Roles of Cell Adhesion and Cytoskeleton Activity in Entamoeba histolytica Pathogenesis: a Delicate Balance
http://www.100md.com
感染与免疫杂志 2005年第3期
Unite de Pathogenie Microbienne Moleculaire
Unite de Biologie Cellulaire du Parasitisme, INSERM U389
Unite d'Histotechnologie et Pathologie
Plateforme d'Imagerie Dynamique, Institut Pasteur, Paris
Laboratoire de Biologie et Contrle des Organismes Parasites, UPRES 398-IFR 75, Faculte de Pharmacie, Universite Paris-Sud, Chatênay-Malabry, France
ABSTRACT
The protozoan parasite Entamoeba histolytica colonizes the human large bowel. Invasion of the intestinal epithelium causes amoebic colitis and opens the route for amoebic liver abscesses. The parasite relies on its dynamic actomyosin cytoskeleton and on surface adhesion molecules for dissemination in the human tissues. Here we show that the galactose/N-acetylgalactosamine (Gal/GalNAc) lectin clusters in focal structures localized in the region of E. histolytica that contacts monolayers of enterocytes. Disruption of myosin II activity impairs the formation of these structures and renders the trophozoites avirulent for liver abscess development. Production of the cytoplasmic domain of the E. histolytica Gal/GalNAc lectin in engineered trophozoites causes reduced adhesion to enterocytes. Intraportal delivery of these parasites to the liver leads to the formation of a large number of small abscesses with disorganized morphology that are localized in the vicinity of blood vessels. The data support a model for invasion in which parasite motility is essential for establishment of infectious foci, while the adhesion to host cells modulates the distribution of trophozoites in the liver and their capacity to migrate in the hepatic tissue.
INTRODUCTION
Cells communicate with their environment through a large variety of cell surface molecules. Contact of cell surface adhesion molecules with their external ligands triggers signaling pathways that lead to cytoskeleton rearrangements necessary both for cell adhesion and for cell displacement. The adhesion sites of higher eukaryotes are complex, dynamic, molecular platforms that link the substratum in the cell outside to the internal cytoskeleton (9, 30). The actomyosin network provides the central cytoplasmic anchor of focal adhesion areas (30). Deregulation of these environment-cell interactions in multicellular organisms accounts for disease processes such as, for example, tumor initiation, invasion, and metastasis of cancer cells in human malignancies (13). Unicellular eukaryotes also rely on cell adhesion and cytoskeleton functions for survival. In particular, parasites encounter different environments when they colonize and invade a multicellular organism. Their motilities, cell adhesion properties, and aggressive behaviors vary, accounting for different physiopathologies that depend on the localization in the host.
Entamoeba histolytica, the etiological agent of human amoebiasis, colonizes the large bowel. Parasite invasion of the intestinal epithelium is characterized by extensive degradation of the extracellular matrix by amoebic secreted proteases, human cell cytolysis, and phagocytosis. Highly motile trophozoites gain access to the bloodstream in the region of ulcer formation and eventually disseminate to other organs, causing most commonly amoebic liver abscesses (ALA) (22). A central contributor in E. histolytica pathogenesis is the highly effective actomyosin cytoskeleton, which allows fast morphological changes associated with amoebic motility and spatial reorganization of cellular components (6). Production of the light meromyosin (LMM) moiety of the myosin II heavy chain in genetically engineered trophozoites (strain LMM) was shown to strongly reduce cell motility in vitro and in vivo, to affect the retrograde motion of amoebic surface molecules, and to abolish cytotoxicity when organisms were incubated in the presence of epithelial cells (2, 4).
The coordinated action of the E. histolytica cytoskeleton and surface adhesion molecules is essential for pathogenesis (25). The major adhesion molecule of the parasite is the immunodominant galactose/N-acetylgalactosamine (Gal/GalNAc) inhibitable lectin (17, 21). The Gal/GalNAc lectin plays a central role in adhesion of the trophozoite to intestinal mucins and to enterocytes (27) and has been shown to be a major virulence factor (21). This lectin is a 260-kDa heterodimeric complex composed of a light subunit (Lgl) and a heavy subunit (Hgl) bound by disulfide bonds (20). The two known isoforms of Lgl (31 and 35 kDa) are extracellular and undergo different posttranslational modifications; the 31-kDa subunit has a carboxyl-terminal acyl-glycosylphosphatidylinositol anchor that probably attaches it to the plasma membrane, while the 35-kDa subunit does not have the glycolipid but undergoes more extensive glycosylation (17, 18). Studies with genetically engineered trophozoites showed that Lgl is required for E. histolytica virulence and plays a role in parasite adhesion and killing of host cells (1, 10). The various Hgl isoforms are integral membrane proteins with a short cytoplasmic domain (Fig. 1) (15, 24). Production of the lectin heavy-chain cytoplasmic domain from either the Hgl2 (strain HGL-2) or Hgl3 (strain HGL-3) isoform that is anchored to the membrane through the transmembrane segment has a dominant negative effect and reduces significantly the adhesion of amoebae to epithelial cells (29; this study). Furthermore, strain HGL-2 has normal motility in vitro but is immobile when it is injected into the liver (4).
Here we studied the interaction of strains LMM and HGL-2 with monolayers of enterocytes and the capacity of these engineered parasites to produce liver abscesses in the hamster animal model after injection by the intraportal route. Our results demonstrated the central roles of myosin II and the Gal/GalNAc lectin in the invasive process of E. histolytica and also revealed different cellular and physiopathological behavior when the function of these molecules was disrupted. Whereas the strain deficient in myosin II activity was avirulent, the strain defective in Gal/GalNAc function disseminated through the blood and massively invaded the hepatic tissue, resulting in a metastatic-like spread of infection (13).
MATERIALS AND METHODS
Bacteria, parasites, cell lines, and cell biology procedures. Escherichia coli strains DH5 and BL21DE3(pLysS) were used for cloning procedures and overproduction of the His6-Hgl3TM-COO– fusion peptide, respectively. E. histolytica strain HM1:IMSS was cultivated axenically in TYI-S-33 (5). Virulent HM1:IMSS trophozoites (referred to as strain HM1 or the wild-type strain below) derived from a single culture that was axenized after passage through hamster liver (23) were transfected by electroporation with the control vector pExEhNeo (7), and its derived constructs were used for production of LMM (2) and Hgl2SP-TM-COO– (Fig. 1) (4). The engineered strains designated LMM and HGL-2 were maintained in the presence of 10 μg of G418 (Gibco-BRL) per ml and were frozen at –80°C for long-term storage. Before each experiment the concentration of drug was increased to 30 μg/ml for 48 h. Production of LMM and of the Hgl carboxyl-terminal region was checked by Western blotting of trophozoite extracts after transfection and regularly thereafter. New transfections were carried out before each series of hamster infestations.
Adhesion of trophozoites to monolayers of polarized Caco-2 cells grown to confluence for 14 days and fixed with 5% formaldehyde was quantified as described previously (19). Adhesion tests were carried out in the presence of 200 mM glucose or 200 mM galactose.
Production and purification of the cytoplasmic tail of the Gal/GalNAc lectin in E. coli. The DNA fragment coding for the putative transmembrane segment and cytoplasmic tail of Hgl3 (coordinates 3736 to 3940 in GenBank accession no. L14815) (15) was amplified by PCR by using oligonucleotides 5'CGCGGATCCGTTGGAGCTATTGCAGCGG (coding strand) and 5'CGGGATCCCTGCAGCTTATCCATTGAATGTTGCTGC (noncoding strand) (the sequence homologous to hgl3 is indicated by boldface type; the BamHI and PstI sites are underlined). The purified PCR fragment (approximately 310 bp) was cleaved with BamHI/PstI and ligated to vector pRSET A (Invitrogen) digested with the same endonucleases. The correctness of the construction was checked by restriction analysis, and the recombinant plasmid was inserted by transformation into E. coli BL21DE3(pLysS) for overproduction of the His6-Hg13TM-COO– fusion peptide.
