Neutrophil extracellular traps modulate inflammatory markers and uptake of oxidized LDL by human and murine macrophages

Neutrophil extracellular traps (NETs) are web-like structures, which are released upon neutrophil activation. It has previously been demonstrated that NETs are present in atherosclerotic lesions of both humans and animal models thus playing a decisive role in atherosclerosis. Besides, macrophages have a crucial role in disease progression, whereby classically activated M1 macrophages sustain inflammation and alternatively activated M2 macrophages display anti-inflammatory effects. Although NETs and macrophages were found to colocalize in atherosclerotic lesions, the impact of NETs on macrophage function is not fully understood. In the present study, we aimed to investigate the effect of NETs on human and murine macrophages in respect to the expression of pro-inflammatory cytokines, matrix metalloproteinases (MMPs) and uptake of oxidized LDL (oxLDL) in vitro. Human THP-1 and murine bone marrow-derived macrophages were cultured under M1 (LPS + IFN-γ)- and M2a (IL-4)-polarizing culture conditions and treated with NETs. To mimic intraplaque regions, cells were additionally cultured under hypoxic conditions. NETs significantly increased the expression of IL-1β, TNF-α and IL-6 in THP-M1 macrophages under normoxia but suppressed their expression in murine M1 macrophages under hypoxic conditions. Notably, NETs increased the number of oxLDL-positive M1 and M2 human and murine macrophages under normoxia, but did not influence formation of murine foam cells under hypoxia. However, oxLDL uptake did not strongly correlate with the expression of the LDL receptor CD36. Besides, upregulated MMP-9 expression and secretion by macrophages was detected in the presence of NETs. Again, hypoxic culture conditions dampened NETs effects. These results suggest that NETs may favor foam cell formation and plaque vulnerability, but exert opposite effects in respect to the inflammatory response of human and murine M1 macrophages. Moreover, effects of NETs on macrophages’ phenotype are altered under hypoxia.


Introduction
Atherosclerosis is recognized as the primary pathophysiology of cardiovascular disease and the leading cause of morbidity and mortality worldwide. Lipid-driven vascular chronic inflammation streptomycin (Sigma Aldrich) at a density of 2.5 × 10 5 cells / ml. For macrophage differentiation, THP-1 cells were treated with 5 ng/ml phorbol 12-myristate 13-acetate (PMA; Sigma Aldrich) for 48 h followed by 96 h of resting in PMA-free RPMI medium. Macrophages were further polarized into M1-like macrophages with 100 ng/ml LPS (Sigma Aldrich) and 20 ng/ ml recombinant human IFN-γ (Peprotech). Polarization into the M2a-like phenotype was obtained by incubation with 20 ng/ml recombinant human IL-4 (Peprotech) for 24 h. Macrophage differentiation and polarization were evaluated by flow cytometry and Real-time PCR (S1 Fig).

Isolation, polarization and characterization of murine macrophages
Eight to twelve weeks old C57BL/6J mice were euthanized by cervical dislocation to minimize suffering. Isolation of bone marrow-derived cells and in vitro differentiation into macrophages were performed as previously reported [22]. Cell isolation from mice was approved by the local ethics committee (Bezirksregierung Köln; Germany; No: 4.17.026; 4.18.015). For in vitro polarization, M0 macrophages were cultured in RPMI 1640 medium with 10% FCS supplemented with 20 ng/ml recombinant murine IFN-γ (Peprotech) and 100 ng/ml LPS (Sigma Aldrich) to induce a M1-like phenotype or in medium supplemented with 20 ng/ml recombinant murine IL-4 (Peprotech) to induce M2a-like macrophages, respectively. Macrophage phenotype was routinely confirmed by flow cytometry and Real-time PCR as previously described by our group [22].

Isolation of human and murine neutrophils
Human neutrophils were isolated by discontinuous density gradient centrifugation on Percoll (GE Healthcare) as previously reported [23]. Isolation of human blood cells was approved by the Ethics Committee of the Medical Faculty of the University of Cologne (#15-393). Study participants provided written informed consent. Murine neutrophils were isolated from the bone marrow of 8-12 weeks old C57BL/6J mice, by flushing cells out of the bone with HBSS (Gibco) + 3% FCS. Erythrocytes were lysed using 0.2% NaCl solution, followed by the addition of 1.2% NaCl to restore isotonicity. Neutrophils were isolated by density gradient centrifugation on Percoll (GE Healthcare) at 1000 × g for 30 min at 15˚C without break. The pellet enriched in neutrophils was washed twice with HBSS + 3% FCS.

