Leishmania infantum Infection of Primary Human Myeloid Cells

Circulating phagocytic cells often serve as cellular targets for a large number of pathogens such as Leishmania parasites. Studying primary human cells in an infectious context requires lengthy procedures for cell isolation that may affect the analysis performed. Using whole blood and a no-lyse and no-wash flow cytometric assay (NoNo assay), we monitored the Leishmania infantum infection of primary human cells. We demonstrated, using fluorescent parasites, that among monocyte cell populations, L. infantum preferentially infects classical (CD14+CD16−) and intermediate (CD14+CD16+) primary human monocytes in whole blood. Because classical monocytes are the preponderant population, they represent the larger L. infantum reservoir. Moreover, we also found that, concomitantly to monocyte infection, a subset of PMNs is infected early in whole blood. Of interest, in whole blood, PMNs are less infected compared to classical monocytes. Overall, by using this NoNo assay, we provided a novel avenue in our understanding of host–leishmania interactions.


Introduction
Leishmania donovani and infantum are intracellular parasites belonging to the Trypanosomatidae family that are the causative agents of visceral leishmaniasis (VL) [1]. This disease affects vulnerable populations, and approximately 90,000 infections occur each year [1,2]. The clinical manifestations of VL are fever, anemia, weight loss, splenomegaly and hepatomegaly, leading to death if left untreated [3]. Leishmania parasites are transmitted as promastigotes (extracellular forms) to the mammalian host during phlebotomine sandfly blood meals. Thus, parasites are introduced into the skin following sand fly inoculation [4,5]. In turn, tissue damage due to the inoculation of parasites causes the recruitment and accumulation of immune cells at the bite's site, involving monocytes and polymorphonuclear neutrophils (PMNs). In addition, it has been reported that sand fly salivary proteins attract and activate host cells, and taking advantage of this inflammatory response, parasites invade their host [6]. PMNs may act as a Trojan horse, silently transmitting parasites to infiltrate mononuclear cells [4,5,[7][8][9][10]. Furthermore, Leishmania hijack host machinery to survive and multiply into host cells [11,12]. Of importance, the parasite has also been detected in the peripheral blood of patients with cutaneous or visceral leishmaniasis [13,14]. Risks of the contamination of leishmaniasis via blood transfusion or organ donation (bone marrow and liver) in endemic regions where asymptomatic cases predominate are not negligible [14][15][16]. However, little attention has been paid regarding the interaction between viscerotropic strains and monocyte subsets. vided informed consent to EFS. In order to monitor the dynamic of interaction between the cells and parasites, the infection rate was assessed at two times (4 and 24 h). Whole blood (1 mL), transferred in a 15 mL tube (Falcon tube, BD Biosciences, Le Pont-de-Claix, France), was infected with either live parasites at three doses (1 × 10 6 , 5 × 10 6 and 10 × 10 6 of parasites per mL) or HK parasites (10 × 10 6 parasites per mL). For each condition, whole blood was incubated at 37 • C and 5% CO 2 before analysis.

Statistical Analyses
All values were expressed as means ± SEM. Statistical analysis was performed using the non-parametric paired Wilconxon test or the Two-Way ANOVA test with a Tukey analysis and a Spearman test for correlation. Comparisons were considered significant when p < 0.05. All statistical analyses were performed with GraphPad Prism 8 software (San Diego, CA, USA).

