Rapid cytoskeleton remodelling in dendritic cells following invasion by Toxoplasma gondii coincides with the onset of a hypermigratory phenotype

Authors

  • Jessica M. Weidner,

    1. Center for Infectious Medicine, Department of Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    2. Swedish Institute for Communicable Disease Control, Stockholm, Sweden
    Search for more papers by this author
  • Sachie Kanatani,

    1. Center for Infectious Medicine, Department of Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    2. Swedish Institute for Communicable Disease Control, Stockholm, Sweden
    Search for more papers by this author
  • Maria A. Hernández-Castañeda,

    1. Center for Infectious Medicine, Department of Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    2. Swedish Institute for Communicable Disease Control, Stockholm, Sweden
    Search for more papers by this author
  • Jonas M. Fuks,

    1. Center for Infectious Medicine, Department of Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    2. Swedish Institute for Communicable Disease Control, Stockholm, Sweden
    Search for more papers by this author
  • Bence Rethi,

    1. Department of Microbiology, Tumor- and Cell Biology, Karolinska Institutet, Stockholm, Sweden
    Search for more papers by this author
  • Robert P. A. Wallin,

    1. Center for Infectious Medicine, Department of Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    2. Department of Microbiology, Tumor- and Cell Biology, Karolinska Institutet, Stockholm, Sweden
    Search for more papers by this author
  • Antonio Barragan

    Corresponding author
    1. Swedish Institute for Communicable Disease Control, Stockholm, Sweden
    • Center for Infectious Medicine, Department of Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    Search for more papers by this author

For correspondence. E-mail antonio.barragan@ki.se; Tel. (+46) 8457 2524. Fax (+46) 8457 3762.

Summary

Host cell manipulation is an important feature of the obligate intracellular parasite Toxoplasma gondii. Recent reports have shown that the tachyzoite stages subvert dendritic cells (DC) as a conduit for dissemination (Trojan horse) during acute infection. To examine the cellular basis of these processes, we performed a detailed analysis of the early events following tachyzoite invasion of human monocyte-derived DC. We demonstrate that within minutes after tachyzoite penetration, profound morphological changes take place in DC that coincide with a migratory activation. Active parasite invasion of DC led to cytoskeletal actin redistribution with loss of adhesive podosome structures and redistribution of integrins (CD18 and CD11c), that concurred with the onset of DC hypermotility in vitro. Inhibition of parasite rhoptry secretion and invasion, but not inhibition of parasite or host cell protein synthesis, abrogated the onset of morphological changes and hypermotility in DC dose-dependently. Also, infected DC, but not by-stander DC, exhibited upregulation of C-C chemokine receptor 7 (CCR7). Yet, the onset of parasite-induced DC hypermotility preceded chemotactic migratory responsesin vitro. Collectively, present data reveal that invasion of DC by T. gondii initiates a series of regulated events, including rapid cytoskeleton rearrangements, hypermotility and chemotaxis, that promote the migratory activation of DC.

Introduction

Toxoplasma gondii is a ubiquitous obligate intracellular parasite that infects humans and animals. In humans, severe clinical manifestations of toxoplasmosis are linked to the systemic dissemination of the parasite, e.g. life-threatening neurological complications in immune-compromised individuals; disseminated congenital infections in the developing fetus and ocular pathology in otherwise healthy individuals (Joynson and Wreghitt, 2001).

Following oral infection, the parasite crosses epithelial and endothelial cellular barriers to enter into circulation and disseminate within the organism. T. gondii can infect a range of host cells, e.g. leukocytes, by active penetration. The process of invasion involves discharge of secretory organelles from the parasite, e.g. rhoptries, and does not primarily rely on the host cell machinery for uptake (Sibley, 2004; Delorme-Walker et al., 2012). While the exact mechanisms of parasite translocation across restrictive biological barriers have yet to be determined, T. gondii may exploit the migratory properties of dendritic cells (DC) for shuttling in the organism by a Trojan horse type of mechanism (Courret et al., 2006; Lambert et al., 2006; Bierly et al., 2008). Recent work has shown that upon infection by tachyzoites, DC exhibit a hypermigratory phenotype (Lambert et al., 2006; 2009) that is potentiated through GABAergic signalling pathways (Fuks et al., 2012).

Mounting evidence indicates that DC play a pivotal role during T. gondii infection as mediators of essential immune responses (Liu et al., 2006; Mashayekhi et al., 2011) and as parasite carriers that facilitate the dissemination of the infection (Courret et al., 2006; Lambert et al., 2006; 2009; Bierly et al., 2008). As a fundamental component of the immune response, DC serve as sensors in peripheral tissues that allow processing and presentation of antigens for initiation of adaptive immune responses and pathogen clearance (Banchereau and Steinman, 1998). The mechanisms underlying DC maturation and migration are complex, and the molecular trafficking signals that govern DC migration are not fully understood (Alvarez et al., 2008). One of the hallmarks of maturing DC is the expression of the C-C chemokine receptor 7 (CCR7) and downregulation of CCR5. CCR7 binding to its ligands (CCL19 and CCL21) guides the migrating cells across interstitial tissues to the secondary lymphoid organs where adaptive immune response is initiated (Alvarez et al., 2008; Forster et al., 2008). The switch from an immature state to a mature state requires major alterations in the actin cytoskeleton of DC, thereby allowing the DC to migrate from the periphery to the lymphatic circulation or from the blood into tissues (Alvarez et al., 2008). A hallmark morphological marker of DC maturation is the loss of actin-rich, cylindrical adhesion structures known as podosomes (Burns et al., 2004). Although podosomes are generally accepted to be important in myeloid lineages for cell adhesion, and therefore motility, their exact role during DC migration remains unclear (Linder and Aepfelbacher, 2003). In human DC, podosomes often localize just behind the leading edge of the cells and form a link between the extracellular matrix and the cytoskeleton through integrins and adaptor proteins (Burns et al., 2001; 2004). Because podosomes limit fast migration by their strong interactions with the extracellular matrix, it is believed that DC need to dissolve their podosomes during maturation, e.g. after encountering microbial antigens or LPS (van Helden et al., 2006; 2010; Friedl and Weigelin, 2008; Lammermann et al., 2008).

