ABSTRACT
Epidemiological studies suggest frequent association of enteropathogenic bacteria with Entamoeba histolytica during symptomatic infection. In this study, we sought to determine if the interaction with enteropathogenic (EPEC) or nonpathogenic Escherichia coli (strain DH5α) could modify the virulence of E. histolytica to cause disease in animal models of amebiasis. In vitro studies showed a 2-fold increase in CaCo2 monolayer destruction when E. histolytica interacted with EPEC but not with E. coli DH5α for 2.5 h. This was associated with increased E. histolytica proteolytic activity as revealed by zymogram analysis and degradation of the E. histolytica CP-A1/5 (EhCP-A1/5) peptide substrate Z-Arg-Arg-pNC and EhCP4 substrate Z-Val-Val-Arg-AMC. Additionally, E. histolytica-EPEC interaction increased EhCP-A1, -A2, -A4, and -A5, Hgl, Apa, and Cox-1 mRNA expression. Despite the marked upregulation of E. histolytica virulence factors, nonsignificant macroscopic differences in amebic liver abscess development were observed at early stages in hamsters inoculated with either E. histolytica-EPEC or E. histolytica-E. coli DH5α. Histopathology of livers of E. histolytica-EPEC-inoculated animals revealed foci of acute inflammation 3 h postinoculation that progressively increased, producing large inflammatory reactions, ischemia, and necrosis with high expression of il-1β, ifn-γ, and tnf-α proinflammatory cytokine genes compared with that in livers of E. histolytica-E. coli DH5α-inoculated animals. In closed colonic loops from mice, intense inflammation was observed with E. histolytica-EPEC manifested by downregulation of Math1 mRNA with a corresponding increase in the expression of Muc2 mucin and proinflammatory cytokine genes il-6, il-12, and mcp-1. These results demonstrate that E. histolytica/EPEC interaction enhanced the expression and production of key molecules associated with E. histolytica virulence, critical in pathogenesis and progression of disease.
INTRODUCTION
In countries of endemicity, the human protozoan parasite Entamoeba histolytica has become a common resident of the large intestine, where it survives by feeding on dead cells and bacteria without causing harm to the host (1, 2). However, for reasons not quite understood, E. histolytica transition to an invasive form and following invasion gives rise to amebic colitis with symptoms ranging from diarrhea, ameboma, and life-threatening extraintestinal invasion to the liver (3, 4). This suggests that changes in the gut environment may contribute to the pathogenesis of E. histolytica, leading to invasive amebiasis (5–7).
Coinfection with pathogenic organisms, especially enteropathogenic bacteria, may be an important factor that contributes to alteration of normal enteric microbiota and immune regulation, enhancing the virulence of E. histolytica in disease pathogenesis (5, 7, 8). Epidemiological studies have reported frequent presence of enteropathogenic bacteria in coinfection with symptomatic intestinal amebic infection (9–11). Under in vitro culture conditions, E. histolytica interaction with enteropathogenic bacteria for as little as 1 h enhanced parasite adherence and cysteine protease activity, with increased cytopathic activity (12–14). Another study showed that short-term coculture (12 h) of a pathogenic Escherichia coli serotype with an E. histolytica strain that had lost its capacity to produce amebic liver abscess (ALA) in hamsters restored parasite virulence by producing ALA (15). Likewise, Entamoeba dispar strain ADO cultured under axenic conditions did not produce ALA, but when it was maintained in culture with microbiota from patients, it produced liver damage in hamsters similar to that produced by axenic E. histolytica (16). The influence of bacteria on intestinal amebiasis has been observed in gnotobiotic athymic mice (17), in which the percentage of cecum colonization with E. histolytica strains HK-9 and NIH:200 increased when they interacted with E. coli or Clostridium perfringens. Similarly, in C3H/Hej mice, cecum damage produced by E. histolytica increased from 17% to 39% when parasites were cocultured with bacterial organisms (18). Furthermore, innate host immune responses also play a key role in susceptibility to E. histolytica infection. In HT-29 cells, exposure to E. histolytica in the presence of E. coli DH5α resulted in a synergistic increase in the expression of interleukin 8 (IL-8), IL-1α, and granulocyte-macrophage colony-stimulating factor (GM-CSF) (19). Bacterial components in the healthy gut maintain protective innate immune responses against E. histolytica, as seen in antibiotic-treated animals, that rapidly increase susceptibility of mice to E. histolytica invasion due to decreased CXCR2 in neutrophils (6). An increased percentage of pathogenic bacteria can lead to dysbiosis and contribute to chronic intestinal inflammation through the induction of proinflammatory cytokines gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), IL-1β, and IL-6 (8).
The aim of this study was to determine if E. histolytica interactions with enteropathogenic E. coli (EPEC) modulated parasite virulence factors and host innate immune responses associated with disease pathogenesis using several novel pathological approaches that are quantifiable to differentiate acute disease that affects gene transcription, proinflammatory cytokine release, and disease progression. Here we show that short-term interaction between E. histolytica and EPEC markedly upregulated cysteine protease, amebapore A, and cyclooxygenase (Cox)-like gene expression and increased parasite adherence and killing of host cells. In animal models of disease, E. histolytica-EPEC interaction enhanced cellular inflammatory reaction, granuloma formation, and necrosis in ALA in hamsters and increased secretory and proinflammatory cytokine responses in closed colonic loops in mice.
