Apical Spore Phagocytosis Is Not a Significant Route of Infection of Differentiated Enterocytes by Encephalitozoon intestinalis
http://www.100md.com
感染与免疫杂志 2005年第11期
Department of Physiology
Department of Microbiology, Biochemistry and Immunology, Morehouse School of Medicine, Atlanta, Georgia 30310
ABSTRACT
Encephalitozoon intestinalis is a microsporidian species that infects the intestinal mucosal epithelium, primarily in immunodeficient individuals. The present study employed undifferentiated and differentiated human colonic carcinoma cell lines to determine if this parasite species infected polarized epithelial cells by spore phagocytosis or by impalement with the deployed spore polar tube. Apical surface spore attachment differed between cell lines such that SW480>HT-29>Caco-2>HCT-8, with attachment being greater to undifferentiated Caco-2 cells than differentiated cells and greater to partially differentiated HCT-8 cells than differentiated HCT-8 cells. Attachment was inhibited by chondroitin sulfate A, suggesting that it was mediated by host cell sulfated glycoaminoglycans. Infection rates 3 days postinfection paralleled spore attachment in the various cell lines. The undifferentiated cell line SW480 and undifferentiated Caco-2 and HCT-8 cells exhibited modest spore phagocytosis while the more differentiated cell line HT29 and differentiated Caco-2 and HCT-8 cells did not. All cell lines were impaled by the polar tubes of germinating spores. When normalized to the number of spores attached to the apical membrane, such impalement was greatest in the more differentiated Caco-2 and HCT-8 cells. The host cell apical surface influenced parasite spore germination, as in populations of large undifferentiated Caco-2 cells to which >3 spores had attached, the frequency distribution of the percentages of spores germinated per cell was bimodal, indicating that the surface of some cells favored germination, while others did not. This study suggests that phagocytosis is not a biologically significant mode of infection in differentiated intestinal epithelial cells.
INTRODUCTION
Microsporidia are obligate intracellular parasites, originally classified as protozoa but now considered to be fungi (9, 27). Over 140 microsporidian genera and over 1,200 species infect host species from almost all phyla of the animal kingdom (26). With the onset of the AIDS pandemic, microsporidian infections began to be identified in seriously immunodeficient patients (6, 8, 18, 22). The majority of these human infections involved the gastrointestinal tract and resulted in a severe wasting diarrhea (6, 8, 18). The microsporidian species most frequently identified in such diarrheal cases was Enterocytozoon bieneusi, while the second most commonly encountered species was Encephalitozoon intestinalis (6). Enterocytozoon bieneusi is not readily cultured, while E. intestinalis can be maintained in several cultured cell lines (31). Electron microscopy of human biopsy samples showed that both Enterocytozoon bienuesi and E. intestinalis primarily infect the intestinal epithelium (enterocytes), although the latter species frequently causes a disseminated infection and spores have been detected in lamina propria macrophages (18, 25).
The numerous diverse microsporidian species share one hallmark stage, an environmentally resistant spore. This spore contains a unique invasion apparatus and the infectious sporoplasm (17). On exposure to an appropriate signal a coiled tube within the spore, the polar tube, rapidly everts and extends. This process is variously known as germination or hatching. The infectious sporoplasm is then forced through the tube and injected into any cell that may have been impaled, continuing the cycle of infection. The force that causes the polar tube to evert and the sporoplasm to be extruded appears to be osmotic (29). A posterior vacuole within the spore plays a role in this process (10).
The major target cell of the microsporidia infecting humans is the polarized epithelium of the gastrointestinal tract and the presumed mechanism of infection is the impaling of target cells by germinating spores. Recently it has been suggested that cultured intestinal epithelial cells, and in particular Caco-2 cells, may become infected by spore phagocytosis (11, 12, 20). Other cell lines have also been shown to phagocytize microsporidian spores (5, 12). While one might expect undifferentiated cultured cells to exhibit phagocytosis, it seems less likely that differentiated and/or polarized cultured cells or enterocytes in vivo would exhibit this type of endocytosis. The present study employed several human intestinal epithelial cell lines and Encephalitozoon intestinalis spores to determine if phagocytosis is a significant mode of infection of enterocytes by intestinal luminal spores, and if host cell differentiation reduces the incidence of any such phagocytosis.
MATERIALS AND METHODS
Cell culture. Human colonic carcinoma cells were obtained from the American Type Culture Collection (Manassas, VA). Caco-2 (ATCC HTB 37), HCT-8 (ATCC CCL 244), and HT-29 (ATCC HTB 38) were cultured at 37°C in 5% CO2. SW480 cells (ATCC CCL 228) were cultured at 35°C without CO2 or antibiotics. The media used to culture all other cell lines contained 50 μg/ml of gentamicin and 2 μg/ml of amphotericin B (Fungizone). The cell culture-specific medium used was that recommended by ATCC, except that fetal bovine serum was used in place of horse serum in the case of HCT-8 cultures (15). Twice a week, each culture was split using 0.25% trypsin-EDTA (Sigma Chemical Co., St. Louis, MO) and changed with fresh culture medium. Culture medium, fetal bovine serum, sodium pyruvate, gentamicin, and amphotericin B were purchased from the Atlanta Biologicals Company (Atlanta, GA). Minimal essential medium (MEM), nonessential amino acids, and 0.25% trypsin-EDTA solution were purchased from Sigma Chemical Company, and glutamine was purchased from Gibco-BRL (Invitrogen, Carlsbad, CA).
Human mononuclear cells were isolated from peripheral blood using Ficoll-Hypaque gradients (Amersham Bioscience, Piscataway, NJ). The cells were plated at a density of 1 x 106 cells per well in a 48-well plate on 15-mm Permanox coverslips (NUNC, Naperville, IL) in Iscove's modified Dulbecco's medium (MediaTech Inc., Herndon, VA) with 10% pooled human male serum (Lampire Biologicals, Inc., Pipersville, PA). After 24 h of incubation at 37°C with 5% CO2, the nonadherent cells were removed by washing with Hanks balanced salt solution. The remaining cells were 98% macrophages as determined by nonspecific esterase staining.
For phagocytosis experiments, cells were plated separately onto eight-well chambered slides for 3 (HCT-8 and Caco-2 cells), 5 (Caco-2, HCT-8 cells, SW480 cells, and HT-29 cells), or 7 (Caco-2 and HCT-8) days. Cultures that were 7 days old contained more differentiated cells than did 3-day-old cultures. Culture media were changed twice a week.
Parasite culture and purification. African green monkey kidney (E6) cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum, 50 μg/ml gentamicin, 2 μg/ml amphotericin B, and 44 mM NaHCO3. Cells were infected with isolated Encephalitozoon intestinalis (ATCC CDC V307) spores and cultured at 37°C in CO2 (31).
Heavily infected cells were collected twice a week by gentle mechanical agitation and stored at 4°C for no more than 14 days prior to spore purification, at which time the infected cells were concentrated by centrifugation for 30 min at 1,200 x g. This and all subsequent steps were performed at 4°C. The supernatant solution was removed and the remaining sample was expressed through a 27-gauge needle three times to rupture host cells. In order to eliminate excessive debris, the sample was filtered through a sterile non-pyrogenic 5-μm-pore-size Cameo filter (GE Osmonics, Minnetonka, MN). The sample was then placed in a discontinuous Percoll gradient (5.5 ml of Percoll and 4.5 ml of phosphate-buffered saline) and centrifuged for 30 min at 1,200 x g. After centrifugation, the pellet containing the intact microsporidian spores was washed three times in phosphate-buffered saline by suspension and centrifugation. Spores were counted with a hemacytometer.
Infection assay. Human colonic carcinoma cells (3, 5, and 7 days old) were cultured to confluence in eight-well chamber slides. Cells were infected with E. intestinalis (1 x 106 spores) for three days. The culture-specific media were changed daily. Infection was assessed by counting the number of cells with clearly defined parasitophorous vacuoles per five fields as illustrated in Fig. 1A.
Spore internalization (phagocytosis) assay. E. intestinalis spores (100 μl serum-free Opti-MEM 1 medium [Gibco-BRL] containing 1 x 106 spores) were placed in each well of eight-well chamber slides containing confluent cultures of SW480, HT-29, Caco-2, or HCT-8 cells. All cultures were maintained for 4 h at 37°C in 5% CO2. Two additional experiments of this type were performed, one to determine the effects of greater spore attachment to target cells and the other to determine if performing the assay with cell line-specific media affected the results. For better spore attachment, this experiment was repeated with the addition of 2 mM manganese to all solutions and media (14). In order to determine the effect of removing the cells from their cell line-specific media, these experiments were also repeated in the media in which the cells were cultured. Qualitatively similar results were obtained with all of these experiments and only those experiments employing Opti-MEM alone are reported here.
