Transepithelial Migration of Toxoplasma gondii Is Linked to Parasite Motility and Virulence

Toxoplasma gondii is an important opportunistic pathogen causing disseminated congenital infections in the developing fetus (1), severe neurological complications in immunocompromised individuals (2), and ocular pathology in otherwise healthy individuals (3). Worldwide genotypic analysis of T. gondii isolates has identified a population structure consisting of three widespread clonal lineages, termed type I, II, and III (4). Although type II strains have a higher overall prevalence in human infections, type I strains are over represented in studies of congenital toxoplasmosis (4, 5), and a perfect correlation between acute virulence in mice and type I strains exists (6).

T. gondii is an obligate intracellular parasite of the phylum Apicomplexa. It enters a wide range of host cells by active penetration, a rapid process that is dependent on the actin-myosin cytoskeleton of the parasite, and does not rely on the host cell machinery for uptake (7, 8). Once inside the host cell, the parasite resides in a vacuole that avoids fusion with host cell endocytic and exocytic vesicular trafficking pathways (9, 10). Apicomplexan parasites lack cilia or flagella and their mode of locomotion, termed gliding motility, relies on the actin-myosin motor of the parasite (8). Gliding motility also plays determinant roles in infections by other apicomplexan that are important causative agents of human and animal diseases, e.g., Plasmodium (11), Cryptosporidium (12), and Eimeria (13).

In natural oral infections, Toxoplasma initially crosses the intestinal epithelium, disseminates into the deep tissues, and traverses biological barriers such as the placenta and the blood-brain barrier to reach immunologically privileged sites where it causes the most severe pathology. Histopathological studies are consistent with invasion of a variety of cell types in the intestine, including intra-epithelial leukocytes (14), and with rapid hematogenous spread of parasites (15–17). Once in the circulation, it remains unknown whether Toxoplasma disseminates as free, extracellular parasites or if it resides inside infected leukocytes, which are permissive hosts that can cross many biological barriers.

Previous study of migration by Toxoplasma has relied on gliding assays on coated surfaces in vitro (8) and focused on the role of gliding motility for host cell invasion. The rapid spread of primary infections beyond the gut (15–17) led us to hypothesize that this mode of locomotion plays a role in another crucial step in infection, i.e., the ability to disseminate in the host by crossing nonpermissive biological barriers. Using in vitro and ex vivo models for migration, we now demonstrate that T. gondii has substantial capacity to cross epithelial barriers. Transmigration is correlated with virulence and likely contributes to dissemination during in vivo infections.

Toxoplasma tachyzoites were maintained by serial 2-d passage in human foreskin fibroblast (HFF)^* cell monolayers as described previously (7). Parasite strains used in this study include RH, RH-lacZ (8), PTG, and the green fluorescent protein–expressing lines PTG-GFPS65T and RH-GFPS65T, provided by Dr. K. Kim, Albert Einstein College of Medicine, Bronx, NY (18). The various strains and clinical isolates of genotype I, II, and III (4, 6) used in this study have been deposited at the American Type Culture Collection. Viability of parasites was determined by plaque formation on HFF monolayers (19) and was defined as the ratio of the number of intracellular replicating parasites/number of parasites inoculated. Madin Darby Canine Kidney cells (MDCK, CCL-34) were obtained from the American Type Culture Collection and cultured according to guidelines.

Cytochalasin D was obtained from Calbiochem. Low gelling temperature agarose was obtained from Sigma-Aldrich. Matrigel^® was obtained from Becton Dickinson. For detection of parasites in tissue sections and gliding trails, polyclonal rabbit anti-toxoplasma tachyzoite serum or mAb DG52 directly conjugated to Alexa 488 fluorophore (Molecular Probes Inc.) were used. Goat anti–rabbit antibodies conjugated to Alexa 488 or to Alexa 594 (Molecular Probes Inc.) were used as secondary antibodies.

Freshly harvested parasites were plated onto HFF monolayers to yield an infection density of ∼1 infected HFF cell per 4 fields of vision (×20). After 24 h, the monolayer was washed twice and the media substituted by 0.75% low-melting temperature agarose in DMEM supplemented with 10% fetal calf serum, 2 mM glutamine, and 20 μg/ml gentamicin. Assessment of migration under agarose was performed ∼18–24 h after egress of parasites from the initial host cell to allow migration and invasion into new host cells. The monolayers were examined using a ZEISS Axioscope equipped for phase-contrast and epifluorescence microscopy. Foci were photographed randomly and distances migrated by individual, intracellular replicating parasites were assessed using image analysis software by measurement of the shortest distance between the original focus of lysis (host cell) and the intracellular parasites (Openlab, Improvision Inc.; and Adobe Photoshop version 5.5, Adobe Systems Inc.).

