INFECTION AND IMMUNITY, Oct. 2004, p. 6125–6131 Vol. 72, No.

10
0019-9567/04/$08.00!0 DOI: 10.1128/IAI.72.10.6125–6131.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Host Cell Tropism Underlies Species Restriction of Human and

Bovine Cryptosporidium parvum Genotypes
Amna Hashim,1,2,3 Marguerite Clyne,1,2 Grace Mulcahy,3 Donna Akiyoshi,4
Rachel Chalmers,5 and Billy Bourke1,2*
The Children’s Research Centre, Our Lady’s Hospital for Sick Children, Crumlin, 1 and Departments of Paediatrics2 and Veterinary
Microbiology and Parasitology, 3 Conway Institute for Biomolecular and Biomedical Research, University College, Dublin, Ireland;
Division of Infectious Diseases, Tufts University School of Veterinary Medicine, North Grafton, Massachusetts4; and
Cryptosporidium Reference Unit, NPHS Microbiology Swansea, Singleton Hospital, Swansea,
Wales, United Kingdom5
Received 31 March 2004/Returned for modification 22 May 2004/Accepted 15 July 2004

It has been recognized recently that human cryptosporidiosis is usually caused by Cryptosporidium parvum
genotype I (“human” C. parvum), which is not found in animals. Compared to C. parvum genotype II, little is
known of the biology of invasion of the human-restricted C. parvum genotype I. The aims of the present study
were (i) to explore and compare with genotype II the pathogenesis of C. parvum genotype I infection by using
an established in vitro model of infection and (ii) to examine the possibility that host-specific cell tropism
determines species restriction among C. parvum genotypes by using a novel ex vivo small intestinal primary cell
model of infection. Oocysts of C. parvum genotypes I and II were used to infect HCT-8 cells and primary
intestinal epithelial cells in vitro. Primary cells were harvested from human endoscopic small-bowel biopsies
and from bovine duodenum postmortem. C. parvum genotype I infected HCT-8 cells with lower efficiency than
C. parvum genotype II. Actin colocalization at the host parasite interface and reduction in levels of invasion
after treatment with microfilament inhibitors (cytochalasin B and cytochalasin D) were observed for both
genotypes. C. parvum genotype II invaded primary intestinal epithelial cells, regardless of the species of origin.
In contrast, C. parvum genotype I invaded only human small-bowel cells. The pathogenesis of C. parvum
genotype I differs from C. parvum genotype II. C parvum genotype I does not enter primary bovine intestinal
cells, suggesting that the species restriction of this genotype is due to host tissue tropism of the infecting
isolate.

Cryptosporidium parvum, an intracellular parasite belonging in terms of oocyst shedding, disease severity, and site of infec-
to the phylum Apicomplexa, is an important cause of diarrheal tion (28, 34). For clarity, we continue to use here the termi-
disease worldwide (10, 13, 17). Although many C. parvum nology “genotype I” and “genotype II.”
infections are self-limiting, severe manifestations may occur in In most human epidemiological studies C. parvum genotype
immunologically incompetent and debilitated individuals (10, I isolates are identified in human stool samples more fre-
16, 17). In the immunodeficient host, in particular those with quently than genotype II (21, 24, 25, 43, 47). However, under-
AIDS, infection may be life-threatening, and the development standing of Cryptosporidium pathogenesis to date has relied on
of effective antimicrobial therapy for human cryptosporidiosis studies with C. parvum genotype II isolates or isolates of un-
is only beginning to evolve (35, 44, 45). known genotype. The emergence of a widely prevalent non-
To develop novel, effective treatments against cryptospori- zoonotic C. parvum genotype has stimulated a major reassess-
diosis, a detailed understanding of the pathogenesis of the ment of our understanding of this disease in humans. In
causative organisms is necessary. Traditionally, cryptosporidi- particular, the possibility that the mechanisms of transmission
osis in humans has been viewed as a zoonotic infection, with and pathogenesis of this organism differ from those associated
cattle being the principal environmental reservoir. However, with C. parvum genotype II has major implications for strate-
recently it has emerged that humans are commonly infected gies aimed at controlling and treating human cryptosporidiosis.
with a C. parvum genotype (genotype I) only very rarely found We have used here a combination of established in vitro tissue
in other species (45). Although C. parvum genotype I is re- culture models and an ex vivo primary intestinal cell model of
stricted to humans, C. parvum genotype II can infect humans, cryptosporidial infection to show that C. parvum genotype I
cattle, and a wide variety of other animals (2, 33, 37, 40). A differs in the pattern and severity of infection compared to C.
compelling case now exists for the redesignation of this subtype parvum genotype II isolates. In addition, we show that bovine
as a novel cryptosporidial species (Cryptosporidium hominis sp. primary cells cannot be infected with C. parvum genotype I
nov.) based on molecular differences at several genetic loci and isolates, indicating that host cell tropism underlies the species
restriction of this C. parvum subtype in vivo.