In order to distinguish internalized from attached extracellular spores, plated cultured cells were fixed with 3.7% paraformaldehyde (pH 7.4) for 15 min. Cells were then washed three times with Tris-buffered saline (TBS), 5 min per rinse. Samples were blocked with 2% bovine serum albumin (BSA; Sigma Chemical Co.) in TBS for 10 min at room temperature. External spores were labeled with a 1:500 dilution of rabbit anti-E. intestinalis polyclonal antibody in TBS containing 2% BSA. Samples were incubated for 1 h at 37°C, washed three times in TBS and labeled with Oregon Green 488 goat anti-rabbit immunoglobulin G (IgG) (Molecular Probes, Eugene, OR) diluted 1:700 in TBS containing 2% BSA, and incubated for 1 h at 37°C. After washing of the samples three times, the cells were permeabilized with 0.02% Triton X-100 in TBS for 10 min, washed three times with TBS, and blocked with 2% BSA in TBS at room temperature. Intracellular structures were labeled with the rabbit polyclonal anti-E. intestinalis antibody (diluted 1:500 in TBS containing 2% BSA) and a mouse anti-zona occludens 1 (ZO-1) monoclonal antibody (BD BioSciences Pharmingen, San Jose, CA) diluted 1:25 in TBS containing 2% BSA. Cells were incubated for one hour at 37°C, washed three times with TBS, and labeled with goat anti-rabbit IgG AlexaFluor 594 (Molecular Probes) (diluted 1:500 in TBS containing 2% BSA) and Biotin-SP conjugate AffiniPure goat anti-mouse antibody (Jackson Immuno Research, West Grove, PA) diluted 1:300 in TBS containing 2% BSA. Cells were washed three times with TBS and then labeled with Streptavidin Alexa Fluor 488 conjugate dye (Molecular Probes) diluted 1:300 in TBS containing 2% BSA. The sample was washed three times with TBS and mounted on glass slides using mounting media. Extracellular and intracellular parasites were visualized with an epifluorescence microscope and counted. Intracellular parasite spores appeared red, target cell tight junction ZO-1 appeared green, and adherent extracellular spores were double labeled and appeared yellow (Fig. 1). For each slide, 300 attached spores and 20+ fields were assessed by using a 63x oil immersion objective.
Assessment of differentiation. SW480 cells are generally considered undifferentiated, and HT-29 cells are poorly differentiated when cultured in the type of medium used here (2). Caco-2 and HCT-8 cells exhibit a continuum of degrees of differentiation depending on the age of the culture and regional cell density. Criteria were established to divide the Caco-2 and HCT-8 cells into undifferentiated, intermediate, and differentiated cells. The criteria included the x-y surface area of each cell, which decreased as the cell become more differentiated and columnar (Table 1) and was readily measured by visualization of the ZO-1 probe outline which localized the tight junction and by the location of the cell on a dome or plateau (villus-like structure) of differentiated cells. Differentiated Caco-2 cells were columnar and small in x-y surface area, often located on domelike structures. Intermediate Caco-2 cells surrounded these differentiated cells. HCT-8 cells were considered differentiated when they were located on the tips of villi and crests, and intermediate cells had a larger x-y surface area and surrounded the villi and crests. Both Caco-2 and HCT-8 undifferentiated cells were large and flat. Undifferentiated HCT-8 cells predominated in 3-day-old cultures and were found in small clusters in 5-day-old cultures. In practice, a transparent template of the cells used to generate Table 1 was placed over the computer screen to act as a guide in determining the degree of differentiation of a field of Caco-2 or HCT-8 cells. Cells were also fixed with cacodylate-buffered 2.5% glutaraldehyde and prepared for scanning electron microscopy (SEM) (30). The apical brush border of undifferentiated and differentiated Caco-2 and HCT-8 cells was then imaged and the brush border development was used to confirm the assessment of differentiation (Fig. 2).
Measurement of the incidence of infection, phagocytosis, impalement, and germination. Labeled cells were visualized by fluorescent microscopy. Data were collected using SPOT 3.4.5 software (Diagnostic Instruments, Sterling Heights, MI). A field of 4,200 μm2 was used as the standard field size referred to here when measuring the incidence of infected, attached, phagocytized, or germinated spores. Germinated spores were further divided into those that had impaled a host cell and those that remained on the cell surface. Fields were randomly chosen, except in those cases where fields of undifferentiated, intermediate, and differentiated Caco-2 and HCT-8 cells were used. In those cases, composites of half and quarter fields were often pooled. The degree of host cell differentiation was first determined by visualizing the ZO-1 labeled outline of a cell population and then the parasite spores were visualized and counted. Germination was not seen with internalized spores during the 4-h time period of these experiments.
Infection and attached and phagocytized spores were distinguished by color. Double-labeled extracellular spores appeared yellow while intracellular parasite stages and internalized spores appeared red (Fig. 1A and B). Intracellular parasitophorous vacuoles were clearly evident from two days following spore exposure (1) and these were used to assess the incidence of infected cells. Individual red (internalized) spores that were seen in 4-h experiments were considered to have been phagocytized (5). Impalement was detected when a polar tube was seen extending from an attached extracellular spore (yellow) and, at the point at which the polar tube entered the host cell, the filament and any extruded sporoplasm appeared red (Fig. 1C). In germinated spores that did not impale a cell, the spore, polar tube and sporoplasm all appeared yellow.
To determine if spore germination was affected by the apical microclimate of individual host cells, the incidence of germination was determined in undifferentiated individual Caco-2 cells to which three or more spores were attached.
Statistical evaluation. Analysis of variance tests were used to determine the significance of differences between groups and Tukey's protected t tests were used as post hoc tests to determine the significance of differences between group mean values. Group mean differences were considered statistically significant when P was <0.05. In one experiment, the germination rates of spores were assessed when multiple spores were attached to individual cells. Frequency distributions of the percentage of spores germinating on individual host cells were plotted for cases where three to eight spores were attached/cell. In cases where the number of spores attached per cell was three, four, five, six, or eight, the distributions was bimodal and these data are illustrated showing the percentage of cells in which either all or none of the attached spores had germinated.
RESULTS
Data are presented for experiments in which the enterocyte exposure to E. intestinalis spores took place in serum-free Opti-MEM medium. Qualitatively similar results were obtained when the 4-h spore exposure took place in the culture medium appropriate to the cell line or the Opti-MEM to which MnCl2 had been added (data not shown). Figures 3, 4, and 5 illustrate the spore attachment (per five fields of 4,200 μm2 each), spore germination occurring per 100 visualized spores, phagocytized spores per 100 visualized spores, and germinated spores that impaled target cells per 100 visualized spores. The SW480 and HT-29 cells were treated as a homogenous monolayer. The Caco-2 cells and HCT-8 cells were divided into areas of undifferentiated, intermediate and differentiated cells on the basis of their x-y surface area and location (Table 1). Spore attachment was significantly different between cell lines: SW480>HT-29>Caco-2>HCT-8 (Fig. 3, 4, and 5). In Caco-2 cells, spore attachment was significantly greater for undifferentiated than for either intermediate or differentiated cells (Fig. 4) or HCT-8 undifferentiated or differentiated cells (Fig. 5), while in HCT-8 cultures, attachment was significantly greater with intermediate than with either undifferentiated or differentiated cells. Thus, in both Caco-2 and HCT-8 cells, as the cells assumed an enterocyte-like appearance, attachment decreased as the cells differentiated.
Host cell surface sulfated glycoaminoglycans have been implicated in E. intestinalis spore attachment (14). When the attachment assay was repeated using Optim-MEM with and without 1 mg/ml chondroitin sulfate A (Sigma Chemical Co.), the addition of this exogenous sulfated glycoaminoglycan significantly reduced attachment with all four cell lines (P < 0.01). The observed reductions were 80.0%, 70.6%, 84.1%, and 81.7% for SW480 cells, HT-29 cells, undifferentiated Caco-2 cells and intermediate HCT-8 cells, respectively.
The differences between and within cultures with respect to 4 -h spore attachment paralleled the degree of infection observed 3 days postinfection (Table 2).
Spore germination was observed in all cell lines. When normalized to the number of attached spores, such germination was significantly greater with differentiated Caco-2 cells than with any other group. Germination in the SW480 cultures was less than in any other, being significantly different from intermediate and differentiated Caco-2 and HCT-8 preparations. In both the Caco-2 and HCT-8 cultures, germination was significantly greater with intermediate and differentiated cells than with undifferentiated cells.
