Over-expression and localization of a host protein on the membrane of Cryptosporidium parvum infected epithelial cells

Abstract

The genus Cryptosporidium includes several species of intestinal protozoan parasites which multiply in intestinal epithelial cells. The impact of this infection on the transcriptome of cultured host cells was investigated using DNA microarray hybridizations. The expression of 14 genes found to be consistently up- or down-regulated in infected cell monolayers was validated with RT PCR. Using immunofluorescence we examined the expression of Protease Activated Receptor-2, which is encoded by one of the up-regulated genes. In infected cells this receptor localized to the host cell membrane which covers intracellular trophozoites and meronts. This observation indicates that the composition of the host cell membrane is affected by the developing trophozoite, a phenomenon which has not been described previously.

1. Introduction

Cryptosporidiosis, a cosmopolitan diarrheal illness, is caused by several species of intestinal protozoa belonging to the genus Cryptosporidium. Cryptosporidium parvum primarily infects the epithelium of the small intestine, which typically results in acute, self-limited diarrhea. Immunodeficient individuals and young children are more vulnerable and may develop a persistent infection [1, 2]. Unlike infections caused by other pathogenic Apicomplexa, such as the parasite causing malaria, toxoplasmosis and coccidiosis, there is no reliable treatment against cryptosporidiosis. Understanding the interaction of the pathogen with the host is expected to facilitate the development of new therapies.

Several studies have documented morphological changes in host cells infected with C. parvum [3-5]. Studies of host cell invasion by C. parvum sporozoites using electron microscopy indicate that the microvilli are incorporated into the host membrane which covers the developing trophozoite [6]. Accumulation of host cell actin and actin-binding proteins at the host parasite interface was observed during parasite attachment and invasion, suggesting that C. parvum also induces the rearrangement of the host cell cytoskeleton [7, 8]. Host cell apoptosis has been demonstrated to occur in response to C. parvum infection in different cell lines [9-12]. Secretion of cytokines and chemokines by infected cells has also been described [13, 14].

Microarrays have been used to study protozoal infections, including those caused by Toxoplasma, Plasmodium, Eimeria, and Trypanosoma species, which all develop intracellularly. These studies have found that the expression of genes associated with cytokine and chemokine production is typically altered in response to infection [15-17]. The expression of host cell genes controlling cellular metabolism, cell adhesion, and DNA repair can also be affected, as is the expression of apoptosis related genes [18, 19]. Up-regulation of cell cycle related genes were identified in Toxoplasma gondii infected cell monolayers [15]. These data not only illustrate the complexity of the cellular processes modulated in response to intracellular protozoa, but also highlight commonalities in the transcriptional response to different intracellular pathogens.

Microarray analyses of the transcriptome of C. parvum infected cells [13, 20, 21] have found genes associated with the inflammatory response, cytoskeleton function, cell proliferation and apoptosis to be differentially expressed in response to C. parvum infection. Here we describe changes in host cell gene expression in C. parvum infected cell monolayers using microarrays, and present evidence that one over-expressed membrane protein localizes to the cellular membrane overlaying the developing parasite. These observations demonstrate for the first time that the composition of the host cell membrane is affected by this infection.

2. Materials and Methods

2.1. Cell culture

Human ileocecal epithelial cells (HCT-8) (American Type Culture Collection #CCL-244) were cultured in RPMI 1640 (GIBCO/Invitrogen, Parsley, United Kingdom) supplemented with 10% heat-inactivated horse serum (GIBCO), 1% penicillin/streptomycin, 2 mM L-glutamine, and 1 mM sodium pyruvate (Sigma, St. Louis, MO). Cells were maintained in a 5% CO₂ atmosphere at 37°C and 85% humidity. Monolayers were seeded in 25 or 75 cm² flasks at 1×10⁶ cells/25 cm² or 3×10⁶/75 cm² and grown to 80% confluence.

2.2. Parasites

C. parvum oocysts were obtained from experimentally infected calves or immunosuppressed rodents. Oocysts were purified from stool as described [22]. Oocysts for inoculation of cell monolayers were surface-sterilized with 10% bleach for 7 min on ice and washed twice with phosphate buffered saline (PBS). Oocysts for mock infections were heat-inactivated at 80°C for 20 min.