Parasites migrating a distance >110 μm on HFF monolayers were cloned using micromanipulation technique. Glass micropipettes (Stoelting) were stretched to have a distal diameter of 25 μm and an Eppendorf Micromanipulator 1570 system and ZEISS Axioscope equipped for phase contrast and epifluorescence microscopy were used. Individual vacuoles containing intracellular replicating parasites located >110 μm from the original focus of egress were picked and plated onto HFF monolayers in 96-well plates.

MDCK cells were plated onto 24-well transwell filters (3 μm pore; Becton Dickinson) and grown for at least 7 d to form a polarized monolayer (resistance >2,500 Ω/cm² using an Ohmmeter; Millipore). Freshly egressed parasites (0.5–1 × 10⁶) were added to the upper well of the Transwell system and incubated at 37°C at indicated times. Alternatively, a layer of 18–22 μm of Matrigel^® mixed with an equal volume of DMEM supplemented with 10% fetal calf serum, 2 mM glutamine, and 20 μg/ml gentamicin was added on top of transwell filter and allowed to solidify. Transmigration of parasites was quantified by colorimetric detection of β-galactosidase activity immediately after incubation as described previously (8) or by visual counting of parasite vacuoles, expressed as CFU in the underlying HFF monolayer using epifluorescence microscopy (green fluorescent protein [GFP]-transfected parasites) or phase-contrast microscopy. Visual counting of vacuoles with intracellular replicating parasites was performed before lysis of the host cells, i.e., 24 to 48 h after invasion. For inhibition of motility, parasite suspensions were pretreated for 10 min at room temperature with 1 μM cytochalasin D and incubated in the presence of 1 μM cytochalasin D during the assay. For heat inactivation, parasites were incubated 30 min at 50°C before use.

Adult female CD1 mice (Charles River Laboratories) were used to obtain ileum that was extensively washed and placed, lumenal side up, into a modified Ussing 2-chamber system. Both the mucosal and serosal reservoirs of the chamber contained DMEM supplemented with 10% fetal calf serum, 2 mM glutamine, and 20 μg/ml gentamicin (pH 7.4). The buffer was gassed (21% O₂–5% CO₂) continuously during the experiment and kept at 37°C. Parasites were introduced in the upper chamber and incubated at 37°C for 8 h. Samples were fixed in 3.7% formaldehyde in PBS, paraffin-embedded, and sections processed for immunohistochemistry as described below. Penetration of parasites to the different histological layers was assessed using phase-contrast or epifluorescence microscopy by scoring parasites in the epithelial cell layer and subepithelial cells layers including the lamina propria, rich in connective tissue, and the submucosa, where the post-capillary circulation is found. All protocols involving animals were approved by the Animal Studies Committee of Washington University.

Serial sections from mouse ileum were deparaffinized, rehydrated, and exposed to primary antibody (polyclonal rabbit anti-toxoplasma tachyzoite serum). Detection was provided using a Vectastain Elite ABC HRP kit, visualized by diaminobenzidine tetrahydrochloride (Vector Laboratories), and counterstained with hematoxylin. In immunofluorescence (IF) experiments, sections were exposed to a polyclonal rabbit anti-toxoplasma serum or mouse mAb DG52 followed by goat anti–rabbit conjugated to Alexa 594 or goat anti-mouse conjugated to FITC, respectively (Molecular Probes). Fluorescein-conjugated Wheat Germ Agglutinin (Vector Laboratories) was used as a surface cell marker when indicated, and nuclear staining was with 4′,6 diamidino-2-phenylindole (DAPI; Vector Laboratories).

Adult female CD1 mice (Charles River Laboratories) were challenged intra-peritoneally with 10⁴ parasites. Parasite numbers in the spleen were determined as described by Mordue et al. (20). Briefly, mice were killed and the spleen was extracted and homogenized under conditions that did not affect parasite viability. The number of viable parasites in the tissue was determined by plaque formation on HFF monolayers.

