detection and counting of cryptosporidium parvum in hct-8 cells by flowcytometry

Drrrcnox AND couNTtNG or CRyprospoRrDtuM pARvurvlII.I HCT-8 cElts BY FLoWcYToMETRY

SummaryrThe oblective of lhe present study wos lo evoluole flowcyromerryonolysis (FCA) os o lool for ropidly ond obiectively estimoting ihepercentoge of cells infected wilh C4yplosporidíum parvum in onin vllro model. We compored ihe results to those obloined withimmunofluorescence ossoy llFA) ond evoìuoted the iniro-osscyvoriobility of bolh ossoys cnd the inter-cssoy voriobility of lFA.Humqn ileocecol odenocorcinomc cells IHCT-81 were infectedwith different doses of excysted oocysts. After 24 hours, cells werecnclysed by FCA ond by IFA using o monoclonol onlibody thotrecognises o C. porvum ontigenic prolein ond q lectin thot bindswilh glycoproleins presenl in lhe porosilophorous vqcuoles, Thecoefficient of voriobility in terms of the percentcge of infecied cellswos lower for FCA (i.e , I 3-14 %) thon for IFA ii,e., 27-38 %when performed by o single operolor snd 19'22 % whenperformecl by three operotors), suggesting thot FCA is moreoccurcte, in ihct it is nol subiecl io operolor experlise, FCA olsohos ihe odvonloge of ollowing the enlire cullure io be exomined,thus ovoiding problems wiih heterogeneity omong microscopicfields. In light of these resulis, this method could olso be used totesl new ctnli-Cryptosporidrum drugs.

Résumé.'DÉrecrtoN ET colvrprAcE oe CnypTospoanruttr pARvuÌr ENcurruRr cELLULAIT (HCT-8) pArì cyroÀ,rÉTRrE DB FLux

Le developpement de lq cuhure in vilro de Cryplosporidiumporvum pose le problène de lo quontifícotion des cellulesinfectées par une nethode ropide et objective. Dons ce trovaÌ\,une onolyse en cytométrie de flux {CF) o élé mise ou poinl pourdétecter et quonfifier les cellules infecÌées por C, povum. Descellules d'un odénocorcinome iléo-coecol hunoin (HCT"8) ont etéinfeclées rn vilro por des doses dífférentes d'oocysles spc:,rulés.Après 24 h de culture, les cellules ont été onolysées en CF et porimmunofluorescence índirecte (lFl) en utilisant un onticorpsmonoclonel dirigé conlre une proléine de C. povum eÌ unet lectines'altochont ò des glycoproÍéines présentes dons /es vocuolesparositophores, Les coeffícienls de voriotion du nombre de cellulesinfectées étoienf compris entre 13 el 14 % pour une CF sur toutelo culture cellulaire, entre 27 el 38 % pour une défection en lFl sur20 chomps microscopiques por un seul opéroteur et entre l9 et22 % pour une détectíon en lFl sur /es mémes 20 chompsmicroscopiques par trois opéroleurs. Les foibles coefficients devoriotion obtenus por lo cytomélrie de flux suggèrent que celteméthode est plus précise el moins dependonte des compélencesde l'opéroteur que l'lFl.

KEY WORDS : Cryptospoiúum poÍvum, HCT-8 cells, flowcytomeky,monoclonol onlibody, in vllro.

INTRODUCTION

,{lryptosporidiunz paruutlx constitutes a majorÍ . cause of diarrhoeal disease in many species of\J mammals, including humans (Fayer et a|.,7997;O'Donogl-rue, 1995), yet little is known about the basicbiology of this organism, and no specific effective the-rapy for C. paruum infection exists. For many years,the srudy of C. paruum and of effective drugs has beenhampered by the lack of specific in uitro systems anda rapid fiìethod for evaluating the infection rate in ahost cell population, However, in recent years, ceil cul-tures for studying the life cycle and metabolic requi-

* Department of infectious, parasitic and immunomediated diseases,IstitLìto Superiore di Sanità, 00161 Rome, Italy.Correspondence: Edoardo Pozio, Department of infectious, parasiticand immunomediated diseases, Istituto Superiore di Sanità, VialeRegina Elena 299, 00út Rome, Italy,Tel,: +39 06 4990 2304 - Fax: +39 06 4938 7065,E-maiL pozio@iss.it

MIOTS CLÉS r Cryplosporidium poruum, cel/u/es HCT-|, cloméIie de flux,onlicorp monoclonol, ìn viko.