Generation of NETs-enriched supernatants
Human or murine neutrophils (2 × 10 6 / well) were cultured in RPMI 1640 medium supplemented with 2% FCS. For induction of NETosis, cells were incubated with 100 nM PMA (Sigma Aldrich) for 3.5 h. Afterwards, cells were carefully washed with PBS and NETs were collected in PBS by vigorous pipetting. To prepare large heterogeneous DNA/NETs fragments [24], Alu I (New England Biolabs) was added to the culture medium for 20 min before NETs harvesting. NETs were stored at -80˚C until use. The concentration of NETs was quantified using a NanoDrop stectrophotometer (Thermo Fisher) at a wavelength of 260 nm. NETs-associated proteins were quantified by BCA assay (Thermo Fisher).

Immunofluorescence staining of NETs
Freshly isolated human or murine bone marrow-derived neutrophils (1.5 × 10 5 ) were seeded on polylysine-coated glass coverslips (Ø 18 mm), allowed to settle for 30 min, and then treated with 100 nM PMA. After 3.5 h of activation, cells were fixed with 4% paraformaldehyde (PFA), blocked with 5% normal goat serum (Cell Signaling Technology) and 1% bovine serum albumin (BSA) (Carl Roth) and incubated with polyclonal antibodies against myeloperoxidase (MPO) (Abcam) for 2 h at room temperature. Cells were further incubated with a secondary anti-rabbit Alexa Fluor 488-conjugated antibody (Cell Signaling Technology) for 1 h, counterstained with DAPI and mounted in Dako fluorescent mounting medium (Dako). Cells were examined with an inverted microscope (Eclipse Ti-U 100, Nikon).

Determination of peroxidase activity
The MPO activity in NETs-enriched supernatants was measured by mixing 50 μl sample with 50 μl detection solution (1:1 TMB Substrate A/TMB substrate B, BioLegend) in a 96-well microplate. After 3 min incubation at room temperature the staining reaction was stopped by adding 50 μl H 2 SO 4 . The absorbance of the reaction product was measured at 450 nm using a microplate reader (Victor X3, Perkin Elmer).

Cell culture conditions
Human THP-1 macrophages were cultured in 6-well plates in RPMI 1640 medium supplemented with 10% FCS and 100 U/ml penicillin and 10 μg/ml streptomycin under normoxic and hypoxic conditions in the presence or absence of human NETs (1000 ng/ml). Murine macrophages were cultured under the same conditions and in the presence of murine NETs (1000 ng/ml). For hypoxia exposure, culture dishes were placed in a hypoxia chamber (STEM-CELL Technologies), pre-flushed with a gas mixture (2% O 2 , 5% CO 2 , and 93%N 2 ). Macrophages were placed at 37˚C in a 21% O 2 , 5% CO 2 , and 74% N 2 humidified incubator and cultured under standard culture conditions in parallel.

Cellular Dil-oxLDL uptake
THP-1 macrophages and murine macrophages were seeded on coverslips (Ø 18 mm) at a density of 1 × 10 5 cells / well and polarized into M1-and M2a-like cells. Then, cells were washed with HBSS and incubated in HBSS + 0.3% BSA in the presence or absence of NETs (1000 ng/ ml) for 24 h. Subsequently the cells were incubated with Dil-oxLDL (20 μg/ml) in HBSS + 0.3% BSA for 6 h. After washing with PBS, cells were fixed with 4% paraformaldehyde for 10 min, counterstained with DAPI and mounted in Dako fluorescent mounting medium (Dako).

Statistical analysis
Data are presented as mean ± standard deviation (SD). Statistical analysis was performed using GraphPad Prism software (GraphPad Software, San Diego, CA). Data sets were assessed for normality using the Shapiro-Wilk test. Nonparametric unpaired data of multiple groups were analyse using Kruskal-Wallis test with Dunn's post-hoc test. Unpaired data from two groups were analysed using t test or Mann-Whitney nonparametric test. One-sample t test (for data that were normally distributed) or Wilcoxon signed-rank test (for data that were not normally distributed) were used when samples were compared with reference control sample (set as 1). When normality could not be assumed, a nonparametric test was performed. A P-value less that 0.05 was considered as statistically significant. The statistical tests used are indicated in the respective figure legends.

