Copolymer functional groups modulate extracellular trap accumulation and inflammatory markers in HL60 and murine neutrophils

Undesirable host responses to implants commonly lead to impaired device function. As the first immune cell to respond to inflammation, activated neutrophils release antimicrobials and neutrophil extracellular traps (NETs) that prime microenvironments for macrophages and other infiltrating cells. This research aims to understand how functional groups in copolymers of isodecyl acrylate (IDA) that are known to modulate healing in vivo, modulate neutrophil cells. Phorbol myristate acetate-activated HL60 cells and bone marrow-derived murine neutrophils (BMDN) were incubated with coatings of IDA copolymerized with, methacrylic acid (MAA films), methyl methacrylate (MM films), or MM functionalized with hexamethylenediamine (HMD films). Cells incubated on HMD films resulted in increased accumulation of NETs at the film’s surface in comparison to other copolymers because of increased adhesion of HL60 onto HMD films or increased rates of NETosis from BMDN. Overall, lower inflammation was observed with cells on MAA films. HL60 cells had no increase in classical inflammatory markers such as tumor necrosis factor alpha and intracellular adhesion molecule-1, whereas HL60 on HMD films had increases in these same markers. Taken together, these studies give important insights into how neutrophils interact differently with functionalized copolymers and the proteins that adsorb to them, with MAA (carboxyl groups) leading to behavior associated with lower inflammation and HMD (amine groups) with higher inflammation and accumulation of NETs.


Introduction
Chronic inflammation and the resulting host response continue to be significant obstacles for medical device implants and biomaterials research. After the implantation of a medical device, local proteins adsorb and inflammatory cells interact with both the surface of the device and the proteins that adsorb. Chronic inflammation surrounding a device results in the formation of a fibrous capsule that encapsulated the device, segregating it from its local microenvironment [1]. In the case of diabetic sensors or other devices that must interact with their local microenvironment, this results in device failure. The proteins that adsorb to device surfaces and the resulting cell-material interactions are influenced by surface properties, including hydrophobicity, surface chemistry, and topography [2]. Notably, some polymeric materials show improved healing after implantations. The concept that materials alone can modulate the healing response in the absence of delivered proteins, drugs, or cells is intriguing and its further development could be applied as a simpler method to improve medical device integration.
Methacrylic acid (MAA) copolymerized with isodecyl acrylate (IDA) forms MAA copolymers (MAAcoIDA) previously shown to improve the healing response to implants relative to methyl methacrylate copolymers (MMcoIDA) by reducing fibrosis and increasing vascularization. For example, porous polypropylene materials coated with 40% MAAcoIDA and implanted subcutaneously into mice promote the development of functional blood vessels that connect to the bloodstream [3]. Also, implanted silicone tubes coated with 40% MAAcoIDA copolymers can vascularize subcutaneous tissue to sustain the growth of islets enabling them to secret insulin at levels that return diabetic Sprague Dawley rats to normoglycemia for up to 70 d [4]. The biological mechanisms involved in the observed vascularization are not fully understood. In vitro, markers of inflammation are also noted to be modulated by the MAAcoIDA copolymers using monocyte cell lines. For example, the differentiated form of the monolytic cell line THP-1 (dTHP-1) macrophage-like cells incubated on the IDA-based copolymer coatings had increased expression in inflammatory genes, such as tumor necrosis factor alpha (TNF-α), interleukin (IL)-6, and IL-1β when incubated on MAAcoIDA in comparison to MMcoIDA copolymers, illustrating the modulation of inflammatory responses in monocytes to surfaces with small changes in surface chemistry [5].
Prior to the arrival of monocytes at implantation sites, neutrophil cells first arrive and interact with the device surface and proteins adsorbed to the device surface. Traditionally in polymeric biomaterials research, the role of neutrophils in host responses is centered on their role to recruit inflammatory cells through degranulation and the removal of pathogens through phagocytosis. However, neutrophils have been observed to polarize, specifically in the tumor microenvironment as pro-tumor or anti-tumor neutrophils, collectively described as tumor-associated neutrophils (TANs) [6,7]. Polarization has also been observed in other circumstances such as myocardial infarction and injury. In these studies, N1 generally describes an inflammatory polarization state and N2 describes an anti-inflammatory polarization state [8,9]. Furthermore, a more recently discovered role of neutrophils in host responses to an implant is NETosis; a process when cells extrude a web of DNA containing cytosolic proteins and primary granules (i.e. myeloperoxidase, neutrophil elastase) induced by inflammatory stimuli [10]. NETosis occurs through either a lytic mechanism, resulting in cell lysis and death upon the release of neutrophil extracellular traps (NETs), or a non-lytic mechanism, where NETs are released by vesicles and the cell maintains its ability to phagocytose, etc [11]. The potency and early arrival of NETs are of considerable interest, but little research has focused on their release from neutrophils interacting with polymeric biomaterials.
Examining the effects of MAAcoIDA copolymers on neutrophil cells will decipher the role of early innate immune responses on implanted polymer devices and how they can be used to improve host responses. MAA and MM copolymers with IDA as described previously [5] and in this work, can be coated onto glass coverglass as films (to form MAA films and MM films) enabling the testing of cell behavior directly on the materials. In this research, HL60 cells and freshly isolated murine bone marrowderived neutrophil cells (BMDN) were incubated on MAA films and MM films as well as hexamethylenediamine (HMD)-modified MM films (HMD films), and their viability and NETosis measured by immunofluorescence. Their behavior was also evaluated by measuring gene expression or cytokine release. The behavior of both cell types was modulated by the different copolymers with MAA resulting in indicators of lower inflammation and HMD higher inflammation.

