-
Gut 1995; 37: 512-518
Attachment of Giardia lamblia trophozoites to acultured human
intestinal cell line
P H Katelaris, A Naeem, M J G Farthing
AbstractAttachment of Giardia lamblia tropho-zoites to
enterocytes is essential forcolonisation of the small intestine and
isconsidered a prerequisite for giardiainduced enterocyte damage.
The precisemechanisms involved are still beingdebated and some
earlier work has beenperformed in models of uncertain biologi-cal
relevance. In this study, co-incubationof giardia with
enterocyte-like differenti-ated Caco-2 celis was used as a model
tostudy the influence of physical andchemical factors on
attachment. Giardiaattachment was maximal between one andeight
hours and stable over pH 7.2-8.2 butit was reduced by
acidification. Attach-ment was dependent on temperature andwas
maximal at 370 and virtually abolishedat 4°C. It was reduced
compared withcontrols (p
-
Attachment ofGiardia
from the parent WB isolate by the technique oflimiting
dilution.'9
CACo-2 CELL CULTURECaco-2 cells (passage 88) were kindly
providedby Dr I Hassan, Ciba-Geigy Pharmaceuticals,Horsham, UK.
Cells were cultured at 37°C in75 cm2 flasks in Dulbecco's modified
Eagle'smedium, supplemented by 10% fetal bovineserum, 1%
non-essential amino acids, 1%L-glutamine, penicillin 100 IU/ml,
andstreptomycin 100 ,ug/ml in an atmosphere of10% CO2 and 90%
air.14 15 The medium waschanged every 48 hours and cells
werepassaged every seven days in a split ratio of 1:4.For
experiments, cells were seeded into 9-6cm2, six-well, tissue
culture treated polystyreneplates (Becton Dickinson, UK) in a split
ratioof 1:18, and the medium was replaced every 48hours. Cells
reached confluency within three tofive days. Cells between passage
92-102 wereused. The number of Caco-2 cells per well wasestimated
by counting cells with an invertedmicroscope using a grid lens with
a fielddiameter of 0.9 mm and multiplying the cellcount obtained by
the number of fields perwell. The median number of cells 16 days
afterpassage was 2.0X 106 cells/well (n= 10).
DIFFERENTIATION OF CACo-2 CELLSFunctional differentiation of
Caco-2 cells wasdetermined by assay of sucrase and maltaseactivity
after a varying duration of culture. Forthis, cell monolayers were
washed twice withchilled 0. 15 M phosphate buffered saline(PBS),
then harvested in 1 ml ice cold distilledwater. Cells were gently
homogenised to a finesuspension by rapid pipetting, and
disacchari-dase activity was determined using modifica-tions of the
glucose oxidase-peroxidasemethod.20 21 Briefly, 10 RI duplicate
aliquots ofappropriately diluted, homogenised Caco-2samples were
incubated with 0.056 M disac-charide substrate in 0.1 M phosphate
buffersolution in 96 well microtitre plates (BectonDickinson, UK)
at 37°C for one hour. TRIS-glucose oxidase reagent (250 1I) was
addedand plates were incubated for a further hour. Asubstrate blank
was included for each sample.Plates were read on a Titertek
Multiscanmicroplate reader (Flow Laboratories, UK) at510 nm and the
enzyme activity was expressedas U/jig protein. Protein was
estimated withthe bicinchoninic acid protein assay reagentmethod22
(BCA, Pierce, USA).The structural differentiation of Caco-2
cells
was assessed 16 days after passage by scanningelectron
microscopy. Cells grown on tissue cul-ture-treated, 10 mm round,
glass cover slipswere double fixed in 3% glutaraldehyde in 0 1M
phosphate buffer at pH 7-4 followed by 1%aqueous osmium tetroxide.
They were dehy-drated, critically point dried using C02,
sputtercoated with gold palladium, and examined in aJEOL SEM 5300.
Electron micrographs werekindly taken by Dr Alan Phillips,
ElectronMicroscopy Department, Queen ElizabethHospital for Sick
Children, London.
GIARDIA-CACo-2 CELL CO-INCUBATIONTo determine the optimal medium
for co-incubation, preliminary attachment assayswere done (as
described below) with four dif-ferent media: supplemented DMEM,
modifiedTYI-S-33, modified RPMI-1640, and modi-fied HSP3 (MHSP3).