As all cultures were maintained in the same medium in these experiments, cell line differences in attached spore germination were considered likely due to some aspect of the host cell apical membrane. In the case of large undifferentiated Caco-2 cells multiple spores were frequently attached to a single cell. The percentage of attached spores germinating on the surface of individual cells was measured. Figure 6 illustrates the percentage of cells in which all or none of the attached spores had germinated for cells to which three, four, five, six, seven, or eight spores had attached. In all but the case of seven spores/cell the frequency distribution of germination was bimodal. For example, in the case of cells to which five spores had attached, in 63% of the cells all or none of the spores had germinated, while in the remaining 37% of the cells, one, two, three, or four of the attached spores had germinated. This bimodal frequency distribution suggests that the apical surface of individual cells differed such that some cell surfaces favored germination while others did not. The examples using large undifferentiated Caco-2 cells shown in Fig. 1B and C exemplify this phenomenon of most of the spores attached to a single cell being either not germinated or germinated, as none of the seven spores seen attached to a single cell in Fig. 1B germinated, while three and possibly four of the spores attached to a single cell in Fig. 1C had germinated.
It was possible to detect when germination resulted in impalement of the target cell. When normalized to the number of attached spores, impalement increased significantly with differentiation in Caco-2 cells (Fig. 4) and was significantly greater in differentiated HCT-8 cells than in intermediate HCT-8, SW480, HT-29, or Caco-2 cells (Fig. 3, 4, and 5). It has been suggested that E. intestinalis infection in Caco-2 cells may result from phagocytosis of sporoplasm at the cell surface, rather than a forceful penetration of the polar tube and subsequent injection of the sporoplasm (11). Scanning electron microscopy failed to show any example of sporoplasm being phagocytized at the cell surface. Typically the polar tube seemed to penetrate the host cell with no suggestion of cell surface envelopment of adherent sporoplasm (Fig. 7).
Because phagocytosis was rarely observed and the variance of the phagocytosis data was so great, no significant differences were detected between mean phagocytosis values measured in the various cultures. In the Opti-MEM experiments no phagocytosis was observed in HT-29 cultures or in differentiated Caco-2 or HCT-8 cells (Fig. 3, 4, and 5). Thus, as host cells became more differentiated, relative to the number of attached spores, the incidence of impaled cells increased and the incidence of phagocytized spores decreased. In experiments in which the 4-h spore exposure was conducted in the culture-specific media, phagocytosis as a percentage of visualized spores was as follows: SW480, 0.38% ± 0.20%; HT-29, 0.10% ± 0.10%; Caco-2 undifferentiated, 0.93% ± 0.24%; Caco-2 differentiated, 0.11% ± 0.11%; HCT-8 undifferentiated, 2.11% ± 0.90%; HCT-8 differentiated, 0.44% ± 0.24%. By comparison, when human macrophages were exposed to E. intestinalis spores at the same density for 2 h rather than 4 h the phagocytosis was 72.0% ± 12.1%.
DISCUSSION
There appear to be at least three major routes by which microsporidia infect host cells: impalement by extracellular spores, which is probably the major route of an initial infection; impalement of a cell by a spore germinating in a adjacent cell, which is probably the reason that infection in cell cultures often occurs as foci; and spore phagocytosis, particularly by macrophages, which is probably the major cause of parasite dissemination among those microsporidian species that cause disseminated human disease. Macrophage phagocytosis of microsporidian spores is usually defensive and results in spore destruction (7). However, there is evidence that if the spores can germinate within the macrophage before lysosomal fusion has begun to destroy the phagosome content, the infectious sporoplasm will escape destruction (12). In addition, some microsporidian species have been demonstrated to be able to block macrophage phagosome acidification and lysosome fusion (32, 33). Studies using cultured cell lines have indicated that a variety of cells are capable of phagocytizing microsporidian spores (5, 11, 12, 20), suggesting that an endocytic process may be important in intestinal microsporidiosis. Invasive microorganisms have devised a rich repertoire of mechanisms to subvert the host cell's endocytic processes, allowing invasion to occur even in cells that are not generally phagocytic (3, 4). The present study was designed to clarify whether phagocytosis was indeed a biologically significant means of infecting polarized intestinal enterocytes, a major target of human microsporidiosis.
The methods we employed allowed us to readily distinguish microsporidian spores that were attached to the host cell surface from those that had been phagocytized (Fig. 1B) and to distinguish attached spores that had germinated but not impaled a host cell from those that had impaled a cell, although the latter may have been undercounted due to difficulties in distinguishing intracellular polar tubes and sporoplasms (Fig. 1C). Spore attachment to host cells differed significantly between cell lines and between cells of the same cell line that were at different stages of differentiation (Fig. 3, 4, and 5). Because of differences in cell surface area and spore attachment between cell lines, we normalized the number of spores used in the assays to well surface area and the count of spore attachment to fields of a known area. The number of spores used in these experiments averaged about four spores/cell, considerably less than that generally used in the literature (e.g., 14). However, the number of spores we used is still probably higher than that found in the intestinal lumen within range of the mucosal epithelium, especially at the onset of an infection.
Hayman and colleagues (14) provided evidence that E. intestinalis spores adhere to surface sulfated glycoaminoglycans of Vero and CHO cells. Such attachment was reduced when a number of exogenous sulfated glycoaminoglycans were included in their adherence assay. Similarly, we found that chondroitin sulfate A inhibited attachment in all four cell lines used in our study, suggesting that the observed cell line and differentiation-dependent differences in attachment were the result of differences in the cell surface sulfated glycoaminoglycan expression. We did not observe the preferential spore attachment to intercellular junctions as has been reported elsewhere (11). Infection sites per field in the various cell lines (Table 2) paralleled attachment, in that the greatest infection was seen in SW480 cells, and undifferentiated Caco-2 and intermediate HCT-8 cell infections were greater than those seen in differentiated Caco-2 and HCT-8 cells. The infection seen in fields of differentiated Caco-2 cells was higher than one would have predicted based on the low spore attachment to such cells. However, the 3-day duration of the infection assay allowed time for infected undifferentiated and intermediate cells to differentiate. While it is known that infecting cells with Encephalitozoon spp. inhibits mitosis (23), the effects of infection on differentiation are not known.
Because there were such large differences in spore attachment between and within cell lines, spore germination was normalized to the number of attached spores. Germination was significantly less when spores were attached to the undifferentiated SW480 cell line and to undifferentiated Caco-2 and HCT-8 cells. This suggests that the degree of host cell differentiation influenced germination and that this effect was separate from any effect of the bulk phase (medium) environment and was probably independent of the cell surface sulfated glycoaminoglycan spore attachment effect. When spore germination was assessed in large, flat undifferentiated Caco-2 cells to which three or more spores had attached, a bimodal distribution was noted that suggested that while the surface of some cells was conducive to the germination of all attached spores, this was not the case for other cells (Fig. 6). These differences could not be attributed to the medium, as all cells in an assay were from the same slides and were exposed to the same bulk phase environment. Most studies of microsporidian spore germination have focused on the environmental effects of medium pH or ionic composition (13, 16, 19, 28) and not on the microenvironment of the unstirred layer that occurs at the brush border.
Impalement of the host cell by the polar tube at germination appears to be the major way in which infection came about in these cell lines. This was particularly true with the differentiated Caco-2 and HCT-8 cells and the moderately differentiated HT-29 cells. This might be explained by the fact that spores attached to undifferentiated cells would be likely to attach so that their long axis paralleled the apical surface while spores attached to a well developed brush border might be more likely to be oriented so that on germinating the polar tube would be pointing towards the target cell (20). The present study was not able to distinguish between cells that had been impaled and cells that had phagocytized the sporoplasm released from a spore germinating at the cell surface as has been suggested (11), although SEM data did not support this suggestion. In SEM studies, we were unable to find a suggestion of phagocytosis of surface sporoplasm but observed numerous examples of clear penetration of polar filaments through the host cell plasma membrane (Fig. 7). There are examples in the literature of germinating microsporidian spores caught in the act of impaling a cell and injecting sporoplasm into that impaled cell (for an example, see Fig. 1B) (5). Schottelius and colleagues (24) have suggested that with Encephalitozoon spp., the host cell plasma membrane forms an invagination at the site of contact with the polar filament, perhaps by some endocytic process. The polar tube then continues into the host cell cytoplasm within a membranous channel into which the sporoplasm is subsequently injected. There is also equivocal transmission electron microscopy data to support such a channel (22). Whether the polar tube impalement of host cells observed here involves any type of endocytosis of the polar tube is not known.
As expected, E. intestinalis whole spore phagocytosis was greater in the least differentiated cell line (SW480) and in the less differentiated Caco-2 and HCT-8 cells. In experiments in which the assay was conducted in serum-free medium, phagocytosis was not observed in HT29 cells or differentiated Caco-2 and HCT-8 cells (Fig. 3, 4, and 5). In experiments in which serum was included in the assay medium there was modest phagocytosis, but again it was significantly less in differentiated than undifferentiated cells and the incidence was very small when compared to that seen in human macrophages exposed to spores for half the time (0.10 to 0.44% versus 72%). More significantly, as Caco-2 and HCT-8 cells became more differentiated the incidence of phagocytosis decreased while the incidence of cells being impaled by germinating spores increased.