Preparation of films
MM or MAA was polymerized with IDA (figure 1(A), (B)) and cast onto glass coverslips as per Wells and Sefton [5] to form MAA and MM films. Mediumcontent 30%-40% MAA or MM was used because previous studies show they modulate dTHP1 cells to higher degrees than lower content 20% [5] and because recent publications show that 40% has a regenerative healing phenotype [3,4]. To synthesize 40% MM copolymerized with 60% IDA, 1% w/w BPO was dissolved in a mixture of 2.6 ml MM (0.024 mol) and 8.7 ml IDA (0.036 mol). 12 ml of endotoxin-free water was added to the flask and the resulting solution was purged with nitrogen gas for 5 min. The flask was placed in a 70 • C water bath under reflux for 1.5 h. The same protocol was used to copolymerize 40% MAA with IDA, 1% w/w BPO was dissolved in 2.1 ml (0.025 mol) of MAA and 9 ml (0.037 mol) IDA. To purify the copolymers, excess water and reagents were decanted and the solid polymer was dissolved in ∼60 ml of THF. The polymer solution was precipitated into excess endotoxin-free water and the precipitate was collected and dried. Proton nuclear magnetic resonance (NMR) of the MM copolymer in CDCl3 was used to monitor the polymerization and the integration of peaks at 3.65 (3H, MM) and 4.07 (2H, IDA) used to determine the percentage of MM in the copolymers and peaks monitored for the proper product as per Wells and Sefton [5]. MAA and IDA cannot be readily distinguished using proton NMR, so back titration of MAA copolymers incubated in NaOH was used to determine the percentage of charged MAA groups [5]. The MM and MAA copolymers were then coated onto 12 and 18 mm-diameter, round glass coverslips by casting solutions of polymer in THF onto coverslips to a final mass of 5.25 mg cm −2 , followed by air drying for 96 h to form MAA films (coatings of MAA copolymer) and MM films (coatings of MM copolymer).
MM films were functionalized with amine groups by an aminolysis reaction with HMD (figure 1(C)) [12]. Each MM film was rinsed twice with isopropanol and then once with endotoxin-free water, and then incubated in 2 ml of a 20% v/v solution of HMD in sodium tetraborate buffer made of 0.1 M sodium tetraborate and 0.15 M sodium chloride at a pH of 12.5 (adjusted with 1 N NaOH) at 37 • C on a rotating plate at 60 rpm for 2 h to produce HMDmodified MM films (HMD films). Samples were then washed twice with endotoxin-free water and air-dried for 48 h. The resulting number of amine groups was quantified using the ninhydrin assay. Briefly, three HMD films (and controls of MM films) were each incubated in 2 ml of ninhydrin reagent (Sigma Aldrich, Oakville ON, CA) for 2 h. Ninhydrin reagent was added in a 1:2 ratio to water and samples were incubated in an oil bath at 100 • C for 10 min. After cooling, 150 µl samples were read at 570 nm by spectrophotometry and the amount was calculated using a calibration curve made with solutions of glycine from 1 × 10 −5 to 1 × 10 −4 M.