Gliding assays were performed as described (8). Briefly, glass chamber slides (Lab-Tek; Nunc) were coated by incubation in 50% fetal bovine serum in PBS (pH 7.4) for 1 h at 37°C. Recently egressed parasites were added to the chamber slides and incubated at 37°C for 30 min before fixation in 3.7% formalin-PBS. Detection of trails was performed using mAb DG52 directly conjugated to Alexa 488 fluorochrome. For time-lapse video microscopy, parasites were allowed to glide on precoated glass-bottom microwells and motion was recorded using a ZEISS Axioscope, an ORCA-ER digital camera (Hamamatsu Photonics), and image capture software Openlab version 3.02 (Improvision Inc.).

Statistical analyses were performed using the Statistical Analysis Systems (SAS, version 8.1; SAS Institute).

To determine the capacity of Toxoplasma to migrate away from a focus of infection, the spread of parasites under soft agarose was measured in vitro. Quantitative differences were observed between the ability of the virulent type I strain RH and the nonvirulent type II strain PTG to migrate in this assay. While PTG parasites always migrated <70 μm, a small but reproducible subpopulation of the RH strain distinguished itself by migrating >50% further than the maximum distance observed for PTG (i.e., >110 μm; Fig. 1, A and B, and Table I). Parasites exhibiting the long distance migration phenotype, termed LDM, were cloned by micromanipulation (Fig. 1, B and B′, see arrows). When tested in the HFF spreading assay, these clones generated exhibited a higher frequency of LDM parasites, reaching levels ≥20% of the parasite population (Fig. 1, C and C′, and Table I). In contrast, type II PTG parasites cloned from the periphery of an infection focus (∼70 μm migration) did not exhibit increased migratory capacity, i.e., no parasites migrated distances >70 μm and no shift in the distribution of migrated distances was observed (data not shown). Collectively, these results show (a) important differences in migration exist between the type I strain RH and the type II strain PTG, (b) the presence of at least two distinct populations with different migratory behavior in the type I strain RH, and (c) migration can be up-regulated in the virulent RH strain.

During in vivo infections, Toxoplasma crosses epithelial barriers by an unknown process. To address the question whether Toxoplasma tachyzoites are capable of actively crossing cellular barriers, migration across polarized MDCK cell monolayers or extracellular matrix in a Transwell system was tested in vitro. Parasites readily crossed the polarized monolayer in a time-dependent fashion with type I strain RH being 10–100-fold more effective than type II strain PTG (Fig. 2, A and C, and Table I). The increased migratory capacity of RH versus PTG was confirmed using extracellular matrix (Matrigel^®), where transit also required active motility on the part of the parasite (Fig. 2, C and D, and Table I). Transmigration across MDCK monolayers was totally inhibited by heat inactivation of parasites (Fig. 2 B) and migration across extracellular matrix was blocked by cytochalasin D treatment (Fig. 2, C and D), indicating it depends on active motility by the parasite.

To examine whether increased migratory capacity of LDM-parasites on HFF monolayers would also confer increased transmigratory capacity, LDM-clones were assessed for their ability to cross-polarized monolayers. For comparison, we analyzed a clone called LDM1 that had an enhanced capacity for long distance migration on HFF monolayers where up to 20% of parasites migrated >110 μm from foci of infection (Fig. 3 A). The LDM1 clone also exhibited an enhanced capacity to traverse MDCK cell monolayers in vitro relative to the type I RH strain from which it was derived (Fig. 3 B, and Table I). As described above, the nonvirulent strain PTG did not display a subpopulation of LDM parasites (Fig. 3 A), and showed significantly lower levels of transmigration (Fig. 3 B). The differences in transmigration were not due to differences in viability or infectivity between the strains, which varied by <20% as shown by plaque formation on HFF monolayers (Fig. 3 C). We conclude from these findings that (a) Toxoplasma is able to actively cross polarized cell monolayers and extracellular matrix, (b) these abilities are greatly enhanced in the virulent type I (RH) strain, and (c) transmigratory capacity of type I strains can be up-regulated.

To investigate whether the difference in migration observed between strains was due to differences in motility, the strains were compared using well established gliding assays (8). Significant quantitative differences in gliding distance were observed between the RH strain and the PTG strain (Fig. 3 D, and Table I). Furthermore, the frequency of gliding parasites was significantly higher for RH (1.3 trails/adhered parasite, SD = 0.20) than PTG (0.047 trails/adhered parasite, SD = 0.012). Time-lapse video microscopy of the PTG strain confirmed that the three types of movement previously described for gliding motility in RH, i.e. upright twirling, circular and helical gliding, were present (data not shown). Thus, the lower overall motility of PTG is not due to qualitative differences in motility. Importantly, the LDM1 clone was only slightly more motile in this in vitro assay when compared with the RH strain (Fig. 3 D). Thus, while the enhanced migration ability of type I strains may be due to greater motility, this cannot account for the increase in transmigration seen in LDM clones.