This study addresses only the mode of microsporidian infection at the apical pole of differentiated intestinal epithelial cells, the presumed site of the initial infection by ingested spores. The phagocytic activity of the basolateral surface of such cells may be different from that of the apical membrane, as has been observed with several enteric bacteria (e.g., Shigella) (21). While spore phagocytosis is undoubtedly the most important mode of entry of many microsporidia into macrophages, the present study supports the concept that the initial invasion of the intestinal epithelium by ingested parasite spores involves spore germination and host cell impalement rather than spore phagocytosis.
ACKNOWLEDGMENTS
This work was supported by U.S. PHS grant R21 DK64573-A1.
We thank Amad Al-Mahmoud, Morehouse School of Medicine Clinical Research Center, for assistance with statistical analyses.
REFERENCES
1. Barry, W. K. 2003. Molecular characterization of Encephalitozoon Microspora) replication kinetics in a murine intestinal cell line. J. Eukaryot. Microbiol. 50:169-174.
2. Chantret, I., A. Barbat, E. Dussaulx, M. G. Brattain, and A. Zweilbaum. 1988. Epithelial polarity, villin expression, and enterocytic differentiation of cultured human colon carcinoma cells: a survey of twenty cell lines. Cancer Res. 48:1936-1942.
3. Chen, X. M., P. L. Splinter, P. S. Tietz, B. Q. Huang, D. D. Billadeau, and N. F. LaRusso. 2004. Phosphatidylinositol 3-kinase and frabin mediate Cryptosporidium parvum cellular invasion via activation of Cdc42. J. Biol. Chem. 279:31671-31678.
4. Cossart, P., and P. J. Sansonetti. 2004. Bacterial invasion: the paradigms of enteroinvasive pathogens. Science 304:242-248.
5. Couzinet, S., E. Cejas, J. Schittny, P. Deplazes, R. Weber, and S. Zimmerli. 2000. Phagocytic uptake of Encephalitozoon cuniculi by nonprofessional phagocytes. Infect. Immun. 68:6939-6945.
6. Coyle, C. M., M. Wittner, D. P. Kotler, C. Noyer, J. M. Orenstein, H. B. Tanowitz, and L. M. Weiss. 1996. Prevalence of microsporidiosis due to Enterocytozoon bieneusi and Encephalitozoon (Septata) intestinalis among patients with AIDS-related diarrhea: determination by polymerase chain reaction to the microsporidian small-subunit rRNA gene. Clin. Infect. Dis. 23:1002-1006.
7. Didier, E. S. 1995. Reactive nitrogen intermediates implicated in the inhibition of Encephalitozoon cuniculi (phylum microspora) replication in murine peritoneal macrophages. Parasite Immunol. 17:405-412.
8. Didier, E. S. 2005. Microsporidiosis: an emerging and opportunistic infection in humans and animals. Acta Trop. 94:61-76.
9. Edlind, T. D., J. Li, G. S. Visvesvara, M. H. Vodkin, G. L. McLaughlin, and S. K. Katiyar. 1996. Phylogenetic analysis of beta-tubulin sequences from amitochondrial protozoa. Mol. Phylogenet. Evol. 5:359-367.
10. Findley, A. M., E. H. Weidner, K. R. Carman, Z. Xu, and J. S. Godbar. 2005. Role of the posterior vacuole in Spraguea lophii (Microsporidia) spore hatching. Folia Parasitol. (Prague) 52:111-117.
11. Foucault, C., and M. Drancourt. 2000. Actin mediates Encephalitozoon intestinalis entry into human enterocyte-like cell line, Caco-2. Microb. Pathog. 28:51-58.
12. Franzen, C., A. Muller, P. Hartmann, and B. Salzberger. 2005. Cell invasion and intracellular fate of Encephalitozoon cuniculi (Microsporidia). Parasitology 130:285-292.
13. Frixione, E., L. Ruiz, J. Cerbon, and A. H. Undeen. 1997. Germination of Nosema algerae (Microspora) spores: conditional inhibition by D2O, ethanol and Hg2+ suggests dependence of water influx upon membrane hydration and specific transmembrane pathways. J. Eukaryot. Microbiol. 44:109-116.
14. Hayman, J. R., T. R. Southern, and T. E. Nash. 2005. Role of sulfated glycans in adherence of the microsporidian Encephalitozoon intestinalis to host cells in vitro. Infect. Immun. 73:841-848.
15. Huschtscha, L., E. Rozengurt, and W. F. Bodmer. 1991. Growth factor requirements of human colorectal tumor cell; relations to cellular differentiation. Eur. J. Cancer 27:1680-1684.
16. Ishihara, R. 1967. Stimuli causing extrusion of polar filaments of Glugea fumiferanae spores. Can. J. Microbiol. 13:1321-1332.
17. Keohane, E. M., and L. M. Weiss. 1998. Characterization and function of the microsporidian polar tube: a review. Folia Parasitol. (Prague) 45:117-127.
18. Kotler, D. P., and J. M. Orenstein. 1998. Clinical syndromes associated with microsporidiosis. Adv. Parasitol. 40:321-349.
19. Leitch, G. J., Q. He, S. Wallace, and G. S. Visvesvara. 1993. Inhibition of the polar filament extrusion of the microsporidium, Encephalitozoon hellem, isolated from an AIDS patient. J. Eukaryot. Microbiol. 40:711-717.
20. Magaud, A., A. Achbarou, and L. Desportes-Livage. 1997. Cell invasion by the microsporidium Encephalitozoon intestinalis. J. Eukaryot. Microbiol. 44:81S.
21. Mounier, J., T. Vasselon, R. Hellio, M. Lesourd, and P. J. Sansonetti. 1992. Shigella flexneri enters human colonic Caco-2 epithelial cells through the basolateral pole. Infect. Immun. 60:237-248.
22. Scaglia, M., L. Sacchi, S. Gatti, A. M. Bernuzzi, P. P. Polver, I. Placentini, G. P. Croppo, A. J. Da Silva, N. J. Pieniazek, S. B. Slemenda, S. Wallace, G. J. Leitch, and G. S. Visvesvara. 1004. Isolation and identification of Encephalitozoon hellem from an Italian AIDS patient with disseminated microsporidiosis. APMIS 102:817-827.
23. Scanlon, M., A. P. Shaw, C. J. Zhou, G. S. Visvesvara, and G. J. Leitch. 2000. Infection by microsporidia disrupts the host cell cycle. J. Eukaryot. Microbiol. 47:525-531.
24. Schottelius, J., C. Schmetz, N. P. Kock, T. Schuler, I. Sobottka, and B. Fleischer. 2000. Presentation by scanning electron microscopy of the life cycle of microsporidia of the genus Encephalitozoon. Microbes Infect. 2:1401-1406.
25. Schwartz, D. A., I. Sobottka, G. J. Leitch, A. Cali, and G. S. Visvesvara. 1996. Pathology of microsporidiosis. Emerging parasitic infections in patients with acquired immunodeficiency syndrome. Arch. Pathol. Lab. Med. 120:173-188.
26. Sprague, V., and J. J. Becnel. 1999. Appendix: checklist of available generic names for microsporidia with type species and type hosts, p. 517-530. In M. Whittner and L. M. Weiss (ed.), The microsporidia and microsporidiosis. ASM Press, Washington, D.C.
27. Thomarat, F., C. P. Vivares, and M. Gouy. 2004. Phylogenetic analysis of the complete genome sequence of Encephalitozoon cuniculi supports the fungal origin of microsporidia and reveals a high frequency of fast-evolving genes. J. Mol. Evol. 59:780-781.
28. Undeen, A. H., and S. W. Avery. 1988. Effects of anions on the germination of Nosema algerae (Microspora; Nosematidae) spores. J. Invert. Pathol. 52:84-89.
29. Undeen, A. H., and E. Frixione. 1990. The role of osmotic pressure in the germination of Nosema algerae spores. J. Protozool. 37:561-567.
30. Visvesvara, G. S., G. J. Leitch, H. Moura, S. Wallace, R. Weber, and R. T. Bryan. 1991. Culture, electron microscopy, and immunoblot studies on a microsporidian parasite isolated from the urine of a patient with AIDS. J. Protozool. 38:105S-111S.
31. Visvesvara, G. S., H. Moura, G. J. Leitch, and D. A. Schwartz. 1999. Culture and propagation of microsporidia, p. 363-392. In M. Whittner and L. M. Weiss (ed.), The microsporidia and microsporidiosis. ASM Press, Washington, D.C.
32. Weidner, E. 1975. Interaction between Encephalitozoon cuniculi and macrophages. Parasitophorous vacuole growth and the absence of lysosomal fusion. Z. Parasitenkd. 47:1-9.
