Fate of Entamoeba histolytica during Establishment of Amoebic Liver Abscess Analyzed by Quantitative Radioimaging and Histology

The protozoan parasite Entamoeba histolytica is the causativeagent of amoebiasis, a human disease characterized by dysenteryand liver abscess. The physiopathology of hepatic lesions canbe satisfactorily reproduced in the hamster animal model bythe administration of trophozoites through the portal vein route.Hamsters were infected with radioactively labeled amoebas foranalysis of liver abscess establishment and progression. Theradioimaging of material from parasite origin and quantificationof the number inflammation foci, with or without amoebas, describedhere provides the first detailed assessment of trophozoite survivaland death during liver infection by E. histolytica. The massivedeath of trophozoites observed in the first hours postinfectioncorrelates with the presence of a majority of inflammatory fociwithout parasites. A critical point for success of infectionis reached after 12 h when the lowest number of trophozoitesis observed. The process then enters a commitment phase duringwhich parasites multiply and the size of the infection fociincreases fast. The liver shows extensive areas of dead hepatocytesthat are surrounded by a peripheral layer of parasites facinginflammatory cells leading to acute inflammation. Our resultsshow that the host response promotes massive parasite deathbut also suggest also that this is a major contributor to theestablishment of inflammation during development of liver abscess.

INTRODUCTION

Top

Introduction

Amoebiasis is a widespread human parasitic disease caused bythe protozoan Entamoeba histolytica. The two major clinicalmanifestations of E. histolytica infection are amoebic colitisand liver amoebic abscess. Yearly, 50 million people developintestinal amoebiasis and it is estimated that 10% of individualswith amebic colitis will develop an amebic liver abscess. Amoebainfection causes 100,000 deaths per year, ranking, second aftermalaria, among the most deadly human parasitic diseases causedby a protozoan (6).

During invasive amoebiasis, highly motile trophozoites invadethe intestinal epithelium, causing extensive tissue damage characterizedby acute inflammation and ulceration with necrosis and hemorrhage.In contrast to intestinal amoebiasis, invasion of the liveris characterized by the presence of nonmotile E. histolyticatrophozoites that cause an acute inflammatory reaction. Well-individualizedinfection foci contain mostly dead hepatocytes, polymorphonuclearleukocytes (PMN), macrophages, and parasites. Differences inpathology between amoebic colitis and amoebic liver abscess(ALA) likely result from a variation of the E. histolytica virulencerepertoire in the two organs and from different responses ofthese organs to amoeba infection (11). Understanding the physiopathologyof amoebiasis thus requires the development of new experimentalstrategies that take into account the multifactorial cross talkbetween the amoeba and the host cells during invasion of a specifichuman organ.

Our current knowledge of ALA development is essentially derivedfrom studies with rodent models. Pioneer studies have shownthat ALA formation after inoculation of E. histolytica intothe portal vein of hamsters evolves through different phases(12). During the early stages, trophozoites brought by the bloodstreampenetrate the liver sinusoids, where they are rapidly surroundedby host defense neutrophils, forming inflammatory foci. Trophozoitesare separated from the hepatocytes by a ring of inflammatorycells. Hepatocytes show degenerative changes and lyse. Duringlater stages, the infection foci acquire a characteristic organizationwith a central necrotic region composed of cellular debris surroundedby two peripheric layers, the first of trophozoites and thesecond of immune cells that separate the infection region fromthe hepatic tissue. Interestingly, direct contact of hepatocyteswith trophozoites is observed only occasionally during infection,indicating that the hepatocytes’ massive death may bepartially caused by diffusible products. It was hypothesizedthat such products originate from lysis of the neutrophils (12).The second rodent model that has been used for ALA developmentstudies is severe combined immunodeficient mice (2), as wellas normal mice or neutrophil-depleted mice (10, 13). The histologicalfeatures of infected livers are similar to those found in thehamster. The death of hepatocytes and of cells of the immunesystem during E. histolytica invasion in these mice resultsnot only from the cytolytic activity of the trophozoites butalso from an apoptotic process (11). The not-yet-identifiedsignaling pathway triggering apoptosis is independent of Fas-Fasligand interactions and of the tumor necrosis factor alpha pathway(9).