Cell-material interactions
Both coated and uncoated coverslips were disinfected by washes with 70% ethanol or by UV treatments prior to use in cell studies. Using the LAL pyrochrome endotoxin test kit and LAL reagent water, purchased from Associates of Cape Cod (East Flamouth, MA), the endotoxin was tested and was below acceptable levels (under 0.25 EU/100 mg). Either the HL60 cell line or freshly isolated BMDN were incubated on the films and then cell behavior was evaluated. The human cell line (HL60) and freshly isolated BMDN cells were incubated with different films to observe the interaction of large homogenous populations of human and murine cells. The human cells will enable the observations of gene expression over short time frames and be comparable to previous studies by Wells and Sefton [5] and the murine cells enable comparisons with implant studies that observed changes over short timeframes [1,2]. After the incubations of the cells with the different films, immunofluorescence of markers of NETosis with adhered cells, gene expression, and protein expression were evaluated. Immunofluorescence was chosen to evaluate cell behavior instead of flow cytometry because cells adhered to materials and undergoing NETosis are fragile and not able to be processed for flow. Flow cytometry is best reserved for analyzing freshly isolated neutrophils in suspension.

HL60 cell culture
The HL60 cell line is from promyeloblasts (adult, female origin) from the American Type Culture Collection (ATCC) (Burlington ON, CA) and cultured in Iscove's Modified Dulbecco's Medium (IMDM) (ATCC, Burlington ON, CA) supplemented with 20% fetal bovine serum (FBS) from Wisent Inc. (Saint-Jean-Baptiste QC, CA) and 1% penicillin/streptomycin solution (Sigma Aldrich, Oakville ON, CA). HL60 cells were maintained at a concentration between 1 × 10 5 and 1 × 10 6 cells ml −1 of medium and incubated at 37 • C and 5% CO 2 . Solutions of 50 nM of phorbol myristate acetate (PMA) in medium were used to differentiate HL60 cells to a neutrophil-like phenotype.
HL60 cells were seeded at 400 000 cells cm −2 on sterile 18 mm-diameter glass coverslips, MM films, MAA films, or HMD films in a 12-well plate. 50 nM PMA was added to each sample to promote cell differentiation and adherence and samples were then incubated at 37 • C and 5% CO 2 for 24 h or 48 h. For 48 h studies, the cell medium was replaced at 24 h with medium without PMA. After incubation with the different copolymers, the alamarBlue assay, quantitative polymerase chain reaction (qPCR), or stains and immunofluorescence were used to assess cell behavior. The chosen times of 24 and 48 h were to allow the evaluation of cells after sustained periods of NETosis (which can start at 4 h) and then after the PMA was removed. qPCR was done at 24 and 48 h to match the immunofluorescence studies and because previous studies using THP1 cells showed changes were noted at 24 and 48 h [5].