33. Weidner, E., and L. D. Sibley. 1985. Phagocytized intracellular microsporidian blocks phagosome acidification and phagosome-lysosome fusion. J. Protozool. 32:311-317.(Gordon J. Leitch, Tarsha )
Department of Microbiology, Biochemistry and Immunology, Morehouse School of Medicine, Atlanta, Georgia 30310
ABSTRACT
Encephalitozoon intestinalis is a microsporidian species that infects the intestinal mucosal epithelium, primarily in immunodeficient individuals. The present study employed undifferentiated and differentiated human colonic carcinoma cell lines to determine if this parasite species infected polarized epithelial cells by spore phagocytosis or by impalement with the deployed spore polar tube. Apical surface spore attachment differed between cell lines such that SW480>HT-29>Caco-2>HCT-8, with attachment being greater to undifferentiated Caco-2 cells than differentiated cells and greater to partially differentiated HCT-8 cells than differentiated HCT-8 cells. Attachment was inhibited by chondroitin sulfate A, suggesting that it was mediated by host cell sulfated glycoaminoglycans. Infection rates 3 days postinfection paralleled spore attachment in the various cell lines. The undifferentiated cell line SW480 and undifferentiated Caco-2 and HCT-8 cells exhibited modest spore phagocytosis while the more differentiated cell line HT29 and differentiated Caco-2 and HCT-8 cells did not. All cell lines were impaled by the polar tubes of germinating spores. When normalized to the number of spores attached to the apical membrane, such impalement was greatest in the more differentiated Caco-2 and HCT-8 cells. The host cell apical surface influenced parasite spore germination, as in populations of large undifferentiated Caco-2 cells to which >3 spores had attached, the frequency distribution of the percentages of spores germinated per cell was bimodal, indicating that the surface of some cells favored germination, while others did not. This study suggests that phagocytosis is not a biologically significant mode of infection in differentiated intestinal epithelial cells.
INTRODUCTION
Microsporidia are obligate intracellular parasites, originally classified as protozoa but now considered to be fungi (9, 27). Over 140 microsporidian genera and over 1,200 species infect host species from almost all phyla of the animal kingdom (26). With the onset of the AIDS pandemic, microsporidian infections began to be identified in seriously immunodeficient patients (6, 8, 18, 22). The majority of these human infections involved the gastrointestinal tract and resulted in a severe wasting diarrhea (6, 8, 18). The microsporidian species most frequently identified in such diarrheal cases was Enterocytozoon bieneusi, while the second most commonly encountered species was Encephalitozoon intestinalis (6). Enterocytozoon bieneusi is not readily cultured, while E. intestinalis can be maintained in several cultured cell lines (31). Electron microscopy of human biopsy samples showed that both Enterocytozoon bienuesi and E. intestinalis primarily infect the intestinal epithelium (enterocytes), although the latter species frequently causes a disseminated infection and spores have been detected in lamina propria macrophages (18, 25).
The numerous diverse microsporidian species share one hallmark stage, an environmentally resistant spore. This spore contains a unique invasion apparatus and the infectious sporoplasm (17). On exposure to an appropriate signal a coiled tube within the spore, the polar tube, rapidly everts and extends. This process is variously known as germination or hatching. The infectious sporoplasm is then forced through the tube and injected into any cell that may have been impaled, continuing the cycle of infection. The force that causes the polar tube to evert and the sporoplasm to be extruded appears to be osmotic (29). A posterior vacuole within the spore plays a role in this process (10).
The major target cell of the microsporidia infecting humans is the polarized epithelium of the gastrointestinal tract and the presumed mechanism of infection is the impaling of target cells by germinating spores. Recently it has been suggested that cultured intestinal epithelial cells, and in particular Caco-2 cells, may become infected by spore phagocytosis (11, 12, 20). Other cell lines have also been shown to phagocytize microsporidian spores (5, 12). While one might expect undifferentiated cultured cells to exhibit phagocytosis, it seems less likely that differentiated and/or polarized cultured cells or enterocytes in vivo would exhibit this type of endocytosis. The present study employed several human intestinal epithelial cell lines and Encephalitozoon intestinalis spores to determine if phagocytosis is a significant mode of infection of enterocytes by intestinal luminal spores, and if host cell differentiation reduces the incidence of any such phagocytosis.
MATERIALS AND METHODS
Cell culture. Human colonic carcinoma cells were obtained from the American Type Culture Collection (Manassas, VA). Caco-2 (ATCC HTB 37), HCT-8 (ATCC CCL 244), and HT-29 (ATCC HTB 38) were cultured at 37°C in 5% CO2. SW480 cells (ATCC CCL 228) were cultured at 35°C without CO2 or antibiotics. The media used to culture all other cell lines contained 50 μg/ml of gentamicin and 2 μg/ml of amphotericin B (Fungizone). The cell culture-specific medium used was that recommended by ATCC, except that fetal bovine serum was used in place of horse serum in the case of HCT-8 cultures (15). Twice a week, each culture was split using 0.25% trypsin-EDTA (Sigma Chemical Co., St. Louis, MO) and changed with fresh culture medium. Culture medium, fetal bovine serum, sodium pyruvate, gentamicin, and amphotericin B were purchased from the Atlanta Biologicals Company (Atlanta, GA). Minimal essential medium (MEM), nonessential amino acids, and 0.25% trypsin-EDTA solution were purchased from Sigma Chemical Company, and glutamine was purchased from Gibco-BRL (Invitrogen, Carlsbad, CA).
Human mononuclear cells were isolated from peripheral blood using Ficoll-Hypaque gradients (Amersham Bioscience, Piscataway, NJ). The cells were plated at a density of 1 x 106 cells per well in a 48-well plate on 15-mm Permanox coverslips (NUNC, Naperville, IL) in Iscove's modified Dulbecco's medium (MediaTech Inc., Herndon, VA) with 10% pooled human male serum (Lampire Biologicals, Inc., Pipersville, PA). After 24 h of incubation at 37°C with 5% CO2, the nonadherent cells were removed by washing with Hanks balanced salt solution. The remaining cells were 98% macrophages as determined by nonspecific esterase staining.
For phagocytosis experiments, cells were plated separately onto eight-well chambered slides for 3 (HCT-8 and Caco-2 cells), 5 (Caco-2, HCT-8 cells, SW480 cells, and HT-29 cells), or 7 (Caco-2 and HCT-8) days. Cultures that were 7 days old contained more differentiated cells than did 3-day-old cultures. Culture media were changed twice a week.
Parasite culture and purification. African green monkey kidney (E6) cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum, 50 μg/ml gentamicin, 2 μg/ml amphotericin B, and 44 mM NaHCO3. Cells were infected with isolated Encephalitozoon intestinalis (ATCC CDC V307) spores and cultured at 37°C in CO2 (31).
Heavily infected cells were collected twice a week by gentle mechanical agitation and stored at 4°C for no more than 14 days prior to spore purification, at which time the infected cells were concentrated by centrifugation for 30 min at 1,200 x g. This and all subsequent steps were performed at 4°C. The supernatant solution was removed and the remaining sample was expressed through a 27-gauge needle three times to rupture host cells. In order to eliminate excessive debris, the sample was filtered through a sterile non-pyrogenic 5-μm-pore-size Cameo filter (GE Osmonics, Minnetonka, MN). The sample was then placed in a discontinuous Percoll gradient (5.5 ml of Percoll and 4.5 ml of phosphate-buffered saline) and centrifuged for 30 min at 1,200 x g. After centrifugation, the pellet containing the intact microsporidian spores was washed three times in phosphate-buffered saline by suspension and centrifugation. Spores were counted with a hemacytometer.
Infection assay. Human colonic carcinoma cells (3, 5, and 7 days old) were cultured to confluence in eight-well chamber slides. Cells were infected with E. intestinalis (1 x 106 spores) for three days. The culture-specific media were changed daily. Infection was assessed by counting the number of cells with clearly defined parasitophorous vacuoles per five fields as illustrated in Fig. 1A.
Spore internalization (phagocytosis) assay. E. intestinalis spores (100 μl serum-free Opti-MEM 1 medium [Gibco-BRL] containing 1 x 106 spores) were placed in each well of eight-well chamber slides containing confluent cultures of SW480, HT-29, Caco-2, or HCT-8 cells. All cultures were maintained for 4 h at 37°C in 5% CO2. Two additional experiments of this type were performed, one to determine the effects of greater spore attachment to target cells and the other to determine if performing the assay with cell line-specific media affected the results. For better spore attachment, this experiment was repeated with the addition of 2 mM manganese to all solutions and media (14). In order to determine the effect of removing the cells from their cell line-specific media, these experiments were also repeated in the media in which the cells were cultured. Qualitatively similar results were obtained with all of these experiments and only those experiments employing Opti-MEM alone are reported here.