NETs modulate cytokine expression in human and murine M1-like macrophages
Human and murine NETs were generated as described in Material and Methods. The formation of NETs was confirmed by DNA, protein and peroxidase activity quantification in NETsenriched supernatants as well as by immunofluorescence (Fig 1).
THP-1 macrophages and murine macrophages (M0) were polarized by LPS + IFN-γ into the M1 state. Polarization of human THP-1 macrophages was further confirmed by flow cytometry and Real-time PCR, respectively (S1 Fig). THP-M1 macrophages expressed high levels of the M1 marker CD197 and lower levels of the M2 markers CD163 and CD206. Besides, M1-like cells showed increased expression of the pro-inflammatory genes IL-1β, TNFα, IL-6 and IL-12. Polarized murine macrophages were previously characterized in detail by our group [22].
To study the capacity of NETs to modulate cytokine expression by macrophages, cells were treated with NETs purified from PMA-stimulated neutrophils in the presence of LPS + IFN-γ under standard culture conditions. As hypoxia is present in atherosclerotic plaques [25], cells were cultured in 2% oxygen in parallel. As depicted in Fig 2, NETs significantly increased the expression of the pro-inflammatory cytokines IL-1β, TNF-α and IL-6 in THP-1 macrophages under M1-polarizing conditions and at high oxygen levels. In turn, the expression of IL-12 was significantly suppressed (Fig 2A). IL-1β levels in culture supernatants also increased significantly ( Fig 2B). Notably, the effects of NETs on cytokine gene expression were less pronounced under hypoxic culture conditions and only IL-1β was significantly upregulated. Different results were obtained after stimulation of murine macrophages with NETs. As shown in Fig  2C, NETs strongly suppressed the expression of the pro-inflammatory cytokines IL-1β, TNF-α and IL-6 under normoxia and hypoxia, but did not alter IL-12 expression. In addition, IL-1β expression was found to be reduced at 2% oxygen, although no significant downregulation of IL-1β secretion could be observed (Fig 2D). These results suggest that NETs might dampen the inflammatory response in murine macrophages, but conversely increase the pro-inflammatory action of THP-1-derived M1-like cells.

NETs amplify oxLDL uptake by macrophages
To further explore the effect of NETs on foam cell formation during atherogenesis, polarized THP-1 and primary murine macrophages were incubated in serum-free culture medium in   (Fig 3A and 3B). For cells cultured under normoxic conditions, oxLDL uptake was additionally confirmed by immunofluorescence (Fig 3C).
In THP-1 macrophages, human NETs significantly increased the number of oxLDL-positive M1-and M2a-like macrophages under both culture conditions. Of note, culture under hypoxia strongly reduced the number of oxLDL-positive M2a-macrophages ( Fig 3A).
Similar to THP-1 macrophages, the number of murine M1-and M2a-like cells, taking up oxLDL, was found to increase after stimulation with NETs, although no NETs effects were observed when cells where culture under hypoxia. In M1-like macrophages, hypoxic culture conditions strongly enhanced foam cell formation (Fig 3B). In general, M2a-polarized macrophages cultured under standard conditions were identified to incorporate higher amounts of Dil-oxLDL when compared to M1 macrophages.
It was previously reported that inhibition of the oxLDL receptor CD36 strongly decreases oxLDL uptake by macrophages [26]. We therefore questioned if increased oxLDL uptake after NETs stimulation is related to upregulated CD36 expression. Cellular CD36 expression was higher in M2a-like macrophages corresponding to increased foam cell formation (Fig 4). Although human NETs did not alter the number of CD36 + cells, significant upregulation of CD36 on the surface of M1-like THP-1 macrophages was detected, but not on M2a-like cells (Fig 4A). However, murine NETs significantly decreased the expression of CD36 on M1-like murine macrophages despite increased oxLDL incorporation. Under hypoxia, murine NETs did not significantly changed the expression of CD36 on M1-and M2a-like macrophages ( Fig 4B). These data indicate that increased oxLDL uptake after NETs stimulation does not seem to strictly depend on upregulated CD36 expression on macrophages.