Isolation of murine neutrophils
Murine neutrophils were isolated from the bone marrow of male C57BL/6J mice, aged 8-12 weeks, purchased from Charles River Laboratory (Wilmington MA, USA). All cell isolations were performed following compliance with the Institution Animal Care Committee and Canadian Council for Animal Care (Protocol 2019-1895). Mice were habituated in the animal facility at Queen's University for at least seven days before sacrifice. Mice were euthanized primarily via CO 2 inhalation, followed by bilateral pneumothorax puncture. Cells were cultured in Roswell Park Memorial Institute 1640 (Gibco, Thermo Fisher Scientific, Whitby ON, CA) supplemented with 10% heat-inactivated FBS from Wisent Inc. (Saint-Jean-Baptiste QC, CA). Neutrophils were enriched from the total population using EasySep™ Mouse Neutrophil Enrichment Kit from StemCell Technologies (Burnaby BC, CA). Cells were stained with Zombie Violet, and Ly6G and CD11b antibodies from Bio-Legend for phenotyping using flow cytometry using the CytoFLEX from Beckmann Coulter. The murine bone marrow population was enriched to approximately 80% Ly6G + CD11b + cells, identified as neutrophils. The remaining 20% of cells were Ly6G − CD11b + but the phenotypes of this population were not identified.
The isolated murine BMDN were seeded in RPMI 1640 (0.5% hi-FBS, 1% P/S) directly onto copolymer film-coated coverglass with no additional treatment. Controls included cells treated with 100 nM PMA or 25 µg ml −1 LPS with coverglass. Cells were incubated for 2 and 16 h at 37 • C and 5% CO 2 . Afterward, stains and immunofluorescence and analysis of protein content in the medium were done to assess cell behavior. The 2 and 16 h incubation times used for the BMDN enabled the evaluation of viable cells. Previous studies show that neutrophils from a variety of murine models have undergone high levels of NETosis-based death after 10 h [13].

AlamarBlue assay for HL60 cells
After incubation of HL60 with the different films for 24 h or 48 h, the medium was removed from each well and the cells were gently rinsed with 1 ml of phosphate buffered saline (PBS, at pH = 7.4). Each sample was incubated at 37 • C with 0.1 ml alamarBlue in 0.9 ml of medium for 5 h. Hundred microliter of each sample was transferred in triplicate to a 96-well plate and absorption was read at 570 and 600 nm using a Synergy H1 plate reader from BioTek. Results show the average absorbance of medium from cells incubated on films normalized to cells incubated on glass.

Quantitative PCR for HL60 cells
As an additional study, after incubation of HL60 with the different films for 24 or 48 h, the cells on the films were moved to fresh well plates (by moving the coverslip with the film) to collect RNA from cells adhered to the films (but not the surrounding well). The 24 and 48 h timepoints were chosen to match the immunofluorescence studies and because studies with dTHP cells on similar coatings showed differences in gene expression at 48 h (albeit, more were noted at 72 h) [5]. Importantly, in general, it takes longer periods for gene expression changes in cells responding to materials versus cells incubated in dissolved cytokines and so longer periods that matched other measurements were the focus. Prior to the isolation RNA, the cells adhered to the coatings were washed gently with 1 ml of PBS. RNA was collected and purified using the Aurum™ Total RNA Mini Kit from Bio-Rad (Mississauga ON, CA). Cells were lysed with lysis solution supplemented with βmercaptoethanol and cell scrapers were used to scrape coverslip surfaces. RNA was purified according to the manufacturer's instructions with DNAse digestion to improve purity. Afterward, cDNA was synthesized from the isolated RNA using the High-Capacity RNAto-cDNA Kit from Thermo Fisher Scientific (Whitby ON, CA). The protocol was followed as per the manufacturer's instructions for a 20 µl synthesis volume and stored at −20 • C until use. We did not pursue gene expression studies with BMDN to minimize the use of animals.
qPCR reactions were performed using SsoAd-vanced™ Universal SYBR® Green Supermix from Bio-Rad (Mississauga ON, CA). Primers for genes of  1). To measure the expression of the goi, in a 384 well plate, a mixture of 0.5 µl of primer stock in 5 µl of SsoAdvanced mix was added to each well, followed by 10 ng of cDNA in 4.5 µl of nuclease-free water (per reaction). The plate was spun and analyzed in a CFX384 Real-Time System from Bio-Rad. The thermal cycling protocol was run according to guidelines. An initial first step of polymerase activation and DNA denaturation was done at 95 • C for 30 s, followed by cycles of amplification (denaturation at 95 • C for 10 s, and annealing and extension at 60 • C for 30 s) and a plate read for a total of 40 cycles. Finally, a melt curve was generated from 65 to 95 • C (+0.5 • C per cycle, 0.5 • C s −1 ).

