Cyclooxygenase (COX)-2 modulates Toxoplasma gondii infection, immune response and lipid droplets formation in human trophoblast cells and villous explants

Congenital toxoplasmosis is represented by the transplacental passage of Toxoplasma gondii from the mother to the fetus. Our studies demonstrated that T. gondii developed mechanisms to evade of the host immune response, such as cyclooxygenase (COX)-2 and prostaglandin E2 (PGE2) induction, and these mediators can be produced/stored in lipid droplets (LDs). The aim of this study was to evaluate the role of COX-2 and LDs during T. gondii infection in human trophoblast cells and villous explants. Our data demonstrated that COX-2 inhibitors decreased T. gondii replication in trophoblast cells and villous. In BeWo cells, the COX-2 inhibitors induced an increase of pro-inflammatory cytokines (IL-6 and MIF), and a decrease in anti-inflammatory cytokines (IL-4 and IL-10). In HTR-8/SVneo cells, the COX-2 inhibitors induced an increase of IL-6 and nitrite and decreased IL-4 and TGF-β1. In villous explants, the COX-2 inhibitors increased MIF and decreased TNF-α and IL-10. Furthermore, T. gondii induced an increase in LDs in BeWo and HTR-8/SVneo, but COX-2 inhibitors reduced LDs in both cells type. We highlighted that COX-2 is a key factor to T. gondii proliferation in human trophoblast cells, since its inhibition induced a pro-inflammatory response capable of controlling parasitism and leading to a decrease in the availability of LDs, which are essentials for parasite growth.

Toxoplasmosis is a disease caused by Toxoplasma gondii, an obligate intracellular protozoan parasite that can infect all warm-blooded vertebrates, such as birds and mammals, including humans, representing a significant public health problem 1,2 . It is estimated that 30 to 50% of the world population is infected with the parasite 3 . Congenital toxoplasmosis is considered one of the most severe forms of the disease, and occurs due to the transplacental passage of T. gondii tachyzoites during pregnancy, reaching circulation and fetal tissues, which may have serious implications from fetus to adulthood [4][5][6] .
The immune response against T. gondii is primarily mediated by T helper 1 (Th1) cells, and requires components of the innate and adaptive immune response 7 . At the beginning of the infection, T. gondii is recognized by innate imune response cells, stimulating synthesis of interleukin (IL)-12 in dendritic, macrophage and neutrophil cells, and induces interferon (IFN)-γ production by natural killer (NK) cells. Additionally, tumor necrosis factor (TNF) can act in synergism with IL-12, optimizing IFN-γ production by these cells 8,9 . The production of these pro-inflammatory cytokines is associated with activation of the CD4 and CD8 T-lymphocyte-mediated adaptive immune response, which act to produce and secrete various inflammatory mediators, such as nitric oxide (NO), and induce an even greater increase in the levels of IL-12 and IFN-γ 10,11 . IFN-γ activates dendritic Results COX-2 inhibitors did not change the viability of BeWo and HTR-8/SVneo cells. Initially, we evaluated the cellular viability of BeWo and HTR-8/SVneo cells treated with COX-2 inhibitors, meloxicam or celecoxib, at different concentrations (Fig. 1).
We observed that both inhibitors did not induce toxicity at any concentration tested when compared to untreated cells (medium), regardless of cell type. In addition, DMSO (0.005%) also did not trigger toxicity in both cells (Fig. 1A,B).
Based on these results, we chose two concentrations of each inhibitor for further experiments, one of low and one of high concentration. For meloxicam, 10 and 250 μg/mL were chosen, while for celecoxib 1 and 5 μg/ mL were chosen. www.nature.com/scientificreports/
In HTR-8/SVneo cells, the meloxicam (10 μg/mL: * P = 0.0261; 250 μg/mL: * P < 0.0001) and celecoxib (1 μg/ mL: * P = 0.0010; 5 μg/mL: * P < 0.0001) also induced a reduction in T. gondii proliferation when compared to untreated and infected cells (T. gondii). In this cell line, meloxicam and celecoxib presented IC 50 of 39.7 ± 13.2, and 1.5 ± 0.2 μg/mL, respectively. Furthermore, our results showed a significant difference between the different doses of meloxicam ( $ P = 0.0007) or celecoxib ( % P = 0.0009) (Fig. 2B), showing that the inhibition was dosedependent in both cells. HTR-8/SVneo (B) cells were treated with meloxicam or celecoxib in increasing concentrations, or untreated (medium) or treated with 0.005% DMSO. After 24 h, the cells were submitted to MTT assay and data were presented as percentage (%) of viable cells (cellular viability) in relation to untreated cells (medium, 100% of cellular viability). Data were shown as mean ± SEM from three independents experiments with eight replicates. Differences between groups were analyzed by Kruskal-Wallis test and Dunn's multiple comparison post-test (GraphPad Prism Sofware version 6.01, https:// www. graph pad. com). Differences were considered significant when P < 0.05. BeWo (A) and HTR-8/SVneo (B) cells were infected with T. gondii for 3 h and treated or not with meloxicam or celecoxib for an additional 24 h. Next, cells were submitted to T. gondii intracellular proliferation assay (β-galactosidase assay), and data were presented as percentage (%) of T. gondii proliferation in relation to untreated and infected cells (100% of intracellular proliferation). Data were shown as mean ± SEM from three independents experiments with eight replicates. Differences between groups were analyzed by Kruskal-Wallis test and Dunn's multiple comparison post-test (GraphPad Prism Sofware version 6.01, https:// www. graph pad. com). Significant differences in relation to untreated and infected cells ( * T. gondii), and between concentrations of meloxicam ( $ ) or celecoxib ( % ). Differences were considered significant when P < 0.05. www.nature.com/scientificreports/ COX-2 inhibitors increased IL-6 and MIF and decreased IL-4 and IL-10 in BeWo cells infected by T. gondii. Next, we evaluated the cytokine production in the supernatant of BeWo cells under different experimental situations (Fig. 3). Our data showed that uninfected cells and treated only with meloxicam (10 μg/mL: * P = 0.004; 250 μg/mL: * P < 0.0002) increased the MIF production when compared to untreated and uninfected cells (Fig. 3A). In addition, we observed an increase in MIF production in untreated and infected cells ( * P = 0.0073) or infected and treated with meloxicam (10 μg/mL: * P = 0.0164; 250 μg/mL: * P = 0.0002) or celecoxib (1 μg/mL: * P = 0.0015; 5 μg/ mL: * P < 0.0001) when compared to untreated and uninfected cells (Fig. 3E). In addition, the higher concentration of both inhibitors induced greater MIF production when compared to lower concentrations (meloxicam: $ P = 0.033; celecoxib: % P < 0.0001) in treated and infected cells, showing a dose-dependent effect (Fig. 3E).