RESULTS
E. histolytica interaction with EPEC increases parasite virulence and cysteine protease activity.To determine if E. coli serotypes can directly alter the virulence of E. histolytica, parasites were exposed to either EPEC or E. coli DH5α (which is nonpathogenic). Phagocytosis of green fluorescent protein (GFP)-labeled bacteria was determined by confocal microscopy and flow cytometry from 1 to 6 h. Maximal phagocytosis by E. histolytica towards EPEC and E. coli DH5α occurred between 2 and 3 h (see Fig. S1A and SB in the supplemental material), and based on this observation, we used 2.5 h as the optimal time for E. histolytica interactions with bacteria for all subsequent studies. Although the optimal times for E. histolytica to phagocytose bacteria were similar, more E. coli DH5α organsims were phagocytized than EPEC (Fig. S1A and B). E. histolytica interaction with EPEC but not with E. coli DH5α significantly increased CaCo2 cell monolayer destruction (Fig. 1A). The increase in cytopathic effect produced by E. histolytica-EPEC interaction correlated with increased E. histolytica adherence to cell monolayers compared to that obtained with E. histolytica-E. coli DH5α and untreated E. histolytica (Fig. 1B).
Phenotypic characterizations of E. histolytica (Eh) following interaction with E. coli DH5α and EPEC for 2.5 h. (A) E. histolytica destruction of CaCo2 monolayers after 30 min of exposure. (B) Adhesion index of CFSE-labeled E. histolytica on CaCo2 cell monolayer after 15 min of exposure. (C) Zymogram analysis of proteolytic activity of E. histolytica (control), E. histolytica-E. coli DH5α, and E. histolytica-EPEC. The arrows point to cleavage products that were enhanced following E. histolytica-EPEC interaction. The histogram on the right represents the band density produced by individual EhCP activity. (D) Kinetic analysis of E. histolytica CP enzymatic degradation on the substrates Z-Arg-Arg-pNA (left), specific for EhCP-A1/5, and Z-Val-Val-Arg-AMC, specific for EhCP-A4 (right). Enzymatic unity for each substrate was calculated as the micromolar concentration per minute per milligram. For statistical analysis, the absorbance of E. histolytica (reference value, control) was compared with that of either E. histolytica-E. coli DH5α or E. histolytica-EPEC. P values were determined using Student’s t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
Even though increased cysteine protease activity has been reported for E. histolytica following exposure to different serotypes of pathogenic bacteria (13, 14, 20), the specific cysteine proteases have not been defined. To quantify the specific cysteine protease, we determined if E. histolytica interaction with EPEC and E. coli DH5α differentially degraded porcine gelatin by zymogram analysis and synthetic substrates. E. histolytica exposed to both bacteria showed increased proteolytic activity toward gelatin; however, E. histolytica-EPEC interaction showed significantly enhanced cleavages at 35 and 27/29 kDa (Fig. 1C, arrows) compared to the case with E. histolytica-E. coli DH5α or only E. histolytica. To determine if the major cysteine proteinases involved in parasite virulence were upregulated following E. histolytica-EPEC and E. histolytica-E. coli DH5α interaction, we quantified the degradation of E. histolytica CP-A1/5 (EhCP-A1/5)-specific synthetic substrate Z-RR and EhCP-A4-specific substrate Z-VVR (21). Surprisingly, both E. histolytica-EPEC and E. histolytica-E. coli DH5α interactions significantly degraded the substrates compared to untreated E. histolytica controls (Fig. 1D). Based on these results, we next determined if bacteria could module the virulence genes in E. histolytica. Interaction with EPEC significantly upregulated several E. histolytica virulence genes in a time-dependent manner (Table 1). Maximal gene expression occurred after 2 h of E. histolytica-EPEC interaction, significantly increasing the expression of EhCP-A1, EhCP-A2, EhCP-A4, EhCP-A5, Gal/GalNAc lectin (Hgl), amebapore A (Apa) and cyclooxygenase (Cox-1)-like mRNA. In contrast, E. histolytica-E. coli DH5α interaction induced only a slight overexpression of Hgl and Apa. Taken together, these results suggest that E. histolytica interaction with EPEC has a stronger effect in modulating E. histolytica virulence genes with increased enzymatic and cytopathic activity.
Expression of E. histolytica virulence genes following interaction with E. coli DH5α and EPECa
E. histolytica-EPEC interaction does not alter macroscopic liver damage but affects the evolution of lesions compared to that with E. histolytica-E. coli DH5.Hamsters inoculated with E. histolytica following 2.5 h of interaction with either EPEC or E. coli DH5α showed ∼25 to 40% ALA at 3 and 7 days postinoculation (p.i.) (Fig. 2A). The lesions produced by E. histolytica-E. coli DH5α and E. histolytica-EPEC showed similar whitish and irregular anatomic changes on the liver surfaces (Fig. 2B, arrows). Lesion morphology was consistent with the normal development of ALA produced by E. histolytica (22, 23). The percentages of liver damage produced by axenic E. histolytica, E. histolytica-E. coli DH5α, and E. histolytica-EPEC were comparable at 3 days p.i.; however, at 7 days p.i. the percent ALA in response to E. histolytica-EPEC was significantly higher (Fig. 2A). Animals inoculated with E. histolytica-EPEC showed noticeable total body weight loss at 3 days p.i. that became worse at day 7 p.i., whereas the body weight of animals inoculated with E. histolytica-E. coli DH5α decreased slightly at day 3 p.i. but thereafter was similar to that of E. histolytica controls. Animals inoculated with EPEC showed an increase in body weight similar to those of animals inoculated with E. histolytica and E. histolytica-E. coli DH5α at 6 days p.i. (Fig. 2C).