We present here the first semiquantitative study of ALA developmentin hamsters. The fate of radioactively labeled trophozoitesinoculated in the portal vein was monitored and quantified byhistological analysis and radioimaging with a Beta-Imager. Thesesensitive procedures allowed us to detect quantitatively theamount of amoebic material in the liver and to determine itsdistribution in tissue sections. We estimated the survival andlysis rate of trophozoites reaching the liver after intraportalinoculation. The results show that a large number of trophozoitesarriving in the liver die and that there is a critical earlytime point at which the success of infection is decided. Thetrophozoites that survive this early stage begin to multiplyand cause significant tissue damage.

MATERIALS AND METHODS

Top

Materials and Methods

E. histolytica cultivation and radioactive labeling. Pathogenic E. histolytica strain HM1:IMSS was cultivated axenicallyin TYI-S-33 medium (4). Highly virulent trophozoites of strainHM1:IMSS (a gift of R. Perez Tamayo via Esther Orozco, CINVESTAV,Mexico City, Mexico) were passed through hamster livers every2 months to maintain virulence.

Trophozoite cultures for intraportal inoculation were grownat 37°C for 48 h in closed bottles filled with medium. At12 h before inoculation the trophozoites were diluted if requiredto ca. 50 to 60% saturation and ³⁵S-protein labeling mix (NEN,Boston, Mass.) was added to 12 µCi/ml of medium. Radioactivelabeling was carried out in complete TYI-S-33 medium becausewe observed that trophozoites were less fit and grew slowerin medium lacking cysteine normally used to increase radioactivityincorporation. Before inoculation, the radioactive medium wasdecanted, and the adherent amoebas were washed twice with prewarmedmedium before being detached by tapping the bottle and counted.

Hepatic inoculation procedure and preparation of infected liver sections. Male Syrian golden hamsters (Mesocricetus auratus), aged 6 weeks,with an average weight of 100 g, were inoculated intraportallywith 4 x 10⁵ HM1:IMSS trophozoites. Treatment of animals andsurgical procedures were done according to the method of Tsutsumiet al. (12). Groups of three animals were sacrificed at 1, 3,6, and 12 h and at 1, 3, 5, and 7 days after inoculation. Controlnoninfested animals were sacrificed after 7 days. The liverswere removed, inspected for the presence of amebic abscesses,and photographed. The infected livers were then treated forhistological analysis and radioimaging analysis.

Quantitative determination of the radioactivity in liver sections. The quantitative determination of the radioactivity in liversections was carried out by using both a Micro-Imager and aBeta-Imager (Biospace, Paris, France) (3, 5). The radioimagingprocedure converts the detected energy produced by ionized particlesinto light and intensifies the corresponding localized lightspot observed by using a charge-coupled device camera (1). Datafrom digitized autoradiograms were collected for 8 to 14 h fromfive liver sections and three animals at each time point. Theradioactivity was determined as the total counts per section.

Histology and immunohistochemistry. Dissected livers were fixed in 10% buffered neutral formalinand embedded in paraffin. Serial 5-µm sections were stainedwith Harris' hematoxylin and periodic acid-Schiff reagent. Immunohistochemistrywas performed on sections from different livers with a humanpolyclonal antibody to E. histolytica. The sections were deparaffinizedand immersed in 200 µl of citrate buffer. They were thenincubated with the human polyclonal antibody, washed three timesin PBS, incubated with an anti-human immunoglobulin G, and visualizedby the streptavidin-peroxidase method with aminoethyl carbazoleas the chromogen.

Tissue preparation for sectioning and light microautoradiographic visualization. The localization of radioactive trophozoites was investigatedby using the paraffin section autoradiography procedure (8).Serial 5-µm sections obtained from infected livers wereused. The number of sections depended on the average size ofinflammatory foci for each time point (see Results). Paraffinfrom these sections was removed with xylene, and slides wereincubated through a descending and then an ascending ethanolseries (from 100 to 50% [vol/vol] and then the reverse). Driedsections were processed for light microscope autoradiographyby overloading a coverslip coated with a nuclear emulsion (KodakNTB2, diluted 1:1 with distilled water). After the slides wereair dried for 8 h at room temperature, they were exposed for2 to 4 weeks in light-proof boxes at 4°C. Radiosensitivecoverslips were developed for 3 min in Kodak D19 developer andfixed for 5 min in Kodak fixer. Sections were counterstainedwith Harris' hematoxylin, dehydrated, and mounted in Eukittmedium. The tissue and overlying silver grains were viewed andphotographed with a Leica photomicroscope equipped with a bright-fieldoptics (DMRD Leitz; Leica) camera.

