Distinct Trypanosoma cruzi isolates induce activation and apoptosis of human neutrophils

Neutrophils are critical players in the first line of defense against pathogens and in the activation of subsequent cellular responses. We aimed to determine the effects of the interaction of Trypanosoma cruzi with human neutrophils, using isolates of the two major discrete type units (DTUs) associated with Chagas’ disease in Latin America (clone Col1.7G2 and Y strain, DTU I and II, respectively). Thus, we used CFSE-stained trypomastigotes to measure neutrophil-T. cruzi interaction, neutrophil activation, cytokine expression and death, after infection with Col1.7G2 and Y strain. Our results show that the frequency of CFSE+ neutrophils, indicative of interaction, and CFSE intensity on a cell-per-cell basis were similar when comparing Col1.7G2 and Y strains. Interaction with T. cruzi increased neutrophil activation, as measured by CD282, CD284, TNF and IL-12 expression, although at different levels between the two strains. No change in IL-10 expression was observed after interaction of neutrophils with either strain. We observed that exposure to Y and Col1.7G2 caused marked neutrophil death. This was specific to neutrophils, since interaction of either strain with monocytes did not cause death. Our further analysis showed that neutrophil death was a result of apoptosis, which was associated with an upregulation of TNF-receptor, TNF and FasLigand, but not of Fas. Induction of TNF-associated neutrophil apoptosis by the different T. cruzi isolates may act as an effective common mechanism to decrease the host’s immune response and favor parasite survival.


Introduction
Polymorphonuclear neutrophil granulocytes play an important role in the first line of defense against pathogens and the activation of subsequent immune responses [1]. The bone marrow a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 were volunteers and provided written informed consent. This work was approved by the Ethical Committee of the Universidade Federal de Minas Gerais, under the protocol# ETIC077/06. Peripheral blood was collected from the donors by venipuncture.

Parasites
Tissue culture-derived trypomastigotes (TCT) of Col1.7G2 and Y strain were isolated from infected monolayers of LLC cells (from ATCC). LLC cells were infected using a ratio of five TCT: one host cell, and kept in DMEM enriched with 1% inactivated fetal calf serum (FCS), supplemented with antibiotics (penicillin at 500μ/mL and streptomycin at 0.5 mg/mL). After approximately 5 days, the TCT were collected from the supernatant, washed once by centrifugation with phosphate-buffered saline (PBS) pH 7.2 at 1000 x g for 10 min at 4˚C, and resuspended in RPMI enriched with 5% of inactivated human serum, antibiotics (penicillin at 500U/mL and streptomycin at 05 mg/mL) and 1mM of L-glutamine to a concentration of 6x10 7 TCT/mL. Parasites obtained in this manner were used for infecting peripheral blood cells from donors and adherent cells to continue the in vitro cycle. Contamination with amastigotes was always below 10%, as determined by light microscopy evaluation of all cultures.

T. cruzi interaction with peripheral blood cells in suspension and flow cytometry
Trypomastigotes from LLC cultures, obtained as described above, were labeled with CFSE (carboxyfluorescein diacetate succinimidyl ester-Molecular Probes C1157), using a protocol previously reported by us [23], with modifications, and incubated with peripheral blood. Briefly, 6 x 10 7 parasites were incubated with 5μM CFSE for 15 min at 37˚C, 5% CO 2 . Labeled parasites were washed three times with cold PBS + 10% of inactivated FCS by centrifugation at 1000 x g for 10 min at 4˚C. It has been previously demonstrated by our group that both strains stain similarly with CFSE [22].
