In Vitro Infection of Human Nervous Cells by Two Strains of Toxoplasma gondii: A Kinetic Analysis of Immune Mediators and Parasite Multiplication

The severity of toxoplasmic infection depends mainly on the immune status of the host, but also on the Toxoplasma gondii strains, which differ by their virulence profile. The relationship between the human host and T. gondii has not yet been elucidated because few studies have been conducted on human models. The immune mechanisms involved in the persistence of T. gondii in the brains of immunocompetent subjects and during the reactivation of latent infections are still unclear. In this study, we analyzed the kinetics of immune mediators in human nervous cells in vitro, infected with two strains of T. gondii. Human neuroblast cell line (SH SY5Y), microglial (CMH5) and endothelial cells (Hbmec) were infected separately by RH (type I) or PRU (type II) strains for 8 h, 14 h, 24 h and 48 h (ratio 1 cell: 2 tachyzoites). Pro-inflammatory protein expression was different between the two strains and among different human nervous cells. The cytokines IL-6, IL-8 and the chemokines MCP-1 and GROα, and SERPIN E1 were significantly increased in CMH5 and SH SY5Y at 24 h pi. At this point of infection, the parasite burden declined in microglial cells and neurons, but remained high in endothelial cells. This differential effect on the early parasite multiplication may be correlated with a higher production of immune mediators by neurons and microglial cells compared to endothelial cells. Regarding strain differences, PRU strain, but not RH strain, stimulates all cells to produce pro-inflammatory growth factors, G-CSF and GM-CSF. These proteins could increase the inflammatory effect of this type II strain. These results suggest that the different protein expression profiles depend on the parasitic strain and on the human nervous cell type, and that this could be at the origin of diverse brain lesions caused by T. gondii.


Introduction
The majority of human Toxoplasma infections in immunocompetent hosts are asymptomatic, After an acute infection, tachyzoites can escape from the immune system, leading to the formation of tissue cysts containing bradyzoites, especially in the brain. However, in the immunocompromised host, latent bradyzoites in cysts revert to tachyzoites, leading to reactivation of chronic toxoplasmosis and development of a toxoplasmic encephalitis [1]. Thus, the severity of Toxoplasma infection obviously depends on the host immune status.
The role of the Toxoplasma strain is more debated. Genotyping of Toxoplasma isolates from all continents revealed a complex population structure. Up to now, 15 haplogroups were described [2]. These haplogroups comprise the 3 main clonal lineages initially described (type I, II and III) and other haplogroups that cluster various atypical strains and new clonal lineages [3,4] [5] [6]. On the basis of lethality in mice, type I strains were classified as virulent, and type II and III as non-virulent. These 3 types differ with respect to their ability to transmigrate across cellular barriers during invasion. Type I strains exhibit a higher migratory capacity than type II strains [7]. In humans, the influence of the strain in the clinical outcome is obvious in the severe cases of toxoplasmosis in immunocompetent patients due to the most divergent strains such as those circulating in the Amazonian forest [6], Its role is also highly suspected in the higher occurrence and severity of ocular toxoplasmosis in South America [8]. But it remains unclear if the Toxoplasma strain has any influence on the development of brain infection. In a study performed on 88 immunocompromised patients, the distribution of type II vs non-type II strains was not significantly different when patients were stratified by underlying cause of immunosuppression, site of infection (cerebral or extracerebral), or outcome [9].
During a toxoplasmic infection, the immune response can firstly reduce the parasite proliferation during acute infection, and then maintains chronic infection in immunocompetent hosts. During acute infection, monocytes, neutrophils, and dendritic cells are recruited to the site of infection [10] [11] [12]. These cells also play a role for migration and dissemination of the parasite in peripheral tissues and the central nervous system (CNS). This process depends on the parasitic strain. The type II strains induce superior migration of infected dendritic cells compared to type I strains [7]. Experimental data on animal models suggest that the immune response type 1 (Th1) is activated against T. gondii to control parasite replication. This immune response leads to production of interferon-gamma (IFN-c) in mice infected with RH (type I) or ME49 (type II) strains [10] [13]. IFN-c is the major mediator of resistance to T. gondii in the murine model; it can inhibit parasite replication, preventing toxoplasmic encephalitis during the late stage of infection in mice. During this host response, other cytokines and chemokines are produced, which can promote infiltration of immune cells to the site of infection [14] [15] [16].
In the mouse brain, microglial cells play a major role in the control of infections caused by T. gondii type II. These cells inhibit efficiently parasite growth and may thus function as important inhibitors of T. gondii propagation within the CNS by IFN-c and NO-independent mechanisms [24] [27]. In vitro comparison has shown that astrocytes are infected more efficiently than neurons and microglia [28] [27]. This is confirmed by in vivo experiments demonstrated that astrocyte is the predominant cell type infected by RH and PRU strains in the brain [26] [29]. Murine astrocytes have also been shown to inhibit in vitro the growth of NTE (type II) Toxoplasma strain [25]. Astrocytes infected by ME49 strain become activated to produce IL-1, IL-6 and GM-CSF [14]. These proinflammatory proteins together with microglia-produced cytokines play an important role in inducing infiltration of immune cells into the brain. In addition, Lüder et al. found that rat astrocytes and neurons are suitable host cells for the intra-cerebral proliferation of PLK strain (type II) [27]. It was described that both tachyzoites and bradyzoites can deregulate the function of murine neurons [30]. IFNc and IL-6 production by murine neurons indicates that these cells contribute to intracellular control of type II and I strains [31] [32] [33].
Most of these results were obtained in the murine model. The role of human neuro-endothelial cells in neuro-invasion by the parasite and the behavior of human microglial cells and neurons after T. gondii infection remains to be shown.
In this study, we analyzed the production of immune mediators during different times following Toxoplasma infection in a human model. For that, three human brain cells: human microglial, human endothelial and a human neuroblast cell lines were used. These cells were infected in vitro by different T. gondii strains. Proteomic analysis and parasite quantification were performed at each time point of infection to identify the various pro-inflammatory proteins produced by human nervous cells infected with different T. gondii strains and their impact on parasitic growth.