2.3. Cell monolayer infection, immunofluorescent labeling and flow cytometry

HCT-8 cells were infected with C. parvum oocysts at a 1/1 of oocyst/host cell ratio and incubated for 24 h. Taurocholic acid at a concentration of 0.05% was added with the oocysts to promote infection [23]. After incubation, cell monolayers were recovered by treating with Accutase (Innovation Cell Technologies, Inc) and processed for immunofluorescence labelling with a Cryptosporidium-specific antiserum as described [24]. Labelled cells were re-suspended at a concentration of 10⁶/ml in PBS and were analyzed by flow cytometry (FACSCalibur; BD Biosciences, San Jose, California) equipped with a 488 nm argon laser. Alexa 488 fluorescence emitted by the secondary anti-rabbit IgG antibody conjugate was acquired in the FL1 channel. FL1 and scatter signal amplifiers were set in log mode. Cell populations were displayed on scatter plots using CellQuest software (BD Biosciences).

2.4. RNA extraction for microarrays

RNA was extracted from about 3×10⁶ cells using 1 ml TRI reagent (Sigma, St. Louis, Missouri). RNA was precipitated in isopropanol, and converted to double-stranded cRNA. cRNA was transcribed with T7 RNA polymerase in the presence of biotinilitated UTP analogue and purified to remove unincorporated nucleotides. Fragmented cRNA was added to hybridization cocktail and heated to 99°C. Oligonucleotide microarrays containing 14,500 human genes (Human Genome U133A 2.0; Affymetrix, Santa Clara, CA) were hybridized at 45°C for 16 h while being rotated at 60 rpm. Arrays were then washed, stained with streptavidin-phycoerythrin (Molecular Probes, Eugene, OR), and scanned.

2.5. Microarrays and data analysis

Background subtraction and normalization of Affymetrix U133A 2.0 microarray data were carried out using the Robust Multichip Average (RMA; [25]) module in BioConductor. A two-class paired Significance Analysis of Microarrays (SAM) module in TIGR Microarray Data Analysis System (MIDAS, version 2.19) [26], which utilizes a modified t-test statistic and sample-label permutations to evaluate statistical significance, was used to identify differentially regulated genes in C. parvum infected vs. mock infected monolayers. The false-discovery rate [27], an estimate of the fraction of falsely called infection-responsive genes, was set at 4.9%.

The list of differentially regulated genes identified in Affymetrix U133A 2.0 was further subdivided into functional categories using the bioinformatics analysis resource DAVID (Database for Annotation, Visualization and Integrated Discovery) of the National Institute of Allergy and Infectious Diseases [28, 29]. The Modified Fisher Exact test was used to measure the probability of observing a number of differentially regulated genes for a specific category given a category’s representation on the array.

2.6. Reverse-transcriptase PCR

Fourteen (of 147) significantly up- or down-regulated genes were randomly selected for reverse-transcriptase (RT) PCR analysis based on them belonging to different biological processes. Primers were designed to amplify fragments of 200-350 basepair from the selected transcripts (Table 1). A portion of 10 ng of the same RNA as used for array hybridization was reverse transcribed using M-MLV RT (Promega, Madison, Wisconsin). Control amplification reactions lacking reverse transcriptase were included to control for any DNA contamination. The cDNA product was added to 10 μl Fast Start DNA Master SYBR Green I master mix (Roche Applied Science, Indianapolis, Indiana) and amplified using a LightCycler® 2.0 thermal cycler. Triplicate samples of a ten-fold dilution series of each cDNA were amplified. Amplification conditions were as follows; initial denaturation at 95°C for 10 min; 40-45 cycles of 95°C for 1 sec, specific annealing temperature for 15 sec, and 1 sec/25 basepair extension at 72°C ; A melting cycle from 65°C to 95°C was applied to acquire melting curves. In all cases HCT-8 DNA was included as a positive PCR control and HCT-8 cDNA was used to draw standard curves.

Table 1.

