Leishmania Promastigotes Enhance Neutrophil Recruitment through the Production of CXCL8 by Endothelial Cells

Endothelial cells represent one of the first cell types encountered by Leishmania promastigotes when inoculated into the skin of the human hosts by the bite of phlebotomine sand flies. However, little is known on their role in the early recruitment of phagocytic cells and in the establishment of the infection. Initially, neutrophils, rapidly recruited to the site of promastigotes deposition, phagocytize Leishmania promastigotes, which elude the killing mechanisms of the host cells, survive, and infect other phagocytic cells. Here, we show that Leishmania promastigotes co-incubated with HMEC-1, a microvascular endothelial cell line, exhibited significant morphological changes and loss of infectivity. Moreover, promastigotes of different Leishmania species stimulated the production of CXCL8 by HMEC-1 in a dose- and TLR4-dependent manner. Interestingly, we observed that the conditioned media from Leishmania-stimulated HMEC-1 cells attracted leukocytes, mostly neutrophils, after 2 h of incubation. After 24 h, a higher percentage of monocytes was detected in conditioned media of unstimulated HMEC-1 cells, whereas neutrophils still predominated in conditioned medium from Leishmania-stimulated cells. The same supernatants did not contain CCL5, a chemokine recruiting T cells and monocytes. On the contrary, inhibition of the production of CCL5 induced by TNF-α was seen. These data indicate that the interaction of Leishmania promastigotes with endothelial cells leads to the production of chemokines and the recruitment of neutrophils, which contribute to the establishment of Leishmania infection.


Introduction
Leishmaniasis are parasitic neglected vector-borne diseases, caused by protozoa of the genus Leishmania. Worldwide, an estimated 700,000 to 1 million new cases occur every year [1]. These diseases are transmitted to the vertebrate host by the bite of an infected female sand fly. Human infections are caused by more than 20 of the 30 species of Leishmania that infect mammals. During blood feeding, Leishmania metacyclic flagellated promastigotes are co-inoculated into the host's skin together with sand fly saliva and midgut content. Sand flies damage the skin of the vertebrate hosts using their proboscis to rip and scratch through the tissues causing bleeding, which immediately activates  :20). Promastigotes of L. infantum were cocultured with HMEC-1 or were grown alone for 24 h (D). Cell viability was measured by MTT assay. Data are expressed as OD (550/650) means ± standard deviation (SD) of three independent experiments. One-way ANOVA, Tukey's multiple comparisons.

Viability, Morphological Analysis and Infectivity of Leishmania spp. in HMEC-1 Culture Conditions
In the first set of experiments, promastigotes from different Leishmania spp. were incubated under endothelial cell culture conditions (HMEC-1 culture medium, 37 °C and 5% CO2) for 4 or 24 h and their viability was determined by MTT assay and lightmicroscopy examination. The optical density was significantly lower at 24 h than at 4 h of incubation (p > 0.01), indicating a reduced viability or a reduced proliferation of promastigotes in the conditions used for in vitro HMEC-1 culturing (Figure 2A). Similar effects were observed for all three different species examined. MTT data were further confirmed by microscopic parasite counting. Starting from a concentration of 5 × 10 6 Leishmania/mL, after 4 h of incubation in cell culture conditions, parasites reached 15 × 10 6 /mL and 14 × 10 6 /mL at 23 °C and 37 °C, respectively. After 24 h of incubation, parasites were 32 × 10 6 /mL and 12 × 10 6 /mL at 23 °C and 37 °C, respectively. These data confirm that promastigotes do not proliferate at 37 °C, and some of them die.
L. infantum morphology was evaluated by Giemsa smears at different incubation times (0, 4, 24 h) in cell culture conditions ( Figure 2B-D). After 4 h, some Leishmania promastigotes appeared enlarged with a visible flagellum, whereas some others were smaller, rounded forms with a short flagellum ( Figure 2C). After 24 h, all the parasites appeared rounded with short or even with no visible flagellum ( Figure 2D). The same morphology was observed when Leishmania promastigotes were co-incubated with HMEC-1 ( Figure 2E). In addition, some parasites appeared with a pale cytoplasm and broken cell membranes (red arrows in Figure 2E).
Next, in order to assess whether the selected incubation conditions might also alter promastigotes infectivity, the ability of L. infantum or L. tropica promastigotes (1:10, cell:parasite ratio) to infect PMA differentiated THP-1 macrophages was examined after 4  :20). Promastigotes of L. infantum were cocultured with HMEC-1 or were grown alone for 24 h (D). Cell viability was measured by MTT assay. Data are expressed as OD (550/650) means ± standard deviation (SD) of three independent experiments. One-way ANOVA, Tukey's multiple comparisons.