Ethics statement
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Committee on the Ethics of Animal Experiments of Limousin, France (Permit Number: 3-07-2012). All efforts were made to minimize suffering.

Parasite strains
Two strains of Toxoplasma gondii parasite were used: RH strain, a type I strain highly virulent in mice (DL90,10), and PRU strain, a type II strain with a non-virulent profile in mice (DL100 = 10 3 ) [34].

Cyst isolation
PRU cysts were obtained from the brains of Swiss mice infected one month earlier. Infected mouse brains were extracted and homogenized by several passages through a 20-gauge needle in Modified Eagle's Medium (MEM) 0.1% (Gibco, Cergy Pontoise, France) Tween 80 (Sigma-Aldrich, Lyon, France), to disrupt tissues. PRU cysts were isolated by Percoll gradient [35]. Briefly, ninety milliliters of Percoll (Amersham Biosciences, Orsay, France) were rendered isotonic by adding 10 mL of 106 MEM. The pH was neutralized by adding HCL 10%. Brain homogenates were diluted in MEM, 0.1% Tween-80%; 1.5 mL of 30% Percoll and then 1.5 mL of 90% Percoll were successively added with a tapered Pasteur pipette under suspension. After centrifugation (1811 g), cysts were recovered in the 30% Percoll layer. The cyst suspension was centrifuged to eliminate the remaining Percoll, and the pellet was recuperated in MEM, 0.1% Tween-80. PRU bradyzoites were released from cysts by trypsin 1X (Gibco) at 37Cu for 10 min and recovered after centrifugation (1811 g-15 min).

Tachyzoite production
RH tachyzoites from mouse ascites and PRU bradyzoites released from cysts were inoculated in human fibroblastic cell cultures (MRC5). After parasite multiplication, tachyzoites from both strains were released by disruption of infected cells by passages through 20 and 27 gauge needles. Cell debris were removed by filtration through polycarbonate membrane 3 mM (Nucleopore Whatman, Versailles, France). Purified parasites were suspended in MEM and enumerated.  [38], [39] and [40]. The properties of these cells are similar to those of endothelial brain cells, with respect to endothelial adhesion molecules, cell surface markers, and morphologic characteristics. Hbmec cells were grown in DMEM, supplemented with 10% FBS inactivated at 56uC for 1 h, 2 mM L-glutamine, 2 mM sodium pyruvate, 100 UI/mL of streptomycin-penicillin, 10 mM Hepes buffer, 4.5 g/mL glucose, 3500 UI/mL heparin and 0.15 mg/mL gentamicin. Hbmec cells were seeded in culture flasks (25 cm 2 ) pre-treated with 2% gelatin at density of 4610 6 cells/mL. Confluent cells were trypsinized in PBS.
All cell lines were cultured at 37uC in humidified air containing 5% CO 2 .

