Shear forces enhance Toxoplasma gondii tachyzoite motility on vascular endothelium.

Toxoplasma gondii is a highly successful parasite that infects approximately one-third of the human population and can cause fatal disease in immunocompromised individuals. Systemic parasite dissemination to organs such as the brain and eye is critical to pathogenesis. T. gondii can disseminate via the circulation, and both intracellular and extracellular modes of transport have been proposed. However, the processes by which extracellular tachyzoites adhere to and migrate across vascular endothelium under the conditions of rapidly flowing blood remain unknown. We used microfluidics and time-lapse fluorescence microscopy to examine the interactions between extracellular T. gondii and primary human endothelial cells under conditions of physiologic shear stress. Remarkably, tachyzoites adhered to and glided on human vascular endothelium under shear stress conditions. Compared to static conditions, shear stress enhanced T. gondii helical gliding, resulting in a significantly greater displacement, and increased the percentage of tachyzoites that invaded or migrated across the endothelium. The intensity of the shear forces (from 0.5 to 10 dynes/cm(2)) influenced both initial and sustained adhesion to endothelium. By examining tachyzoites deficient in the T. gondii adhesion protein MIC2, we found that MIC2 contributed to initial adhesion but was not required for adhesion strengthening. These data suggest that under fluidic conditions, T. gondii adhesion to endothelium may be mediated by a multistep cascade of interactions that is governed by unique combinations of adhesion molecules. This work provides novel information about tachyzoite interactions with vascular endothelium and contributes to our understanding of T. gondii dissemination in the infected host. IMPORTANCE Toxoplasma gondii is a global parasite pathogen that can cause fatal disease in immunocompromised individuals. An unresolved question is how the parasites circulate in the body to tissues to cause disease. T. gondii parasites are found in the bloodstream of infected animals and patients, and they have been shown to adhere to and cross the endothelial cells that line blood vessel walls. To investigate these interactions, we devised a microfluidic system to visualize parasites interacting with vascular endothelium under conditions similar to those found in the bloodstream. Interestingly, parasite migration was significantly influenced by the mechanical force of shear flow. Furthermore, we identified a role for the parasite surface protein MIC2 in the initial phase of adhesion. Our study is the first to document T. gondii interactions with endothelium under shear stress conditions and provides a foundation for future studies on the molecules that mediate parasite interaction with the vasculature.

INTRODUCTION

Toxoplasma gondii is a single-celled eukaryotic parasite that is found worldwide (1). Infection predominantly occurs through the ingestion of parasite tissue cysts or oocysts from contaminated food or water (2). In immunocompetent individuals, a robust cellular and humoral immune response controls the infection, and the parasite establishes a latent state. In immunosuppressed individuals, however, infection can result in severe toxoplasmosis, due either to reactivation of encysted parasites or to acute infection (1).

The spread of T. gondii from the intestine into secondary tissues is critical for pathogenesis (3). Infection results in colonization of the heart, brain, eye, skeletal muscle, and kidney (2, 4). Many recent studies on T. gondii dissemination have focused on an intracellular mode of transport in motile leukocytes (5–9). However, free T. gondii parasites are found in the blood of acutely infected mice (5) and humans (10), and in vitro assays have shown that extracellular tachyzoites can adhere to and cross polarized epithelial cell barriers (11) and retinal endothelium (12), suggesting that an extracellular mode of dissemination may also be employed, depending on the tissue or site in the body.

T. gondii uses an actomyosin motor complex to power substrate-dependent gliding motility and invasion (13). The parasite secretes adhesion proteins from apical secretory organelles called the micronemes. These adhesins are anchored to the parasite surface via their cytoplasmic domains, which associate with the parasite actin cytoskeleton via an aldolase bridge (14). The adhesins “treadmill” along the length of the parasite, propelling it forward, and are cleaved at the posterior end. T. gondii gliding motility has been largely studied in vitro by the examination of adhesion trails (15, 16) or by live-cell microscopy (17, 18) on protein-coated glass. Håkansson et al. defined three major types of parasite motility: circular and helical gliding, which produce net forward movement, and twirling, which does not (17). Similar modes of movement have also been described for other members of the phylum Apicomplexa, including Plasmodium (19) and Cryptosporidium sporozoites (20).

T. gondii dissemination in the body occurs via the circulation. To leave the bloodstream and enter tissues, the parasites must adhere to vascular endothelial cells and cross the endothelial barrier under conditions of blood flow. Whether T. gondii parasites are capable of completing this processes as extracellular tachyzoites has not been examined. Using a system combining live-cell microscopy with microfluidics, we performed an in-depth analysis of parasite adhesion and motility under conditions of physiologic shear stress. Remarkably, T. gondii tachyzoites adhere and glide on human endothelium under flow conditions. Fluidic shear stress significantly influenced motility dynamics and the outcome of parasite interactions with endothelium. Lastly, we identified a distinct role for the parasite microneme protein 2 (MIC2) in the initial adhesion events of T. gondii to endothelium under conditions of shear flow.

RESULTS

Extracellular T. gondii tachyzoites adhere to endothelium and are motile under shear stress conditions.

