*3.8. E. histolytica-Induced NETosis Occurs Independently of Mitochondrial Derived ROS*

We decided to explore whether mitochondrial ROS are produced during the NETosis triggered by *E. histolytica* trophozoites. As expected, PMA did not induce mitochondrial ROS, whereas the calcium ionophore A23187 induced a significant increase of these molecules. It is noteworthy that amoebas induced mitochondrial ROS in neutrophils in a dose-dependent manner (Figure 9A). To determine whether mitochondrial ROS are necessary for amoebic-induced NETosis, we used the specific scavenger mitoTEMPO, which did not affect PMA-induced NETosis. As shown in Figure 9B, NETosis was not affected by mitoTEMPO at any ratio tested, suggesting that mitochondrial ROS are not involved.

**Figure 9.** Mitochondrial ROS are not necessary for *E. histolytica*-induced NETosis. (**A**) MitoSOXTM red-pretreated neutrophils (1 <sup>×</sup> <sup>10</sup>5) were cultured in RPMI-1640 medium supplemented with 5% FBS and then stimulated with PMA (50 nM), A23187 (10 <sup>μ</sup>M) or *E. histolytica* trophozoites at ratios 1:100, 1:50 or 1:20. Fluorescence was read at 2 h. (**B**) Neutrophils (1 <sup>×</sup> <sup>10</sup>5) were pretreated with 400 μM mitoTEMPO (Mito) or DMSO for 30 min. Posteriorly, cells were transferred to RPMI-1640 medium added with 5% FBS and 500 nM SYTOX® Green and then stimulated with PMA (50 nM), A23187 (10 μM) or *E. histolytica* trophozoites at ratios of 1:100, 1:50, 1:20 and 1:10. Fluorescence was read at 4 h. ROS and NET amount are expressed in fluorescence relative units (FRU). Values are means ± SD of three independent experiments. # *p* < 0.001 with respect to the control, \* *p* < 0.01.

#### **4. Discussion**

Neutrophil extracellular traps (NETs) were initially described by Brinkmann et al. [1] as a novel effector mechanism used by neutrophils to entrap and kill bacteria. Since then, many works have explored the mechanism underlying the DNA extrusion to extracellular space, process known as NETosis. Fuchs et al. [27] provided one of the first approaches to understand how NET release takes place, showing that oxidative metabolism directed by NADPH oxidase is involved. Nevertheless, the finding of calcium ionophores triggering NETosis opened the possibility that other mechanisms could also lead to the release of DNA, as they do not require the activity of an NADPH oxidase, but rather require PAD4 activity [44]. The case of NETosis induced by the *E. histolytica* trophozoites is intriguing, as we showed that the process occurs through a non-classical mechanism, independent of NADPH-ROS and PAD4 activity [41,45]. In this work, we performed a set of experiments to continue with the characterization of the amoeba-induced NETosis and found that it is dependent on ROS from *E. histolytica* trophozoites and on the activity of the MPO from neutrophils present in the surface of the parasites.

