Voltage-dependent calcium channel signaling mediates GABAA receptor-induced migratory activation of dendritic cells infected by Toxoplasma gondii

The obligate intracellular parasite Toxoplasma gondii exploits cells of the immune system to disseminate. Upon T. gondii-infection, γ–aminobutyric acid (GABA)/GABAA receptor signaling triggers a hypermigratory phenotype in dendritic cells (DCs) by unknown signal transduction pathways. Here, we demonstrate that calcium (Ca2+) signaling in DCs is indispensable for T. gondii-induced DC hypermotility and transmigration in vitro. We report that activation of GABAA receptors by GABA induces transient Ca2+ entry in DCs. Murine bone marrow-derived DCs preferentially expressed the L-type voltage-dependent Ca2+ channel (VDCC) subtype Cav1.3. Silencing of Cav1.3 by short hairpin RNA or selective pharmacological antagonism of VDCCs abolished the Toxoplasma-induced hypermigratory phenotype. In a mouse model of toxoplasmosis, VDCC inhibition of adoptively transferred Toxoplasma-infected DCs delayed the appearance of cell-associated parasites in the blood circulation and reduced parasite dissemination to target organs. The present data establish that T. gondii-induced hypermigration of DCs requires signaling via VDCCs and that Ca2+ acts as a second messenger to GABAergic signaling via the VDCC Cav1.3. The findings define a novel motility-related signaling axis in DCs and unveil that interneurons and DCs share common GABAergic motogenic pathways. T. gondii employs GABAergic non-canonical pathways to induce host cell migration and facilitate dissemination.


Introduction
The obligate intracellular parasite Toxoplasma gondii chronically infects a large portion of the global human population and is capable of infecting any warm-blooded vertebrate [1]. The dissemination of the parasite from the point of entry in the intestinal tract plays a determinant role in the pathogenesis of toxoplasmosis. Although chronic infection is generally considered asymptomatic in otherwise healthy individuals, reactivated infection in the central nervous system (CNS) of immune-compromised individuals may be fatal. Congenital toxoplasmosis occurs by transmission to the fetus from the infected mother and can result in serious disabilities or death of the unborn child [2].
Previous studies have demonstrated that active invasion of dendritic cells (DCs) by T. gondii tachyzoites rapidly (within minutes) induces a hypermigratory phenotype in parasitized DCs [3]. This migratory activation is characterized by cytoskeletal rearrangements and dramatically enhanced cellular locomotion, termed hypermotility [4], and enhanced transmigratory activity in vitro [5]. These phenotypes have been linked to enhanced dissemination and parasitic loads in mice for different species of apicomplexan parasites [5][6][7]. The initiation of the hypermigratory phenotype in DCs is related to the discharge of secretory organelles during parasite invasion and does not depend on de novo protein synthesis in the host cell [4]. It is mediated through non-canonical GABAergic signaling pathways, and is independent of MyD88-mediated TLR signaling and chemotaxis [3][4][5]8].
Within the context of the host-parasite interaction, we have recently shown that DCs possess functional GABA A receptors, and the capability to synthesize and secrete γ-aminobutyric acid (GABA) [8]. Challenge with T. gondii triggered GABA secretion in the invaded DCs and inhibition of GABA A receptors, GABA synthesis or GABA transport abrogated the T. gondiiinduced hypermigratory phenotype [8]. Along these lines, mounting evidence shows that GABA, the main inhibitory neurotransmitter in the vertebrate brain, participates outside the CNS in diverse functions including cell migration, immunomodulation and metastasis [9][10][11].
GABA A receptors are ionotropic chloride channels whose functions are regulated by cation-chloride co-transporters [12]. Membrane depolarization secondary to GABA receptor activation can elicit opening of voltage-dependent Ca 2+ channels (VDCCs, also termed voltage-gated Ca 2+ channels, VGCCs) that are normally closed at physiologic or resting membrane potential [13][14][15]. Thus, GABA-mediated Ca 2+ influx via VDCCs is a well-established concept in neuronal cells but remains unexplored in immune cells. While various Ca 2+ signaling pathways have been implicated in the regulation of multiple DC functions, including activation, maturation and formation of immunological synapses with T cells (reviewed in [16]), knowledge on the role of VDCCs in DCs remains limited [17,18].
Here, we show how an obligate intracellular pathogen takes advantage of a hitherto uncharacterized Ca 2+ signaling axis in DCs to modulate the migration of parasitized host cells. We demonstrate that the hypermigratory phenotype induced in DCs by T. gondii is predominantly dependent on the L-type VDCC subtype Ca v 1.3, which is activated by GABAergic signaling upon T. gondii invasion.