Here, we have assessed the morphological and functional alterations in DC that follow invasion by T. gondii. We describe that a sequence of rapid morphological events accompany T. gondii invasion of DC, that concur with the onset of a hypermigratory phenotype, and that differ in time and characteristics from LPS-induced maturation.

Results

Toxoplasma infection induces motility-related morphological alterations in DC that differ in time from LPS-induced activation

We previously described that T. gondii-infected DC, but not LPS-matured DC, exhibit a hypermigratory phenotype in absence of chemotactic cues (Lambert et al., 2006). To determine the underlying cellular processes, we stained F-actin filaments and analysed the cytoskeletal morphology of human monocyte-derived DC following challenge with T. gondii (Fig. 1A). Tachyzoite-infected DC were consistently characterized by absence of podosomes and a rounded phenotype (Fig. 1A and B). When cells were assessed for morphological criteria – cell shape, podosomes, dendrite-like extensions and membrane veils/ruffles – we observed significant total score differences between tachyzoite-infected DC and non-infected DC or LPS-treated DC after 6–8 h (Fig. 1C). However, in line with previous reports (van Helden et al., 2006) prolonged exposure to LPS (24 h) led to morphological changes with podosome loss in DC (Fig. S1A and B). Next, a motility analysis was performed in order to address a possible link between the morphological changes observed and the previously described hypermotility in T. gondii-infected DC (Lambert et al., 2006). Tachyzoite-infected DC migrated significantly longer distances at higher velocities compared with non-infected DC or LPS-treated DC (Fig. 1D and E). These data show that DC challenged with T. gondii tachyzoites for 6–8 h undergo dramatic morphological changes accompanied by an increase in motility.

Figure 1.

Changes in DC morphology and induction of a hypermotile phenotype in response to T. gondii infection. Human monocyte-derived DC were challenged with freshly egressed RFP-expressing T. gondii tachyzoites (PRU-RFP, MOI 3), treated with LPS (100 ng ml−1) or maintained in complete medium (non-infected), for 6–8 h.

A. DC were stained with Alexa Fluor 488 Phalloidin to detect F-actin as indicated under Experimental procedures. In representative micrographs shown, arrows indicate actin condensation in podosome structures. Scale bar = 10 μm.

B. For each condition, graph shows percentage of cells (mean ± SEM) that exhibited absence of podosomes and rounded cell shape, respectively, related to the total cell population. For each condition, a total of 100 cells were analysed from 3 different donors. Asterisks indicate significant differences (*P < 0.001, **P < 0.00001, Student's t-test).

C. Cells were graded according to morphological criteria: the cell shape, the presence of podosomes, dendrite-like extensions and membrane veils/ruffles and scored as specified under Experimental procedures. Graph shows, for each condition, the distribution of the total scores obtained represented as percentage related to the total population. For each condition, a total of 100 cells from 3 donors were assessed. Significant total score differences were observed between tachyzoite-infected DC (median score ± SD: 5 ± 1.20) and non-infected DC (1 ± 1.09; P < 0.0001; Fisher's exact test) or LPS-treated DC (2 ± 0.96; P < 0.0001; Fisher's exact test), while differences between non-infected DC and LPS-treated DC were non-significant (P > 0.05; Fisher's exact test).

D. For each condition, cell track analyses were performed as outlined under Experimental procedures. Plots depict the motility plots (μm) of 100 single cells during 45 min from one representative donor.

E. Velocity analysis of cells under same conditions as (D). Data represent compiled analysis of 100 cells (± SEM) from 3 different donors. Asterisks indicate significant differences (*P < 0.001, **P < 0.00001, Student's t-test).

Time-course analysis reveals a rapid onset of morphological changes in DC following tachyzoite invasion

To further determine the kinetics of the morphological changes induced by T. gondii, the DC morphology was scored over time and related to the intracellular localization of tachyzoites (Fig. 2A). Strikingly, after 10 min incubation, the vast majority of the DC containing intracellular tachyzoites lacked podosomes (85.7 ± 8.1%) while only a minor portion of the non-infected by-stander DC population lacked these structures (8.2 ± 4.4%). Statistical analyses demonstrated a significant association between intracellular parasite localization and absence of podosomes in DC (P < 0.0001, Fisher's exact test; Fig. 2B and C). In line with this, a rapid dissolution of podosomes after parasite invasion was observed by time-lapse microscopy (Videos S1 and S2). To determine whether these rapid morphological changes were linked to a migratory activation, we performed a motility assay on DC after 10 min challenge with T. gondii tachyzoites. When compared with non-infected DC, as early as 10 min post infection, the infected DC exhibited a significant increase in their motility (Fig. 2D and E). This rapid onset of hypermotility was confirmed by time-lapse microscopy (Videos S3 and S4). When cell motility was analysed for 60 min, non-significant differences in motility were observed in DC assayed during the first and last 30 min (P > 0.05, Student's t-test; data not shown). We conclude that DC containing intracellular tachyzoites undergo rapid morphological changes that accompany the onset of a hypermotile phenotype.

Figure 2.

Rapid induction of morphological changes following T. gondii infection.

A. Human monocyte-derived DC were challenged with freshly egressed RFP-expressing T. gondii tachyzoites (PRU-RFP, MOI 10) for 10 min, fixed and stained with rabbit polyclonal anti-T. gondii tachyzoite antibodies (blue). Cells were washed, permeabilized and stained with Alexa Fluor 488 Phalloidin (green) to detect F-actin as indicated under Experimental procedures. Double-stained tachyzoites (red + blue) indicate extracellular localization (arrows) and only-red coloured tachyzoites indicate intracellular localization (arrowheads). Scale bar: 10 μm.

B. Frequency of podosomes in relation to intracellular or extracellular parasite localization in DC challenged with T. gondii for 10 min as in (A). For each category, the frequency of cell features related to the total population is indicated (% ± SD). Arrow indicates double-stained extracellular tachyzoite and arrowheads indicate only-red intracellular tachyzoites. A total of 150 cells from 2 donors were assessed.

C. Graphic representation of the distribution of morphological features (podosomes) of cells analysed as in (B). For each category, the frequency of each combination of features related to the total population is indicated (% ± SD). A significant association was observed between the conditions infection and absence of podosomes (P < 0.0001; two tailed Fisher's exact test).