A 1-liter culture of the strain overproducing His6-Hgl3TM-COO– was grown at 37°C to a density of approximately 108 CFU/ml and induced with 5 mM isopropyl--D-thiogalactopyranoside (IPTG). After 2 h the cells were sedimented, washed with cold medium, and resuspended in 20 ml of lysis buffer (20 mM HEPES-KOH [pH 7.6], 500 mM KCl, 1.5% [wt/vol] Sarkosyl) freshly supplemented with 0.5 mg of lysozyme per ml, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (Sigma), and 10 mM leupeptin (Sigma). All subsequent manipulations were carried out at 4°C. Cells were lysed by sonication (four 15-s pulses with 15-s intervals between pulses), and cell debris was eliminated by centrifugation at 18,000 x g for 1 h. The supernatant was added to 3 ml of cobalt affinity matrix (Talon; Clontech), which was previously equilibrated with lysis buffer, and mixed for 1 h in a rotating wheel. The resin was allowed to settle by gravity and washed twice with lysis buffer (15 min each), and this was followed by two stringent washes (30 min each) with lysis buffer containing 10 mM imidazole. His6-Hgl3TM-COO– was eluted with 500 mM imidazole in lysis buffer, which yielded a peptide that was more than 90% pure (data not shown). The fusion peptide was used to immunize rabbits for production of polyclonal antibodies. The serum obtained recognized the carboxyl terminus of both Hgl2 and Hgl3.
Parasite immunostaining and confocal microscopy. Trophozoites grown to approximately 80% saturation in 25-cm2 bottles were washed and resuspended in 6 ml of Dulbecco modified Eagle medium (DMEM) prewarmed to 37°C by gentle tapping in the bottle. The trophozoite suspension was divided into aliquots and mixed with DMEM containing glucose (or galactose in control experiments, as specified below) at a final concentration of 200 mM. Trophozoites were counted, and 1 x 105 to 2 amoebae were added to Caco-2 cells grown for 14 days in coverslips previously washed twice with DMEM lacking serum. Cocultures were incubated for 30 min at 37°C under 10% CO2. The medium was carefully decanted, and the coculture was fixed for 30 min at 37°C with 3.7% paraformaldehyde prepared in phosphate-buffered saline. Cells were permeabilized for 3 min with 0.1% Triton X-100 in phosphate-buffered saline. Immunolabeling with the anti-lectin monoclonal antibody CD6 (12) visualized with the secondary anti-monoclonal antibody coupled to Cy3 (Molecular Probes) and staining of filamentous actin with phalloidin-fluorescein isothiocyanate (1:25 dilution; Sigma) were carried out as described previously (2, 16, 26). Fluorescent samples were observed with a confocal laser scanning microscope (16, 26). Optical sections were acquired on serial planes (0.5 μm) starting from top of the epithelial cell monolayer that contacted the amoebae to at least two-thirds of the trophozoite height.
Animal model experiments. Intraportal infestation of male Syrian golden hamsters (Mesocricetus auratus) with wild-type or engineered strains of E. histolytica, animal dissection, histological analysis, and quantitative analysis of inflammatory foci were carried out as described previously (23). The significance of statistical differences was assessed by using the method of Fisher. The production of LMM and of the Hgl C-terminal region in trophozoites recovered from liver abscesses was checked by Western blotting by using a monoclonal antibody against the vesicular stomatitis virus epitope (2) and the rabbit polyclonal serum against His6-Hgl3TM-COO– (this study).
RESULTS
Characterization of engineered trophozoites deficient in motility and adhesion. In order to investigate the impact of trophozoite motility and adhesion in the biology and pathogenesis of E. histolytica, we characterized the behavior of parasite strains LMM (2) and HGL-2 (4).
Production of the carboxyl-terminal domain of myosin II (LMM fragment), which is involved in association of this mechanochemical enzyme, has a dominant negative effect on the activity of the cell actomyosin cytoskeleton (2, 3). E. histolytica trophozoites transfected with a construct express the LMM coding region so that the ratio of LMM protein fragment to myosin II is 4:1 (strain LMM) (2). The amoebae have low motility, are not cytotoxic for Caco-2 cells, and exhibit a reduction in the multiplication rate compared to trophozoites transfected with vector pExEhNeo (mock strain) when they are grown at high drug concentrations (>20 μg of G418 per ml) (2). The adhesion of the LMM strain to monolayers of Caco-2 cells is 53.9% ± 8.7% (n = 3) compared to the adhesion of the mock strain.
Production of the cytoplasmic tail from the Hgl3 isoform of the E. histolytica Gal/GalNAc lectin (Fig. 1) inhibits parasite adhesion to epithelial cells (29). We constructed plasmid pEhC-Hgl2, in which the extracellular region of hgl2 (24) was replaced by a FLAG epitope directing production of a 21-kDa peptide that contained the signal peptide, the transmembrane segment, and the cytoplasmic tail of Hgl2 (Hgl2SP-TM-COO–) (Fig. 1a). The peptide was detected in trophozoites transfected with pEhC-Hgl2 by using anti-FLAG and by using serum against His6-Hgl3TM-COO– (data not shown). The cytoplasmic tail of Hgl2 differed from Hgl3 at two positions; at one of these positions Ile1270 in Hgl3 was replaced by a threonine in Hgl2 in the sequence Thr-Ile-Thr that Vines et al. (29) showed to be important for trophozoite adhesion to epithelial cells (Fig. 1b). However, trophozoites transfected with pEhC-Hgl2 (strain HGL-2) also exhibited a reduction in adhesion to Caco-2 monolayers (46.8% ± 16.6% adhesion; n = 3), demonstrating that the cytoplasmic tail of Hgl2 cross talks with intracellular factors involved in parasite adhesion. The HGL-2 strain remained cytotoxic, as evaluated by inspection of cocultures with Caco-2 monolayers, and its growth was normal compared to the growth of the mock strain (data not shown).
Inhibition of adhesion in engineered trophozoites correlates with a reduced surface of contact with epithelial cell monolayers. The adhesion of E. histolytica to epithelial cells is strongly inhibited by millimolar amounts of galactose or N-acetylgalactosamine (14), a feature explained to a significant extent by the Gal/GalNAc lectin adhesion activity (references 17, 21, and 25 and references therein). Immunofluorescence labeling of the trophozoites for the Gal/GalNAc lectin in cocultures with Caco-2 cell monolayers showed that in the presence of galactose the lectin was distributed homogeneously around the surface of the few amoebae that remained in contact with the monolayer (Fig. 2a). When adhesion was not inhibited (in the presence of glucose instead of galactose), the trophozoites exhibited an extended surface of interaction with the monolayer, where focal clusters of surface Gal/GalNAc lectin made intimate contact with the epithelial cells (Fig. 2b) (8). The interaction sites likely represented focal adhesion points assembled at the trophozoite surface that were functional analogues of structures found in adhesive animal cells (9, 30). The human cells in the monolayer, in turn, showed a significant increase in subcortical actin filaments underneath the region of contact with the parasite (data not shown) (3).
The specific contributions of the actomyosin cytoskeleton and of the Gal/GalNAc lectin to the formation of the trophozoite adhesion structures were investigated by confocal microscopy of Caco-2 monolayers cocultured with trophozoite strains transfected with the vector pExEhNeo (control), LMM, and HGL-2 (Fig. 2c to e). Control trophozoites bound to the enterocytes through a large surface with numerous focal adhesion points (Fig. 2c), as observed for wild-type parasites (Fig. 2b). LMM and HGL-2 parasites were not firmly attached to the apex of enterocytes. Approximately 55% of these parasites were washed away from the monolayer as a consequence of their low-adhesion phenotype (see above). The LMM and HGL-2 trophozoites that remained in contact with the surface of enterocytes might have had weak phenotypes because E. histolytica transfectants are not homogeneous populations due to variability in the copy number of the transfected plasmid. This heterogeneity was best appraised by other immunofluorescence studies of a population of transfected trophozoites that were shown to produce different levels of the protein C-PAK encoded by an episomal vector (11; E. Labruyere, unpublished observations). Confocal analysis of the LMM trophozoites bound to enterocytes revealed their rounded shape and a contact surface with the monolayer that was smaller than that of wild-type amoebae (Fig. 2d). These features correlated with enrichment of myosin II in the parasite cytoskeleton (data not shown). The results suggest that disruption of the mechanoenzyme function might lead to rigidification of the actomyosin cytoskeleton, accounting for the altered shape of the trophozoite. The HGL-2 parasites remaining in contact with the enterocytes had an intermediate phenotype, in which the contact surface between the trophozoites and the monolayer appeared to be smaller than that observed for wild-type trophozoites and some Gal/GalNAc lectin clusters in the region of trophozoite-enterocyte contact were still present. The phenotypes described for the different strains were confirmed by inspection of confocal images in a blind test. The results revealed the importance of myosin II and of the Gal/GalNAc lectin for binding to host cells and that disruption of the activity of these molecules has a distinct effect on the interaction.