Scientific Reports
In relation to IL-6, we observed that uninfected cells and treated with the higher concentration of meloxicam ( * P = 0.0005) or celecoxib ( * P < 0.0001) increased the production of this cytokine when compared to untreated and uninfected cells, showing a dose-dependent effect (Fig. 3B). In untreated and infected cells ( * P = 0.034) or infected and treated with meloxicam (10 μg/mL: * P = 0.0431; 250 μg/mL: * P < 0.0001) or celecoxib (1 μg/mL: * P = 0.021; 5 μg/mL: * P < 0.0001), there was an increased production of this cytokine when compared to untreated and uninfected cells (Fig. 3F). Also, we observed an increase in IL-6 production in infected cells and treated with higher concentration of meloxicam ( # P = 0.0111) and celecoxib ( # P < 0.0001) when compared to untreated and infected cells (Fig. 3F). Furthermore, the increase IL-6 production in treated and infected cells also had a dose-dependent effect (Fig. 3F).
Regarding to IL-10, we observed that uninfected cells and treated with both concentrations of meloxicam (10 μg/mL: * P = 0.0031; 250 μg/mL: * P = 0.0107), there was an increased production of this cytokine when compared to untreated and uninfected cells, and a dose-dependent effect was observed in relation to celecoxib (Fig. 3D). In untreated and infected cells, we observed an increase in IL-10 production when compared to untreated and uninfected cells ( * P = 0.0367) (Fig. 3H). However, infected cells and treated with 250 μg/mL of meloxicam ( # P = 0.041) or both concentrations of celecoxib (1 μg/mL: # P = 0.0303; 5 μg/mL: # P = 0.0002) decreased the production of this cytokine when compared to untreated and infected cells (Fig. 3H).
In summary, our data show that COX-2 inhibitors increase MIF and IL-6, and decrease IL-4 and IL-10 production in BeWo cells, triggering a pro-inflammatory profile in BeWo cells.
IL-8, TGF-β1, TNF-α and nitrite showed values below detection levels in all conditions tested in BeWo cells (data not shown).

COX-2 inhibitors increased IL-6 and nitrite and decreased IL-4 and TGF-β1 in HTR-8/SVneo cells infected by T. gondii.
Next, we also evaluated the cytokine production and nitrite in the supernatant of HTR-8/SVneo cells under different experimental conditions (Fig. 4).
In relation to IL-6, we observed that uninfected cells and treated with higher concentration of meloxicam ( * P = 0.0241) or celecoxib ( * P = 0.0004) increased the production of this cytokine when compared to untreated and uninfected cells (Fig. 4B). Untreated and infected cells not changed IL-6 levels, but infected cells and treated with both concentrations of meloxicam (10 μg/mL: * P = 0.0110; 250 μg/mL: * P = 0.0222) or with 5 μg/mL of celecoxib ( * P < 0.0001) increased the IL-6 production when compared to untreated and uninfected cells (Fig. 4F). Furthermore, infected cells and treated with higher concentration of meloxicam ( # P = 0.031) or celecoxib ( # P < 0.0001) increased the IL-6 production when compared to untreated and infected cells, and celecoxib showed a dosedependent effect (Fig. 4F).
Moreover, we observed the nitrite production in these cells ( Supplementary Fig. S1). We observed that uninfected cells and treated with 10 μg/mL of meloxicam or both concentrations of celecoxib decreased the nitrite production when compared to untreated and uninfected cells ( * P < 0.0001), except the concentration 250 μg/mL of meloxicam ( Supplementary Fig. S1A). However, infected cells and treated with both concentrations of meloxicam (10 μg/mL: # P = 0.0203; 250 μg/mL: # P < 0.0001) or celecoxib (1 and 5 μg/mL: # P < 0.0001) increased the nitrite production when compared to untreated and infected cells, with meloxicam presenting a dose-dependent effect ( Supplementary Fig. S1B). www.nature.com/scientificreports/

T. gondii increased lipid droplets formation in BeWo and HTR-8/SVneo cells. After evaluating
the parasite proliferation and immune response in BeWo and HTR-8/SVneo cells infected and treated with COX-2 inhibitors, we evaluated the LDs formation in these cells infected or not with T. gondii (Fig. 5).