RESULTS

Top

Results

Radioimaging examination of ALA development. Intraportal infestation of young adult hamsters with trophozoitesis a well-established method for reproducing the different stagesof ALA formation. Parasites with ³⁵S-labeled proteins were injectedinto the hamster portal vein. Sections of the infected liverwere prepared from animals sacrificed at different times postinfestationranging from 3 h to 7 days. The total radioactivity and itsdistribution in the complete tissue section were examined byusing a Micro-Imager (Fig. 1). Fine detail was obtained by histologicaland immunological staining combined, when required, with detectionof radioactivity emission by silver precipitation (Fig. 1A,panel 3). Radioactive material is strictly found in infectedlivers colocalizing with parasites or with ghosts of dead amoebasin a basically background-free environment. This method allowstracing the fate of the original trophozoites inoculated inthe animal.

View larger version (33K):
[in this window]
[in a new window]
FIG. 1. Radioimaging analysis of liver abscess formation. (A) Micrographs of dissected liver after 24 h of infection examined by Micro-Imager and histology: 1, optical acquisition (x10 magnification); 2, radiological acquisition of the same section, allowing localization of radioactive foci; and 3, radioimmunohistochemical photomicrograph, allowing the identification of E. histolytica (immunodetection), radioactive amoeba proteins (silver precipitation), and host cells (histological staining) (x1,000 magnification). (B) Distribution of ³⁵S radioactivity in liver slices of animals infected with E. histolytica for the periods indicated in the micrographs. The radioactive material concentrates in discrete areas of the liver until 24 h, and then it exhibits a diffuse distribution before the signal becomes very weak after 72 h (results not shown). The bar under the panel indicates the intensity of the radioactivity by a colored scale from dark (no signal) to red (intense signal). (C) Radioactivity in total liver sections at several time points postinfection. Quantification made in the Beta-Imager is an average of the signal obtained during 1 h from five tissue sections from three different animals and for each time point. The surface of liver sections was of 250 mm² ± 50 mm². The bars represent the means plus the standard deviations of 15 determinations for each time point.

Visualization of amoeba radioactive material was also done byusing the Beta-Imager. Analysis of complete tissue sectionsof infected liver showed strong individualized signals at 3h postinfestation, indicating that parasites had reached theliver parenchyma at this early stage of the infection (Fig.1B). As infection proceeds the pattern of focal radioactivitydistribution is maintained for 24 h, but then it gets diffusedthroughout the complete section. The total amount of radioactivityin the tissue sections, however, decreases slowly until 3 dayspostinfestation (Fig. 1C). The data suggest that most of theoriginal injected amoebas die or undergo cell division until24 h postinfestation but that their radioactive proteins orbreakdown products are mostly retained in the organ tissue untilday 3. Radioactive products are then progressively eliminatedfrom the liver by metabolism.

Kinetics of liver infiltrates in the presence of E. histolytica. The indication that a significant number of original trophozoitesreaching the liver do not survive, a feature of ALA not yetinvestigated, led us to analyze quantitatively their fate. Thesequence of events occurring at the tissue level during ALAformation was monitored by histology (Fig. 2). The descriptivehistology of disease development reproduces particularly wellthe seminal work of Tsutsumi et al. (12). Early stages of infectionare characterized by acute cellular infiltration leading toirregular liver lesions ranging from 70 to 150 µm in diameter.Infectious foci present a large number of inflammatory cellssurrounding normally a single parasite. In a significant numberof cases no trophozoite is found in the inflammatory focus.At these early stages, parasites can be also seen in blood vesselssurrounded by PMN, monocytes, and a few lymphocytes. The volumeof infiltrates increases steadily as infection proceeds. Necrosisappears at 24 h and becomes obvious at 48 h. The hepatic lesionsat this stage exhibit the characteristic features of ALA: thecenters of the abscesses are filled with amorphous necroticdebris and edema, while viable amoebas form a peripheral layerthat is surrounded by inflammatory cells delimiting the borderbetween the area of necrosis and hepatic parenchyma (Fig. 2,days 3 to 7).

