Nanoparticles Equipped with α2,8-Linked Sialic Acid Chains Inhibit the Release of Neutrophil Extracellular Traps

Neutrophils can combat the invasion of pathogens by the formation of neutrophil extracellular traps (NETs). The NET mechanism is not only an effective tool for combating pathogens, but is also associated with diseases. Therefore, NETs are a potential target for combating pathologies, such as cystic fibrosis and thrombosis. We investigated the potential of nanoparticles, which were modified with α2,8-linked sialic acid chains, to modulate NET release during phorbol myristate acetate stimulation. Interestingly, when these nanoparticles were applied, the formation of reactive oxygen species was partly inhibited and the release of NET was counteracted. However, although the release of NET fibers was prevented, the nuclei still lost their characteristic segmented structure and became swollen, indicating that only the release, and not complete activation was suppressed. Intriguingly, coincubation of α2,8-sialylated particles with free sialic acid chains prevented the outlined inhibitory effects. Thus, the sialic acid chains must be attached to a linker molecule to generate an active bioconjugate that is able to inhibit the release of NET.

to the transfer of neutrophil elastase to the nucleus, where neutrophil elastase degrades histones, such as histone H4, promoting the decondensation of DNA. During the last step, granular vesicles, in addition to plasma membranes, rupture and a mixture, which consists of antimicrobial biomolecules (e.g., myeloperoxidases, neutrophil elastases, lactoferrin, defensins, and cytotoxic histones) is released to combat the invading pathogens [9,16].
In mammals, several biomolecules have the capacity to modulate immunological mechanisms. Since all our cells are surrounded by a glycocalyx, consisting of glycolipids and glycoproteins, glycan-dependent mechanisms frequently take place during processes of innate and adaptive immunity [24]. In mammals, glycans are often terminated with sialic acid residues [25]. These sialylated structures can be recognized by immune cells using, for example, sialic acid-binding immunoglobulin-like lectins (siglecs), which are important immunoregulatory elements in vertebrates [26][27][28][29]. Intriguingly, inhibitory siglecs can counteract kinase-dependent activation of immune cells by recruiting SHP1 and SHP2 [28,30]. For example, Varki and colleagues described that, in the bloodstream, siglec-9 on neutrophils inhibits neutrophil activation by binding sialylated glycoproteins on erythrocytes [31]. Thus, siglecs became a target for modulating immunological events [32]. One possibility is to decorate nanoparticles with sialylated glycans. For instance, Spence et al. showed that nanoparticles coated with dimers of α2,8-linked sialic acid residues, which are the ligand for murine siglec-E, decreased inflammation driven by lipopolysaccharide (LPS) in murine macrophages [33]. Human neutrophils are known to express siglec-3, -5, -9, and -14, and interestingly, siglec-5 is able to recognize α2,8-linked sialic acid residues, such as siglec-E in mice [34][35][36].
Thus, we hypothesized that α2,8-linked sialic acid residues can be used to modulate the activation of neutrophils and NETosis. The obtained results demonstrate that the release of NETs is inhibited by the application of particles containing α2,8-linked sialic acid chains, indicating that α2,8-sialylated nanoparticles are a tool for manipulating the formation of NET.

Human Neutrophils
All volunteers provided written informed consent and all samples were anonymized. The use of human neutrophils was approved by the local ethics office of the University of Giessen, School of Medicine (05/00).

Coupling of Sialic Acids on Latex Beads
The separated sialic acid chains, as well as Neu5Ac (sialic acid monomers; monoSia) (Carbosynth, Compton, UK), were used for coupling reaction. Aliphatic amine latex particles (1% w/v, 0.1 µm; Thermo Fisher Scientific) were washed twice prior usage. A total of 150 µL latex particles were resuspended in 123 µL PBS and homogenization was performed by using ultrasonic. A total of 200 µg of dried sugars were dissolved in 25 µL PBS and then added to the latex particles. A total of 1.5 µL 5 M NaCNBH 3 was added and the coupling reaction took place for 2 h at 65 • C and 250 rpm. After this, sialic acid-coupled particles were washed twice and resuspended in 100 µL PBS. An aliquot was taken to quantify the coupling reaction, as described in the previous chapter. The used nanoparticles showed no cytotoxic characteristics as tested previously in [40]. Regarding the stability of the sialylated particles, it should be noted that the sialylation status is reduced by 30% when stored at 8 • C for one month, resulting in a decreased activity.