* Corresponding author. Mailing address: Professorial Unit, The

MATERIALS AND METHODS
Children’s Research Centre, Our Lady’s Hospital for Sick Children,
Crumlin, Dublin 12, Ireland. Phone: 35314556901. Fax: 35314555307. C. parvum oocysts. C. parvum genotype II oocysts (Iowa strain) were obtained
E-mail: billy.bourke@ucd.ie. from a commercial source (Pleasant Hill Farms, Troy, Idaho). This genotype II

6125
6126 HASHIM ET AL. INFECT. IMMUN.

strain was originally isolated from a calf by Harley Moon. It has been passaged Cytokeratin staining of epithelial cells. To confirm that the isolated primary
through calves and purified from the fecal material by ether extraction, followed cells were epithelial in origin, the cells were stained for cytokeratin. Monolayers
by a one-step sucrose gradient. were fixed with 4% formaldehyde in PBS for 30 min. Cells were permeabilized
C. parvum genotype I isolates, 5942 and TU502, were isolated and genotyped with 1% (vol/vol) Triton X-100 in PBS for 10 min and were blocked with 0.5%
at the Cryptosporidium Reference Laboratory and the Tufts University School of (wt/vol) bovine serum albumin (BSA) for another 10 min. Monolayers were then
Veterinary Medicine, respectively. For the C. parvum 5942 isolate oocysts were incubated with monoclonal anti-PAN cytokeratin (Sigma) in 0.5% (wt/vol) BSA
purified from Cryptosporidium-positive feces collected by the Cryptosporidium for 30 min, followed by goat anti-mouse fluorescein isothiocyanate (FITC; Dako)
Reference Unit at the Singleton Hospital as part of the National Collection of for another 30 min. The coverslips were mounted by using fluorescence mount-
Cryptosporidium Oocysts. Fecal isolates were confirmed by microscopic inspec- ing medium (Dako) and examined with a fluorescence microscope. Control cells
tion of modified Ziehl-Neelsen-stained smears (3) and PCR-restriction fragment were stained as described above, but the primary antibody was omitted.
length polymorphism analysis of the Cryptosporidium oocyst wall protein gene Infection assay. The maintenance medium was removed and 2 " 105 Crypto-
(38). Briefly, oocysts were separated by flotation from fecal debris by using a sporidium oocysts in 2 ml of growth medium were added to the cell monolayers.
saturated sodium chloride solution and centrifugation for 8 min at 1,600 " g (36). The multiplicity of infection was approximately five oocysts/cell. The plates were
The floating material containing the oocysts was washed with deionized oocyst- placed at 37°C in a 5% CO2–95% air humidified incubator. After incubation for
free water, the oocysts were resuspended in deionized, oocyst-free water and 3 h the infected monolayers were washed with PBS to remove unexcysted oo-
stored at 4°C. To extract DNA, 200 #l oocyst suspension was incubated at 100°C cysts, empty oocyst walls, and toxic materials that may have been liberated from
for 60 min, and DNA was extracted by using proteinase K digestion in lysis buffer the oocysts. Then, 2 ml of growth medium containing 100 U of penicillin/ml and
at 56°C by using a spin-column filtration technique (QiAMP DNA minikit; 100 #g of streptomycin/ml was added. Primary cells were infected for up to 24 h,
Qiagen). DNA extracts were stored at $20°C prior to use. The primers cry-15 and HCT-8 cells were infected for up to 72 h.
and cry-9 were used to amplify a 550-bp region of the Cryptosporidium oocyst wall Lectin VVL staining. Parasites were detected in the monolayers by using
protein gene, which was then subjected to restriction endonuclease digestion by Lectin VVL (a plant lectin from Vicia villosa) staining, as previously described
RsaI. The digestion products were separated by agarose (3% [wt/vol]) gel elec- (18). Monolayers were fixed with 4% formaldehyde in PBS for 30 min, perme-
trophoresis, visualized by using ethidium bromide (0.1 mg/100 ml), and recorded abilized with 1% Triton X-100 in PBS for 10 min, and blocked with 0.5% (wt/vol)
with a digital camera and KDS1D analysis software (Kodak). Product sizes were BSA for an additional 10 min. The monolayers were then incubated with 1 #g of
confirmed by comparison with a DNA molecular weight standard marker (Life conjugated lectin VVL-biotin (B1235; Vector Laboratories)/ml in 0.5% (wt/vol)
Technologies). Oocyst suspensions confirmed as containing C. parvum genotype BSA, followed by 1 #g of streptavidin-CY3 (Sigma)/ml. The coverslips were
1 isolates were resuspended in phosphate-buffered saline (PBS), and those con- mounted by using fluorescence mounting medium (Dako) and then examined by
taining sufficient oocysts after visual inspection of a sample by bright-field mi- fluorescence microscopy with a "50 water immersion lens. The infected cells in
croscopy were used for infection assays. C. parvum TU502 is a well-characterized 10 high-power fields were counted, and the percent rate of infection (number of
genotype I isolate that has been passaged through gnotobiotic piglets (1). Before infected cells/total number of cells " 100) was calculated. All experiments were
host cells were infected, oocysts were decontaminated by an established tech- performed in triplicate and expressed as the mean percent cells infected % the
nique (22). Briefly, they were washed twice with distilled water and then incu- standard deviation of the mean. Uninfected monolayers were used as a control.
bated with freshly prepared 10% (vol/vol) Chlorox bleach (Sigma-Aldrich) for 10 Host cell actin was stained by using phalloidin conjugated to FITC. We used