Previous studies demonstrated that murine peritoneal macrophages increased the LDs production when treated with mouse serum, decreasing the microbicidal capacity against T. gondii 45 . Then, in a firts step of experiments, we evaluated the effect of fetal bovine serum (FBS) in LDs production in uninfected and infected cells for 24 h. Uninfected or infected BeWo and HTR-8/SVneo cells, in presence of FBS, increased the LDs production when compared to uninfected cells FBS-free ( * P = 0.0012; * P = 0.0243, respectively) or infected cells FBS-free ( $ P < 0.0001) (Fig. 5A,B).

Lipid droplets formation decreased in BeWo and HTR-8/SVneo cells infected and treated with COX-2 inhibitors.
Next, we evaluated the LDs production under different experimental conditions with COX-2 inhibitors (Fig. 6). In BeWo cells, we observed that untreated and infected cells ( * P < 0.0001) induced an increase in LDs production when compared to untreated and uninfected cells (Fig. 6A). On the other hand, the treatment of infected cells with both concentrations of meloxicam ( # P < 0.0001) or celecoxib (1 μg/mL: # P < 0.0001; 5 μg/mL: # P = 0.0041) reduced significantly the LDs production in comparison to untreated/infected cells (Fig. 6A). The concentrations 250 μg/mL of meloxicam ( $ P = 0.0009) and 1 μg/mL of celecoxib ( % P < 0.0001) reduced more the LDs production, showing a dose-dependent effect (Fig. 6A).

COX-2 inhibitors were not toxic and reduced the T. gondii intracellular proliferation in human villous explants.
After evaluating the effects of COX-2 inhibitors in in vitro models, we evaluated the influence of COX-2 inhibitors in human villous explants, an ex vivo model of maternal-fetal interface (Fig. 7).
Firstly, based on the measurement of the LDH enzyme, we observed that no concentration of both inhibitors changed the villous viability (Fig. 7A). Additionally, we performed the MTT assay and observed the same effect (Fig. 7B). Based on these data, we chosed the concentration of 250 μg/mL meloxicam and 5 μg/mL celecoxib for further ex vivo experiments. Representative photomicrographs of untreated villous (Fig. 7C), or villous treated with 250 μg/mL meloxicam (Fig. 7D) or 5 μg/mL celecoxib (Fig. 7E), where it is possible to observe the morphology of the syncytiotrophoblast (arrow) and mesenchyme (M) preserved when compared to untreated villous (medium).

COX-2 inhibitors increased MIF and decreased TNF-α and IL-10 in human villous explants infected by T. gondii.
Next, we evaluated the cytokine profile in supernatant of untreated and uninfected, treated and uninfected, untreated and infected, and infected and treated villous with meloxicam or celecoxib (Fig. 9).
We observed that uninfected villous and treated with both inhibitors not induced MIF production when compared to untreated and uninfected villous (Fig. 9A). In untreated and infected villous, the production of this cytokine decreased when compared to untreated and uninfected villous ( * P = 0.0094) (Fig. 9D). However, infected villous and treated with meloxicam ( # P = 0.0238) or celecoxib ( # P = 0.0098) increased the MIF production when compared to untreated and infected villous (Fig. 9D).
In relation the TNF-α production, we observed that only uninfected villous and treated with meloxicam produced this cytokine (Fig. 9B). In untreated and infected villous, the TNF-α production increased, but infected www.nature.com/scientificreports/ villous and treated with meloxicam decreased the production of this cytokine when compared to untreated and infected villous ( # P = 0.0445) (Fig. 9E).
Regarding to IL-10 production, we observed that uninfected villous and treated with both COX-2 inhibitors did not alter the production of this cytokine when compared to untreated and uninfected villous (Fig. 9C). In untreated and infected villous, the IL-10 production increased when compared to untreated and uninfected villous ( * P < 0.0001). However, infected villous and treated with meloxicam ( # P < 0.0001) or celecoxib ( # P = 0.0007) decreased the production of this cytokine when compared to untreated and infected villous (Fig. 9F).
In summary, our data show that COX-2 inhibitors increase MIF and decrease TNF-α and IL-10 production in human villous explants.
IL-4, IL-8 and TGF-β1 showed values below detection levels in all conditions tested (data not shown).

Proposed model for the action of COX-2 inhibitors during T. gondii infection in human trophoblast cells and human villous explants.
Based on our results, a proposed model of the actions induced by COX-2 inhibitors in in vivo and ex vivo experimental models is shown in Fig. 10. When BeWo cells were infected with T. gondii and treated with COX-2 inhibitors, there was a reduced parasite intracellular proliferation, upregulation of pro-inflammatory cytokines (MIF and IL-6), downmodulation of anti-inflammatory cytokines (IL-4 and IL-10), and a decrease in LDs production (Fig. 10A, left panel). In addition, HTR-8/SVneo cells infected with T. gondii and treated with COX-2 inhibitors also reduced parasite intracellular proliferation, upregulated IL-6 and nitrite, downmodulated IL-4 and TGF-β1, and also reduced the LDs production (Fig. 10A, right panel). When we evaluated the role of COX-2 in ex vivo model, we observed that human villous explants infected with T. gondii and treated with COX-2 inhibitors reduced T. gondii intracellular proliferation, upregulated MIF and downmodulated TNF and IL-10 ( Fig. 10B). Thus, COX-2 is an important inflammatory mediator for the T. gondii proliferation, since its inhibition induced a pro-inflammatory response and decreased the production of LDs, favoring the control of the parasite growth.