L. infantum Infects Monocytes in Whole Blood
The method that we used to monitor infection exploited the difference between red blood cells that contain hemoglobin, which absorbs violet laser (405 nm) light, and leukocytes, which do not. Thus, by coupling 405 nm (violet) and 488 nm (blue) laser excitations to study whole-blood cells, a unique side scatter (SSC) pattern was observed ( Figure 1A(a)), separating erythrocytes from leukocytes and platelets [42]. Cell populations were then discriminated using a cocktail of specific antibodies. The gating strategy used to identify the myeloid populations is shown in Figure 1. After selecting white cells from whole blood, leukocytes were defined on SSC and FSC parameters ( Figure 1A(b)), and the analysis was performed on single cells ( Figure 1A(c)) and excluded dead cells using the LIVE/DEAD probe. Then, T lymphocytes (CD3 + ) were excluded ( Figure 1A(d)), as well as the CD3 − CD8 + (natural killer cells) and CD20 + (B cells) populations ( Figure 1A(e)). To discriminate myeloid cell populations, we used HLA-DR expression, because it was only expressed on monocytes. Thus, by combining HLA-DR and CD16 markers, ( Figure 1A(f)), we separated PMNs (CD16 + DR − ) from monocytes (DR + ) independently to be CD16 − or CD16 + . we separated PMNs (CD16 + DR − ) from monocytes (DR + ) independently to be CD16 − or CD16 + .  To distinguish between the infection versus uptake of parasites via myeloid cells, L. infantum was killed at 60 • C. Firstly, we validated the CFSE staining of labeled live and dead (heat-killed) parasites showing similar fluorescence intensities after in vitro treatment for both live (geometric mean fluorescence intensity (geo MFI), 5782) and HK (geo MFI, 5938) parasites ( Figure 1B). Then, we assessed the percentages of infected monocytes (DR + ) via flow cytometry after 4 h of infection with an increasing dose of parasites in whole blood ( Figure 1C(a)). Hence, with 1 × 10 6 parasites per mL, 11.4 ± 2.4% of monocytes were infected ( Figure 1C(a-b)), whereas with higher doses (5 × 10 6 or 10 × 10 6 parasites), the percentages of L. infantum-infected monocytes were relatively similar (31.4 ± 5.0% and 39.8 ± 5.0%, respectively). The infection of monocytes increased with the time of parasite infection ( Figure 1C(b)). Thus, the percentages of infected cells at 24 h increased to 22.2 ± 8.0% with 1 × 10 6 parasites. With 5 × 10 6 and 10 × 10 6 parasites, these rates reached a plateau of 50.6 ± 9.3% and 52.5 ± 8.5%, respectively ( Figure 1C(b)). Of note, the percentages of HK-labeled monocytes were 30.1 ± 3.7% at 4 h and 39.9 ± 10.3% at 24 h, which were lower compared to live parasites, particularly early after infection. Thus, L. infantum promastigotes can infect monocytes in whole blood without cell isolation.

L. infantum Preferentially Infects CL and ITM Monocytes in Whole Blood
To further characterize the nature of infected monocyte cell subsets, we analyzed the three main subsets, CL, ITM and NCL ( Figure 2A). Furthermore, the double-negative population (DN: HLA-DR + CD14 − CD16 − ) was also analyzed ( Figure 2A). Firstly, we analyzed the distribution of the parasite among monocyte cell subsets ( Figure 2B). According to the intensity of L. infantum-CFSE + in each quadrant of the heat-map view ( Figure 2B), 4 h after infection, most of the L. infantum parasites were localized in the CL monocytic region (88.4 ± 1.8% at the lowest dose), whereas less than 5% of the parasites were observed in the other regions. Increasing the doses of parasites, we did not observe major changes in the distribution of parasites between the different subsets ( Figure 2C). HK parasites were also mostly observed in the region of CL (78.3 ± 5.5%) compared to in ITM (3.5 ± 0.8%), NCL (8.1 ± 2.4%) and DN subsets (10.0 ± 4.4%) ( Figure 2C). To validate the concept that parasites infect cells in whole blood and do not remain at the surface, we purified monocytes using anti-CD14 beads which included both CL and ITM. The fraction of cells was analyzed using confocal microscopic analysis. The images showed that both live and dead L. infantum parasites were detected inside the cells ( Figure 2D). Interestingly, the images showed the transition of the living parasites into amastigotes (1-5 µm), while, as expected, the HK parasites remained in the promastigote form (10-15 µm).
Because CL monocytes are the preponderant population in whole blood, using flow cytometry, we assessed the frequency of infected cells in each subset at two time points (4 and 24 h, Figure 3). It is interesting to note that at higher doses (5 × 10 6 and 10 × 10 6 parasites), the rates of infection were quite similar at both time points ( Figure 3A,B), reaching 60.8 ± 10% of CL infection with 10 × 10 6 parasites at 4 h and 74.6 ± 10.5% at 24 h ( Figure 3B), whereas this frequency of infected CL was 15.2 ± 3.7% with 1 × 10 6 parasites at 4 h and tended to be higher at 24 h (26.6 ± 10% p-value = 0.1484).  cytometry, we assessed the frequency of infected cells in each subset at two time points (4 and 24 h, Figure 3). It is interesting to note that at higher doses (5 × 10 6 and 10 × 10 6 para sites), the rates of infection were quite similar at both time points ( Figure 3A,B), reaching 60.8 ± 10% of CL infection with 10x10 6 parasites at 4 h and 74.6 ± 10.5% at 24 h ( Figure 3B) whereas this frequency of infected CL was 15.2 ± 3.7% with 1 × 10 6 parasites at 4 h and tended to be higher at 24 h (26.6 ± 10% p-value = 0.1484).  Interestingly, whereas the ITM population is not the predominant population in whole blood, we found that the frequency of L. infantum-infected ITM populations was elevated at 4 h (51.0 ± 9.3% of ITM populations were infected with 10 × 10 6 parasites) and quite similar to that observed for CL populations (60.8 ± 10%). After 24 h, the rate of infection increased with a maximum reached with 10 × 10 6 parasites (74.6 ± 10.5% for CL and 65.0 ± 12.3% for ITM populations). Of importance, at lower doses, this increase was particularly significantly higher after 24 h compared to 4 h (29.4 ± 9.8% versus 12.1 ± 2.3%, p-value = 0.0391) of infection, reaching percentages closer to those observed in CL populations ( Figure 3A,B). Of note, while the same number of parasites was used for HK and live leishmania to infect whole blood, the frequencies HK parasites' uptake by the cells were lower at 4 h (CL, 49.8 ± 10.4% and ITM, 35.6 ± 9.3%) to those observed with live parasites in CL (60.8 ± 10%) and ITM (51.0 ± 9.3%) subsets (CL, p-value = 0.0391 and ITM, p-value = 0.1875).
In contrast to that observed for CL and ITM populations, early after infection (4 h) the frequencies of L. infantum-infected NCL and DN populations were low ( Figure 3A,B), even at the higher dose of infection (15.6 ± 2.6% and 6.1 ± 1.6%, respectively). After 24 h, the rates of infection increased for these two subsets with a maximum reached with 10 × 10 6 parasites (25.5 ± 9.0% for NCL, and 21.1 ± 6.6% for DN). We also analyzed the geo MFI of infected CL populations at 4 and 24 h, showing quite similar intensities, except for HK, in which the geo MFI was higher at 24 h compared to live parasites ( Figure 3C).
Altogether, our results demonstrated that CL and ITM populations are preferentially infected cell subsets in whole blood despite the fact that the latter is less abundant compared to CL populations.

