Differential Response to the Course of Cryptosporidium parvum Infection and Its Impact on Epithelial Integrity in Differentiated versus Undifferentiated Human Intestinal Enteroids

ABSTRACT

Cryptosporidium is a leading cause of diarrhea and death in young children and untreated AIDS patients and causes waterborne outbreaks. Pathogenic mechanisms underlying diarrhea and intestinal dysfunction are poorly understood. We previously developed stem-cell derived human intestinal enteroid (HIE) models for Cryptosporidium parvum which we used in this study to investigate the course of infection and its effect on intestinal epithelial integrity. By immunofluorescence and confocal microscopy, there was robust infection of undifferentiated and differentiated HIEs in two and three-dimensional (2D, 3D) models. Infection of differentiated HIEs in the 2D model was greater than that of undifferentiated HIEs but lasted only for 3 days, whereas infection persisted for 21 days and resulted in completion of the life cycle in undifferentiated HIEs. Infection of undifferentiated HIE monolayers suggest that C. parvum infects LGR5+ stem cells. Transepithelial electrical resistance measurement of HIEs in the 2D model revealed that infection resulted in decreased epithelial integrity which persisted in differentiated HIEs but recovered in undifferentiated HIEs. Compromised epithelial integrity was reflected in disorganization of the tight and adherens junctions as visualized using the markers ZO-1 and E-cadherin, respectively. Quantitation using the image analysis tools Tight Junction Organizational Rate and Intercellular Junction Organization Quantification, measurement of monolayer height, and RNA transcripts of both proteins by quantitative reverse transcription PCR confirmed that disruption persisted in differentiated HIEs but recovered in undifferentiated HIEs. These models, which more accurately recapitulate human infection, will be useful tools to dissect pathogenic mechanisms underlying diarrhea and intestinal dysfunction in cryptosporidiosis.

INTRODUCTION

The intestinal parasite Cryptosporidium is a leading cause of diarrheal disease and death in young children and untreated HIV/AIDS patients in low- and middle-income countries (1). It is also a major cause of waterborne outbreaks of disease in the United States (2). Due to the potential for waterborne contamination, Cryptosporidium was classified as a category B pathogen for biodefense (3). Despite this global burden of disease, there is only one FDA approved drug, nitazoxanide, for treatment of Cryptosporidium infection. Unfortunately, this drug is not effective in those who are immunocompromised, such as untreated HIV/AIDS patients, and is not approved for treatment in infants (4). Therefore, there is a critical need for new interventions to better treat those who are infected and protect the immunocompromised from severe disease and possibly death.

The scarcity of interventions for cryptosporidiosis is in part due to a lack of appropriate human model systems to study pathogenic mechanisms of the disease (5). Most studies have been done on immortalized intestinal epithelial cell (IEC) lines such as HCT-8, Caco-2, and HT29 cells, which can only sustain infection for 3 to 5 days and in which the parasite can only undergo asexual reproduction (6). We previously developed a three-dimensional silk scaffold-based model using Caco-2 and HT29 cells which allowed for up to 3 weeks of infection and completion of the life cycle of the parasite (7). However, transformed cell lines are known to have significant differences in gene expression levels compared to primary cells and can also undergo chromosomal rearrangements, mutations, and epigenetic changes that alter their phenotypes (8, 9). In addition, most transformed IEC lines only contain a single epithelial cell type, and do not recapitulate the different cell types present in the normal human small intestine (10). Studies in primary human IEC have also shown that C. parvum infection lasts only for 3 to 5 days (11, 12). Thus, there is a need for human IEC models which support long-term infection and completion of the life cycle of the parasite, and which can be used to study pathogenic mechanisms.

Development of stem-cell based organoid systems has revolutionized primary human models for study of many intestinal diseases (13). Seminal work from Sato and Clevers demonstrated that leucine-rich repeat-containing G-protein coupled receptor 5 positive (LGR5⁺) stem cells can be isolated from intestinal crypts of mice or humans and cultured to form self-organizing mini-guts (14, 15). These stem cells can differentiate into enterocytes, goblet cells, enteroendocrine cells, Paneth cells, and tuft cells and accurately recapitulate normal physiology of the human intestinal epithelium (16). Subsequently, other investigators have referred to small intestinal organoids as human intestinal enteroids (HIEs) (17, 18). When grown as monolayers on transwells, HIEs polarize with formation of apical brush border microvilli and display intact tight junctions and basolateral expression of NA+/K+ ATPase, NKCC1, β-catenin, and E-cadherin (10, 17).

The intestinal epithelium provides the first line of defense and plays a critical role in activating and orchestrating host responses to various intestinal disorders such as inflammatory bowel disease (19) as well as infection with intestinal pathogens (20) including C. parvum (21, 22). Diarrhea and inflammation resulting from these disorders are associated with disruption of intestinal epithelial integrity (20). We (23) and others (24) have reported increased intestinal epithelial permeability in Cryptosporidium infection in children in low- and middle-income countries. Studies of C. parvum infection in transformed IEC and a mouse model have also demonstrated disruption of intestinal epithelial barrier integrity (21, 22). However, no studies of intestinal epithelial barrier integrity have been performed using C. parvum infection of organoids or enteroids.

Intestinal epithelial integrity is maintained by the apical junctional complex (AJC), which consists of tight junctions, adherens junctions, and desmosomes (25, 26) (Fig. 1B). The tight junction protein zonula occludens 1 (ZO-1) acts as a scaffold for multiprotein complexes within the cytoplasm and interacts directly with the transmembrane proteins occludin and claudins, anchoring them to the actin/myosin cytoskeleton (26). Adherens junction proteins include E-cadherin and α and β catenins, which play a pivotal role in maintaining structure and mechanical strength (27). Desmosomes include the transmembrane proteins desmoglein and desmocollin, which bind to plakophilin and plakoglobin, respectively, to interact with intermediate filaments within the cells (28).

FIG 1.