The latter two mediahave been reported to support growth of
bothgiardia and mammalian cells.1223 Co-incuba-tion was in a ratio
of giardia:Caco-2 cells of1:10, at 37°C in 50/o C02, 95% air for
onehour. Giardia motility 24 hours after co-incubation in these
media was assessedqualitatively. Giardia attachment was: 24 (5)%in
supplemented DMEM, 27 (4)% in modifiedTYI-S-33, 13 (3)% in modified
RPMI-1640and 42 (4)% in MHSP3, (n=6 for -eachmedium). On the basis
of these results,MHSP3 was selected for use in experiments,
asattachment was maximal (p
-
Katelaris, Naeem, Farthing
120-
2 Maltase1- 100 / Mut80 _
* 60
C)40 Sucrase
s 20 -
. 0
oz 3 9 14 21Time after passage (d)
Figure 1: The time course for the expression ofdisaccharidase
activity in Caco-2 cells (n= 6for eachpoint).
divalent cation dependency of attachment wasdetermined by
co-incubating giardia in thepresence of EDTA 2.5 mM. These
experi-ments and those following, unless otherwisestated, were done
with isolate WB, on Caco-2cells 15-17 days after passage in 5% CO2
and95% air. The percentage attachment wascompared between three
different giardia iso-lates and a clone. As bile promotes growthof
trophozoites in vitro,'8 24 an effect onattachment was sought by
comparing attach-ment of trophozoites passaged in the presenceand
absence of bovine bile in the culturemedium.To determine the role
of components of the
giardia cytoskeleton in attachment, assayswere done in the
presence of microtubuleinhibitors colchicine (dissolved in PBS)and
mebendazole (1.0-100 ,ug/ml, SigmaChemical Co, St Louis, USA) and
after 15minutes' preincubation of trophozoites withthe
microfilament inhibitor, cytochalasin B(dissolved in DMSO 1%). The
role of giardialectin in attachment was studied by preincu-bating
giardia for 15 minutes with D-mannose(5-100 mM),
mannose-6-phosphate (3.5-70mM), and D-glucose (5-300 mM).
Glucosestudies were done using PBS for co-incuba-tion. Attachment
was also determined aftergiardia had been pre-incubated in
trypsin(0.01-10-0 mg/ml) for 20 minutes (Sigma typeXIII) to
determine whether previouslyreported lectin activation by trypsin
in giardiasonicate is also evident in whole tropho-zoites.25
Further studies were done pre-incu-bating Caco-2 cells for 15
minutes withconcanavalin A (10-100 ,ug/ml), which bindsmannosyl
residues.
CONTROLSFor all experiments, the results were comparedwith an
equal number of control attachmentassays. These assays were done at
the sametime as test wells, under identical conditionsbut without
the alteration of co-culture para-meter on the compound being
tested. Wherepossible controls were included in the sameculture
plates as test wells. Further controlexperiments were done in
medium containingDMSO 1%.
SCANNING ELECTRON MICROSCOPYScanning electron micrographs of
giardia afterco-incubation with Caco-2 cells for one hourand 16
hours were obtained, using themethods described above.
STATISTICAL ANALYSISResults are expressed as mean
(SEM).Differences between means were comparedusing two-tailed
Student's t tests or analysis ofvariance where appropriate. All
data representa minimum of n=6 for each data point andexperiments
were carried out on at least threeseparate occasions.
Results
CACO-2 CELLS MODELThe time course for the expression of
disac-charidase activity is shown in Figure 1.Enzyme activity rose
sharply from day 9 afterpassage and reached a plateau after day
14,indicating functional differentiation.'7 Thepresence of
confluent monolayers with micro-villi on the apical surface of
cells providedevidence of structural differentiation. As hasbeen
previously observed,'5 17 the density ofthe microvilli varied
between cells indicatingthe heterogeneity of the cell population.
Somecells showed dense microvilli, others hadclusters, while a few
cells had sparsely presentmicrovilli (Fig 2). On the basis of
thesefindings, cells 15-17 days after passage wereused for
attachment experiments as functionaland structural differentiation
was evident.At the end of the co-incubation period,
Caco-2 cell monolayers remained confluentand appeared intact
with normal morphologywhen examined by light microscopy and
scan-ning electron microscopy. Giardia remainedviable after
co-incubation as trophozoites weremotile and could be successfully
subculturedafter recovery from culture wells.