T. gondii-induced DC hypermotility and transmigration depend on physiological extracellular Ca 2+ concentrations
Shortly after parasite entry, DCs exhibit a dramatic migratory activation [5,8]. Based on the implication of GABAergic signaling and on the rapid onset of the hypermigratory phenotype minutes after parasite invasion [4], we hypothesized a role for Ca 2+ signal transduction for the induction and maintenance of T. gondii-induced migration. When DC motility was assessed in a Ca 2+ -deprived medium, individual cell track analysis of infected DCs showed a reduction in migrated distances ( Fig 1A) and a significant downward shift in the distribution of migrated distances (Fig 1C and 1D). In low Ca 2+ medium with 1% FBS, the median velocity of T. gondiiinfected DCs was significantly reduced, and addition of Ca 2+ (CaCl 2 ) at physiological concentration reconstituted hypermotility in infected DCs (Fig 1B). Similarly, base-line motility of non-challenged DCs was reduced upon Ca 2+ deprivation and reconstituted by addition of Ca 2+ (Fig 1B and S1 Fig). In line with motility assays, the relative transmigration frequencies of infected DCs, and non-infected DCs, across a transwell porous-membrane were significantly reduced in low extracellular Ca 2+ and reconstituted upon addition of Ca 2+ at physiological concentration ( Fig 1E). Next, we assessed the motility of infected DCs in the presence of NiCl 2 , which blocks plasma membrane Ca 2+ channels. NiCl 2 dose-dependently reduced the velocity of hypermotile infected DCs, that reached velocities comparable with non-infected DCs ( Fig 1F). Altogether, the present data indicate that the T. gondii-induced hypermigratory phenotype of DCs is dependent on the entry of extracellular Ca 2+ through plasma membrane Ca 2+ channels.

Stimulation of DCs with GABA elicits Ca 2+ influx transients in the DC cytosol
We have previously established that infection by T. gondii induces motility-related GABAergic signaling pathways in DCs [8]. Because hypermotile Toxoplasma-infected DCs exhibited dependency on Ca 2+ and the established links between GABA receptor activation and Ca 2+ responses in neuronal cellular systems [13,14], we tested whether GABA A receptor activation led to Ca 2+ responses in DCs. Perfusion of GABA elicited cytosolic Ca 2+ elevations in DCs, visualized by fluorescent Ca 2+ indicators (Fig 2A and S1 Video). Stimulation of DCs with GABA led to a simultaneous and transient Ca 2+ influx (Fig 2B and 2C) in~20% of the tested DC population at a given time point and, for the reference stimulus ATP, in~42% of DCs (S1 Table). Ca 2+ transients induced by GABA had relatively similar longevity and relatively lower amplitude than responses to ATP (Fig 2B and 2C), which were in line with ATP responses previously characterized in various types of DCs [19,20]. Upon repeated stimulations with GABA and at varying GABA concentrations, consecutive Ca 2+ responses were observed in individual cells (S2 Fig). Altogether, the data is in line with the previously recorded GABA-induced membrane potential changes by patch-clamping [8] and demonstrate that GABA stimulation of DCs is followed by influx of Ca 2+ and transiently increased cytosolic Ca 2+ concentration.   [21]. Inhibition of GABA synthesis and secretion (SC, SNAP inhibitors, respectively) selectively reduced the m/z 104,2 peak signal ( Fig 3A) and abrogated the hypermotility of T. gondiiinfected DCs, which was reconstituted by addition of exogenous GABA (Fig 3B and 3C). This provided further specificity to previously reported elevations of GABA secretion in Toxoplasma-infected DCs, as quantified by GABA-ELISA under the same conditions [8]. Next, to test the impact of Ca 2+ influx in DC hypermotility under GABA-deprived conditions, a cell membrane Ca 2+ channel/ L-type VDCC agonist (BayK8644) was added to the cells. Importantly, the abrogated hypermotility of infected DCs, generated by GABAergic inhibition, was rescued by addition of BayK8644 (Fig 3B and 3C). A moderate but significant increase in cell  motility was also observed in naïve DCs in presence of BayK8644 (Fig 3B and 3C). We conclude that, upon GABAergic inhibition, Ca 2+ channel (VDCC) agonism leading to Ca 2+ entry in DCs can reconstitute hypermotility in Toxoplasma-infected DCs.
Upon inhibition of voltage-dependent Ca 2+ channels (VDCCs), exogenous GABA fails to reconstitute hypermigration in Toxoplasmainfected DCs VDCCs are known to respond with Ca 2+ permeability to membrane potential changes. Because GABA A receptor activation by GABA elicits membrane potential changes in Toxoplasma-infected DCs [8] and GABA elicited Ca 2+ influx (Fig 2), we investigated the putative involvement of VDCCs in DC hypermotility. L-type VDCC inhibition by nifedipine abolished hypermotility ( Fig 4A) and significantly reduced transmigration ( Fig 4B). In sharp contrast, inhibition of purinergic Ca 2+ receptors by PPADS at high concentrations [22] had non-significant effects on hypermotility and transmigration of infected DCs (Fig 4A and 4B), despite that activation of purinergic receptors by ATP caused a significant Ca 2+ influx and increased cytosolic Ca 2+ levels in DCs (Fig 2). This indicated that VDCC-related effects governed hypermigration. We therefore explored further the function of VDCCs in relation to GABAergic signaling. We previously reported that inhibition of GABA synthesis (SC) and/or transport (SNAP) significantly reduced GABA secretion and transmigration of T. gondii-infected DCs [8]. Extending these observations, addition of exogenous GABA rescued the hypermotility of infected DCs under GABAergic inhibition (Fig 4C and 4D). In sharp contrast, VDCC inhibition by nifedipine treatment caused a significant decrease in the motility of infected DCs that was not restored by exogenous GABA (Fig 4C and 4D), indicating implication of L-type VDCCs downstream of GABAergic signaling.
At resting membrane potential VDCCs are normally closed and, respond with Ca 2+ permeability upon membrane depolarization. To relate the effect of GABAergic signaling to that of membrane depolarization, we treated GABA-deprived infected DCs with the depolarizing agent KCl. Upon blockade of GABA synthesis and secretion, KCl treatment fully restored hypermotility in Toxoplasma-infected DCs (Fig 4E), thus mimicking the effects obtained by addition of exogenous GABA (Figs 4D and 3C). Importantly, hypermotility was not restored by KCl in the presence of the L-type VDCC inhibitor nifedipine ( Fig 4E). Taken together with the effects of VDCC agonism (Fig 3), these data demonstrate a link between L-type VDCCs and the hypermigratory phenotype of T. gondii-infected DCs downstream of GABAergic signaling.