Nonmotile E. histolytica is avirulent, while the adhesion-deficient strain exhibits exacerbated virulence and a distinct histopathology in hamster liver abscesses. The distinct behavior of the LMM and HGL-2 strains in cocultures with enterocytes prompted us to correlate the cell biology results with the pathogenicity of the strains in vivo, as assayed in the hamster liver abscess model (28). Young adult hamsters were inoculated by the intraportal route with wild-type (nontransfected strain HM1), control, LMM, and HGL-2 trophozoites. Macroscopic abscesses, increases of 30 to 40% in weight due to edema, and extended inflammation were observed in livers of animals sacrificed 5 days after infestation except in the case of hamsters infected with the LMM strain (Fig. 3). In the latter case the appearance of the liver remained normal, and histology studies showed that the hepatic tissue was intact (Fig. 3).
The development of ALA was monitored by histology and immunolabeling of parasites. Serial sections of the infected livers were prepared from animals sacrificed at times ranging from 6 h to 5 days postinfection. The number of inflammatory foci containing parasites was determined at 6 h postinfection in five different sections from lobes 1 and 2 of each liver (lobe numbering is shown in Fig. 3 for the liver infected with the LMM strain). This stage of ALA was characterized by the presence of a large number of inflammatory foci (70%), the majority of which lacked trophozoites (23). The number of foci increased, and the size increased progressively. The foci were characterized by the presence of numerous polymorphonuclear cells (PMN), most of which were necrotic. Similar results were observed after infestation of animals with the wild-type and control strains (Fig. 4). LMM infection yielded a significantly lower number of foci, and amoebae were found only in very rare cases (Fig. 4). The small foci present at 6 h postinfestation did not lead to detectable lesions after 5 days, showing that the amoebic infection was eradicated (Fig. 3). In contrast, infection with HGL-2 revealed a high capacity for successful liver invasion with a unique histopathology. A high number of small inflammatory foci were detected at 6 h postinfection (Fig. 4). The majority of the foci contained parasites at this stage of infection, a situation opposite the situation found in animals challenged with wild-type amoebae (Fig. 4). Thus, HGL-2 had an increased capacity to survive the host immune response that was most likely responsible for the high number of abscesses disseminated throughout the liver at 5 days postinfection (Fig. 3). Parasites were distributed with a relatively disorganized pattern in the interior of the abscess necrotic area and close to its periphery. This observation was also distinct from the observation made for normal abscesses of E. histolytica, in which a peripheral layer of trophozoites was surrounded by inflammatory cells, mostly PMN, delimiting the border between the area of necrosis and the hepatic parenchyma (Fig. 3) (23, 28).
In order to understand the different histopathologies of liver abscesses caused by the wild-type and HGL-2 strains, we examined the formation of liver abscesses at early stages postinfection (6 and 24 h). The adhesion-defective strain caused more inflammatory foci, but the progression was slower than that with wild-type amoebae, as shown by the significantly smaller diameter after 24 h of infection (Fig. 5a and b). Interestingly, more than 30% of the HGL-2 foci were contiguous to blood vessels (Fig. 5a and c), and the sinusoids exhibited a layer of host PMN resulting from inflammation (Fig. 5a, panel 4).
DISCUSSION
Motility and the capacity to adhere to host cells are essential for pathogenesis of the human parasite E. histolytica. However, the cell biology of these processes remains poorly understood. In this study we observed that the main adhesion molecule of E. histolytica, the Gal/GalNAc lectin, forms clusters in the region of interaction with enterocytes, as previously described (8). The clusters make intimate contact with the enterocyte surface, as demonstrated by the overlapping labeling of the parasite Gal/GalNAc and the enterocyte subcortical actin filaments (Fig. 2). We propose that the numerous focal points present at the parasite-host cell interface are functional analogues of adhesion structures found in multicellular organisms. These structures are dynamic macromolecular platforms that anchor transmembrane adhesion proteins to the intracellular cell cytoskeleton (9, 30). Studies with transfected trophozoites suggest that there is similar cross talk in E. histolytica. Disruption of the Gal/GalNAc function leads to a weaker interaction of trophozoites with the enterocyte monolayer. A stronger effect leading to reduced contact between the trophozoite and the monolayer was found when the activity of the cytoskeletal component myosin II was disrupted. Inspection of confocal images showed that these phenotypes correlate with a decrease in the number of focal adhesion points (Fig. 2b to e; data not shown). A quantitative assessment of this correlation requires a methodology that we are currently developing to determine the three-dimensional structure of individual trophozoites based on the confocal microscopy data. The present work revealed that molecules of E. histolytica involved in adhesion and in actomyosin cytoskeleton activity participate in the assembly of structures for the parasite-host interaction. Identification of the other protein components involved and of the signaling pathways that control the dynamics of these structures should be a major asset for characterizing the molecular mechanisms that support E. histolytica binding and motility in monolayers of enterocytes.
Other studies and our studies showed that disruption of myosin II (strain LMM) and Gal/GalNAc (strains HGL-2 and HGL-3) activities have a major effect on trophozoite adhesion and motility (2, 4, 21, 29; this study). Interestingly, the phenotypes observed in vitro may not be directly applied to the complex situation of parasite invasion in vivo. This is exemplified by strain HGL-2, which has normal motility in monolayers of enterocytes but is immobile when it is injected into the liver tissue (4). Thus, we studied the physiopathology of ALA development to correlate our cell biology studies on parasite adhesion with the invasion of the liver parenchyma by the adhesion-deficient strains LMM and HGL-2. Administration of parasites by the portal vein route implies that the trophozoites are transported by the bloodstream to the liver network of irrigation vessels. Successful invasion of the liver parenchyma requires (i) that the trophozoite binds strongly enough to the endothelium to counteract the pressure imposed by the bloodstream, (ii) extravasation, and (iii) migration through the three-dimensional structure of the hepatic tissue. The trophozoite has to survive the host inflammatory response at each of these steps.
Intraportal injection of wild-type and control strains caused ALA progression with the histopathology described previously (references 23 and 28 and references therein). Examination of liver sections from animals infected with strain LMM showed the presence of very few trophozoites, and no inflammation was observed at 24 h. The normal tissue architecture was recovered at 5 days postinfection (Fig. 3). The LMM strain thus displays a low efficiency of tissue penetration and is unable to survive in the liver. This phenotype correlates with the inability of the trophozoites to penetrate Caco-2 cell monolayers and with a lack of cytotoxicity in vitro (2). Both of these features probably contribute to preventing migration of the LMM parasites in the liver tissue and to facilitating killing of the parasites by the host immune response. The properties of strain LMM indicate that it is a good candidate for vaccination studies to elicit a host immune response against amoebiasis.
HGL-2 trophozoites injected by the portal route infect all lobes of the hamster liver, causing a large number of infection foci. The foci differ from those found in wild-type infections by their smaller size, by their disorganized structure, and by their preferential localization close to blood vessels with normal endothelial cells (Fig. 3 and 5). The destruction of vessels is an early event in ALA caused by wild-type E. histolytica. The only difference detected in vitro between the two strains was the reduced adhesion of HGL-2 to monolayers of enterocytes, while the cytotoxicities and motilities were identical. However, HGL-2 trophozoites injected intrahepatically remained immobile, showing that the Gal/GalNAc lectin is essential for migration of E . histolytica in the three-dimensional structure of the liver parenchyma (4). This is most likely due to the role of this lectin in adhesion to host cells. The HGL-2 strain might also have a defect in the contact-dependent transfer of the Gal/GalNAc lectin to the host cell (12), a process that was proposed to contribute to invasion of the host tissue (12, 25). The data suggest that during infection the HGL-2 trophozoites brought by the portal vein are spread throughout the liver by the blood vessel network. The reduced adhesion of the trophozoites might limit binding to the vessel wall and extravasation to the hepatic tissue, therefore facilitating dissemination in the organ through the bloodstream. When trophozoites penetrate the liver parenchyma, their very low mobility leads to the establishment of infection foci that are concentrated in regions adjacent to the blood vessels (Fig. 5). These foci progress into small abscesses, in which the amoebae are distributed both at the periphery, as in wild-type ALA, and in the necrotic region. In spite of the difficulty that they have penetrating the liver tissue, the HGL-2 trophozoites resist very well the host immune response. How this feature relates to their low-adhesion phenotype remains to be established. The histopathology associated with HGL-2 infection reveals the complex effect of reduced trophozoite adhesion in host-pathogen interactions resulting in enhanced spread of the parasite and in exacerbated virulence.