Viability, Morphological Analysis and Infectivity of Leishmania spp. in HMEC-1 Culture Conditions
In the first set of experiments, promastigotes from different Leishmania spp. were incubated under endothelial cell culture conditions (HMEC-1 culture medium, 37 • C and 5% CO 2 ) for 4 or 24 h and their viability was determined by MTT assay and light-microscopy examination. The optical density was significantly lower at 24 h than at 4 h of incubation (p > 0.01), indicating a reduced viability or a reduced proliferation of promastigotes in the conditions used for in vitro HMEC-1 culturing (Figure 2A). Similar effects were observed for all three different species examined. MTT data were further confirmed by microscopic parasite counting. Starting from a concentration of 5 × 10 6 Leishmania/mL, after 4 h of incubation in cell culture conditions, parasites reached 15 × 10 6 /mL and 14 × 10 6 /mL at 23 • C and 37 • C, respectively. After 24 h of incubation, parasites were 32 × 10 6 /mL and 12 × 10 6 /mL at 23 • C and 37 • C, respectively. These data confirm that promastigotes do not proliferate at 37 • C, and some of them die.
L. infantum morphology was evaluated by Giemsa smears at different incubation times (0, 4, 24 h) in cell culture conditions ( Figure 2B-D). After 4 h, some Leishmania promastigotes appeared enlarged with a visible flagellum, whereas some others were smaller, rounded forms with a short flagellum ( Figure 2C). After 24 h, all the parasites appeared rounded with short or even with no visible flagellum ( Figure 2D). The same morphology was observed when Leishmania promastigotes were co-incubated with HMEC-1 ( Figure 2E). In addition, some parasites appeared with a pale cytoplasm and broken cell membranes (red arrows in Figure 2E). h of incubation in different culture conditions: (i) standard parasite conditions (complete Schneider's Drosophila Medium, 23 °C); (ii) cell culture conditions (HMEC-1 medium, 37 °C and 5% CO2); (iii) co-incubation with HMEC-1 in cell culture conditions. The percent of infected macrophages was then evaluated after 24 h of incubation. Figure 2F clearly shows that cell culture conditions and, especially, co-incubation with HMEC-1, significantly impaired the parasites' ability to infect human macrophages.

CXCL8 Production by Endothelial Cells Treated with Leishmania spp. Promastigotes
HMEC-1 were left untreated (control) or co-incubated with promastigotes of L. infantum, L. tropica or L. braziliensis at different cell:Leishmania ratio for 24 h. Thereafter, CXCL8 levels were measured in cell supernatants. Exposure of HMEC-1 to promastigotes of Leishmania spp. stimulated the production of CXCL8 in a concentration-dependent manner ( Figure 3A-C). A cell:Leishmania ratio of 1:10 was sufficient for inducing significant amount of CXCL8, compared to untreated controls. In addition, time course experiments indicated that after 2 h of co-incubation, a significant amount of CXCL8 was induced by both L. infantum and L. tropica ( Figure 3D). When, LPS or TNFα, were used as positive controls, only TNFα, but not LPS, induced significant levels of CXCL8 after 2 h of stimulation ( Figure 3E). Next, in order to assess whether the selected incubation conditions might also alter promastigotes infectivity, the ability of L. infantum or L. tropica promastigotes (1:10, cell:parasite ratio) to infect PMA differentiated THP-1 macrophages was examined after 4 h of incubation in different culture conditions: (i) standard parasite conditions (complete Schneider's Drosophila Medium, 23 • C); (ii) cell culture conditions (HMEC-1 medium, 37 • C and 5% CO 2 ); (iii) co-incubation with HMEC-1 in cell culture conditions. The percent of infected macrophages was then evaluated after 24 h of incubation. Figure 2F clearly shows that cell culture conditions and, especially, co-incubation with HMEC-1, significantly impaired the parasites' ability to infect human macrophages.