In order to distinguish internalized from attached extracellular spores, plated cultured cells were fixed with 3.7% paraformaldehyde (pH 7.4) for 15 min. Cells were then washed three times with Tris-buffered saline (TBS), 5 min per rinse. Samples were blocked with 2% bovine serum albumin (BSA; Sigma Chemical Co.) in TBS for 10 min at room temperature. External spores were labeled with a 1:500 dilution of rabbit anti-E. intestinalis polyclonal antibody in TBS containing 2% BSA. Samples were incubated for 1 h at 37°C, washed three times in TBS and labeled with Oregon Green 488 goat anti-rabbit immunoglobulin G (IgG) (Molecular Probes, Eugene, OR) diluted 1:700 in TBS containing 2% BSA, and incubated for 1 h at 37°C. After washing of the samples three times, the cells were permeabilized with 0.02% Triton X-100 in TBS for 10 min, washed three times with TBS, and blocked with 2% BSA in TBS at room temperature. Intracellular structures were labeled with the rabbit polyclonal anti-E. intestinalis antibody (diluted 1:500 in TBS containing 2% BSA) and a mouse anti-zona occludens 1 (ZO-1) monoclonal antibody (BD BioSciences Pharmingen, San Jose, CA) diluted 1:25 in TBS containing 2% BSA. Cells were incubated for one hour at 37°C, washed three times with TBS, and labeled with goat anti-rabbit IgG AlexaFluor 594 (Molecular Probes) (diluted 1:500 in TBS containing 2% BSA) and Biotin-SP conjugate AffiniPure goat anti-mouse antibody (Jackson Immuno Research, West Grove, PA) diluted 1:300 in TBS containing 2% BSA. Cells were washed three times with TBS and then labeled with Streptavidin Alexa Fluor 488 conjugate dye (Molecular Probes) diluted 1:300 in TBS containing 2% BSA. The sample was washed three times with TBS and mounted on glass slides using mounting media. Extracellular and intracellular parasites were visualized with an epifluorescence microscope and counted. Intracellular parasite spores appeared red, target cell tight junction ZO-1 appeared green, and adherent extracellular spores were double labeled and appeared yellow (Fig. 1). For each slide, 300 attached spores and 20+ fields were assessed by using a 63x oil immersion objective.
Assessment of differentiation. SW480 cells are generally considered undifferentiated, and HT-29 cells are poorly differentiated when cultured in the type of medium used here (2). Caco-2 and HCT-8 cells exhibit a continuum of degrees of differentiation depending on the age of the culture and regional cell density. Criteria were established to divide the Caco-2 and HCT-8 cells into undifferentiated, intermediate, and differentiated cells. The criteria included the x-y surface area of each cell, which decreased as the cell become more differentiated and columnar (Table 1) and was readily measured by visualization of the ZO-1 probe outline which localized the tight junction and by the location of the cell on a dome or plateau (villus-like structure) of differentiated cells. Differentiated Caco-2 cells were columnar and small in x-y surface area, often located on domelike structures. Intermediate Caco-2 cells surrounded these differentiated cells. HCT-8 cells were considered differentiated when they were located on the tips of villi and crests, and intermediate cells had a larger x-y surface area and surrounded the villi and crests. Both Caco-2 and HCT-8 undifferentiated cells were large and flat. Undifferentiated HCT-8 cells predominated in 3-day-old cultures and were found in small clusters in 5-day-old cultures. In practice, a transparent template of the cells used to generate Table 1 was placed over the computer screen to act as a guide in determining the degree of differentiation of a field of Caco-2 or HCT-8 cells. Cells were also fixed with cacodylate-buffered 2.5% glutaraldehyde and prepared for scanning electron microscopy (SEM) (30). The apical brush border of undifferentiated and differentiated Caco-2 and HCT-8 cells was then imaged and the brush border development was used to confirm the assessment of differentiation (Fig. 2).
Measurement of the incidence of infection, phagocytosis, impalement, and germination. Labeled cells were visualized by fluorescent microscopy. Data were collected using SPOT 3.4.5 software (Diagnostic Instruments, Sterling Heights, MI). A field of 4,200 μm2 was used as the standard field size referred to here when measuring the incidence of infected, attached, phagocytized, or germinated spores. Germinated spores were further divided into those that had impaled a host cell and those that remained on the cell surface. Fields were randomly chosen, except in those cases where fields of undifferentiated, intermediate, and differentiated Caco-2 and HCT-8 cells were used. In those cases, composites of half and quarter fields were often pooled. The degree of host cell differentiation was first determined by visualizing the ZO-1 labeled outline of a cell population and then the parasite spores were visualized and counted. Germination was not seen with internalized spores during the 4-h time period of these experiments.
Infection and attached and phagocytized spores were distinguished by color. Double-labeled extracellular spores appeared yellow while intracellular parasite stages and internalized spores appeared red (Fig. 1A and B). Intracellular parasitophorous vacuoles were clearly evident from two days following spore exposure (1) and these were used to assess the incidence of infected cells. Individual red (internalized) spores that were seen in 4-h experiments were considered to have been phagocytized (5). Impalement was detected when a polar tube was seen extending from an attached extracellular spore (yellow) and, at the point at which the polar tube entered the host cell, the filament and any extruded sporoplasm appeared red (Fig. 1C). In germinated spores that did not impale a cell, the spore, polar tube and sporoplasm all appeared yellow.
To determine if spore germination was affected by the apical microclimate of individual host cells, the incidence of germination was determined in undifferentiated individual Caco-2 cells to which three or more spores were attached.
Statistical evaluation. Analysis of variance tests were used to determine the significance of differences between groups and Tukey's protected t tests were used as post hoc tests to determine the significance of differences between group mean values. Group mean differences were considered statistically significant when P was <0.05. In one experiment, the germination rates of spores were assessed when multiple spores were attached to individual cells. Frequency distributions of the percentage of spores germinating on individual host cells were plotted for cases where three to eight spores were attached/cell. In cases where the number of spores attached per cell was three, four, five, six, or eight, the distributions was bimodal and these data are illustrated showing the percentage of cells in which either all or none of the attached spores had germinated.
RESULTS
Data are presented for experiments in which the enterocyte exposure to E. intestinalis spores took place in serum-free Opti-MEM medium. Qualitatively similar results were obtained when the 4-h spore exposure took place in the culture medium appropriate to the cell line or the Opti-MEM to which MnCl2 had been added (data not shown). Figures 3, 4, and 5 illustrate the spore attachment (per five fields of 4,200 μm2 each), spore germination occurring per 100 visualized spores, phagocytized spores per 100 visualized spores, and germinated spores that impaled target cells per 100 visualized spores. The SW480 and HT-29 cells were treated as a homogenous monolayer. The Caco-2 cells and HCT-8 cells were divided into areas of undifferentiated, intermediate and differentiated cells on the basis of their x-y surface area and location (Table 1). Spore attachment was significantly different between cell lines: SW480>HT-29>Caco-2>HCT-8 (Fig. 3, 4, and 5). In Caco-2 cells, spore attachment was significantly greater for undifferentiated than for either intermediate or differentiated cells (Fig. 4) or HCT-8 undifferentiated or differentiated cells (Fig. 5), while in HCT-8 cultures, attachment was significantly greater with intermediate than with either undifferentiated or differentiated cells. Thus, in both Caco-2 and HCT-8 cells, as the cells assumed an enterocyte-like appearance, attachment decreased as the cells differentiated.
Host cell surface sulfated glycoaminoglycans have been implicated in E. intestinalis spore attachment (14). When the attachment assay was repeated using Optim-MEM with and without 1 mg/ml chondroitin sulfate A (Sigma Chemical Co.), the addition of this exogenous sulfated glycoaminoglycan significantly reduced attachment with all four cell lines (P < 0.01). The observed reductions were 80.0%, 70.6%, 84.1%, and 81.7% for SW480 cells, HT-29 cells, undifferentiated Caco-2 cells and intermediate HCT-8 cells, respectively.
The differences between and within cultures with respect to 4 -h spore attachment paralleled the degree of infection observed 3 days postinfection (Table 2).
Spore germination was observed in all cell lines. When normalized to the number of attached spores, such germination was significantly greater with differentiated Caco-2 cells than with any other group. Germination in the SW480 cultures was less than in any other, being significantly different from intermediate and differentiated Caco-2 and HCT-8 preparations. In both the Caco-2 and HCT-8 cultures, germination was significantly greater with intermediate and differentiated cells than with undifferentiated cells.