Discussion
COX-2 and PGE 2 can be stored and produced in lipid droplets 63 , and studies show that many pathogens are able to increase the production of these organelles to evade the host immune response 39,64 . However, the role of COX-2 in T. gondii infection, as well as its importance in LDs production is not well understood in many cells type, including in maternal-fetal interface. Thus, in this work we evaluated the functional role of COX-2 during infection by T. gondii (RH strain, clone 2F1) and its effect on the LDs production in human trophoblastic cells. Firstly, we analyzed the possible cytotoxicity effect of COX-2 inhibitors (meloxicam and celecoxib) in BeWo and HTR-8/SVneo cells. We observed that both inhibitors did not decrease the viability of both cell lines. These results are in agreement with studies by our group, where the same concentrations of both inhibitors did not alter the viability of peritoneal macrophages 53 . In other study using aspirin and celecoxib as COX-2 inhibitors, they observed that the concentrations used also did not alter the viability of cardiac myoblast cells 65 .
Next, we analyze whether COX-2 inhibitors would modulate the T. gondii intracelular proliferation. Our results demonstrated that both inhibitors were able to decrease parasitism in dose-dependent manner in both cell types, suggesting that susceptibility to T. gondii is modulated by COX-2 in villous and extravillous human trophoblast cells, and this effect is probably related to a decrease of PGE 2 . Our findings are in agreement with many other studies. Previous studies have demonstrated an association of COX-2/PGE 2 in susceptibility in Leishmania spp. and T. cruzi infection. In macrophages coinfected with HIV-1 and Leishmania amazonensis, the TAT viral protein induced greater COX-2 expression and PGE 2 production, favoring the growth of L. amazonensis, but this effect was reversed when COX-2 was inhibited by celecoxib 66 . Moreover, in murine mononuclear phagocyte derived Villous explants were infected with T. gondii for 24 h, and treated with meloxicam or celecoxib for an additional 24 h. Next, the villous explants were macerated for β-galactosidase assay, and data were presented as percentage (%) of T. gondii proliferation in relation to untreated and infected villous (100% of intracellular proliferation). Data were shown as mean ± SEM from two independents experiments with eight replicates. Differences between groups were analyzed by Kruskal-Wallis test and Dunn's multiple comparison post-test (GraphPad Prism Sofware version 6.01, https:// www. graph pad. com). Significant differences in relation to untreated and infected villous ( * T. gondii). Differences were considered significant when P < 0.05. www.nature.com/scientificreports/ www.nature.com/scientificreports/ from B-1 cells (B-1CDP) infected with L. major, the treatment with COX-2 inhibitors (aspirin, indomethacin or NS-398) decreased the parasites number 67 . Also, cells-enriched peripheral human blood cells population treated with celecoxib promoted a decrease in T. cruzi infection, however, when these cells were treated with celecoxib plus PGE 2 , the rate of infected cells with T. cruzi was restored when compared to cells treated only with celecoxib, but aspirin (more selective for COX-1 than COX-2) did not present the same effect 68 . In contrast, Swiss mice infected with T. cruzi and treated with aspirin reduce parasitemia during an acute or chronic phases 69 . Recently, our research group demonstrated that Calomys callosus rodents, murine peritoneal macrophages and human monocytes infected by T. gondii (ME-49 or RH strains) and treated with COX-2 inhibitors (celecoxib or meloxicam) decreased parasitism 53 . However, when human monocytes were treated with COX-2 inhibitors plus PGE 2 , the parasitism was restored, highlighting the importance of these inflammatory mediators in T. gondii infection 53 . However, there are no studies about COX-2 and T. gondii in maternal-fetal interface of any specie, and it is widely established by scientific literature that maternal-fetal interface is an environment completely different in aspects about immune response, physiology and several other mechanisms 70 . Thus, our present study is innovator, since is the first showing the role of COX-2 in the susceptibility of T. gondii infection in human trophoblastic cells. Then, we investigate some possible mechanisms that COX-2 inhibitors would be interfering in T. gondii infection. Studies have showed that PGE 2 can inhibit the pro-inflammatory immune response (Th1 profile). Therefore, the COX-2 inhibitors and, consequently, the decrease of PGE 2 production, could restore the Th1 profile, leading to parasitism control 52,71 . In this sense, we evaluated the cytokines profile secreted by human trophoblastic cells. In BeWo cells, both COX-2 inhibitors increased MIF and IL-6, and decreased IL-4 and IL-10, while in HTR-8/ SVneo cells, both COX-2 inhibitors increased IL-6 and nitrite, and decreased IL-4 and TGF-β1. Although BeWo and HTR-8/SVneo cells presented a pro-inflammatory profile induced by COX-2 inhibitors, they demonstrated different expression of cytokines. Additonally, BeWo cells and HTR-8/SVneo showed below detection levels of IL-8, TGF-β1, and TNF, or IL-8, TNF and IL-10, respectively. Regarding to T. gondii infection in experimental models of maternal-fetal interface, previous studies also have reported, by ELISA assay, undetectable levels of TGF-β1, IL-12p70, TNF-α, and IFN-γ cytokines produced by BeWo cells, even after parasite infection 56,57 . In addition, a weak expression of IL-8 is commonly observed in unstimulated BeWo cells 72 .