CXCL8 Production by Endothelial Cells Treated with Leishmania spp. Promastigotes
HMEC-1 were left untreated (control) or co-incubated with promastigotes of L. infantum, L. tropica or L. braziliensis at different cell: Leishmania ratio for 24 h. Thereafter, CXCL8 levels were measured in cell supernatants. Exposure of HMEC-1 to promastigotes of Leishmania spp. stimulated the production of CXCL8 in a concentration-dependent manner ( Figure 3A-C). A cell: Leishmania ratio of 1:10 was sufficient for inducing significant amount of CXCL8, compared to untreated controls. In addition, time course experiments indicated that after 2 h of co-incubation, a significant amount of CXCL8 was induced by both L. infantum and L. tropica ( Figure 3D). When, LPS or TNFα, were used as positive controls, only TNFα, but not LPS, induced significant levels of CXCL8 after 2 h of stimulation ( Figure 3E).  To verify whether CXCL8 production was induced by phagocytosis of Leishmania promastigotes, HMEC-1 were co-incubated with L. infantum for 24 h. Cell monolayers were then extensively washed to remove non-internalized parasites and further stained with Giemsa. As shown in Figure 4, before washing, many Leishmania promastigotes were present around endothelial cells ( Figure 4A). However, after washing, all Leishmania promastigotes were successfully removed, showing that cells were not infected by parasites ( Figure 4B). To verify whether CXCL8 production was induced by phagocytosis of Leishmania promastigotes, HMEC-1 were co-incubated with L. infantum for 24 h. Cell monolayers were then extensively washed to remove non-internalized parasites and further stained with Giemsa. As shown in Figure 4, before washing, many Leishmania promastigotes were present around endothelial cells ( Figure 4A). However, after washing, all Leishmania promastigotes were successfully removed, showing that cells were not infected by parasites ( Figure 4B).
Since Leishmania promastigotes can interact with Toll-like receptors (TLRs), the involvement of TLR4 in the production of CXCL8 from Leishmania-induced HMEC-1 was subsequently evaluated. In the presence of an anti-TLR4 antibody, the production of CXCL8 induced by Leishmania infantum or LPS (the main ligand of TLR4) was reduced by 20.4 and 28.3%, respectively ( Figure 4C). Since TLR-4 activates the NF-κB pathway to regulate the expression of proinflammatory mediators, artemisinin, a known inhibitor of NF-κB was used. Artemisinin reduced the Leishmania-induced CXCL-8 production by 16.8%. When LPS was used as positive control, artemisinin induced a significant reduction of CXCL-8.
Unlike CXCL8, promastigotes of L. infantum did not alter the basal production of CCL5 ( Figure 5A), although they reduced the production of CCL5 induced by TNF-α in a dose-dependent manner ( Figure 5B). This effect seems specific for CCL5 since L. infantum did not reduce the production of CXCL8 induced by TNF-α ( Figure 5C). promastigotes, HMEC-1 were co-incubated with L. infantum for 24 h. Cell monolayers were then extensively washed to remove non-internalized parasites and further stained with Giemsa. As shown in Figure 4, before washing, many Leishmania promastigotes were present around endothelial cells ( Figure 4A). However, after washing, all Leishmania promastigotes were successfully removed, showing that cells were not infected by parasites ( Figure 4B).  Since Leishmania promastigotes can interact with Toll-like receptors (TLRs), the involvement of TLR4 in the production of CXCL8 from Leishmania-induced HMEC-1 was subsequently evaluated. In the presence of an anti-TLR4 antibody, the production of CXCL8 induced by Leishmania infantum or LPS (the main ligand of TLR4) was reduced by 20.4 and 28.3%, respectively ( Figure 4C). Since TLR-4 activates the NF-κB pathway to regulate the expression of proinflammatory mediators, artemisinin, a known inhibitor of NF-κB was used. Artemisinin reduced the Leishmania-induced CXCL-8 production by 16.8%. When LPS was used as positive control, artemisinin induced a significant reduction of CXCL-8.
Unlike CXCL8, promastigotes of L. infantum did not alter the basal production of CCL5 ( Figure 5A), although they reduced the production of CCL5 induced by TNF-α in a dose-dependent manner ( Figure 5B). This effect seems specific for CCL5 since L. infantum did not reduce the production of CXCL8 induced by TNF-α ( Figure 5C).