Infection by L. infantum Impacts Monocyte Cell Subsets in Whole Blood
Because L. infantum could impact the nature of infected monocyte cell subsets, we analyzed their percentages in whole blood. As mentioned, CL was the most abundant monocyte subset (73.9 ± 3.6%) in the whole blood of healthy donors, whereas ITM and NCL subsets represented 3.2 ± 0.4% and 8.0 ± 0.4%, respectively ( Figure 4). The DN subsets represented 15.5 ± 4.1% of blood monocytes ( Figure 4A). With 10 × 10 6 parasites, CL represented 64.5 ± 6.3% in the presence of live parasites and 54.1 ± 9.0% for HK in comparison to 73.9 ± 3.6% in non-infected cells. We did not observe such significant differences for ITM and NCL populations ( Figure 4A). Thus, both live and HK parasites induced the lowest CL levels in whole blood. This decrease in the percentage of CL was associated with an increase in the percentage of DN, which may be suggestive of a CD14 down-regulation in CL (Figure 4(Bb)). The absence of a significant difference between live parasites was related to the fact that a decrease was only observed in four donors out of the six. Thus, we found that the percentage of CL was negatively correlated with the rate of CL infection 4 h post-infection ( Figure 4A,(Ba)). Thus, the lowest infection rate was associated with the highest percentage of CL. Therefore, the increase rate of DN infection that we observed in whole blood may represent infected, CL in which CD14 is lost.

L. infantum Infects PMNs in Whole Blood
Because this assay allowed us to monitor concomitantly with the monocyte PMN population, we assessed the rate of PMN infection in the same whole blood samples. Early after infection, we detected PMN infection ( Figure 5A). The percentages of CFSE + -PMN reached 10 ± 4.3% at the dose of 1x10 6 ( Figure 5B). Little impact on the infection rate was observed with higher doses, where 5 × 10 6 and 10 × 10 6 parasites led to 14.4 ± 5.0% and 17.0 ± 5.1% of infected PMNs, respectively. These percentages were lower compared to the extent of infection of CL (60.8 ± 10% at the same time using 10 × 10 6 parasites). Interestingly, PMNs were capable to phagocytize HK parasites (25.37 ± 10.9% of PMNs at 4 h) ( Figure 5B). Altogether, these results clarify the impact of L. infantum infection on monocyte cell subsets in whole blood.