A. Two- and three-dimensional (2D and 3D) human intestinal enteroids (HIE) models for C. parvum infection. A. Intact 3D HIEs grown in Matrigel (a) are dissociated into single cells (b) and grown as differentiated (d) or undifferentiated (e) monolayers on transwell filters and infected apically with C. parvum in the 2D model (Modified from reference 33). Alternatively, intact HIEs are fragmented, infected with C. parvum (c), replated in fresh Matrigel and grown as differentiated (f) or undifferentiated (g) intact HIEs in the 3D model. Figure created in part with BioRender.com. B. Apical junctional complex organization: The apical junctional complex is composed of tight and adherens junctions and desmosomes. The tight junction protein zonula occludens 1 (ZO-1) and the adherens junction protein E-cadherin (which are used as representative proteins of these junctions in this study) are linked to the perijunctional actinomyosin ring. Filamentous (f-actin) filaments are present in the core of microvilli. Figure created with BioRender.com. C and D. Markers of differentiated and undifferentiated HIEs in 2D and 3D models: Differentiated or undifferentiated HIEs in 2D (C) or 3D models (D) were fixed and analyzed for intestinal epithelial cell differentiation or stem cell markers by immunofluorescence assays (IFA) and confocal microscopy using antibodies to Sucrose isomaltase (SI; green); chromogranin A (green); mucin 2 (Muc 2; red); lysozyme, (green) and leucine-rich repeat-containing G-protein coupled receptor 5 (LGR5; green). Nuclei were stained with Hoechst 33258 (blue). Scale bars = 10 μM.

Cryptosporidium infection starts with ingestion of oocysts in contaminated food or water. While in the upper intestine, excystation occurs and the oocysts release sporozoites which attach to and invade intestinal epithelial cells. Following invasion, the sporozoites undergo asexual and sexual reproduction within a parasitophorous vacuole that is located in a unique intracellular but extracytoplasmic niche in the microvillous brush border of intestinal epithelial cells (29). Recent studies utilizing intestinal organoids in two (2D) and/or three-dimensional (3D) (30–32) model systems have demonstrated that these models can support C. parvum growth and replication. Using epithelial organoids derived from human small intestine and lung tissues, Heo et al. (31) showed that C. parvum can infect, propagate, and complete its life cycle within the 3D organoids. Using an air liquid interface (ALI) system, Wilke et al. (30) showed that a monolayer of murine intestinal epithelial stem cells reinforced by 3T3 feeder cells grown on transwell filters can sustain C. parvum infection and completion of the life cycle. However, neither of these studies investigated the effect of infection on intestinal barrier integrity in these systems.

In this study, we used these 2D and 3D HIE models (Fig. 1A) to study the course of C. parvum infection in differentiated and undifferentiated cells and its effect on intestinal barrier integrity and apical junctional complex organization (Fig. 1B). In addition, we evaluated whether the 2D HIE model could support long term infection of C. parvum and completion of the life cycle and examined whether the parasite infects LGR5⁺ intestinal stem cells in undifferentiated HIEs in addition to enterocytes in differentiated HIEs.

RESULTS

Two- and three-dimensional HIE models for C. parvum infection.

We used our previously described differentiated and undifferentiated 2D and 3D HIE models (33, 34) (Fig. 1A) for this study. The HIEs used in our study were derived from the ileum (which is the preferred site of infection of C. parvum), whereas Heo et al. (31) used organoids derived from the duodenum. Using specific antibodies, we found that differentiated HIEs grown on transwell filters in the 2D model expressed markers typical of absorptive and secretory cell lineages (Fig. 1C, panels a-d) as described previously (10). Specifically, enterocytes were identified by sucrase-isomaltase (SI), goblet cells by mucin 2 (Muc2), enteroendocrine cells by chromogranin A (ChgA), and Paneth cells by lysozyme. Paneth cells were also present in small numbers in undifferentiated monolayers (Fig. 1C, panel i). LGR5⁺ stem cells were only detected in small numbers in differentiated HIE monolayers (Fig. 1C, panel e). In contrast, most cells in the undifferentiated HIEs were LGR5⁺ (Fig. 1C, panel j). Similar results were obtained in differentiated (Fig. 1D, panels a-d) and undifferentiated HIEs in the 3D models (Fig. 1D, panels e-h; staining for LGR5⁺ stem cells not done). These results are consistent with previous studies on small intestinal organoids or enteroids (35).

2D and 3D models of HIEs support C. parvum infection, which is greater in differentiated HIEs but persists for longer and results in completion of the life cycle in undifferentiated HIEs in the 2D model.

We first determined whether C. parvum could infect differentiated as well as undifferentiated HIEs in both 2D and 3D models. C. parvum was detected in both differentiated and undifferentiated HIEs in the 3D as well as the 2D models at 24, 48 (not shown), and 72 (Fig. 2A) hours postinfection (hpi). At higher magnifications (Fig. 2A, panels c, i, f and l), we observed intracellular stages in both differentiated and undifferentiated HIEs in the 3D model. A similar pattern of infection was seen in both undifferentiated and differentiated HIEs in the 2D model at 72 hpi. In the 2D model, as expected, since C. parvum infects the microvillus brush border of IECs, intracellular parasites can be seen in the apical region of the HIE cells in the orthogonal XZ view (Fig. 2A, panels f and l).

FIG 2.

2D and 3D HIE models support C. parvum infection, which is greater in differentiated HIEs but persists for longer and results in completion of the life cycle in undifferentiated HIEs in the 2D model. A. Differentiated or undifferentiated HIEs in 3D or 2D models were left uninfected or infected with C. parvum for 72 h. Magnified (2.5×) insets c, f, i, and l from panels b, e, h, and k, respectively, show intracellular stages of C. parvum. Arrows in f and l indicate the same parasite in XY and XZ views. B. Differentiated HIE monolayers in the 2D model were infected with C. parvum for 12 h, 24 h, and 72 h. Arrows indicate condensed or fragmented nuclei suggestive of cell death. Undifferentiated HIE monolayers in the 2D model were analyzed after infection with C. parvum for 3, 7, and 21 days. Fixed HIE monolayers (A and B) were analyzed by IFA. C. parvum was stained with mAb 4E9 (green) (A [b, c, h, i]) or anti-gp40 (green) (A [e, f, k, l] and B [a–f]). Nuclei were stained with Hoechst 33258 (blue). Scale bars in A, B and C = 10 μM. C. C. parvum infection in Fig. 2B was quantified in differentiated (blue) and undifferentiated (red) HIE monolayers in the 2D model. Data shown are the mean +/– SE of 3 independent experiments, each performed in triplicate. Statistical analysis of differences in the number of parasites per field between differentiated and undifferentiated HIEs were determined using the Mann-Whitney test (*, P < 0.05; **, P < 0.005; ****, P < 0.0005). D. Supernatants from (a) undifferentiated uninfected HIEs, (b) undifferentiated HIEs infected with C. parvum oocysts, or (c) purified sporozoites for 21 days were fixed on glass slides and analyzed by IFA and oocyst walls stained with Sporo-Glo (green).