PCR primers

ATAD4a	FORb	5′- GAG CAG TCG GGA TTA CAG -3′
	REV	5′- CTT CTC CAG GCG GGT GTT A -3′
BTG2	FOR	5′- CAT TCG CAT CAA CCA CAA -3′
	REV	5′- CTG GAG ACT GCC ATC A -3′
CCL5	FOR	5′- CCT CGG ACA CCA CAC CCT -3′
	REV	5′- GGA CAA GAG CAA GCA GAA AC -3′
EGFR	FOR	5′- CCA AGG CAC GAG TAA CAA G -3′
	REV	5′- GAC TGC TAA GGC ATA GGA A -3′
EMP1	FOR	5′- GTA TCA CCA CGG CTA TTC -3′
	REV	5′- GGT TTG GGA TTT GAC CTC -3′
F2RL1	FOR	5′- ATG GCA ACA ACT GGA TTT AT -3′
	REV	5′- TCA CGA CAT ACA AAG GG -3′
FGF18	FOR	5′- CTG CTG CTG TGC TTC CA -3′
	REV	5′- GGC ATA CTT GTC CCC AT -3′
FOS	FOR	5′- CTT CCT TCG TCT TCA CCT A -3′
	REV	5′- ATG CGT TTT GCT ACA TCT CC -3′
GAPDH	FOR	5′- AAT GAC CCC TTC ATT GAC C -3′
	REV	5′- TGG TGC AGG AGG CAT TGC -3′
KLK10	FOR	5′- CCA TAA AGT CAT ACG CTC C -3′
	REV	5′- GCC CAA AGT CAC ACA G -3′
KRT4	FOR	5′- GCA GTT CTT AGA GCA ACA -3′
	REV	5′- CTC CGC ATC ATA GAG G -3′
MALL	FOR	5′- ATG AAC TTG CTC TTC TAG -3′
	REV	5′- GCT AAT ACG GTG AAA CC -3′
PHLDA1	FOR	5′- GTT TGG ACA TCA CCC TAC T -3′
	REV	5′- GCT GGC ACA ACA ATG AAA G -3′
PLAUR	FOR	5′- GTG TAA GAC CAA CGG GGA TTG -3′
	REV	5′- GCT TCG GGA ATA GGT GAC AG -3′
SDHA	FOR	5′- CAT CCA CTA CAT GAC GGA GCA -3′
	REV	5′- ATC TTC CCA TCT TCA GTT CTG CTA -3′
TNFRSF12A	FOR	5′- GGA CCT GGA CAA GTG C -3′
	REV	5′- GAA TGG ATG AAT GAA TGA TGA GT -3′

^aFor gene titles see Table 2.

^bFOR, forward primer; REV, reverse primer

2.7. Immunofluorescence of Protease Activated Receptor-2

HCT-8 cells were infected with C. parvum oocysts (ratio 1/1) and processed 24 h post-infection for immunofluorescent labeling as described above for C. parvum labelling with the following modifications: methanol fixed cells were incubated with a 1:200 dilution of Cryptosporidium specific 2E5 mouse monoclonal primary antibody and a 1:200 dilution of anti-Protease Activated Receptor-2 (PAR₂) rabbit antibody (abcam, Inc.). This antibody recognizes a peptide located in the PAR₂ cytoplasmic carbox-terminal region. The sequence of the peptide is not disclosed by the company. Subsequently, the cells were incubated with a 1:1000 dilution of secondary Alexa Fluor 488 labeled rabbit anti-mouse IgG antibody pre-adsorbed against human IgG and a 1:1000 dilution of Alexa Fluor 568 conjugated goat anti-rabbit IgG antibody pre-adsorbed against goat, mouse, and human IgG. Labelled cells were examined by fluorescence microscopy. Uninfected cells labeled with anti-PAR₂ or anti-Cryptosporidium antibody were used to assess background fluorescence. The lack of cross-reactivity of the anti-PAR₂ antibody with parasite antigen was investigated by Western blotting. Briefly, lysates of 10⁶ oocysts or of 3 × 10⁵ infected, mock-infected, and uninfected HCT-8 cells were resolved by SDS-PAGE under reducing conditions on a 4-12% PAGE Bis-Tris gel and protein electrophoretically transferred to nitrocellulose. The blot was incubated with the anti-PAR₂ antibody or control rabbit serum as negative control. Bound antibody was visualized with a horseradish peroxidase secondary anti-rabbit antibody and a luminol based enhanced chemiluminescence peroxidase substrate.