ACKNOWLEDGMENTS
We acknowledge E. Labruyere for communication of unpublished results. We thank E. Tannich for kindly providing the plasmid carrying cloned hgl2. A. Leroy and G. Bailey are acknowledged for providing monoclonal antibody CD6. We are grateful to P. Sansonetti for his continuous support and interest in this project.
This work was supported in part by grants from the French Ministry of National Education through the PRFMMIP program. P.T. was supported by EMBO and FCT postdoctoral fellowships.
REFERENCES
1. Ankri, S., F. Padilla-Vaca, T. Stolarsky, L. Koole, U. Katz, and D. Mirelman. 1999. Antisense inhibition of expression of the light subunit (35 kDa) of the Gal/GalNac lectin complex inhibits Entamoeba histolytica virulence. Mol. Microbiol. 33:327-337.
2. Arhets, P., Olivo J.-C., P. Gounon, P. Sansonetti, and N. Guillen. 1998. Virulence and functions of myosin II are inhibited by overexpression of light meromyosin in Entamoeba histolytica. Mol. Biol. Cell 8:1537-1547.
3. Burns, C. G., M. Reedy, J. Heuser, and A. De Lozanne. 1995. Expression of light meromyosin in Dictyostelium blocks normal myosin II function. J. Cell Biol. 130:605-612.
4. Coudrier, E., F. Amblard, C. Zimmer, P. Roux, J.-C. Olivo, M.-C. Rigothier, and N. Guillen. 2005. Myosin II and the Gal-GalNAc lectin play a crucial role in tissue invasion by Entamoeba histolytica. Cell. Microbiol. 7:19-27.
5. Diamond, L. S., D. R. Harlow, and C. C. Cunnick. 1978. A new medium for the axenic cultivation of Entamoeba histolytica and other Entamoeba. Trans. R. Soc. Trop. Med. Hyg. 72:431-432.
6. Guillen, N. 1996. Role of signalling and cytoskeletal rearrangements in the pathogenesis of Entamoeba histolytica. Trends Microbiol. 4:191-196.
7. Hamann, L., I. R. Nicke, and E. Tannich. 1995. Transfection and continuous expression of heterologous genes in the protozoan parasite Entamoeba histolytica. Proc. Natl. Acad. Sci. USA 92:8975-8979.
8. Hughes, M. A., C. W. Lee, C. F. Holm, S. Ghosh, A. Mills, L. A. Lockhart, S. L. Reed, and B. J. Mann. 2003. Identification of Entamoeba histolytica thiol-specific antioxidant as a GalNAc lectin-associated protein. Mol. Biochem. Parasitol. 127:113-120.
9. Jockusch, B. M., P. Bubeck, K. Giehl, M. Kroemker, J. Moschner, M. Rothkegel, M. Rudiger, K. Schluter, G. Stanke, and J. Winkler. 1995. The molecular architecture of focal adhesions. Annu. Rev. Cell. Dev. Biol. 11:379-416.
10. Katz, U., S. Ankri, T. Stolarsky, Y. Nuchamowitz, and D. Mirelman. 2002. Entamoeba histolytica expressing a dominant negative N-truncated light subunit of its gal-lectin are less virulent. Mol. Biol. Cell 13:4256-4265.
11. Labruyere, E., C. Zimmer, V. Galy, J.-C. Olivo-Marin, and N. Guillen. 2003. EhPAK, a member of the p21-activated kinase family, is involved in the control of Entamoeba histolytica migration and phagocytosis. J. Cell Sci. 116:61-71.
12. Leroy, A., G. de Bruyne, M. Mareel, C. Nokkaew, G. B. Bailey, and H. Nelis. 1995. Contact-dependent transfer of the galactose-specific lectin of Entamoeba histolytica to the lateral surface of enterocytes in culture. Infect. Immun. 63:4253-4260.
13. Leroy, A., M. Mareel, G. De Bruyne, G. Bailey, and H. Nelis. 1995. Metastasis of Entamoeba histolytica compared to colon cancer: one more step in invasion. Invasion Metastasis 14:177-191.
14. Li, E., A. Becker, and S. L. Stanley, Jr. 1988. Use of Chinese hamster ovary cells with altered glycosylation patterns to define the carbohydrate specificity of Entamoeba histolytica adhesion. J. Exp. Med. 167:1725-1730.
15. Mann, B. J., B. E. Torian, T. S. Vedvick, and W. A. Petri, Jr. 1991. Sequence of a cysteine-rich galactose-specific lectin of Entamoeba histolytica. Proc. Natl. Acad. Sci. USA 88:3248-3252.
16. Marion, S., P. Tavares, P. Arhets, and N. Guillen. 2004. Signal transduction through the Gal-GalNAc lectin of Entamoeba histolytica involves a spectrin-like protein. Mol. Biochem. Parasitol. 135:31-38.
17. McCoy, J. J., B. Mann, and W. A. Petri, Jr. 1994. Adherence and cytotoxicity of Entamoeba histolytica or how lectins let parasites stick around. Infect. Immun. 62:3045-3050.
18. McCoy, J. J., B. Mann, T. S. Vedvick., Y. Pak, D. B. Heimark, and W. A. Petri, Jr. 1993. Structural analysis of the light subunit of the Entamoeba histolytica galactose-specific adherence lectin. J. Biol. Chem. 268:24223-24231.
19. Padilla-Vaca, F., S. Ankri, R. Bracha, L. A. Koole, and D. Mirelman. 1999. Down regulation of Entamoeba histolytica virulence by monoaxenic cultivation with Escherichia coli O55 is related to a decrease in expression of the light (35-kilodalton) subunit of the Gal/GalNAc lectin. Infect. Immun. 67:2096-2102.
20. Petri, W. A., Jr., M. D. Chapman, T. Snodgrass, B. J. Mann, J. Broman, and J. I. Ravdin. 1989. Subunit structure of the galactose and N-acetyl-D-galactosamine-inhibitable adherence lectin of Entamoeba histolytica. J. Biol. Chem. 264:3007-3012.
21. Petri, W. A., Jr., R. Haque, and B. J. Mann. 2002. The bittersweet interface of parasite and host: lectin-carbohydrate interactions during human invasion by the parasite Entamoeba histolytica. Annu. Rev. Microbiol. 56:39-64.
22. Ravdin, J. I. (ed.). 1988. Amebiasis. Human infection by Entamoeba histolytica. John Wiley & Sons, New York, N.Y.
23. Rigothier, M. C., H. Khun, P. Tavares, A. Cardona, M. Huerre, and N. Guillen. 2002. The fate of Entamoeba histolytica during establishment of amoebic liver abscess analyzed by quantitative radio-imaging and histology. Infect. Immun. 70:3208-3215.
24. Tannich, E., F. Ebert, and R. D. Horstmann. 1991. Primary structure of the 170-kDa surface lectin of pathogenic Entamoeba histolytica. Proc. Natl. Acad. Sci. USA 88:1849-1853.
25. Tavares, P., P. Sansonetti, and N. Guillen. 2000. Cell adhesion and polarization in a pathogenic protozoan: role of the Entamoeba histolytica Gal/GalNAc lectin. Microbes Infect. 2:643-649.
26. Tavares, P., P. Sansonetti, and N. Guillen. 2000. The interplay between receptor capping and cytoskeleton remodelling in Entamoeba histolytica. Arch. Med. Res. 31:S140-S142.
27. Tse, S.-K., and K. Chadee. 1991. The interaction between intestinal mucus glycoproteins and enteric infections, Parasitol. Today 7:163-172.
28. Tsutsumi, V., R. Mena-Lopez, F. Anaya-Velazquez, and A. Martinez-Palomo. 1984. Cellular bases of experimental amebic liver abscess formation. Am. J. Parasitol. 117:81-91.
29. Vines, R. R., G. Ramakrishnan, J. B. Rogers, L. A. Lockhart, B. J. Mann, and W. A. Petri. 1998. Regulation of adherence and virulence by the Entamoeba histolytica lectin cytoplasmic domain, which contains a beta 2 integrin motif. Mol. Biol. Cell 9:2069-2079.