The dominantly expressed VDCC, Ca v 1.3, is linked to hypermigration of Toxoplasma-infected DCs
In order to determine putative VDCCs mediating the nifedipine-sensitive GABA reconstitution effect, we performed a screen of VDCCs expressed in DCs. RT-PCR analyses indicated transcriptional expression of the L-type VDCC Ca v 1.3 in DCs, similar to brain homogenate ( Fig 5A). A real-time quantitative PCR (qPCR) screen of VDCCs confirmed a consistent high relative expression of Ca v 1.3 transcripts in 6 mice tested over time, and also less abundant relative expression of Ca v 2.2 (Fig 5B and 5C). Other VDCC types, e.g. Ca v 1.1, Ca v 1.4, Ca v 2.1, Ca v 3.1, exhibited low, undetectable or inconsistent relative expression (Fig 5B and 5C). In Toxoplasma-challenged DCs, Ca v 1.3 remained the predominantly expressed VDCC type over other types (Fig 5D and S4A Fig) [23], and with similar relative expression in DCs and Toxoplasma-infected DCs (Fig 5E and S4C Fig). Immunocytochemistry using a mAb to a predicted sub-membranous Ca v 1.3 epitope yielded a distinct fluorescence signal in non-infected and in infected permeabilized DCs ( Fig 5F). Altogether, we conclude that Ca v 1.3 was the predominantly expressed VDCC in murine bone marrowderived DCs and that the relative VDCC expression profile varied between mice or varied over time. Upon Toxoplasma-infection, Ca v 1.3 remains the predominant transcriptionally expressed VDCC.
To functionally assess the relative contribution of Ca v 1.3 to hypermigration in relation to other putatively expressed VDCCs, we took advantage of a pharmacological antagonist with high specificity for Ca v 1.3, CPCPT [24], and a broad inhibitor of L, N and T type VDCCs, benidipine [25]. Both inhibitors similarly abolished the hypermotility of infected DCs ( Fig 5G  and S4D Fig). While CPCPT significantly reduced transmigration of DCs from different mice, benidipine was a more consistent abrogator of transmigration ( Fig 5H). Jointly, these data suggest that VDCCs play a significant role in T. gondii-induced hypermotility of DCs. As Ca v 1.3 appeared to be the most abundantly expressed VDCC, these data suggested that CPCPT and benidipine might act primarily on Ca v 1.3.

Ca v 1.3 gene silencing abrogates the hypermotility of T. gondii-infected DCs
To test the functional implication of Ca v 1.3 in T. gondii-induced hypermotility, we employed an RNA interference approach. Calcium signaling in Toxoplasma-infected hypermigratory DCs expression were selectively reduced in DCs exposed to shCa v 1.3 and that lentiviral transduction generates enhanced IL-12 mRNA expression in primary DCs.
To assess the impact of Ca v 1.3 silencing on hypermotility, we first optimized the approach using the murine DC line (JAWS II). JAWS II cells and DCs expressed a similar VDCC profile, with Ca v 1.  (Fig 6E). In line with results obtained upon pharmacological L-type VDCC inhibition ( Fig 4D), exogenous GABA restored motility in mock-treated GABA-inhibited DCs but failed to restore motility in the shCa v 1.3-transduced cells ( Fig 6F). We conclude that selective silencing of Ca v 1.3 abolishes T. gondii-induced hypermotility in DCs.

