayers in vitro, as descrnbedhere, may provide a model system for studying thepathogenicity of T vaginalis. Other local factors pre-valent in the vagina, however, such as the influx ofgranulocytes, the release of enzymes, and the effect ofother micro-organisms, must be taken into account to

explain the onset of the severe inflammatory responsethat is characteristic of the disease.

We are indebted to Dr Lebech, head of the gynaecologicaldepartment, Frederiksberg Hospital, for the specimens ofvaginal tissue, and to Dr Mary Andersen, specialist ingynaecology, for the T vaginalis isolated from a patient withsevere trichomoniasis.

245

S E Rasmussen, MH Nielsen, I Lind4 and JM Rhodes

References

1. Honigberg BM. Trichomonas of importance in humanmedicine. In: Kreier JP, ed. Parasitic protozoa Vol 2. NewYork: Academic Press, 1978:275-454.

2. Krieger JN. Urologic aspects of trichomoniasis. InvestigativeUrology 1981 ;18:411-7.

3. Farris VK, Honigberg BM. Behaviour and pathogenicity ofTrichomonas vaginalis Donne in chick liver cell cultures. JParasitol 1970;56:849-82.

4. Alderete JF, Pearlman E. Pathogenic Trichomonas vaginaliscytotoxicity to cell culture monolayers. British Journal ofVenereal Diseases 1984;60;99-105.

5. Heath E. Behaviour and pathogenicity of Trichomonasvaginalis in epithelial cell cultures. British Journal ofVenerealDiseases 1981;57:106-17.

6. Christian RT, Miller NF, Ludovici PP, Riley GM. A story ofTrichomonas vaginalis in human cell culture. Am J ObstetGynecol 1963;85:947-54.

7. Nielsen MH. The fine structure of cells of Trichomonasvaginalis Donne obtained from the exponential phase ofgrowthand from stationary culture. Acta Pathol Microbiol Scand B1976;84:205-16.

8. Diamond LS. The establishment of various trichomonads ofanimal and man in axenic cultures. J Parasitol 1957;43:488-98.

9. Kamovsky MJ. A formaldehyde glutaraldehyde fixative of highosmolarity for use in electron microscopy. J Cell Biol1965;27:137A.

10. Krieger JN, Poisson MA, Rein MV. Beta-hemolytic activity ofTrichomonas vaginalis correlates with virulence. InfectImmun 1983;41:1291-5.

11. Hogue MJ. The effect of Trichomonas vaginalis on tissueculture cells. American Journal ofHygiene 1943;37:142-52.

12. Nielsen MH, Nielsen R. Electron microscopy of Trichomonasvaginalis Donne. Interaction with vaginal epithelium in humantrichomoniasis. Acta Pathol Microbiol Scand B 1975;83:305-20.

246