Macroscopic changes in livers of hamsters inoculated with E. histolytica (1 × 106 in 200 μl of TYI-S-33) following interaction with nonpathogenic E. coli (strain DH5α) or EPEC for 2.5 h. (A) ALA induced by E. histolytica following interaction with E. coli DH5α and EPEC at 3 and 7 days p.i. No significant damage was observed in the livers of animals inoculated with only EPEC or E. coli DH5α. **, P < 0.01. (B) Gross anatomy of liver lesions produced by E. histolytica following interaction with E. coli DH5α or EPEC after 3 and 7 days p.i. Note the whitish granular amebic lesions on the liver surfaces (arrows). At least six hamsters were evaluated for each time and condition. Animals inoculated with TYI-S-33 were used as controls. (C) Changes in hamster body weights at 3 and 7 days after infection with axenic E. histolytica and following interaction with E. coli DH5α and EPEC. Animals inoculated with EPEC only presented changes in body weight similar to those observed in the E. histolytica and E. histolytica-DH5α groups. P value was calculated by Student’s t test. ***, P < 0.001.
Despite the clear absence of gross anatomic differences in ALA development, histological analysis (Fig. 3A to C) showed marked cytopathological differences among the groups at early stages. At 3 h p.i., animals inoculated with E. histolytica-EPEC presented inflammatory foci with more dense, greater, and compact infiltration of polymorphonuclear leukocytes (PMNs) associated with E. histolytica (Fig. 3A). Comparative two-dimensional (2D) morphometric analysis of the inflammatory foci between E. histolytica-EPEC and E. histolytica-E. coli DH5α at 3 h p.i. showed significant differences in sizes of the foci, whereas no differences were observed between E. histolytica-E. coli DH5α and E. histolytica (Fig. 3D). Besides showing larger foci in E. histolytica-EPEC, each contained more PMNs, whereas livers inoculated with E. histolytica/E. coli DH5α or axenic E. histolytica showed smaller foci and less PMN infiltration (Fig. 3A, black arrows). Also, the inflammatory foci at the early stages of E. histolytica-EPEC showed the presence of E. histolytica organisms scattered inside each focus, sometimes with two or more E. histolytica organisms, in contrast to E. histolytica-E. coli DH5α or controls inoculated with axenic E. histolytica, which usually showed one E. histolytica organism at the center of inflammatory focus. Structural evolution of hepatic lesions at day 3 after infection with E. histolytica or E. histolytica-E. coli DH5α showed typical amebic granulomas constituted by epithelioid cells at the outer limit with E. histolytica located between the palisade cells and the necrotic center; no clear difference in the extents of damage was observed between these two groups (Fig. 3B). In contrast, hamsters inoculated with E. histolytica-EPEC showed fewer, compact granulomas but with a liver parenchyma substituted by an extensive necrotic and ischemic areas peripherally surrounded by acute and chronic inflammatory reactions with irregular outer limits (Fig. 3B). Lesions produced at 3 h and day 3 after infection with E. histolytica-EPEC were similar to the lesions reported at 6 h and 7 days after infection, respectively, with only axenic E. histolytica (22, 23). In hamsters inoculated only with E. coli DH5α or EPEC (without E. histolytica), a relatively small perivascular inflammatory infiltrate that progressed to compact small areas of ischemia was observed at 24 h p.i., and by day 3 p.i. the inflammatory response was completely cleared, presenting a normal liver parenchymal appearance (Fig. 3C). Although bacteria alone did not produce liver lesions, they induced early proinflammatory responses that could promote the survival of E. histolytica and progression of ALA.
Liver histopathology of hamsters inoculated with axenic E. histolytica, E. histolytica-E. coli DH5α, E. histolytica-EPEC, and E. coli DH5α or EPEC. (A) Inflammatory foci produced at 3 h after intrahepatic inoculation demonstrating E. histolytica (arrows) surrounded by PMNs. Inflammatory foci with E. histolytica-EPEC show more abundant inflammatory cells than with E. histolytica-DH5α or E. histolytica. (B) Three days after intrahepatic inoculation. Numerous inflammatory granulomas can be observed clearly in liver from E. histolytica-inoculated hamsters. E. histolytica-DH5α-inoculated animals show more irregular granulomas, whereas livers from animals inoculated with E. histolytica-EPEC present larger necrotic areas with small dense granulomas and intense inflammatory reaction. (C) Perivascular inflammatory infiltration around small ischemic zones produced by EPEC at 24 h p.i. and normal hepatic tissue 3 days after EPEC intrahepatic inoculation. Hematoxylin-eosin stain was used. (D) Histogram of 2D morphometry of inflammatory foci at 3 h post-intrahepatic inoculation. E. histolytica-EPEC yielded significantly larger foci than E. histolytica and E. histolytica-DH5α. P value was calculated by Tukey’s multiple-comparison test. ***, P < 0.05.