To examine T. gondii tachyzoites under conditions of fluidic shear stress, we modified our previously described fluidic system (8) to a microfluidic scale. We characterized parasite interactions with primary human umbilical vein endothelial cells (HUVEC) and with fetal bovine serum (FBS)-coated glass, which is commonly used to examine T. gondii motility (17, 21, 22). Following syringe lysis, T. gondii tachyzoites were added to wells of a chamber slide (static condition) or perfused into a microfluidic channel (flow condition) at 0.5 dyne/cm2, and movements were tracked at the apical end of the parasite for the 30 s after contact with either HUVEC or FBS. Interestingly, extracellular T. gondii were capable of adhesion and motility under shear stress conditions on both HUVEC and FBS (see Videos S1 and S2, respectively, in the supplemental material). Under both static and shear stress conditions, parasites that adhered to FBS moved at higher velocities than those that adhered to HUVEC. In addition, parasites that adhered to FBS traveled greater distances from where they initiated adhesion than those that adhered to HUVEC (Fig. 1; also, see Fig. S1 in the supplemental material). The total pathlength is the total distance a parasite travels, whereas the maximum displacement is the greatest distance a parasite travels from the point where it initially adheres. Interestingly, the introduction of shear flow significantly increased the maximum displacement of tachyzoites on both substrates, without affecting their total pathlength (Fig. 1B).

FIG 1 

Extracellular T. gondii tachyzoites adhere to and are motile on HUVEC under static and shear stress conditions. T. gondii tachyzoites were added to FBS-coated glass or confluent monolayers of HUVEC in either a chamber slide (static conditions) or in the microfluidic channel (flow conditions). (A) Representative cell traces of extracellular tachyzoites on FBS or HUVEC under static or flow conditions are shown. Each trace begins at the origin and shows the path of a single parasite (n = 15 for all conditions). The direction of flow is from left to right. (B) The average velocity, total pathlength traveled, and maximum displacement reached by each parasite during the first 30 s post-adhesion are shown. Each data point represents a single parasite. Three independent experiments were performed, and the data were pooled (nFBS, Static = 50; nFBS, Flow = 35; nHUVEC, Static = 50; nHUVEC, Flow = 49). The red bar indicates the mean for the group. ANOVA with a Tukey comparison of all means was used to compare multiple means.

We next wanted to determine if shear force influenced the likelihood of parasites remaining attached to the endothelial surface or of invasion or migration across the endothelial barrier. To address this, we observed tachyzoites interacting with endothelium under flow conditions for 12 min and categorized their location every 30 s after their adhesion to the monolayer. We found that parasites remained attached to the endothelial surface, moved into or below the endothelium (identified as dim by differential interference contrast [DIC] illumination), or detached from the monolayer (Fig. 2A). Tachyzoite movement in the z direction was most dynamic immediately after adhesion: most of the parasites that entered the monolayer did so within 4 min, and beyond this time point, there were very few changes in parasite location in the z direction. To determine whether the addition of shear force influenced parasite invasion or migration across the monolayer, we compared parasite location under static and flow conditions within this dynamic period. Under static conditions, 85% of the parasites remained on the surface of the HUVEC at 3 min post-adhesion, whereas 15% had entered the endothelium (Fig. 2B). Notably, we observed a nearly 2.5-fold increase in the percentage of parasites that migrated into or below the HUVEC monolayer under conditions of shear stress (Fig. 2B). Collectively, these data suggest that shear forces significantly affect parasite movement and the outcome of tachyzoite interactions with endothelium.

FIG 2 

Influence of shear stress on T. gondii movement into and/or across the endothelium over time. (A) T. gondii tachyzoites were flowed over HUVEC in the microfluidic channel for 12 min, and their interactions with the endothelium were analyzed. Parasites were considered adherent when they were in contact with the endothelial surface for a minimum of 4 s. The location of each adherent parasite was assessed at 30-s intervals post-adhesion. The percentages of parasites that remained attached to (red circles), moved in/below (blue squares), or detached from (green triangles) the HUVEC are shown. Three independent experiments were performed, and the data were pooled (n = 211). Bars represent means ± standard errors of the means (SEM). (B) T. gondii tachyzoites were added to HUVEC in either a chamber slide (static conditions) or in the microfluidic channel (flow conditions). At 3 min post-adhesion, the percentages of parasites that remained attached to (red), moved in/below (blue), or, in the flow condition, detached from (green) the HUVEC are shown (nStatic = 431; nFlow = 211). For each condition, at least two independent experiments were performed, and the data were pooled. Bars represent means and SEM. nd, no data. Student’s two-tailed t test with Welch’s correction was used for pairwise comparisons.

Characterization of T. gondii motility on HUVEC under shear stress conditions.

To qualitatively and quantitatively examine parasite migration under flow conditions, a detailed analysis of motility was performed. Under conditions of shear stress, parasites performed all three types of T. gondii movement (17): helical gliding, circular gliding, and twirling. Tachyzoites also moved “nonproductively,” i.e., they exhibited movement characterized by stretching and retracting along the longitudinal axis and oscillation, which did not produce net forward movement along the plane of the substrate. Motionless tachyzoites were also observed and were characterized as stationary. Time-lapse images and videos of representative parasites performing helical gliding, circular gliding, or twirling on endothelium under shear stress conditions were acquired (Fig. 3; also, see Videos S3 to S5 in the supplemental material). The path and velocity of each parasite, as determined by tracking at the apical end, are shown. Tracking was done at the apical end because this allowed clear resolution of the transitions between the different modes of parasite motility. Moreover, apical-end tracking allowed us to gain information about parasite movements that did not necessarily result in net forward motion. For instance, tracking twirling parasites at the apical end permitted an analysis of twirling “velocity,” which varied for individual parasites. During helical gliding (Fig. 3A; also, see Video S3 in the supplemental material), a characteristic 180° rotation around the longitudinal axis occurred between 31 and 34 s, producing a spike in instantaneous velocity (Fig. 3A), as has been previously reported (17). Two additional helical turns occurred between 43 and 46 s and between 49 and 52 s. After each turn, the parasite reoriented, moving at velocities of up to 2 µm/s. During circular gliding, the clockwise circular motion of the parasite was evident in the shape of the trace produced from tracking analysis (Fig. 3B). In contrast to helical gliding, circular gliding occurred at relatively low velocities (typically less than 4 µm/s) and was characterized by momentary pauses (see Video S4 in the supplemental material). During twirling, the parasites were oriented vertically, with the posterior end anchored to the substrate, and a tilting rotation was observed (Fig. 3C; also, see Video S5 in the supplemental material). On average, the apical end of parasites under shear stress conditions twirled at relatively low velocity (2 to 4 µm/s) with sporadic peaks in instantaneous velocity, indicative of a reorientation event similar to that of helical gliding, but without net forward movement parallel to the endothelial monolayer. A minority of the population was also observed performing higher-velocity twirling (>7 µm/s).