To determine whether infection in HIEs persisted longer than in transformed or immortalized IEC lines (3 to 5 days) we analyzed infection in differentiated as well as undifferentiated HIEs in the 2D transwell model for extended periods of time. By immunofluorescence assays (IFA) and confocal microscopy, infection in differentiated HIE monolayers appeared to increase from 12 to 24 h but decreased by 72 hpi (Fig. 2B, panels a-c). At 72 hpi, in addition to fewer intracellular parasites, there were fewer host cell nuclei, many of which were condensed or fragmented, suggesting programmed cell death of host cells previously reported in C. parvum-infected transformed IEC (36). After 72 h, HIEs appeared mostly dead with very few parasites seen (not shown). In contrast, infection persisted for up to 21 days in undifferentiated HIEs with the numbers of intracellular parasites apparently similar at all time points (Fig. 2B, panels d-f). In addition, the number of HIE cells (as indicated by nuclei) appeared constant throughout infection, with no indication of condensed or fragmented nuclei suggestive of cell death.

To quantify these observations regarding infection over time in both differentiated and undifferentiated HIE monolayers, we counted the number of parasites per image at each time point (Fig. 2C). In differentiated HIE monolayers, there was a steady and significant increase in parasite numbers from 8 to 48 h followed by a steep drop at 72 h. In contrast, in undifferentiated HIE monolayers, infection was significantly less than in differentiated HIE monolayers at 12 (P < 0.001) 24 (P < 0.001) and 48 h (P < 0.001) postinfection. However, infection in the undifferentiated HIE monolayers increased slightly after 3 days (P < 0.001), 7 days (P < 0.005), and 14 days (P < 0.001) and remained steady until 21 (P < 0.005) days postinfection.

Since infection persisted for up to 21 days in undifferentiated HIEs, we next determined whether infection in this model leads to formation of new oocysts and completion of the parasite life cycle. To do this, we examined supernatants from the apical side of transwell filters containing undifferentiated HIE monolayers 21 days after infection with either oocysts or purified sporozoites (Fig. 2D). IFA and confocal microscopy using the oocyst wall-specific monoclonal antibody (mAb) Crypt-a Glo showed that oocysts were identified in apical contents of undifferentiated HIEs infected with oocysts or purified sporozoites (Fig. 2D, panels b-c). Although unlikely (since the medium is replaced every 2 days), it is possible that intact oocysts may remain in the apical contents of transwells with oocyst-infected undifferentiated HIE monolayers. However, the finding that intact oocysts were detected in the supernatants of HIEs infected with purified sporozoites clearly indicates completion of the life cycle in this model. We did not assess viability of the newly formed oocysts or infectivity in mice since that was not the objective of this part of the study.

These results indicate that both differentiated and undifferentiated HIEs can support infection and replication of C. parvum. Although infection is less overall in undifferentiated compared to differentiated HIEs, it persists for longer and results in completion of the life cycle. Taken together, these HIE models may provide a more physiological system than culture methods employing transformed or immortalized cell lines.

C. parvum infects SI⁺ enterocytes in differentiated HIEs and LGR5⁺ stem cells in undifferentiated HIEs in the 2D model.

Since infection in undifferentiated HIE monolayers in the 2D model persisted for up to 3 weeks (Fig. 2B, panels d to f) and since we found that most cells in the 2D HIE model were LGR5⁺ stem cells (Fig. 1B), we wanted to determine whether C. parvum directly infects these cells. In addition, we sought to determine if, as expected, the parasite infects enterocytes in differentiated HIEs. We used LGR5 and SI as markers of stem cells and enterocytes, respectively. Costaining of uninfected differentiated and undifferentiated HIEs at 72 hpi revealed that SI was exclusively expressed in differentiated HIEs and LGR5 in undifferentiated HIEs. (Fig. 3A, panels b and c and 3B, panels h and i). Upon infection of differentiated HIE monolayers, as predicted, different intracellular stages of C. parvum were seen infecting SI+ cells (Fig. 3A, panels d, e and f). A likely macrogamont stage of the parasite is indicated by the arrow in Fig. 3A, panels d, e and f, with the arrow in the XZ view showing the apical location. Fig. 3B, panels j, k and l show infection of an LGR5⁺ cell with a likely meront stage, which is also located apically in the XZ view. Also of note is the finding that there were fewer parasite stages seen in undifferentiated compared to differentiated HIEs (Fig. 3, panels d, e, j and k) confirming what we found in Fig. 2C. These results show that the parasite infects SI+ enterocytes in differentiated HIEs and has the capability to infect LGR5⁺ intestinal stem cells in undifferentiated HIEs.

FIG 3.

C. parvum infects SI⁺ enterocytes in differentiated HIEs and LGR5⁺ stem cells in undifferentiated HIEs. Differentiated (A) or undifferentiated (B) HIE monolayers in the 2D model were infected with C. parvum for 72 h, fixed, and analyzed by IFA. Monolayers were costained with anti-LGR5 (red) for stem cells; anti-SI (green) for enterocytes and anti gp40 (cyan) for C. parvum. Nuclei were stained with Hoechst 33258 (blue). Merged images of C. parvum-infected (cyan) HIE monolayers stained with anti-SI (green) indicating enterocytes and anti-LGR5 (red) indicating stem cells are shown in e and k. Magnified images (2.5×) of dotted boxes in e and k are shown in f and l. Arrows indicate the location of the same parasite in the orthogonal XZ plane and the XY plane in the Z-stack images. Scale bars = 10 μM. Images shown are representative of three independent experiments.

C. parvum infection induces disruption of the intestinal epithelial barrier in HIE monolayers which persists in differentiated HIE but recovers in undifferentiated HIEs.

Previous clinical investigation (23, 24) and studies in transformed human IEC (21, 22) indicated that Cryptosporidium infection results in disruption of intestinal epithelial barrier integrity. We therefore sought to determine whether intestinal barrier integrity was also disrupted during C. parvum infection of differentiated or undifferentiated HIE monolayers in the 2D model. To do this, we monitored transepithelial electrical resistance (TEER) across the monolayer from apical to basolateral compartments of transwells (37) prior to and during infection in both differentiated and undifferentiated HIE monolayers. TEER values increased rapidly in uninfected differentiated HIE monolayers, reaching a peak on day 4 (Fig. 4A). However, once monolayers were infected with C. parvum, TEER values decreased significantly by 10% (P < 0.05), 51% (P < 0.001), and 71% (P < 0.001) at 12, 24, and 48 hpi, respectively. Although TEER values of uninfected undifferentiated HIE monolayers did not reach as high levels as differentiated HIEs, infection with C. parvum led to a significant decrease of 30% (P < 0.005) and 68% (P < 0.005) in TEER values at 12 and 24 hpi, respectively (Fig. 4B). In contrast to undifferentiated HIE monolayers, however, there was a steep recovery of TEER values to above baseline (P > 0.05) at 48 hpi. These data indicate that C. parvum infection of both differentiated and undifferentiated HIEs results in disruption of intestinal barrier integrity. However, whereas barrier disruption persists in differentiated HIEs, barrier function recovers by 48 hpi in undifferentiated HIEs.