Scientific
Studies have shown that IL-6 is a pro-inflammatory cytokine due to its protective role against pathogens, such as T. gondii 13,54 . Furthermore, corroborating our findings, C. callosus rodents infected with T. gondii and treated with both COX-2 inhibitors also increased IL-6 production 53 , showing the importance of this cytokine in the parasitism control. BeWo and HTR-8/SVneo cells showed increase IL-6 when treated with COX-2 inhibitors, and it was an important mechanism shared by both population of trophoblast to control the parasite replication. Also, both cells secreted IL-6 in basal conditions. IL-6 is a multifunctional cytokine produced by human trophoblast, since is associated with normal placental development and successful pregnancy [73][74][75] . Therefore, it is plausible that BeWo and HTR-8/SVneo modulate similarly IL-6 to ensure adequate physiology of the maternal-fetal interface. MIF is another important pro-inflammatory cytokine in the immune response against pathogens 76,77 , and studies www.nature.com/scientificreports/ showed their protective role in T. gondii infection 14,15,53,56 . Although studies showing its protective role against T. gondii, including in BeWo cells 14,54,78 , recently a study by our group showed that inhibition of MIF decreases the T. gondii proliferation in HTR-8/SVneo cells 59 . Moreover, the addition of recombinant MIF increased COX-2 expression, favoring the T. gondii growth in HTR-8/SVneo cells 59

. Our findings in HTR-8/SVneo cells showed that COX-2 inhibitors did not increase MIF production when compared to only infected cells, but the inhibitors induced an upregulation of this cytokine in BeWo cells, suggesting that parasitism control in HTR-8/SVneo
is not related to this cytokine. Thus, different population of human trophoblast can present difference expression of cytokines, as MIF, since villous trophoblast can control T. gondii infection using MIF, but extravillous trophoblast do not present the same ability. In addition, our findings showed that infected HTR-8/SVneo cells decreased nitrite production, but when COX-2 was inhibited, its production increased, corroborating with other studies, where COX-2 inhibition also increased nitrite production 53,65 , showing the importance of this inflammatory mediator in infection control. On the other hand, nitrite production was not observed in BeWo cells. It is possible to speculate that nitrite is harmful for immune tolerance in placental environment (villous trophoblast), however previous study showed active role of nitrite in HTR-8/SVneo cells (extravillous trophoblast) 79 , once again evidencing that differences of inflammatory mediators between trophoblasts cells are normal and expected.
Although there is a strong immune response to control the infection, the parasite seeks mechanisms to evade the host immune response, such as the induction of anti-inflammatory cytokines. Studie by our group showed that the susceptibility to T. gondii infection in BeWo cells was increased by IL-10 and TGF-β1 80 . Our findings demonstrated that the infection by T. gondii induced the IL-4, IL-10 and TGF-β1 production, however, COX-2 inhibitors decreased these cytokines and also contributed to parasitism control. Corroborating our findings, studies showed that human monocytes infected with T. gondii and B-1CDP infected with L. major and treated with COX-2 inhibitors (meloxicam, celecoxib or aspirin) decreased the production of IL-10 when compared to only infected cells 53,67 . Moreover, spleen cells of BALB/c mice infected with T. cruzi and treated with COX-2 inhibitor (sodyum salicilate or meloxicam) decreased the IL-4 and IL-10 production 81 . Both BeWo and HTR-8/ SVneo cells modulated IL-4, but they were different in relation to IL-10 and TGF-β1. Our previous study showed that HTR-8/SVneo is able to modulate TGF-β1, and we could not verify IL-10 in these cells 58 , although we demonstrated that BeWo can modulate TGF-β1 and IL-10 80 . We have to consider that, in the present study, both cell lines were submitted to COX-2 inhibitors, and this is the first study to demonstrate the role of this enzyme in maternal-fetal interface, thus is plausible that different population of trophoblast can express different cytokines.
It has been demonstrated that intracellular tachyzoites of T. gondii are able to upregulate the LDs production by host cells and recruit these organelles to close the parasitophore vacuole, which together favors the parasite replication 40,41 . Moreover, literature findings show that LDs are source of COX-2 and PGE 2 82 and that the presence of COX-2 and PGE 2 in LDs may be associated with the success in the parasite proliferation 64 . Illustrating that, some pathogens modulate the expression of the COX-2/PGE 2 axis to subvert the host immune response 39,64 . In this line of discussion, our data demonstrated that the COX-2 inhibition caused a reduction in the number of intracellular tachyzoites in both cell types and in villous explants. Thus, to understand the underlying mechanisms associated with the parasite control mediated by COX-2 inhibitors, besides immune response, we assessed the impact of the COX-2 inhibition in the LDs production in BeWo and HTR-8/SVneo cells infected by T. gondii. Interestingly, our findings showed that the COX-2 inhibition resulted in a decrease of LDs in both cell lines; thus, these data highlight that the blockage of the COX-2 activity reduce the availability of LDs to T. gondii, which could explain, in part, the impairment in the parasite replication. Thus, this is the first study to show the influence of COX-2 in LDs expression during Toxoplasma infection in maternal-fetal interface.
Lovo-Martins et al. 83 showed that C57BL/6 mice treated with extracellular vesicles from T. cruzi of Y strain (EV Y) had increased cardiac parasitism and high internalization of the parasite in bone marrow-derived macrophages, and these effects were associated with the LDs formation and PGE 2 production. Macrophages infected with T. cruzi increased the LDs production, leading to a higher PGE 2 production and contributing to the growth of the parasite. However, the COX-2 inhibitor reversed this effect 35 , corroborating our findings. Thus, the parasitism control in our results may also be related to the decrease in LDs. However, our data showed that the lower concentrations of both inhibitors, except meloxicam in BeWo cells, were more effective in reducing the LDs production than the higher concentrations, different of data from parasite intracellular proliferation, where the higher concentrations of COX-2 inhibitors were more efficient in controlling parasitism. However, the dosedependent effect of COX-2 inhibitors in controlling T. gondii infection in BeWo and HTR-8/SVneo cells can be explained by the dose-dependent effect of the inhibitors in inducing a strong pro-inflammatory immune response, such as upregulation of IL-6 and MIF. Thus, it is possible to suggest that the immune response is more efficient in controlling parasitism in human trophoblastic cells than the amount of LDs. Furthermore, since T. gondii can consume the content of these organelles, it is possible that in higher doses of COX-2 inhibitors, because there are fewer parasites, there was higher lipid droplets availability.