During in vivo infection, Toxoplasma crosses the intestinal epithelium to lead to disseminated infection. The capacity of virulent and nonvirulent strains to penetrate tissue was assessed using mouse intestine in a modified Ussing two-chamber system. Parasites readily invaded the intestinal mucosa, where a majority of them (>96%) took up residence in epithelial cells (Fig. 4). However, significantly more type I strain RH parasites penetrated into deeper tissue layers than type II PTG parasites (Table I). RH parasites were more successful in entering the lamina propria, penetrating the submucosa, and in some cases were able to reach the vascular endothelium (Fig. 4 D). The notable capacity for transmigration of the RH strain thus enables the parasite to gain rapid transfer across the intestinal epithelium.

To investigate whether the differences in migration observed between LDM clones and the parental line RH, resulted in differences in dissemination during acute infection, parasites were injected intraperitoneally and the number of viable parasites in the spleen was monitored daily. Both LDM1 and RH parasites were found in the circulation and the spleen as early as 12 h after infection (data not shown). Importantly, LDM1 parasites reached the spleen in higher numbers shortly after the challenge, showing more efficient dissemination at days 2 and 3 (Fig. 5). Despite reaching the spleen more slowly the parental line RH reached similar levels of parasite burden as LDM1 from day 4–5 after infection, suggesting the maximum level of parasite growth is limited by host factors. The extreme lethality of the parental line RH (LD₁₀₀ = 1) precluded analysis of differences in mortality. Nevertheless, it is clear that upregulation of the LDM-phenotype results in more rapid dissemination in vivo.

T. gondii strains are highly clonal and distinct biological traits, such as virulence, are characteristic of each clonal type. To investigate if our observations on differences in migration could be extended to other type I, II, and III strains, 17 different strains and clinical isolates were tested in migration assays. In every case, type I strains showed significantly greater transmigration than types II and III, crossing the barrier on average with up to three orders of magnitude greater efficiency (Fig. 6). Importantly, parasites migrating >110 μm (LDM-phenotype) were only found in type I strains, extending our findings on the RH strain to all type I strains tested. In contrast, no type II or type III strain exhibited subpopulations of parasites that migrated >110 μm (LDM). Collectively, these results demonstrate a strong association of the virulence trait (type I strains) with increased migratory capacity.

During in vivo infections, Toxoplasma crosses epithelial barriers to gain access to deeper tissues including the brain, the placenta and the retina. We now demonstrate that Toxoplasma tachyzoites use active motility to cross polarized cell layers and extracellular matrix. Using in vitro and ex vivo systems that mimic biological barriers in vivo, we show that the virulent type I strains consistently exhibit a superior migratory capacity compared with the nonvirulent type II and type III strains. The potential for this capacity to be up-regulated, and its association with virulence, suggests that active parasite migration is an important component of dissemination during toxoplasmosis

Our studies establish that (a) Toxoplasma is able to actively migrate across polarized cell monolayers and extracellular matrix in vitro and across intestinal epithelium ex vivo. Furthermore, we demonstrate that (b) type I strains consistently exhibit a higher migratory capacity than type II and III strains and that (c) the migratory capacity of the highly virulent type I strain RH can be up-regulated. The enhanced transmigration capacity of type I strains, which was up to three orders of magnitude greater, could not be explained by differences in viability or efficiency of invasion. Furthermore, parasites with LDM-phenotype were present in all type I strains tested but not in type II and III strains. The migratory capacity of parasite populations was easily up-regulated in the type I strain (RH) by selecting parasites that migrated further.

Our findings demonstrate that active motility is not only used for cell invasion but also provides the parasite with an effective mechanism of dissemination in its microenvironment within tissues. Moreover, ex vivo data indicates important differences in the ability to penetrate tissues, resulting in dramatic differences in the efficiency of reaching the vascular endothelium. The ability to rapidly cross epithelial barriers and reach the circulation may be an important component of dissemination in vivo, particularly to sites of immune privilege, e.g., the central nervous system and the developing fetus.