FIG 4.

C. parvum infection induces disruption of intestinal epithelial barrier integrity in both differentiated and undifferentiated HIE monolayers which persists in differentiated HIEs but recovers in undifferentiated HIEs. Differentiated (A) and undifferentiated (B) HIE monolayers in the 2D model were infected with C. parvum (red) or kept uninfected (blue) and TEER measurements recorded at the indicated time points. Data shown are representative of three independent experiments. Each data point is the mean +/– SD of 3 TEER measurements of 3 independent experiments. Statistical analysis of differences in TEER values between uninfected and infected HIE monolayers at each time point was determined using the Mann-Whitney test. *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0005.

C. parvum infection induces disorganization of tight junctions which persists in in differentiated HIEs but recovers in undifferentiated HIEs.

Previous studies in C. parvum-infected transformed human IEC have demonstrated altered protein or RNA expression of tight and adherens junctions (21, 22). Therefore, we sought to determine whether C. parvum infection also induces alterations in these junctions in infected differentiated and undifferentiated HIEs. We used the tight junction protein ZO-1 and the adherens junction protein E-cadherin as representatives of each junction.

As seen in Fig. 5A, panels a-d, uninfected differentiated HIE monolayers showed consistent organization of ZO-1 networks at each time point, although there were fewer cells at later time points (likely due to naturally occurring cell death within the epithelium) (38). ZO-1 was located in the apical region of the HIEs with the staining for ZO-1 alternating with the phalloidin-stained filamentous (f)-actin layer as seen in the XZ view. Upon infection, however, there was disorganization of the network patterning and aggregation of ZO-1 at 24 h, which worsened at 48 and 72 h postinfection (Fig. 5A, panels f-h). In the XZ view, we also observed a decrease in the monolayer height in infected compared to uninfected HIE monolayers. To quantify the disorganization of the ZO-1 patterning, we used the tight junction organization rate (TiJOR) image analysis technique developed by Terryn et al. (39). Infection with C. parvum led to a significantly sharp decrease in TiJOR in infected compared to uninfected HIEs at 24 (P < 0.001), 48 (P < 0.001), and 72 (P < 0.05) hpi (Fig. 5B, panel a). Monolayer height of infected compared to uninfected HIE monolayers also decreased significantly at 24 (P < 0.001), 48 (P < 0.001), and 72 (P < 0.05) hpi (Fig. 5B, panel b). To determine if ZO-1 RNA levels were also impacted, we performed qRT PCR on RNA extracted from HIEs in the 3D model (since it is difficult to obtain enough RNA in the 2D model) to compare ZO-1 expression in uninfected and infected HIEs. As seen in Fig. 5B, panel c, there was a significant decrease in RNA expression of ZO-1 at 24 (P < 0.05), 48 (P < 0.005), and 72 (P < 0.005) hpi in infected compared to uninfected (at 12 h) HIEs.

FIG 5.

C. parvum infection induces disorganization of tight junctions in differentiated HIEs. A. Differentiated HIE monolayers in the 2D model were left uninfected or infected with C. parvum for 12, 24, 48, or 72 h, fixed, and analyzed by IFA. Tight junctions were stained with anti-ZO-1 (red), f-actin was stained with phalloidin (cyan), C. parvum was stained with anti-gp40 (green), and nuclei were stained with Hoechst 33528 (blue). Scale bars = 10 μM. Disorganization of tight junctions was quantified using TiJOR (B) and monolayer height (C) measured at the indicated time points. HIEs in the differentiated 3D model were left uninfected for 12 h or infected with C. parvum for 12, 24, 48, or 72 h. RNA was extracted, and ZO-1 expression determined by qRT PCR. Data shown represent mean +/– SE of 3 independent experiments, each performed in duplicate. Data were analyzed using the 2^-ΔΔ CT method and expressed relative to GAPDH (D) which was used as an endogenous reference. Statistical analysis of differences in ZO1 expression between uninfected (at 12 h) and infected HIEs at each time point was determined using the Friedman test with multiple comparisons. *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0005.

Next, we wanted to determine if ZO-1 levels in undifferentiated HIE monolayers were affected by C. parvum infection in the same way. Overall, the numbers of HIE cells in undifferentiated HIEs was greater than in differentiated HIEs, resulting in a tighter network with smaller outlines of ZO-I staining than in differentiated HIEs. In uninfected, undifferentiated monolayers, the ZO-1 network was evenly distributed throughout the monolayer and located in the apical region of the monolayer in the XZ plane (Fig. 6A, panels a-d). Upon infection, we observed increasing disorganization and aggregation of ZO-1 at 12, 24, and 48 hpi (Fig. 6A, panels e, f, g). However, by 72 h, ZO-1 network patterning in the monolayer seemed to have recovered with a more even pattern similar to the uninfected monolayers (Fig. 6A, panel h). Correspondingly, at 24 h postinfection, there was a significant decrease in TiJOR (P < 0.001) values that was rescued by 48 and 72 h (Fig. 6B, panel a). In addition, there was a significant decrease in monolayer height as early as 12 h (P < 0.005) that persisted at 24 (P < 0.005) and 48 (P < 0.001) hours but was rescued by 72 h (P > 0.05) postinfection (Fig. 6B, panel b). ZO-1 RNA expression showed a similar pattern with transcript levels significantly reduced at 24 h (P < 0.05) but rescued by 48 h (P < 0.05) postinfection (Fig. 6B, pannel c). Although expression appeared to further increase at 72 h, this difference was not statistically significant.

FIG 6.