D and E. Cell motility assay on DC in complete medium (D) and DC challenged with T. gondii tachyzoites (MOI 10) for 10 min (E). The histograms show the frequency of accumulated distances (μm) related to the total cell population, recorded during 60 min as indicated under Experimental procedures. Red lines indicate median migrated distance (11.4 and 28.4 μm respectively). Significant differences in migrated distances were observed between the two conditions (P < 0.0001; Mann–Whitney U-test). Insets show track analyses of individual cells. A total of 75 cells were tracked from one representative donor. Performed on multiple donors with similar results.

Tachyzoite invasion is required for induction of morphological changes and hypermotility of DC

To determine whether internalization of whole parasites or parasite fractions was sufficient to induce morphological changes and migratory activation, DC were exposed to heat-inactivated tachyzoites or soluble Toxoplasma antigen (STAg). Heat-inactivated tachyzoites were readily internalized and localized in LAMP1-expressing vacuoles (Fig. S2). The hypermotility phenotype was absent in DC containing heat-inactivated tachyzoites and in DC exposed to STAg (Fig. 3A–C). Discrete morphological changes, similar to those described for LPS-treated DC and distinct from those observed for Toxoplasma-infected DC (Fig. 1), were observed for both treatments (Fig. S3A and B). This indicates that phagocytic uptake of tachyzoites or exposure to soluble Toxoplasma antigens was not sufficient to induce the hypermotility phenotype and the morphological changes observed following active tachyzoite invasion.

Figure 3.

Active invasion of T. gondii is necessary to induce morphological changes and the onset of the hypermotile phenotype.

A. DC were challenged with freshly egressed T. gondii tachyzoites (PRU-RFP, MOI 5) heat-inactivated (H.I.) tachyzoites, STAg (10 μg ml−1) or complete medium (non-infected). After 4 h, the cell motility was assessed as outlined under Experimental procedures. Cell track analysis depicts the motility plots (μm) of single cells during 45 min from one representative donor.

B. Cell velocity of T. gondii-infected DC in relation to challenge with heat-inactivated tachyzoites. Bar graph shows average velocity (± SEM) of 75 cells analysed from one representative donor. Asterisks indicate significant differences (**P < 10−5; Student's t-test). NS indicate non-significant differences (P > 0.05; Student's t-test).

C. Cell velocity of T. gondii-infected DC in relation to challenge with STAg (10 μg ml−1). Performed as in (C) with a different donor.

D. Freshly egressed T. gondii tachyzoites (PRU-RFP) were pretreated with 4-BPB (0–10 μM) for 15 min. DC were challenged with pretreated tachyzoites (MOI 10) for 5 h whereafter cell velocity was assessed for 45 min as indicated under Experimental procedures. Bar graph shows average velocity (± SEM) of 75 cells analysed from one representative donor. Asterisks indicate significant differences versus 0 μM (*0.1 μM: P < 0.01; **1.0 μM and 10 μM: P < 10−5; Student's t-test).

E. Cell track analyses performed as in (D). For each condition, plots show track analyses of 75 individual cells from one representative donor.

Next, we determined whether the hypermotility and the morphological changes were triggered by binding of tachyzoites to the host cell surface, or whether active tachyzoite invasion was required. Assays were performed using actin polymerization inhibitors that block parasite invasion but allow discharge of secretory organelles, e.g. cytochalasin D. The observed effects on host cell cytoskeleton, and podosomes (van den Dries et al., 2013), precluded reliable analysis of DC morphology and motility (data not shown). Instead, we took advantage of the drug, 4-bromophenacyl bromide (4-BPB), that has been shown to specifically block invasion and rhoptry secretion dose-dependently, but not tachyzoite adhesion to the host cell membrane and microneme secretion (Ravindran et al., 2009). When DC were challenged with freshly egressed tachyzoites pretreated with 4-BPB, the onset of the hypermotility phenotype was inhibited dose-dependently (Fig. 3D and E). The majority of 4-BPB pretreated tachyzoites remained extracellular or in close contact with the host cell membrane. The portion of DC that contained intracellular tachyzoites, regardless of the 4-BPB concentration (Ravindran et al., 2009), exhibited morphological changes identical to DC infected with untreated tachyzoites, i.e. absence of podosomes and rounded morphology (Fig. S4). Altogether, we conclude that live intracellular parasites or the process of active parasite invasion per se, including discharge of secretory organelles, are necessary for induction of DC hypermotility.

De novo protein synthesis by the host cell or the parasite is not required for induction of morphological changes and hypermotility

To determine whether de novo protein synthesis was necessary for induction of the migratory phenotype, tachyzoites and DC were pretreated with protein synthesis inhibitors. DC were challenged with tachyzoites pretreated with the parasitostatic drug, pyrimethamine, for 24 h. Under these conditions, tachyzoites can invade host cells but do not replicate (Meneceur et al., 2008). Interestingly, DC containing intracellular pyrimethamine-treated tachyzoites exhibited hypermotility similar to DC challenged with non-treated tachyzoites (Fig. 4A and B). Further, DC pretreated with the protein synthesis inhibitor, cycloheximide, and challenged with T. gondii exhibited hypermotility for up to 6 h post infection, whereafter a decrease in motility was observed (Fig. 4C and D). Altogether, these data indicate that de novo protein synthesis in the invading tachyzoite and in the infected host cell is likely not required for the initiation of DC hypermotility.

Figure 4.

Effects of protein synthesis inhibitors on T. gondii-induced DC hypermotility.

A. Replicating tachyzoites (PRU-RFP) in HFF monolayers were treated with pyrimethamine (50 μM) for 24 h. DC were challenged with force-lysed pyrimethamine-treated tachyzoites (T. gondii + Pyr, MOI 5), freshly egressed tachyzoites (T. gondii, MOI 1) or, maintained in complete medium (non-infected). For DC challenged with treated and untreated tachyzoites, cell track analysis was performed on RFP-expressing DC. A total of 25 cells from one representative donor were tracked for 60 min.

B. Bar graph shows average velocity (± SEM) of 25 cells analysed from one representative donor as in (A). Asterisks indicate significant differences (**P < 10−5, Student's t-test). NS, non-significant.