L. infantum Infects PMNs in Whole Blood
Because this assay allowed us to monitor concomitantly with the monocyte PMN population, we assessed the rate of PMN infection in the same whole blood samples. Early after infection, we detected PMN infection ( Figure 5A). The percentages of CFSE + -PMN reached 10 ± 4.3% at the dose of 1 × 10 6 ( Figure 5B). Little impact on the infection rate was observed with higher doses, where 5 × 10 6 and 10 × 10 6 parasites led to 14.4 ± 5.0% and 17.0 ± 5.1% of infected PMNs, respectively. These percentages were lower compared to the extent of infection of CL (60.8 ± 10% at the same time using 10 × 10 6 parasites). Interestingly, PMNs were capable to phagocytize HK parasites (25.37 ± 10.9% of PMNs at 4 h) ( Figure 5B).  Our results also indicated that the rate of infection was quite similar between 4 and 24 h with 1 × 10 6 parasites (9.2 ± 5.5%), while these proportions declined using higher doses (10 × 10 6 , 8.8 ± 2.8% compared to 17.0 ± 5.1% (p-value = 0.0234) ( Figure 5A,B). The geo MFI indicated the lower intensity of infected PMNs with live parasites at 24 h compared to 4 h as well as for HK parasites (Figure 5C), and the intensity differed to that observed in the same conditions for CL geo MFI ( Figure 3C).
As shown here, among PMNs, we identified two populations, in which one expressed a higher level of CD16 expression (CD16 bright ). By plotting CD14 against CD16, this population was represented in whole blood (49.5 ± 9.9% ranging from 21 to 77.3% of PMNs in the non-infected condition) ( Figure 5D). Thus, we defined PMNs as CD14 low CD16 + and CD14 bright CD16 bright cell subsets. We observed that the percentages of CD14 bright CD16 bright PMNs were lower in the presence of increased doses of parasites compared to the noninfected condition ( Figure 5D-E). Thus, more than 40% of this population disappeared 4 h after infection and even more drastically after 24 h at the highest doses (5 × 10 6 and 10 × 10 6 parasites). Analyzing the tropism of L. infantum-infected PMNs in more detail, we observed that this CD14 bright CD16 bright population displayed most of the CFSE-labeled L. infantum parasites early after infection, as shown by the heat-map view ( Figure 5D). Thus, the diminution of these PMN subsets at the highest dose may support the observation of the lowest infection rate observed at 24 h ( Figure 5B). When increasing the number of parasites, most CFSE-labeled L. infantum parasites were then detected in the CD14 low CD16 + population ( Figure 5D). Of interest, our results highlighted that at lower doses of parasites (1 × 10 6 parasites per mL), this PMN population persisted even after 24 h of infection in whole blood.
Collectively, our results highlighted that PMNs can be infected in whole blood, although the percentages of infected cells were lower as compared to CL monocytes. We further demonstrated the selective tropism of L. infantum infection for PMN subsets, in which the majority of the parasites at lower doses infected and persisted in CD14 bright CD16 bright , which may represent a "sanctuary" for L. infantum.