30. Wehrle-Haller, B., and B. Imhof. 2002. The inner lives of focal adhesions. Trends Cell Biol. 12:382-389.(Paulo Tavares, Marie-Chri)
Unite de Biologie Cellulaire du Parasitisme, INSERM U389
Unite d'Histotechnologie et Pathologie
Plateforme d'Imagerie Dynamique, Institut Pasteur, Paris
Laboratoire de Biologie et Contrle des Organismes Parasites, UPRES 398-IFR 75, Faculte de Pharmacie, Universite Paris-Sud, Chatênay-Malabry, France
ABSTRACT
The protozoan parasite Entamoeba histolytica colonizes the human large bowel. Invasion of the intestinal epithelium causes amoebic colitis and opens the route for amoebic liver abscesses. The parasite relies on its dynamic actomyosin cytoskeleton and on surface adhesion molecules for dissemination in the human tissues. Here we show that the galactose/N-acetylgalactosamine (Gal/GalNAc) lectin clusters in focal structures localized in the region of E. histolytica that contacts monolayers of enterocytes. Disruption of myosin II activity impairs the formation of these structures and renders the trophozoites avirulent for liver abscess development. Production of the cytoplasmic domain of the E. histolytica Gal/GalNAc lectin in engineered trophozoites causes reduced adhesion to enterocytes. Intraportal delivery of these parasites to the liver leads to the formation of a large number of small abscesses with disorganized morphology that are localized in the vicinity of blood vessels. The data support a model for invasion in which parasite motility is essential for establishment of infectious foci, while the adhesion to host cells modulates the distribution of trophozoites in the liver and their capacity to migrate in the hepatic tissue.
INTRODUCTION
Cells communicate with their environment through a large variety of cell surface molecules. Contact of cell surface adhesion molecules with their external ligands triggers signaling pathways that lead to cytoskeleton rearrangements necessary both for cell adhesion and for cell displacement. The adhesion sites of higher eukaryotes are complex, dynamic, molecular platforms that link the substratum in the cell outside to the internal cytoskeleton (9, 30). The actomyosin network provides the central cytoplasmic anchor of focal adhesion areas (30). Deregulation of these environment-cell interactions in multicellular organisms accounts for disease processes such as, for example, tumor initiation, invasion, and metastasis of cancer cells in human malignancies (13). Unicellular eukaryotes also rely on cell adhesion and cytoskeleton functions for survival. In particular, parasites encounter different environments when they colonize and invade a multicellular organism. Their motilities, cell adhesion properties, and aggressive behaviors vary, accounting for different physiopathologies that depend on the localization in the host.
Entamoeba histolytica, the etiological agent of human amoebiasis, colonizes the large bowel. Parasite invasion of the intestinal epithelium is characterized by extensive degradation of the extracellular matrix by amoebic secreted proteases, human cell cytolysis, and phagocytosis. Highly motile trophozoites gain access to the bloodstream in the region of ulcer formation and eventually disseminate to other organs, causing most commonly amoebic liver abscesses (ALA) (22). A central contributor in E. histolytica pathogenesis is the highly effective actomyosin cytoskeleton, which allows fast morphological changes associated with amoebic motility and spatial reorganization of cellular components (6). Production of the light meromyosin (LMM) moiety of the myosin II heavy chain in genetically engineered trophozoites (strain LMM) was shown to strongly reduce cell motility in vitro and in vivo, to affect the retrograde motion of amoebic surface molecules, and to abolish cytotoxicity when organisms were incubated in the presence of epithelial cells (2, 4).
The coordinated action of the E. histolytica cytoskeleton and surface adhesion molecules is essential for pathogenesis (25). The major adhesion molecule of the parasite is the immunodominant galactose/N-acetylgalactosamine (Gal/GalNAc) inhibitable lectin (17, 21). The Gal/GalNAc lectin plays a central role in adhesion of the trophozoite to intestinal mucins and to enterocytes (27) and has been shown to be a major virulence factor (21). This lectin is a 260-kDa heterodimeric complex composed of a light subunit (Lgl) and a heavy subunit (Hgl) bound by disulfide bonds (20). The two known isoforms of Lgl (31 and 35 kDa) are extracellular and undergo different posttranslational modifications; the 31-kDa subunit has a carboxyl-terminal acyl-glycosylphosphatidylinositol anchor that probably attaches it to the plasma membrane, while the 35-kDa subunit does not have the glycolipid but undergoes more extensive glycosylation (17, 18). Studies with genetically engineered trophozoites showed that Lgl is required for E. histolytica virulence and plays a role in parasite adhesion and killing of host cells (1, 10). The various Hgl isoforms are integral membrane proteins with a short cytoplasmic domain (Fig. 1) (15, 24). Production of the lectin heavy-chain cytoplasmic domain from either the Hgl2 (strain HGL-2) or Hgl3 (strain HGL-3) isoform that is anchored to the membrane through the transmembrane segment has a dominant negative effect and reduces significantly the adhesion of amoebae to epithelial cells (29; this study). Furthermore, strain HGL-2 has normal motility in vitro but is immobile when it is injected into the liver (4).
Here we studied the interaction of strains LMM and HGL-2 with monolayers of enterocytes and the capacity of these engineered parasites to produce liver abscesses in the hamster animal model after injection by the intraportal route. Our results demonstrated the central roles of myosin II and the Gal/GalNAc lectin in the invasive process of E. histolytica and also revealed different cellular and physiopathological behavior when the function of these molecules was disrupted. Whereas the strain deficient in myosin II activity was avirulent, the strain defective in Gal/GalNAc function disseminated through the blood and massively invaded the hepatic tissue, resulting in a metastatic-like spread of infection (13).
MATERIALS AND METHODS
Bacteria, parasites, cell lines, and cell biology procedures. Escherichia coli strains DH5 and BL21DE3(pLysS) were used for cloning procedures and overproduction of the His6-Hgl3TM-COO– fusion peptide, respectively. E. histolytica strain HM1:IMSS was cultivated axenically in TYI-S-33 (5). Virulent HM1:IMSS trophozoites (referred to as strain HM1 or the wild-type strain below) derived from a single culture that was axenized after passage through hamster liver (23) were transfected by electroporation with the control vector pExEhNeo (7), and its derived constructs were used for production of LMM (2) and Hgl2SP-TM-COO– (Fig. 1) (4). The engineered strains designated LMM and HGL-2 were maintained in the presence of 10 μg of G418 (Gibco-BRL) per ml and were frozen at –80°C for long-term storage. Before each experiment the concentration of drug was increased to 30 μg/ml for 48 h. Production of LMM and of the Hgl carboxyl-terminal region was checked by Western blotting of trophozoite extracts after transfection and regularly thereafter. New transfections were carried out before each series of hamster infestations.
Adhesion of trophozoites to monolayers of polarized Caco-2 cells grown to confluence for 14 days and fixed with 5% formaldehyde was quantified as described previously (19). Adhesion tests were carried out in the presence of 200 mM glucose or 200 mM galactose.
Production and purification of the cytoplasmic tail of the Gal/GalNAc lectin in E. coli. The DNA fragment coding for the putative transmembrane segment and cytoplasmic tail of Hgl3 (coordinates 3736 to 3940 in GenBank accession no. L14815) (15) was amplified by PCR by using oligonucleotides 5'CGCGGATCCGTTGGAGCTATTGCAGCGG (coding strand) and 5'CGGGATCCCTGCAGCTTATCCATTGAATGTTGCTGC (noncoding strand) (the sequence homologous to hgl3 is indicated by boldface type; the BamHI and PstI sites are underlined). The purified PCR fragment (approximately 310 bp) was cleaved with BamHI/PstI and ligated to vector pRSET A (Invitrogen) digested with the same endonucleases. The correctness of the construction was checked by restriction analysis, and the recombinant plasmid was inserted by transformation into E. coli BL21DE3(pLysS) for overproduction of the His6-Hg13TM-COO– fusion peptide.