Peripheral blood infections were performed with CFSE-labeled parasite in a proportion of 10 parasites/cell and incubated for 3 hours at 37˚C in 5% CO 2 . After the incubation period, cells were washed by centrifugation with PBS at 600 x g for 10 min at 4˚C to remove extracellular parasites and re-incubated for additional 12 hours with RPMI enriched with 5% of inactivated human serum, antibiotics (penicillin at 500U/mL and streptomycin at 05 mg/mL) and 1mM of L-glutamine. Brefeldin A (1μg/mL) was added for the last four hours of infection to prevent protein secretion. After incubation cells were washed with PBS by centrifugation at 600 x g for 10 min at 4˚C. At the end of the centrifugation, the erythrocytes were lysed using RBC "Lysing buffer" (Bio Legend, CA, USA), at a concentration of 20mL/1mL of peripheral blood, by incubating for 15 min at 20˚C in the dark. After the incubation, cells were washed three times by centrifugation and resuspended with PBS.
Analysis of expression of surface molecules and cytokines by peripheral blood using flow cytometry After lysis of erythrocytes, cells were immunostained and analyzed using multiparametric flow cytometry. We used a combination of monoclonal antibodies directed to surface molecules (PE-labeled CD282, CD284, and Annexin; PeCy7-labeled CD120b; APC-labeled CD14; BV421-Fas-L) and 7AAD. Intracellular cytokine expression and FAS expression were evaluated using PE labeled antibodies against Fas, TNF, IL-12/IL-23p40 and IL-10. For surface and intracellular straining, cells were treated as previously described by us [23]. After staining, cells were resuspended in PBS and acquired using a FACSCanto II (Becton & Dickinson, San Jose, CA, USA). An average of 100,000 events were acquired for each experiment. The analyses were performed using FlowJo 7.6.5 software (Tree Star Inc., Ashland, OR, USA).
Previous studies have shown that neutrophils present a low expression of CD14 (CD14low), whereas monocytes express higher intensity of CD14 (CD14high) [24,25]. Thus, selection of the neutrophil population was made according to this strategy, by selecting granulocytes, based on FSC vs SSC plot, followed by selection of CD14low cells (Fig 1A). After selecting neutrophils, we gated on CFSE+ and CFSE-neutrophils, to evaluate the influence of parasite infection on the different parameters.

Statistical analysis
All samples were submitted to ROUT test to identify outliers. We compared our results using One-Way Anova or Kruskal-Wallis, as indicated by the results obtained from running the Kolmogorov-Smirnov normality test. Correlation analyses were done using Pearson's correlation coefficient. All analyses were performed using Graph Pad Prism Software (LaJolla, CA, USA). Differences that returned p values equal or less than 0.05 were considered statistically significant from one another. Evaluation of the frequency and intensity of CFSE+ neutrophils exposed to Y and Col1.7G2 trypomastigotes. Trypomastigotes from Col1.7G2 and Y strain were previously stained with CFSE and exposed in vitro to human neutrophils for 15 hours (n = 7) (A) The first panel shows a representative figure of CD14 expression versus granularity indicating that neutrophils express low CD14, and defining neutrophil, monocyte and lymphocyte gate. The second to fourth panels show representative plots of gating strategy for defining the neutrophil population. The second panel shows the selection of granulocytes in the size versus granularity (SSCxFSC), the third panel shows the selection of CD14low population, corresponding to neutrophils, and the fourth panel shows the histogram of CFSE and the selection of CFSE-and CFSE+ cells.

Results
Col1.7G2 and Y strain trypomastigotes interact with human neutrophils, and do so with similar intensity Trypomastigotes obtained from cell cultures were previously labeled with CFSE in order to determine the rate of interaction of the different T. cruzi isolates with human neutrophils. We observed a similar frequency of CFSE+ neutrophils after exposure to Col1.7G2 or Y strain in 15 hour cultures (Fig 1B). The intensity of infection on a cell per cell basis, as measured by the mean fluorescence intensity (MFI) of CFSE was also similar comparing neutrophils exposed to the different strains ( Fig 1C). As previously shown by us (22), the frequencies and intensity of infection by human monocytes with either strain is similar (Fig 1B and 1C). The frequency of CFSE+ lymphocytes is much lower than neutrophils or monocytes ( Fig 1B).