Neutrophils Recruitment by Supernatant of Endothelial Cells Treated with Leishmania Promastigotes
To verify whether the supernatants of the endothelial cells incubated with Leishmania promastigotes contain either active CXCL8, other chemokines or both, able to attract

Neutrophils Recruitment by Supernatant of Endothelial Cells Treated with Leishmania Promastigotes
To verify whether the supernatants of the endothelial cells incubated with Leishmania promastigotes contain either active CXCL8, other chemokines or both, able to attract leukocytes, cell migration was evaluated in a transwell system using human PBLs and conditioned medium from HMEC-1 treated with L. tropica or L. infantum promastigotes. Conditioned medium from HMEC-1 alone or HMEC-1 stimulated with LPS or TNF-α were Pathogens 2021, 10, 1380 7 of 14 used as negative and positive controls, respectively. After 2 h of incubation, the number of migrated leukocytes and the percentage of the different leukocyte subpopulations was further evaluated. As shown in Figure 6A, the number of migrated leukocytes towards Leishmania-stimulated HMEC-1 conditioned medium was higher than those migrated to the conditioned medium from unstimulated control HMEC-1. The fold change in the mean of migrated cells towards Leishmania conditioned medium relative to unstimulated control cells ranged from 3.6 to 5.2 (n = 4) and 1.9 to 5.6 (n = 4) for L. infantum and L. tropica, respectively. The fold change for LPS and TNF-α relative to unstimulated cells ranged from 3.0 to 6.7 (n = 3) and from 2.9 to 5 (n = 3). The number of migrated leukocytes towards Leishmania-stimulated HMEC-1 conditioned medium was comparable to that induced by conditioned medium from TNF-α or LPS-stimulated HMEC-1, suggesting a strong production of functional chemokines induced by the parasites.
Pathogens 2021, 10, x FOR PEER REVIEW 7 of 14 leukocytes, cell migration was evaluated in a transwell system using human PBLs and conditioned medium from HMEC-1 treated with L. tropica or L. infantum promastigotes. Conditioned medium from HMEC-1 alone or HMEC-1 stimulated with LPS or TNF-α were used as negative and positive controls, respectively. After 2 h of incubation, the number of migrated leukocytes and the percentage of the different leukocyte subpopulations was further evaluated. As shown in Figure 6A, the number of migrated leukocytes towards Leishmania-stimulated HMEC-1 conditioned medium was higher than those migrated to the conditioned medium from unstimulated control HMEC-1. The fold change in the mean of migrated cells towards Leishmania conditioned medium relative to unstimulated control cells ranged from 3.6 to 5.2 (n = 4) and 1.9 to 5.6 (n = 4) for L. infantum and L. tropica, respectively. The fold change for LPS and TNF-α relative to unstimulated cells ranged from 3.0 to 6.7 (n = 3) and from 2.9 to 5 (n = 3). The number of migrated leukocytes towards Leishmania-stimulated HMEC-1 conditioned medium was comparable to that induced by conditioned medium from TNF-α or LPS-stimulated HMEC-1, suggesting a strong production of functional chemokines induced by the parasites. The percentages of the different migrated leukocyte subpopulations were then determined in each group. As shown in Figure 6B, after 2 h of migration, migrated cells were mostly neutrophils in all groups: 73% in medium from unstimulated cells, and more than 93% in L. infantum and L. tropica conditioned medium. After 24 h of chemotaxis, a higher percentage of monocytes and lymphocytes was observed in all groups. However, neutrophils continued to be the predominant subpopulation in conditioned media from stimulated cells ( Figure 6C).
Neutrophils migrated after 2 h of chemotaxis were then recovered and incubated with L. infantum and L. tropica promastigotes in order to verify their ability to phagocytize parasites. After 30 min of incubation, some amastigotes were visible inside neutrophils (arrow in Figure 6D).