Previously, we described that amoebic trophozoites triggered NETosis on human neutrophils when co-incubated at trophozoite:neutrophil ratios of 1:20 [41,45–47]. Here, we demonstrated that lower amounts of amoebas (ratios 1:100 and 1:50) were also capable of leading NET release in a dose-dependent manner, whereas ratios higher than 1:20 did not induce more NETs. This result indicated that *E. histolytica* trophozoites are one of the most potent parasites to induce NETosis, since other protozoa require greater numbers to trigger significant DNA release. Thus, *Toxoplasma gondii* was used at MOI of 5:1, *Trypanosoma cruzi* at a 1:1 ratio and *Leishmania chagasi*, *L. major* or *L. amazonensis* at 10:1 to 1:1 ratios [12,48–50]. In contrast, we observed that 1 amoeba per 100 neutrophils is sufficient to induce NETosis. The reason is unknown but the size of the parasites, and therefore the density of NETosis triggering molecules, could be involved, since *T. gondii* and *Leishmania* forms (amastigotes and promastigotes) do not exceed 15 μm in length and *T. cruzi* trypomastigotes measure 12–30 μm, which is small compared with *E. histolytica* trophozoites measuring up to 60 μm [34,51–53]. While neutrophils can phagocyte small parasites enlisted above [54–56], they cannot phagocyte amoebas. Instead, we have observed that trophozoites engulf these leukocytes [45]. Neutrophils probably sense the pathogen size through dectin-1, to decide between phagocytosis or NETosis by sequestering of NE [57]. During phagocytosis, NE is moved to the phagolysosome compartment. In contrast, during NETosis, NE is guided to the nucleus for chromatin decondensation. This correlates with our previous observation in which NE is translocated to nuclei during the neutrophil-amoeba interaction [41]. It is conceivable that small parasites mainly drive neutrophils towards phagocytosis, whereas greater pathogens, such as amoebic trophozoites, preferentially induce NETosis.

NETosis mechanisms are generally divided into two groups: dependent on NADPH oxidase activity and independent of NADPH oxidase activity. Fuchs et al. [27] was the first to report that NADPH oxidase inhibition prevented NET release by PMA, and different authors have described the same mechanism for other stimuli [58–60]. Later, it was described that calcium ionophores trigger NETosis independently of NADPH oxidase activity, but this mechanism requires PAD4 [44]. We previously reported that *E. histolytica* trophozoites induce NETosis by a non-classical mechanism, independent of NADPH oxidase and PAD4 activities, since apocynin and GSK484, as respective inhibitors, failed to reduce the NET amount. Our previous observations also showed that amoebic trophozoites at 1:20 ratio completely suppressed the oxidative burst in neutrophils [41,46], which has also been reported by others [61]. Interestingly, in this work we found that lower numbers of amoebic trophozoites (1:50 and 1:100 ratios) did not completely abolish ROS generation in these leucocytes. This suggested that neutrophil ROS inhibition by amoebas depends on the density of parasites trespassing a threshold that may cause citrullination of proteins. Accordingly, Zhou et al. [62] showed that dysregulated calcium influx in neutrophils activates PAD4 that citrullinates the cytoplasmatic units p47phox and p67phox, blocking the assemble of the NADPH oxidase complex and, in turn, preventing ROS generation. In this context, we previously reported that *E. histolytica* trophozoites trigger calcium influx on human neutrophils and when incubated at a 1:20 ratio [46] and citrullinated proteins were detected [41], which would cause the inactivation of NADPH oxidase. When smaller numbers of trophozoites are confronted, this process may not happen; however, additional studies are required on this. It is worth mentioning that PAD4-independent citrullination of proteins has also been observed during the NETosis induced by *Candida albicans* [63].

Previously, we showed that heat-killed and fixed trophozoites failed to induce NETosis, which was confirmed in this work (Figure 3A) [41]. This data suggests that some products of the *E. histolytica* trophozoites metabolism are responsible for inducing NETosis. In this context, ROS derived from pathogens have been identified as molecules that lead NET release independently of ROS produced by neutrophils [33,43]. Noteworthy, we previously showed that *Entamoeba dispar*, a non-pathogen human amoeba, does not produce ROS and does not trigger NETosis [43,47]. Since most studies suggest that NETosis requires some source of ROS that does not come from neutrophils in this case, we considered the possibility that ROS produced by *E. histolytica* trophozoites induce NETosis. Here we

report that viable trophozoites produced basal ROS levels, which agrees with a previous report [64], whereas amoebas killed by heat or fixation with paraformaldehyde produced scarce ROS (Figure 3B). Even though cell death processes have been associated to an increase in ROS production [65], the low level detected by us could be explained by deathinducing agents that were used, since both reduce enzymatic activity and formaldehyde, causing protein cross-linking and the development of heat denaturalizing proteins [66].