After evaluating the role of COX-2 inhibitors in in vitro models, we investigated the effect of these inhibitors in ex vivo model. We observed that both inhibitors did not alter the viability of human villous explants and were also able to strongly reduce the T. gondii intracellular proliferation. Studies demonstrate that the annexin A1 peptide was able to reduce parasitism in villous explants infected by T. gondii, and this was related to a decrease in COX-2/PGE 2 production 84 . Regarding to cytokines production, we observed that only infected villous decreased the MIF production and increased IL-10, suggesting a possible escape mechanism for T. gondii, besides inducing a high production of TNF-α. However, both inhibitors reversed these effects, corroborating the parasitism findings. Our research group have demonstrated the protective role of MIF against T. gondii in human placenta 56,61,62,85 . The placental microenvironment is rich in IL-10, an anti-inflammatory cytokine that contributes to maternal-fetal tolerance, which can facilitate infection by T. gondii 80,86 , as shown in this study. Interestingly, COX-2 inhibitors decreased the production of this cytokine, contributing to the control of the parasitism. In addition, www.nature.com/scientificreports/ we also demonstrated that the infection of human villous by T. gondii induced a high TNF-α production, a proinflammatory abortogenic cytokine that can disrupt the maternal-fetal immune balance and cause miscarriage 87 .
In contrast, COX-2 inhibitors decreased the TNF-α release, which may be important for the maintenance of pregnancy. Literature findings have demonstrated a relationship between TNF-α and COX-2 [88][89][90] . Illustrating that, it was reported that TNF-α is capable of stimulating the COX-2 activation resulting in an augmentation of PGE 2 by first trimester trophoblasts; surprinsigly, this phenomenon was reversed by the use of COX-2-selective inhibitors 88 . In this sense, combining our findings with literature, we hypothesized that the reduction in TNF-α levels, as result of COX-2 inhibition, may culminate in a decrease of PGE 2 production by infected villous, which could explain the impairment of the parasite growth. Taken together, in agreement with published studies, our data reveal a different cytokine profile between human trophoblast cells and villous explants from human third trimester pregnancy in the context of T. gondii 56,57,60 . In this context, we suggested that these differences among both models may be explained, in part, due to the fact that human villous is a tissue composed of mesenchyme, villous and extravillous trophoblast and syncytiotrophoblast, culminating in a distinct cytokine production.
In conclusion, we demonstrated that COX-2 inhibitors impair T. gondii intracellular proliferation in human trophoblastic cells and human chorionic villous, since upregulates pro-inflammatory mediators and reduce the lipid droplets production in cells, dampening the environment for parasite survival. In this sense, the present findings are novel and highly essential in the context of placental environment.
However, future studies are necessary to clarify specific mechanisms involved, as signaling pathways triggered by modulated cytokines and associated to lipid droplets production in trophoblast cells. Thus, COX-2 may be a future therapeutic target for the treatment of congenital toxoplasmosis. Human villous explants culture. Third-trimester human placentas (36 to 40 weeks of pregnancy, n = 2) were acquired from pregnant patients after elective cesarean section deliveries at the Clinics Hospital of the Federal University of Uberlândia (HC-UFU), MG, Brazil. Exclusion criteria included pre-eclampsia, hypertension, infectious disease such as toxoplasmosis, chronic renal disease, cardiac disease, connective tissue disease, diabetes and other manifestations which could interfere with the results of this study. After being collected, placental tissues were washed in sterile phosphate-buffered saline (PBS, pH 7.2) to remove the excess of blood, and the dissection of the villous was performed using a stereomicroscope to remove endometrial tissue and fetal membranes up to 1 h after collection. Then, floating terminal chorionic villi containing five to seven free tips per explant were collected, placed in 96-well plates (one per well) and cultured in 200 µL RPMI 1640 medium with 10% FBS, penicillin (100 U/mL) and streptomycin (100 μg/ml) for 24 h at 37 °C and 5% CO 2 for future experiments. The volume of the villous explants was ∼ 10 mm 3 56,91 . Parasites. Tachyzoites of T. gondii (2F1 clone) were provided from Dr. Vern Carruthers (Medical School of Michigan University, Ann Arbor, MI, USA). This clone is derived from highly virulent RH strain and express the β-galactosidase gene, which enable a colorimetric assay to measure the parasites in cells or tissues 54,56 . These parasites were propagated in BeWo cells maintained in RPMI 1640 medium supplemented with penicillin (100 U/mL) and streptomycin (100 μg/ml) and 2% FBS at 37 °C and 5% CO 2 .
After 24 h of treatment, supernatants were collected and frozen at -80 °C for later measurements of cytokines and nitrite. In parallel, the 3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl2H-tetrazolium bromide (MTT, Sigma) assay was performed on the cells 92  gondii tachyzoites in a proportion of one parasite per cell (1:1) in medium FBS-free, and after 3 h of culture, infected cells were treated with 10 and 250 µg/mL meloxicam or 1 and 5 µg/mL celecoxib in medium FBS-free for an additional 24 h. As controls, cells were infected and untreated (T. gondii). After 24 h, supernatants were collected and stored at -80 °C for later measurements of cytokines and nitrite. In parallel, T. gondii intracellular proliferation assay was performed in infected cells by β-galactosidase colorimetric assay, as previously describe 54 .