A 1-liter culture of the strain overproducing His6-Hgl3TM-COO– was grown at 37°C to a density of approximately 108 CFU/ml and induced with 5 mM isopropyl--D-thiogalactopyranoside (IPTG). After 2 h the cells were sedimented, washed with cold medium, and resuspended in 20 ml of lysis buffer (20 mM HEPES-KOH [pH 7.6], 500 mM KCl, 1.5% [wt/vol] Sarkosyl) freshly supplemented with 0.5 mg of lysozyme per ml, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (Sigma), and 10 mM leupeptin (Sigma). All subsequent manipulations were carried out at 4°C. Cells were lysed by sonication (four 15-s pulses with 15-s intervals between pulses), and cell debris was eliminated by centrifugation at 18,000 x g for 1 h. The supernatant was added to 3 ml of cobalt affinity matrix (Talon; Clontech), which was previously equilibrated with lysis buffer, and mixed for 1 h in a rotating wheel. The resin was allowed to settle by gravity and washed twice with lysis buffer (15 min each), and this was followed by two stringent washes (30 min each) with lysis buffer containing 10 mM imidazole. His6-Hgl3TM-COO– was eluted with 500 mM imidazole in lysis buffer, which yielded a peptide that was more than 90% pure (data not shown). The fusion peptide was used to immunize rabbits for production of polyclonal antibodies. The serum obtained recognized the carboxyl terminus of both Hgl2 and Hgl3.
Parasite immunostaining and confocal microscopy. Trophozoites grown to approximately 80% saturation in 25-cm2 bottles were washed and resuspended in 6 ml of Dulbecco modified Eagle medium (DMEM) prewarmed to 37°C by gentle tapping in the bottle. The trophozoite suspension was divided into aliquots and mixed with DMEM containing glucose (or galactose in control experiments, as specified below) at a final concentration of 200 mM. Trophozoites were counted, and 1 x 105 to 2 amoebae were added to Caco-2 cells grown for 14 days in coverslips previously washed twice with DMEM lacking serum. Cocultures were incubated for 30 min at 37°C under 10% CO2. The medium was carefully decanted, and the coculture was fixed for 30 min at 37°C with 3.7% paraformaldehyde prepared in phosphate-buffered saline. Cells were permeabilized for 3 min with 0.1% Triton X-100 in phosphate-buffered saline. Immunolabeling with the anti-lectin monoclonal antibody CD6 (12) visualized with the secondary anti-monoclonal antibody coupled to Cy3 (Molecular Probes) and staining of filamentous actin with phalloidin-fluorescein isothiocyanate (1:25 dilution; Sigma) were carried out as described previously (2, 16, 26). Fluorescent samples were observed with a confocal laser scanning microscope (16, 26). Optical sections were acquired on serial planes (0.5 μm) starting from top of the epithelial cell monolayer that contacted the amoebae to at least two-thirds of the trophozoite height.
Animal model experiments. Intraportal infestation of male Syrian golden hamsters (Mesocricetus auratus) with wild-type or engineered strains of E. histolytica, animal dissection, histological analysis, and quantitative analysis of inflammatory foci were carried out as described previously (23). The significance of statistical differences was assessed by using the method of Fisher. The production of LMM and of the Hgl C-terminal region in trophozoites recovered from liver abscesses was checked by Western blotting by using a monoclonal antibody against the vesicular stomatitis virus epitope (2) and the rabbit polyclonal serum against His6-Hgl3TM-COO– (this study).
RESULTS
Characterization of engineered trophozoites deficient in motility and adhesion. In order to investigate the impact of trophozoite motility and adhesion in the biology and pathogenesis of E. histolytica, we characterized the behavior of parasite strains LMM (2) and HGL-2 (4).
Production of the carboxyl-terminal domain of myosin II (LMM fragment), which is involved in association of this mechanochemical enzyme, has a dominant negative effect on the activity of the cell actomyosin cytoskeleton (2, 3). E. histolytica trophozoites transfected with a construct express the LMM coding region so that the ratio of LMM protein fragment to myosin II is 4:1 (strain LMM) (2). The amoebae have low motility, are not cytotoxic for Caco-2 cells, and exhibit a reduction in the multiplication rate compared to trophozoites transfected with vector pExEhNeo (mock strain) when they are grown at high drug concentrations (>20 μg of G418 per ml) (2). The adhesion of the LMM strain to monolayers of Caco-2 cells is 53.9% ± 8.7% (n = 3) compared to the adhesion of the mock strain.
Production of the cytoplasmic tail from the Hgl3 isoform of the E. histolytica Gal/GalNAc lectin (Fig. 1) inhibits parasite adhesion to epithelial cells (29). We constructed plasmid pEhC-Hgl2, in which the extracellular region of hgl2 (24) was replaced by a FLAG epitope directing production of a 21-kDa peptide that contained the signal peptide, the transmembrane segment, and the cytoplasmic tail of Hgl2 (Hgl2SP-TM-COO–) (Fig. 1a). The peptide was detected in trophozoites transfected with pEhC-Hgl2 by using anti-FLAG and by using serum against His6-Hgl3TM-COO– (data not shown). The cytoplasmic tail of Hgl2 differed from Hgl3 at two positions; at one of these positions Ile1270 in Hgl3 was replaced by a threonine in Hgl2 in the sequence Thr-Ile-Thr that Vines et al. (29) showed to be important for trophozoite adhesion to epithelial cells (Fig. 1b). However, trophozoites transfected with pEhC-Hgl2 (strain HGL-2) also exhibited a reduction in adhesion to Caco-2 monolayers (46.8% ± 16.6% adhesion; n = 3), demonstrating that the cytoplasmic tail of Hgl2 cross talks with intracellular factors involved in parasite adhesion. The HGL-2 strain remained cytotoxic, as evaluated by inspection of cocultures with Caco-2 monolayers, and its growth was normal compared to the growth of the mock strain (data not shown).
Inhibition of adhesion in engineered trophozoites correlates with a reduced surface of contact with epithelial cell monolayers. The adhesion of E. histolytica to epithelial cells is strongly inhibited by millimolar amounts of galactose or N-acetylgalactosamine (14), a feature explained to a significant extent by the Gal/GalNAc lectin adhesion activity (references 17, 21, and 25 and references therein). Immunofluorescence labeling of the trophozoites for the Gal/GalNAc lectin in cocultures with Caco-2 cell monolayers showed that in the presence of galactose the lectin was distributed homogeneously around the surface of the few amoebae that remained in contact with the monolayer (Fig. 2a). When adhesion was not inhibited (in the presence of glucose instead of galactose), the trophozoites exhibited an extended surface of interaction with the monolayer, where focal clusters of surface Gal/GalNAc lectin made intimate contact with the epithelial cells (Fig. 2b) (8). The interaction sites likely represented focal adhesion points assembled at the trophozoite surface that were functional analogues of structures found in adhesive animal cells (9, 30). The human cells in the monolayer, in turn, showed a significant increase in subcortical actin filaments underneath the region of contact with the parasite (data not shown) (3).
The specific contributions of the actomyosin cytoskeleton and of the Gal/GalNAc lectin to the formation of the trophozoite adhesion structures were investigated by confocal microscopy of Caco-2 monolayers cocultured with trophozoite strains transfected with the vector pExEhNeo (control), LMM, and HGL-2 (Fig. 2c to e). Control trophozoites bound to the enterocytes through a large surface with numerous focal adhesion points (Fig. 2c), as observed for wild-type parasites (Fig. 2b). LMM and HGL-2 parasites were not firmly attached to the apex of enterocytes. Approximately 55% of these parasites were washed away from the monolayer as a consequence of their low-adhesion phenotype (see above). The LMM and HGL-2 trophozoites that remained in contact with the surface of enterocytes might have had weak phenotypes because E. histolytica transfectants are not homogeneous populations due to variability in the copy number of the transfected plasmid. This heterogeneity was best appraised by other immunofluorescence studies of a population of transfected trophozoites that were shown to produce different levels of the protein C-PAK encoded by an episomal vector (11; E. Labruyere, unpublished observations). Confocal analysis of the LMM trophozoites bound to enterocytes revealed their rounded shape and a contact surface with the monolayer that was smaller than that of wild-type amoebae (Fig. 2d). These features correlated with enrichment of myosin II in the parasite cytoskeleton (data not shown). The results suggest that disruption of the mechanoenzyme function might lead to rigidification of the actomyosin cytoskeleton, accounting for the altered shape of the trophozoite. The HGL-2 parasites remaining in contact with the enterocytes had an intermediate phenotype, in which the contact surface between the trophozoites and the monolayer appeared to be smaller than that observed for wild-type trophozoites and some Gal/GalNAc lectin clusters in the region of trophozoite-enterocyte contact were still present. The phenotypes described for the different strains were confirmed by inspection of confocal images in a blind test. The results revealed the importance of myosin II and of the Gal/GalNAc lectin for binding to host cells and that disruption of the activity of these molecules has a distinct effect on the interaction.