rements of C, paruutm have been developed (Arrc,woodet al,, 7994; Griffiths et al,, 1.994; Upton et al,, 7994a;Upton et al., 7994b; Upton et al., 7995), and it has beenproposed that these cell cultures be used instead ofanimal models for pharmacological screening (Voodset al., 1.995; Yang et al., 1996; Deng & Cliver, 1998),Furihermore, sevefal studies have quantified C, paruuminfection in uitro using immunofluorescence assay(IFA) (Slifko et al., 7999), enzyme linked immunosor-bent assay (Woods et al., 7995), or real-time pc,lyme-rase chain 1'eaction (Fontaine & Guillot, 2002).The objective of the present study was to evaluateflowcytometry analysis (FCA) as a tool for npidly andobjectively estimating the percentage oî C. paraunx-infected cells in an in uitro model, -ùle compared theresults to those obtained with IFA and evaluatr:d theintra-assay vatiability of both assays (i.e., che varia-bility among replicates analysed by a single ope-rator) and the inter-assay vaúabrlity of IFA (i.e,, thevaliability among replicates anaiysed by three ope-fatofs).

MATERIALS AND METHODS

Iu wTno CULTURE

uman ileoce cal adenocarcinoma ceils (HCT-8)were cultured in RPMI 1640 (Hyclone) sup-plemented with 5 o/o foetal calf serum (FCS,

Hyclone), 200 mM L-glutamine (Sigma), 1 % sodiumpyruvate (Sigma), 5 % penicillin, and 5 % streptomycin,and they were maintained in tissue-culrure flasks in a

5 o/o Co, atmosphere at 37o C and 85 % humidity' \X/henthe cells reached 80-90 % confluence, they were deta-ched by incubating them in PBS with 0,05 o/o Trypsinand 0,02 % EDTA (Er-rroclone) at 37o C in the CO,atmosphere for 5-8 minutes, The cell suspen.sìon wasvigorously and repeatedly pipetted using a 200 plpipette tip r-rntil the cells were separated. The cellswere then centfifuged at 500 xg for' 10 minutes; thepellet was resuspended in the maintenance medium(Upton et al,, 1.995).

Excvstatto N oF C Rwros PIRIDIUM PAR1,z1l4 oocYSTS

Crypt ospal"i diurn p a'ruurlx oocysts (isolate code ISSC4)were obtained from experimentally infected calvesafter faeces purlfication by sucrose and Percoll@ den-siry gradients (Rossi et al,, 7990), For excystation, theoocysts were resuspended in 10 mM HCI and incuba-ted at 37o C for 1,0 minutes, according to a previouslypublished protocol (Gut & Nelson, 1999). Briefly, thesuspension was centrifuged at 3,000 49 for five minu-tes, and the pellet was resuspended in 2 mM sodiumtaurocholate (Sigma) in PBS and incubated at 15o C for10 minutes and followedby a 5-8 minutes incubationat 37" C. The level of excystation was 90 o/o when 2-3week-old oocysts were used and decreased whenusing older oocysts; thus 2-3 week-old oocysts wereused throughout the study. This protocol allows for thesynchronous excystation of at leasi 108 oocysts in 1'5 mlmicrofuge tubes. For a smaller quantify of oocysts, thetubes are pre-coated with bovine serum albumln toreduce sporozoite loss. The excystation mk was cen-tlifuged at 1,000 xg for rwo rninutes. The excystedoocysts were resuspended, aliquoted in the gfosithmedium, and tt'ansferred into a Petli dish.Before the infection with excysted oocysts, the HCT-8cells were plated in Petri dishes at 1 x 10Ó cells/100 mmdiameter; the FCS concentration was increased to 10 0/0,

ancl the cells were grown to 80-90 7o confluence. TheHCT-S cells were infected with 0.1, 0.5, 1, 2, 3,5, 74,and 20 excysted oocysts pel cell, Nter 24 hours ofincubation, we were able to observe a direct correla-don berween the infective dose and the quantiry ofinfected cells, because parasites of the first generationgrown in uitro were not yet able to infect other cells(Slifko et al., 7999),