Interaction with Col1.7G2 and Y strain increases CD282 and CD284 expression by human neutrophils
Previous studies have shown that neutrophil activation can be accompanied by increased expression of Toll-like receptors [24,26]. Thus, in order to investigate the activation of neutrophils after in vitro exposure to Col1.7G2 and Y strain, we assessed the expression of CD282 and CD284 by flow cytometry. Fig 2A shows the gating strategy for the analysis of CD282 and CD284 expression. Interaction with both strains led to a higher CD282 and CD284 expression (Fig 2B and 2D). A higher expression of CD284 was observed by interaction with Col1.7G2 compared to Y strain ( Fig 2D). Increases in CD282 and CD284 were dependent on the parasite interaction, since CFSE-cells did not up regulate these molecules (Fig 2C and 2E).
Interaction with Col1.7G2 and Y strain trypomastigotes increased IL-12 and TNF, but not IL-10, expression by human neutrophils We then questioned whether interaction with the different parasite strains and subsequent activation influenced the expression of cytokines by human neutrophils. It was observed that interaction with Col1.7G2, but not Y strain, led to a significant increase in expression of IL-12/ IL-23p40 in CFSE+ neutrophils ( Fig 3A). Infection with either Col1.7G2 or Y strain did not change the expression of IL-10 ( Fig 3C). A higher expression of TNF in CFSE+ neutrophils in relation to non-infected culture was observed after contact with both isolates (Fig 3E). No change in the expression of cytokines was found in CFSE-neutrophils after exposure to either isolate ( Fig 3B, 3D and 3F). Moreover, cytokine expression was lower in CFSE-, as compared to CFSE+ neutrophils (Fig 3, comparing left and right panels), and induction of IL-12 was totally dependent on parasite presence, since CFSE-neutrophils did no express any IL-12.

Col1.7G2 and Y strain trypomastigotes decrease viability of human neutrophils and induce apoptosis in these cells
The life span of circulating neutrophils is typically short (8-20h), but certain factors such as recruitment to an inflammation site and phagocytosis of infectious agents may change their lifespan [27,28]. In our study, we observed that the percentage of granulocytes decreased after 15 hours of exposure to both Col1.7G2 and Y strain when compared to media control ( Fig  4A). We then evaluated the expression of Annexin V and 7AAD to determine cell viability and access the occurrence of apoptosis of neutrophils. Gating strategy for the analysis of 7AAD and Annexin V expression is shown in Fig 4B. We observed that the decrease in the granulocyte population was associated with a decreased viability of infected cells, since the frequency of viable cells (annexinV-7AAD-) was lower within the CFSE+ population ( Fig 4C). This Results are expressed as average ± standard deviation (n = 7). The symbols *, ** and **** indicate p < 0.05, p < 0.01 and p<0.0001 between groups, respectively. Comparisons between groups were performed using One-Way Anova (Panel D) or Kruskal-Wallis test (panels B, C and E) according to Kolmogorov-Smirnov normality test.
https://doi.org/10.1371/journal.pone.0188083.g002 (F) TNF in CFSE-neutrophils. Results are expressed as average ± standard deviation (n = 7). The symbols * and ** indicate p < 0.05 and p < 0.01 between groups, respectively. The plots show the selection of IL-12/IL-23p40 or IL-10 or TNF positive cells versus forward scatter from CFSE+ and CFSE-gated cells in media, as well as cultures infected with different isolates. Comparisons between groups were performed using One-Way Anova or Kruskal-Wallis test according to Kolmogorov-Smirnov normality test.  versus AnnexinV in media, positive control, as well as cultures exposed to different isolates in CFSE+ and CFSE-neutrophils. (C) Frequency of neutrophils in media, and CFSE+ and CFSE-that do not express Annexin V or 7AAD (viable cells); insert shows viability of monocytes exposed to the different strains. (D) Frequency of neutrophils in media, and CFSE+ and CFSEthat express Annexin V but not 7AAD (apoptotic cells); insert shows apoptosis of monocytes exposed to the different strains. Results are expressed as average ± standard deviation. The symbols *, **, *** and **** indicates p < 0.05, p < 0.01, p < 0.001 and p<0.0001 between groups, respectively. Comparisons between groups were performed using One-Way Anova test according to Kolmogorov-Smirnov normality test. https://doi.org/10.1371/journal.pone.0188083.g004 T. cruzi induces human neutrophil apoptosis decrease in viability was not observed in monocytes exposed to either strain (Fig 4C, upper  right panel).