Isolation of Human Neutrophils
Human neutrophils were isolated as described previously by Saffarzadeh et al. [48]. Therefore, a density gradient using a Histopaque-1077 and a Histopaque-1119 was applied at 37 • C and 700 × g for 30 min. The neutrophil containing layer was washed with PBS. After erylysis (lysis buffer: pH 7.5; 0.15 M NH 4 Cl, 0.1 mM EDTA, 1 mM KHCO 3 ) cells were washed with PBS. The erylysis was performed two times. As a final step, the received pellet was washed and resuspended in RPMI 1640 (Thermo Fisher Scientific) with 1% penicillin/streptomycin (PenStrep; Thermo Fisher Scientific) and 1% fetal bovine serum (FBS; Thermo Fisher Scientific). A total of 30,000 cells/well were dissolved in RPMI 1640 with 1% PenStrep, and 1% FBS and incubated 1 h at 37 • C and 5% CO 2 before the NETs were induced.

NETosis Stimulation and Inhibition Assay
The stimulation of neutrophils with 20 nM PMA took place in Poly-l-Lysine coated 12-well chamber slides for 2 h/4 h at 37 • C and 5% CO 2 . To inhibit NETosis cells were incubated with PMA and a final concentration of 10 µg mL −1 sialic acid was coupled on latex particles. Coincubations of α2,8-sialylated particles with 30 µg mL −1 free sialic acid chains of an inherent chain length were performed under equal conditions.

Immunofluorescence Staining
After 2 h or 4 h of PMA stimulation cells were fixed with 4% paraformaldehyde (PFA) for 30 min at 4 • C. After washing, cells were incubated with 0.5% Triton X-100 for 1 min, followed by three further washing steps. Blocking was performed for 30 min at 37 • C with 2% IgG free bovine serum albumin (BSA) (Carl Roth, Karlsruhe, Germany) and anti-Neutrophil Elastase (Abcam, Cambridge, UK) was incubated overnight at 4 • C. After further washing steps, the secondary antibody (FITC anti-rabbit, 1 h, room temperature (RT)) was added before the nuclei staining with 4 ,6-Diamidin-2-phenylindol (DAPI) was done (Carl Roth, 1 µg mL -1 ). Fixation was performed with 2% PFA for 20 min at RT. Samples were then mounted and analyzed using fluorescence microscopy (Carl Zeiss confocal laser scanning microscope LSM 800). The determination of the NET area per cell was based on the release of decondensed DNA fibers during NETosis (DAPI staining) and was performed using Cellprofiler 2.2.0. Three pictures were randomly taken of each biological sample.

Measuring of the Production of Reactive Oxygen Species via Dihydrorhodamine 123 (DHR)
Since the production of reactive oxygen species reaches its maximum after 30 min of PMA incubation [49], 35,000 cells/well were seeded and DHR (Thermo Fisher Scientific) was coincubated with PMA with a final concentration of 1.445 µM diluted in RPMI 1640, 1% FBS, and 1% PenStrep for 30 min at 37 • C and 5% carbon dioxide. Measurement was performed with a Gallios, Beckman Coulter, at an excitation of 488 nm and with the emission at 525 nm ± 25 nm (argon laser).

Determination of the Membrane Potential with Bis-(1,3-Dibutylbarbituric Acid)trimethine Oxonol (DiBAC 4 )
The determination of the membrane potential was performed with DiBAC 4 . (Thermo Fisher Scientific). Therefore, 35,000 cells/well were seeded and NETosis was induced with 20 nM PMA for 3.5 h and cells were further incubated 30 min with 250 nM DiBAC 4 . Measuring took place with a Gallios, Beckman Coulter, at an excitation of 488 nm and with the emission at 525 nm ± 25 nm (argon laser), as described previously by Löhrke et al. [50].