INornect IÀ4MUNoFLUoRESCENCE ASSAY

HCT-8 cells were plated on tisstte culture glass s'lides(Falcon) and grown to 80-90'%,confluence,. HCT-8 lllono-layels were then infected with 0,1, 0.5, or 1 excystedoocyst per celi, Nter 24 hours of incubration, ttre cel'lswere fixed in 4 o/o forrn-aldehy'de in PBS fo1' 20 minutesand treated with 1 0/o Triton X-100 in PBS fc,r 10'minu-tes and then with 2 % BSA' in PBS for 30 rr'rinutes. TnLe

fixed cells were incubated af room tempelatut'e for30 minutes with 0.5 pglml of a ,monoclonal a,r'ltibodythat recognises the SA35 antigen, which l'ras breenidentified in the sporozoite stage oÎ C. paruwn (Tosiniet al,, 1999) and also in the endocellular stages (unpu-blished data), and with 0.5 pglm1 of a trectin (\"fl,-biotin) (Vector Laboratories, 8-1235) that binds glyco-proteins present ìn the parasitophorous vacuoles (Gut& Nelson, 1999), Samples were then incubated withl pg/ffi of goat anti-mouse FITC and streptavidin-PE a't

room temperature for 30 minutes. As negative controls,non-infected cells and C. paruuru-infected cells incu-bated with FITC and PE conjugates wel'e used' Slideswere observed under epifluorescence mk:roscopy at400 X magnification, The level of infection was expres-sed as the mean number of parasitophoror:s vacuolesper field out of a total of 20 fields (which correspondto 14 0/o of the total well), Each experiment was per-formed in triplicate, For the inter-assay test, each expe-riment was carried out by three diffelent operators.

FrovcvtouETRY ANALYSIS

HCT-8 monolayers were infected with 0,1, 0.5, !,2,3,10, and 20 excysted oocysts per cell, Aftr:r 24 houlsof incubation, the cells were detached by tleatingthem with 0,05 o/o Trypsin and 0'02 % EDTA (Euro-clone) af 37o C In a CO, atmosphere, for ji-8 minutes.The cell suspension was vigorously and repeatedlypipetted using a 200 pl pipette tip until th'e cells wereseparated, The cel1s were fixed in 4 0/o formaldehydein PBS for 20 minutes and treated with 1 7o Triton X-100 in PBS for 10 minutes and then with 2 0/o BSA inPBS for 30 minutes. The fixed ce1ls were incubatedwith 0.5 pglm1 of anti-SA35 and 0,5 pglml of V\/I--biotin at room temperature for 30 minutes. Sampleswere then incubated with 1 pglm1 of goat anti-mouse-FITC and streptavidin-PE at room temperature for30 minuies. As negative controls, non-infected cel1s andC,. paruun't-infected ce1ls incubated with goat anfr'mouse-FITC and streptavidin-PE coniugates were used.The infected cells were analysed by a Facs Calibur(Becton Dickinson). The instrument was set up tomeasure forward-angle light scatter (FSC-H), side-anglelight scatter (SSC-H), and the fluorescence intensitiesof anti-SA35-FITC (FL1-H) and aWl-bjotin-strepta-vidin-PE (FL2-H), For each oocysts,/cell ratio, from 15

to 20 replicates were Performed.

Parasite, 2003, 10, 297 -302

RESULTS

ANarysrs or C, aARz-IM-TNFECTED cens ev IFAfS"lhe infection level of HCT-8 cells was monitoredT,I bV iFA r.rsing the anti-SA35-FITC and aWL-

l* biotin-streptavidin-PE, Botl-r SA35 and VW-biotinallowed infecteJ ceils to be distinguished from non-infected cells (data not shown) and the number ofinfected ce1ls increased with the infective dose, Sinceonly dor"rble stained cells are considered positives, thepLobability to get false positives diminished, Table Ishows the results of the intra-assay test, i,e,, the varia-biiity in the number of parasitophorous vacuolesamong the 20 fields examined (by a single operatol'and for each replicate), as determined by the standarddeviation of the mean and the coefficient of variabi-lity (CV). The variabiliff (consideled as an indicator of

the accr-rracy of the mean value) increased with theincreasing dose, with the CV ranging from 27'o/o forthe lowest close to 38 0/o fot the highest dose. Toconduct the inter'-assay test, for each replicate, thenumber of vacuoles in the same 20 fields are countedby three operators, The results (Table I) showed thatthe CV nnged flom 19 % for the lowest dose to 22 o/o

for the highest dose,

ANerysrs oF C, zARzzM-TNFECTED cELLS By FCAFor the morphometric analysis, the sample was ana-lysed with a threshold for FSC-H, in order to select onlythe HCT-S population (i.e,, discarding debris, freeoocysts, and sporozoites), Figure 1 shows the dotplots of the FCA of the FSC-H and SSC-H of non-infected cells (Fig. 1A) and of celis infected wiih anoocyst,/cell ratio of 0.5, 1, 3, and 10 (Figs 18, C, D,and E, respectively). For the infected cells, thr: light-

Intra-assay Inter.assay

Oocyst/cell ratio at infectionùlean number * SD of parasitophorous

vacuoles pel microscopic fieldCV (%o)

o.L/14xL.1

27

0.5/r33x12

36

1./158!22

38

0.1./1.