min on ice. After two further washes with ice-cold distilled water and once with 5 #g of FITC-phalloidin (Sigma)/ml, together with 1 #g of Lectin VVL/ml, for 30
RPMI 1640 (Bio-Whittaker) oocysts were resuspended at a concentration of 2 " min. Slides were examined by using a laser scanning confocal microscope (Bio-
105 oocysts/ml and then used to infect cell monolayers. Rad MRC 1024).
Preparation of HCT-8 cells. Human ileocecal adenocarcinoma cells (HCT-8; Effect of cytoskeletal inhibitors on C. parvum invasion of gastrointestinal cells.
ATCC CCL 244) were obtained from the American Type Culture Collection. To investigate the importance of cytoskeletal components during invasion, the
Cells were maintained in 75-cm2 tissue culture flasks in RPMI 1640 and 10% actin-disrupting drugs cytochalasin D and cytochalasin B (Sigma) and microtu-
(vol/vol) fetal bovine serum (FBS; Sigma). The cells were grown as adherent bule depolymerizing drugs colchicine and vincristine (Sigma) were used. Stock
monolayers in a 37°C and 5% CO2 $95% humidified incubator. At 24 h prior to solutions of the four drugs were prepared and diluted to their final concentration
infection, the monolayers were treated with trypsin-EDTA (Bio-Whittaker) for by using culture medium. Three separate experiments were performed. In the
15 min at 37°C. The cells were grown on 25-mm Thermonax coverslips in six-well first set of experiments, HCT-8 cells and primary cells were incubated for 1 h in
Costar tissue culture plates (Gibco) to 80 to 85% confluence. an assay medium containing either 1 #g of cytochalasin D/ml, 10 #g of cytocha-
Isolation of epithelial cells from human and bovine gastrointestinal cells. lasin B/ml (7), or a range of concentrations of colchicine and vincristine (48) for
Epithelial cells from human and bovine duodenal biopsies were isolated as 1 h at 37°C. The treated monolayers were washed twice with PBS and infected
described previously (9). Human small bowel biopsy tissue was obtained from with 2 " 105 Cryptosporidium oocysts/ml for either 48 h (HCT-8 cells) or 22 h
children undergoing endoscopy at Our Lady’s Hospital for Sick Children after (primary cells). In a second set of experiments cell monolayers were infected as
parental consent was obtained. Ethical approval for the present study was ob- described above but without removing the cytoskeletal inhibitor during the in-
tained from the Ethics Committee at the hospital. The biopsy tissue was placed fection assay. Finally, in order to assess the effect of cytoskeletal inhibition on the
in RPMI 1640 medium containing 10% (vol/vol) FBS and transported directly to parasite, oocysts were incubated with a range of concentrations of either colchi-
the laboratory. Bovine small duodenal tissue was removed from cattle younger cine, vincristine, cytochalasin D, or cytochalasin B for 15 min at 37°C. Oocysts
than 30 months immediately after slaughter at a local abattoir. Small specimens were then washed twice with PBS before being used to infect monolayers.
were taken from the tissue by using a biopsy forceps; placed in RPMI 1640 Untreated monolayers or oocysts were used as a control.
containing 10% (vol/vol) FBS, 100 #g of streptomycin/ml, 100 U of penicillin/ml, Statistical analysis. All results were expressed as the mean percentage of cells
and 2.5 #g of amphotericin B (Sigma)/ml; and transported on ice to the labo- infected % the standard deviation of triplicate experiments. Means were com-
ratory. Cells from both tissues were then isolated by using similar procedures. pared by using the nonparametric Mann-Whitney U test. A P value of &0.05 was
The tissue was placed in Hanks balanced salt solution (Bio-Whittaker) without considered statistically significant.
Ca2! or Mg2! containing 0.1 mM EDTA (Sigma) and 0.1 mM dithiothreitol
(Sigma) and incubated at 37°C for 10 min with vigorous shaking. The superna-
tant was removed, and the tissue was placed in RPMI 1640 containing 0.05% RESULTS
(wt/vol) collagenase (Sigma) and incubated at 37°C for 15 min with vigorous
shaking. After this procedure, the supernatant consisted of intestinal crypts and Invasion of HCT-8 cells by Cryptosporidium isolates. The
single cells. The supernatant was removed with a Pasteur pipette, and the cells
patterns and degrees of invasiveness of C. parvum genotypes I
and crypts were pelleted by centrifugation at 800 " g for 5 min and then
resuspended in RPMI 1640 medium containing 10% (vol/vol) FBS. Fresh colla- and II were compared. Both genotypes invaded HCT-8 cells.
genase solution was added to the biopsy tissue, and the procedure repeated three However, genotype I entered HCT-8 cells with significantly
to four times until no more cells could be isolated. The isolated cells and crypts lower efficiency than did genotype II. Specifically, 55.8% %
were kept on ice until 500 #l was plated in 24-well Costar culture plates on 1.01% of cells were infected by genotype I, and 88.9% % 1.54%
13-mm plastic coverslips with Dulbecco modified Eagle medium–Ham F-12 plus
10% (vol/vol) FBS, 8 #g of insulin/ml, 10 #g of gentamicin/ml, 50 #g of hydro-
of cells were infected with genotype II (P ' 0.002). The inva-
cortisone/ml, 100 #g of streptomycin/ml, 100 U of penicillin/ml, and 2.5 #g of sion efficiency of HCT-8 cells by both isolates of genotype I was
amphotericin B/ml. similar. The pattern of infection differed according to the or-
VOL. 72, 2004 HOST CELL TROPISM AND SPECIES RESTRICTION OF C. PARVUM 6127