FIG 3 

Modes of T. gondii motility on HUVEC under shear stress conditions. T. gondii tachyzoites were flowed over HUVEC at 0.5 dyne/cm2. Cell tracking software was used to track the movements at the apical end of the parasite (designated with a colored circle). Representative time-lapse images of parasites performing helical gliding (A), circular gliding (B), or twirling (C) are shown. The path of the apical end of the parasite and the instantaneous velocity at the apical end are shown for each mode of motility. The colored dots in the velocity-versus-time plots correspond to those in the time-lapse images. The direction of flow is from left to right. Bars, 10 µm.

We frequently observed tachyzoites exhibiting multiple types of motility over time (Fig. 4A and B; also, see Video S6 in the supplemental material). To better understand these complex parasite movements and the effects of substrate and fluidic shear stress, parasite motility was classified at either 10 s or 60 s after adhesion to either FBS or HUVEC and under static or flow conditions (Fig. 4C and D). The percentages of parasites that were moving, nonproductive, or stationary under each condition are shown (Fig. 4C). For parasites that moved productively in Fig. 4C, the absolute numbers of parasites performing helical or circular gliding or twirling were graphed (Fig. 4D). For parasites on FBS, the percentage of the parasite population that was motile remained relatively constant over time (Fig. 4C), but the addition of shear stress resulted in a significantly larger portion of parasites performing helical gliding (Fig. 4D). In contrast, for parasites that adhered to HUVEC, shear stress conditions increased the proportion of motile parasites immediately following adhesion (10 s) and specifically enhanced helical and circular gliding; however, these highly dynamic modes of motility declined over time (60 s). When we examined the motility of a representative subset of 15 parasites during their entire first minute post-adhesion, we observed similar trends (Fig. 5). We also observed a fraction of parasites that reinitiated motility after prolonged stationary periods. These data indicate that flow conditions increased dynamic, productive motility (helical and circular gliding) immediately after attachment, which may contribute to the increased displacement observed for parasites under shear stress (Fig. 1).

FIG 4 

Characterization of parasite motility under static and shear stress conditions. T. gondii tachyzoites were added to FBS or HUVEC in either a chamber slide (static conditions) or the microfluidic channel (flow conditions). (A) The trace generated by a tachyzoite exhibiting three modes of motility within 1 min post-adhesion to HUVEC under flow conditions is shown. (B) A velocity plot over time is shown for the parasite examined in panel A. (C) The percentages of motile, nonproductive, and stationary parasites at either 10 s or 60 s after adhesion to FBS or HUVEC under static or flow conditions are shown (nFBS, Static = 50; nFBS, Flow = 35; nHUVEC, Static = 50; nHUVEC, Flow = 49). (D) The absolute numbers of motile parasites (from panel C) undergoing helical or circular gliding or twirling at either 10 s or 60 s after adhesion to FBS or HUVEC under static or flow conditions are shown. nFBS, Static, 10 s = 21; nFBS, Flow, 10 s = 28; nFBS, Static, 60 s = 23; nFBS, Flow, 60 s = 26; nHUVEC, Static, 10 s = 11; nHUVEC, Flow, 10 s = 17; nHUVEC, Static, 60 s = 12; nHUVEC, Flow, 60 s = 3.

FIG 5 

Parasite motility over time under static and shear stress conditions. T. gondii tachyzoites were added to FBS or HUVEC in either a chamber slide (static conditions) or the microfluidic channel (flow conditions). The motility of a representative subset of 15 parasites under each condition over time is shown, with each bar representing the motility of a single parasite.

Initial and sustained parasite adhesion at increasing shear force.

The initial adhesion and in-depth motility analyses of extracellular tachyzoites were performed at 0.5 dyne/cm2, a relatively low physiological shear stress (23). To address the effects of higher levels of shear stress on parasite adhesion, T. gondii tachyzoites were flowed over confluent monolayers of HUVEC at 0.5, 2, 5, or 10 dynes/cm2 for 3 min, and adhesion events were counted. An adhesion event was defined as a parasite that appeared on the endothelial surface and remained attached in the plane of focus for a minimum of 4 s. As shear force increased, we observed a decrease in the number of adhesion events (Fig. 6A; also, see Fig. S2 in the supplemental material).

FIG 6 

Initial and sustained adhesion with increasing shear force. T. gondii tachyzoites were flowed over confluent monolayers of HUVEC in the microfluidic channel. (A) The total numbers of parasite adhesion events at shear forces of 0.5, 2, 5, and 10 dynes/cm2 are shown. The graph represents aggregated data from 4 independent microfluidic experiments (n = 468). (B) T. gondii tachyzoites were flowed over HUVEC at 0.5 dyne/cm2 for a 3 min “loading” phase. The shear force was then maintained at 0.5 dyne/cm2 or switched to 2, 5, or 10 dynes/cm2. Parasites that were attached to the endothelial surface at the time of the switch were monitored for an additional 3 min. The percentage of adherent parasites that remained attached to, moved inside/below, or detached from the HUVEC monolayer during the “monitoring” phase is shown. Bars represent means and SEM. n = 276 from 4 independent experiments.