PHYSICOCHEMICAL FACTORS IN ATTACHMENTUsing an inoculum of 2.OX
105 trophozoites at37°C and pH 7.2, giardia attachment to Caco-2
cells increased with time up to 60 minutes,then reached a plateau.
The range of attach-ment over one to eight hours was 40-46 (6)%of
the total number added. Attachment wasstill evident by 24 hours.
The time course ofattachment is shown in Figure 3. The totalnumber
of organisms recovered was between75-1 15% of the estimated
inoculum added fortime periods up to eight hours. After 24
hoursco-incubation, giardia numbers had increasedto 5.1 x 105
indicating multiplication oftrophozoites (generation time= 17
hours).Multiplication was not evident after shorterperiods of
co-incubation. Attachment wastemperature dependent, being maximal
at37°C (43.5 (4)%, reduced by 59% at 21°C(17.8 (3)%) and virtually
abolished at 40C (0 3(0.3)%). Attachment occurred over a range ofpH
but was maximal at pH 7.2-8.2 (Fig 4).The total number of
trophozoites attached
514
on June 24, 2021 by guest. Protected by copyright.
http://gut.bmj.com
/G
ut: first published as 10.1136/gut.37.4.512 on 1 October 1995.
D
ownloaded from
http://gut.bmj.com/
-
Attachment of Giardia
N _U y .. ...Figure 2: Scanning electron micrographs of Caco-2
cell monolayers 16 days after passage. (A) Structural
differentiation isevidenced by the presence of microvilli on the
apical surface of the cells. These cells have a dense array of
microvilli.(Magnification X 7500; bar= I ,m.) (B) These cells
display clusters of microvilli on the apical surface.
(MagnificationX 7500; bar= 1 ,m.)
after 60 minutes at pH 7.2 and 37°C increasedwith increasing
numbers of giardia seeded,whereas the percentage attached was
similar atall giardia:Caco-2 cell ratios (Fig 5). The fore-going
experiments established the optimal con-ditions for the attachment
assay. Unlessotherwise stated, the results following werederived
using a parasite:Caco-2 cell ratio of1:10, over 60 minutes at 37°C,
pH 7-2, in 5%CO2 and 95% air.
CYTOSKELETAL INHIBITION AND CHELATION OFDIVALENT
CATIONSChelation of divalent cations with EDTA 2.5mM reduced
attachment by 32 (4)% com-pared with controls (p
-
Katelaris, Naeem, Farthing
50 r
40 ;-
c
E
cJCot
30 F-
20 F-
10
n is i1:100 1:20 1:10
Giardia Iambia:Caco-2 cell ratioFigure 5: The relationship
between Giardia lamblia:Caco-2 celland absolute numbers of
trophozoites attached (n= 6-12 for each
giardia was not differenthad been grown in the piabsence of bile
(45 (3)%,
ISOLATE VARIATIONThe proportion of trophnot vary significantly
betisolates. Using the standsattachment was: WB 46 (and VNB3 47
(5)%. Theattached to the same disolate (44 (5)%).
SCANNING ELECTRON MIC.Giardia were seen in capparently attached
tomost trophozoites were
25;: 5c* 1251225 125- 10X 00
Figure 6: The effect of EDTA, colchicine, mebendazole,
cytochal.
dimercaptosuccinic acid (DMSO) 1 on attachment of Giardia
point, *p
-
Attachment of Giardia 517
Figure 7: Scanning electron micrographs ofGiardia lamblia
coincubated with Caco-2 cells. (A) G lamblia trophozoite inclose
apposition and apparently attached to a Caco-2 cell. The ventral
surface of the trophozoite is applied to themonolayer, which
displays sparse microvilli (1 h co-incubation). (Magnification X
5000; bar= 1 ,gm.) (B) G lambliatrophozoites in dorsal and ventral
orientation to Caco-2 cells (16 h co-incubation). (Magnification X
5000; bar= 1 ,gm.)
Caco-2 cells as a model for attachment,reported a biphasic
increase of attachment toCaco-2 cells after inhibition of
contractileproteins with cytochalasin B, a finding which isat odds
with other previous studies and thecurrent work and may be the
result of a lownumber of repeat experiments. This alsoconflicts
with a previous report from the samegroup using this model, which
describedreduced attachment to Caco-2 cells afterinhibition of
contractile proteins by chelationof divalent cations.12 The current
study showsthe importance of both divalent cation deple-tion with
EDTA and microfilament inhibitionwith cytochalasin B in reducing
attachment ofgiardia to a human cell line. This is consistentwith
data from models using plastic surfacesand cell lines that are not
human9 11 andsupports the role of contractile filaments
inattachment.