To examine the time course PAR₂ expression, HCT-8 cells were infected with C. parvum oocysts (oocysts/host cell ratio 1/1) and processed 2, 6, 10, 15, 20, 24, 30, 35, 40, and 48 h post-infection for immunofluorescence labelling with C. parvum anti-2E5 antibody or anti-PAR₂ antibody as described above. Labelled cells were examined by microscopy and flow cytometry. CellQuest Pro Software was used to analyze the flow cytometry data.

3. Results

3.1. Analysis of global gene expression

In triplicate experiments with Affymetrix HG-U133A 2.0 arrays, and based on a false-discovery rate of 4.95%, a total of 147 genes were found to be differentially regulated 24 h after infection with C. parvum, including 73 up-regulated and 74 down-regulated genes. Fig. 1 shows the relative expression of 14 randomly selected genes. At 24 h post-infection the following gene ontology (GO) biological process terms were found to be over-represented among up-regulated genes (p-value ≤ 5.2E-4, EASE Score): Cell Communication, Signal Transduction, Organ Development, and Response to External Stimulus. Biological process terms found to be over-represented among down-regulated genes (p-value ≤ 8.6E-3, EASE Score) include Morphogenesis of an Epithelium and Amino Acid Metabolic Processes.

Fig. 1. Gene expression assessed by microarrays and quantitative RT PCR.

Fourteen mRNA species were RT PCR amplified using the primers shown in Table 1. For each sample, expression values were normalized against SDHA mRNA and relative expression values obtained by comparing RNA extracted from corresponding infected and uninfected monolayer. RT PCR (grey bars) was compared with microarray data (black bar) (n = 3 experiments). For gene symbols see Table 2.

3.2. Confirmatory RT PCR analysis

Portions of the same RNA samples analyzed in triplicate Affymetrix microarrays were used as substrate for RT PCR to compare the expression of the 14 selected genes (Table 2) in infected and mock infected monolayers. These genes were randomly selected among the 30 most up-regulated and 17 most down-regulated genes. The concentration of each mRNA transcript was estimated using a standard curve and normalized against the concentration of GAPDH (not shown) or succinate dehydrogenase complex subunit A (SDHA) mRNA. The RT PCR results were in good agreement with the microarray data (Fig. 1). Normalization against GAPDH or SDHA did not significantly change the results.

Table 2.

Gene symbol and title of 14 genes selected for RT PCR analysis

Gene Symbol	Gene Title	fold change (log2 )	GenBank accession
ATAD4	ATPase family, AAA domain containing 4	-1.32	NM_024320
BTG2	BTG family, member 2	-1.21	NM_006763
CCL5	chemokine (C-C motif) ligand 5	0.87	NM_002985
EGFR	epidermal growth factor receptor	0.84	NM_005228
EMP1	epithelial membrane protein 1	0.58	NM_001423
F2RL1	coagulation factor II (thrombin) receptor-like 1	0.94	NM_005242
FGF18	fibroblast growth factor 18	-0.60	NM_003862
FOS	v-fos FBJ murine osteosarcoma viral oncogene homolog	-1.87	NM_005252
KLK10a	kallikrein-related peptidase 10	0.76	NM_002776
KRT4	keratin 4	-1.09	NM_002272
MALLa	mal, T-cell differentiation protein-like	1.56	NM_005434
PHLDA1	pleckstrin homology-like domain, family A, member 1	1.42	NM_007350
PLAUR	plasminogen activator, urokinase receptor	0.77	NM_002659
TNFRSF12A	tumor necrosis factor receptor superfamily, member 12A	1.16	NM_016639

^aThese two genes are not among the 147 most differentially regulated genes. They were included in the RT PCR validation based on a preliminary set of microarray data.