We next investigated the adhesion strength of tachyzoites by examining their response to an increase in shear stress. This can be addressed by allowing adhesion to occur under static or low-shear conditions and then increasing the flow rate (24). To do this, tachyzoites were flowed over HUVEC at 0.5 dyne/cm2 for 3 min to allow for adhesion, and then the shear stress was maintained at 0.5 dyne/cm2 or increased to 2, 5, or 10 dynes/cm2 for an additional 3 min. Parasites could adhere at any point during the first 3 min of flow; however, only those that were attached to the HUVEC surface at the switch in shear force were analyzed. The percentage of adherent tachyzoites that remained attached, entered the monolayer (in/below), or detached after the switch in shear stress was determined. We found that the percentage of adherent parasites that detached from the monolayer increased as shear stress increased (Fig. 6B). The increase in shear stress, however, did not affect the proportion of parasites that entered the monolayer (“in/below” category in Fig. 6B). These data indicate that the intensity of the shear stress influences the number of parasites that adhere to the endothelium and their ability to sustain adhesion; however, it does not appear to affect their likelihood of moving into or below the monolayer, at least at shear forces up to 10 dynes/cm2.

Invasion or transmigration of T. gondii under shear stress conditions.

We reasoned that parasites migrating into or below the monolayer may be invading the endothelium, as has been previously reported (25), or transmigrating across the endothelial barrier via a paracellular or transcellular route. We also wanted to determine if shear force influenced these possible outcomes of movement into the monolayer. For the static condition, tachyzoites were incubated on HUVEC monolayers for 20 min and then washed and fixed. For shear stress conditions, T. gondii were flowed over the endothelium for 1 to 2 min, and a Y-valve connector was used to flow medium alone without disrupting flow for a total of 20 min. Fluidic channels were then perfused with fixative. Under both conditions the cells were then stained with antibodies against the parasite dense granule protein GRA7, as a marker for invasion. In extracellular tachyzoites, GRA7 is found in the dense granules, whereas in infected host cells, GRA7 localizes to the parasitophorous vacuole (PV) (26). In both static and flow conditions, we observed tachyzoites that appeared dim by DIC illumination (in/below) and were either surrounded by GRA7 staining or not (Fig. 7A). Under static conditions, 81% of DIC-dim parasites localized to GRA7+ vacuoles, whereas 19% were not associated with GRA7 staining. Interestingly, flow conditions significantly increased the percentage of DIC-dim parasites that were GRA7− (35% of the population) (Fig. 7B). The DIC-dim GRA7− population likely included parasites that had completed transmigration and were under the monolayer. Indeed, in live-cell imaging experiments, we have observed tachyzoites that continued to migrate under the endothelium after entering the monolayer. We cannot exclude the possibility that some of the DIC-dim GRA7− tachyzoites were in the process of transcytosis at the time of analysis (in a vesicle that lacked GRA7) or had invaded immediately prior to fixation and had yet to secrete a sufficient amount of GRA7 into the PV to be detectable by immunofluorescence assay (IFA). Collectively, these data indicate that the majority of tachyzoites that enter the endothelium under shear stress and static conditions invade the monolayer and form a PV, and that flow conditions may result in a greater percentage of parasites that breach the barrier and undergo transmigration.

FIG 7 

Invasion or transmigration of T. gondii under static and shear stress conditions. T. gondii tachyzoites were incubated for 20 min with a HUVEC monolayer under either static or shear stress conditions. Coverslips or fluidic channels were then fixed and stained with antibodies against GRA7. (A) Sample micrographs depict DIC-dim parasites (previously categorized as in/below in Fig. 2 and 6B) that are either GRA7+ (white arrowheads) or GRA7−. Bar, 10 µm. (B) The percentages of DIC-dim parasites associated with GRA7 are shown. For each condition, two independent experiments were performed and the data were pooled (nStatic = 499; nFlow = 309). Bars represent means ± SEM. Student’s two-tailed t test with Welch’s correction was used for pairwise comparisons. *, P < 0.01 for comparisons between the static and flow conditions.

Role of MIC2 in adhesion to HUVEC under shear stress conditions.

We next sought to address the role of parasite adhesion proteins in interactions with the endothelium under shear stress conditions and the possibility that different parasite adhesins may contribute to distinct stages of the T. gondii adhesion cascade in shear stress. One of the best-characterized T. gondii surface proteins associated with adhesion and motility is the microneme protein MIC2 (21, 27–30). Under static assay conditions, MIC2 was found to be important for attachment to host cells, helical gliding, and invasion (21). However, a role for MIC2 in adhesion to vascular endothelium or under conditions of shear stress has not yet been examined. To address this possibility, we employed a conditional MIC2 knockdown system developed and generously provided by the Carruthers lab (21). Parental tTA-dhfr parasites and the Δmic2e/mic2i conditional MIC2 knockdown parasite line, which uses a tetracycline-responsive promoter to control the expression of MIC2, were used. The addition of anhydrotetracycline (ATc) reduced the expression of MIC2 as determined by both immunofluorescence assay (Fig. 8A) and Western blotting (Fig. 8B). The tTA-dhfr and ATc-treated Δmic2e/mic2i tachyzoites were labeled with different fluorescent dyes, mixed 1:1, and perfused over a HUVEC monolayer, and adhesion events under flow conditions were counted along the entire length of the fluidic channel. A sample of the mixture was used to determine the precise input ratio of the two strains by both flow cytometry and by microscopy (data not shown). The ratio of tTA-dhfr and ATc-treated Δmic2e/mic2i tachyzoite adhesion events in the channel relative to the input ratio was calculated. A value of 1 indicates equivalent adhesion of the two populations. We observed a consistent underrepresentation of the ATc-treated Δmic2e/mic2i parasites compared to the tTA-dhfr parasites (Fig. 8C). To control for the possibility that treatment with ATc affected adhesion, we evaluated the adhesion of ATc-treated and untreated tTA-dhfr parasites under static conditions and found no significant difference in their adhesion (see Fig. S3A in the supplemental material). In addition, when we examined adhesion in a static assay, we observed a similar underrepresentation of ATc-treated Δmic2e/mic2i parasites compared to the tTA-dhfr parasites (see Fig. S3B). These data indicate that MIC2 contributes to T. gondii adhesion to endothelium in shear stress; however, a portion of ATc-treated Δmic2e/mic2i parasites were still capable of adhesion under these conditions. Although we observed that MIC2 levels were substantially reduced in ATc-treated parasites, it is possible that a low level of MIC2 expression in these treated parasites contributed to this residual adhesion.