FIG. 1. Comparison of the pattern of infection of C. parvum genotypes I and II in HCT-8 cells. Genotype I C. parvum invasion occurred in
clusters (A), whereas genotype II C. parvum invasion was evenly distributed throughout the monolayer (B). Slides were examined by using a
fluorescence microscope. Magnification, "400.

igin of the isolates. Whereas C. parvum genotype II infected permeability of the oocyst wall to the drugs since invasion of
HCT-8 cells evenly (Fig. 1B), internalized genotype I parasites the cells in the presence of either inhibitor markedly reduced
formed discreet clusters throughout the cell monolayers (Fig. infection regardless of the genotype or inhibitor type (Fig. 3B).
1A). Preexposure of oocysts of both genotypes to cytochalasin D
Host cell actin cytoskeletal rearrangement is similar after and cytochalasin B did not affect subsequent invasiveness.
invasion of HCT-8 cells by both genotype I and genotype II Pretreatment of oocysts with microtubule inhibitors atten-
Cryptosporidium isolates. The effects on the host cell actin uates cattle but not human genotype invasion of HCT-8 cells.
cytoskeleton after invasion with both genotypes were com- In order to determine whether C. parvum genotype I also
pared by using confocal scanning microscopy to visualize co- requires an intact microtubule cytoskeleton for efficient infec-
localization of sporozoites with actin. Actin accumulation oc- tion of HCT-8 cells, we examined the effect of the microtubule-
curred in association with internalized sporozoites of both depolymerizing agents, colchicine and vincristine, on infection.
types (Fig. 2). As described in previous studies (7), inhibition of host cell
In addition, disruption of the host actin cytoskeleton with microtubules did not affect C. parvum invasion (data not
cytochalasin B or cytochalasin D significantly inhibited entry of shown). However, pretreatment of genotype II oocysts with
both genotypes. Pretreatment of HCT-8 cells for 1 h with either microtubule inhibitor for 1 h reduced infection of
either microfilament inhibitor substantially reduced infection HCT-8 cells in a dose-dependent manner (Fig. 4A and B).
(Fig. 3A). The reason for the different efficiencies of the two Infection in the presence of either 10$4 M colchicine or 10$4
inhibitors is unclear. However, it may reflect differences in the M vincristine throughout the infection period reduced the in-

FIG. 2. Confocal laser scanning micrographs of host-actin colocalization during invasion of genotype I (A) and genotype II (B) into HCT-8
cells. The green color shows the stained actin, and the red color indicates the intracellular stages of C. parvum. Colocalization of host actin and
the parasite appears yellow (arrows). Slides were examined by using a laser scanning confocal microscope. Magnification, "400.
6128 HASHIM ET AL. INFECT. IMMUN.

FIG. 3. Effect of cytochalasin D and cytochalasin B on infection of genotype I and genotype II C. parvum into HCT-8 cells. Cells were either
treated for 1 h with cytochalasin D or cytochalasin B (A) prior to infection or infected in presence of cytochalasin D or cytochalasin B for 48 h
(B). All values are given as the mean % the standard deviation of triplicate experiments. ❋, P & 0.005 compared to control (no inhibitor).

fection of genotype II oocysts from 88.4% % 1.59% to 30.7% C. parvum genotype II but not genotype I infects primary
% 1.14% and 7.7% % 2.5%, respectively (Fig. 4C). In contrast, bovine intestinal cells. In order to explore the possibility that
neither vincristine nor colchicine treatment of C. parvum ge- the restriction of C. parvum genotype I to human hosts reflects
notype I oocysts affected the rate of infection of HCT-8 cells species-specific tissue tropism for this genotype, we isolated
(Fig. 4A and B). These data suggest that either the mechanism and infected in vitro bovine intestinal cells. The viability of
of host cell entry utilized by C. parvum genotype I sporozoites isolated bovine intestinal cells was monitored by the trypan
is independent of the parasite’s microtubule structures or ge- blue exclusion assay. Microscopically, after 24 to 48 h, at which
notype I oocysts are impermeable to microtubule-depolymer- time they were used for infection studies, these cells appeared
izing drugs used. healthy and viable. C. parvum genotype II infected bovine cells
Culture of primary human cells and invasion by C. parvum with an efficiency comparable to infection of primary human
genotypes I and II. Isolated cells and crypts attached to the cells and HCT-8 cells (92.3% % 1.34%). However, genotype I
coverslips and spread within 24 to 48 h. The cells grew as (C. parvum 5942) organisms did not infect primary bovine cells.
isolated epithelial colonies and propagated by growing out No parasites were seen in association with the bovine cells
from the crypts. The epithelial origin of the cultured cells was either adherent (after 1 h without washing) or after a 22-h
demonstrated by immunofluorescence labeling with cytokera- invasion assay (Fig. 6). In order to corroborate this finding, we
tin. used a second C. parvum isolate. As observed with isolate 5942,
C. parvum genotype II entered primary human cells with an TU502 efficiently invaded primary human cells but did not
efficiency comparable to the invasion of HCT-8 cells. However, infect primary bovine intestinal cells (data not shown).
C. parvum genotype I infected primary human intestinal cells
significantly more efficiently than was the case for HCT-8 cells. DISCUSSION
In addition, the clustered pattern of infection observed for C.
parvum genotype I entry into HCT-8 cells was not evident in Studies of the pathogenesis of C. parvum in vitro to date
primary cells, with infection being equally distributed through- have relied on the use of C. parvum genotype II isolates or
out the cell monolayer (Fig. 5). human-derived isolates of unknown genotype, (5–8, 12, 23).