Nonmotile E. histolytica is avirulent, while the adhesion-deficient strain exhibits exacerbated virulence and a distinct histopathology in hamster liver abscesses. The distinct behavior of the LMM and HGL-2 strains in cocultures with enterocytes prompted us to correlate the cell biology results with the pathogenicity of the strains in vivo, as assayed in the hamster liver abscess model (28). Young adult hamsters were inoculated by the intraportal route with wild-type (nontransfected strain HM1), control, LMM, and HGL-2 trophozoites. Macroscopic abscesses, increases of 30 to 40% in weight due to edema, and extended inflammation were observed in livers of animals sacrificed 5 days after infestation except in the case of hamsters infected with the LMM strain (Fig. 3). In the latter case the appearance of the liver remained normal, and histology studies showed that the hepatic tissue was intact (Fig. 3).
The development of ALA was monitored by histology and immunolabeling of parasites. Serial sections of the infected livers were prepared from animals sacrificed at times ranging from 6 h to 5 days postinfection. The number of inflammatory foci containing parasites was determined at 6 h postinfection in five different sections from lobes 1 and 2 of each liver (lobe numbering is shown in Fig. 3 for the liver infected with the LMM strain). This stage of ALA was characterized by the presence of a large number of inflammatory foci (70%), the majority of which lacked trophozoites (23). The number of foci increased, and the size increased progressively. The foci were characterized by the presence of numerous polymorphonuclear cells (PMN), most of which were necrotic. Similar results were observed after infestation of animals with the wild-type and control strains (Fig. 4). LMM infection yielded a significantly lower number of foci, and amoebae were found only in very rare cases (Fig. 4). The small foci present at 6 h postinfestation did not lead to detectable lesions after 5 days, showing that the amoebic infection was eradicated (Fig. 3). In contrast, infection with HGL-2 revealed a high capacity for successful liver invasion with a unique histopathology. A high number of small inflammatory foci were detected at 6 h postinfection (Fig. 4). The majority of the foci contained parasites at this stage of infection, a situation opposite the situation found in animals challenged with wild-type amoebae (Fig. 4). Thus, HGL-2 had an increased capacity to survive the host immune response that was most likely responsible for the high number of abscesses disseminated throughout the liver at 5 days postinfection (Fig. 3). Parasites were distributed with a relatively disorganized pattern in the interior of the abscess necrotic area and close to its periphery. This observation was also distinct from the observation made for normal abscesses of E. histolytica, in which a peripheral layer of trophozoites was surrounded by inflammatory cells, mostly PMN, delimiting the border between the area of necrosis and the hepatic parenchyma (Fig. 3) (23, 28).
In order to understand the different histopathologies of liver abscesses caused by the wild-type and HGL-2 strains, we examined the formation of liver abscesses at early stages postinfection (6 and 24 h). The adhesion-defective strain caused more inflammatory foci, but the progression was slower than that with wild-type amoebae, as shown by the significantly smaller diameter after 24 h of infection (Fig. 5a and b). Interestingly, more than 30% of the HGL-2 foci were contiguous to blood vessels (Fig. 5a and c), and the sinusoids exhibited a layer of host PMN resulting from inflammation (Fig. 5a, panel 4).
DISCUSSION
Motility and the capacity to adhere to host cells are essential for pathogenesis of the human parasite E. histolytica. However, the cell biology of these processes remains poorly understood. In this study we observed that the main adhesion molecule of E. histolytica, the Gal/GalNAc lectin, forms clusters in the region of interaction with enterocytes, as previously described (8). The clusters make intimate contact with the enterocyte surface, as demonstrated by the overlapping labeling of the parasite Gal/GalNAc and the enterocyte subcortical actin filaments (Fig. 2). We propose that the numerous focal points present at the parasite-host cell interface are functional analogues of adhesion structures found in multicellular organisms. These structures are dynamic macromolecular platforms that anchor transmembrane adhesion proteins to the intracellular cell cytoskeleton (9, 30). Studies with transfected trophozoites suggest that there is similar cross talk in E. histolytica. Disruption of the Gal/GalNAc function leads to a weaker interaction of trophozoites with the enterocyte monolayer. A stronger effect leading to reduced contact between the trophozoite and the monolayer was found when the activity of the cytoskeletal component myosin II was disrupted. Inspection of confocal images showed that these phenotypes correlate with a decrease in the number of focal adhesion points (Fig. 2b to e; data not shown). A quantitative assessment of this correlation requires a methodology that we are currently developing to determine the three-dimensional structure of individual trophozoites based on the confocal microscopy data. The present work revealed that molecules of E. histolytica involved in adhesion and in actomyosin cytoskeleton activity participate in the assembly of structures for the parasite-host interaction. Identification of the other protein components involved and of the signaling pathways that control the dynamics of these structures should be a major asset for characterizing the molecular mechanisms that support E. histolytica binding and motility in monolayers of enterocytes.
Other studies and our studies showed that disruption of myosin II (strain LMM) and Gal/GalNAc (strains HGL-2 and HGL-3) activities have a major effect on trophozoite adhesion and motility (2, 4, 21, 29; this study). Interestingly, the phenotypes observed in vitro may not be directly applied to the complex situation of parasite invasion in vivo. This is exemplified by strain HGL-2, which has normal motility in monolayers of enterocytes but is immobile when it is injected into the liver tissue (4). Thus, we studied the physiopathology of ALA development to correlate our cell biology studies on parasite adhesion with the invasion of the liver parenchyma by the adhesion-deficient strains LMM and HGL-2. Administration of parasites by the portal vein route implies that the trophozoites are transported by the bloodstream to the liver network of irrigation vessels. Successful invasion of the liver parenchyma requires (i) that the trophozoite binds strongly enough to the endothelium to counteract the pressure imposed by the bloodstream, (ii) extravasation, and (iii) migration through the three-dimensional structure of the hepatic tissue. The trophozoite has to survive the host inflammatory response at each of these steps.
Intraportal injection of wild-type and control strains caused ALA progression with the histopathology described previously (references 23 and 28 and references therein). Examination of liver sections from animals infected with strain LMM showed the presence of very few trophozoites, and no inflammation was observed at 24 h. The normal tissue architecture was recovered at 5 days postinfection (Fig. 3). The LMM strain thus displays a low efficiency of tissue penetration and is unable to survive in the liver. This phenotype correlates with the inability of the trophozoites to penetrate Caco-2 cell monolayers and with a lack of cytotoxicity in vitro (2). Both of these features probably contribute to preventing migration of the LMM parasites in the liver tissue and to facilitating killing of the parasites by the host immune response. The properties of strain LMM indicate that it is a good candidate for vaccination studies to elicit a host immune response against amoebiasis.
HGL-2 trophozoites injected by the portal route infect all lobes of the hamster liver, causing a large number of infection foci. The foci differ from those found in wild-type infections by their smaller size, by their disorganized structure, and by their preferential localization close to blood vessels with normal endothelial cells (Fig. 3 and 5). The destruction of vessels is an early event in ALA caused by wild-type E. histolytica. The only difference detected in vitro between the two strains was the reduced adhesion of HGL-2 to monolayers of enterocytes, while the cytotoxicities and motilities were identical. However, HGL-2 trophozoites injected intrahepatically remained immobile, showing that the Gal/GalNAc lectin is essential for migration of E . histolytica in the three-dimensional structure of the liver parenchyma (4). This is most likely due to the role of this lectin in adhesion to host cells. The HGL-2 strain might also have a defect in the contact-dependent transfer of the Gal/GalNAc lectin to the host cell (12), a process that was proposed to contribute to invasion of the host tissue (12, 25). The data suggest that during infection the HGL-2 trophozoites brought by the portal vein are spread throughout the liver by the blood vessel network. The reduced adhesion of the trophozoites might limit binding to the vessel wall and extravasation to the hepatic tissue, therefore facilitating dissemination in the organ through the bloodstream. When trophozoites penetrate the liver parenchyma, their very low mobility leads to the establishment of infection foci that are concentrated in regions adjacent to the blood vessels (Fig. 5). These foci progress into small abscesses, in which the amoebae are distributed both at the periphery, as in wild-type ALA, and in the necrotic region. In spite of the difficulty that they have penetrating the liver tissue, the HGL-2 trophozoites resist very well the host immune response. How this feature relates to their low-adhesion phenotype remains to be established. The histopathology associated with HGL-2 infection reveals the complex effect of reduced trophozoite adhesion in host-pathogen interactions resulting in enhanced spread of the parasite and in exacerbated virulence.