To determine whether the observed viability reduction was associated with the occurrence of apoptosis, we determined the frequency of Annexin V+ 7AAD-cells and observed that the frequency of CFSE+ neutrophils, indicative of interaction with the different isolates, that underwent apoptosis (%AnnexinV+7AAD-) was significantly higher than CFSE-neutrophils ( Fig 4D). Again, despite the fact that monocytes also get infected with trypomastigotes from either strain (Fig 1), they do not undergo apoptosis (Fig 4D, upper right panel). CFSE+ neutrophils exposed to Col1.7G2 and Y strain express higher percentage of TNF-receptor and FasL, but no significant changes in Fas expression In order to investigate the pathway associated with apoptosis after exposure to Col1.7G2 and Y strain, we evaluated the expression of Fas and TNF-receptor II. Both molecules are related to the activation of caspase 8 and induction of apoptosis through the extrinsic pathway [29]. TNF R II has greater affinity for TNF and also a greater half-life of binding TNF than TNF R I [30]. Fig 5A shows the gating strategy used in the analysis of Fas, FasL and TNF receptor in CFSE+ neutrophils. Although interaction with Col1.7G2 and Y strain does not lead to a higher percentage or intensity of expression of Fas in CFSE+ neutrophils (Fig 5B and 5C), expression of FasL was upregulated in CFSE+ neutrophils (Fig 5D and 5E). Strikingly, interaction with both strains induced a higher percentage of expression of TNF-receptor in these cells (Fig 5F) and only Y strain induced higher intensity of expression of TNF-receptor (Fig 5G).
Higher intensity of CFSE expression is correlated with higher TNF, IL-10 and apoptosis in neutrophils exposed to Y strain trypomastigotes In order to determine if the intensity of neutrophil-trypomastigote interaction was related to cytokine expression and apoptosis, we performed correlation analysis between the frequency of neutrophils expressing IL-12, TNF or IL-10, as well as the frequency of annexinV+7AADcells, and the intensity of CFSE expression. We observed a statistically significant positive correlation between the frequencies of TNF, IL-10 and annexinV+7AAD-cells with mean intensity of CFSE expression after exposure to Y strain trypomastigotes, suggesting that the higher the parasite interaction, the higher the expression of these cytokines and the occurrence of apoptosis (Fig 6). No correlation was observed between IL-12 expression and CFSE intensity, and no statistically significant correlations were observed between the different parameters in cultures exposed to Col1.7G2 (Fig 6).

Discussion
In this work, we demonstrated that T. cruzi trypomastigotes belonging to DTUs I and II (Col1.7G2 and Y strain, respectively) are capable of interacting with and activating human neutrophils, and induce their apoptosis. These DTUs are the main parasite groups associated with Chagas disease in Latin America and this mechanism may be important for parasite survival in the human host, by decreasing the immune response.