Parasite-cell co-cultures
The required number of cells were cultured 48 h before infection, time needed for cell adhesion. After cell adhesion, 6610 6 cells/mL of CMH5 were infected by 1.2610 7 tachyzoites/ mL. For Hbmec and SH SY5Y, 4610 6 cells/mL were infected by 8610 6 tachyzoites/mL (ratio: 1 cell/2 tachyzoites) of RH or PRU strains for 8 h, 14 h, 24 h and 48 h. Uninfected control cells were cultured under the same conditions. All co-cultures were performed separately in culture flasks (25 cm 2 ). The experiments were performed four times, in the same conditions but at different times to ensure reproducibility of results. Infections with different strains were performed at different periods to avoid any contamination between strains.
After infection, supernatants were recovered and frozen at 280uC; cells were trypsinized for 10 min at 37uC. After centrifugation at 804 g for 10 min, cell pellets were recovered and frozen at 280uC.

Cytokine, chemokine & growth factor analysis
Cytokine, chemokine & growth factor production profiles were detected by Proteome Profiler array kit, type human cytokines array panel A (R&D, Lille, France). Each nitrocellulose membrane contains duplicated spots of 36 different antibodies anti-cytokines, chemokines, growth factors and adhesion proteins. The kit was used according to manufacturer's instructions. Each membrane was incubated with a mixture chemiluminesence substrate A&B (Promega, Lyon, France) and exposed to G-BOX (Syngene, Versailles, France). The image was captured using GeneSnap software (Syngene). Results appear as dark spots, protein expression levels were determined by quantifying the intensity of coloration, expressed as pixel units, using GeneTools software (Syngene). The negative control quantification value was subtracted from the quantification value of each spot.

RNA extraction & quantification of T. gondii burden by semi-quantitative reverse transcription polymerase chain reaction (RT-PCR)
RNAs from each co-culture pellet were extracted using the RNeasy Mini Kit (Qiagen, Essones, France), according to manufacturer's instructions. DNAc was used to quantify the viable parasite burden in each co-culture by amplifying a T. gondii 529 pb repeat gene [43] [44]. qRT-PCR was performed using one step SYBR Green RT-PCR Kit (Qiagen). Each RNA sample (# 100 ng) was added to PCR tubes containing SYBR master mix (12.5 mL), RT Mix (0.25 mL), specific primers (Sigma-Aldrich) 529 pb (0.6 mM) sense: 59AGGCGAGGTGAGGATGA39; antisense: TCGTCTCGTCTGGATCGAAT 39 [44] and sterile water for final volume of 25 mL. Following assay optimization, negative samples were used. qRT-PCR was performed with a Rotor-Gene 6000 (Qiagen) using cycling conditions: reverse transcription 10 min at 55uC, PCR initial activation 5 min at 95uC, denaturation 5 sec at 95uC and annealing/extension 10 sec at 60uC for 40 cycles. Quantification was performed using a range of tachyzoites from 5610 4 to 1 Toxoplasma amplified by primers for the 529 pb gene. In addition, to verify the stage obtained after PRU culture in MRC5 cells, BAG-1, a bradyzoite-specific gene, was amplified by qRT-PCR using BAG-1 primers sense: 59-TGA GCG AGT GTC CGG TTA TT-39 and anti-sense: 59-ATT CCG TCG GGC TTG TAA T-39 [45] (Sigma-Aldrich).

Statistical analysis
All experiments were performed four times and values of corrected pixels were obtained from pixel values of infected cells minus pixel values of uninfected cells. They were expressed by 6 mean standard error. Statistical analysis was performed using a parametric ANOVA test and Tukey HSD (Honestly Significant Difference) test for multiple comparisons of means, with 95% confidence level. A value of p,0.05 was considered statistically significant.
In each kinetic study, the comparison of immune mediator production was performed between three groups: cells, strain and infection times. This was carried out firstly between each variable of each group: cell (CMH5 vs SH SY5Y) (SH SY5Y vs HBMEC) (HBMEC vs CMH5), strain (RH vs PRU) and idem for each period of infection. All interactions between the three groups were also studied (two by two).
The parasite burdens were evaluated as the ratio of final tachyzoite number/initial tachyzoite number. When the ratio was less than 100%, the tachyzoite mortality was higher than their proliferation. Based on the fact that 1 RH strain tachyzoite gives 2 tachyzoites after 7 h [46], 128 tachyzoites must be obtained after 48 h which represents a maximal ratio of 12800% parasite burden. Data on parasite multiplication rate were analyzed by Tukey HSD test, with 95% confidence level. A value of p,0.05 was considered statistically significant.