C. DC were pretreated with cycloheximide (CHX, 1 μg ml−1) for 1 h and then challenged with freshly egressed tachyzoites (PRU-RFP, MOI 3) in the presence of CHX. Bar graph shows average velocity (± SEM) of 50 cells analysed at indicated time points from one representative donor as in (A). Asterisks indicate significant differences (**P < 10−5, Student's t-test). NS, non-significant.

D. Cell track analysis plots after 4 h incubation performed as in (C). Compiled tracking data of 100 cells is from 2 different donors.

Distribution of integrins and podosome dynamics in DC challenged with T. gondii

The distribution of integrins at the cell membrane and the podosome dynamics are tightly linked to the migratory properties of DC (Barreiro et al., 2007; Monypenny et al., 2011). We therefore assessed the distribution of the integrin subunits CD11c and CD18 in DC. Upon challenge with T. gondii, a redistribution of F-actin intensity staining to the edges of the cell membrane was observed, coinciding with disappearance of podosomes and appearance of membrane veils (Fig. 5). In infected DC, the F-actin staining and the β2 integrin CD18 staining concentrated to the periphery of the cell membrane (Fig. 5A) while the αX subunit CD11c staining appeared most prominent in veil structures (Fig. 5B). Co-staining of vinculin and F-actin showed characteristic podosome morphology in non-infected DC and confirmed the dissolution of podosomes in Toxoplasma-infected DC (Fig. 5C).

Figure 5.

Redistribution of integrin and actin staining following T. gondii infection.

A and B. DC were seeded on poly-l-lysine-coated coverslips and challenged with freshly egressed tachyzoites (MOI 3) for 4 h. Cells were stained with Alexa Fluor 488 Phalloidin and CD18 (A) or CD11c (B) monoclonal antibodies as indicated under Experimental procedures. T. gondii was stained with rabbit polyclonal anti-T. gondii tachyzoite antibodies and anti-rabbit Alexa 350 (blue), indicated by arrows. In non-infected DC, F-actin in podosome structures is observed. Scale bar: 10 μm.

C. DC were stained with vinculin monoclonal antibody and Alexa Fluor 488 Phalloidin as indicated under Experimental procedures. Upper panel: for each condition, the inset image shows a magnification of the podosome structures observed in the boxed area of the micrograph. Lower panel: arrows indicate T. gondii tachyzoites stained as in (A). Scale bar: 10 μm.

In LPS-matured DC, podosome dissolution was linked to the secretion of the prostaglandin PGE2 and reverted by inhibition of the synthesis enzyme cyclooxygenase (COX-1 and 2) (van Helden et al., 2010). To assess the role of this pathway in Toxoplasma-induced cytoskeletal remodelling, DC were pretreated with indomethacin (COX-1 and −2 inhibitor) and nimesulide (COX-2 inhibitor). Interestingly, high-dose pretreatment of DC with inhibitors and presence of inhibitors during the assay did not reverse the dissolution of podosomes following DC invasion by T. gondii, while an inhibition was observed for LPS-treated DC as previously reported (van Helden et al., 2010) (Fig. 6A and B). Also, the motility and transmigration of infected DC was not reduced by COX inhibition (Fig. 6C and D). We conclude that morphological changes induced by T. gondii were not inhibited by agents that stabilize podosomes in LPS-treated DC. As signalling via cell surface-located TLR4 did not seem to mediate the observed changes, we assessed the effects of endosomal TLR activation. DC treated with the TLR3 agonist Poly (I:C) or with the TLR7/8 agonist CL075 exhibited absence of hypermotility or morphological changes (Fig. S5 and data not shown).

Figure 6.

Actin cytoskeleton remodelling in DC after T. gondii infection is independent of prostaglandin signalling and GABAergic signalling.

A and B. DC were seeded on poly-l-lysine coated coverslips, pretreated with indomethacin (Indo, 50 μM) or nimesulide (Ni, 32 μM) for 1 h and then challenged with freshly egressed tachyzoites (PRU-RFP, MOI 3) or treated with LPS (50 ng ml−1) for 4 h (A) or 16 h (B). Podosomes were visualized by staining with Alexa Fluor 488 Phalloidin and assessed as indicated under Experimental procedures. Bar graphs show the frequency of podosome loss related to the total population of 50 cells (± SEM) from 2 donors. NS indicates non-significant differences (P > 0.05, Student's t-test) and asterisk indicates significant differences (*P < 0.05, Student's t-test).

C. DC were treated as in (A) and average cell velocity (± SEM) of 50 cells from 2 donors was analysed as indicated under Experimental procedures. Asterisks indicate significant differences (**P < 0.001, Student's t-test).

D. DC were pretreated with indomethacin and then challenged with freshly egressed tachyzoites for 6 h. Transmigration of DC was assessed in a transwell system after 16 h as indicated under Experimental procedures. Bar graph shows transmigration frequency (% ± SEM) related to total added cells. Performed in triplicate from one representative donor. NS indicates non-significant differences (P > 0.05, Student's t-test) and asterisk indicate significant difference (*P < 0.05, Student's t-test).

E. DC were pretreated with GABA (1 μM), GABAergic inhibitors semicarbazide (SC, 50 μM), SNAP 5114 (SNAP, 50 μM) for 30 min and then challenged with freshly egressed tachyzoites (PRU-RFP, MOI 3). After 4 h, F-actin was stained with Alexa Fluor 488 Phalloidin and podosomes were assessed as indicated under Experimental procedures. Bar graphs show the frequency of podosome loss related to the total population of 50 cells (± SEM) from 2 donors. NS indicates non-significant differences (P > 0.05, Student's t-test).

Next, we assessed the impact of GABAergic signalling in the morphological changes observed in infected DC. We have recently reported that inhibition of GABAergic signalling abrogates hypermotility of DC (Fuks et al., 2012). In presence of γ-aminobutyric acid (GABA), no morphological effects were observed in DC and GABAergic inhibition did not abrogate morphological changes in DC (Fig. 6E). Thus, while hypermotility was abrogated by GABAergic inhibition, the morphological changes observed in infected DC were not. Altogether this indicates that upon T. gondii infection, the actin cytoskeleton remodelling appears independent of prostaglandin signalling and GABAergic signalling.