VDCC inhibition in adoptively transferred infected DCs delays the dissemination of T. gondii in mice
We have previously shown that adoptive transfer of T. gondii-infected DCs in mice leads to rapid dissemination of parasites and to exacerbation of the infection compared to infection with free tachyzoites [5,26], and that GABAergic inhibition blocks this exacerbated dissemination [8]. To assess if VDCC inhibition impacted on parasite loads, benidipine pre-treated infected DCs were adoptively transferred to mice intraperitoneally. When the infections were monitored by in vivo bioluminescence, photonic emissions indicated dissemination of parasites to spleen and mesenteric lymph nodes (MLN) (Fig 7A and 7B). Plaquing assays of homogenized whole organs (non-perfused) revealed overall reduced mean parasite loads in the spleens of mice challenged with infected DCs (+) benidipine compared to mice challenged with infected DCs (-) benidipine (Fig 7C). To analyze the contribution of parasites in the blood circulation to the total parasite loads in organs, parasite loads were analyzed after blood perfusion. Perfused spleens exhibited overall reduced parasite loads, and reduced or abolished differences in parasite loads between the benidipine-treated and the non-treated conditions ( Fig 7C). This showed that both removal of blood and benidipine treatment had a reducing impact on parasite loads in the spleen. In contrast, blood perfusion yielded more discrete relative reductions of parasite loads in MLNs, which are indirectly linked to the blood circulation via the lymphatic system ( Fig 7D). In the perfused mice, parasites were consistently detected in all brains from day 4 versus day 3 in non-perfused mice (Fig 7E), indicating a contribution of parasites that were displaceable by blood perfusion to the total parasite loads in non-perfused mice. Calcium signaling in Toxoplasma-infected hypermigratory DCs Because benidipine-treatment had an impact on splenic parasite loads early during infection and this effect appeared linked to parasites in blood, we analyzed the fate of parasites and DCs within 24 h post-inoculation intraperitoneally. Upon benidipine pre-treatment of infected DCs, significantly reduced parasite numbers were measured in spleen by 24 h (Fig  8A), with a non-significant reduction of parasites in blood and non-significant differences in peritoneum (Fig 8A). Similarly, flow cytometry analyses identified reduced numbers of cellassociated GFP-expressing parasites (GFP + ) in the spleen upon benidipine treatment ( This indicated direct transport to the spleen by infected DCs and also rapid transfer of parasites to new leukocytes in peritoneum and spleen. Benidipine treatment yielded non-significant effects on DC viability, infection frequencies and parasite viability (S10 Fig). Altogether, the data show that adoptively transferred infected DCs rapidly entered the circulation and that VDCC inhibition led to reduced numbers of parasite-associated cells in spleen during the early phase of infection. VDCC inhibition delayed the appearance of parasites in circulation and, thereby, also the systemic dissemination of T. gondii.