Our data showed that the frequency of CFSE+ neutrophils, indicative of interaction between neutrophil and trypomastigotes, as well as the intensity of CFSE expression on a cellper-cell basis, were similar when comparing Y strain and Col1.7G2 (Fig 1B and 1C). Although monocytes are preferentially infected with T. cruzi [22], neutrophils have also been shown to internalize trypomastigotes [19,23]. Comparing the infection of neutrophils versus monocytes, T. cruzi induces human neutrophil apoptosis we observed that the percentage of neutrophils infected by either isolate of T. cruzi was similar to the frequency of monocytes infected by the same strains (Fig 1B). However, the intensity of infection in neutrophils was significantly lower when compared to the values observed in human monocytes (Fig 1C). Sanderson and colleagues suggested that amastigotes of T. cruzi are capable of replicating in macrophages but not in neutrophils, as confirmed by the higher intensity of infection found in macrophages in that study [31]. This could explain the lower intensity of infection in neutrophils observed in this work, despite the short-term culture. Another possibility is that, although neutrophils and monocytes present a similar phagocytic capacity [32], it is known that the expression of CD282, is higher in the latter [24]. This difference in CD282 expression could be responsible for the lower intensity of neutrophil infection, since its activation is related to the internalization of T. cruzi [33]. Further studies are required to directly address this issue.
Our next goal was to evaluate the activation status of neutrophils after exposure to the different isolates. For this, we assessed the expression of CD282 and CD284 after exposure to trypomastigotes of Y strain and Col1.7G2. It has been shown that T. cruzi expresses several CD282 and CD284 agonists [34]. Both receptors are important in the activation of the innate immune system, and CD282 acts as an immunoregulator in the early stage of infection. CD282 (-/-) mice showed increased production of IL-12/IL-23p40 and IFN-gamma after T. cruzi infection [35]. CD282 activation has also been implicated in phagocytosis of trypomastigotes [36]. Moreover, CD284 expression is critical in the control of parasites during the acute phase. It has been shown that CD284 signaling triggers an important early parasiticidal event against T. cruzi, which is dependent on the formation of NO and ROS [37]. We observed an increased expression of CD282 and CD284 following exposure to either isolates, compared to media control. Bystander neutrophils (CFSE-) had a lower expression of CD282 and CD284 compared to CFSE+, demonstrating that expression of CD282 and CD284 was dependent on the presence of the parasite. Furthermore, exposure to Co1.7G2 led to a higher CD284 expression compared to infection with Y strain.
Since the increase in CD282 and CD284 expression in CFSE+ neutrophils indicates an activation of these cells, we sought to evaluate the expression of the immunoregulatory cytokines IL-12/IL-23p40, IL-10 and TNF. In agreement with the results of CD282 and CD284 expression, only CFSE+ neutrophils had an increased expression of cytokines. Interestingly, our data showed a statistically significant increase in IL-12/IL-23p40 expression only after infection with Col1.7G2, which could be related to higher CD284 expression, as previously suggested [34]. While TNF was increased in neutrophils exposed to both isolates, no alterations in IL-10 expression were observed. It was previously shown that infection with Col1.7G2 induces high expression of IL-10 by human monocytes, as compared to Y strain, and that this increase was not dependent on parasite contact, since it was observed in CFSE+ and CFSE-cells [19]. This shows that different cell types react differently to infection, possibly due to activation of distinct surface receptors and/or intracellular pathways.
Previous studies have shown that TNF is an apoptosis inducer in many different cell types [38,39]. It is noteworthy that the population of neutrophils in which we observed the higher rate of apoptosis (CFSE+) is also the one expressing higher TNF. TNF expression also triggers the formation of neutrophils extracellular traps (NET) [40], an important control mechanism of the innate immune response, elicited by different protozoa [16,41].
Our results showed a decrease in the percentage of granulocytes after interaction with the two T. cruzi isolates. We also demonstrated that CFSE+ neutrophils exposed to both Col1.7G2 and Y strain display a decrease in viability compared with CFSE-neutrophils (Fig 4). Similar results were found in murine neutrophils infected with Leishmania amazonensis [42]. On the other hand, infection of human peripheral blood neutrophils in vitro with Leishmania major has been shown to inhibit apoptosis [43]. Thus, induction of neutrophil apoptosis may play a role in some parasitic infections but not others. Decreased viability was not observed in monocytes, despite the fact that these cells are highly infected by these strains ( [22] and Fig 1), suggesting that this is a specific activity of T. cruzi over neutrophils.