Toxoplasma-infected DC acquire chemotactic properties

We have previously shown that parasite-induced hypermotility of DC does not depend on chemotactic stimuli (Lambert et al., 2006). In agreement, upon screening a panel of chemokine receptors, we found that the expression levels of CCR1, CCR2, CCR6, CCR9, CCR10 and CXCR6 remained chiefly unaffected in T. gondii-infected DC by 6 h, i.e. after the onset of hypermotility and morphological changes (Fig. S6). Yet, migratory responses in mature DC are often directed by chemotactic cues, e.g. CCR7 ligands. We therefore assessed the chemotactic responses of DC challenged with T. gondii over time. In presence of LPS, DC exhibited a distinct chemotactic response to CCR7 ligand (CCL19) after 12 h, while Toxoplasma-infected DC acquired this chemotactic response by 24 h (Fig. 7A). Interestingly, a chemotactic response was observed in Toxoplasma-infected DC (RFP+), but not in by-stander non-infected DC (RFP) (Fig. 7A, lower panels). To address if chemokine receptors were modulated by the infection over time, we assessed the expression of CCR7 and CCR5. A time-dependent upregulation of CCR7 was observed in infected DC (RFP+) and a downregulation of CCR5 was observed in both infected DC (RFP+) and by-stander DC (RFP) (Fig. 7B and C). This pattern was consistently observed in the 5 human donors tested (Fig. 8). Interestingly, and consistent with the chemotactic responses, a significant upregulation of CCR7 was observed in infected (RFP+) DC by 24 h, but not in by-stander DC (RFP) (Fig. 8B). In LPS-treated DC, a significant upregulation of CCR7 was present from 6 h (Fig. 8A) and a clear chemotactic response was observed by 12 h. A rapid downregulation of CCR5 was observed in LPS-treated DC (Fig. 8C). For DC challenged with T. gondii, a significant downregulation of CCR5 was observed for both infected (RFP+) and non-infected DC (RFP) compared with non-treated cells (P < 0.0001, paired t-test). Infected DC exhibited a tendency to lower expression of CCR5 compared with non-infected by-stander DC with non-significant differences (P > 0.05, paired t-test) (Fig. 8D). We conclude that CCR5 and CCR7 expression is modulated in Toxoplasma-infected DC but follows a different kinetics compared with LPS-induced maturation. Altogether, this shows that Toxoplasma-infected DC acquire the ability to respond chemotactically to CCL19 in vitro and that the onset of parasite-induced hypermotility precedes in time the chemotactic responses of infected DC in vitro.

Figure 7.

Induction of chemotaxis is distinct from the onset of hypermotility. DC were challenged with freshly egressed tachyzoites (PRU-RFP; 6 h: MOI 3; 12 and 24 h: MOI 1) or treated with LPS (200 ng ml−1) and subjected to a chemotaxis assay (A) or flow cytometry (B) to determine the cell surface expression of CCR7 or CCR5 as indicated under Experimental procedures.

A. Chemotaxis tracking analyses of DC. DC in complete medium (non-infected), DC exposed to LPS (200 ng ml−1) or challenged with freshly egressed RFP-expressing tachyzoites (T. gondii) were placed in chemotaxis chambers with a CCL19 gradient and tracked for 60 min as described under Experimental procedures. In the bottom panels, cells were separated into RFP+ (red tracks) and RFP (black tracks). For each condition, percentages represent the proportion of cells migrating towards chemoattractant in complete medium (CCL19) or complete medium (CM) respectively. The tracking data are a composite of two separate experiments (2 donors) performed in duplicate (n = 75 cells).

B. Histograms show CCR7 and CCR5 expression of DC treated as in (A). Flow cytometry was performed at indicated time points as described under Experimental procedures. Each time point is indicated by coloured lines (grey 6 h, black 12 h, red or blue 24 h). In each histogram, thick coloured lines represent CCR7 or CCR5 stainings and thinner lines of the same colour represent corresponding isotype control stainings. Displayed data are representative from 4 independent experiments and 5 donors (CCR7) or 4 donors (CCR5).

C. Density plots show CCR7 and CCR5 expression of DC challenged with RFP-expressing T. gondii (PRU-RFP, MOI 1, 24 h). The upper and lower panels show the expression of CCR7 and CCR5, respectively, in the RFP+ (infected) and RFP (non-infected) DC populations. Displayed data are representative of 4 independent experiments.

Figure 8.

Expression of CCR7 and CCR5 in DC challenged with T. gondii, over time, among human donors. DC were challenged with freshly egressed tachyzoites (PRU-RFP, MOI 1) or treated with LPS (200 ng ml−1). At indicated time-points, the cell surface expression of CCR7 (A and B) and CCR5 (C and D) was determined by flow cytometry as indicated under Experimental procedures. For all conditions, each symbol series represents one individual human donor (n = 4–5).

A and C. LPS-treated samples are indicated by the closed symbols. Mock-treated samples in complete medium are indicated by the open symbols.

B and D. For DC challenged with T. gondii, cells were separated by RFP positivity, with the infected DC (RFP+) indicated by the closed symbols and the by-stander DC (RFP) indicated by the open symbols. Closed and open bar symbols indicate average MFI for the corresponding sample group. Asterisks indicate significant differences (*P < 0.05; **P < 0.01; paired t-test).

Discussion

We here report that a hypermotility phenotype is induced in immature human monocyte-derived DC very shortly after invasion by T. gondii. Our data also show that infection with T. gondii led to a surprisingly rapid morphological transformation of DC, setting in within minutes after invasion. This comprised of the disappearance of podosome structures and the appearance of a rounded cellular morphology with veils, ruffles, filopodia and lamellipodia. Invasion of DC by T. gondii induced morphological changes and migratory activation significantly more rapidly than LPS, even in the presence of chemokine. As a consequence of T. gondii infection, DC acquired features consistent with high-speed ameboid or interstitial migration in vitro (Friedl and Weigelin, 2008). Altogether, this indicates a link in time between parasite invasion, cytoskeletal remodelling and the onset of hypermotility in DC.