Discussion
In this study we investigated the molecular signaling mechanisms that govern how T. gondii hijacks the migratory properties of DCs. Building on previous work showing that a hypermigratory phenotype sets in within a few minutes after T. gondii invasion of DCs [4] and depends on GABAergic signaling [8], we addressed the role of Ca 2+ signaling in these processes.
Our studies establish that Ca 2+ signaling in murine bone marrow-derived DCs is indispensable for T. gondii-induced hypermotility and transmigration in vitro. The observation that the onset of the hypermigratory phenotype was abrogated at sub-physiological extracellular Ca 2+ concentrations or by blocking plasma membrane Ca 2+ channels underpinned a role for membrane-bound Ca 2+ channels. However, Ca 2+ mediates signal transduction to multiple cellular pathways. It was therefore crucial to determine its putative interaction with the GABAergic system of DCs. We previously showed that inhibition of GABA synthesis, GABA secretion or GABA A receptor blockade in Toxoplasma-infected DCs abolishes hypermigration [8]. Here, we demonstrate that hypermotility and transmigration are restored in GABA-deprived infected DCs by (i) addition of exogenous GABA, (ii) by cell membrane depolarization with KCl and (iii) by L-type VDCC agonism. Consequently, (iv) L-type VDCC blockade hindered reconstitution of hypermotility by GABA and KCl. This pinpointed a role for VDCCs downstream of GABAergic signaling. Further, inhibition of purinergic Ca 2+ channels (P2 receptors) yielded non-significant effects on hypermotility despite a measurable Ca 2+ influx in response to ATP. Also, while Ca 2+ -deprivation led to a similar proportional reduction of motility of unchallenged DCs (baseline motility) and infected DCs (hypermotility), selective L-type  VDCC inhibition abolished hypermigration but had non-significant effects on the baseline motility of DCs. Altogether, this indicated that L-type VDCCs primarily mediated the GABAevoked motility-related Ca 2+ influx and that extracellular Ca 2+ influx per se into the cell or increased cytosolic Ca 2+ levels per se was not sufficient to induce hypermigration.
To our knowledge, the findings demonstrate for the first time that murine DCs express the L-type VDCC subtype Ca v 1.3, with a functional implication in motility. Ca v 1.3 appeared to be the predominant transcriptionally expressed VDCC in primary DCs, a feature also maintained by the DC line JAWS II. Importantly, silencing of Ca v 1.3 by shRNA or selective pharmacological antagonism of Ca v 1.3 abrogated the hypermigratory phenotype in Toxoplasma-infected DCs, while baseline motility and morphology of DCs remained intact related to mock-treated and non-infected DC. A caveat of lentiviral transduction in primary DCs is that the lentivirus vector may have activation effects on the DCs [27], yet without reported apparent inhibitory effects on functionality [28], thereof the requirement of appropriate control experiments. We validated and confirmed our results in two additional cell lines. IL-12 mRNA expression indicated activation by the lentiviral vector primarily in DCs, to a lesser extent in JAWS II and, non-significant effects on the NE-4C line. Silencing of Ca v 1.3 expression abolished the hypermigratory phenotype, in contrast to Ca v 1.2 silencing. This, together with its apparent predominant expression, attributes a primary role in Toxoplasma-induced hypermotility to the VDCC subtype Ca v 1.3. However, despite that we did not observe compensatory up-regulation of other VDCCs/Ca v 1.2 upon Ca v 1.3 silencing in DCs, the data do not exclude a contributive role for other VDCC subtypes. In fact, the neuronal VDCC family members often display overlapping functions in mediating signal transduction [29]. This may apply to VDCCs in murine DCs too, as relative variations in transcription of several VDCC subtypes were detected in different mice over time, yet conserving a relative predominant expression of the subtype Ca v 1.3. Altogether, the data at hand defines a role for Ca v 1.3 in Toxoplasma-induced DC hypermotility and establish Ca 2+ as a second messenger to GABAergic signaling in DCs.
We have previously shown that GABA induces GABA A receptor-activated currents in DCs [8]. Here, we demonstrate that DCs can sense membrane voltage changes caused by depolarization (KCl or GABA) and can respond to GABA by a Ca 2+ transient. The analogous hypermotility restoration effect of exogenous GABA and depolarization by KCl, together with the opposite effects of the VDCC inhibitors (benidipine, nifedipine, CPCPT) and the agonist BayK8644 (a structural analog of nifedipine with positive inotropic activity) strongly suggests that GABA mediates membrane depolarization-induced Ca 2+ release via VDCCs. Also by analogy to findings in neurons [12], it is plausible that GABA A receptors and chloride homeostasis are regulated by cation-chloride co-transporters in DCs. Altogether, our findings provide evidence of a direct link between GABA receptor signaling, Ca v 1.3 activation and hypermotility.
Although modulated by Toxoplasma infection, functional GABA A receptors appear to be constitutively expressed by murine and human DCs [8]. The effects of GABA on Ca 2+ signaling via VDCCs / Ca v 1.3 has not been previously addressed in immune cells [30]. However, Ca 2+ channels mediate some of the most rapid biological processes described and VDCC signaling allows for immediate cellular responses to external stimuli [31]. This is in agreement with the features attributed to the hypermigratory phenotype of Toxoplasma-infected DCs [3]: fluid 24 h post-inoculation of 5x10 6 T. gondii (PTGluc, GFP-expressing)-challenged DCs ± benidipine, as indicated under Materials and Methods. Bivariate dot plots show, for each condition, GFP + cells and CD11c + cells, following gating on live CD3 -CD19 -GR1 -NK1.1 -CD11b + cells. Gatings indicate percentage of GFP + cells related to the total cell population. Bar graphs indicate, for each condition, the percentage of GFP + cells from 3 independent experiments (*: p < 0.05, ns: p ! 0.05, Student´s t-test, n = 5-6). for example, its rapid onset, cytoskeletal remodeling and switch to amoeboid-type of migration within minutes after T. gondii invasion of the host DC in vitro [4] and is also in line with the observed rapid migration of adoptively transferred DCs in vivo [8,26].