T. gondii intracellular proliferation was obtained according to a reference curve containing free tachyzoites (from 1 × 10 6 to 15.6 × 10 3 ). The data were expressed as a percentage (%) of T. gondii proliferation: the mean number of tachyzoites from infected and untreated (T. gondii) corresponded to 100% parasite proliferation, then the number of tachyzoites from each treatment condition was transformed in percentage according to 100% of parasite proliferation from the control (T. gondii) 56 . The half-maximum antiparasitic activity (IC 50 ) were calculated by extrapolation of the corresponding dose-response curve on a log-linear plot employing the portions of the curve that transected the 50% response point; the values were expressed in μg/mL ± standard desviation (SD) 93 . Three independent experiments in eight replicates were performed.
Lipid droplets staining. In order to evaluate the production of lipid droplets in BeWo and HTR-8/SVneo cells, we performed Oil Red O (Sigma) staining 94 .
In the first step of the experimets, BeWo and HTR-8/SVneo cells were cultured on 13 mm glass coverslips (Sigma) (5 × 10 4 cells/200 µL in 24-well plates), infected with T. gondii tachyzoites (1:1) in medium in absence or presence of FBS for 3 h, and then treated with medium in absence or presence of FBS for adittional 24 h. As control, cells were uninfected and maintained in medium FBS-free. In a second step, BeWo and HTR-8/SVneo cells were cultured on 13 mm glass coverslips (5 × 10 4 cells/200 µL in 24-well plates), infected with T. gondii tachyzoites (1:1) in medium FBS-free and, after 3 h of culture, cells were treated with 10 and 250 µg/mL meloxicam or 1 and 5 µg/mL celecoxib in medium FBS-free for an additional 24 h. As controls, cells were uninfected and untreated (uninfected medium) or infected and untreated (infected medium) in medium FBS-free. After 24 h, cells from all experimental conditions were processed for analysis of lipid droplets. The glass coverslips with adherent cells were washed with sterile PBS, fixed with formaldehyde 3.7% in HBSS (pH 7.4) for 15 min, incubated with propylene glycol (100%) for 5 min, and stained with 0.5% Oil Red O (Sigma) diluted in 100% propylene glycol for 10 min at 60 °C. Next, cells were washed with 85% propylene glycol for 5 min and washed thrice with distilled water. In addition, as representative images, one glass coverslip from each condition was counter stained with Harris-hematoxylin for 30 s. The glass coverslips stained only with Oil Red O were analyzed under a light microscope (BX40, Olympus, Tokyo, Japan) (40 images from different fields of each condition). Images were analyzed in Image J software version 1.50i (National Institutes of Health, USA, https:// imagej. nih. gov/ ij). The data were expressed as the percentage of LDs (% area LDs/field) in each condition in comparison to untreated and uninfected cells (100% of LDs area in the field). Three independent experiments in four replicates were performed.
Immunofluorescence. In addition, also to evaluate the production of lipid droplets in BeWo and HTR-8/ SVneo cells infected, we performed confocal fluorescence microscopy assay.
BeWo and HTR-8/SVneo cells were cultured on 13 mm glass coverslips (Sigma) (5 × 10 4 cells/200 µL in 24-well plates), infected with T. gondii tachyzoites (1:1) in medium FBS-free for 24 h. As control, cells were uninfected and maintained in medium FBS-free (medium). After, the glass coverslips with adherent cells were washed with sterile PBS, fixed with formaldehyde 10% for 10 min, washed twice with sterile PBS and incubated with Nile Red (Sigma) (diluted 1:1000 in sterile PBS 1X) to label lipid droplets for 30 min in the dark at room temperature. Next, coverslips were washed twice with sterile PBS and incubated overnight with a mouse monoclonal primary anti-Toxoplasma gondii antibody [SAG1/p30] (Abcam TP3 #ab8313, Cambridge, United Kingdom) (diluted 1:500 in PGN-0.01% saponin solution). Then, coverslips were washed thrice with sterile PBS and incubated with Alexa Fluor 488-conjugated anti-mouse IgG (Invitrogen, #A11001, Carlsbad, CA, USA) (diluted 1:500 in PGN-0.01% saponin solution) and TOPRO-3 Iodide (Life Technologies) (diluted 1:500 in PGN-0.01% saponin solution) for 1 h in the dark at room temperature to label tachyzoites of T. gondii and nuclei, respectively. Coverslips were mounted on glass slides and 20 images from different fields of each condition were analyzed by confocal fluorescence microscopy (Zeiss, LSM 510 Meta, Germany) with an inverted microscope (Zeiss Axiovert 200 M). Images were analyzed in Image J software version 1.50i (National Institutes of Health, USA, https:// imagej. nih. gov/ ij), and the data were expressed as the percentage of total intensity of LDs (% total intensity LDs/field) in each condition in comparison to untreated and uninfected cells (100% of total intensity LDs/field).