View larger version (158K):
[in this window]
[in a new window]
FIG. 2. Histology of ALA progression. Sections of tissue liver from hamsters infected with E. histolytica for the periods indicated in the panels were stained with hematoxylin to visualize the hamster cells (in blue) and immunolabeled to visualize the trophozoites (in red). Note the presence of infiltrates containing PMN present already during the early stages of the infection. Several of these inflammatory foci appear without parasites. Infiltration is massive after 1 day postinfection, and necrosis is very well established at 2 days. Coalescent foci generate abscesses that show a central region of necrosis delimited by a ring of trophozoites surrounded by cells from the immune response, mainly PMN and macrophages. Bar, 20 µm.

The highest number of inflammatory foci is found in the hamsterliver during the early stages after intraportal inoculationof trophozoites (3 to 6 h), and this number decreases as ALAdevelops. A majority of the foci do not contain visible parasites(Fig. 3A). However, detection of radioactive material by silverprecipitation in the histology sections demonstrated the presenceof significant levels of radioactive protein of parasite originin these foci (Fig. 4 and data not shown), originating mostlikely from dead trophozoites. We attribute the massive amoebakilling to the acute inflammatory response mediated by hostimmune system cells.

View larger version (18K):
[in this window]
[in a new window]
FIG. 3. Quantitative analysis of inflammatory foci during development of ALA. (A) The total number of inflammatory foci (open bars) and foci with trophozoites (patterned bars) were counted in total liver sections processed as as described in Fig. 2. Values correspond to the average from counts of five tissue sections from three different animals for each time point indicated. The standard deviations are presented as a thin line at the top of each bar. No inflammatory foci were found in the liver sections of control animals that were not infected. (B) Serial sectioning experiment to quantify the total number of trophozoites per inflammatory foci. The 5-µm tissue sections stained for histology observation are presented as shaded bars, and the section number is on the right side. The average dimensions of inflammatory foci at 3, 6, and 12 h are drawn to scale (see text) as dashed lines whose shape provides an idea of the focus form. One trophozoite is schematized in the interior of each focus to illustrate its dimensions relative to the foci and to the tissue sectioning strategy.

View larger version (137K):
[in this window]
[in a new window]
FIG. 4. Distribution of ³⁵S-labeled material of trophozoite origin during ALA formation. E. histolytica was identified by immunolabeling, inflammatory cells were identified by staining, and radioactivity was visualized by using a radiosensitive emulsion. Strong precipitation of silver colocalizes with trophozoites, showing that radioactive material is brought in by amoebas. At early stages, silver grains are deposited on the surface of endothelial cells (magnification view at 1 h). While infiltration develops, radioactive debris can be found in numerous inflammatory foci associated in a number of cases with discernible ghosts of dead amoebas. The ghosts are characterized by a diffuse antiamoeba antibody staining that overlaps with the silver grain concentration; this can be seen on the micrograph illustrating the 24-h infection silver grain concentration. Bar, 20 µm.

Histological observations suggest that an important number ofinflammatory foci do not contain parasites. A quantitative assessmentof the number of inflammatory foci with or without intact trophozoitesrequires a statistically representative counting and serialsectioning of the foci to check at different planes whetheramoebas are present. A preliminary single-section analysis wasused to determine the average diameter of inflammatory focifor each time point of the infestation kinetics. Visual inspectionconfirmed that inflammatory foci without trophozoites are normallyobserved at 3, 6, and 12 h postinfection (Fig. 2). We then madeserial sections of 5 µm that span >130% of the averageheight of the foci for these three time points and also for24 h. Sections spaced by 20 µm (the average diameter ofa trophozoite is 25 µm) were prepared for histology staining(Fig. 3B). Observation of serial histology sections provideda three-dimensional view of the inflammatory foci at a resolutionat which the presence of trophozoites can be detected accurately.The study was only continued for the time points 3, 6, and 12h because the vast majority of foci contained at least one trophozoiteat 24 h postinfection, as was already suspected from single-sectionobservation (Fig. 2; data not shown). The foci had a regularspherical or oval three-dimensional shape at 3 and 6 h postinfection,whereas they exhibited a more irregular form with lobular protuberancesafter 12 h. The total number of foci without trophozoites (referredto as "negative" here) and with one or more trophozoites ("positive")that were present in each serial section was scored first. Second,foci that were negative in one section were examined by inspectionof serial sections along their complete height. In this way,the percentage of negative foci in one section that containeda parasite at a different plane of the inflammatory focus wasestimated. Starting from a central section in the series (e.g.,section 9 for 3 and 6 h; Fig. 3B) and from negative foci inthat section, we analyzed 30 complete foci for 3 h postinfection:31 for 6 h and 26 for 12 h. Among these 20, 29, and 4% werefound to contain one or more trophozoites for the time points3, 6, and 12 h, respectively. These percentages were used ascorrection factors for the average counts of individual serialsections (Table 1). The results with these corrected data agreewell with those from Fig. 3A and demonstrate the validity ofquantifications based on tissue section inspection.