qPCR relative expression and statistics
Quantification cycles (Cq) were determined using Bio-Rad CFX Maestro software. The efficiency used for the calculation of relative quantities (RQ) of each gene was assumed 100% (E = 2) because the primer efficiencies were between 90 and 110 (supplemental  table 1 and supplemental table 2) [19]. The RQ (equation (1)) of the goi was normalized to the geometric mean RQ of the reference genes, yielding the normalized relative quantity (NRQ, equation (2)). The relative expression ratio (equation (3)) is the ratio of the NRQ of the treated biological group to the NRQ of the control group (uncoated coverslips). An analysis of variance (ANOVA) on the logtransformed NRQ values (logNRQ) was performed to test for statistical differences in gene expression among the treatment groups at 95% confidence level. The gene expression of intracellular adhesion molecule 1 (ICAM-1), TNF-α, arginase 1 (Arg-1) and chemokine ligand 17 (CCL17) are presented as normalized relative expression ratios of one sample to another using TBP and RPS18 as internal reference genes. Arg-1 expression did not have any significant changes in cells on the different surfaces and had a large standard error (SE) due to high Cq values. These results indicate Arg-1 is lowly expressed in HL60 cells at these time points, and the data is not reported, A gene was considered significantly different if the relative expression ratio was less than 0.67 or greater than 1.5 (0.67 > ratio > 1.5) and the logNRQ p-values were less than 0.05 (p < 0.05). Error bars are presented as SE of the ratio, calculated using equation (4) as described by Rieu and Powers [20], The resulting slides were imaged with EVOS FL fluorescent microscope from Life Technologies (ThermoFisher Scientific, Whitby ON, CA). Five to six discrete areas of each coverslip (with film) were imaged at 10× using 4 ′ ,6-diamidino-2-phenylindole (DAPI), green fluorescent protein (GFP), and Texas Red channels, for independent and overlaid images. Total blue, green, red, and blue-green spots were counted using Image J (downloaded from https:// imagej.nih.gov/ij/). The total cells adhered was estimated as the total blue, green, and blue-green costained cells, while dead cells were green. Cells costained with citH3 and either NucBlue or Sytox Green were considered as undergoing NETosis. CitH3 and Sytox Green were normalized to the total adhered cells to account for significant differences in adhered cells between films.

Protein quantification of BMDN cell supernatants
After the incubation of BMDN with the films, the supernatants were individually collected and centrifuged at 300 g to remove debris, and the supernatant was transferred to fresh vials for storage at −20 • C. A custom five-panel MILLIPLEX MAP Mouse Cytokine/Chemokine Magnetic Bead Panel from Millipore Sigma was used to detect quantities of IL-10, IL-1β, monocyte chemoattractant protein 1 (MCP-1 or CCL2), macrophage inflammatory protein 2 (MIP-2 or CXCL2), and macrophage inflammatory protein 1 alpha (MIP-1α or CCL3) in the murine neutrophil supernatants. IL-10, IL1β, MCP-1 or CCL2, MIP1-α/CCL3 and MIP-2/CXCL2 are produced by murine neutrophils in vivo and in vitro [22]. The assay was performed as per manufacturer's instructions, incubating the beads with samples for 16 h, per the two day protocol. Concentrations were measured on a Luminex 200 (Luminex Corporation, Toronto ON, CA).