Proinflammatory cytokine responses were more pronounced in the liver of hamsters inoculated with E. histolytica-EPEC.E. histolytica-EPEC inoculations not only affected granuloma development but also stimulated robust proinflammatory cytokine expression during the induction of hepatic amebiasis. At 3 h p.i., E. histolytica-EPEC was found to have elicited higher expression of tnf-α, ifn-γ, and il-1β than did E. histolytica-E. coli DH5α and E. histolytica (Fig. 4A). Surprisingly, kc expression was significant higher towards EPEC, E. histolytica, and E. histolytica-DH5α. Proinflammatory cytokine responses were low towards EPEC and E. coli DH5α. The anti-inflammatory cytokine gene il-10 was overexpressed only at 3 h p.i. in all groups (Fig. 4A). By 12 h p.i., only tnf-α and il-1β were significantly increased, but at lower levels than at 3 h p.i., and they remained higher in the E. histolytica-EPEC group (Fig. 4B). At 12 and 24 h p.i., kc and il-10 gene expressions returned to basal levels in all groups (Fig. 4B and C). At 24 h p.i., except for ifn-γ, which was still upregulated in the E. histolytica-EPEC group, all the cytokines returned to basal level in all groups (Fig. 4C). These findings indicate that the proinflammatory milieu evoked by E. histolytica-EPEC might be critical in shaping the early development and progression of ALA.
RT-qPCR for cytokine expression in the liver of hamsters inoculated with E. histolytica, E. histolytica-E. coli DH5α, E. histolytica-EPEC, and EPEC. (A) At 3 h p.i., genes for all proinflammatory cytokines were overexpressed in animals inoculated with E. histolytica, E. histolytica-E. coli DH5α, and E. histolytica-EPEC. EPEC inoculation induced il-8 and il-10 overexpression, and E. coli DH5α only induced il-10 overexpression. (B) At 12 h p.i., expression of tnf-α decreased and ifn-γ increased in animals inoculated with E. histolytica, E. histolytica-DH5α, and E. histolytica-EPEC. (C) At 24 h p.i., ifn-γ was still overexpressed in E. histolytica-EPEC-inoculated hamsters. Hamsters inoculated with only E. histolytica culture medium were used as a basal expression control (Ctl); the GAPDH gene was used as a housekeeping gene to normalize mRNA levels. Data correspond to means ± SDs from three independent experiments repeated twice. Animals inoculated with TYI-S-33 were used as a control. P value was calculated by two-way ANOVA. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (compared with values for E. histolytica). Differences over Ctl (*), E. histolytica (+), and E. histolytica-E. coli DH5α (•) are shown.
E. histolytica-EPEC inhibited Muc2 secretory lineage with a corresponding increase in activity in the distal colon driven by proinflammatory cytokines.To determine if E. histolytica-E. coli DH5α and E. histolytica-EPEC interaction altered the virulence of E. histolytica, we tested this using a short-term colonic loop model with Math1GFP mice. Math1 protein is a transcription factor that promotes the differentiation of multipotential colonic cells to secretory goblet cells whose lineage is lost by the ablation of this factor. We have recently shown that a decrease in Math1GFP activity (signal) correlates with E. histolytica-induced proinflammatory responses and microbial translocation in the colon (24). Interestingly, EPEC and E. histolytica-EPEC inoculation in the proximal colon inhibited Math1GFP activity in both the proximal and distal colon (Fig. 5A and B). However, reduction in Math1 mRNA expression was higher in E. histolytica-EPEC- than in EPEC-inoculated loops. This contrasts with colonic loops inoculated with E. histolytica, E. coli DH5α, or E. histolytica-E. coli DH5α, which showed a significant decrease in Math1GFP activity in the proximal colon with a corresponding increase in activity in the distal colon (Fig. 5B and C) compared to that in phosphate-buffered saline (PBS)-inoculated controls. Predictably, Muc2 mRNA expression was the highest towards E. histolytica-EPEC in both the proximal and distal colon (Fig. 5D). E. histolytica, E. coli DH5α, and E. histolytica-E. coli DH5α stimulated modest Muc2 mRNA expression in the proximal colon, with a similar increase in the distal colon.
Math1 and Muc2 secretion and gene expression in colons of mice inoculated with E. histolytica, EPEC, E. coli DH5α, E. histolytica-EPEC, and E. histolytica-E. coli DH5α. (A) Math1GFP activity in closed proximal colonic loops from control animals and animals inoculated with E. histolytica, E. coli DH5α, EPEC, E. histolytica-EPEC, and E. histolytica-E. coli DH5α. The dotted line indicates the colonic loop ligations. The images show an overlay of the black-and-white colon tissue and the GFP signal (fluorescence images are shown in color, superimposed over a black-and-white image of the dissected colon ex vivo). The fluorescence emission signal in the green channel was quantified and corrected for background. (B) Histogram representation of Math1GFP signal in the proximal and distal colon. (C) Relative q-PCR for Math1 mRNA in the proximal and distal colonic tissues from panel A. (D) Differences in relative Muc2 expression evaluated by RT-qPCR in the proximal and distal colon. Gene expression levels were normalized using β-actin. Animals inoculated with PBS were used as the control. P value was calculated by Student’s t est. AU, arbitrary units. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
As host protection from E. histolytica colonic invasion has been attributed mainly to neutrophil infiltration via oxygen-free radicals induced by the presence of TNF-α and IFN-γ (25, 26), we quantified changes in neutrophil recruitment using myeloperoxidase (MPO) activity as a readout and proinflammatory responses after 3 h in colonic loops inoculated with E. histolytica, E. histolytica-EPEC, or E. histolytica-E. coli DH5α. However, no significant differences were observed in MPO activity induced by inoculating E. histolytica, E. histolytica-E. coli DH5α, or E. histolytica/EPEC (Fig. 6A). Moreover, colonic tissue cytokine gene expression was upregulated in loops inoculated with E. histolytica-EPEC; the highest expression was found for il-6, il-12, mcp-1, tnf-α, and il-1β (Fig. 6B). EPEC inoculation alone significantly increased the expression of il-6, il-1β, and mcp-1 (Fig. 6B), suggesting that the overexpression of these cytokines was a direct or partial response to bacterial components in the E. histolytica-EPEC inoculum. There were no discernible differences in loops inoculated with E. histolytica or E. histolytica-E. coli DH5α. ifn-γ expression slightly increased in loops inoculated with E. histolytica and E. histolytica-E. coli DH5α and remained at basal levels in response to E. histolytica-EPEC or EPEC (Fig. 6B). il-10 levels were unaltered regardless of the inoculum. Taken together, these results demonstrate that E. histolytica-EPEC interaction enhanced parasite virulence in the colon to dampen Math1 and upregulate Muc2 and proinflammatory cytokine expression.