FIG. 4. Effect of vincristine (vin) and colchicine (col) on infection of HCT-8 cells by genotype I and genotype II C. parvum. Genotype I and
genotype II C. parvum oocysts were treated with a range of concentrations of colchicine (A) and vincristine (B) for 15 min at 37°C prior to their
use to infect HCT-8 cells. (C) HCT-8 cells were infected by genotype II C. parvum in the presence of colchicine and vincristine for 48 h.
VOL. 72, 2004 HOST CELL TROPISM AND SPECIES RESTRICTION OF C. PARVUM 6129

FIG. 5. Lectin VVL staining of primary intestinal human epithelial cells infected with C. parvum genotype I (A) and genotype II (B). Epithelial
cells were infected with 2 " 105 oocysts/ml of genotype I and genotype II. C. parvum and stained with lectin VVL. Background staining by the lectin
VVL also revealed cell morphology. The efficiency and pattern of infection was similar for both C. parvum genotypes (96.2% % 0.36% for genotype
I and 97.4% % 0.1% for genotype II). Slides were examined by using a fluorescence microscope. Magnification, "400.

Given that C. parvum genotype I is the predominant genotype Primary intestinal cells have a number of advantages over
in the majority of studies of sporadic and outbreak cases of immortalized tissue culture cell lines for the study of host-
human cryptosporidiosis (15, 16, 24, 25), it is essential to char- parasite interactions in vitro. As a biologically meaningful sur-
acterize and compare its pathogenesis with that of C. parvum rogate model of direct challenge studies in vivo, primary cells
genotype II. have provided important novel insights into other gastrointes-
In the present study a combination of established in vitro tinal pathogens, including Campylobacter and Helicobacter spp.
tissue culture and an ex vivo primary intestinal cell model of (9, 26). Comparison of Cryptosporidium infection of primary
Cryptosporidium infection was used to compare infection by C. and conventional cell lines in the present study demonstrates
parvum genotype I and genotype II. The data presented here the potential usefulness of primary cells for the study of C.
indicate that C. parvum genotype I differs from genotype II parvum infection. For example, our results indicate that the
with regard to the efficiency and pattern of infection and sus- apparent efficiency of C. parvum infection in vitro may be
ceptibility to microtubule depolymerizing drugs. In addition, critically dependent on the infection model used. A previous
we have shown that C. parvum genotype I isolates are unable to study suggested that genotype I appeared to be more aggres-
infect bovine intestinal cells ex vivo, indicating that host cell sive in its growth in HCT-8 cells than genotype II (19). How-
tropism underlies species restriction among human and bovine ever, in the present study, entry of a C. parvum genotype I
genotypes of C. parvum. isolate into human-derived HCT-8 cells was less efficient than

FIG. 6. Infection of primary bovine intestinal cells with genotype I C. parvum isolate 5942 (A) and genotype II C. parvum (B). Whereas
genotype II C. parvum efficiently invaded bovine cells, genotype I did not. Slides were examined by using a fluorescence microscope. Magnification,
"400.
6130 HASHIM ET AL. INFECT. IMMUN.