Lectin mediated attachment was evident inthis system,
inhibitable by D-mannose, man-nose-6-phosphate, and concanavalin A,
consis-tent with a mannosyl target for binding. Noenhancement of
attachment was seen afterexposure to trypsin which suggests that
thetrypsin activation of giardia lectin described ingiardia
sonicate25 28 does not have biologicalimportance for attachment.
Evidence suggeststhat lectin mediated binding is not the
primarymode of attachment of giardia. Firstly, attach-ment to
synthetic surfaces is avid and is notdependent on receptor-ligand
mediated bind-ing. Secondly, the magnitude of the reductionin
attachment is greater after inhibition ofcytoskeletal function than
with competitiveinhibitin of lectin mediated binding.
Lastly,although giardia are found in various orienta-tions to
epithelial cells, most trophozoites areobserved ventral surface
down. Avidity forbinding of giardia to rat intestinal cells hasbeen
shown to be greater with more proximalintestinal cells.8 It is
possible that lectinmediated attachment serves to localise
giardiato its preferred intestinal site and that tropho-zoites
subsequently reorientate and attach viathe ventral disc.
This model allows physiological manipula-tion of the co-culture
medium to mimic someconditions in the small intestine. For
example,
pancreatic bicarbonate secretion into theduodenum results in
periodically elevatedluminal pH and giardia survive in
thisalkalinised milieu. It is not known whether thisis because
giardia can resist a raised pH orbecause trophozoites may avoid
exposure toluminal contents by burrowing under themucus layer. With
this model, attachment hasbeen demonstrated to be maintained
despitedirect exposure to alkaline conditions up to apH of 8.2.
In summary, the giardia-Caco-2 cell co-incubation model seems to
be a relativelysimple and useful vehicle for the study of
theattachment of giardia in vitro. A predominantrole for mechanical
attachment via cytoskeletalmechanisms is suggested by these
studies,although lectin associated binding is presentand may also
have a role in vivo.This study was supported by the Wellcome Trust,
London. Partof this work was presented at the American
GastroenterologicalAssociation meeting in Boston 1993
(Gastroenterology 1993;104: A721).
1 Erlandsen SL. Scanning electron microscopy of
intestinalgiardiasis: Lesions of the microvillous border of the
villusepithelial cells produced by trophozoites of Giardia,pp.
775-82. In: Johari 0, ed. Scanning electron microscopy.Chicago: IIT
Research Institute, 1974.
2 Smith PD. Pathophysiology and immunology of giardiasis.Ann Rev
Med 1985; 36: 295-307.
3 Holberton DV. Arrangement of subunits in microribbonsfrom
Giardia. J Cell Sci 1981; 47: 167-85.
4 Peattie DA, Alonso RA, Hein A, Caulfield JP.Ultrastructural
localisation of giardins to the edges of diskmicroribbons of
Giardia lamblia and the nucleotide anddeduced protein sequence of
alpha giardin. J Cell Biol1989; 109: 2323-35.
5 Feely DE, Schollmeyer JV, Erlandsen SL. Giardia: distri-bution
of contractile proteins in the attachment organelle.Exp Parasitol
1982; 53: 145-54.
6 Holberton DV. Attachment of Giardia - a hydrodynamicmodel
based on flagellar activity. J Exp Biol 1974; 60:207-21.
7 Farthing MJG, Periera MEA, Keusch GT. Description
andcharacterization of a surface lectin from Giardia lamblia.Infect
Immun 1986; 51: 661-7.
8 Inge PMG, Edson CM, Farthing MJG. Attachment ofGiardia lamblia
to rat intestinal epithelial cells. Gut 1988;29: 795-801.
9 Feely DE, Erlandsen SL. Effect of cytochalasin-B, lowCa'
concentration, iodoacetic acid and quinacrine-HClon the attachment
of Giardia trophozoites in vitro.J Parasitol 1982; 68: 869-73.
10 Gillin FD, Reiner DS. Attachment of the flagellate
Giardialamblia: role of reducing agents, serum, temperature,
andionic composition. Mol Cell Biol 1982; 2: 369-77.
11 McCabe RE, Yu GSM, Conteas C, Morrill PR, McMorrowB. In vitro
model of attachment of Giardia intestinalistrophozoites to IEC-6
cells, an intestinal cell line.Antimicrob Agents Chemother 199 1;
35: 29-35.