Proinflammatory responses in mouse proximal colon inoculated with E. histolytica, E. histolytica-EPEC, E. histolytica-E. coli DH5α, or EPEC. (A) Acute inflammation was evaluated by MPO activity; PBS-inoculated colons were used as controls. (B) Cytokine mRNA in the proximal colon (site of E. histolytica inoculation) was quantified by RT-qPCR. Gene expression levels were normalized using β-actin. For both panels, 5 mice were analyzed for each condition in two independent experiments. Animals inoculated with PBS were used as a control. P value was calculated by Student’s t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Differences over Ctl (*), E. histolytica (+), and E. histolytica-E. coli DH5α (•) are shown.
DISCUSSION
Here we report that E. histolytica interaction with EPEC significantly increased parasite virulence, which led to enhanced destruction of CaCo2 cell monolayers and aggressive immune responses in hamster liver and mouse colon. E. histolytica-EPEC interactions markedly upregulated E. histolytica virulence genes EhCP-A1, EhCP-A2, EhCP-A4, EhCP-A5, Hgl, Apa, and Cox-1 in a time-dependent manner. In contrast, E. histolytica interaction with nonpathogenic E. coli DH5α modestly induced the expression of Hgl and Apa without modifying parasite virulence. Correlated with gene expression, high cysteine protease activity and adhesion to host cells mediated by the Gal/GalNAc lectin were induced by E. histolytica-EPEC interaction. These virulence factors have been directly associated with E. histolytica killing cells in vitro (27–30) and in vivo to alter secretory responses and tight junction proteins in the gut (31, 32).
In animal models of disease, increased expression of the Gal/GalNAc lectin and EhCP genes has been observed during ALA development (30, 33–39). However, despite a considerable increase in proteolytic activity and adherence to mammalian cells induced by E. histolytica-EPEC interaction, no significant differences were observed in the percentages of ALA produced in hamsters inoculated with E. histolytica-EPEC and with E. histolytica alone. The lack of correlation between the ability to induce ALA with E. histolytica proteolytic activity was previously observed (40) using two clones of HM1:IMSS_B that were not able to produce ALA in gerbils but had high CP activity. These findings are partially in contrast to those of a previous study (15) that restored virulence of E. histolytica strain HMI:IMMS to produce ALA in hamsters when interacting with a pathogenic serotype of E. coli. This indicates that E. histolytica interaction with pathogenic E. coli may not increase E. histolytica capacity to produced larger lesions, but it could enhance its adaptation and survival in the host by reinforcing E. histolytica pathogenesis. In accordance with this, preincubation of E. histolytica with live E. coli O55 increased parasite resistance to oxidative stress by reversing changes in the expression of genes involved in protein synthesis, intracellular trafficking, nutrition, homeostasis, and peroxidases altered by oxidative stress (41). It is well known that EhCPs play important roles in disease pathogenesis, as demonstrated by the inability of E. histolytica to induce lesions in gerbil liver when transfected with an antisense gene for EhCP-A5, despite conserving the cytopathic capacity over mammalian cell monolayers (42). Additionally, E. histolytica invasion in human intestinal xenografts was reduced >95% when EhCP-A1 was blocked with the specific inhibitor WRR483 or with EhCP-A4-specific inhibitor WRR605 in C3H/HeJ mice (43, 44). Therefore, it is possible that E. histolytica interaction with EPEC improved survival of E. histolytica under host conditions by increasing EhCP production and adhesion to mammalian cells.
Histological analysis of hepatic lesions suggested a faster progression of the lesion in response to E. histolytica-EPEC. This could be related to a more aggressive immune response expressed by greater numbers of PMNs surrounding E. histolytica, observed exclusively at a very early stage of E. histolytica inoculation, and higher expression of the proinflammatory cytokines IL-1β, IFN-γ, and TNF-α than with E. histolytica-E. coli DH5α or E. histolytica alone. No significant upregulation of inflammatory cytokines in response to EPEC occurred, suggesting that the sole presence of EPEC was not the cause for the more aggressive immune response during ALA. More likely, the combination of E. histolytica virulence factors and host immune response to pathogenic bacterial components was responsible for the robust acute inflammatory response. Overexpression of the proinflammatory cytokines IL-1β, IL-8, IFN-γ, and TNF-α was previously observed during the first 3 to 12 h after infection with E. histolytica in the livers of hamsters and in a model of human liver tissue explants (33, 45). These results suggest that the expression of proinflammatory cytokines participated in tissue damage during ALA by inducing a more aggressive response of PMNs and other immune cells recruited in the liver.