Tight junction disorganization induced by C. parvum infection of undifferentiated 2D HIE monolayers recovers after 48 h. A. Undifferentiated 2D HIE monolayers were left uninfected or infected with C. parvum for 12, 24, 48, and 72 h and analyzed by IFA. Tight junctions were stained with anti-ZO-1 (red), f-actin was stained with phalloidin (cyan), C. parvum was stained with anti-gp40 (green) and nuclei were stained with Hoechst 33528 (blue). Scale bars = 10 μM. B. Disorganization of tight junctions was quantified using (B) TiJOR and (C) monolayer height measured at the indicated time points. 3D HIEs were left undifferentiated and uninfected for 12 h or infected with C. parvum for 12, 24, 48, or 72 h. RNA was extracted, and ZO-1 expression determined by qRT PCR. Data shown represent mean +/– SE of 3 independent experiments, each performed in duplicate. Data were analyzed using the 2^-ΔΔ CT method and expressed relative to GAPDH (D) which was used as an endogenous reference. Statistical analysis of differences in ZO1 expression between uninfected (at 12 h) and infected HIEs at each time point was determined using the Friedman test with multiple comparisons. *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0005.

These results indicate that C. parvum infection induces disorganization of ZO-1 (and consequently tight junctions) and a reduction in monolayer height and ZO-1 RNA expression which persist in differentiated HIE monolayers but recover by 48 to 72 hpi in undifferentiated HIE.

C. parvum infection induces disorganization of adherens junctions which persists in differentiated HIEs but recovers in undifferentiated HIEs.

To determine whether C. parvum infection induces disorganization of adherens junctions in addition to tight junctions, we monitored the effect of infection on E-cadherin (as a representative of the adherens junction) organization over time in differentiated and undifferentiated HIE monolayers using a similar approach as for tight junctions. In uninfected, differentiated monolayers, we observed an even distribution of E-cadherin network patterning at 12, 24, and 48 h with slightly increased aggregation at 72 h (Fig. 7A, panels a-d). Unlike ZO-1, E-cadherin was located below the Phalloidin-stained f-actin and appeared to form a longitudinal wall between HIE cells (Fig. 7A, panels a-d, XZ views). In infected HIE monolayers, as early as 12 h and increasing until 72 hpi, E-cadherin network patterning was distorted with increasing aggregation (Fig. 7A, panels e-h). This pattern of disorganization was different from that of ZO-1 and, consequently, the TiJOR image analysis method did not perform as well. To quantify this distorted network patterning, we therefore used the recently published image analysis method intercellular junction organization quantification (IJOQ) which was developed specifically for adherens junctions by Mo et al. (40). We observed a significant decrease in E-cadherin organization at 12 (P < 0.05), 24 (P < 0.005), 48 (P < 0.005), and 78 h (P < 0.005) postinfection in infected compared to uninfected HIE monolayers (Fig. 7A, panel a and B). There was a similar significant decrease in monolayer height at 24 (P < 0.005), 48 (P < 0.005), and 72 h (P < 0.05) postinfection in infected compared to uninfected HIE monolayers. When we measured E-cadherin transcript levels in the 3D model, we observed a significant and persistent reduction from 24 to 72 hpi (P < 0.05, 0.05, 0.005, respectively) (Fig. 7A, panel c and B).

FIG 7.

C. parvum infection induces disorganization of adherens junctions in differentiated HIEs. A. Differentiated HIE monolayers in the 2D model were left uninfected or infected with C. parvum for 12, 24, 48, or 72 h, fixed, and analyzed by IFA. Adherens junctions were stained with anti-E-cadherin (red), f-actin was stained with phalloidin (cyan), C. parvum was stained with anti-gp40 (green), and nuclei were stained with Hoechst 33528 (blue). Scale bars = 10 μM. B. Disorganization of adherens junctions was quantified using (B) intercellular junction organization quantification (IJOQ) and (C) monolayer height measured at the indicated time points. Differentiated 3D HIEs were left uninfected for 12 h or infected with C. parvum for 12, 24, 48, or 72 h. RNA was extracted, and ZO-1 expression determined by qRT PCR. Data were analyzed using the 2^-ΔΔ CT method and expressed relative to GAPDH (D) which was used as an endogenous reference. Data shown are representative of 3 independent experiments. Statistical analysis of differences in E-cadherin expression between uninfected (at 12 h) and infected HIEs at each time point was determined using the Friedman test with multiple comparisons. *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0005.

Compared to the differentiated monolayers, the undifferentiated monolayers showed a different pattern of E-cadherin disruption, as was observed for ZO-1. The uninfected monolayers showed the same even distribution of E-cadherin network patterning (Fig. 8A, panels a-d). Upon infection, as early as 12 h, and progressively worsening until 48 h postinfection, network patterning was disorganized (Fig. 8A, panels e-g). However, this disorganization was partly rescued by 72 hpi (Fig. 8A, panel h). When quantified by IJOQ, there was a significant decrease at 24 h (P < 0.005) which was rescued to baseline by 48 and 72 hpi (Fig. 8B, panel a). There was a significant decrease in monolayer height at 12 (P < 0.005), 24 (P < 0.005), and 48 (P < 0.005) hrs but no longer at 72 h (P > 0.05) postinfection (Fig. 8B, panel b). E-cadherin transcript levels decreased dramatically at 24 h postinfection (P < 0.005). However, at 48 h, transcript levels started to increase (P < 0.05) and continued until 72 h (P > 0.5) (Fig. 8B, panel c). These results indicate that like tight junctions, C. parvum infection leads to the disorganization of adherens junctions which progressively increases until 72 hpi in differentiated HIEs but which is rescued after 48 to 72 hpi in undifferentiated HIEs.

FIG 8.

Adherens junction disorganization induced by C. parvum infection of undifferentiated 2D HIE monolayers recovers after 48 h. A. Undifferentiated HIE monolayers in the 2D model were left uninfected or infected with C. parvum for 12, 24, 48, or 72 h, fixed, and analyzed by IFA. Adherens junctions were stained with anti-E-cadherin (red), f-actin was stained with phalloidin (cyan), C. parvum was stained with anti-gp40 (green), and nuclei were stained with Hoechst 33528 (blue). Scale bars = 10 μM. B. Disorganization of adherens junctions was quantified using (B) IJOQ and (C) monolayer height measured at the indicated time points. Undifferentiated 3D HIEs were left uninfected for 12 h or infected with C. parvum for 12, 24, 48, or 72 h. RNA was extracted, and E-cadherin expression determined by qRT PCR. Data were analyzed using the 2^-ΔΔ CT method and expressed relative to GAPDH (D) which was used as an endogenous reference. Data shown are representative of 3 independent experiments. Statistical analysis of differences in E-cadherin expression between uninfected (at 12 h) and infected HIEs at each time point was determined using the Friedman test with multiple comparisons. *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0005.