NETose-Microscopy of Living Cells
After the isolation of human neutrophils, 50,000 cells/well were seeded in a Poly-l-Lysine -coated 12-well chamber slide. To de-stress cells from the isolation procedure, the following step was an incubation in RPMI 1640 for 1 h at 37 • C and 5% carbon dioxide. Afterwards, the cells were washed twice with RPMI 1640 and stained with 1 µg mL -1 DAPI (Carl Roth) for 30 min before a further washing step was done. To stain the membrane, Deep Red Cell Membrane Stain (1:1000; Thermo Fisher Scientific) was added 15 min before NETosis induction, with 50 nM PMA. Pictures were taken at an interval of 2 per min.

Determination of Membrane Integrity via Propidium Iodide (PI) Staining
Membrane integrity was analyzed using PI (Miltenyi Biotec, Bergisch Gladbach, Germany; Annexin V-FITC Kit) in a final concentration of 1 µg mL -1 . After removing PI, nuclei were stained with Hoechst (Sigma Aldrich, St. Louis, MO, USA) in a final concentration of 1 µg mL -1 . Cells were directly analyzed using fluorescence microscopy (Carl Zeiss confocal laser scanning microscope LSM 800). Three pictures were randomly taken from three independently performed experiments. The total cell number and PI positive cells were counted.

Statistical Analysis
Data sets were analyzed with Graph Pad Prism 7.0 software using ANOVA and a multiple-comparison Tukey test or a t-test with Welch's correction, when the calculated values passed the Shapiro-Wilk normality test. Otherwise, the Kruskal-Wallis test in combination with Dunn's test for multiple comparisons was applied. Differences were considered statistically significant at p ≤ 0.05. Statistically significant differences are given the labels * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; and **** p ≤ 0.0001.

α2,8-Sialylated Particles Decrease the Production of Reactive Oxygen Species and Membrane Depolarization
NADPH oxidase represents a key enzyme in the signaling cascade during the formation of NET induced by PMA [51]. To test whether nanoparticles containing α2,8-linked sialic acid chains modulate the activation of neutrophils, ROS production was analyzed. To quantify ROS, DHR was used. In the presence of ROS, DHR is oxidized to cationic rhodamine 123, which exhibits green fluorescence. As shown in Figure 1A, in comparison to the neutrophils treated with PMA, unstimulated neutrophils assemble approximately 98% less ROS. When nanoparticles with sialic acid monomers were applied during the PMA stimulation, no effect on the PMA-induced ROS generation was observed. In contrast, the α2,8-sialylated nanoparticles statistically significantly decrease the ROS production that had been initiated by PMA. However, the production of ROS could not be completely abolished. Thus, chemically coupled α2,8-linked sialic acid chains seem to influence the activation of neutrophils.

The Swelling of Neutrophil Nuclei Is Not Influenced by α2,8-Sialylated Nanoparticles
The formation of ROS is directly linked to the translocation of neutrophil elastase into the nucleus because ROS serves as a substrate for myeloperoxidase. Subsequently, the increase in HOCl initiates the detachment of neutrophil elastase from the myeloperoxidase/neutrophil elastase complex, enabling the transfer of neutrophil elastase into the nucleus and triggering decondensation of DNA [4,9]. The result is that the nuclei lose their typical segmented structure and swell.
When we compared the nuclear structure of unstimulated neutrophils with that of the activated neutrophils in the presence of α2,8-sialylated nanoparticles, we observed that nearly all nuclei lost their segmented structure and the nuclei were swollen 4 h after the beginning of the PMA stimulation ( Figure 2). Thus, the results suggest that the decondensation of DNA could not be prevented. The occurrence of membrane depolarization after activation of the NADPH oxidase complex was described in the 1980s [52][53][54]. Experiments focused on inhibition of NADPH oxidase complex-induced electron transport showed that the inhibition of electron transport leads to attenuated membrane depolarization [54]. In line with these experiments, neutrophils of patients with chronic granulomatous disease (CGD) showed no membrane depolarization after PMA stimulation, as this disease is characterized by an impaired NADPH oxidase complex [52,55,56]. Since α2,8-sialylated nanoparticles inhibit ROS production, we also examined whether membrane depolarization is influenced. To this end, the polarization status of the cells was determined with DiBAC 4 , a substance that can enter depolarized cells. Here, DiBAC 4 can bind intracellular proteins exhibiting enhanced fluorescence and a red spectral shift. In Figure 1B, it is apparent that the stimulation of neutrophils with PMA results in cell membrane depolarization. Statistically significant differences between non-treated and stimulated cells were detected, which could not be influenced by monosialylated nanoparticles. However, in line with the ROS experiments, α2,8-sialylated nanoparticles inhibited membrane depolarization considerably.
In sum, the outlined results of the cell assays suggested that the essential steps during PMA stimulation, which trigger the formation of NET, are impaired by α2,8-sialylated nanoparticles.