It has been previously reported that VDCCs may play a role in DC maturation [18] and T cell activation [32]. VDCC-related activity on DCs has been implicated in engulfment of apoptotic bodies, IL-12-production and up-regulation of major histocompatibility complex II [17,18], all of which are important immune functions of DCs. In line with these observations, the hypermigratory phenotype induced by T. gondii appears to rely on receptors and channels expressed by naïve DCs [8], but additionally requires the active invasion of a T. gondii tachyzoite [4], which is confirmed here by the observation that exogenous GABA per se is not motogenic on naïve DCs [8] (while VDCC agonism is). Altogether, this also advocates that T. gondii primes the host cell for responsiveness to GABA and is consistent with the idea that GABAergic activation occurs in an autocrine fashion with minimal by-stander effect [8]. Notably, the vast majority of GABA-responding DCs also responded to ATP or to consecutive stimuli with GABA ranging from micromolar to millimolar concentrations, indicating that GABA does not render the DCs refractory to other Ca 2+ -related stimuli and that intracellular Ca 2+ homeostasis is rapidly restored. We cannot exclude the involvement of additional mechanisms for Ca 2+ entry in DCs [33] acting sequentially or in parallel. However, their possible contribution to the hypermigratory phenotype should be secondary or posterior to Ca v 1.3 activation, as silencing of Ca v 1.3 in both primary DCs and the DC cell line JAWS II or selective pharmacological inhibition of Ca v 1.3 [24] abrogated T. gondii-induced hypermotility. The posterior involvement of intracellular Ca 2+ stores is also likely.
Ca 2+ also controls a number of critical processes in apicomplexan parasites, including gliding motility, cell invasion and egress [34][35][36]. It is unlikely that these mechanisms play in the interpretation of our results as inhibitors were added posterior to parasite invasion and non-significant effects were observed on parasite viability, reinvasion after egress or after forced release from treated host cells. On the other hand, our observations suggest that, through activation of the GABAergic system, T. gondii modulates the Ca 2+ homeostasis of the infected host cell, albeit transiently and locally. Induction of Ca 2+ signaling offers the advantage of bypassing transcriptional regulation in the host cells and thereby accelerating effector functions, i.e. rapid migratory activation of the invaded DC and, thereby, dissemination. We have previously shown that the onset of T. gondii-induced hypermotility precedes chemotactic responses in DCs in vitro and that, after the onset of chemotaxis, GABA/ GABA A receptor-mediated hypermotility and CCR7-mediated chemotaxis can cooperatively enhance the migration of infected DCs in vitro [4,8]. Thus, Ca 2+ entry in DCs, secondary to GABAergic activation, could hypothetically also influence Ca 2+ -dependent chemotaxis, with propagation of the signal to intracellular Ca 2+ stores. Future research needs to determine if Ca v 1.3 is involved in the cytoskeletal rearrangements that accompany the onset of hypermotility, some of which are independent of GABAergic signaling [4], e.g. the dissolution of adhesion-related podosomes [4] and the switch to amoeboid-like high velocity migration [37].
Our data demonstrate that VDCC inhibition in adoptively transferred infected DCs delays the dissemination of T. gondii tachyzoites in mice. VDCC inhibition reduced the parasite numbers in circulation and in the spleen early after inoculation, likely by delaying the outmigration of infected DCs from the peritoneal cavity [38]. The data advocates that the early presence of parasites in blood is important for setting the parasite loads in mice and that VDCC inhibition delayed this process. In line with this, perfusion experiments showed that the circulating pool of parasites contributes to the total parasite loads in organs and to dissemination during acute infection. Also, the high variability in leukocyte-associated parasitemias between mice 24 h post-inoculation is in contrast with the lower variability of parasite loads in the organs later during infection, and may indicate that parasitemia is intermittent early after infection. Because the spleen is an early site of T. gondii replication during acute infection [39], this mobilizable pool of parasites (by blood perfusion) may be important for the systemic dissemination observed at later time points. Also, DCs and monocytic cells are parasitized early during infection [26,38,40] and, both leukocyte-associated tachyzoites [41] and extracellular (free) tachyzoites are detected in blood later during acute infection (day 4) [41]. Our data show that adoptively transferred infected DCs reach the circulation and spleen rapidly but also that the transfer of replicating tachyzoites to new leukocytes is rapid and can occur in the peritoneal cavity, in line with previous observations [42,43]. Toxoplasma tachyzoites replicate in adoptively transferred DCs with lysis of infected DCs occurring within 48 h [5] and VDCC inhibition did not abrogate this process. This, together with the observed absence of parasites in brain parenchyma before day 4, indicates that it is unlikely that the adoptively transferred DCs transported parasites into the brain parenchyma. Rather, the observed delay in penetration to the parenchyma upon benidipine-treatment may be a consequence of delayed or lower parasitemias. Yet, DCs infiltrate the brain parenchyma during toxoplasmic encephalitis [44] and transportation of parasites to the brain by CD11b + leukocytes has ben shown [40]. However, more recent findings show that replication of tachyzoites in the endothelium is necessary before passage to the brain parenchyma [41]. Our studies contribute to elucidating the role of infected DCs in circulation and their impact on systemic dissemination, which indirectly impacts on parasitic loads in the brain parenchyma, but do not specifically address the mechanisms of passage of T. gondii tachyzoites across the blood-brain barrier [8]. Jointly, mounting evidences show that Toxoplasma utilizes combined strategies for systemic dissemination [6], by hijacking leukocytes [5,38,40] and as free parasites [26,41], and also with significant differences between Toxoplasma genotypes [26,45]. Additionally, intracellular localization of tachyzoites in migratory leukocytes may offer a safe intracellular niche for replication and delivery to organs and vasculature.
To the best of our knowledge, this constitutes the first report showing that the VDCC signaling axis can be utilized by an intracellular pathogen to modulate host cell migration and potentiate systemic dissemination. Based on the data at hand, we propose a model for the initiation of the hypermigratory phenotype in DCs by T. gondii, mediated by GABAergic signaling and with Ca 2+ acting as a second messenger (Fig 9). Initially, tachyzoite invasion triggers activation of the GABAergic system-GABA synthesis, transport and activation of GABA A receptors. Autocrine secretion of GABA by parasitized DCs leads to a membrane depolarization that activates the VDCC Ca v 1.3, with entry of Ca 2+ as a result. Finally, entry of Ca 2+ activates downstream signaling pathways that lead to cytoskeletal rearrangements and hypermotility. Mounting evidence indicates that, rather than being passively transported, intracellular microorganisms induce refined molecular orchestrations to manipulate the signaling pathways that modulate the migration of infected immune cells [6,46,47]. Continued investigations into how intracellular pathogens manipulate host cell Ca 2+ signaling pathways may identify new targets for inhibiting processes associated to pathogenesis.