The present data demonstrate that the process of active tachyzoite invasion is a requisite for induction of hypermotility and morphological changes in DC. The attachment of parasites to the cell surface, phagocytic uptake of tachyzoites or exposure to STAg was not sufficient to induce these features. Further, pretreatment of parasites with 4-BPB, an inhibitor of rhoptry secretion and also of invasion, dose-dependently (Ravindran et al., 2009) abrogated their onset. In contrast, pretreatment of parasites with protein synthesis inhibitors did not abolish the induction of hypermotility or morphological changes in DC. Altogether, this may indicate that the process of active invasion with the discharge of secretory organelles is crucial for the induction, and that pre-synthesized effector molecules may mediate these effects in DC. This assumption is also in line with the observed rapid onset of hypermotility and cytoskeletal remodelling within minutes after parasite invasion. The effects of actin-targeting inhibitors of parasite invasion on the host cell cytoskeleton hampered a distinct discrimination between the role of invasion and the discharge of secretory organelles. Thus, further studies need to determine if, e.g. rhoptry secretion in absence of invasion is sufficient to induce these events, or if intracellular parasite localization is required. The perseverance of the hypermotility phenotype in DC throughout the infection (> 24 h) (Lambert et al., 2006; Fuks et al., 2012) poses the question whether maintained secretion after invasion is required. Also, the inability of heat-inactivated tachyzoites and STAg to induce detectable phenotypic effects indicates that the form of delivery of effector molecules into the host cell may be crucial. Further, infected DC exhibited hypermotility after inhibition of protein synthesis for up to 6 h, indicating that de novo protein synthesis in the host cell is not necessary to initiate or temporally maintain the hypermotility phenotype. Altogether, present data are consistent with the notion that T. gondii exerts important functions in host cells through secreted effector proteins during and after invasion (Hunter and Sibley, 2012).

Shortly after invasion, dramatic morphological changes were observed in infected DC with loss of podosomes being a prominent and easily quantifiable parameter. Because podosomes limit fast migration by their strong interactions with the extracellular matrix, the presence of podosomes appears incompatible with the high-speed ameboid migration observed in mature DC (De Vries et al., 2003; van Helden et al., 2006; 2010). This amoeboid motility of DC occurs independently of adhesion to specific substrates and extracellular matrix degradation (Lammermann et al., 2008) and is required for efficient migration (Calle et al., 2006). Thus, in line with our observations using live imaging, the dissolution of podosomes and the switch in the mode of motility in Toxoplasma-infected DC needs to be tightly controlled for proper migration. Also, we previously reported that integrin expression levels in infected DC are maintained or down-modulated (Lambert et al., 2006). Consistent with the above, here we observed a more prominent integrin staining at the edges of infected DC. Thus, the podosome dissolution, the rounding-up of DC, the integrin redistribution and the hypermotility observed in DC very shortly after T. gondii invasion are well in line with features exhibited by DC during high-speed ameboid migration.

A role for prostaglandins in podosome dissolution during LPS-mediated DC maturation has been described (van Helden et al., 2006; 2008; 2010). As prostaglandins do not normally exist preformed in any cellular reservoir, when cells are activated, prostaglandin E2 (PGE2) is synthesized de novo and released into the extracellular space to activate receptors on the cell membrane (Kalinski, 2012). Interestingly, in our system, the induction of morphological changes, e.g. podosome dissolution, was not reversed by inhibitors of the prostaglandin synthesis enzymes COX-1 and COX-2, as reported for TLR-4-mediated activation of DC exposed to LPS (van Helden et al., 2006). Also, in sharp contrast to infected DC, by-stander DC exhibited normal podosome frequency and supernatants from infected DC did not affect podosome frequency in DC. Altogether, this and the very rapid onset (minutes) advocates against a role for secreted prostaglandins as the main cause for podosome dissolution. It also indicates that alternative pathways may be induced by T. gondii. In fact, other actin regulatory pathways may affect the formation and turnover of podosomes, but the precise regulatory mechanisms remain unknown (van Helden et al., 2008). Future approaches need to address the role of key components in DC migration, e.g. WASp, N-WASp, Vav1, also in the context of their partly redundant functions, and the impact of the present observations on infections in vivo (de Noronha et al., 2005; Spurrell et al., 2009; Isaac et al., 2010; Murphy and Courtneidge, 2011).

We recently described that GABA secreted from Toxoplasma-infected DC provides a potent motogenic effect and mediates hypermotility through GABAergic signalling pathways (Fuks et al., 2012). Thus, while GABAergic inhibition efficiently abrogates hypermotility, here we report that exogenous GABA or GABAergic inhibition does not significantly affect podosome frequency or cell morphology exhibited by non-infected or infected DC. This indicates that GABAergic signalling per se does not directly mediate the morphological changes observed in infected DC and may either operate downstream of the initial signalling following Toxoplasma invasion or constitute a separate pathway. Yet, the two phenomena are linked in time, setting in within minutes after parasite invasion, and may jointly contribute to the rapid migratory activation of the infected DC. The identification of parasite-derived effector molecules awaits further investigation.

The finding that Toxoplasma-induced hypermotility of DC sets in within minutes after infection also corroborates that the hypermigratory phenotype does not depend on classical chemotactic activation, e.g. in LPS-matured DC (Lambert et al., 2006; Fuks et al., 2012). Interestingly, and not in contraposition, infection of DC by T. gondii led to an upregulation of CCR7 (Fuks et al., 2012). Here, we extend these studies further and report that a distinct downregulation of CCR5 and an upregulation of CCR7 were observed for the vast majority of human donors tested. In line with the previously described upregulation of co-stimulatory molecules and cell surface maturation markers in infected DC (Lambert et al., 2006), this indicates that infection also leads to maturation events that make the DC responsive to chemotactic cues. Importantly, upregulation of CCR7 was not observed in by-stander DC, indicating a direct effect of intracellular parasite localization and that activation required prolonged exposure. In contrast, a downregulation of CCR5 was observed in both infected DC and by-stander DC shortly after exposure to T. gondii. A hypothetical explanation for this general down-modulatory effect may be related to T. gondii cyclophilin (C-18), that is released by extracellular tachyzoites and that acts as a chemokine mimic by binding to CCR5 (Aliberti et al., 2000; 2003). In contrast to the observed relatively rapid (6 h) down-modulation of CCR5, the expression levels of a number of chemokine receptors screened here remained chiefly unaffected by T. gondii infection after the onset of morphological changes and hypermotility. Yet, a modulation of additional chemokine receptors over time or synergistic effects of multiple chemokine receptors (John et al., 2011) cannot be excluded and awaits further investigation. Also, we have previously shown that Toxoplasma-induced hypermotility occurs independently of MyD88-mediated TLR signalling in DC (Lambert et al., 2006). Here, we add to this line of evidence that MyD88-independent TLR3 activation did not mediate hypermotility in DC.