Traversal of biological barriers, including the intestine, the blood-brain barrier and the placenta, leading to dissemination within the host is a requisite for the establishment of T. gondii infections. Ingested parasites (oocysts and bradyzoites in tissue cysts) invade the intestine and differentiate into tachyzoites followed by spread of the organism hematogenously and via lymphatics (15). Tachyzoites play an important role in pathogenesis during acute toxoplasmosis (1, 3) and upon reactivation of chronic infections in humans (21). Thus, our studies on the migratory characteristics of Toxoplasma tachyzoites are directly relevant to the role this stage plays in dissemination-related acute pathology and in reactivation of chronic infections.

T. gondii has a highly unusual population structure comprised of three clonal lineages (I, II, and III) and the parasite genotype influences the pathogenicity of the strain in mice (4, 6). Type I strains have been associated to acute virulence in the murine model and a relative over-representation of type I strains in congenital toxoplasmosis has been described (4, 5). One key feature of infections by type I strains in mice, is that they rapidly disseminate and reach high tissue burdens even from a very low initial inoculum (20, 22). The enhanced ability of type I strains to cross biological barriers likely represents an important advantage in the establishment of disseminated infection.

The strong association of the LDM-phenotype and transmigration efficiency with the parasite genotype I indicates an underlying genetic basis controls this trait. Nonetheless, the degree to which the LDM-phenotype is expressed is clearly influenced by environmental factors as it was up regulated in cloned populations (20%), but waned again over a period of 20–21 infection cycles (∼160 mitotic cycles) in the absence of selection (data not shown). LDM clones were also capable of faster migration in vivo. After oral infection, it is possible that the ability for deep tissue migration is up-regulated, for example by selection for those parasites that successfully cross the intestinal epithelial barrier. Establishment of this early wave of rapidly migrating parasites may be critically important to dissemination of infection before the onset of an effective immune response.

Natural immunity against T. gondii is dependent on the induction of strong parasite-specific immunity in the host (23). Because neutrophils rapidly migrate to a site of infection, this initial response may be important for parasite control before adaptive immunity is established (24, 25). Toxoplasma actively infects leukocytes in vitro (26) and tachyzoites have been identified within leukocytes in the murine intestine (14). Thus, trafficking of leukocytes could contribute to dissemination of intracellular parasites via a Trojan horse type of mechanism (27, 28). However, the present finding that Toxoplasma directly penetrates epithelial monolayers and extracellular matrix indicates that migration across barriers may be primarily an active process. Moreover, only active penetration by the parasite can adequately explain transplacental infection, a common feature of acute toxoplasmosis, as maternal cells do not routinely traffic to the fetus. Enhanced migration by LDM parasites was associated with a specific increase in transmigration and not simply an overall increase in motility or invasion of cells, suggesting that the parasite actively crosses biological barriers using a specific pathway. Future investigations of migration across endothelial and epithelial barriers will be important to establish the mechanism of transit.

A number of invasive bacterial, viral and fungal microorganisms have been described to cross biological barriers in the host and to cause disseminated infections. The precise mechanisms by which microbial pathogens gain entry into the target tissue are only partly understood and include the induction of endocytosis/phagocytosis (29, 30) or host actin-based translocation (31). In contrast, Toxoplasma, and other apicomplexans, actively penetrate their host cells in vitro (8) and this process is also instrumental in vivo. After subcutaneous injection by an infected mosquito, sporozoites of Plasmodium actively cross vascular endothelium and Kupffer cells to reach hepatocytes where they reside (32–36).

Toxoplasma establishes infection following oral ingestion, and consequently, to cause a systemic infection it must cross the intestinal epithelium, basement membrane, and lamina propria. Active traversal of cellular barriers represents a novel means of tissue dissemination used to gain access to biologically restricted organs, e.g., the brain and the placenta. We propose that the inherent property of virulent T. gondii type I strains to effectively migrate across barriers and up regulate the migratory capacity is an important component of pathogenesis in toxoplasmosis. Elucidation of mechanisms underlying transmigration may lead to the identification of important virulence factors in infections by T. gondii and other apicomplexan parasites.

We are grateful to Jeff Diffenderfer for expert tissue culture, to Dr. Thomas Steinberg, Dr. Jan Andersson, and Dr. Daniel Goldberg for critical evaluation of the manuscript, and to Dr. Eva Varisella for expert advice on micromanipulation.

Supported in part by a grant from the National Institutes of Health (grant no. A134036) to L.D. Sibley. A. Barragan is the recipient of a fellowship from the Swedish Foundation for International Cooperation in Research and Higher Education (STINT). L.D. Sibley is the recipient of a Scholar Award in Molecular Parasitology from the Burroughs Wellcome Fund.