12 Favennec L, Chochillon C, Meillet D, Magne D, Savel
J,Raichvarg D, et al. Adherence and multiplication of
on June 24, 2021 by guest. Protected by copyright.
http://gut.bmj.com
/G
ut: first published as 10.1136/gut.37.4.512 on 1 October 1995.
D
ownloaded from
http://gut.bmj.com/
-
518 Katelaris, Naeemn, Farthing
Giardia intestinalis on human enterocyte-like
differentiatedcells in vitro. Parasitol Res 1990; 76: 581-4.
13 Pinto M, Robine-Leon S, Appay MD, Kerdinger M,Triadou N,
Dussaulx E, et al. Enterocyte-like differentia-tion and
polarisation of the human colon carcinoma cellline Caco-2 culture.
Biol Cell 1983; 47: 323-30.
14 Hidalgo IJ, Raub TJ, Borchardt RT. Characterisation of
thehuman colon carcinoma cell line (Caco-2) as a modelsystem for
intestinal epithelial permeability.Gastroenterology 1989; 96:
736-49.
15 Wilson G, Hassan IF, Dix CJ, Williamson I, Shah R, MackayM,
Transport and permeability properties of humanCaco-2 cells: an in
vitro model of the intestinal epithelialcell barrier. Journal of
Controlled Release 1990; 11: 25-40.
16 Matsumoto H, Erickson RH, Gum JR, Yoshioka M, GumE, Kim YS.
Biosynthesis of alkaline phosphatase duringdifferentiation of the
human colon cancer cell line Caco-2.Gastroenterology 1990; 98:
1199-207.
17 Vachon PH, Beaulieu J-F. Transient mosaic patterns
ofmorphological and functional differentiation in the Caco-2 cell
line. Gastroenterology 1992; 103: 414-23.
18 Keister DB. Axenic culture of Giardia lamblia in
TYI-S-33media supplemented with bile. Trans Roy Soc Trop MedHvg
1983; 77: 487-8.
19 Baum KF, Berens RL, Jones RH, Marr JJ. A new methodfor
cloning Giardia la?nblia, with a discussion of thestatistical
considerations of limiting dilution. _7 Parasitol1988; 74:
267-9.
20 Belosevic M, Faubert GM, MacLean JD. Disaccharidase
activity in the small intestine of gerbils
(Merionesunguiculatus) during primary and challenge infectionswith
Giardia lamblia. Gut 1989; 30: 1213-9.
21 Dahlqvist A. Assay of intestinal disaccharidases. AnalBiochem
1968; 22: 99-107.
22 Smith PK, Krohn RI, Hermanson GT, Mallai AK, GartnerFH,
Provenzano MD, et al. Measurement of protein usingbicinchoninic
acid. Anal Biochemn 1985; 150: 76-85.
23 Guy RA, Bertrand S, Faubert GM. Modification of RPMI-1640 for
use in in vitro immunological studies of host-parasite interactions
in giardiasis. J Gln Microbiol 1991;29: 627-9.
24 Farthing MJG, Varon SR, Keusch GT. Mammalian bilepromotes
growth of Giardia lainblia in axenic culture.Trans Roy Soc Trop Med
Hyg 1983; 77: 467-9.
25 Lev B, Ward H, Keusch GT, Pereira MEA. Lectin activa-tion in
Giardia lamblia by host protease: A novel host-parasite
interaction. Science 1986; 232: 71-3.
26 Edlind TD, Hang TL, Chakraborty PR. Activity of
theanthelmintic benzimidazoles against Giardia lamibl/ia invitro. J
Infect Dis 1990; 162: 1408-11.
27 Magne D, Favennec L, Chochillon C, Gorenflot A, MeilletD,
Kapel N, et al. Role of cytoskeleton and surface lectinsin Giardia
duodenalis attachment to Caco2 cells. ParasitolRes 1991; 77:
659-62.
28 Ward HD, Lev BI, Kane AV, Keusch GT, Pereira
ME.Identification and characterization of taglin, a mannose
6-phosphate binding, trypsin-activated lectin from Giardialamiblia.
Bi(cheinistrv 1987; 26: 8669-75.
on June 24, 2021 by guest. Protected by copyright.
http://gut.bmj.com
/G
ut: first published as 10.1136/gut.37.4.512 on 1 October 1995.
D
ownloaded from
http://gut.bmj.com/