When we treated H2DCFDA-stained amoebas with hydrogen peroxide, they exhibited a stronger fluorescence (Supplementary Figure S3), suggesting that H2DCFDA can be used as an indicator for hydrogen peroxide. Using this approach, we found that pretreatment of amoebas with pyrocatechol, a ROS scavenger of hydrogen peroxide [67,68], reduced the hydrogen peroxide detected in viable amoebas, which is produced as a response to detoxify oxygen through diverse enzymes including NADPH:flavin oxidoreductase (Eh43), thioredoxin reductase (TrxR), NADPH-dependent oxidoreductases (NO1/2) or Fesuperoxide dismutase (FeSOD) [64,69]. The hydrogen peroxide reduction in trophozoites by pretreatment with pyrocatechol impacted the NETosis directly, since a smaller amount of DNA was detected in the extracellular medium. Moreover, NETosis was completely abolished when amoebas were pretreated with pyrocatechol at 200 μM in 1:100 and 1:50 ratios. In accordance with this, pyrocatechol blocked NET release induced by PMA, which is triggered by ROS produced in neutrophils. All these data indicate that amoebic trophozoites, instead of neutrophils, are the source of ROS responsible for leading NETosis involving this parasite. This is the first report regarding the importance of ROS from *E. histolytica* trophozoites for NETosis.

Kenny et al. [33] proposed that hydrogen peroxide produced by *C. albicans* was able to enter to neutrophils to trigger NETosis. In our case, however, the addition of catalase (which possesses high specificity for hydrogen peroxide) [70] to the media in the cocultures amoeba-neutrophil failed to reduce NETosis. In contrast, catalase reduced significantly NET release induced by PMA, which has been linked to extracellular production of hydrogen peroxide by NADPH oxidase in the plasma membrane [27,71]. As amoebic hydrogen peroxide is important for amoeba-induced NETosis but catalase did not affect the process, we proposed that another ROS, probable derived from hydrogen peroxide released by the trophozoites but produced in the extracellular media, might be involved. This ROS may be produced by the activity of a neutrophil product released very early after the contact with amoebas. In this context, neutrophil MPO has been related to some mechanisms of NETosis [72], mainly with non-phagocytosed stimuli [73,74]. This enzyme produces hypochlorous acid (HClO) from hydrogen peroxide and chloride during oxidative burst in neutrophils [75]. Although luminol can react with other oxidants, Gross et al. [76] reported that its luminescence depends substantially on MPO activity. Here we observed that luminol-pretreated trophozoites when incubated with neutrophils exhibit MPO activity denoted by an increase in luminol signal. It is noteworthy that the MPO maximum activity was detected when the DNA extrusion started (approximately 20 min), suggesting that this could be the triggering stimulus. In addition, amoebas did not produce a luminol signal in the absence of neutrophil, indicating that the signal detected in amoebas corresponds to the activity of neutrophil´s MPO and not to other molecules produced by trophozoites or neutrophils. The role of MPO activity on amoeba-induced NETosis was confirmed by taking advantage of the luminol ability to scavenge HClO [77]. Luminol, at all concentrations tested here, reduced NET release induced by the amoebic trophozoites at 1:50, 1:20 and 1:10 ratios. As pretreatment of amoebas with luminol showed similar results but increased ROS levels in the parasite, the data together indicated that NET reduction was due to the scavenging activity of luminol on MPO derived ROS (HClO), instead of a decrease in ROS from amoebas. The reason why luminol increased the ROS of amoeba is unknown. We suspect that luminol can cause intensive stress in the trophozoites without affecting their viability, but this is a subject for further study in our laboratory. It is also interesting that luminol was unable to reduce NETosis at a 1:100 ratio and only reduced but did not abolish NETosis at higher ratios, suggesting that other mechanisms take place to

compensate for NET release depending on the culture conditions. In this context, it has been reported that some pathogens such as *L. amazonensis* and *C. glabrata* can trigger NETosis by different mechanisms [50,59].