The Swelling of Neutrophil Nuclei Is Not Influenced by α2,8-Sialylated Nanoparticles
The formation of ROS is directly linked to the translocation of neutrophil elastase into the nucleus because ROS serves as a substrate for myeloperoxidase. Subsequently, the increase in HOCl initiates the detachment of neutrophil elastase from the myeloperoxidase/neutrophil elastase complex, enabling the transfer of neutrophil elastase into the nucleus and triggering decondensation of DNA [4,9]. The result is that the nuclei lose their typical segmented structure and swell.
When we compared the nuclear structure of unstimulated neutrophils with that of the activated neutrophils in the presence of α2,8-sialylated nanoparticles, we observed that nearly all nuclei lost their segmented structure and the nuclei were swollen 4 h after the beginning of the PMA stimulation ( Figure 2). Thus, the results suggest that the decondensation of DNA could not be prevented. The translocation of neutrophil elastase to the nuclei is an important step during the decondensation of DNA and the formation of NETs [16]. To test whether the localization of neutrophil elastase was also unchanged due to the application of α2,8-sialylated nanoparticles, the enzyme was visualized using a polyclonal Ab against neutrophil elastase. Based on published data concerning the time point of the translocation of neutrophil elastase into the nucleus, immunofluorescence staining of neutrophil elastases was performed 2 h after the PMA treatment [16].
In contrast to the unstimulated neutrophils, neutrophil elastase was transferred into the nucleus after stimulation (Figure 3). In addition, the segmented structure of the nucleus was resolved. Comparable results were obtained using α2,8-sialylated particles. Thus, the reduction of ROS to approximately 40%, which was induced by α2,8-sialylated nanoparticles, does not prevent the transformation of the nuclei. The translocation of neutrophil elastase to the nuclei is an important step during the decondensation of DNA and the formation of NETs [16]. To test whether the localization of neutrophil elastase was also unchanged due to the application of α2,8-sialylated nanoparticles, the enzyme was visualized using a polyclonal Ab against neutrophil elastase. Based on published data concerning the time point of the translocation of neutrophil elastase into the nucleus, immunofluorescence staining of neutrophil elastases was performed 2 h after the PMA treatment [16].
In contrast to the unstimulated neutrophils, neutrophil elastase was transferred into the nucleus after stimulation (Figure 3). In addition, the segmented structure of the nucleus was resolved. Comparable results were obtained using α2,8-sialylated particles. Thus, the reduction of ROS to approximately 40%, which was induced by α2,8-sialylated nanoparticles, does not prevent the transformation of the nuclei.