We provide evidence that Toxoplasma-induced hypermotility of DC and chemotactic migration of infected DC are two mechanisms that can be distinctly separated in time in vitro. Hypermotility set in within minutes after parasite invasion while chemotactic response to CCL19 set in earliest 12–24 h after infection or LPS stimulation. Importantly, after the onset of the chemotactic response, hypermotility and chemotaxis worked in conjunction, potentiating speed and directional motility of infected DC in vitro. Our results may have implications for how we conceive dissemination of T. gondii. Hypothetically, GABA/GABAA receptor-mediated hypermotility and CCR7-mediated chemotaxis could cooperatively enhance the migratory potential of infected DC in vivo, and consequently, potentiate the dissemination of the parasite. This is well in line with the observed enhanced dissemination upon adoptive transfers of infected DC (Lambert et al., 2006; 2011; 2009) and the inhibitory effects on DC migration by pertussis toxin treatment (Lambert et al., 2006; John et al., 2011) or GABAergic inhibition (Fuks et al., 2012). Similarly, in malignancies, several types of metastasizing cells can also express CCR7, that contributes to their dissemination (Mburu et al., 2006; Raman et al., 2007).

The strategies for dissemination described above are likely not limited to T. gondii. Recently, we described that another member of the apicomplexan parasite family, Neospora caninum, induces hypermotility of DC with an impact on parasite dissemination in vivo and on transplacental passage (Collantes-Fernandez et al., 2012). In contrast to the hypermotility phenotype observed in DC and macrophages for T. gondii and N. caninum, the motility of macrophages infected by the apicomplexan Theileria annulata was reduced in vitro and persistence of podosomes was observed (Baumgartner, 2011). This indicates that Apicomplexa may target the migratory machinery of host cells via different mechanisms. Importantly, these effects are not restricted to protozoa. Microbial pathogens manipulate the host cell cytoskeleton for adherence, uptake, motility through the cytoplasm, exit and dissemination (Gruenheid and Finlay, 2003; Worley et al., 2006). Likewise, viruses of the pox family can induce cell migration and induce membrane projections in infected host cells (Sanderson and Smith, 1999). In this context, we here provide evidence that infection of DC by T. gondii tachyzoites initiates a series of regulated cellular events that promote the migratory activation of the host cell. Elucidation of the molecular effector mechanisms and host cell targets may provide new insights in how intracellular pathogens manipulate host cells to establish infection and propagate in their hosts.

Experimental procedures

Parasites and cell lines

Tachyzoites from the T. gondii PRU-RFP-expressing line (type II) (Pepper et al., 2008) were maintained by serial 2-day passaging in human foreskin fibroblast (HFF) monolayers. HFFs were propagated in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) with 10% fetal bovine serum (FBS), gentamicin (20 μg ml−1; Gibco), glutamine (2 mM; Gibco) and Hepes (0.01 M; Gibco), referred to as complete medium (CM).

Primary cells

To generate human DC, buffy coats from healthy blood donors were incubated with Monocyte Enrichment cocktail (RosetteSep™, StemCell Technologies), followed by centrifugation on Lymphoprep (Axis-Shield PoC AS) as previously described (Lambert et al., 2006). The cell population obtained was composed mainly of CD14+ (DakoCytomation) with < 1% CD3+/CD19+ cells (BD), as evaluated by flow cytometry (FACSCalibur; BD). DC were generated by culturing the cells in DMEM media supplemented with 100 ng ml−1 GM-CSF (PeproTech) and 20 ng ml−1 IL-4 (PeproTech) for 6 days. Fresh cytokines were added after 3 days in culture. The Regional Ethics Committee, Stockholm, Sweden, approved protocols involving human cells. All donors received written and oral information upon donation of blood at the Karolinska University Hospital. Written consent was obtained for utilization of white blood cells for research purposes.

Reagents

All chemicals were purchased from Sigma Aldrich unless otherwise stated. The following chemicals were used: cytochalasin D (1 μM), 4-Bromophenacyl bromide (4-BPB, as indicated), LPS (as indicated), pyrimethamine (50 μM), cycloheximide (1 μg ml−1), indomethacin (50 μM), nimesulide (32 μM), semicarbazide (50 μM), SNAP 5114 (50 μM), GABA (1 μM), Poly (I:C) (1 μg ml−1), CL075 (1 μg ml−1, Invivogen). Heat-inactivated (H.I.) parasites were generated by incubating freshly egressed tachyzoites at 56°C for 30 min. Soluble Toxoplasma antigen (STAg) was prepared as previously described (Aliberti et al., 2000).

Immunocytochemistry and intracellular/extracellular parasite stainings

Host cells and parasites were cultured on poly-l-lysine-coated glass coverslips (Sigma). Fixation was performed with 0.3% glutaraldehyde (Sigma) in BRB80 buffer (80 mM PIPES, pH 6.9; 1 mM MgCl2; 1 mM EGTA) for 10 min at room temperature (RT). The cells were permeabilized using 1% PBS-Triton X-100 (PBST, Sigma). To visualize host cell F-actin and podosomes, cells were stained with Alexa Fluor 488-conjugated phalloidin (Invitrogen) and vinculin monoclonal antibody (Sigma). For detection of integrins, cells were processed as described (Burns et al., 2004). Briefly, after fixing in 4% paraformaldehyde in PBS/3% glucose for 15 min at RT, cells were permeabilized using 0.5% Triton X-100 in PBS for 5 min and blocked with 1% BSA in PBS for 30 min at RT. Integrins were stained with CD11c or CD18 monoclonal antibody (BD Pharmingen) for 60 min, followed by Alexa 594 (Invitrogen) secondary antibody. T. gondii tachyzoites were stained with rabbit polyclonal anti-T. gondii tachyzoite antibodies (R14; a gift from E. Linder, Statens Bakteriologiska Laboratorium, Solna, Sweden), followed by Alexa 350 (Invitrogen) donkey anti-rabbit secondary antibody.