Results
For all infected cells by PRU strain, semi-quantitative RT-PCR for BAG-1 gene shown absence of bradyzoites in the inoculum.

Kinetics of pro-inflammatory proteins and T. gondii burden in infected human endothelial cells (Hbmec)
As shown in Figure 1A Figure 1C). G-CSF and GM-CSF production increased slightly from 14 h to 48 h post infection (NS) ( Figure 1E). In the period when all protein levels declined, the RH parasite burden rose sharply in endothelial cells ( Figure 1A, 2C).

Kinetics of pro-inflammatory proteins and T. gondii burden in an infected human neuroblastoma cell line (SH SY5Y)
As shown in Figure 3A

Discussion
Different experimental approaches were used to analyze expression of immune mediators during T. gondii infection in brain cells. For example, regarding human brain cells, Xiao et al, used a human neuroepithelioma cell line (neural cells, line SK-N-MC) infected with 3 canonical T. gondii strains to show differences in a wide range of biological functions such as the immune response [47]. They noticed a modulation of the expression of proinflammatory gene like IL-8 depending of the strain type.
The originality of our work resides in the comparison of the different protein levels of cytokines, chemokines and other immune factors in three human nervous cells infected in vitro by RH (type I) and PRU (type II) Toxoplasma strains during the 48 first hours post-infection by tachyzoites. Proteome Profiler Arrays allowed analyzing simultaneously a broad spectrum of proinflammatory proteins.
In this study, we identified many pro-inflammatory factors like cytokines, chemokines, and growth factors secreted by different human nervous cells after Toxoplasma infection. All these factors have roles in the host immune response to parasite challenge: chemokines and their receptors are important in the control of parasite replication and CNS inflammation, and trigger proinflammatory responses to many microbial pathogens [15]. Among the pro-inflammatory proteins, were chemokines (MCP1, GROa, MIF and Rantes) designed to recruit immune cells and cytokines (IL-6 and IL-8) able to activate immune responses. Growth factors (G-CSF and GM-CSF) are not only hematopoietic colony-stimulating factor but also neuronal growth factors with strong anti-apoptotic actions on neurons. In parallel to this pro-inflammatory protein synthesis, we observed a repeatable decrease in parasitic burden for each T. gondii strain at 24 h post infection in microglial cells and neurons. This suggests that this decrease may be due to production of various immune mediators involved in the early control of T. gondii multiplication in human nervous cells. At 48 h post infection, parasite burden increased in a large part of infected cells, in RH-infected cells whatever the cell type. We also observed a lysis of cells, which could explain the decrease of immune response.
Endothelial cells constitute the blood-brain barrier through which the tachyzoites pass into the CNS. Our results show that these cells are stimulated with Toxoplasma to produce immune mediators mainly at 8 h post infection. This reaction rapidly decreases between 14 h and 48 h post infection, when parasite burden increased. This suggests that these cells do not react strongly against Toxoplasma infection compared to neurons and microglial cells. The production of interleukins and chemokines in the early hours of the infection might activate the immune response. In these cells, PRU strain induces more inflammatory response than RH strain, as shown by a higher secretion of chemokines and growth factors between 14 h and 48 h. Endothelial cells specifically express sICAM-1 in the presence of PRU strain, but not after infection with RH strain (data not shown). Barragan et al. demonstrated that this soluble form of ICAM, sICAM-1, has an inhibitory effect on transmigration of T. gondii across BeWo (human placenta), Caco2 (human intestine) and MDCK (canine kidney) cells, whereas it did not significantly affect host cell invasion by the parasite [19]. According to these results, in our model, sICAM-1 could inhibit the transmigration process of type II strain through brain endothelial cells.
After infection by both T. gondii strains, all cells (microglial cells, endothelial cells, and neurons) synthesized mainly IL-6, IL-8, MCP-1 and GROa. In CMH5 and SH SY5Y, the expression of cytokines, chemokines and growth factors was variable but significantly higher at 24 h post infection, the time when parasite burden (RH and PRU) declined. At 24 h post infection, microglial cells infected by both strains produced high levels of MCP-1. This chemokine can participate in the control of T. gondii infection as described in human fibroblastic cells (MRC5) [48] and human astrocyte cells [49]. In vivo, it increases inflammation, promoting recruitment of monocytes and lymphocytes. The immune responses could be amplified in these cells by secretion of GROa. Synthesis of GROa was also