Although the liver is a good organ to study E. histolytica pathogenesis, the natural environment of E. histolytica is the colon, where it colonizes and preferentially feeds on bacterial components of the normal microbiota by phagocytosis (46). In our study, we evaluated whether the interaction of E. histolytica with E. coli DH5α or EPEC induced changes in stimulating mucus secretion, Muc2 mucin gene expression, and repression of the transcription factor Math1, events critical in the pathogenesis of intestinal amebiasis (24). For these studies we used Math1GFP mice, which express the green fluorescent protein (GFP) in goblet cells, to determine differences in intestinal homeostasis alteration. Somewhat surprisingly, inoculating E. histolytica or E. histolytica-E. coli DH5α caused a shift in Math1GFP signal from the proximal colon, where E. histolytica was inoculated, towards the distal colon. In contrast, with E. histolytica-EPEC inoculation, Math1GFP signal was repressed in both the proximal and distal colon, indicating a greater imbalance in intestinal cellular homeostasis, related to the repression of the secretory cell linage and robust proinflammatory responses. Muc2 mucin mRNA expression was high in both the proximal and distal colon, with a corresponding increase in mucus secretion when E. histolytica-EPEC was inoculated. Muc2 is the main component of the intestinal mucus that forms the first line of innate host defense against pathogenic organisms in the large intestine; deficiency of this glycoprotein induces sensitivity and greater damage to the epithelium by E. histolytica (47–49). We have recently shown (50) that high production of Muc2 mucin in response to dextran sodium sulfate (DSS) caused oxidative stress and apoptosis of goblet cells that led to the loss of the protective mucus barrier.
During the establishment of intestinal amebiasis, regulation of the immune response by cytokines is widely associated with tissue damage and/or protection against E. histolytica invasion. In our study, we observed that E. histolytica-EPEC induced greater expression of tissue il-12, il-6, il-1β, tnf-α, and mcp-1 and inhibited the expression of ifn-γ that remained at basal levels. Elimination of E. histolytica infection is associated with the presence of IFN-γ, which plays a critical role in the destruction of E. histolytica and inhibition of EhCP activity (51–53). In contrast, tnf-α and il-4 are correlated with the progression of disease in murine models of amebic colitis, where it was found that the Th2 adaptive response determines the establishment of infection in the host. In our study, mice inoculated with E. histolytica-EPEC showed high expression of Mcp-1 protein, which is produced by epithelial cells and functions as a potent chemoattractant for monocytes, naive dendritic cells, and basophils and is important for the recruitment of antigen-presenting cells (APCs) to the site of damage, where IL-6 and transforming growth factor β (TGF-β) are secreted and induce the differentiation of naive CD4+ cells to Th17 type cells (54–56). Th17 cells secrete IL-23, contributing to expansion of the Th17 response, and the combination of IL-17 and IL-23 is apparently important for the recruitment of neutrophils, induction of mucus secretion, and inhibition of the Th1 response (57). Interestingly, in this study, the cyclooxygenase enzyme that produced prostaglandin 2 (PGE2) was overexpressed when E. histolytica interacted with EPEC. PGE2 can condition dendritic cells to induce IL-23 and the migration of neutrophils through the Th17 immune response, simultaneously inhibiting Th1 responses, which is necessary for host defense against E. histolytica (57–59). However, despite high upregulation of the cyclooxygenase enzyme in E. histolytica-EPEC and overexpression of Mcp-1 and IL-1β, no differences were observed in neutrophil recruitment in mice inoculated with axenic E. histolytica, E. histolytica-E. coli DH5α, E. histolytica-EPEC, and EPEC.
Coinfection with pathogenic bacteria normally occurs in intestinal amebiasis, and ALAs harbor bacteria from the gut (10, 11, 60); however, it is unclear if pathogenic bacteria affect E. histolytica virulence. Unfortunately, an animal model that allows us to study in detail how the prevalence of pathogenic bacteria affects the establishment and progression of intestinal amebiasis has not yet been established. In summary, the results of this work show that E. histolytica interaction with EPEC or nonpathogenic E. coli markedly influences parasite virulence by upregulating key virulence genes critical in disease pathogenesis. In particular, E. histolytica-EPEC interaction increased EhCP-A1, EhCP-A2, EhCP-A4, EhCP-A5, Hgl, Apa, and Cox-1 mRNA expression associated with increased parasite adherence to and killing of CaCo2 cell monolayers. E. histolytica-EPEC interactions increased cellular necrosis in hamster liver and evoked intense mucus secretagogue activity and downregulation of Math1 activity associated with high expression of proinflammatory cytokines in mouse colon. These results demonstrate that E. histolytica interaction with EPEC during parasite colonization in asymptomatic hosts can potentially influence the expression and production of key molecules associated with E. histolytica virulence, critical in disease pathogenesis and progression of the disease.
MATERIALS AND METHODS
E. histolytica culture.E. histolytica strain HM1:IMSS, which produces large ALA in hamsters, was used in this study. E. histolytica were axenically cultured at 37°C in TYI-S-33 medium supplemented with 20% adult bovine serum (ABS). For all experiments, cultures of E. histolytica were harvested and used in logarithmic growth phase (72-h culture).
Bacterial cultures.The nonpathogenic E. coli strain DH5α and the enteropathogenic E. coli (EPEC) strain B171-0111:NM, donated by Teresa Estrada (CINVESTAV, Mexico), were grown in Luria-Bertani (LB) agar plates at 37°C for 15 to 18 h. For bacterium-E. histolytica interaction, a colony was taken and grown in LB broth at 37°C with shaking for 15 to 18 h. A total of 500 μl of the culture was seeded in 5 ml of TYI-S-33 medium without ABS and incubated at 37°C with shaking for 4 to 6 h, until reaching an optical density (OD) at 600 nm of 0.6 to 0.8.