3.3. Immunofluorescence analysis of F2RL1 gene products

Based on the microarray and RT PCR quantification described above, immunofluorescence was used to examine the expression of PAR₂, the protein encoded by gene F2RL1. This choice was motivated by the fact that PAR₂ may play an important role in the innate immune response, as further discussed below. From a practical point of view, this analysis was facilitated by the availability of antibody specific for human PAR₂. Infected HCT-8 cells were processed for immunofluorescence 24 h post-infection, and reacted with fluorescent anti-PAR₂ antibody and anti-C. parvum antibody as described in Materials and Methods. Labelling with secondary anti-rabbit Alexa 568 conjugate antibody and anti-mouse Alexa 488 conjugate revealed that PAR₂ fluorescence was localized to the cell membrane which surrounds the intracellular stages of the parasite (Fig. 2). Interestingly, PAR₂ fluorescence was strongly concentrated within this well-defined region of host cell membrane. The absence of cross-reaction between anti-PAR₂ rabbit antibody and secondary anti-mouse IgG antibody, or between the anti-Cryptosporidium monoclonal mouse antibody and secondary anti-rabbit IgG antibody was confirmed in control monolayers (not shown). Uninfected cells were used as the negative control and labeled with 2E5 or PAR₂ antibodies, together with their fluorescent secondary antibody. No fluorescence from antibodies against 2E5 or PAR₂ was observed on the uninfected cells (not shown). Infected cells labelled separately with each antibody combination were also examined to confirm that the green and red immunofluorescent signals were specific and that no bleed-over from red to green and vice versa took place. Moreover, a Western blot of lysed C. parvum oocysts, and of infected, uninfected and mock-infected HCT-8 cells were probed with anti-PAR₂ antibody. The three cell lysates displayed an identical profile indicating that the antibody does not recognize any antigen produced by intracelluar parasite stages.

Fig. 2. Immunofluorescence analysis of PAR₂ expression on infected cells.

Cells recovered from a HCT-8 monolayer 24 h post-infection were double-labelled for C. parvum (2E5 monoclonal antidoby) and PAR₂ as indicated. In unlabelled cells the parasites are difficult to discern and their position are indicated with arrows.

The change in PAR₂ fluorescent signal in C. parvum infected cell monolayers was investigated over a 48-h period by flow cytometry using infected and uninfected cells labelled separately with anti-PAR2 or anti-Cryptosporidium antibody. Each cell population was then reacted separately with Alexa 488 conjugated anti-rabbit and anti-mouse IgG, respectively, and green fluorescence emitted by each cell population displayed on histograms (Fig. 3). The parasite immunofluorescence signal increased during the 48-h period, which is consistent with previous reports indicating parasite proliferation in the initial 48-72-h [11, 30]. At 48 h post-infection, the magnitude of parasite immunofluorescence as compared to that emitted by uninfected cells indicates that a majority of host cells was infected. PAR₂ fluorescence also increased during this period, and the PAR₂ signal paralleled parasite fluorescence, reaching a maximum 48 h post-infection. These observations are consistent with the microarray and microscopic observations in showing PAR₂ over-expression in response to the infection.

Fig. 3. Time course of PAR₂ and C. parvum immunofluorescence.

Cells from an infected monolayer were fixed at different time points post-infection and two portions of the cell suspensions immunolabelled for C. parvum antigen or PAR₂, respectively. C. parvum and PAR₂ immunofluorescence was acquired by flow cytometry in the FL1 channel. Uninfected cells labeled with C. parvum or PAR₂ specific antibody were used as negative control. Histogram overlay shows fluorescence of infected (black) and uninfected cells (grey).

4. Discussion

Cell monolayers infected with Cryptosporidium parasites are complex systems in which two eukaryotic species interact. The system is difficult to standardize, as Cryptosporidium oocysts cannot be produced in vitro, but are purified from the feces of experimentally infected animals. Even when conventional laboratory rodent models, such as inbred, immuno-suppressed [31] or -deficient mice [32] are used to propagate C. parvum, the oocysts recovered from the feces are a heterogeneous population which can greatly vary in infectivity [33]. C. parvum isolates are also genetically heterogeneous and readily generate new recombinants with different infectivity phenotypes [34]. Oocyst age has also been shown to affect infectivity [30]. As far as the host cells are concerned, the age of the monolayer can modulate the infection [35]. These variables may have negatively affected the correlation between replicate microarray experiments.