α2,8-Sialylated Particles Inhibit the Release of NET
In addition, the release of NET was examined. The stimulation of NETosis leads to the formation of the DNA meshwork, which can be visualized with DAPI. As shown in Figure 4A,B, 4 h after PMA stimulation, the expected NET filaments are formed. In contrast, the unstimulated neutrophils retained their shape, including their segmented nuclei. However, the application of α2,8-sialylated nanoparticles prevented NET release although the nuclei were swollen. The calculation of the NET area shows that NET expansion was statistically significantly reduced.
The same experimental test setup was also performed with monosialylated beads, which showed no impact on the activation by PMA ( Figure 5). In addition, experiments were performed to investigate whether free sialic acid chains have the same effect or whether sialic acid chains have to be attached to a linker to inhibit the release of NETs. To this end, we used free sialic acid chains together with the α2,8-sialylated beads. Again, α2,8-sialylated nanoparticles had the capability to prevent NET release ( Figure 6). However, this effect was inhibited by free sialic acid chains. Thus, a linker molecule is necessary to inhibit the release of NETs.

α2,8-Sialylated Particles Inhibit the Release of NET
In addition, the release of NET was examined. The stimulation of NETosis leads to the formation of the DNA meshwork, which can be visualized with DAPI. As shown in Figure 4A,B, 4 h after PMA stimulation, the expected NET filaments are formed. In contrast, the unstimulated neutrophils retained their shape, including their segmented nuclei. However, the application of α2,8-sialylated nanoparticles prevented NET release although the nuclei were swollen. The calculation of the NET area shows that NET expansion was statistically significantly reduced.
The same experimental test setup was also performed with monosialylated beads, which showed no impact on the activation by PMA ( Figure 5). In addition, experiments were performed to investigate whether free sialic acid chains have the same effect or whether sialic acid chains have to be attached to a linker to inhibit the release of NETs. To this end, we used free sialic acid chains together with the α2,8-sialylated beads. Again, α2,8-sialylated nanoparticles had the capability to prevent NET release ( Figure 6). However, this effect was inhibited by free sialic acid chains. Thus, a linker molecule is necessary to inhibit the release of NETs.

α2,8-Sialylated Particles Prevent Cell Membrane Perforation
For the release of NETs, cell membrane rupture is necessary [9]. Using Deep Red Cell Membrane Staining, it is possible to visualize the cellular membrane during NETosis and its explosive bursts to release the DNA network (illustrated in Figure 7 and Supplemental 1; video). The resulting hole allowed the release of intracellular substances and the penetration of molecules into the cell. Interestingly, the intensity of the membrane staining became stronger during the alteration of the nuclear structure, indicating that the membrane properties of the neutrophils had changed before the release of NETs.

α2,8-Sialylated Particles Prevent Cell Membrane Perforation
For the release of NETs, cell membrane rupture is necessary [9]. Using Deep Red Cell Membrane Staining, it is possible to visualize the cellular membrane during NETosis and its explosive bursts to release the DNA network (illustrated in Figure 7 and Supplemental 1; video). The resulting hole allowed the release of intracellular substances and the penetration of molecules into the cell. Interestingly, the intensity of the membrane staining became stronger during the alteration of the nuclear structure, indicating that the membrane properties of the neutrophils had changed before the release of NETs. In order to test whether the outer membranes were still intact we used PI, a substance that can pass only perforated, not intact, biomembranes. Once PI enters the cell, it interacts with DNA, resulting in a red fluorescent staining. The calculation was performed by counting the total number of cells compared to the number of PI-positive cells. The results indicate that, compared to untreated neutrophils, PMA led to an increasing number of perforated cells (Figure 8). Similar results were obtained when monosialylated beads were added. However, the application of α2,8-sialylated beads resulted in a significantly reduced number of PI-positive cells, demonstrating that α2,8-sialylated beads inhibit the perforation of the biomembrane, and thus, the release NET. In order to test whether the outer membranes were still intact we used PI, a substance that can pass only perforated, not intact, biomembranes. Once PI enters the cell, it interacts with DNA, resulting in a red fluorescent staining. The calculation was performed by counting the total number of cells compared to the number of PI-positive cells. The results indicate that, compared to untreated neutrophils, PMA led to an increasing number of perforated cells (Figure 8). Similar results were obtained when monosialylated beads were added. However, the application of α2,8-sialylated beads resulted in a significantly reduced number of PI-positive cells, demonstrating that α2,8-sialylated beads inhibit the perforation of the biomembrane, and thus, the release NET.