that of a genotype II isolate, and the genotype I isolate showed sporozoite lectin-carbohydrate interactions (7) that subserve
a distinct, focal pattern of infection in the monolayer. The these distinct attachment specificities will be an important area
reason for this difference in pattern of infection of HCT-8 cells for future investigation.
by genotype I and genotype II C. parvum is uncertain, but it C. parvum genotype I isolates have only rarely been de-
may reflect different adhesin-receptor interactions subserving scribed in nonhuman primates, sheep, and dugongs (14, 27, 39,
invasion by the two genotypes of these immortalized cells. On 46). Under experimental conditions genotype I isolates can
the other hand, in a primary human cell model the two isolates infect gnotobiotic piglets, and mixed (genotype I and II) infec-
were indistinguishable in terms of the pattern and efficiency of tions have been described following propagation studies in
infection. More importantly, our primary cell model clearly calves (41). However, from an epidemiological perspective this
shows that bovine cells are resistant to infection by C. parvum subtype of C. parvum is considered to infect humans almost
genotype I. exclusively (45). Under both clinical and experimental condi-
In order to avoid the confounding effect of variation in tions C. parvum genotype I and II isolates maintain a separate
isolate pathogenicity (30, 31, 42) on the interpretation of our reproductive cycle, indicating a lack of genetic recombination
results, we used two geographically distinct isolates of C. par- between them (45). Based on these and other molecular ge-
vum genotype I for studies on primary cells. Neither isolate netic and biological data (28), there now exists compelling
could enter primary intestinal bovine cells. Both entered pri- evidence that C. parvum genotype I is a separate species from
mary human small bowel cells with similar efficiency. That the C. parvum genotype II. Our results showing species-restricted
observed tropism was not simply a result of host cell adaptation host cell tropism adds further support to the proposal that C.
is underlined by the fact that genotype I isolate, TU502, was parvum genotype I should be redesignated as a novel species,
extensively passaged through gnotobiotic neonatal pigs (a spe- Cryptosporidium hominis (28, 45).
cies successfully infected with C. parvum genotype I experi- In conclusion, the data presented here show that the patho-
mentally) before use. genesis of C. parvum genotype I differs from that of genotype
The results of previous studies of the effects of microtubule II. C. parvum genotype I was less infective in HCT-8 cells, and
inhibitors on Cryptosporidium infectivity have been contradic- its infectivity was not affected by the addition of microtubule
tory (7, 48). Our data support existing reports indicating that depolymerizing drugs. Furthermore, C. parvum genotype I was
the host microtubule cytoskeleton is not involved in C. parvum not able to infect primary bovine intestinal cells, indicating that
infection (7). In addition, our results are in agreement with the host species restriction of this genotype reflects tissue tro-
those of Wiest et al. (48), who showed that pretreatment of C. pism in vivo. It will be important now to investigate and com-
parvum oocysts with colchicine and vincristine attenuated in- pare with C. parvum genotype II the relative effects of C.
fection in a dose-dependent manner. However, in the present parvum genotype I on signal transduction events (6, 7, 11, 12),
study the infectivity in vitro of a C. parvum genotype I isolate cell death (5, 8), and the inflammatory response (20) in existing
was not affected by preincubation with colchicine or vincris- immortalized cell models and primary intestinal cells.
tine. It is possible that genotype I does not require its micro-
tubule cytoskeleton for cell invasion. However, an alternative ACKNOWLEDGMENTS
explanation could be a variation in oocyst permeability of the
This study was supported by the Irish Health Research Board and
two C. parvum genotypes. Whether isolates of C. parvum differ the Children’s Research and Medical Foundation.
in their requirement for microtubule function during the in- We thank Kristin Elwin and Anne Thomas for maintenance and
fection process or vary in their sensitivity to the effects of genotyping of the National Collection of Oocysts at the Cryptospo-
inhibiting agents requires further study. However, these data ridium Reference Laboratory. We are grateful to Saul Tzipori, De-
suggest that the parasite’s microtubule structure is unlikely to partment of Geographical Medicine, Tufts University, for providing
TU502 C. parvum isolates.
be an attractive target for developing novel human anticrypto-
sporidial therapies. REFERENCES
It is possible that C. parvum genotype II uses a pathway 1. Akiyoshi, D. E., X. Feng, M. A. Buckholt, G. Widmer, and S. Tzipori. 2002.
distinct from that of C. parvum genotype I for infection of both Genetic analysis of a Cryptosporidium parvum human genotype 1 isolate
human and bovine intestinal cells. Alternatively, C. parvum passaged through different host species. Infect. Immun. 70:5670–5675.
2. Awad-el-Kariem, F. M., H. A. Robinson, D. A. Dyson, D. Evans, S. Wright,
genotype II may have evolved a specialized mechanism of M. T. Fox, and V. McDonald. 1995. Differentiation between human and
host-parasite interaction specific for infecting bovine intestinal animal strains of Cryptosporidium parvum using isoenzyme typing. Parasitol-
ogy 110:129–132.
cells that is not shared by C. parvum genotype I. Other api- 3. Casemore, D. 1991. Broadsheet 128: laboratory methods for diagnosing
complexan invasion mechanisms have been shown to have a cryptosporidiosis. J. Clin. Pathol. 44:445–451.
high degree of redundancy (29, 32). It is possible that genotype 4. Chen, X., S. A. Levine, P. L. Splinter, P. S. Tietz, A. L. Ganong, C. Jobin,
G. J. Gores, C. V. Paya, and N. F. LaRusso. 2001. Cryptosporidium parvum
I has become specifically adapted for human cells and there- activates nuclear (B in biliary epithelia preventing epithelial cell apoptosis.
fore has less redundancy in invasion pathways than genotype Gastroenterology 120:1774–1783.
II. Our studies with microfilament inhibition and confocal mi- 5. Chen, X. M., G. J. Gores, C. V. Paya, and N. F. LaRusso. 1999. Cryptospo-
ridium parvum induces apoptosis in biliary epithelia by a Fas/Fas ligand-
croscopy showed that, like C. parvum genotype II (6, 7), the dependent mechanism. Am. J. Physiol. 277:G599–G608.
human genotype requires the participation of the host cell 6. Chen, X. M., B. Q. Huang, P. L. Splinter, H. Cao, G. Zhu, M. A. McNiven,
and N. F. LaRusso. 2003. Cryptosporidium parvum invasion of biliary epithe-
actin cytoskeleton for entry into cells. This, together with the lia requires host cell tyrosine phosphorylation of cortactin via c-Src. Gastro-
finding that C. parvum genotype I did not attach to bovine enterology 125:216–228.
intestinal cells, strongly suggests that the species tropism we 7. Chen, X. M., and N. F. LaRusso. 2000. Mechanisms of attachment and
internalization of Cryptosporidium parvum to biliary and intestinal epithelial
observed occurred at the level of receptor-ligand interaction cells. Gastroenterology 118:368–379.
during the initial stages of infection. Exploration of potential 8. Chen, X. M., S. A. Levine, P. Tietz, E. Krueger, M. A. McNiven, D. M.
VOL. 72, 2004 HOST CELL TROPISM AND SPECIES RESTRICTION OF C. PARVUM 6131