Discussion
After infected sand fly bites, neutrophils are rapidly recruited to the site of inoculation, representing the first cells infected by Leishmania promastigotes [10]. Still, the immune mechanisms governing the sustained and intensified neutrophil recruitment remain mostly undefined. It is expected that different factors deriving from the vector, parasite itself and host's cells are all involved in the onset of infection. It is known that sandfly-derived factors, including salivary proteins and gut microbiota, act as chemoattractants for neutrophils [13,20]. Recently, it was demonstrated that members of sand fly yellow salivary proteins can induce in vitro chemotaxis of neutrophils [11]. Moreover, bacteria egested from sandflies activate the inflammasome along with the production of IL-1β, which acts as a chemotactic factor [13]. In the present study, we provide in vitro evidence that endothelial cells actively induce neutrophil chemotaxis by producing CXCL8, one of the most effective chemoattractants for neutrophils.
Metacyclic flagellated promastigotes, present in the anterior part of the midgut of the sand fly vector, differentiate into amastigotes inside the mammalian host's phagocytic cells. This differentiation is modulated by environmental changes, such as pH and temperature, but also by H 2 O 2 and iron uptake [21][22][23]. Indeed, changes in pH and temperature can induce loss of Leishmania viability through the production of reactive oxygen species [24]. Here, we also show in vitro, in cell culture conditions (mammalian cell culture medium, 37 • C, 5% CO 2 ), that the parasites assume an amastigote-like morphology, stop to proliferate and some of them die. However, only 4 h in cell culture conditions resulted in a significant reduction in the parasites' ability to infect macrophages. Infectivity was further decreased by co-culturing the parasites in the presence of HMEC-1. Therefore, even if the parasites assumed an amastigote-like morphology, they lost infectivity, suggesting that the differentiation process was not complete. This is consistent with the observation that amastigotes exhibit higher infectivity than promastigotes from the same Leishmania species [25]. The causes of this loss of infectivity albeit associated with an amastigotelike phenotype, are presently unknown. In co-culture conditions, endothelial cells may deprive the medium of the nutrients necessary for parasite viability and differentiation. Alternatively, HMEC-1 could produce toxic mediators affecting parasite viability. Most likely, however, the parasites' loss of infectivity might reflect the fact that the parasites must quickly infect host cells and thus interact with cells present in the microenvironment, such as endothelial cells, to recruit phagocytic cells. We have indeed demonstrated that endothelial cells incubated with promastigotes of different Leishmania species produced CXCL8, a potent neutrophil chemotactic cytokine capable of delaying their apoptosis [16,26]. The induction of CXCL8 by Leishmania has also been demonstrated in a murine model and in human infections. Upon experimental infection with L. major, macrophage inflammatory protein (MIP)-2 and keratinocyte-derived cytokine (KC; also known as CXCL1), the functional murine homologues of human CXCL8 are rapidly produced in the skin [10]. Moreover, immunohistochemistry studies demonstrated strong expression of CXCL8 in dermal lesions of patients infected with L. tropica [27]. CXCL8 has also been shown to be produced by human monocytes stimulated by sandfly salivary gland homogenates [28].
Here, we showed that as little as 2 h of incubation in the presence of Leishmania promastigotes is enough for inducing CXCL8 secretion by HMEC-1. CXCL8 in endothelial cells is primarily stored in the secretory organelles, the Weibel-Palade bodies that can be rapidly exocytosed in response to different stimuli, such as thrombin and histamine [29]. The rapid release of CXCL8 from HMEC-1 may indicate that Leishmania promastigotes stimulate HMEC-1 to release the preformed chemokine present in Weibel-Palade organelles. This is different from LPS-induced production of CXCL8, which peaks after 24 h of incubation indicating neo-synthesis of the chemokine.
To investigate the mechanisms by which Leishmania promastigotes induce the production of CXCL8 by HMEC-1, the role of phagocytosis and of TLR4 receptor was investigated. Although endothelial cells are non-professional phagocytes, they can internalize apoptotic neutrophils, apoptotic bodies, platelets and pathogens such as Listeria monocytogenes [30][31][32][33].
In our experiments, HMEC-1 did not internalize Leishmania promastigotes, and, to the best of our knowledge, there is no evidence of Leishmania phagocytosis by endothelial cells in vivo.