FIG 8 

Role of MIC2 in adhesion to HUVEC under shear stress conditions. (A) The Δmic2e/mic2i parasites were treated with ATc for 2 days. Extracellular tachyzoites were stained with antibodies against MIC2 or SAG1 and imaged by DIC and by fluorescence for MIC2 expression. Bar, 5 µm. (B) tTA-dhfr and Δmic2e/mic2i parasites were untreated or treated with ATc for 2 days. Western blotting was performed to detect MIC2 or SAG1 in the parasite lysates. (C) tTA-dhfr tachyzoites were mixed 1:1 with ATc-treated Δmic2e/mic2i tachyzoites and flowed over HUVEC. The parasites were allowed to adhere for 3 min, and DIC and fluorescence images were captured along the entire length of the channel while the cells were still under flow. Parasite adhesion relative to the input ratio is shown. A value of 1.0 would represent equivalent adhesion of the two populations. ntTA-dhfr = 1,120; nATc-treated Δmic2e/mic2i = 1,073 (3 independent experiments). (D) A mixture of tTA-dhfr and ATc-treated Δmic2e/mic2i parasites were flowed over HUVEC in the microfluidic channel at 0.5 dyne/cm2 for 3 min, and then the shear stress was increased to 10 dynes/cm2 for an additional 3 min. The percentage of adherent parasites that remained attached or detached after the increase in shear stress is shown. ntTA-dhfr = 70; nATc-treated Δmic2e/mic2i = 88 (2 independent experiments). (C and D) Error bars represent means and SEM, and Student’s two-tailed t test with Welch’s correction was used for pairwise comparisons. ns, not significant.

We next investigated a potential role for MIC2 in adhesion strengthening under conditions of increasing shear stress. To do this, a mixture of tTA-dhfr and ATc-treated Δmic2e/mic2i tachyzoites were perfused over a HUVEC monolayer at 0.5 dyne/cm2 to allow adhesion to occur. The shear stress was then increased to 10 dynes/cm2. The percentage of tachyzoites that remained attached or that detached from the monolayer after the increase in shear force was calculated. Interestingly, the tTA-dhfr and ATc-treated Δmic2e/mic2i tachyzoites remained similarly attached to the endothelium after the increase in shear stress to 10 dynes/cm2 (Fig. 8D): for both populations of parasites, approximately 50% of the tachyzoites detached after the transition to increased shear force. Collectively, these data suggest that MIC2 plays an important role in the initial phase of attachment to the endothelium but is not required for adhesion strengthening and that other parasite adhesins may mediate these sustained interactions with the endothelium.

DISCUSSION

Dissemination of T. gondii from the primary site of infection to secondary tissues is thought to occur through both intracellular and extracellular mechanisms (3). To leave the bloodstream, tachyzoites interacting with blood vessel walls must contend with the mechanical forces of shear stress. Live-cell microfluidic systems provide a highly tunable method for characterizing cellular adhesion and motility and analyzing the role of specific adhesins in attachment to human cellular substrates under flow conditions. Our current study shows that tachyzoites can adhere to and migrate on vascular endothelium under conditions of physiologic shear stress. In addition, we found that shear forces enhance dynamic motility and the likelihood of invasion and/or transmigration, both of which are critical components of pathogenesis. These findings support the relevance of including mechanical force as an additional experimental parameter in understanding patterns of parasite migration. In addition, recent work on T. gondii motility in a three-dimensional system revealed that tachyzoites move through Matrigel in a corkscrew-like trajectory distinct from that observed in two-dimensional assays (31), further reinforcing the value of incorporating conditions that more closely mimic in vivo environments.

Multistep adhesion cascades are used by both circulating mammalian cells (32) and pathogens (33, 34) to exit the bloodstream. However, the precise manner in which these interactions occur appears to vary. Initial adhesion interactions, often referred to as “tethering” or “capture,” are transient, occur at high velocities, and serve to decelerate the adhering cell. In Borrelia burgdorferi, tethering is mediated by the fibronectin-binding domain of the bacterial adhesin BBK32 and is the essential first interaction that decelerates the bacterium and brings it into close contact with the endothelial cell surface (34). For Plasmodium berghei, initial sporozoite adhesion to glass flow chambers requires the parasite adhesins TRAP and S6 (33). Following initial adhesion, transient adhesion associated with a phase of rolling or active motility occurs. P. berghei sporozoites wave or glide (33), whereas B. burgdorferi organisms perform dragging mediated by the glycosaminoglycan-binding domain of BBK32 (34). Subsequently, a transition to firm adhesion occurs. For B. burgdorferi a period of “stationary adhesion” near endothelial cell junctions immediately follows dragging and precedes transmigration and escape from the vasculature (35). A similar period of sustained firm adhesion prior to extravasation has been documented for P. berghei exiting blood vessels in the mouse dermis (36). We frequently observed that T. gondii gliding on endothelium was punctuated by periods of motionless activity, sometimes lasting several minutes, prior to reinitiation of migration. One potential explanation for these patterns is fluctuation in intracellular cation concentration. Intraparasitic calcium mobilization has been shown to coordinate microneme secretion (37) and influence motility (38). In addition, parasite egress is regulated by both potassium and calcium flux (39).