Statistics
Data was compiled using three biological replicates and technical replicates as specified and error is reported as the standard deviation in HL60 experiments unless otherwise indicated. At the 2 h time-point, BMDN staining studies used an n = 4 for control groups (untreated cells on glass, LPS-and PMA-treated cells) as a preliminary study was included, while n = 3 for MM, HMD, and MAA films. The 16 h BMDN studies used an n = 4 for all groups except MAA (n = 3). SPSS was used to perform a one-way ANOVA with a Tukey's HSD posthoc when comparing multiple treatment groups. Statistical significance was assessed at a 95% confidence interval assuming normal distribution and unequal variance. Murine experiments report the error as the standard error of the mean.

Results
The synthesized MMcoIDA had 44 ± 1.37% methyl groups and MAAcoIDA 32.4 ± 4.27% carboxyl groups. The coverslips were coated with 5.94 mg of copolymer resulting in films with a thickness of 30.5 µm with approximately 112 nmol of COOH groups at the surface of the resulting MAA films, 160 nmol methyl groups for the MM films, and 56 nmol amine groups for the HMD films. Previous studies with films with 40% MAA or MM (copolymerized with IDA) coated using the same process had similar roughnesses with Ra = 114 ± 16 nm for MAAcoIDA and ∼20 nm for MMcoIDA and x-ray photoelectron spectroscopy (XPS) in these studies established low contamination from silicone [5]. The behavior of HL60 and BMDN on the different chemistries identified trends in neutrophil-material interactions. Figure 2 shows the alamarBlue absorbance ratios for HL60 on different films, seeded at 400 000 cells cm −2 for 24 and 48 h with 50 nM PMA (the medium was replaced at 24 h to remove the PMA). The values represent relative live cell numbers, although metabolic activity may be altered during NETosis [23] and may also influence the overall observations. Cells incubated on MAA films show significantly higher alamarBlue ratios than those incubated on MM films (p < 0.001 at 24 h, p = 0.002 at 48 h). AlamarBlue ratios of cells incubated with HMD films were also significantly higher than MM films after 24 h (p < 0.001) but was significantly lower than MAA films (p = 0.013) at 48 h.

Viability and NETosis of HL60 cells incubated on the films
HL60 cells were stained after 24 and 48 h incubations with the different films, and representative images are shown in figure 3. The absolute counts of adhered cells are quantified in figure 3(A). Interestingly, after 24 h, fewer cells were adhered to the polymer films relative to glass, with significantly more cells adhered to glass than all films: MM films (p = 0.011), MAA films (p = 0.013), and HMD films (p = 0.012). Of note was a high, significant increase  NETosis was visualized by antibodies to citH3, the product of converting arginine to citrulline in the histone H3 protein during NETosis. NETosis is indicated by the ratio of citH3-stained cells to total adhered cells on each surface ( figure 3(C)). The number of adhered cells undergoing NETosis on MAA films (36 ± 11% at 24 h, 20 ± 7% at 48 h) is consistently lower than both MM films (55 ± 13% at 24 h, 40 ± 1% at 48 h) and HMD films (57 ± 1% at 24 h, 36 ± 2% at 48 h), though only statistically significant compared to MM at 48 h (p = 0.022). However, because there is high adherence of cells to HMD films, the overall quantity of NETs present on HMD films after 48 h is much larger than either MM films or MAA films.