NETs upregulate the expression and secretion of MMP-9 in macrophages
The expression of MMPs substantially influences the vulnerability of atherosclerotic plaques. As NETs and macrophages were reported to colocalize in atherosclerotic plaques [27], we next investigated the expression of MMPs in macrophages after exposure to NETs. In THP-M1 cells, we found gene expression of MMP-8 and MMP-9 to be strongly increased by NETs, whereas profound inhibition of MMP-2 and MMP-8 expression was detected under M2apolarizing conditions (Fig 5A). However, MMP-9 expression was increased in NETs-treated THP-M2a macrophages. In turn, hypoxia attenuated NETs effects suggesting that their biological effects strongly depend on oxygen levels. Concomitant with highly upregulated gene expression, THP-M1 cultured under normoxia secreted significantly elevated levels of MMP-9 in the culture supernatant after stimulation with NETs ( Fig 5B).
Murine M1-like macrophages exposed to murine NETs under normoxia showed upregulated MMP-9 expression, whereas M2a-like macrophages displayed reduced MMP-2 and elevated MMP-9 expression (Fig 6A). However, NETs stimulation resulted in increased MMP-9 secretion from macrophages, regardless of the polarization state of the cells and oxygen levels ( Fig 6B).

Discussion
Macrophages represent key players in the genesis and progression of atherosclerosis in mice and humans. It has previously been reported that macrophages exhibit an M2 phenotype at early disease stages and become M1 during plaque progression [28]. The M2 phenotype was associated with the resolution of plaque inflammation and regression of atherosclerosis [8,9]. During the past decade, many studies reported the presence of NETs in human and murine atherosclerotic lesions [19,20,27], but their biological functions remain incompletely understood.   In the present in vitro study, we have compared the impact of NETs on human and murine macrophages in respect to the expression of cytokines and MMPs as well as foam cell formation. One important finding of our study was the opposite effect of NETs on the expression of pro-inflammatory cytokines by human and murine macrophages. NETs enhanced the expression of pro-inflammatory genes in THP-1 macrophages cultured under M1-polarizing conditions and standard oxygen levels but not in murine M1-like cells. On the contrary, the expression of IL-1β, TNF-α and IL-6 was significantly suppressed in murine M1-like cells under hypoxia largely supporting the results of our previous report [21]. Gene expression levels of IL-1β visibly correlated with protein secretion. These findings contradict the results of an earlier study showing an increase of pro-inflammatory IL-1α, IL-1β and IL-6 in atherosclerotic lesions of mice containing NETs [19]. Similarly, Josefs et al. recently reported that NETs-positive areas in atherosclerotic plaques of mice express higher levels of the M1 marker gene iNOS suggesting that NETs promote an M1-like phenotype [27]. However, the presence of NETs in atherosclerotic lesions does not provide supporting evidence for their pro-inflammatory actions. The effect of NETs was shown to become reversed by the treatment of mice with DNase I [27] or peptidylarginine (PAD) inhibitors [29,30]. It is noteworthy to consider that DNase I does not specifically degrade NETs but rather cell-free DNA (cfDNA) released by necrotic cells, which in turn acts highly pro-inflammatory [21,31]. Additionally, a wide range of side effects of commonly used PAD inhibitors in mice, like inhibition of T cell proliferation and suppression of dendritic cells and smooth muscle cells activation needs to be considered for the interpretation of findings obtained with PAD inhibitors [32][33][34]. Of note and in accordance with prior work [35], our data presented here also demonstrate that human NETs significantly enhance pro-inflammatory cytokine expression in human THP-M1 macrophages.
Studies have indicated that foam cells play a major role in plaque formation during atherosclerosis and promote disease progression [36,37]. Here, we show for the first time that NETs strongly increase the number of macrophages that take up oxLDL independently on macrophages' phenotype. In addition, previous studies reported NETs formation after stimulation of human [38,39] neutrophils with oxLDL. As neutrophils and monocytes / macrophages come in close cell-to-cell contact within the atherosclerotic lesions [27,40], NETs derived from e.g. oxLDL-stimulated neutrophils may probably contribute to oxLDL uptake by macrophages and foam cell formation in both mice and humans. In line with the findings of a previous report [41], the number of oxLDL-positive M2a-like macrophages was found to be higher compared