Our data suggest that T. gondii tachyzoite adhesion to endothelium under shear stress conditions could also be divided into a multistep cascade. It is likely that a combination of adhesion molecules, potentially with overlapping functions, mediate T. gondii adhesion under shear stress conditions. In our studies, MIC2 clearly played a role in this process. However, under conditions of increasing shear stress the MIC2-deficient parasites remained as adherent as the parental control parasites, implying a role for MIC2 in initial parasite adhesion that may be distinct from adhesion strengthening. One potential ligand for MIC2 on the endothelium is ICAM-1. MIC2 has been shown to interact with ICAM-1 to mediate T. gondii transepithelial migration (30), and ICAM-1-dependent transmigration across retinal endothelium has been observed (12). In our assays, a portion of the MIC2-deficient parasites were capable of adhesion to endothelium under shear stress conditions, indicating that other parasite factors may contribute to this process. T. gondii has been shown to use a broad range of sulfated glycosaminoglycans to bind to substratum and host cells in static systems (40). Other microneme proteins are also known to contribute to parasite adhesion, invasion, and virulence in mice (41). MIC1 binds host glycans, and a conserved microneme adhesive repeat (MAR) recognizes sialic acid and contributes to invasion (42, 43). The chitin binding-like domain of MIC3 is required for attachment to host cells (44, 45). Single gene disruptions of either MIC1 or MIC3 produce mild virulence defects; however, double knockouts produce a severe virulence defect in vivo (45). Also, T. gondii surface antigen 3 (SAG3) can mediate the recognition of cell surface proteoglycans, such as heparin (46), and SAG3-deficient parasites have a 50 to 60% adhesion deficit in in vitro assays and an 80% reduction in virulence in vivo (47). Whether carbohydrate-mediated adhesion interactions contribute to tachyzoite capture, motility, sustained adhesion, or transmigration under shear stress conditions remains to be investigated.

Interestingly, T. gondii migrated farther and at higher velocities on FBS than on endothelial cells, suggesting that the endothelium itself may be a barrier to motility. This may be due to the varied topography of the endothelium and to the adhesion molecules expressed on the endothelial surface. In addition, vascular endothelium is heterogeneous in both structure and function (48), suggesting that the processes by which tachyzoites adhere to distinct endothelial cell types may also differ. Endothelial permeability is also highly modulated by acute and chronic inflammation. T. gondii infection induces an inflammatory response characterized by the expression of interleukin 12 (IL-12), tumor necrosis factor alpha (TNF-α), and IL-6 (49). T. gondii infection of brain endothelial cells upregulates the expression of ICAM-1 (50), and infection of human retinal epithelial cells upregulates ICAM-1 and the inflammatory cytokines IL-1 and IL-6 (51). The stimulation of endothelial cells with proinflammatory mediators induces a proadhesive phenotype termed “endothelial cell activation” (52, 53). “Activated” endothelium expresses elevated levels of adhesion molecules and has decreased tight-junction integrity, resulting in an increase in barrier permeability (54). The unique response of endothelium to infection-induced inflammatory mediators in a given tissue may impact the likelihood of parasite invasion of that tissue. Ultimately, understanding how these interactions allow T. gondii dissemination out of the circulation will provide novel insight into mechanisms of parasite spread into tissues and the development of disease.

MATERIALS AND METHODS

Mammalian and parasite cell culture.

HUVEC were cultured in endothelial growth medium 2 (EGM-2) with EGM-2 SingleQuot supplements and growth factors (Lonza, Allendale, NJ). Green fluorescent protein (GFP)-expressing type I T. gondii (RH) tachyzoites were maintained in human foreskin fibroblasts (HFF) in Dulbecco’s modified Eagle medium (DMEM) with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin (D-10%), as previously described (55). Parasites were syringe lysed, washed, and processed in a high-potassium buffer and then immediately transferred into D-10% for experimentation to prevent microneme exhaustion prior to adhesion and motility assays (56). All parasite and human cell cultures were tested monthly for Mycoplasma contamination and confirmed to be negative.

Fabrication of fluidic systems.

Fluidic channels (height [H], 1.5 mm; width [W], 3 mm; length [L], 20 mm) were fabricated as previously reported (8) with the following modification: polydimethylsiloxane (PDMS) was adhered to cover glass by plasma bonding. New microfluidic channels (H, 0.1 mm; W, 1.5 mm; L, 10 mm) were fabricated using standard soft lithographic techniques (57). Briefly, a photomask containing arrays of a 1.5-mm by 10-mm pattern was generated by high-resolution laser printing on a transparency (CAD/Art Services Inc., Bandon, OR). A 100-µm thick layer of negative SU-8 photoresist (MicroChem Corp., Newton, MA) was deposited onto a silicon wafer and exposed to UV through the photomask to polymerize. Unpolymerized SU-8 was removed to create relief structures of the channels on the master. PDMS was then cast from the master, oxygen-plasma cleaned (Harrick O2; Harrick Plasma, Ithaca, NY), and permanently bonded to glass slides.

Fluidic experiments.