Discussion
In 2015, Spence et al. discovered that nanoparticles coated with the murine siglec-E ligand, α2,8-linked sialic acid residues, decrease macrophage-driven inflammation [33]. Based on this discovery, we decided to investigate the potential of α2,8-sialylated particles to inhibit the release of NETs. Interestingly, the application of α2,8-sialylated particles decreased the production of ROS in

Discussion
In 2015, Spence et al. discovered that nanoparticles coated with the murine siglec-E ligand, α2,8-linked sialic acid residues, decrease macrophage-driven inflammation [33]. Based on this discovery, we decided to investigate the potential of α2,8-sialylated particles to inhibit the release of NETs. Interestingly, the application of α2,8-sialylated particles decreased the production of ROS in PMA-stimulated neutrophils, while monosialylated particles showed no effect. In line with the reduced ROS, depolarization was also inhibited by these sugar-coated nanoparticles. Since the activation of the NADPH oxidase complex includes PKC and the Raf-MEK-ERK pathway, resulting in the formation of ROS [14,57], and siglecs are known to inhibit the immune response by counteracting kinases via the recruitment of the phosphatases SHP1 and SHP2 [28,30], the involvement of siglecs may trigger these observations. As siglec-5 is known to prefer α2,8-linked sialic acids, decreased ROS production and depolarization might be the result of the interaction between α2,8-linked chains and siglec-5 of neutrophils [34][35][36]. However, the impaired ROS production and depolarization did not prevent swelling of the nuclei. This indirectly indicates that neutrophil elastases still translocate to the nuclei, because neutrophil elastase is necessary for decondensation of DNA and, therefore, for nuclei swelling as described previously [16].
Nevertheless, the release of NETs was retained and the α2,8-sialylated particles prevented the perforation of the outer membrane ( Figure 8). Interestingly, a comparable effect was described for lactoferrin [58]. In line with α2,8-sialylated beads, elastase translocates to the nucleus in the presence of lactoferrin and nucleus swelling takes place. In addition, in the case of lactoferrin, the last step, the release of NET is inhibited. Okubo and colleagues observed that lactoferrin accumulates on the surface of neutrophils [58]. They suggested a physical blocking barrier, which is mediated via a "lactoferrin-shell". The first released DNA aggregates with lactoferrin forming a plug, which prevents the release of NETs [58]. As recently shown, polysialic acid (polySia) can interact with lactoferrin and supports the effect of lactoferrin to inhibit NET release [59]. However, whether the membrane was still intact was not tested.
In addition to lactoferrin, α2,8-linked polymers of sialic acids are known to induce membrane interactions, because these polymers can interact with phospholipids [60]. This interaction usually takes place in ordered regions of membranes, such as lipid rafts [61]. The staining of cell membranes with Deep Red during NETosis (Figure 7) suggested that fluid and dynamic processes occur in the biomembrane, which might mediate such an interaction. Thus, we suggest that the membrane is stabilized by the multivalency of α2,8-sialylated nanoparticles, which might act as "cross struts". Furthermore, perhaps an interaction with membrane proteins, such as siglec-5, supports the stabilization of the cell membrane in a comparable fashion. These hypotheses are supported by the observation that free sialic acid chains inhibit the effects of the α2,8-sialylated nanoparticles.
However, in contrast to lactoferrin, α2,8-sialylated beads influence ROS production [58], which is essential for the formation of NETs induced by PMA. The observed inhibition of ROS formation might be the result of an activation of siglec-5 by the α2,8-sialylated beads. The lower ROS content might additionally reduce the dynamic of the neutrophil burst. Presumably, the combination of these two effects (the "cross struts" effect and lower ROS levels) leads to stabilization of the biomembrane, which prevents the cell membrane from rupturing.
Although the exact mechanism of action has been unknown until now, α2,8-sialylated nanoparticles are a promising option for modulating NETosis during NET-related pathologies. Furthermore, for application as a pharmacological tool, the α2,8-sialylated nanoparticles must be tested in different in vivo models where NETs are the main cause that induces pathology.