DISCUSSION

Despite recent advances in our understanding of the biology of Cryptosporidium (5, 41, 42), the pathogenesis of diarrhea and inflammation in cryptosporidiosis remains poorly understood. Human intestinal stem-cell derived enteroids/organoids may provide ex-vivo models which more accurately recapitulate human physiology to study pathogenic mechanisms underlying the disease induced by this parasite. Using these models in 2D and 3D formats, we showed that both differentiated, and undifferentiated HIEs were able to support C. parvum infection. However, there were key differences in the amount and duration of C. parvum infection and its effect on intestinal epithelial integrity in undifferentiated HIEs, which contain predominantly stem cells, and differentiated HIEs, which contain mainly enterocytes. In addition, our data suggest that C. parvum infects LGR5⁺ stem cells in undifferentiated HIE monolayers.

Our study indicates that C. parvum infection of differentiated HIEs in the 2D model was greater than that of undifferentiated HIEs but lasted only for 3 days whereas infection persisted for up to 21 days and resulted in completion of the life cycle in undifferentiated HIEs. In addition, we observed that the number of cells in infected differentiated HIEs decreased considerably by 72 h postinfection with condensed and fragmented nuclei suggestive of an accelerated form of cell death. Previous studies in transformed IEC or biliary epithelial cells have shown that C. parvum induces apoptosis in these cells by Fas-Fas ligand-mediated interactions (36). Early in infection, apoptosis of infected cells is inhibited by antiapoptotic factors BCL-2, survivin, and osteoprotegerin (induced by the parasite, presumably to allow its growth and replication), whereas later, this inhibition is removed and apoptosis progresses. This accelerated cell death may in part explain why infection does not proceed beyond 3 to 5 days in transformed IEC. Ongoing studies are directed at confirming whether apoptosis or another form of cell death occurs in C. parvum infection of differentiated HIEs in the 2D model. A more recent study using genetically engineered strains of C. parvum reported that a block in gamete fusion in transformed HCT-8 cells prevents the development of new oocysts leading to an arrest in culture (43). It remains to be determined whether this block in gamete fusion occurs in differentiated HIEs.

In contrast to differentiated HIEs, the numbers of infected undifferentiated HIEs did not decrease over time and there was no evidence of fragmented or condensed nuclei, suggesting that cell death may not frequently occur in these cells. In addition, infection in undifferentiated HIEs persisted at lower levels for up to 21 days (the longest time tested) and resulted in the formation of new oocysts, suggesting that the block in gamete fusion found in transformed HCT-8 cells (43) may not occur in these HIEs. Heo et al. (31) also found that infection in differentiated organoids was greater than in expanding organoids and that infection of expanding organoids resulted in prolonged infection and formation of new oocysts. Nikolaev et al. (32) used tissue engineering to induce small intestinal organoids to form a tubular epithelium with a lumen and crypt and villus-like domains. Luminal introduction of C. parvum in these structures also resulted in infection of the organoids with completion of the life cycle. However, it is not clear whether infection occurred in expanding or differentiated organoid cells. Neither of these studies employed infection of organoids in transwells in the 2D format. In addition, the formats used in these studies are not compatible with the studies that can be performed in transwells.

Our findings that undifferentiated HIEs in the 2D model contained exclusively LGR5⁺ cells by IFA and that C. parvum infected these HIEs, raised the possibility that the parasite directly infects these cells. We used costaining of infected undifferentiated HIE monolayers with LGR5 (for stem cells) and SI (for enterocytes) to address this possibility. The finding that intracellular stages were located within HIEs that were LGR5 positive, but SI negative strengthened this possibility. However, immunoelectron microscopy with specific antibodies is needed to confirm that C. parvum infects LGR5⁺ stem cells. Other enteric pathogens have been shown to influence LGR5⁺ stem cells during infection. Rotavirus infection of mice has been shown to lead to activation of specific LGR5⁺ intestinal stem cell populations (44, 45). Listeria monocytogenes infection of murine small intestinal organoids was recently reported to lead to a reduction in the number of LGR5⁺ stem cells (46). However, we are not aware of direct infection of LGR5⁺ intestinal stem cells by other enteric pathogens. Additional studies are required to determine whether C. parvum infection impacts LGR5⁺ stem cell replication in HIEs.

Our data also showed differences in the impact of C. parvum infection on intestinal epithelial integrity in differentiated versus undifferentiated HIE monolayers in the 2D model. Infection resulted in decreased intestinal epithelial integrity which persisted in differentiated HIEs but was rescued in undifferentiated HIEs. Impaired intestinal epithelial integrity was reflected in disorganization of both tight and adherens junctions using the proteins ZO-1 and E-cadherin, respectively, as surrogate markers. In infected HIE monolayers, disorganization was seen in directly infected cells as well as in uninfected cells. This finding suggests that the disruption of tight and adherens junctions may be induced by secreted host or parasite factors but not the parasite directly. Additional studies are needed to confirm this and determine whether host or parasite factors are involved. Disorganization persisted in differentiated HIEs but recovered in undifferentiated HIEs. The reasons for these differences are not entirely known. However, it is likely that accelerated cell death observed at later time points in differentiated HIE monolayers prevents recovery of intercellular contacts and restoration of tight and adherens junctions, whereas continuous cell replication over time in undifferentiated HIEs permits rescue of impaired epithelial integrity and restores the apical junctional complex organization. Additional studies are needed to confirm this possibility and to determine whether C. parvum infection impacts desmosomes in addition to tight and adherens junctions in differentiated and undifferentiated HIEs. Previous studies on C. parvum infection in human intestinal organoids (31, 32) did not study epithelial integrity in these systems, although Nikoleav et al. (32) devised a FITC-dextran-based method to measure epithelial integrity in their mini-intestines model.

Intestinal epithelial integrity has previously been investigated in transformed human IEC. Kumar et al. (21) showed that C. parvum infection of transformed Caco-2 cells increased paracellular permeability using FITC-dextran flux and decreased protein but not mRNA expression of tight junction proteins occludin, claudins 3 and 4, and ZO-1, as well as adherens junctions proteins E-cadherin and α-catenin. Another study by de Sablet et al. (22) found that C. parvum infection of transformed Caco-2 and HCT-8 cell monolayers grown in the transwell system resulted in a decrease in TEER at 24 and 48 hpi. However, they did not observe any changes in the distribution of the tight junction proteins claudin-1, occludin, or ZO-1, but found disruption of the adherens junctions proteins E-cadherin and β-catenin and reduced protein expression of the latter but not the former. The differences between these two studies and between their studies and ours may reflect the known discrepancies in transformed compared to primary human intestinal epithelial cells (8, 9).