The role of mitochondrial ROS (mitROS) in amoeba-induced NETosis was also explored. Douda et al. [28] stated that ROS derived from mitochondria are required for the NETosis induced by the calcium ionophores ionomycin and A23187, results that have been replicated [78]. Since then, other NET inducers such as *Leishmania* parasites or UV light have been shown to require mitROS [50,79]. Here we observed for the first time that *E. histolytica* trophozoites induce production of mitROS on human neutrophils in a dose dependent manner. How they are generated is unknown but the recognition of amoebic LPPG by TLR2 and TLR4 [80] and the stimulation of calcium influx plus a pathological stimulus (such as PAAR detected through TLRs) could be involved, as suggested elsewhere [81,82]. Although mitROS were detected in neutrophils in contact with amoebas, the scavenger mitoTEMPO failed to reduce NET releases, suggesting that they do not participate in the process.

Immunofluorescence performed on amoeba-neutrophil cocultures exhibited that DNA from neutrophils is released as aggregated cloudy NETs in accordance with a previous classification [83]. We observed extensive areas covered by DNA that entrap trophozoites (Figure 1B). Differing from other stimuli such as LPS, monosodium urate crystals or S-nitroso-N-acetyl-D,L-penicillamine, [84–86], NETs induced by trophozoites do not exhibit an homogeneous distribution of MPO and NE; in contrast, these proteins are usually visualized as spots located in reduced areas, suggesting that at least part of the neutrophil MPO and NE proteins are released by degranulation, and bound to trophozoites before, or at the same time, as NETs [47], which can explain the scarce proteins associated with DNA. Binding of MPO to *E. histolytica* trophozoites has be described previously by Pacheco–Yépez et al. [87]. They reported that purified MPO interacts with amoebic trophozoites, leading to morphological changes and loss of viability, which were also observed by us in amoebas entrapped in NETs [47].

Finally, detection of MPO covered trophozoites during neutrophil-amoeba interaction raise a question about the origin of this enzymatic activity. Therefore, we decided to use isoluminol, a hydrophobic isomer of luminol, to explore the scavenging of extracellular HClO [88]. Isoluminol significantly reduced NET amount released in response to *E. histolytica* trophozoites at the same level as luminol, indicating that extracellular MPO activity is the responsible for NETosis. This result is in accordance with previous observations indicating that exogen HClO and hypochlorite are sufficient to induce NETosis on human neutrophils [31,89]. Although the role of MPO activity in NETosis has been controversial, with some reports indicating that is dispensable [90,91] and others indicating that is required [73,77], our results support the latter. Taken together, the mechanism that we propose for amoebic-induced NETosis is shown in Figure 10.

**Figure 10.** Mechanism proposed for NETosis induced by *E. histolytica*. (**A**) Amoebic trophozoites lead neutrophils to degranulation and NET release. (**B**) MPO derived from azurophilic cytoplasmatic granules is released and it binds to the surface of trophozoites. (**C**) During the oxidative metabolism of *E. histolytica*, H2O2 is generated to detoxify O2 by diverse enzymes such as Eh43, TrxR, NO1/2 or FeSOD. H2O2 is rapidly converted to HClO through MPO bounded to cell surface of amoebas. HClO probably enters neutrophils by an unknown mechanism and it starts NETosis, promoting NE translocation to the nucleus. Pyrocatechol blocked NETosis scavenge H2O2 inside amoebas, while luminol and isoluminol react with HClO. Image was made in BioRender.com. Eh43 (NADPH:flavin oxidoreductase), TrxR (thioredoxin reductase), NO1/2 (NADPH-dependent oxidoreductases), FeSOD (Fe-superoxide dismutase), H2O2 (hydrogen peroxide), HOCl (hypochlorous acid), Azu Cx (azurosome complex), NE (neutrophil elastase).