5!1.

20

0.5/r26!5

19

1/L53!12

22

Table I. - lntra-assay and inter'-assay tests for immunofluorescence analysis, Each expeliment was perfolmed in tt'iplicate by either oneoperator (intra-assay test) or by three operatols (inter-assay test) to count the numbel of parasitophofous vacuoles detected in IZO fields.Data ate expressed as the mean number + standard deviation (SD) and coefficient of variability (CV) of the triplicates.

@

+(oaìa)@o

\T

C\

Fig. 1. - Dot plots of the flowcyto-metry analysis of the forward-anglelight scatter (FSC-H) and the side-angle light scatter (SSC-H) of: (A) non-infected HCT-8 cells and of cellsinfected with a ratio of (B) 0.5, (C) 1,(D) 3, and (E) 10 Cryptospoúclitr.rnpaî'uLtnx excysted oocysts per cell.

400 600FSC-H

@

aCI8

ol

ffiT0 200 400 600 800FSC.H

400 600FSC-H

200 800 200 800

Palasite, 2003, 10, 297-302

Oocystlcell ratio at the 0,1/Iinfection

Mean percentage t SD of 3.3 t 0 5

C, parurnt -infected cellsfor the entite cultttre

CV (%) I' 1.3

Table 11. - Intra-assav test on the lesults of flowc)'tometly Mean (15-

20 leplicates for each oocyst/cell ratio) of the percentage of HCT-8 cells infected wirh Crptosporiditn1l pcltuît'nz ! standard deviation(SD) and coefficient of variability (CV) for the entire cultule

scattel pattern Yaried according to the oocyst/cell ratio,as a result of the cells becoming morphologically morecomplex when pluli-parasitised (Rosales et al., 1'993)'

Specifically, both the FSC-H and the SSC-H peaked atthe latio of one oocyst per cell (Fig' 1C); they thenprogresslvely decreased, suggesting that the infectedcells became increasingly granular and then burst(Figs 1D and 1E).For the evaluation of the infection level by FCA, C'paruum-infected cells were labelied in the same wayas used for IFA yet after having detached the cells andperforming all successive steps in suspension' Theresults of the intra-assay test (Table II) showed that thepercentage oî C, paruum-infected ce1ls increased withincreases in the ratio of excysted oocysts per cell(Table II), Figure 2 illustrates, Îor a representativesample of HCT-S cells analysed by FCA, both a dotplot (double fluorescence anti-SA35-FITC and aVW-tiotin-streptavidin-PE) and a histogram plot (singlefluolescence anti-SA35-FITC). FoL the non-infectedce11s (Figs 2A and 2B), both the dot plot and the his-togram piot show only one cell population, which wasnegative for both FL1 and FL2 (i,e., fluorescence inten-siry of less than 101)' At an oocyst,/cell ratio of 0 5(Figs 2C and2D), two populations were identified: oneconsisted of non-infected celis, which was negative forboth fluorochromes and which 1'epresented 80 % of thetotal population, and the othel consisted of infectedcells, which was positirre for both fluorocht'omes andwhich l'epresented 20 o/o of the total population' At a

ratio of one oocyst/cell (Figs 2E and 2F), there wasonly a population of infected cells' At a ratio of thleeoocysts/ ceil (Figs 2G and 2H), the infected cell popu-lation showed a strong increase in fluorescence inten-sity for both fluorochromes, suggesting that the numberof pluri-parasitised cells had increased. At oocyst/cellrati,cs higher than three (data not shown), the fluo-rescence intensity had decleased because the highamount of parasites had greatly damaged the celimonolayer, dlminishlng its ability to support the infec-tion.Figr-rre 3 shows, for each oocyst/ceil ratio, the fluo-rescence intensity measured as the peak channel offluorescence, for cel1s labelled with anti-SA35 com-pared to those labelled with aVW-biotin' The peak

104 100 101 10' 103 104

FL1.H

0.5/r t/\20,3 t 2,6 86.0 I 12,0 Éo'