Ethics statement
The Regional Animal Research Ethical Board, Stockholm, Sweden, approved experimental procedures and protocols involving extraction of cells from mice (N135/15, N78/16), following proceedings described in EU legislation (Council Directive 2010/63/EU).

Ca 2+ imaging
DCs (2x10 5 ) were seeded on 5% 3-aminopropyltriethoxysilane coating glass bottom dish and incubated at 37˚C with 5% CO 2 for 15 min. DCs were then loaded with 2 μM Fluo-8H/AM (AAT Bioquest) in CM at 37˚C with 5% CO 2 for 15 min, and washed with Krebs-Ringer's solution (150 mM NaCl, 6 mM KCl, 1.5 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES and 10 mM Dglucose) with 5% FBS. Time-lapse imaging was performed 2.5 s/frame, at 37˚C with 5% CO 2 on 200x magnification (Zeiss LSM 780 microscope equipped with a definite focus function). Cells were perfused with Krebs-Ringer's solution with 5% FBS via a peristaltic pump (0.5 ml/ min), which was also used to deliver pharmacological agents. The signals from individual cells were analyzed with ImageJ (version 1.46r, ROI Multi Measure). Each trace was normalized against the minimum value of all time points and a responding cell was defined as a signal exceeding 20% above baseline.

Western blotting
To determine the expression of the Ca V 1.3 protein in DCs and ACs, cells were lysed in RIPA buffer (150 mM NaCl, 50 mM Tris, 0.1% Triton, 0.5% deoxycholic acid, 0.1% SDS) with protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific) followed by sonication, addition of 4 x laemmli sample buffer and boiling. Proteins were separated by 8% SDS-PAGE, and blotted onto PVDF membrane (Millipore), blocked in 2.5% BSA followed by Western blotting with monoclonal anti-Ca V 1.3 C-terminal (Abcam), anti-GAPDH (Millipore) and anti-rabbit HRP (Cell signaling). Proteins were revealed by enhanced chemiluminescence (GE Healthcare) in a BioRad ChemiDoc XRS+.

Polymerase chain reaction (PCR)
Total RNA was extracted using TRIzol reagent (Life Technologies). First-strand cDNA was synthesized using Superscript III Reverse Transcriptase (Life Technologies). Real time quantitative PCR (qPCR) was performed in triplicates using SYBR green PCR master mix and a 7900HT Fast Real Time PCR system (Applied Biosystems). Products were analyzed with ABI 7900HT Sequence Detection System (Applied Biosystems) or Rotor gene (Corbett). 2 -ΔCt values are used to calculate the relative expression levels of 9 VDCC subtypes, with TATA box binding protein (TBP) as reference gene (S2 Table). For quantification of Ca v 1.3 knockdown, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and actin were used as reference genes (S2 Table). For quantification of T. gondii in tissues, the B1 gene was used (S2 Table).