Viability assay in human villous explants after COX-2 inhibitors treatment. The human villous explants were collected as described above and cultured for 24 h in medium with 10% FBS at 37 °C and 5% CO 2 . After 24 h, villous were tretated with meloxicam (10, 250, 500 and 1000 μg/mL) or celecoxib (5 and 10 μg/mL) in medium FBS-free for an additional 24 h. As control, villous explants were treated with only medium FBS-free (medium), corresponding to 100% viability. COX-2 inhibitors concentrations were based on a previous study by our research group 53  www.nature.com/scientificreports/ conversion of pyruvate to lactate in the presence of NADH. The absorbance was measured in a DU-70 spectrophotometer (Beckman, Brea, CA., USA) at 340 nm, for 2 min at 37 °C. The decrease in absorbance at 340 nm due to the oxidation of NADH is proportional to the activity of LDH in the sample. Data were expressed as U/L of enzyme activity. In addition, to verify explant integrity with COX-2 inhibitors, they were collected for morphological analysis using hematoxylin and eosin staining 56,62 . Two placentas were used, and two independent experiments with eight replicates were performed. Furthermore, the MTT assay was performed on the villous 57,92,95 . After 24 h of treatment, the supernatants were collected and frozen at -80 °C for later measurements of cytokines. The villous were incubated with 20 µL of MTT (5 mg/mL) plus 180 µL medium with 10% FBS in the same culture condition at 37 °C and 5% CO 2 for 4 h. The formazan crystals were solubilizied by adding a solution containing 10% SDS and HCl (0.01 M) at 37 °C and 5% CO 2 for 18 h. The villous were removed from each well and the optical densities measured at 570 nm (Titertek Multiskan Plus, Flow Laboratories, McLean, VA, USA). Data were expressed as the percentage of viable villous (% viability of villous) in comparison untreated villous (100% of villous viability). Two placentas were used, and two independent experiments with eight replicates were performed.
T. gondii intracellular proliferation in human villous explants after COX-2 inhibitors treatment. The villous explants were cultured in 96-well plates for 24 h in medium with 10% FBS at 37 °C and 5% CO 2 . Next, the villous were infected with T. gondii tachyzoites in a proportion of 1 × 10 6 parasites to each well in medium FBS-free for 24 h. Next, the villous were treated with meloxicam (250 µg/mL) or celecoxib (5 µg/mL) in medium FBS-free for an additional 24 h, according to the results of the toxicity assay (LDH e MTT assay) and morphological analyses. As control, villous explants were infected and untreated (T. gondii). After 24 h, the supernatants were collected and frozen at -80 °C for later measurements of cytokines. In parallel, the T. gondii intracellular proliferation assay was performed in the infected cells by the β-galactosidase colorimetric reaction, as previously described 56,62 . The parasite intracellular proliferation in villous explants samples was performed by adding 150 µL RIPA buffer [50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 1% (w/v) sodium deoxycholate, and 0.1% (w/v) sodium dodecyl sulfate (SDS), pH 7.5] supplemented with protease inhibitor cocktail (Complete, Roche Diagnostic, Mannheim, Germany) to each villous and homogenizing the samples in ice for protein extraction. The homogenate was centrifuged at 21.000 × g for 15 min at 4 °C and the supernatant was collected to measure the protein total (µg/mL) using the Bradford assay 96 . Aliquots of 20 µL of each sample were used to determine T. gondii intracellular proliferation by β-galactosidase assay, as described above. Next, the data of number of tachyzoites were normalized according to the protein concentration of each villous, showing the number of tachyzoites per µg of tissue. The data were expressed as percentage (%) of T. gondii proliferation: the mean number of tachyzoites from controls (untreated and infected villous) corresponded to 100% of parasite proliferation, and the number of tachyzoites from each treatment condition was transformed into a percentage according to 100% of parasite proliferation from the control. Two independent experiments with eight replicates were performed.
Cytokines and nitrite. The human cytokines IL-4, IL-6, IL-8, IL-10, TNF-α, MIF and TGF-β1 were measured in supernatants of BeWo and HTR-8/SVneo cells and villous explants by ELISA, according with manufacturer's instructions (R&D Systems, Minneapolis, MN, USA; or BD Biosciences, San Diego, CA, USA). Data were expressed in pg/mL according to a standard curve of each cytokine for BeWo and HTR-8/SVneo cells, while for villous explants, the data were normalized according to the protein concentration of each villous. Then, for villous explants, the data about cytokines were obtained by the ratio between concentration of cytokines in pg/mL and concentration of total protein from Bradford assay in µg/mL, resulting in pg/mg of tissue 56,57,91 .
In addition, the cells supernatants were used to measure nitrite by the Griess method 97 . The supernatant were added to 96-well plates, incubated with Griess reagent (1% sulfanilamide dihydrochloride and 0.1% naphthylenediamide dihydrochloride in 2.5% H 3 PO 4 ) for 10-20 min, and the absorbance was read in a plate reader (Titertek Multiskan Plus, Flow Laboratories, McLean, VA, USA) at 570 nm. The nitrite concentration was determined with reference to a standard curve of sodium nitrite (µM/mL). Statistical analysis. Statistical analysis was performed using the GraphPad Prism Sofware version 6.01 (GraphPad Software Inc., San Diego, CA, USA, https:// www. graph pad. com). All data were expressed as mean ± standard error of mean (SEM). Differences between groups were assessed by One-Way ANOVA test with Sidak's multiple comparison post-test for the parametric data, or Kruskall-Wallis test with Dunn's multiple comparison post-test for the nonparametric data, or Student's t test, when appropriate. Statistical differences were considered significant when P < 0.05. Ethical approval. The present research protocol using human tissue samples was performed in accordance with relevant guidelines and regulations, and the experimental protocols were approved by the Ethics Committee of the Federal University of Uberlandia, MG, Brazil, with approval number 1.585.342. A consent term was obtained from all subjects, and when the subjects were under 18, a parent and/or legal guardian was acquired. In addition, informed consent was obtained from all participants and/or their legal guardians. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.