View this table:
[in this window]
[in a new window]
TABLE 1. Quantitative determination of the number of inflammatory foci without trophozoites, with one trophozoite, or with more than one trophozoite at 3, 6, and 12 h postinfection^a

In the early stages after intraportal inoculation of trophozoites(3 to 6 h), 40 to 60% of the inflammatory foci in the hamsterliver do not contain visible parasites (Fig. 3A and Table 1).The success of liver abscess establishment has a commitmentstep at 12 h postinfestation. A minimum number of foci are reachedat this time point, and only $~$ 10% of these contained live trophozoitesthat certainly account for the success of ALA development (Fig.3A and Table 1). Later, the number of infection foci increases,and the majority of them contain more than one live amoeba,indicating that trophozoite cell division has resumed, and thereforethe size of the lesion increases (Fig. 2, 3, and 4 and datanot shown). Well-organized infectious foci can be counted upto 3 days, after which they start to coalesce until the formationof uniform necrotic areas.

Quantification of infectious foci with ³⁵S-labeled trophozoites. The fate of trophozoites and of their ³⁵S-labeled proteins wasexamined by immunolabeling and silver precipitation. Parasitesrevealed by an antiamoebic antibody were labeled specificallywith radioactivity (Fig. 4). At the early stages, during whichthe amoeba invades the liver, radioactivity is often encounteredon the surface of endothelial cells (Fig. 4, 1 h, left and rightpanels). The radioactive ghosts of dead amoebas and/or debriscan then be found in numerous inflammatory foci. Ghosts arecharacterized by diffuse antibody staining that overlaps withsilver grains (Fig. 4, e.g., 3 and 12 h). Detection of radioactivematerial by silver precipitation in the histology sections demonstratedthe presence of significant levels of radioactive protein ofparasite origin in these foci (Fig. 4 and data not shown), originatingmost likely from dead trophozoites. We attribute the massiveamoeba killing to the acute inflammatory response mediated byhost immune system cells. This observation reconciles the constantlevel of radioactivity detected in complete liver sections (Fig.1C) with the very significant variations observed in inflammatoryfoci and amoeba counts during the first days postinfestation(Fig. 3A and Table 1). At advanced stages of ALA, labeling isfound in the central necrotic area of the lesion (Fig. 4, 24h). Silver spots are also detected in the hepatic tissue closeto regions with amoebas. These are attributed to products leftbehind during amoeba motion or to proteins secreted by the trophozoitesafter their arrival in the liver that may promote tissue destruction.

DISCUSSION

Top

Discussion

Liver invasion by E. histolytica with production of abscessesis the most common extraintestinal manifestation of amoebiasis.Infection with radioactive parasites of immunocompetent hamstersby the intraportal route allowed a description of abscess developmentby radioimaging combined with histological analyses. Our studyshows that E. histolytica infection of the liver has a fasttemporal program in which the parasite crosses the endotheliumof liver sinusoids, disseminates through the organ, and adaptsto the new environment before it starts to divide and establishitself successfully to cause massive organ damage. An importantfinding was that parasites are largely destroyed by the nativeimmune system; we determined that the period between 6 and 12h is critical for amoeba survival and subsequent ALA development.