Jefferson, M. Mahle, and N. F. LaRusso. 1998. Cryptosporidium parvum is hominis n.sp. (Apicomplexa: Cryptosporidiae) from Homo sapiens. J. Eu-
cytopathic for cultured human biliary epithelia via an apoptotic mechanism. karyot. Microbiol. 49:433-440.
Hepatology 28:906–913. 29. Morrissette, N. S., and L. D. Sibley. 2002. Cytoskeleton of apicomplexan
9. Clyne, M., and B. Drumm. 1993. Adherence of Helicobacter pylori to primary parasites. Microbiol. Mol. Biol. Rev. 66:21-38.
human gastrointestinal cells. Infect. Immun. 61:4051–4057. 30. Okhuysen, P. C., S. M. Rich, C. L. Chappell, K. A. Grimes, G. Widmer, X.
10. Deng, M. Q., and D. O. Cliver. 1998. Cryptosporidium parvum development Feng, and S. Tzipori. 2002. Infectivity of a Cryptosporidium parvum isolate of
in the BS-C-1 cell line. J. Parasitol. 84:8–15. cervine origin for healthy adults and interferon-gamma knockout mice. J. In-
11. Elliot, D., D. J. Coleman, M. A. Lane, R. C. May, L. M. Machesky, and D. P. fect. Dis. 185:1320-1325.
Clark. 2001. Cryptosporidium parvum infection requires host actin polymer- 31. Okhuysen P. C., C. L. Chappell, J. H. Crabb, C. R. Sterling, and H. L.
ization. Infect. Immun. 69:5940–5942. DuPont. 1999. Virulence of three distinct Cryptosporidium parvum isolates
12. Elliott, D. A., and D. P. Clark. 2000. Cryptosporidium parvum induces host for healthy adults. J. Infect. Dis. 180:1275-1281.
cell actin accumulation at the host-parasite interface. Infect. Immun. 68: 32. Opitz, C., and D. Soldat. 2002. “The glideosome”: a dynamic complex pow-
2315–2322. ering gliding motion and host cell invasion by Toxoplasma gondii. Mol.
13. Fayer, R., C. A. Speer, and J. P. Dubey. 1997. The general biology of Microbiol. 45:597-604.
Cryptosporidium, p. 1–42. In R. Fayer (ed), Cryptosporidium and cryptospo- 33. Peng, M. M., M. L. Wilson, R. E. Holland, S. R. Meshnick, A. A. Lal, and L.
ridiosis. CRC, Press, Boca Raton, Fla. Xiao. 2003. Genetic diversity of Cryptosporidium spp. in cattle in Michigan:
14. Giles, M., K. A. Webster, J. A. Marshall, J. Catchpole, and T. M. Goddard. implications for understanding the transmission dynamics. Parasitol. Res.
2001. Experimental infection of a lamb with Cryptosporidium parvum geno- 90:175-180.
type 1. Vet. Rec. 149:523-525. 34. Pereira, S. J., N. E. Ramirez, L. Xiao, and L. A. Ward. 2002. Pathogenesis of
15. Glaberman, S., J. E. Moore, C. J. Lowery, R. M. Chalmers, I. Sulaiman, K. human and bovine Cryptosporidium parvum in gnotobiotic pigs. J. Infect. Dis.
Elwin, P. J. Rooney, B. C. Millar, J. S. Dooley, A. A. Lal, and L. Xiao. 2002. 186:715-718.
Three drinking water-associated cryptosporidiosis outbreaks, Northern Ire- 35. Rossignol, J. F., A. Ayoub, and M. S. Ayers. 2001. Treatment of diarrhea
land. Emerg. Infect. Dis. 8:631-633. caused by Cryptosporidium parvum: a prospective randomized, double-blind,
16. Griffiths, J. K. 1998. Human cryptosporidiosis: epidemiology, transmission, placebo-controlled study of nitazoxanide. J. Infect. Dis. 184:103-106.
clinical disease, treatment and diagnosis. Adv. Parasitol. 40:38-72. 36. Ryley, J. F., R. Mead, J. Hazelhurst, and T. E. Robinson. 1976. Methods in
coccidiosis research: separation of oocysts from faeces. Parasitology 73:311-
17. Guerrant, R. L. 1997. Cryptosporidiosis: an emerging, highly infectious
326.
threat. Emerg. Infect. Dis. 3:51-57.
37. Sestak, K., L. A. Ward, A. Sheoran, X. Feng, D. E. Akiyoshi, H. D. Ward, and
18. Gut, J., and R. G. Nelson. 1999. Cryptosporidium parvum: lectins mediate
S. Tzipori. 2002. Variability among Cryptosporidium parvum genotype 1 and
irreversible inhibition of sporozoite infectivity in vitro. J. Eukaryot. Micro-
2 immunodominant surface glycoproteins. Parasite Immunol. 24:213-219.
biol. 46:48S–49S.
38. Spano, F., C. Puri, L. Ranucci, L. Putignani, and A. Cristanti. 1997. Cloning
19. Hijjawi, N. S., B. P. Meloni, U. M. Morgan, and R. C. Thompson. 2001.
of the entire COWP gene of Cryptosporidium parvum and ultrastructural
Complete development and long-term maintenance of Cryptosporidium par-