Fluidic channels were coated with 20 µg/ml fibronectin and seeded with HUVEC overnight. Microfluidic channels were either coated with a 1:1 mixture of FBS in phosphate-buffered saline (PBS) for a minimum of 1 h or coated with 20 µg/ml fibronectin and seeded with HUVEC for a minimum of 7 h prior to experimentation. Channels were connected to an open-loop flow system as previously described (8), with the flow rate controlled by an 11 Plus dual syringe pump (Harvard Apparatus, Holliston, MA). The device was mounted inside a UNO environmental chamber on the microscope stage (OkoLab, Ottaviano, NA, Italy) and maintained at 37°C and 5% CO2. Tachyzoites were loaded into the syringe pump and perfused through the channels, and live time-lapse imaging was started immediately after flow was initiated. Imaging was performed on a Nikon Eclipse Ti inverted fluorescence microscope using NIS Elements acquisition software. Fluidic assays were performed at a shear force of 0.5 dyne/cm2 unless otherwise noted.

Static experiments.

Wells of an 8-well chamber slide were either coated with a 1:1 mixture of FBS in PBS for a minimum of 1 h or coated with 20 µg/ml fibronectin and seeded with HUVEC. The chamber slide was maintained at 37°C and 5% CO2 in a UNO environmental chamber (9), and the wells were inoculated with tachyzoites. Time-lapse DIC imaging was initiated immediately after the addition of parasites. For the MIC2 adhesion assays in static conditions, the parasites were allowed to adhere to HUVEC for 15 min at 37°C, and the cells were washed 3 times with PBS to remove any unadherent parasites. Monolayers were then fixed with 4% paraformaldehyde (PFA; Electron Microscopy Sciences, Hatfield, PA) for 20 min at room temperature. DIC and fluorescence images were acquired in 6 random fields of view per well using a 20× objective.

Parasite tracking.

The tracking module of the Nikon NIS Elements AR 3.22.11 software was used to manually track the movements of the apical end of individual tachyzoites. For assays in shear stress, every parasite that entered the field of view was tracked. For static assays, at least the same number of parasites that were tracked under flow conditions were chosen at random and tracked as they came into focus. For plotting the traces of individual parasites, the origin (0.0) was set to the parasite’s apical end when it first came into contact with the endothelium. The x and y coordinates of the apical end in the following frames were then used to plot each parasite’s path on the endothelium. The average velocity, total pathlength, and maximum displacement of individual parasites during their first 30 s post-adhesion were plotted using Prism (GraphPad Software, La Jolla, CA).

Motility and adhesion assays.

To assess parasite adhesion and motility on HUVEC or FBS-coated glass, time-lapse DIC images were acquired using a 20× objective with 1.5× zoom. For the motility assay data shown in Fig. 1, 3, 4, and 5 and in Fig. S1 in the supplemental material, images were acquired every second for up to 10 min, and an adhesion event was defined as the presence of a tachyzoite on the endothelial surface for a minimum of 3 s. Velocity was defined as the distance traveled from one frame to the next divided by the time-lapse. For the sustained adhesion assays in Fig. 2, images were acquired every 4 s for up to 15 min, and an adhesion event was defined as the presence of a tachyzoite on the endothelial surface for a minimum of 4 s. Parasite location in the z direction relative to the endothelium was determined by DIC illumination during live-cell imaging. In the initial interactions of tachyzoites with the surface of the endothelium, they appeared bright. Migration into or below the HUVEC monolayer was characterized by the dimming of DIC illumination. By using this change in DIC illumination, the parasites were categorized as remaining attached to, migrating in/below, or detaching from the endothelium. For the adhesion-strengthening assay data shown in Fig. 6 and in Fig. S2 in the supplemental material, perfusion was initiated at 0.5 dyne/cm2. Three minutes later, the shear stress in the channel was either maintained or increased to 2, 5, or 10 dynes/cm2. Parasites that adhered during the initial 3 min and remained attached at the surface at the change in shear stress were monitored and categorized as adherent to, detached from, or in/below the endothelium at the conclusion of 6 min of imaging. De novo adhesion events that occurred after the switch in shear force were also quantified.

Immunofluorescence microscopy.

For GRA7 staining of parasites under static conditions, 5 × 105 GFP-expressing RH tachyzoites were added to confluent HUVEC monolayers on coverslips and incubated for 20 min. Coverslips were then washed with PBS and fixed with 4% PFA. The cells were permeabilized and stained for vacuolar GRA7 as previously described (58). Anti-GRA7 antibody (12B6) was a generous gift from P. Bradley (University of California, Los Angeles, CA). For shear stress conditions (Fig. 7), 5 to 10 × 106/ml tachyzoites were flowed at 0.5 dyne/cm2 into fluidic channels over HUVEC. After 1 to 2 min, a Y-valve connector was used to perfuse D-10% through the device, without disrupting flow, for an additional 18 to 20 min. The channels were then perfused with PBS, followed by 4% PFA, and permeabilized and stained as described above. Fluorescence and DIC micrographs were acquired using a 40× or 60× objective.

For the analysis of MIC2 expression, HFF infected with tTA-dhfr or Δmic2e/mic2i tachyzoites were cultured in the presence or absence of ATc for 2 days. Tachyzoites were harvested by syringe lysis, added to polylysine-coated coverslips, and fixed with 4% PFA. Samples were stained with anti-MIC2 (6D10) and rabbit anti-SAG1 polyclonal serum (59), followed by Alexa Fluor 305-conjugated goat anti-mouse IgG and Alexa Fluor 594-conjugated goat anti-rabbit IgG (Life Technologies) in PBS containing 3% FBS. Coverslips were mounted with Vectashield (Vector Laboratories, Burlingame, CA). Anti-MIC2 (6D10) was kindly provided by V. Carruthers (University of Michigan, Ann Arbor, MI). To determine the input ratio of fluorescently dyed Δmic2e/mic2i and tTA-dhfr parasites, a wet mount of each mixture was analyzed. Images were acquired using a 10× objective with a 1.5× zoom and counted manually.