Discussion
This flow cytometric assay demonstrated the efficacy of monitoring human myeloid cells' infection via visceral Leishmania species directly on whole blood. This designed strategy allowed us to accurately study L. infantum target cells, demonstrating that promastigotes infect human whole-blood cells, in which CL and ITM populations are the preferential targets along with a preferential subset of PMNs. These results represent, to our knowledge, the first demonstration of human monocytic tropism via visceral parasite strains in whole blood.
Whereas numerous studies have described macrophage infection, less attention has been paid regarding human monocytic cell subsets. Classical monocytes represent 80-95% of circulating human monocytes, whereas ITM and NCL populations represent 2-5% and 2-10%, respectively [17][18][19]. Our results demonstrated that from whole blood, L. infantum infects CL and ITM populations preferentially. Thus, more than 50% of cell subsets were infected at the highest dose of parasites. This preference was also highlighted by the fact that even at lower doses, these two populations were targeted by parasites compared to NCL or DN cell subsets. However, given the proportion of CL populations in the blood, this population represents a major target for parasites. This propensity of infection differs from previous studies using a mucocutaneous Leishmania species. Indeed, the authors observed that more than 85% of the cells were infected with L. braziliensis amastigotes without any specific preference among the different monocytic cell subsets [20]. In contrast, Polari et al. [26]. reported that the infection of PBMCs via L. braziliensis promastigotes leads to a preferential infection of CL and ITM populations compared to NCL populations. A difference in the rate of infection could be related to the use of amastigote vs. promastigote parasites. Herein, L. infantum promastigotes also preferentially infected CL and ITM populations in whole blood. CL populations have a wide range of functions, including a great capacity for phagocytosis, ROS production and immune sensing [17], while ITM populations participate in Ag-presentation and subsequently in T-cell stimulation [18,19]. Furthermore, CL and ITM monocytes were shown to be better in phagocytizing GFPexpressing E. coli [43] in comparison to NCL monocytes. In our study, similarly to live parasites, we found that HK parasites were engulfed by CL and ITM populations better than NCL populations. Novais et al. [20] proposed that CL monocytes control L. braziliensis infection by ROS production, while ITM populations promote pathogenesis [20,43]. In the case of viscerotropic strains, it was revealed that monocytes of patients with VL had a decrease in ROS production [44]. Given the ability of CL and ITM populations to migrate into inflamed or damaged tissues, this preferential infection may provide support to parasite dissemination within tissues. In a study performed in rodents, the authors showed that pro-inflammatory monocytes (Ly6C Hi ), besides being a preferential target for L. donovani, were associated with a pathogenic role [45], promoting dissemination in the liver and spleen. Whether the infection of ITM populations contributes to favor parasite visceralization in humans remains an open question.
Concomitantly to monocyte infection, we also reported the infection of PMNs. Whereas other studies, using purified PMNs, reported that L. infantum can infect 60% of PMNs [44,46], our results indicated lower levels (less than 25%). In the setting of the experiment performed, using whole blood and not purified PMNs, we cannot exclude the possibility that those lower levels of PMN infection reflect competition between the different target cells. Moreover, our results also highlighted the preferential targeting of CD14 bright CD16 brigh PMNs. Thus, this approach by using whole blood revealed interesting findings. Firstly, the observation of this PMN subset can be explained by the absence of density gradient centrifugation, which modulates cell surface expression [40]. Secondly, this specific PMN population was lost by increasing the quantity of parasites and time of exposure, suggesting the short life of cells once infected at higher doses of parasites. This could be consistent with CD16 lowering, a marker of activated and apoptotic PMNs [47]. The role of Leishmania in modulating apoptosis as a mechanism employed by the parasites to manipulate PMNs and guarantee the establishment of a perennial infection [48] seems to be more complex. Several reports have proposed that PMNs are more prone to die when infected with L. braziliensis [49], L. infantum or L. major [50], whereas others have proposed that L. major mediates parasite survival [51]. Thus, human PMN activation and function are dependent on the parasite strain used [11,52,53]. Whereas at higher doses, this population, CD14 bright CD16 brigh , disappeared, at lower quantities (1.10 6 parasites), this population persisted and was infected. Recent advances suggest, like for monocytes, the existence of distinct PMN subtypes including N1 and N2-like phenotypes, in which the N2 subset is unable to kill parasites [27]. The observation of parasite persistence in this CD14 bright CD16 brigh at lower quantities may be indicative that this population represents the N2 population. Further experiments are needed to characterize this PMN population more in depth.
By using this approach to infect whole blood, our results also indicated that L. infantum infection impacts the proportion of CL monocytes in culture. Whereas it is not clearly defined whether L. infantum parasites bind to the membrane CD14 of myeloid cells, a lower level of CL populations was positively correlated with the rate of infection. This lower level may reflect CD14 shedding, which is a hallmark of activated monocytes [54,55]. Indeed, CD14 is a glycosylphosphatidyl-anchored membrane protein that can be a direct ligand for LPS and act as a co-receptor for TLR4 [54][55][56]. A classic explanation for CD14 shedding is that it could be a way for cells to protect themselves from activation. Thus, sCD14 could be in competition for PAMPs in binding limiting inflammation [55,57]. It was reported in some studies that increasing levels of sCD14 was correlated with severity [11,58] and was associated with the outcome of VL [54]. In this sense, parasites could, by an unknown mechanism, promote the release of sCD14 and thus decrease the inflammatory response by inhibiting macrophage activation [59,60]. Our results regarding lower levels of CD14 were different to the report from Melo et al. [54], which may reflect the culture conditions, since, instead of using purified macrophages, we used whole-blood cells, which include PMNs. Indeed, the human leukocyte elastase, a serine proteinase produced by PMNs, is directly involved in the cleavage of CD14 [61,62]. Thus, by using whole blood as a model of Leishmania infection, this may have reflected the in vivo situation, in which higher levels of sCD14 were reported in VL patients [58] or co-infected with HIV [63]. It is interesting to note that a decrease in the marker Ly6C has also been also observed in murine inflammatory monocytes infected with L. major, supporting the hypothesis that such negative modulation may provide a strategy for the parasite to colonize its host [9]. Thus, by analyzing the infection of whole-blood cells and by limiting cell manipulation and isolation, we provided novel insights about viscerotropic strains infecting primary human myeloid cell subsets. This flow-cytometric approach may provide a novel assay for the monitoring of Leishmania infection.