to M1-like cells showing positive correlation with increased CD36 expression on these cells. In contrast, NETs-mediated foam cell formation does not seem to depend on surface CD36 expression. As NETs diminished the expression of pro-inflammatory cytokines in M1-like murine macrophages, it is likely that NETs-triggered transdifferentiation from M1 cells to M2 cells [21] may, at least in mice, contribute to lipid accumulation. However, we did not investigate the expression of additional oxLDL receptors or determined the impact of NETs on cholesterol efflux, representing a limitation of this study. Besides, in human THP-1 cells, different mechanisms could be involved and it is still unclear which molecules regulate NETs-mediated foam cell formation. In this respect, a role of NETs-associated citrullinated histones has recently been suggested. The authors demonstrated that citrullinated histones accelerate LDL aggregation and foam cell formation in dependence of the citrulline content [42]. As PMA, which has been used in this study to stimulate neutrophils, was found to enhance histone citrullination [43,44] too, it is likely that increased oxLDL incorporation is mediated by NETs-associated citrullinated histones.
Atherosclerotic lesions represent a hypoxic milieu that favor the uptake of oxLDL by macrophages [25]. Therefore, in human atherosclerotic plaques, the hypoxic regions are rich in macrophages and foam cells [45]. Here, hypoxia did not change oxLDL uptake by human THP-1 macrophages and M2a-like murine cells, but highly increased oxLDL incorporation by M1-polarized murine macrophages. Additionally, the effect of NETs on foam cell formation also depended on the hypoxic environment, whereby lower oxygen levels suppressed NETsmediated oxLDL uptake by murine macrophages.
MMPs produced by macrophages contribute to plaque rupture, atherothrombosis and MI [46]. Here, NETs increased the expression and secretion of MMP-9 in both human and murine macrophages, independently on their phenotype. MMP-9 is involved in all stages of atherosclerosis and MMP-9 deletion is associated with reduced plaque size, macrophage content and collagen deposition in aortic lesions of ApoE -/mice [47]. In humans, plasma MMP-9 levels were reported to be higher in patients with ruptured plaque compared with patients without ruptured plaques. Therefore, MMP-9 was suggested to represent an independent predictor of plaque rupture [48][49][50]. On the other hand, NETs were found to diminish MMP-2 gene expression in human and murine macrophages under M2a-polarizing culture conditions. It is known that MMP-2 plays a role in promoting atherosclerosis, since atherogenesis is reduced in MMP-2-deficient ApoE −/− mice [51]. Thus, by regulating the expression of MMPs, it may be assumed that NETs exert protective or harmful effects depending on the context.
The results of this study reflect the discrepancy between human and murine macrophages, which has already been characterized in detail. Moreover, the use of THP-1 monocytes instead of peripheral blood-derived monocytes represents an important limitation of our study. Although previous research has shown that human peripheral blood-derived macrophages and THP-1 macrophages were more closely related to each other than to mouse macrophages, some discrepancies between these cells have also been observed [52]. In this respect, by comparing human THP-1 macrophages and human monocytes-derived macrophages some investigators concluded that THP-1 macrophages represent an appropriate model to study the M1 polarization but less the M2 polarization [53,54]. For this reason, additional studies should be conducted in the future to prove the effect of NETs on primary human macrophages or iPS cells-derived macrophages, which were found to share many characteristics with primary human cells and display less heterogeneity in comparison to primary macrophages [52].
Overall, our results presented herein suggest that the effects of NETs on macrophages are manifold, depending on the origin of the cells, polarization state and oxygen level. NETs increase the expression of pro-inflammmatory cytokines in human THP-M1 macrophages but strongly suppress their expression in murine bone marrow-derived M1 macrophages. Additionally, we provide evidence for NETs-mediated increased oxLDL uptake by M1 and M2 macrophages in vitro. However, hypoxia may dampen or alter the effects of NETs on macrophages to some extent. Taking into account that NETs enhanced MMP-9 expression and release from macrophages in vitro, we propose that NETs play an important role in plaque stability, vulnerability and progression. Our findings may provide new insights on NETs-mediated atherogenesis and disease progression.