In the hamster model system the parasites reach the liver inthe bloodstream through the portal vein. During early stagesof ALA development, amoebic radiolabeled material is depositedon the surface of endothelial cells (Fig. 4), suggesting thatthese cells are rapidly attacked by elements of the cytotoxicsystem of parasites. This might be the initial trigger for theinflammatory response. It was found that in humans sufferingfrom amoebiasis the liver endothelial cells secrete proinflammatoryfactors after contact with E. histolytica (14). The proinflammatoryfactors lead to recruitment and activation of immune cells thatare responsible for the killing of parasites at early stagespostinfection. Our preliminary results indicate that endothelialcells of the hamster sinusoids undergo apoptosis in the presenceof trophozoites (unpublished results). The resulting disturbanceof the endothelial barrier likely facilitates the passage oftrophozoites between the blood circulation and the hepatic tissue.

The variation in numbers and redistribution of trophozoitesand host immune cells in the first 24 h of ALA development demonstratea highly dynamic interaction between these different cells.Inflammatory cells are rapidly recruited after the arrival ofparasites to the liver, forming large foci around the trophozoitesand eliminating the majority of them. Consequently, most ofthe foci do not contain a live amoeba after 12 h of infestation(Fig. 3). Among the cells of the immune system, neutrophilsprobably play the central role in the elimination of E. histolyticasince parasite debris are normally surrounded by neutrophilsin the hamster animal model (Fig. 2 and 4), and liver lesionsare more prominent in neutrophil-depleted mice than in normalmice (10, 13).

After the initial killing of trophozoites that leads to a significantreduction of positive foci, there is an increase of at leastsixfold in the number of foci with amoebas between 12 and 24h postinfection. This is due to the multiplication of survivortrophozoites (at 12 h postinfection the few positive foci detectedhave more than one amoeba in numerous cases [Fig. 3 and Table1]) and to their migration in the liver tissue to positionswhere new inflammatory foci are formed. Trophozoites leavingthe foci or localized in their periphery can be easily observedin histology sections of infected livers (Fig. 2). The significantvariations in the numbers of positive and negative foci duringthe first 24 h of ALA development (Fig. 3) reflect a very dynamicmovement of both parasites and cells of the immune system inthe infected hamster liver.

The results presented here show that E. histolytica trophozoitesthat were grown under axenic conditions and injected via theportal route are susceptible to the immunoresponse upon theirarrival into the liver. After 12 h of infection the trophozoitesmultiply and the immune system is not able to oppose the fastdissemination of the parasite. It is possible that the adaptationof trophozoites to the liver environment requires the expressionof a repertoire of genes necessary for survival and for escapefrom the immune system. Evidence suggesting that complementresistance has to be acquired by E. histolytica to survive duringinvasion of the human liver has been reported, for example,by Reed et al. and Walderich et al. (7, 15). There is thereforea commitment phase necessary for ALA development that relieson the presence of well-adapted survivor trophozoites. A mostinteresting question is whether, during amoebiasis, the trophozoitesthat cross the intestinal parenchyma and enter the blood circulation,which is the normal route for reaching the liver, are readyto colonize the liver or whether they have to undergo the sameadaptation discussed above. Regardless of this issue, the hamsterexperimental model of amoebiasis provides a very valuable systemfor identifying the requirements for E. histolytica pathogenesisand for investigating the molecular basis for how the parasiteadapts its genetic program depending on the tissue environment.Understanding this program will pave the way for the identificationof major parasite factors that account for the aggressive behaviorof E. histolytica at different stages of amoebiasis.

ACKNOWLEDGMENTS

We thank R. Perez-Tamayo and E. Orozco for kindly providingus with the indispensable E. histolytica strain for the realizationof these experiments. We also acknowledge I. Fernandez for discussions,M. Mavris for manuscript advice, and P. Sansonetti for continuoussupport and interest in the project.

This work was supported by grants from the French Ministry ofNational Education through the PRFMMIP program. P.T. was supportedby EMBO and FCT postdoctoral fellowships.

FOOTNOTES

* Corresponding author. Mailing address: Unité de Biologie Cellulaire du Parasitisme and Unité de Pathogénie Microbienne Moléculaire, INSERM U389, Institut Pasteur, 28 Rue du Dr. Roux, Paris 75724, Cedex 15, France. Phone: 331-45-68-86-75. Fax: 331-45-68-86-74. E-mail: nguillen{at}pasteur.fr.

Editor: W. A. Petri, Jr.

Present address: Laboratoire de Génétique desVirus, CNRS, 91198 Gif-sur-Yvette Cedex, France.

REFERENCES

Top