As all cultures were maintained in the same medium in these experiments, cell line differences in attached spore germination were considered likely due to some aspect of the host cell apical membrane. In the case of large undifferentiated Caco-2 cells multiple spores were frequently attached to a single cell. The percentage of attached spores germinating on the surface of individual cells was measured. Figure 6 illustrates the percentage of cells in which all or none of the attached spores had germinated for cells to which three, four, five, six, seven, or eight spores had attached. In all but the case of seven spores/cell the frequency distribution of germination was bimodal. For example, in the case of cells to which five spores had attached, in 63% of the cells all or none of the spores had germinated, while in the remaining 37% of the cells, one, two, three, or four of the attached spores had germinated. This bimodal frequency distribution suggests that the apical surface of individual cells differed such that some cell surfaces favored germination while others did not. The examples using large undifferentiated Caco-2 cells shown in Fig. 1B and C exemplify this phenomenon of most of the spores attached to a single cell being either not germinated or germinated, as none of the seven spores seen attached to a single cell in Fig. 1B germinated, while three and possibly four of the spores attached to a single cell in Fig. 1C had germinated.
It was possible to detect when germination resulted in impalement of the target cell. When normalized to the number of attached spores, impalement increased significantly with differentiation in Caco-2 cells (Fig. 4) and was significantly greater in differentiated HCT-8 cells than in intermediate HCT-8, SW480, HT-29, or Caco-2 cells (Fig. 3, 4, and 5). It has been suggested that E. intestinalis infection in Caco-2 cells may result from phagocytosis of sporoplasm at the cell surface, rather than a forceful penetration of the polar tube and subsequent injection of the sporoplasm (11). Scanning electron microscopy failed to show any example of sporoplasm being phagocytized at the cell surface. Typically the polar tube seemed to penetrate the host cell with no suggestion of cell surface envelopment of adherent sporoplasm (Fig. 7).
Because phagocytosis was rarely observed and the variance of the phagocytosis data was so great, no significant differences were detected between mean phagocytosis values measured in the various cultures. In the Opti-MEM experiments no phagocytosis was observed in HT-29 cultures or in differentiated Caco-2 or HCT-8 cells (Fig. 3, 4, and 5). Thus, as host cells became more differentiated, relative to the number of attached spores, the incidence of impaled cells increased and the incidence of phagocytized spores decreased. In experiments in which the 4-h spore exposure was conducted in the culture-specific media, phagocytosis as a percentage of visualized spores was as follows: SW480, 0.38% ± 0.20%; HT-29, 0.10% ± 0.10%; Caco-2 undifferentiated, 0.93% ± 0.24%; Caco-2 differentiated, 0.11% ± 0.11%; HCT-8 undifferentiated, 2.11% ± 0.90%; HCT-8 differentiated, 0.44% ± 0.24%. By comparison, when human macrophages were exposed to E. intestinalis spores at the same density for 2 h rather than 4 h the phagocytosis was 72.0% ± 12.1%.
DISCUSSION
There appear to be at least three major routes by which microsporidia infect host cells: impalement by extracellular spores, which is probably the major route of an initial infection; impalement of a cell by a spore germinating in a adjacent cell, which is probably the reason that infection in cell cultures often occurs as foci; and spore phagocytosis, particularly by macrophages, which is probably the major cause of parasite dissemination among those microsporidian species that cause disseminated human disease. Macrophage phagocytosis of microsporidian spores is usually defensive and results in spore destruction (7). However, there is evidence that if the spores can germinate within the macrophage before lysosomal fusion has begun to destroy the phagosome content, the infectious sporoplasm will escape destruction (12). In addition, some microsporidian species have been demonstrated to be able to block macrophage phagosome acidification and lysosome fusion (32, 33). Studies using cultured cell lines have indicated that a variety of cells are capable of phagocytizing microsporidian spores (5, 11, 12, 20), suggesting that an endocytic process may be important in intestinal microsporidiosis. Invasive microorganisms have devised a rich repertoire of mechanisms to subvert the host cell's endocytic processes, allowing invasion to occur even in cells that are not generally phagocytic (3, 4). The present study was designed to clarify whether phagocytosis was indeed a biologically significant means of infecting polarized intestinal enterocytes, a major target of human microsporidiosis.
The methods we employed allowed us to readily distinguish microsporidian spores that were attached to the host cell surface from those that had been phagocytized (Fig. 1B) and to distinguish attached spores that had germinated but not impaled a host cell from those that had impaled a cell, although the latter may have been undercounted due to difficulties in distinguishing intracellular polar tubes and sporoplasms (Fig. 1C). Spore attachment to host cells differed significantly between cell lines and between cells of the same cell line that were at different stages of differentiation (Fig. 3, 4, and 5). Because of differences in cell surface area and spore attachment between cell lines, we normalized the number of spores used in the assays to well surface area and the count of spore attachment to fields of a known area. The number of spores used in these experiments averaged about four spores/cell, considerably less than that generally used in the literature (e.g., 14). However, the number of spores we used is still probably higher than that found in the intestinal lumen within range of the mucosal epithelium, especially at the onset of an infection.
Hayman and colleagues (14) provided evidence that E. intestinalis spores adhere to surface sulfated glycoaminoglycans of Vero and CHO cells. Such attachment was reduced when a number of exogenous sulfated glycoaminoglycans were included in their adherence assay. Similarly, we found that chondroitin sulfate A inhibited attachment in all four cell lines used in our study, suggesting that the observed cell line and differentiation-dependent differences in attachment were the result of differences in the cell surface sulfated glycoaminoglycan expression. We did not observe the preferential spore attachment to intercellular junctions as has been reported elsewhere (11). Infection sites per field in the various cell lines (Table 2) paralleled attachment, in that the greatest infection was seen in SW480 cells, and undifferentiated Caco-2 and intermediate HCT-8 cell infections were greater than those seen in differentiated Caco-2 and HCT-8 cells. The infection seen in fields of differentiated Caco-2 cells was higher than one would have predicted based on the low spore attachment to such cells. However, the 3-day duration of the infection assay allowed time for infected undifferentiated and intermediate cells to differentiate. While it is known that infecting cells with Encephalitozoon spp. inhibits mitosis (23), the effects of infection on differentiation are not known.
Because there were such large differences in spore attachment between and within cell lines, spore germination was normalized to the number of attached spores. Germination was significantly less when spores were attached to the undifferentiated SW480 cell line and to undifferentiated Caco-2 and HCT-8 cells. This suggests that the degree of host cell differentiation influenced germination and that this effect was separate from any effect of the bulk phase (medium) environment and was probably independent of the cell surface sulfated glycoaminoglycan spore attachment effect. When spore germination was assessed in large, flat undifferentiated Caco-2 cells to which three or more spores had attached, a bimodal distribution was noted that suggested that while the surface of some cells was conducive to the germination of all attached spores, this was not the case for other cells (Fig. 6). These differences could not be attributed to the medium, as all cells in an assay were from the same slides and were exposed to the same bulk phase environment. Most studies of microsporidian spore germination have focused on the environmental effects of medium pH or ionic composition (13, 16, 19, 28) and not on the microenvironment of the unstirred layer that occurs at the brush border.
Impalement of the host cell by the polar tube at germination appears to be the major way in which infection came about in these cell lines. This was particularly true with the differentiated Caco-2 and HCT-8 cells and the moderately differentiated HT-29 cells. This might be explained by the fact that spores attached to undifferentiated cells would be likely to attach so that their long axis paralleled the apical surface while spores attached to a well developed brush border might be more likely to be oriented so that on germinating the polar tube would be pointing towards the target cell (20). The present study was not able to distinguish between cells that had been impaled and cells that had phagocytized the sporoplasm released from a spore germinating at the cell surface as has been suggested (11), although SEM data did not support this suggestion. In SEM studies, we were unable to find a suggestion of phagocytosis of surface sporoplasm but observed numerous examples of clear penetration of polar filaments through the host cell plasma membrane (Fig. 7). There are examples in the literature of germinating microsporidian spores caught in the act of impaling a cell and injecting sporoplasm into that impaled cell (for an example, see Fig. 1B) (5). Schottelius and colleagues (24) have suggested that with Encephalitozoon spp., the host cell plasma membrane forms an invagination at the site of contact with the polar filament, perhaps by some endocytic process. The polar tube then continues into the host cell cytoplasm within a membranous channel into which the sporoplasm is subsequently injected. There is also equivocal transmission electron microscopy data to support such a channel (22). Whether the polar tube impalement of host cells observed here involves any type of endocytosis of the polar tube is not known.
As expected, E. intestinalis whole spore phagocytosis was greater in the least differentiated cell line (SW480) and in the less differentiated Caco-2 and HCT-8 cells. In experiments in which the assay was conducted in serum-free medium, phagocytosis was not observed in HT29 cells or differentiated Caco-2 and HCT-8 cells (Fig. 3, 4, and 5). In experiments in which serum was included in the assay medium there was modest phagocytosis, but again it was significantly less in differentiated than undifferentiated cells and the incidence was very small when compared to that seen in human macrophages exposed to spores for half the time (0.10 to 0.44% versus 72%). More significantly, as Caco-2 and HCT-8 cells became more differentiated the incidence of phagocytosis decreased while the incidence of cells being impaled by germinating spores increased.