Ro

b

b

ò

b

I-o

è

fIr14

+-ait o

b

:2ac

+.JFL

N

Éo,îo

1oo lor 102 103

FL1.H

Fig. 2 - Dot plots of the fluorescence of anti-sA35-Ftrc antibody(fif -H) ana a\rvl-biotin-stteptavidin-PE (FL2-H) and histogram plotsof a single fluorescence (FL1-H) Non-infected HCT-8 cells (A andB); cells infected with 0.5 oocyst/cell (C and D); cells infected withone oocyst/cell (E and F)i and cells infected with three oocysts/cell(G and H),

oocysts/cell ratio

Fig. 3. - Fluorescence intensity (peak channel of fluorescence) at24 houls after infection for HCT-8 cells infected with excystedoocysts of Cryptosporidit'r'n't paruul1x, labelled with rantiSA35-FITC(open bars) ànd aWl-biotin-streptavidin-PE (black bars), bY

oocyst/cell ratio,

Loc rcu(ú

or< 100o)

50

20100.50.1

B

c

#$

D

H

channels followed a similar pattern when cornpar.ingthe fwo labelling merhods, althougir those labelled withanti-SA35 showed higher peak channels, suggestingthat antiSA35 and aY\rL-biotin recognise diffelent epi-topes with different frequencies, For oocyst/cel1 ratiosof 0.1-1,0, the peak channels progressively increased;they then levelled-off for rarios of 2,0 and 3,0, sr-rg-gesting that all of the cells were infected, At higher.oocyst/ceil ratios, the peak channel decreased becausethe level of infection significantly damaged the ce1lmonolayer, diminishing its ability ro suppofi the infec-tion,

DISCUSSION

he extent of parasitism of cells, which isva:riable over time, is difficult to quantify byIFA, in that parasitised celis usually detach

from the substlate and cannot be counted, Further-more) parasites are not uniformly distributed in thecell monolayer, and a pluri-parasitised cell can befound in contact with non-parasitised cells (Barneset al,, 1998; Gut & Nelson, I99D, For this i'eason, ahigh number of arbttrarily chosen fields needs to beanalysed when estimating the infection level withIFA. Althor-rgh oul analyses were conducted on20 fields, the consistently high CV indicares thar agreater number is required, The finding that thenon-parametric Spearman correlation coefficient wasapproximately f (t = 0.98, P < 0,01) when compa-ring IFA to FCA in terms of the level of infection atdifferent. oocyst,/cell 1'arios suggests that bothmethods can be used to evaluate the infection levelbecause they have a similar sensitivity at iheoocyst/cell ratio tested, HoweveL, the advantage ofFCA is that it allou,s the entire sample to be exa-mined within a short period of time, hence overco-ming the problem of heterogeneity within thesample, In fact, ln the intra-assay test, the CV waslower for FCA than fol IFA. Moreovet', the infection1eve1 estimated by IFA is subjecr ro the expertise ofthe operator, as shown by the CV values obtainedwith the inter-assay test,In FCA, the cell autofluorescence and the non-spe-cific fluorescence of antibodies are evaluated by sub*tracting the values detected in non-infected cells fr.omthose detected in infected cells. Free oocysts, sporo-zoites, and debris, which could interfere with themeasure of fluorescence of intracellr,rlar parasites, canalso be excluded. In conclusion, FCA allows thenurnber of infected cells in uitro to be analysed andquantified in a short period of time with high 1.ep1.o-ducibllity and without any bias due to operator exper-tlse,

ACKNCWLEDGEMENTS

b tlrank Daniele Tonanzi for technlcal assis-tance in preparing pr-rlified oocysts of C.pa.ruLn1x, Grant supporti Istituto Superiore di

Sanità, Minístero della Sanità (National AJDS project,contract 50D.6),

REFERENCES

ARRo\rooD M,J., Xrn L,T, & Hunn M,R, h,t. uitro assays of madu-ramicin activity against Cryp t o sp o 1"1 d.iuln p a.ruLt 11x. J'o r,u"na Iof Eukaryotic Mícrobiologlt, L994, 41, 235.

Bamrs D,A,, BoNNTN A., HuANG J.X,, Gousser L., Gur J,lX/,,Dovre P,, Dunruvnrz J.F,, \X/ARD H, & PBrnRsow C, A novelmr-rlti-domain mr"rcin-like glycoprotein of A"yptospotidir,m.tparuullx mediates invasion. Molecular ancl BiocbemicalParasitologyt, 7998, 96, 93-710,

DENc M.Q, & CrrwR D.O. Clyptospotid.ittm. paruLlnx deye-lopment in t1-re BS-C-1 cell iine. Jotnnal of Parasitology,1998, 84, B-75.