E. histolytica-bacterium interaction and phagocytosis kinetics.E. histolytica organisms (1 × 106) in logarithmic phase of growth were transferred to TYI-S-33 medium without ABS and 1 × 108 CFU of bacteria from one of the strains of E. coli were added for an E. histolytica-bacterium ratio of 1:100. The E. histolytica-bacterium combinations were incubated at 37°C for the desired time, the supernatant was discarded, and PBS was added and chilled on ice for 10 min. E. histolytica was recovered by centrifugation at 200 × g and 4°C for 5 min, washed 3 times with cold PBS containing 200 μg/ml of gentamicin, and centrifuged at 200 × g for 5 min. E. histolytica was washed twice with PBS to remove antibiotic residue. To quantify bacteria phagocytosed by E. histolytica, bacteria were transfected to express the green fluorescent protein (GFP) using the plasmid pEGFP and incubated with 5 × 105 E. histolytica organisms for 1, 2, 3, 4, 5, and 6 h at 37°C. After incubation, E. histolytica organisms were recovered by centrifugation and washed with PBS-gentamicin. E. histolytica organisms were fixed with 3.7% formaldehyde for 20 min, washed 2 twice with PBS by centrifuging at 200 × g, and suspended in 500 μl of PBS. Intracellular fluorescence was measured by confocal microscopy and flow cytometry.
CaCo2 cell culture, cytopathic effect, and E. histolytica adhesion.CaCo2 cells were cultured in Dulbecco modified Eagle minimal essential medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 100 U/ml of penicillin-streptomycin and incubated at 37°C with 5% CO2. The cells were passaged with 0.25% trypsin-EDTA once they reached 90% confluence. For cytopathic effect, cells were seeded in 24-well plates in triplicate at a density of 5 × 105 and cultured until 80% confluence. Cells were then incubated with 2.5 × 105 E. histolytica organisms for each condition for 30 min and 1 h at 37°C, then placed on ice for 15 min to detach E. histolytica, washed with cold PBS, and stained with 0.1% methylene blue. Cells were extracted with 0.1 M HCl, and the samples were read at 650 nm with a spectrophotometer. The quantified absorbance was compared with that given by intact cell monolayers not exposed to E. histolytica, which were considered a control (0% damage). For the adhesion assay, E. histolytica for each condition was labeled with 8 μM CellTrace carboxyfluorescein succinimidyl ester (CFSE) for 15 min and washed twice with PBS prior to exposure to the cell monolayer. After 30 min of incubation at 37°C, nonattached E. histolytica organisms were gently washed with warm PBS, and the plate was read in a fluorometer at an excitation wavelength of 492 nm and an emission wavelength of 517 nm.
Zymogram of E. histolytica proteases and degradation of synthetic peptide substrates.E. histolytica lysed for each condition was obtained by three cycles of freeze-thawing. Two micrograms of E. histolytica lysate freed from debris by centrifugation at 2,000 × g was loaded in 1% gelatin–12% polyacrylamide gels, and the electrophoresis was developed at 70 V for 15 min and 80 V for 2 h at 4°C. The gel was washed twice for 15 min with shaking in 2.5% Triton X-100 solution, incubated overnight at 37°C in activation buffer at pH 7 (Tris-OH solution at 100 mM), washed with distilled water, and stained with Coomassie blue. The clear areas in the gels revealed the digestion of the gelatin by the cysteine proteinase activity. The gels were scanned with the SigmaGel program. The densitometry analysis of the bands was done using the software ImageJ (http://rsb.info.nih.gov/nih-image/). Proteolytic activity in E. histolytica lysates was evaluated by the degradation of the synthetic peptide substrate Z-Arg-Arg-pNA (Bachem), which is degraded by EhCP-A1/5, and the specific substrate for EhCP4, Z-Val-Val-Arg-AMC (21) (Enzo Life Sciences, NY). The reaction mixture consisted of a 0.1 mM concentration of the substrate in the reaction buffer (5 mM EDTA, 50 mM NaCl [pH 7]) followed by E. histolytica proteins (25 μg). The release rate of pNA was measured by absorbance at 405 nm, while the release of AMC was evaluated in a fluorometer with excitation at 365 nm and emission at 440 nm, every 2 min for 20 min at room temperature, for both substrates. One unit of enzymatic activity was defined as the number of micromoles of digested substrate per minute per milligram of protein.
Experimental hepatic amebiasis.Two-month-old male hamsters (Mesocricetus auratus) weighing 80 to 100 g were fasted for 24 h prior to surgery. Subsequently, they were anesthetized initially with 3% isoflurane and then with 1.5% of the same anesthetic during the surgical procedure. The abdominal surface of the hamster was shaved and a longitudinal incision of the abdominal wall was made in two planes, skin and muscle wall, to expose the liver. E. histolytica (1 × 106 in 200 μl of TYI-S-33 without ABS) of each group (axenic E. histolytica, E. histolytica-E. coli DH5α, and E. histolytica-EPEC after three washes with PBS plus gentamicin) was then directly inoculated in the left lobe using a tuberculin syringe. The abdominal incision was sutured in two planes with 2/0 silk thread. After 3, 12, 24, 72, and 96 h of inoculation, the hamsters were sacrificed with an overdose of anesthetic; the whole liver was first weighed, and then only the liver lesion was dissected and weighed to calculate the percentage of the damaged liver in relation to the total liver weight (percent ALA). Representative tissue fragments were obtained for RNA extraction, and other fragments of the same section were fixed with 4% paraformaldehyde (pH 7.2) for histological analysis.