Although we identified important differences in C. parvum infection of differentiated and undifferentiated HIEs, there are some limitations to our study which can be addressed in future studies. We only used the 2D model for studies on the amount and duration of C. parvum infection and for the effect of infection on intestinal epithelial integrity. In future studies we could use the 3D model as well. However, this model is not conducive to measurement of transepithelial electrical resistance (TEER). In our study we used separate 2D models of undifferentiated HIEs (LGR5+ stem cells) and differentiated (SI+ enterocytes). In the human small intestine, the crypt-villus architecture and intestinal stem cell niche permit rapidly dividing stem cells at the base of the crypt to differentiate into various cell types, predominantly enterocytes, which move up the villus from where they are shed into the intestinal lumen after 3 to 5 days (47). However, this type of villus-crypt architecture is not maintained in either the 2D or 3D models used in our study or the expanding or differentiated organoid 3D model of Heo et al. (31). The bioengineered mini-intestines system of small intestinal organoids developed by Nikolaev et al. (32) does feature some aspects of the human in vivo crypt-villus architecture. This or other similar systems may provide a more physiological model for investigating the course of C. parvum infection and its effect on intestinal epithelial integrity.

Proinflammatory cytokines such as IL1β (22), TNF-α (22), IL-18 (48), and IFN-λ3 (49) produced by IEC or by immune cells in the underlying mucosa have been shown to be expressed during C. parvum infection in animal models. These cytokines have been shown to induce apical junctional complex disruption by upregulation of myosin light chain kinase 1 and/or Rho kinase 1 (50–52). In addition, the gut microbiota are known to modulate intestinal epithelial integrity and tight junctions via the release of metabolites such as short-chain fatty acids (53). Incorporation of immune cells and gut microbiota into bioengineered intestinal organoid/enteroid systems may further refine the physiological relevance of these model systems to facilitate further investigation of the mechanisms underlying the impact of C. parvum infection on intestinal epithelial integrity. Further, these models may be used to investigate inhibitors of intestinal barrier disruption as potential agents to prevent disease development, similar to Divertin, which blocks TNF-α-induced MLCK-1 recruitment and subsequent intestinal barrier disruption and diarrhea (50).

MATERIALS AND METHODS

Parasites and Infection.

C. parvum oocysts (Iowa isolate, obtained from Bunch Grass Farms, Deary, ID) were stored in sterile phosphate-buffered saline (PBS) with penicillin and streptomycin at 4°C and used within 3 months. Prior to infection, the desired number of oocysts were surface sterilized with 10% (vol/vol) commercial bleach solution (sodium hypochlorite) as previously described (34). Oocysts were then resuspended in 1 mL of PBS containing 0.75% (wt/vol) sodium taurocholate and incubated for 10 min at 15°C. The number of oocysts were counted and the mixture of excysted oocysts and sporozoites was washed twice with either differentiation or growth medium. The excystation rate was determined as described previously (34). The parasites were used to infect the apical site of the monolayers at a multiplicity of infection (MOI) of 0.15 of oocysts to HIE cells.

Isolation of C. parvum sporozoites for infection of HIEs was performed as previously described (34). To verify the presence of purified sporozoites and ensure that no oocysts remained, an aliquot of the preparation was stained with an oocyst-specific mAb Crypt-a-Glo and a sporozoite-specific mAb Sporo-Glo according to the manufacturer’s instructions (both from Waterborne, Inc, New Orleans, LA).

Human intestinal enteroids (HIEs).

Ileal HIEs derived from deidentified human intestinal tissue biopsy specimens were obtained from the Texas Medical Center Digestive Diseases Center Enteroid Core through a Materials Transfer Agreement with Baylor College of Medicine. All work with HIEs and C. parvum were carried out under BSL-2 conditions as approved by the Tufts University Institutional Biosafety Committee.

3D HIE model.

Our procedures for maintaining and passaging 3D HIEs have been previously described (34) except that in this study, we used growth medium which contains conditioned medium harvested from l-WRN cells engineered to produce Wnt-3a, R-spondin and Noggin (54, 55) and other previously described components (34) to culture undifferentiated HIEs in 24-well plates. Growth medium was changed every 2 days and the expanding HIEs were passaged every 7 days (34). To differentiate HIEs, growth medium was switched to differentiation medium (34) for 4 days. For all experiments, HIEs were cultured at 37°C, in a 5% CO₂, humidified atmosphere for 4 to 10 days.

2D HIE model.

HIEs cultured for 7 days in growth medium were trypsinized and a single cell suspension obtained. The single cell suspension was seeded on to 0.4 μM transwell filters at a concentration of 7.5× 10⁵ HIE cells per transwell which were placed in 24-well plates (Corning Life Sciences, Corning, NY) and cultured for 2 days in growth medium containing the rho kinase inhibitor (Y2-7632), which increases the survival of the cells (14). For differentiated HIEs, the growth medium was removed after 2 days and replaced with differentiation medium for 4 days. The medium was changed every 2 days, and the HIE cells were infected apically with surface-sterilized oocysts at an MOI of 0.15 after 4 days of culture in differentiation medium (34). To culture undifferentiated monolayers, HIEs were grown for 7 days as above. They were then trypsinized and seeded onto 0.4 μM transwells as above but were continued to be cultured in growth medium. The monolayers were infected as above on the 6th day after seeding.

Transepithelial electrical resistance measurements.

Transepithelial electrical resistance (TEER) was measured in infected and uninfected 2D HIE monolayers using a Millicell-ERS probe (EMD Milipore, Burlington, MA) according to the manufacturer’s instructions. Measurements were taken daily after seeding HIEs on transwells up to the day of infection, and then on various other indicated time points. TEER values were multiplied by 0.33 to account for area of the transwells and expressed as ohms/cm².

Immunofluorescence assays and confocal microscopy.

HIE monolayers in the 2D model (differentiated or undifferentiated, uninfected or infected) on transwell filters were fixed at each time point with 4% paraformaldehyde (PFA) in PBS for 15 min at room temperature or with 100% ice-cold methanol at 20°C for 20 min. The transwell filters were washed three times with PBS and the PFA-fixed transwell filters permeabilized with 0.25% Triton X-100 and 0.1% TWEEN 20 in PBS. Nonspecific binding was blocked with 5% normal goat serum (NGS) in PBS overnight and then washed three times with PBS. Fixed HIE monolayers were stained with Hoechst 33258 (Invitrogen, Fisher Scientific, Waltham, MA) at a concentration of 2 μg/mL in PBS for 30 min in the dark at room temperature (RT). HIEs in the 3D model were cultured in growth medium and infected with C. parvum as previously described (34). Differentiated, undifferentiated, uninfected, or infected 3D HIEs were harvested and cryosectioned at various time points, and IFA and confocal microscopy were performed as described above.