To differentiate between intracellular and extracellular tachyzoites, cells were washed twice with PBS and incubated with rabbit polyclonal anti-T. gondii tachyzoite antibodies (R14) followed by anti-rabbit Alexa Fluor-conjugate (Invitrogen). After washing, cells were permeabilized with 1% PBS-Triton X-100 for 10 min, and incubated with Alexa Fluor 488-conjugated Phalloidin (Invitrogen). Lysosomes/endosomes were stained with rabbit polyclonal LAMP1 antibodies (Abcam). Coverslips were mounted using mounting medium with or without DAPI (Vectashield, Vector Laboratories) and assessed by epifluorescence microscopy (Leica DMRB) or structured illumination microscopy (Zeiss Imager.Z1 equipped with ApoTome). The images were processed using Adobe Photoshop 7.0.

Scoring of cell morphology

Phalloidin-stained DC were monitored by epifluorescence microscopy (Leica DMRB; 100× objective) equipped with a CCD camera (Retiga EXi, Qimaging). For each preparation, micrographs (Openlab version 5.0.2) from 20–30 randomly chosen fields of view were acquired for assessment. An average of 100 cells from each individual donor was graded and verified independently by two microscopists. The DC were scored based on four criteria (A–D):

  1. Cell shape – elongated (score 0) versus rounded (score 1).
  2. Podosome structures – present (score 0) versus absent (score 1).
  3. Distinct dendrite-like extensions – absent (score 0), present on one-third of the cell surface (score 1), present on two-thirds of the cell surface (score 2), present on the entire cell surface (score 3).
  4. Presence of membrane veils and/or ruffles – same criteria as with (C).

A final score resulted in the addition of all four (A–D) criteria and therefore could range from 0–8.

Motility assays

DC (3–5 × 104) were challenged with freshly egressed tachyzoites or treated with LPS as indicated. Bovine collagen I (0.75 mg ml−1, Invitrogen) was added to the cells and the cell/collagen mixture was immediately loaded onto a labtech chamber slide (Nalge Nunc International) where the medium chambers were removed, but gaskets were left intact. A glass coverslip was placed on top of the gaskets, and the slide was spun at 1000 rpm for 5 s to concentrate the cells to the bottom of the slide. The cells were imaged every min for 45–60 min (Zeiss AxioImager). Motility patterns were compiled using ImageJ (image stabilizer software and manual tracking plugins). For the 4-BPB studies, tachyzoites were pretreated with increasing concentrations of 4-BPB for 15 min. The drug was removed by washings, as previously described (Ravindran et al., 2009), before adding tachyzoites to DC.

Transmigration assays

Infection of cells and quantification of migrated cells were conducted as previously described (Lambert et al., 2006). Briefly, DC were plated at a density of 1 × 106 cells/well and incubated with freshly egressed T. gondii tachyzoites (MOI 3) for 4 h at 37°C and 5% CO2. DC were then transferred into transwell filters (8 μm pore size; BD biosciences) and incubated for 16 h at 37°C and 5% CO2. Migrated DC were quantified in a hematocytometer.

Chemotaxis assays

Chemotaxis assays were performed using μ-slide 3Dchemotaxis slides (Ibidi) with 1.5 mg ml−1 Collagen I and following the manufacturer's instructions and as previously described (Fuks et al., 2012). To establish a gradient, 1.5 mg ml−1 CCL19 (Peprotech) was used. Cell migration was monitored every min for 1 h (Zeiss AxioImager). Cell tracking and chemotaxis analysis were performed using ImageJ (Manual Tracking and Chemotaxis Tool plugins).

Flow cytometry

Chemokine receptor expression of human monocyte-derived DC was assessed using FITC-labelled anti-CCR7, APC-labelled anti-CCR5, CCR1, CCR2, CCR6, CCR9, CCR10, CXCR6 antibodies and mouse isotype control antibodies (all from R&D Systems). Samples were analysed using a CyAn ADP (Beckman Coulter) or a FACSCalibur (BD Bioscience) flow cytometer and the FlowJo software 9.2 (Tree Star Inc., Ashland, OR).

Transduction of DC

GFP actin retrovirus was produced by transiently transfecting Phoenix Eco 293T packaging cell line with LZRS-EGFP-actin (Kinsella and Nolan, 1996). The supernatant of the packaging cells was harvested at 48 and 72 h and filtered through a 0.8 micron syringe filter. The virus particles were concentrated by centrifugation (6000 g ON at 4°C) after which 80–90% of the supernatant was discarded and the virus resuspended in the remaining media. Cells from bone marrow of C57BL/6 mice were grown in RPMI containing 10 ng ml−1 recombinant mouse GM-CSF (Peprotech). On day 2 and day 3, the DC culture media was replaced with concentrated virus supernatant and 8 μg ml−1 polybrene, followed by centrifugation for 1 h at 2000 rpm (wide rotor) at 32°C. Cells were then incubated in 5% CO2 at 37°C for 20 min and the virus supernatant was replaced with DC media. DC were used in assays on day 6–7. The transduction frequency was 50–80%.

Time-lapse microscopy

GFP-actin-transduced DC seeded in a 96-well imaging plate (BD) were placed in a 37°C environment chamber enclosing the microscope base of a Deltavision Spectris live cell imaging system (Applied Precision, GE healthcare) and imaged with a 60× high NA oil objective. Imaging started within 2–3 min of adding freshly egressed RFP-expressing tachyzoites (PRU-RFP). Three optical sections were acquired per field and time point in a time-lapse of 60 s between time points. Image deconvolution was applied using the systems image restoration algorithms and maximum projections were made of the optical planes (softWoRx v.2.0, GE healthcare). The image sequences were contrast adjusted and the movies were generated using ImageJ software. The time-lapse is played at a speed of 3 frames per second in the movies.

Statistical analyses

The statistical analyses were performed using R Stats Package version 2.14.1 (R Foundation for Statistical Computing, Vienna, Austria).

Acknowledgements

This study was supported by grants from the Swedish Research Council (to A.B.). J.M.W. is the recipient of a postdoctoral fellowship from the Wenner Gren Foundation. M.A.H.C. is the recipient of a Francisco Jose de Caldas stipend, Institute for the Development of Science and Technology (COLCIENCIAS), administered by the Academic and Professional Program for the Americas, Harvard University (LASPAU).

Competing interests

The authors have declared that no competing interests exist.

Ancillary