We found that neutrophils exposed to both isolates display higher percentage of apoptosis compared with non-infected ones. Interestingly, Freire-de-Lima and colleagues demonstrated that macrophages in contact with apoptotic cells favor the intracellular parasite growth and increase parasitemia [44,45]. More recently, it was demonstrated that apoptotic neutrophils increase T. cruzi replication in murine macrophages [11]. Therefore, the induction of neutrophil apoptosis by T. cruzi may enhance phagocytosis by monocytes and macrophages, favoring infection. It is also known that internalization and/or binding of apoptotic cells down-regulates the response of macrophages [9,44,46,47] and monocytes [48,49]. The induction of phagocytosis of apoptotic neutrophils also acts as an escape mechanism of the immune system since macrophages that have ingested apoptotic cells in vitro inhibit production of pro-inflammatory mediators [9,47,49].
In order to investigate whether Col1.7G2 and Y strain induce apoptosis in a similar manner, we evaluated the expression of Fas and TNF-receptor molecules. The Fas-FasL and TNFR-TNF binding induce apoptosis through activation of caspase 8 and 10, corresponding to the extrinsic pathway of apoptosis [29]. Lula and colleagues described a strong correlation between soluble TNF superfamily ligands and heart failure parameters in chronic Chagas patients presenting functional ventricular disturbances [50]. Strikingly, interaction with both isolates induced higher expression of TNF-receptor by neutrophils. Increased expression of TNF and its receptor is an indicative that both Col1.7G2 and Y strain induce apoptosis through this pathway, even though further studies are necessary to investigate whether T. cruzi infection induces apoptosis through other activation pathways as well. Another possibility would be the Fas-FasL pathway. While no differences were observed in Fas expression by neutrophils after infection, Fas-L was upregulated. However, given the very low expression of Fas by these cells (~0.2%), it is unlikely that this pathway plays an important function, favoring the TNF-mediated mechanism. The upregulation of FasL by both isolates indicates that these infected cells may induce apoptosis of other cell populations, also providing a mechanism of parasite survival. Cabral-Piccin and colleagues demonstrated that apoptotic CD8 lymphocytes favor M2 phenotype and T. cruzi replication in macrophages. They also showed that blocking apoptosis through anti-FasL treatment can restrict parasite growth and increase NO production [45]. Our analysis also showed that the higher the intensity of CFSE expression, indicating interaction with a higher number of parasites, the higher the expression of cytokines and the occurrence of apoptosis. Interestingly, this correlation was only statistically significant when performed with Y strain trypomastigotes and not with Col1.7G2 isolate. Previous studies have shown that acute Chagas patients and individuals with disease reactivation due to HIV infection or transplantation display leukopenia [51][52][53]. This suggests that neutropenia could indeed be occurring in vivo, presenting an important correlation with our data.
Taken together our results show that Tc I and Tc II T. cruzi isolates are capable of interacting and activating human neutrophils. Interaction with both isolates induces apoptosis, likely through the TNF-TNFreceptor extrinsic pathway. The induction of apoptosis may act as an effective escape mechanism of the host immune response. The fact that both isolates lead to a similar reduction of cell viability, rate of apoptosis, and expression of TNF and TNF-receptor in human neutrophils indicate that this may be a highly conserved mechanism for the species, since the isolates are representatives of the two most abundant DTUs, TcI (clone Col cl1.7) and TcII (Y strain). Although different DTUs exhibit different biological properties, it is known that there is some overlap between strains from different DTU. Moreover, heterogeneity is also observed amongst strains belonging to the same DTU. This finding inspires new studies of infection control by targeting parasite-neutrophil interaction.
Although this study provides important new information regarding neutrophil-T. cruzi interaction and its possible implications in the immune response and parasite survival, some limitations need to be taken into account. The exact apoptosis signaling and pathway still remain to be determined. Moreover, in order to truly validate the induction of apoptosis as a mechanism common to the species, it is critical to evaluate other isolates of the parasite, belonging to the same DTUs, as well as to other DTUs. This limitations open perspectives for further studies.