The comparison of infected and control monolayers using microarrays and RT PCR detects host transcripts which are up-regulated in infected cells, as well as transcripts which are up-regulated as a result of changes in the monolayer induced by the infection. Widespread apoptosis, extrusion of infected cells and induction of mitosis to restore monolayer integrity have been documented [9, 36-38]. Consistent with these changes, two of the three published microarray analyses have found that genes involved in cell proliferation and apoptosis are particularly abundant among up-regulated genes in C. parvum infected monolayers [20, 21]. Liu et al. [21] reported that pro-apoptotic genes were induced and anti-apoptotic genes were down-regulated at 24 h post-infection. In contrast, we observe an over-representation of apoptosis-related genes among up- and down-regulated genes at the same time point, which is similar to the observations of Deng et al. [20]. Castellanos-Gonzalez et al. reported the over-expression of five genes in C. parvum or C. hominis infected cell monolayers [13]. Among these genes they found the osteoprotegerin gene, which in our data shows only a modest level of up-regulation (0.31 log₂) and does therefore not feature in the supplemental table. Among the four other up-regulated genes found by Castellanos-Gonzalez and co-workers, only CD24 is found among the up-regulated genes in our study (Supplemental Table). There are thus significant differences among sets of microarray data from C. parvum infected cells, which is not surprising given the complexity of the experimental system and the lack of standardized conditions. More studies performed under similar conditions are needed to discriminate between genes which are modulated as a result of the infection and experimental noise.

Since in C. parvum infected monolayers many cells remain uninfected, we were interested in investigating whether up-regulation of F2RL1 transcription was specific to infected cells, and whether this process resulted in increased PAR₂ expression. PAR₂ was of particular interest in this context as it is widely distributed in the gastrointestinal tract and is involved in regulating mucus production [39-41], a process which could be relevant to the innate host defense against intestinal pathogens. The PAR₂ immunofluorescence analysis confirms a direct link between infection and up-regulation and is significant in showing that the host cell membrane is modified during infection. PAR₂, a G-protein-coupled receptor, is activated by proteolytic cleavage of the N-terminal domain. PAR₂ is expressed in many tissues, including the intestinal epithelium [42-44]. Its involvement in the regulation of salivary gland secretion, pancreatic juice secretion and small ion transport [39, 45, 46] make it potentially relevant for understanding the host’s innate defense against intestinal pathogens. Expression of PAR₂ has also been reported to increase cytosolic Ca²⁺ concentration [43, 46-50] but preliminary observations in our system using Calcium Green-1, AM (Molecular Probes, Eugene, Oregon) failed to demonstrate changes in intracellular Ca²⁺ (data not shown). The observation that agonists of PAR₂ stimulate the secretion of pro-inflammatory cytokines [51, 52], induce NFκB pathway activation, IL-8 production, and activate p-38, MAPK and ERK1/2 signaling [53, 54] may be relevant to understanding the mechanisms involved in C. parvum induced release of cytokines and chemokines from infected cells [55, 56]. Whether these changes are regulated by PAR₂ signaling remains to be investigated. Infections with Helicobacter pylori and Pseudomonas aeruginosa pneumonia can also activate PAR₂ signaling [57, 58]. However, the functional significance of PAR₂ activation in response to Cryptosporidium infection, or how many proteins respond in this manner, is unknown.

The over-expression and localization of host membrane proteins in C. parvum infected cells may have practical implications as a better understanding of the physiology of the cell membrane may shed light on the lack of anti-cryptosporidial activity of many compounds which are effective against other intracellular pathogens [59]. If resistance to such compounds is caused by the inability of certain molecules to reach the trophozoite and meront stages of the parasite, understanding the membrane composition may facilitate the targeted modification of existing anti-protozoal compounds into a form which may better penetrate the cell membrane.

List of abbreviations

RT

reverse-transcriptase

PAR₂

Protease Activated Receptor-2

PAGE

polyacrylamide gel electrophoresis

SDS

sodium dodecyl sulfate

PBS

phosphate buffered saline