Lentiviral vector production and transduction
Self-complementary hairpin DNA oligos targeting the Ca V 1.2 (Cacna1c) mRNA, Ca V 1.3 (Cac-na1d) mRNA, and a non-related sequence (luciferase, Luc) were chemically synthesized (DNA Technology, Denmark), aligned and ligated in a self-inactivating lentiviral vector (pLL3.7) containing a CMV-driven EGFP reporter and a U6 promoter upstream of cloning restriction sites (HpaI and XhoI) [50] (S3 Table). Restriction enzyme analysis and direct DNA sequencing confirmed the correct insertion of short hairpin RNA (shRNA) sequences. Lentivirus production was done using lipofectamine transfection. Briefly, shCa V 1.2, shCa V 1.3 or shLuc vectors were co-transfected with psPAX2 packaging vector and pCMV-VSVg envelope vector into Lenti-X 293T cells (Clontech) and the resulting supernatant was harvested after 60 h. Recovered lentiviral particles were centrifuged to eliminate cell debris, filtered through 0.45-mm cellulose acetate filters and concentrated by ultracentrifugation. Titers were determined by infecting Lenti-X 293T cells with serial dilutions of concentrated lentivirus. NE-4C cells, JAWS II cells and DCs (day 3) were transduced by spinoculation at 1000 g for 30 min in presence of hexadimethrine bromide (Polybrene, 8 μg/ml; Sigma Aldrich). Three to 5 days posttransduction, EGFP-expression was verified by epifluorescence microscopy before the cells were used in experiments. Transduction frequency was defined as the number of EGFPexpressing cells related to the total numbers of cells in five representative fields of view.

Adoptive transfers of Toxoplasma-infected DCs
Adoptive transfers were performed as previously described [8]. Briefly, DCs were challenged with freshly egressed PTGluc tachyzoites (6 h, MOI 3). Extracellular parasites were removed by centrifugation. Following resuspension in RPMI, tachyzoite-infected DCs or freshly egressed tachyzoites were adoptively transferred intraperitoneally into recipient C57BL/6 mice. Total number of colony-forming units (cfu) injected into animals was confirmed by plaquing assays. Benidipine (40 μM) was added to DCs for the last 3 h of the 6 h challenge with tachyzoites and replenished (40 μM) prior to injection in mice. When indicated, cells were stained with CMTMR following manufacturer´s instructions (Invitrogen).

In vivo bioluminescence imaging (BLI)
Eight-10 week old C57BL/6 mice were inoculated i.p. with freshly egressed PTGluc tachyzoites, or with PTGluc-infected DC ± benidipine. 3 mg D-luciferin potassium salt (Caliper Life Sciences, Hopkinton, MA, USA) was injected i.p. and mice were anesthetized with 2.3% isoflurane prior to BLI. Ten min after injection of D-luciferin, biophotonic images were acquired for 180 s (medium binning) with an In Vivo Imaging System (Spectrum CT, Perkin Elmer). For ex vivo imaging, organs are extracted and assessed as above. Analysis of images and assessment of photons emitted from a region of interest (ROI) was performed with Live Imaging Software (version 4.2; Caliper Life Sciences).

Plaquing assays
Plaquing assays were performed as described [8]. Briefly, organs were extracted and homogenized under conditions that did not affect parasite viability. The number of parasites was determined by plaque formation on fibroblast monolayers. When indicated, tachyzoites where released from infected DCs by repeated passages through a hypodermic needle (gauge 27), previous to plaquing. Transcardial blood perfusion was performed by injection of 25 ml PBS in the left ventricle after incision of the right atrium. Peritoneal lavage was performed by intraperinoneal perfusion and aspiration of 10 ml PBS using a hypodermic needle.

Statistical analysis
Statistical analyses were performed using R Stats Package version 3.0.2 (R Foundation for Statistical Computing, Vienna, Austria). Normality was tested by the Shapiro-Wilks test. Pvalues > 0.05 were defined as non-significant.
Supporting information S1  . Data represent mean ± SD of 3 independent experiments (ns: p ! 0.05, Student´s t-test). (C) Plaquing assays of T. gondii-challenged DCs ± benidipine (40μM), nifedipine (30μM) or CPCPT (10μM). Following challenge with freshly egressed tachyzoites (PTGluc, MOI 3, 6 h) and treatment (3 h post-challenge), DCs were force-lyzed before plaquing as indicated under Materials and Methods. Equal volumes of force-lyzed challenged DC suspensions were plated and plaques were counted by epifluorescence microscopy. Bar graph shows compiled analyses of plaque counts related to the force-lyzed and mock-treated controls with complete medium (plaque counts in the mocktreated control was set to 1). Data represent mean ± SD of 3 independent experiments (ns: p ! 0.05, Student´s t-test). (TIF) S1 Movie. DCs respond to GABA and ATP with a transient cytosolic Ca 2+ increase. Live cell Ca 2+ imaging of DCs loaded with 3 μM Fluo-4/AM as described in Materials and Methods. GABA (10 mM) and ATP (50 μM) were perfused after 3 min and 6.5 min, respectively. Data are representative of 3 independent experiments. (MOV)