Western blotting.

Parasite cell lysates were separated on 10% SDS-PAGE, transferred to polyvinylidene difluoride (PVDF) membranes, and immunoblotted with mouse anti-MIC2 (6D10) and rabbit anti-SAG1 (59). The membranes were then stained with horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit IgG (Jackson Immuno Research, West Grove, PA), developed using enhanced chemiluminescence (ECL) (GE Healthcare Life Sciences, Pittsburgh, PA), and detected using a Nikon camera as previously described (60).

MIC2 assays.

Δmic2e/mic2i and tTA-dhfr parasites (21) were a generous gift from V. Carruthers (University of Michigan, Ann Arbor, MI). For the MIC2 knockdown experiments, Δmic2e/mic2i tachyzoites were cultured in the presence of 2 mg/ml of ATc (Clontech, Mountain View, CA) for 2 days, and the tTA-dhfr tachyzoites were cultured in parallel. Parasites were syringe lysed from the HFF cultures, washed in high-potassium buffer, and stained with either 20 µM CellTracker CMPTX (Life Technologies, Carlsbad, CA) or 10 µM carboxyfluorescein succinimidyl ester (CFSE) (Life Technologies) in high-potassium buffer. In control experiments, we confirmed that these dyes do not have an effect on tachyzoite adhesion or motility. Nonetheless, we alternated the dyes on the parental and knockdown parasites in each experimental run to control for this possibility. A 1:1 mixture of parasites at 5 × 106 to 10 × 106/ml was perfused at 0.5 dyne/cm2 into the microfluidic channel. For adhesion assays (Fig. 8C), the parasites were perfused for 3 min, and live images were acquired along the entire length of the channel using a 10× objective. The total number of adherent parasites in both populations was then quantified and normalized to the input ratio, as determined by flow cytometry and fluorescence microscopy. The sustained adhesion assays (Fig. 8D) were performed as described above except that a mixture of dyed tTA-dhfr and ATc-treated Δmic2e/mic2i tachyzoites was used.

Statistical analysis.

Prism Software was used to perform all statistical analyses. A one-way analysis of variance (ANOVA) with a Tukey comparison of all means was used to compare multiple means in Fig. 1B. Student’s two-tailed t test with Welch’s correction was used for pairwise comparisons in Fig. 2B, 7B, 8C, and 8D and in Fig. S3 in the supplemental material.

Individual traces of tachyzoite movement. Traces of extracellular tachyzoites on FBS under static conditions (A) or flow conditions (B) and on HUVEC under static conditions (C) or flow conditions (D) are shown. Each trace begins at the origin and shows the path of a single parasite per plot. These traces are grouped by condition and shown as combined data in Fig. 1A. Download

Figure S1, TIF file, 1 MB

Adhesion events in increasing shear stress. Data collected from 4 independent microfluidic experiments show the total number of adhesion events on HUVEC during flow at increasing shear stress. n = 468. Download

Figure S2, TIF file, 0.5 MB

Adhesion in static conditions. The indicated parasite populations were differentially labeled with CellTracker dyes, mixed, and added to HUVEC in a chamber slide for 15 min. The cells were washed 3 times with PBS to remove nonadherent parasites, fixed, and imaged in DIC and fluorescence. Parasite adhesion relative to the input ratio is shown. A value of 1.0 would represent equivalent adhesion of the two populations. (A) tTA-dhfr parasites treated with ATc for 2 days were mixed 1:1 with untreated tTA-dhfr parasites. nUntreated tTA-dhfr = 1,316; nATc-treated tTAdhfr = 1,278 (from 2 independent experiments). (B) tTA-dhfr parasites were mixed 1:1 with ATc-treated Δmic2e/mic2i parasites. ntTA-dhfr = 1,598; nATc-treated Δ = 932 (from 2 independent experiments). Error bars represent means ± SEM, and Student’s two-tailed t test with Welch’s correction was used for pairwise comparisons. ns, not significant. Download

Figure S3, TIF file, 0.4 MB

Tachyzoite adhesion and gliding on HUVEC under shear stress conditions. DIC micrographs were acquired every 3 s and played back at 10 frames/s. The direction of flow is from left to right. Bar, 20 µm. Download

Video S1, MOV file, 0.5 MB

Tachyzoite adhesion and gliding on FBS under shear stress conditions. DIC micrographs were acquired every second and played back at 10 frames/s. The direction of flow is from left to right. Bar, 20 µm. Download

Video S2, MOV file, 0.5 MB

Tachyzoite performing helical gliding on HUVEC under shear stress conditions. DIC micrographs were acquired every second and played back at 7 frames/s. The red arrow indicates the apical end of the parasite. The direction of flow is from left to right. Bar, 10 µm. Download

Video S3, MOV file, 0.1 MB

Tachyzoite performing circular gliding on HUVEC under shear stress conditions. DIC micrographs were acquired every second and played back at 7 frames/s. The red arrow indicates the apical end of the parasite. The direction of flow is from left to right. Bar, 10 µm. Download

Video S4, MOV file, 0.1 MB

Tachyzoite performing twirling on HUVEC under shear stress conditions. DIC micrographs were acquired every second and played back at 7 frames/s. The red arrow indicates the apical end of the parasite. The direction of flow is from left to right. Bar, 10 µm. Download

Video S5, MOV file, 0.1 MB

Tachyzoite performing complex motility on HUVEC under shear stress conditions. DIC micrographs were acquired every second and played back at 7 frames/s. The mode of motility is indicated in red in the bottom left corner throughout the duration of movement. The direction of flow is from left to right. Bar, 10 µm. Download

Video S6, MOV file, 0.2 MB