This study addresses only the mode of microsporidian infection at the apical pole of differentiated intestinal epithelial cells, the presumed site of the initial infection by ingested spores. The phagocytic activity of the basolateral surface of such cells may be different from that of the apical membrane, as has been observed with several enteric bacteria (e.g., Shigella) (21). While spore phagocytosis is undoubtedly the most important mode of entry of many microsporidia into macrophages, the present study supports the concept that the initial invasion of the intestinal epithelium by ingested parasite spores involves spore germination and host cell impalement rather than spore phagocytosis.
ACKNOWLEDGMENTS
This work was supported by U.S. PHS grant R21 DK64573-A1.
We thank Amad Al-Mahmoud, Morehouse School of Medicine Clinical Research Center, for assistance with statistical analyses.
REFERENCES
1. Barry, W. K. 2003. Molecular characterization of Encephalitozoon Microspora) replication kinetics in a murine intestinal cell line. J. Eukaryot. Microbiol. 50:169-174.
2. Chantret, I., A. Barbat, E. Dussaulx, M. G. Brattain, and A. Zweilbaum. 1988. Epithelial polarity, villin expression, and enterocytic differentiation of cultured human colon carcinoma cells: a survey of twenty cell lines. Cancer Res. 48:1936-1942.
3. Chen, X. M., P. L. Splinter, P. S. Tietz, B. Q. Huang, D. D. Billadeau, and N. F. LaRusso. 2004. Phosphatidylinositol 3-kinase and frabin mediate Cryptosporidium parvum cellular invasion via activation of Cdc42. J. Biol. Chem. 279:31671-31678.
4. Cossart, P., and P. J. Sansonetti. 2004. Bacterial invasion: the paradigms of enteroinvasive pathogens. Science 304:242-248.
5. Couzinet, S., E. Cejas, J. Schittny, P. Deplazes, R. Weber, and S. Zimmerli. 2000. Phagocytic uptake of Encephalitozoon cuniculi by nonprofessional phagocytes. Infect. Immun. 68:6939-6945.
6. Coyle, C. M., M. Wittner, D. P. Kotler, C. Noyer, J. M. Orenstein, H. B. Tanowitz, and L. M. Weiss. 1996. Prevalence of microsporidiosis due to Enterocytozoon bieneusi and Encephalitozoon (Septata) intestinalis among patients with AIDS-related diarrhea: determination by polymerase chain reaction to the microsporidian small-subunit rRNA gene. Clin. Infect. Dis. 23:1002-1006.
7. Didier, E. S. 1995. Reactive nitrogen intermediates implicated in the inhibition of Encephalitozoon cuniculi (phylum microspora) replication in murine peritoneal macrophages. Parasite Immunol. 17:405-412.
8. Didier, E. S. 2005. Microsporidiosis: an emerging and opportunistic infection in humans and animals. Acta Trop. 94:61-76.
9. Edlind, T. D., J. Li, G. S. Visvesvara, M. H. Vodkin, G. L. McLaughlin, and S. K. Katiyar. 1996. Phylogenetic analysis of beta-tubulin sequences from amitochondrial protozoa. Mol. Phylogenet. Evol. 5:359-367.
10. Findley, A. M., E. H. Weidner, K. R. Carman, Z. Xu, and J. S. Godbar. 2005. Role of the posterior vacuole in Spraguea lophii (Microsporidia) spore hatching. Folia Parasitol. (Prague) 52:111-117.
11. Foucault, C., and M. Drancourt. 2000. Actin mediates Encephalitozoon intestinalis entry into human enterocyte-like cell line, Caco-2. Microb. Pathog. 28:51-58.
12. Franzen, C., A. Muller, P. Hartmann, and B. Salzberger. 2005. Cell invasion and intracellular fate of Encephalitozoon cuniculi (Microsporidia). Parasitology 130:285-292.
13. Frixione, E., L. Ruiz, J. Cerbon, and A. H. Undeen. 1997. Germination of Nosema algerae (Microspora) spores: conditional inhibition by D2O, ethanol and Hg2+ suggests dependence of water influx upon membrane hydration and specific transmembrane pathways. J. Eukaryot. Microbiol. 44:109-116.
14. Hayman, J. R., T. R. Southern, and T. E. Nash. 2005. Role of sulfated glycans in adherence of the microsporidian Encephalitozoon intestinalis to host cells in vitro. Infect. Immun. 73:841-848.
15. Huschtscha, L., E. Rozengurt, and W. F. Bodmer. 1991. Growth factor requirements of human colorectal tumor cell; relations to cellular differentiation. Eur. J. Cancer 27:1680-1684.
16. Ishihara, R. 1967. Stimuli causing extrusion of polar filaments of Glugea fumiferanae spores. Can. J. Microbiol. 13:1321-1332.
17. Keohane, E. M., and L. M. Weiss. 1998. Characterization and function of the microsporidian polar tube: a review. Folia Parasitol. (Prague) 45:117-127.
18. Kotler, D. P., and J. M. Orenstein. 1998. Clinical syndromes associated with microsporidiosis. Adv. Parasitol. 40:321-349.
19. Leitch, G. J., Q. He, S. Wallace, and G. S. Visvesvara. 1993. Inhibition of the polar filament extrusion of the microsporidium, Encephalitozoon hellem, isolated from an AIDS patient. J. Eukaryot. Microbiol. 40:711-717.
20. Magaud, A., A. Achbarou, and L. Desportes-Livage. 1997. Cell invasion by the microsporidium Encephalitozoon intestinalis. J. Eukaryot. Microbiol. 44:81S.
21. Mounier, J., T. Vasselon, R. Hellio, M. Lesourd, and P. J. Sansonetti. 1992. Shigella flexneri enters human colonic Caco-2 epithelial cells through the basolateral pole. Infect. Immun. 60:237-248.
22. Scaglia, M., L. Sacchi, S. Gatti, A. M. Bernuzzi, P. P. Polver, I. Placentini, G. P. Croppo, A. J. Da Silva, N. J. Pieniazek, S. B. Slemenda, S. Wallace, G. J. Leitch, and G. S. Visvesvara. 1004. Isolation and identification of Encephalitozoon hellem from an Italian AIDS patient with disseminated microsporidiosis. APMIS 102:817-827.
23. Scanlon, M., A. P. Shaw, C. J. Zhou, G. S. Visvesvara, and G. J. Leitch. 2000. Infection by microsporidia disrupts the host cell cycle. J. Eukaryot. Microbiol. 47:525-531.
24. Schottelius, J., C. Schmetz, N. P. Kock, T. Schuler, I. Sobottka, and B. Fleischer. 2000. Presentation by scanning electron microscopy of the life cycle of microsporidia of the genus Encephalitozoon. Microbes Infect. 2:1401-1406.
25. Schwartz, D. A., I. Sobottka, G. J. Leitch, A. Cali, and G. S. Visvesvara. 1996. Pathology of microsporidiosis. Emerging parasitic infections in patients with acquired immunodeficiency syndrome. Arch. Pathol. Lab. Med. 120:173-188.
26. Sprague, V., and J. J. Becnel. 1999. Appendix: checklist of available generic names for microsporidia with type species and type hosts, p. 517-530. In M. Whittner and L. M. Weiss (ed.), The microsporidia and microsporidiosis. ASM Press, Washington, D.C.
27. Thomarat, F., C. P. Vivares, and M. Gouy. 2004. Phylogenetic analysis of the complete genome sequence of Encephalitozoon cuniculi supports the fungal origin of microsporidia and reveals a high frequency of fast-evolving genes. J. Mol. Evol. 59:780-781.
28. Undeen, A. H., and S. W. Avery. 1988. Effects of anions on the germination of Nosema algerae (Microspora; Nosematidae) spores. J. Invert. Pathol. 52:84-89.
29. Undeen, A. H., and E. Frixione. 1990. The role of osmotic pressure in the germination of Nosema algerae spores. J. Protozool. 37:561-567.
30. Visvesvara, G. S., G. J. Leitch, H. Moura, S. Wallace, R. Weber, and R. T. Bryan. 1991. Culture, electron microscopy, and immunoblot studies on a microsporidian parasite isolated from the urine of a patient with AIDS. J. Protozool. 38:105S-111S.
31. Visvesvara, G. S., H. Moura, G. J. Leitch, and D. A. Schwartz. 1999. Culture and propagation of microsporidia, p. 363-392. In M. Whittner and L. M. Weiss (ed.), The microsporidia and microsporidiosis. ASM Press, Washington, D.C.
32. Weidner, E. 1975. Interaction between Encephalitozoon cuniculi and macrophages. Parasitophorous vacuole growth and the absence of lysosomal fusion. Z. Parasitenkd. 47:1-9.
33. Weidner, E., and L. D. Sibley. 1985. Phagocytized intracellular microsporidian blocks phagosome acidification and phagosome-lysosome fusion. J. Protozool. 32:311-317.(Gordon J. Leitch, Tarsha )