significantly higher at 24 h post infection in microglial cells compared to other cells. This chemokine has a major role in neutrophil activation. Kikumura et al. [50] showed that GROa was up-regulated in the retina and brain of infected mice and involved in neutrophil infiltration. Thus, microglial cells seems to be major cerebral immune cells, able to amplify immune reactions against the parasite, promoting production of the main chemokines. IL-6 and IL-8 production was predominant in microglial cells and neurons compared to endothelial cells, especially at 24 h post infection, which might suggest that these two interleukins have a role in controlling proliferation of RH and PRU strains in microglial cells and neurons. IL-6 is known to control parasite burden and inflammatory activity, in the brain together with other cytokines (such as IL-10) [14]. IL-6 is also a mediator responsible for the production of acute phase proteins and increases cytotoxic activity of NK cells. For the first time, we demonstrated a local production for IL-8 in human nervous cells that appeared to have a crucial role in neutrophil recruitment to the brain as described in patients with retinochoroiditis. IL-8 has a role in the inflammatory mechanisms of acute toxoplasmic retinochoroiditis [51]. It is noteworthy that mice lacking CXCR2, the high affinity receptor of IL-8, show a significant increase in tachyzoite proliferation and a large number of cysts in the brain [52].
Besides the previously described chemokines and interleukins, infected microglial cells produced high levels of MIF compared to other cells; this production can boost the control of parasite proliferation in these cells. MIF has an essential role in controlling T. gondii type II proliferation and the severity of inflammation in a mouse model [53].
G-CSF and GM-CSF were produced by all cells, especially those infected by PRU strain. In addition to their role in growth, differentiation and maturation of macrophages, monocytes and dendritic cells, they have the capacity to inhibit in vivo neutrophil apoptosis, and the ability to lyse tachyzoites in vitro in the presence of specific antibodies [54]. G-CSF and GM-CSF may thus extend the inflammatory process in the brain by enhancing phagocytosis [55] and cytotoxic activities [56]. Infected astrocytes and endothelial cells also become activated to produce GM-CSF [57]. Studies using T. gondii infection in mice, which can cause CNS inflammation, reported a detrimental role for GM-CSF by increasing parasitic burden in peritoneal macrophages [58]. Therefore, GM-CSF is a potent pro-inflammatory cytokine that plays a pathogenic role in the CNS inflammatory disease, experimental autoimmune encephalomyelitis [59], suggesting that G-CSF and GM-CSF may amplify the pro-inflammatory immune reaction during PRU infection in brain.
For the first time, we showed the up-regulation of serine protease inhibitor (SERPIN E1) in infected human nervous cells. It occurred particularly in neurons and microglial cells a few hours post infection. Many members of the serine protease inhibitor superfamily play an important role in physiological and pathological processes, and can be regarded as protease inhibitors, which are involved in coagulation reactions, fiber dissolution, angiogenesis, complement activation, and immune and inflammatory reactions. It has been suggested that SERPINs (SERPIN B3, B4) could inhibit host cell apoptosis that may inhibit replication and decrease T. gondii viability. Therefore, the persistence of upregulated SERPIN E1 in microglial cells and neurons could also be involved in limiting the growth of T. gondii and preventing death of infected human nervous cells. In addition, SERPIN E1, plasminogen activator inhibitor 1, inhibits the uPA/uPAR pathway which promotes active MMP-9 forms secreted by Toxoplasma infected macrophages [60], which could limit the invasion of immune cells infected by T. gondii into the brain.
In addition to several studies which have shown that murine and rat microglial cells contribute in vitro to the inhibition of T. gondii replication [27], our results suggest that both human microglial cells and neurons can react against Toxoplasma by producing IL-6, IL-8, MCP-1, GROa and SERPIN E1 [20] [17] [18]. These immune mediators may play a role in the control of Toxoplasma multiplication at 24 h post infection in the human brain, as shown by the decrease in parasitic burden at this time of cell infection. G-CSF and GM-CSF were specifically produced in all cells infected with PRU, suggesting a pro-inflammatory effect of this strain in human nervous cells compared to RH strain. In this work we did not detect a production of IFN-c in infected human nervous cells. This could be explained by the lack of interaction between the effector immune cells. For this reason, this work will be complemented by a study of IFN-c effect on immune mediators and on parasite multiplication rate in each infected cells.