The involvement of the innate immune receptor TLR4 was investigated by using an anti-TLR4 antibody. TLR4, the sensing receptor for LPS, is expressed in immune and non-immune cells, including endothelial cells [34], being involved in the activation of the proinflammatory response and the production of cytokines. In the murine model of leishmaniasis, TLR4 is required for efficient parasite control [35] and in vitro studies demonstrated that GP29, a L. donovani derived glycoprotein, induced TNF-α and IL-12 production through TLR4 activation [36]. By showing a partial reduction of CXCL8 production, we can hypothesize that TLR4 is involved in CXCL8 production induced by Leishmania. The NF-κB pathway can be activated by pattern-recognition receptors, such as TLR, leading to the modulation of a large array of genes involved in inflammatory responses [37]. Artemisinin, an antimalarial agent, known to inhibit nuclear translocation of NF-kB complex [38] inhibited the production of CXCL-8 induced by Leishmania, suggesting the involvement of NF-κB pathway. However, since the reduction of CXCL-8 production was only partial, we cannot exclude that other signals or other TLRs could contribute to cell activation. It is known that Leishmania lipophosphoglycan (LPG), the major parasite ligand for macrophage adhesion, activates innate immune signaling pathways via TLR2 [39]. Furthermore, endosomal TLR9 can recognizes unmethylated CpG DNA sequences of Leishmania [40].
Interestingly, all three of the species used in this study, L. infantum, L. tropica and L. brazilensis, induced CXCL8 production. This indicates that the ability to stimulate endothelial cells is not species-specific, and not even related to the different pathogenesis, but it is a general feature of different Leishmania species. This contrasts with the host's specific immunity to Leishmania, which is often species-specific and can either promote or control the infection [41]. However, being obligate intracellular parasites, promastigotes are rapidly destroyed in the extracellular tissues [42]. Therefore, all Leishmania species need to rapidly colonize host cells in order to survive and establish infection.
L. infantum promastigotes failed to induce CCL5 by HMEC-1 cells, but inhibited its production induced by TNF-α. This seems to be specific for CCL5, since CXCL8 production induced by TNF-α was not affected by L. infantum. CCL5, also known as RANTES, is important for the recruitment and development of Th1 cells, which are responsible for the control of the infection [43]. In addition, it attracts and activates many different immune cells including T cells, dendritic cells and NK cells to the sites of infection. Indeed, CCL5 induces IL-12 [44] and IFN-γ [45]. Moreover, it has been described that CCL5 contributes to the resistance to L. major infection [46]. In fact, in vivo treatment with Met-RANTES, an antagonist of CCR1 and CCR5, resulted in animals being more susceptible to the infection and in an increase in lesion size [46]. Therefore, the observed inhibition of CCL5 by the parasite, may strongly contribute to the establishment of the infection.
The presence of functional chemokines in the conditioned medium from endothelial cells treated with Leishmania promastigotes was confirmed with the transwell chemotaxis assay. The Leishmania-HMEC-1 conditioned medium recruited higher numbers of leukocytes than medium from unstimulated HMEC-1 cells. Furthermore, the total number of migrating leukocytes was comparable to that of leukocytes recruited by conditioned medium from HMEC-1 stimulated with LPS or TNF-α, two potent proinflammatory stimuli. After 2 h of chemotaxis, the migrated cells were mostly neutrophils in all studied groups. Neutrophils, the most abundant leukocyte subpopulation in the blood, express different surface receptors, which help them control their migration and behavior [47], and often serves as the first responders to a variety of inflammatory stimuli. After 24 h of migration, a higher percentage of monocytes was detected in conditioned media from unstimulated cells, whereas neutrophils still predominated in conditioned medium from Leishmania-stimulated cells. Macrophages represent the ultimate host cells for Leishmania, where parasites can survive and multiply.
In conclusion, our data provide novel insights into how Leishmania spp. promastigotes interact with endothelial cells, and generate a microenvironment able to attract phagocytic cells through the production of chemokines. Parasites must quickly invade phagocytic cells before losing their ability to infect. It is likely that a combination of signals from vector, parasite and host contribute to the early steps of natural infection. In addition to the sandfly saliva stimuli, the high production of CXCL8 by endothelial cells at the site of parasite deposition, may indeed contribute to the recruitment of PMNs, which provide shelter to parasites and allow them to survive and multiply, supporting the development of the disease. The concomitant reduction of CCL5 may further contribute to parasite survival and adaptation to the new host.