Frrnn R,, Spnen C.A, & Dusry J.P, General biology o.î Ayp-tospoti;liunx, iru Ayptospotidiutn and cryprosporìdiosis.Fayer R. (ed.), CRC Press Boca Raton, 1997, 1.-42.

FoNraiNn M. & Gurlror E. Developrnent of a TaqMan quan-titative PCR assay for CryptosporidiLulx paruLnlx, FEMSMict"obiologic Letter, 2002, 214, 13-1,7.

Gnrnrrrss J,K,, Mooru R., Doorev S,, Klusn G,T. & Tzrponr S,Cryptospot4d.iLLln pal^uLnlx infection of Caco-2 cell mono-layers induces an aplcal monolayer defect, selerctivelyrncreases transmolayer permeability and causes epithelialcell deatlr. Infection and. Im.ntut'tity, 1994, 62, 450t5-4574.

Gur J, & NrrsoN R,G. Ctyptosporid.icun. paruuflz: synchronizedexcystation in uitro and evaluation of spor.ozoite ìnfecti-vity witlr a new Lectin,Based Assay, Jotnrtal. EukaryotícMícrobiology, 1999, 46, 565-575,

O'DoNocuun P.J. Ctyptosporidit tn-t and cryptospor.idiosis inman and anirnals. Inten t ational Jownal for Paras ito logy,L995, 25, L39-195.

Rosares M., CrnueNres J. & Masc,rno C, Cryptosporid.iLlllx p6tr-uumt; cultvre in MDCI( cells, Experinzental Parasitology,1993, 76,209-272,

Rossr P,, Pozro 8., Bnssr M,G., Goivnz Monales M.A. & L,r lìosa G.Experimental cryptosporidiosis in hamsters. Jotu nal of Clí-nical Microbíology, 7990, 28, 3i6-357.

Srrmo T.R., Fnrnoruarr' D., Rosn J.B, & Jarunorvsrr W. Anin oík"o method for detecting infectious Cryptospor"iclirm.t.oocysts with cell culture. Applied and Enuít"onnzentalMicrobiologlt, 1997, 63, 3669-3675,

Srrnro T.R., Honnuex D,E, & Rosr J.B. A tnost probablenr"rrnber assay for enumeration of infectlous Ctyptospoti-cliunt. paruunn oocysts. Applied. and Enuirotztnental Micro-biology, 1999, 65, 3936-3947.

TosINr F., C.rccrò S,, TarqeunRrNr A., La. Ros.q. G, & Pozro E.Identification and characterisation of three antigenic pr.o-teins from Cryptospor4diunx paruutlx sporozoites using a

DNA library expressing poly-histidine tagged peptides'Interuational Jou'rnal for Parasitologlt, 1999, 29, 1925-1933.

Urron S.J., Îlmv M, & Bntruunr D.B, Comparative deve-lopment of Cqtptosporid'iu.n't' (Ap'tcomplexa) in 11 conti-nuous lrost cell lines. FEMS Microbíology Letter"s, 1))4a,115, 233-236,

Uerox SJ., Ttunv M, & Brunnanr D.B, Effect of select mediumsupplements on in uitro development of Cryptosporidiunnpa.ruunx in HCT-8 cells. Jou,mal of Clinical Microbiologlt,1995, 33,377-375.

Unron S.J., Tntev M., NrstennNKo M'V, & Bnunrl't D,B. Asimple and leiiable method of producing in uitro inÎec'tions of Cryptosporidiun't' paruunx (Apicomplexa), FEMSMicrobiologlt Letters, 7994b, 1 18, 45-49.

\foops K.M., Nesremnro M,V, & Uprox S.J. Development ofa microtitre ELISA to quantify development of Ctyptospo-rid.ium paruum in uitro, FEtuIS Mícrobiology letter, 1995,128,89-94.

Yeir'c S., Hrarev M,C., Du C. & ZHaNc J. Complete develop-ment of Cryptosporid'trt'nxparuunx in bovine fallopian tubeepithelial cells. Infectlon and InxnxLtnity, 1996, 64,349-354.

Regu 1e 29 mars 2003Accepté le 19 septembre 2003

Palasite, 2003, 10, 297-3oz