Experimental intestinal amebiasis.Mice used were 10 to 12 weeks old and kept in cages with filter-sterilized tops. C57BL/6 mice and Math1GFP mice (strain 013593) with a C57Bl/6 genetic background (Jackson Laboratory, Bar Harbor, ME) were used for colonic loop infections with E. histolytica. To inoculate mice, a laparotomy was performed on anesthetized mice to expose the proximal colon; a colonic loop was created by suturing from the cecum-colon junction to 3 cm down the length of the colon. Mice were then inoculated with 1 × 106 log-phase E. histolytica organisms that interacted or not with E. coli DH5α or EPEC for 2.5 h, washed twice with PBS, and suspended in 100 μl of PBS. As controls, 100 μl of PBS and E. coli DH5α and EPEC cultures were used. Infections were carried out for 3 h and the colonic loop was excised. The intestinal content of the colonic loop was recovered in a 1.5-ml Eppendorf tube, homogenized, and centrifuged at 3,000 × g to remove debris; the supernatant was recovered and normalized to 1 mg/ml. Cytokine release in the intestinal lumen was evaluated by Luminex addressable laser bead-based immunoassay (mouse focused 13-plex Discovery Assay; Eve Technologies, Calgary, AB, Canada).
Whole-colon imaging.To screen whole tissue ex vivo, the colons from naive or infected Math1-GFP mice were dissected and then visualized using an InVivo Xtreme 4MP whole-body imaging system (Bruker, Billerica, MA) as previously described (24, 61). The green channel was visualized using 470-nm excitation and 535-nm emission wavelengths. The imaging protocol contained two steps: reflectance (black and white) imaging (2-s exposure time) and fluorescent imaging at the given wavelength (5-s exposure time). Images from the InVivo Xtreme were acquired and analyzed using the Bruker molecular imaging software MI SE (version 7.1.3.20550). The fluorescence intensity was quantified by measuring the mean fluorescence signal intensity (corrected by background) in a constant region of interest for each individual organ.
RT-qPCR for virulence factors in E. histolytica and host cytokines.The TRIzol method (Invitrogen Life Technologies, Carlsbad, CA) was used for RNA extraction. For E. histolytica, 800 μl of TRIzol was used for 5 × 106 parasites, homogenized by repetitive pipetting and incubated for 5 min at room temperature. For tissue samples, 100 to 200 μg was used, 1 ml of TRIzol was added, and the tissues were homogenized. For all samples, RNA extraction was performed according to the manufacturer’s instructions. For reverse transcription-quantitative PCR (RT-qPCR), the First Strand cDNA synthesis retrotranscription kit (FERMENTAS, Thermo Scientific, Waltham, MA) was used, based on oligo(dT) sequences as primers. For quantitative amplification by reverse transcription, the reactions were performed in a Rotor-Gene 3000 real-time PCR system (Nainital, Uttarakhand, India). Each reaction mixture (total, 20 μl) contained 100 ng of cDNA, 10 μl of a master mix for 2× SYBR green qPCR (catalog number 204072; Qiagen, Venlo, Netherlands), and primers at 1 μM (2 μM). The amplification conditions were as follows: 40 cycles of 95°C for 15 s, melting temperature for 20 s, 68°C for 20 s, and an adjacent melting step (67 to 95°C). The relative differences in gene expression were calculated using the threshold cycle (2−ΔΔCT) methods with the Rotor-Gene software. For E. histolytica gene expression, E. histolytica ribosomal DNA (rDNA) was used as a control, and for hamster and mouse tissues the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and β-actin genes, respectively, were used as housekeeping genes. Primers used for the detection of different cytokines and virulence factors are shown in Table S1.
Statistical analysis.All statistical analyses were performed using GraphPad Prism version 5.01 (GraphPad Software, San Diego, CA). Student’s t test and two-way analysis of variance (ANOVA) with Bonferroni post-t tests were used. Significant differences between groups were assessed at a P value of <0.05. Results are presented as the means of three independent experiments, along with standard deviation (±SD), unless otherwise indicated.
Ethics statement.The Centre for Research and Advanced Studies (CINVESTAV) fulfills the standard of the Mexican Official Norm (NOM-062-ZOO-1999) Technical Specifications for the Care and Use of Laboratory Animals. The Institutional Animal Care and Use Committee (IACUC/ethics committee) from CINVESTAV has reviewed and approved all animal experiments (protocol number 0505-12, CICUAL 001). Mouse studies were approved under the University of Calgary (Calgary, AB, Canada) Animal Care Committee, which adheres to the principles and policies regarding the care and use of experimental animals of the Canadian Council on Animal Care.
ACKNOWLEDGMENTS
We thank Enrique González for excellent technical support and Björn Petri from the Mouse Phenomics Resources Laboratory and the Live Cell Imaging Facility in the Snyder Institute at the University of Calgary for imaging.
This work was funded in part by the Miguel Aleman Foundation (2016, 2017) and by PRODEP, Program for Professional Development, Secretary of Education, Mexico, in awards to V.T. and by a discovery grant (RGPIN-2019-04136) from the Natural Sciences and Engineering Research Council of Canada awarded to K.C. L.A.F.-L. and A.L.-C. received scholarship support from the National Council of Science of Technology, Mexico, during their Ph.D. studies (321615 and 314101, respectively).
We declare no conflict of interest.
FOOTNOTES
- Received 11 April 2019.
- Returned for modification 27 May 2019.
- Accepted 26 August 2019.
- Accepted manuscript posted online 16 September 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00279-19.
- Copyright © 2019 American Society for Microbiology.