Gene expression of HL60 cells incubated with the films
Gene expression measured changes in HL60 behavior when they were activated by PMA onto the different films. As shown in figure 4, ICAM-1 and TNF-α are expressed in HL60 cells with significant differences in their expression on the different surface chemistries. After 24 h, ICAM-1 expression is decreased in cells incubated on MAA films relative to MM films (p = 0.006) and expression is increased almost four-fold in cells incubated with HMD films relative to MAA films (p = 0.014). After 48 h, ICAM-1 expression is significantly upregulated on HMD films relative to both MM films (p = 0.004) and MAA films (p = 0.026). MAA films demonstrated consistent downregulation of ICAM-1 at both time points, while HMD films showed consistent upregulation in all comparisons.
An initial increase in expression of TNF-α is observed in cells incubated on MAA films after 24 h, though not significant, and decreases by 48 h to show no change in expression relative to MM films (figure 4). Expression on MM films is unchanged relative to glass at both timepoints, however expression on HMD films shows a seven-fold increase than on glass (p = 0.005) and a five-fold increase than on MAA films (p = 0.026) after 48 h.
Expression of CCL17 was highly variable at 24 h on the different surface chemistries. CCL17 had high Cq values (>32) and is lowly expressed in HL60 cells. Though expression was very low, a decrease in CCL17 is seen in cells on HMD films relative to glass (p = 0.038) after 48 h, where Cq values had less deviation. CCL17 and Arg-1 (data not shown due to low expression with Cq values >35 after 24 h and Cq >32 after 48 h) were chosen to explore potential pro-healing characteristics, however, activating HL60s with PMA likely interferes with their expression profiles.

Viability and NETosis of BMDN incubated with the films
Because HL60 is a cell line that requires PMA to adhere to the films and there is some controversy over its innate ability to promote NETosis, BMDN behavior on the films was also measured. The BMDN adhered to all the films (and control coverglass) without any added activators, and representative images of BMDN are shown in figure 5. The total number of cells adhered to films and controls (coverglass with PMA, LPS) are shown in figure 5. Large variation was observed in the total adhered cells counted on the different films and controls, and no statistically significant differences were observed. After 16 h, fewer cells remain adhered than at 2 h, and cells on MAA films have a reduction in adherence relative to all other samples.
The ratio of live cells to adhered BMDN was similar between all groups at 2 h ( figure 5(B)). By 16 h, cells incubated with LPS had higher ratios with 64 ± 6% of the adhered cells viable, which was significantly higher than all the film groups: MM films (p = 0.047), MAA films (p = 0.004) and HMD films (p = 0.005). Viable cell ratios were also increased when treated with PMA than when incubated on MAA films (p = 0.026) and HMD films (p = 0.037).
The ratio of NETs released by BMDN incubated on the different films was similar with one notable difference ( figure 5(C)). At 2 h, a significant increase in citH3-positive cells is seen with cells BMDN incubated on HMD films compared to PMA (p = 0.032) or LPS-stimulated cells (p = 0.005) and those incubated with MAA films (p = 0.006) or MM films (p = 0.033). At 16 h, 48 ± 8% of cells on HMD films are citH3-positive, and though this is increased relative to cells incubated with MAA films (18 ± 7%), it is not significantly higher (p = 0.295).

Protein expression of BMDN
Cell supernatants from BMDN incubated on coated and uncoated coverslips for 2 or 16 h were assessed for five proteins. Of the proteins measured, only MIP1α after 16 h, and MIP-2 after 2 and 16 h were notably released by BMDN incubated with the films, at a wide range in concentrations per trial. LPSstimulated BMDN as the positive control produced detectable levels of MIP-2, MIP1-α, and IL1-β at both timepoints, and MCP-1 and IL-10 at 16 h (supplemental figure 2). MIP1-α, or CCL3, is a chemoattractant for neutrophils and monocytes. Large variation was seen between samples; however, secretion was markedly increased upon stimulation with LPS and PMA. After 2 h, cells that were incubated with LPS secreted 55 ± 7 pg ml −1 . No MIP1-α was detected from cells incubated with PMA or on the various films after 2 h. However, after 16 h small concentrations were present from cells incubated on all surfaces as seen in figure 6(A). MIP-2, or CXCL2, is the equivalent of IL-8 in human cells and is known to be produced by neutrophils at high levels and early time-points after injury. Detection of MIP-2 from unstimulated neutrophils on glass and coatings was highly varied, with MIP-2 detected in some samples per treatment group, but not all. Cells in all groups produced detectable concentrations of MIP-2 with no significant differences among cells incubated on the different films as shown in figure 6(B).