Endothelial Cells Culture
The long-term cell line of dermal microvascular endothelial cells (HMEC-1) immortalized by SV 40 large T antigen [48] was kindly provided by the Center for Disease Control, Atlanta, GA. Cells were maintained in MCDB 131 medium supplemented with 10% fetal calf serum, 10 ng/mL of epidermal growth factor, 1 µg/mL of hydrocortisone, 2 mM glutamine, 100 units/mL of penicillin, 100 µg/mL of streptomycin and 20 mM Hepes buffer, pH 7.4.

Morphological Studies of Leishmania Promastigotes
Promastigotes of L. infantum were counted and 10 6 parasites/well were seeded in 24-well flat bottom tissue culture clusters. Promastigotes were then cultured in standard conditions (complete Schneider's Drosophila Medium, 23 • C) or in cell culture conditions (cell medium, 37 • C, 5% CO 2 ) for 4 and 24 h. After incubation, parasites were recovered and a thin layer of parasites smeared on a slide. Slides were fixed with methanol and stained with Giemsa. Morphology was evaluated by light microscopy using a 100× oil immersion objective.

Cell and Leishmania spp. Viability
Cell and Leishmania spp. viability was evaluated using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay [49]. After incubation of the cells in different conditions, 20 µL of MTT solution (5 mg/mL in PBS) were added to the cells for 3 h at 37 • C in the dark. The supernatants were then discarded and the dark blue formazan crystals dissolved using 100 µL of lysis buffer containing 20% (wt/vol) sodium dodecylsulfate, 40% N,N-dimethylformamide (pH 4.7 in 80% acetic acid). The plates were then read on a Synergy 4 (Biotek) microplate reader at a test wavelength of 550 nm and at a reference wavelength of 650 nm.
At the end of each treatment, supernatants were collected and used for chemokine determinations or transwell migration assay (conditioned media).

Chemokine Determination
CXCL8 and CCL5 were measured in cell supernatants by DuoSet ELISA Kit (R&D System) following the manufacturer's instructions.

Isolation of Peripheral Blood Lymphocytes (PBL) and Transwell Migration Assay
PBLs were isolated from fresh peripheral blood of human donors. Blood was diluted 1:1 with RPMI1640, layered on cold Ficoll-Paque and centrifuged at 500× g for 30 min [50]. PBLs were recovered, washed with RPMI and counted.
Supernatants from endothelial cells treated with different stimuli (conditioned media) were introduced into the lower compartment of a 24 transwell plate (3 µm pores). PBLs (1 × 10 5 ) were added to the upper compartment and the transwell plate was incubated for 2 and 24 h at 37 • C in 5% CO 2 incubator. At the end of incubation, the insert was carefully taken out to remove non-migrated cells. Migrated cells on the lower side were counted by trypan blue using light microscopy and 2 × 10 5 cells in 100 µL were used for cytospin preparation. Cells were centrifuged at 500 rpm for 5 min in a cytospin centrifuge. Slides were fixed with methanol, stained with Giemsa and the percentages of the different leucocyte subpopulations counted by microscopic observation.

Statistical Analyses
All data were obtained from three independent experiments and the results are shown as mean ± standard deviation or as a representative experiment. Differences between groups were analyzed for statistical significance by using 1-way or 2-way ANOVA tests followed by post hoc multiple comparison tests (Dunnett's, Sidak's or Tukey's). Statistical significance was set at p < 0.05.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.