ACKNOWLEDGMENTS
We acknowledge E. Labruyere for communication of unpublished results. We thank E. Tannich for kindly providing the plasmid carrying cloned hgl2. A. Leroy and G. Bailey are acknowledged for providing monoclonal antibody CD6. We are grateful to P. Sansonetti for his continuous support and interest in this project.
This work was supported in part by grants from the French Ministry of National Education through the PRFMMIP program. P.T. was supported by EMBO and FCT postdoctoral fellowships.
REFERENCES
1. Ankri, S., F. Padilla-Vaca, T. Stolarsky, L. Koole, U. Katz, and D. Mirelman. 1999. Antisense inhibition of expression of the light subunit (35 kDa) of the Gal/GalNac lectin complex inhibits Entamoeba histolytica virulence. Mol. Microbiol. 33:327-337.
2. Arhets, P., Olivo J.-C., P. Gounon, P. Sansonetti, and N. Guillen. 1998. Virulence and functions of myosin II are inhibited by overexpression of light meromyosin in Entamoeba histolytica. Mol. Biol. Cell 8:1537-1547.
3. Burns, C. G., M. Reedy, J. Heuser, and A. De Lozanne. 1995. Expression of light meromyosin in Dictyostelium blocks normal myosin II function. J. Cell Biol. 130:605-612.
4. Coudrier, E., F. Amblard, C. Zimmer, P. Roux, J.-C. Olivo, M.-C. Rigothier, and N. Guillen. 2005. Myosin II and the Gal-GalNAc lectin play a crucial role in tissue invasion by Entamoeba histolytica. Cell. Microbiol. 7:19-27.
5. Diamond, L. S., D. R. Harlow, and C. C. Cunnick. 1978. A new medium for the axenic cultivation of Entamoeba histolytica and other Entamoeba. Trans. R. Soc. Trop. Med. Hyg. 72:431-432.
6. Guillen, N. 1996. Role of signalling and cytoskeletal rearrangements in the pathogenesis of Entamoeba histolytica. Trends Microbiol. 4:191-196.
7. Hamann, L., I. R. Nicke, and E. Tannich. 1995. Transfection and continuous expression of heterologous genes in the protozoan parasite Entamoeba histolytica. Proc. Natl. Acad. Sci. USA 92:8975-8979.
8. Hughes, M. A., C. W. Lee, C. F. Holm, S. Ghosh, A. Mills, L. A. Lockhart, S. L. Reed, and B. J. Mann. 2003. Identification of Entamoeba histolytica thiol-specific antioxidant as a GalNAc lectin-associated protein. Mol. Biochem. Parasitol. 127:113-120.
9. Jockusch, B. M., P. Bubeck, K. Giehl, M. Kroemker, J. Moschner, M. Rothkegel, M. Rudiger, K. Schluter, G. Stanke, and J. Winkler. 1995. The molecular architecture of focal adhesions. Annu. Rev. Cell. Dev. Biol. 11:379-416.
10. Katz, U., S. Ankri, T. Stolarsky, Y. Nuchamowitz, and D. Mirelman. 2002. Entamoeba histolytica expressing a dominant negative N-truncated light subunit of its gal-lectin are less virulent. Mol. Biol. Cell 13:4256-4265.
11. Labruyere, E., C. Zimmer, V. Galy, J.-C. Olivo-Marin, and N. Guillen. 2003. EhPAK, a member of the p21-activated kinase family, is involved in the control of Entamoeba histolytica migration and phagocytosis. J. Cell Sci. 116:61-71.
12. Leroy, A., G. de Bruyne, M. Mareel, C. Nokkaew, G. B. Bailey, and H. Nelis. 1995. Contact-dependent transfer of the galactose-specific lectin of Entamoeba histolytica to the lateral surface of enterocytes in culture. Infect. Immun. 63:4253-4260.
13. Leroy, A., M. Mareel, G. De Bruyne, G. Bailey, and H. Nelis. 1995. Metastasis of Entamoeba histolytica compared to colon cancer: one more step in invasion. Invasion Metastasis 14:177-191.
14. Li, E., A. Becker, and S. L. Stanley, Jr. 1988. Use of Chinese hamster ovary cells with altered glycosylation patterns to define the carbohydrate specificity of Entamoeba histolytica adhesion. J. Exp. Med. 167:1725-1730.
15. Mann, B. J., B. E. Torian, T. S. Vedvick, and W. A. Petri, Jr. 1991. Sequence of a cysteine-rich galactose-specific lectin of Entamoeba histolytica. Proc. Natl. Acad. Sci. USA 88:3248-3252.
16. Marion, S., P. Tavares, P. Arhets, and N. Guillen. 2004. Signal transduction through the Gal-GalNAc lectin of Entamoeba histolytica involves a spectrin-like protein. Mol. Biochem. Parasitol. 135:31-38.
17. McCoy, J. J., B. Mann, and W. A. Petri, Jr. 1994. Adherence and cytotoxicity of Entamoeba histolytica or how lectins let parasites stick around. Infect. Immun. 62:3045-3050.
18. McCoy, J. J., B. Mann, T. S. Vedvick., Y. Pak, D. B. Heimark, and W. A. Petri, Jr. 1993. Structural analysis of the light subunit of the Entamoeba histolytica galactose-specific adherence lectin. J. Biol. Chem. 268:24223-24231.
19. Padilla-Vaca, F., S. Ankri, R. Bracha, L. A. Koole, and D. Mirelman. 1999. Down regulation of Entamoeba histolytica virulence by monoaxenic cultivation with Escherichia coli O55 is related to a decrease in expression of the light (35-kilodalton) subunit of the Gal/GalNAc lectin. Infect. Immun. 67:2096-2102.
20. Petri, W. A., Jr., M. D. Chapman, T. Snodgrass, B. J. Mann, J. Broman, and J. I. Ravdin. 1989. Subunit structure of the galactose and N-acetyl-D-galactosamine-inhibitable adherence lectin of Entamoeba histolytica. J. Biol. Chem. 264:3007-3012.
21. Petri, W. A., Jr., R. Haque, and B. J. Mann. 2002. The bittersweet interface of parasite and host: lectin-carbohydrate interactions during human invasion by the parasite Entamoeba histolytica. Annu. Rev. Microbiol. 56:39-64.
22. Ravdin, J. I. (ed.). 1988. Amebiasis. Human infection by Entamoeba histolytica. John Wiley & Sons, New York, N.Y.
23. Rigothier, M. C., H. Khun, P. Tavares, A. Cardona, M. Huerre, and N. Guillen. 2002. The fate of Entamoeba histolytica during establishment of amoebic liver abscess analyzed by quantitative radio-imaging and histology. Infect. Immun. 70:3208-3215.
24. Tannich, E., F. Ebert, and R. D. Horstmann. 1991. Primary structure of the 170-kDa surface lectin of pathogenic Entamoeba histolytica. Proc. Natl. Acad. Sci. USA 88:1849-1853.
25. Tavares, P., P. Sansonetti, and N. Guillen. 2000. Cell adhesion and polarization in a pathogenic protozoan: role of the Entamoeba histolytica Gal/GalNAc lectin. Microbes Infect. 2:643-649.
26. Tavares, P., P. Sansonetti, and N. Guillen. 2000. The interplay between receptor capping and cytoskeleton remodelling in Entamoeba histolytica. Arch. Med. Res. 31:S140-S142.
27. Tse, S.-K., and K. Chadee. 1991. The interaction between intestinal mucus glycoproteins and enteric infections, Parasitol. Today 7:163-172.
28. Tsutsumi, V., R. Mena-Lopez, F. Anaya-Velazquez, and A. Martinez-Palomo. 1984. Cellular bases of experimental amebic liver abscess formation. Am. J. Parasitol. 117:81-91.
29. Vines, R. R., G. Ramakrishnan, J. B. Rogers, L. A. Lockhart, B. J. Mann, and W. A. Petri. 1998. Regulation of adherence and virulence by the Entamoeba histolytica lectin cytoplasmic domain, which contains a beta 2 integrin motif. Mol. Biol. Cell 9:2069-2079.
30. Wehrle-Haller, B., and B. Imhof. 2002. The inner lives of focal adhesions. Trends Cell Biol. 12:382-389.(Paulo Tavares, Marie-Chri)