localization of the protein during sexual parasite development. Parasitology
vum human and cattle genotypes in cell culture. Int. J. Parasitol. 31:1048-
114:427-437.
1055.
39. Spano, F., L. Putignani, A. Cristanti, P. Sallicandro, U. M. Morgan, S. M. Le
20. Laurent, F., D. McCole, L. Eckmann, and M. F. Kagnoff. 1999. Pathogenesis Blancq, L. Tchack, S. Tzipori, and G. Widmer. 1998. Multilocus genotypic
of Cryptosporidium parvum infection. Microbes Infect. 1:141-148. analysis of Cryptosporidium parvum isolates from different hosts and geo-
21. Leav, B. A., M. R. Mackay, A. Anyanwu, R. M. O’Connor, A. M. Cevallos, G. graphical origins. J. Clin. Microbiol. 36:3255-3259.
Kindra, N. C. Rollins, M. L. Bennish, R. G. Nelson, and H. D. Ward. 2002. 40. Sulaiman, I. M., L. Xiao, C. Yang, L. Escalante, A. Moore, C. B. Beard, M. J.
Analysis of sequence diversity at the highly polymorphic Cpgp40/15 locus Arrowood, and A. A. Lal. 1998. Differentiating human from animal isolates
among Cryptosporidium isolates from human immunodeficiency virus-in- of Cryptosporidium parvum. Emerg. Infect. Dis. 4:681-685.
fected children in South Africa. Infect. Immun. 70:3881-3890. 41. Tanriverdi, S., M. O. Arslan, D. E. Akiyoshi, S. Tzipori, and G. Widmer.
22. O’Donoghue, P. J. 1995. Cryptosporidium and cryptosporidiosis in man and 2003. Identification of genotypically mixed Cryptosporidium parvum popula-
animals. Int. J. Parasitol. 25:139-195. tions in humans and calves. Mol. Biochem. Parasitol. 130:13-22.
23. McCole, D. F., L. Eckmann, F. Laurent, and M. F. Kagnoff. 2000. Intestinal 42. Teunis, P. F., C. L. Chappell, and P. C. Okhuysen. 2002. Cryptosporidium
epithelial cell apoptosis following Cryptosporidium parvum infection. Infect. dose response studies: variation between isolates. Risk Anal. 22:175-183.
Immun. 68:1710-1713. 43. Tumwine, J. K., A. Kekitiinwa, N. Nabukeera, D. E. Akiyoshi, S. M. Rich, G.
24. McLauchlin, J., S. Pedraza-Diaz, C. Amar-Hoetzeneder, and G. L. Nichols. Widmer, X. Feng, and S. Tzipori. 2003. Cryptosporidium parvum in children
1999. Genetic characterization of Cryptosporidium strains from 219 patients with diarrhea in Mulago Hospital, Kampala, Uganda. Am. J. Trop. Med.
with diarrhea diagnosed as having a sporadic cryptosporidiosis. J. Clin. Mi- Hyg. 68:710-715.
crobiol. 37:5153-5158. 44. Tzipori, S., and J. K. Griffiths. 1998. Natural history and biology of Crypto-
25. McLauchlin, J., C. Amar-Hoetzeneder, S. Pedraza-Diaz, and G. L. Nichols. sporidium parvum. Adv. Parasitol. 40:5-36.
2000. Molecular epidemiological analysis of Cryptosporidium species in the 45. Tzipori, S., and H. Ward. 2002. Cryptosporidiosis: biology, pathogenesis, and
United Kingdom: results of genotyping Cryptosporidium spp. in 1,705 fecal disease. Microbes Infect. 4:1047-1058.
samples from human and 105 samples from livestock animals. J. Clin. Mi- 46. Widmer, G., D. Akiyoshi, M. A. Buckholt, X. Feng, S. M. Rich, K. M. Deary,
crobiol. 38:3984-3990. C. A. Bowman, P. Xu, Y. Wang, G. A. Buck, and S. Tzipori. 2000. Animal
26. Mooney, A., C. Byrne, M. Clyne, K. Johnson-Henry, P. Sherman, and B. propagation and genomic survey of a genotype 1 isolate of Cryptosporidium
Bourke. 2003. Invasion of human epithelial cells by Campylobacter upsalien- parvum. Mol. Biochem. Parasitol. 108:187-197.
sis. Cell. Microbiol. 5:835-847. 47. Widmer, G., S. Tzipori, C. J. Fichtenbaum, and J. K. Griffiths. 1998. Geno-
27. Morgan, U. M., L. Xiao, B. D. Hill, P. O’Donoghue, J. Limor, A. Lal, and typic and phenotypic characterization of Cryptosporidium parvum isolates
R. C. A. Thompson. 2000. Detection of Cryptosporidium parvum “human” from people with AIDS. J. Infect. Dis. 178:834-840.
genotype in a dugong (Dugong dugon). J. Parasitol. 86:1352-1354. 48. Wiest, P. M., J. H. Johnson, and T. P. Flanigan. 1993. Microtubule inhibitors
28. Morgan-Ryan, U. M., A. Fall, L. A. Ward, N. Hijjawi, I. Sulaiman, R. Fayer, block Cryptosporidium parvum infection of a human enterocyte cell line.
R. C. Thompson, M. Olson, A. Lal, and L. Xiao. 2002. Cryptosporidium Infect. Immun. 61:4888-4890.

Editor: W. A. Petri, Jr.