Markers of IEC differentiation were detected using monoclonal antibodies (mAbs) against chromogranin A, lysozyme, mucin 2 (MUC-2), and sucrose isomaltose (all from Santa Cruz, Dallas, TX). A mAb to LGR5 (Invitrogen, Fisher Scientific, Waltham, MA) was used as a marker of stem cells.

To detect infection, a rabbit polyclonal antibody to C. parvum gp40, a surface associated glycoprotein (56), or 4E9, a mouse IgM mAb that recognizes two C. parvum surface glycoproteins gp40 and gp900 (57) were used.

Mouse mAbs to ZO-1 (Invitrogen, Fisher Scientific, Waltham, MA), or E-cadherin (BD-Biosciences, Woburn, MA) were used as markers of tight and adherens junctions, respectively. Combinations of primary antibodies to ZO-1 or E-cadherin and gp40 or 4E9 were incubated with HIE monolayers overnight for 18 h at 4°C. Monolayers were then washed three times with PBS and incubated with goat antimouse IgG or IgM or goat antirabbit IgG, as appropriate, conjugated to fluorescent dyes Alexa Fluor 488, 594, or 647 (Invitrogen, Fisher Scientific, Waltham, MA) for 2 h at room temperature, followed by 3 washes with PBS. Hoechst 33342 and Phalloidin (Invitrogen, Fisher Scientific, Waltham, MA) were used to stain nuclei and f-actin, respectively.

For double staining of infected HIEs with SI and LGR5, PFA-fixed and permeabilized monolayers were incubated with rabbit anti-gp40 and mAb to LGR5 overnight for 18 h at 4°C. After 3 washes with PBS, HIEs were incubated with the appropriate secondary antibodies, goat anti-rabbit IgG (Alexa Fluor 647) and goat anti-mouse IgG. (Alexa Fluor 594), respectively. The mAb to SI was directly conjugated to an Alexa Fluor 488 fluorescently labeled F’ab using the Zenon labeling system according to the manufacturer’s instructions (Thermo Fisher, Waltham, MA). The conjugated mAb to SI was used to stain the previously labeled cells for 3 h at room temperature. Following 3 washes with PBS, nuclei and f-actin were stained as described above.

Transwell filters were excised from transwells, placed onto microscope slides and Fluoro-Gel mounting medium (Electron Microscopy Sciences, Hatfield, PA) was added to prevent photobleaching. Coverslips were sealed onto the slides using transparent nail polish. Images were obtained using a Zeiss LSM800 or a Nikon1R laser scanning confocal microscope and analyzed using ImageJ (58, 59).

Infected and uninfected, differentiated, and undifferentiated 3D HIEs were snap-frozen in optimal cutting temperature (OCT) medium, cryo-sectioned, fixed with paraformaldehyde, and IFA was performed using antibodies for markers of differentiation as described for 2D HIEs (34).

To determine if fresh oocysts formed after infection of undifferentiated HIEs in the 2D model, supernatants were collected from the apical compartments of the transwells after 21 days. Aliquots of the supernatant were spotted onto poly-l-lysine-coated 8-well slides (Sigma, Waltham, MA), dried, fixed in ice cold methanol for 30 min at RT, then washed with PBS. Fixed supernatants were incubated with Crypt-a-Glo for 30 min at RT. After washing with PBS, Fluoro-Gel mounting medium was added and coverslips were sealed onto the slides using transparent nail polish. Imaging was performed by differential interference contrast (DIC) or fluorescence microscopy.

Quantitation of C. parvum infection.

Since obtaining sufficient RNA for RT PCR quantitation of C. parvum infection (7) in the 2D model was a limitation, infection was quantified by image analysis as follows. Z-stack images of the infected HIE monolayers in the 2D model were obtained following IFA and confocal microscopy, and the images were processed using Image J. A threshold was set to identify anti-gp40-stained intracellular C. parvum stages by size (1 to 5 μM). The Image J software then automatically counted the total number of parasite stages in the image. This process was repeated for 10 representative images at each time point and the mean recorded.

Image analysis of tight and adherens junctions using tight junction organizational rate (TiJOR) and intercellular junction organization quantification (IJOQ).

To quantify the disruption of tight junctions, the method previously described by Terryn et al. was used (39). Briefly, using image J, a polygon is drawn on the confocal image at the level of the ZO-1 network and extended to successive larger polygons. The software calculates the size of the polygon, and the number of ZO-1 peaks it crosses using the TiJOR algorithm. Disruption of adherens junctions was quantified using intercellular junction organization quantification (IJOQ), a newly described (40) automated algorithm which analyzes and quantifies the degree of organizational disruption of intercellular junctions. Briefly, the algorithm normalizes within-image and between-image brightness variations that stem from variations in staining and imaging before measuring intercellular junction disruption. Disruption is measured by tracing linear paths through the image and recording the frequency of intercellular junctions detected while tracing these paths. A system with disrupted intercellular junctions is expected to have gaps. These gaps decrease the likelihood that intercellular junctions are detected while tracing any given path, hence disruption leads to a lowered frequency of detected junctions.

Monolayer height measurements.

Monolayer heights (60) were measured as follows. Briefly, Z-stacks (0.5 μM per stack) of the HIE monolayers were obtained and orthogonal views created using Image J (58). A line was drawn from the basolateral to the apical surface of the monolayers and the length was measured. Three measurements were obtained for each image and the mean was recorded as the monolayer height.

Quantitative reverse transcriptase PCR (qRT PCR).

3D HIEs were cultured in growth medium for 2 days, then switched to differentiation medium for 4 days. They were then infected with C. parvum for various times as described (34). Total RNA was extracted from infected and uninfected HIEs using a RNeasy Plus minikit (Qiagen, Valencia, CA). RNA (100 ng) was reverse transcribed using a high-capacity reverse transcription kit (Applied Biosystems, Foster City, CA). PCR was then performed using 10 ng of cDNA and QuantiTect SYBR green PCR Mastermix (Qiagen) on a Mx3000P qPCR system (Agilent Technologies, Santa Clara, CA). Briefly, the reaction mixtures were heated to 95°C for 15 min and then subjected to 40 thermal cycles (94°C for 30 s, 58°C for 30 s, and 72°C for 30 s) of PCR amplification. After amplification, melting curve analysis was performed at temperatures of between 60°C and 95°C to assess the specificity of the reactions. Primer sequences for ZO-1 (61), and E-cadherin (62) were previously described.