Discussion
This research aimed to better understand how neutrophils interact with copolymers films previously shown to have different healing responses in vivo and to modulate macrophage-like cell responses in vitro. HL60 and freshly isolated murine neutrophil (BMDN) behavior when incubated with films made from copolymers of IDA with MM, MAA, or modified MM to have amine groups (HMD) enabled the exploration of large homogenous cell populations. MAA films are of interest because MAA copolymers have promoted angiogenesis in different mouse models in vivo [4,24], altered the inflammatory cytokine profile of macrophage cells in vitro [5], and were associated with the infiltration of neutrophils at early time points after subcutaneous injections of beads [24]. The resulting vasculature is functional and able to support the growth of functional islets that reduce glucose levels [3]. In general, the MAA and MM films have a well-established safety profile through numerous in vivo and in vitro studies, whereas the HMD films serve as a comparison, knowing that aminated surfaces often have worse outcomes in vivo with fibrous encapsulation or increased inflammation compared to carboxyl or hydroxyl groups (for example) [25,26]. Understanding the role of early innate immune responses to these polymer coatings will enable a deeper understanding that can be applied to the development of new materials and devices.

Polymer film composition modulates HL60 viability and NETosis
HL60 cells incubated with the activator PMA will adhere to cell culture surfaces and undergo lytic NETosis [11]. Figure 3 shows that HMD-modified coatings increase the number of HL60 cells adhered to the coating and therefore the overall accumulation of NETs at the surface of the coatings. Importantly, the presence of NETs or an increased amount of NETs can modulate other immune cells indicating the importance of observations in figure 3. For example, THP1 cells incubated on surfaces pretreated with NETs will be stimulated to adhere in the absence of differentiating factors, such as PMA, and these THP1 cells are also shown to have reduced viability [27].
The alamarBlue assay assesses cell viability through aerobic respiration and metabolic activity. Figure 2 shows increased alamarBlue ratios of HL60 cells on MAA films coatings, implying higher viability of cells when incubated with MAA films compared to MM films, which agrees with findings of dTHP-1 and HUVEC viability with these coatings observed by Wells and Sefton [5]. Interestingly, the live cell ratios shown in figure 3 did not show the same trends and this may be due to loosely adhered cells being washed off during the staining process. ICAM-1 expression, which has been shown to regulate cell adhesion [28], is significantly upregulated in cells on HMD films and downregulated in cells on MAA films relative to the other coatings, particularly after 48 h. Loosely adhered live cells on the MAA films may have washed off during staining but not prior to the ala-marBlue assays. Furthermore, neutrophils produce energy primarily through glycolysis, and it has been shown that glucose or pyruvate is required for human neutrophils to release NETs and that glucose uptake is increased in PMA-stimulated neutrophils which may further modulate results [23].

Inflammatory genes TNF-α and ICAM-1 are upregulated in HL60 cells on HMD films
Gene expression studies explored genes indicative of polarization as seen in the tumor microenvironment and after myocardial infarction [6,8]. TNF-α and ICAM-1 were chosen to measure an inflammatory neutrophil phenotype while CCL17 and Arg-1 were chosen to measure a pro-healing phenotype [6,29]. However, because the HL60 cells were stimulated to become inflammatory with PMA, there was little expression of CCL17 and Arg-1 (data not shown).
The expression of ICAM-1 increased in HL60 incubated on HMD films and decreased in HL60 on MAA films relative to MM films. While ICAM-1 is most often implicated in the infiltration stage of inflammation to enable cell-cell adhesion and rolling of neutrophils on the endothelium to enable transendothelial migration, ICAM-1 also binds certain extracellular matrix proteins, such as fibrinogen [28]. As ICAM-1 may play a role in cell-cell and cell-tissue adhesion [28], the upregulation of ICAM-1 seen in figure 4 on cells grown on HMD films provides some insight into the increased cell adherence on HMD films seen in figure 3.
The expression of TNF-α in HL60 incubated with HMD films was five times that of MM films after 48 h. TNF-α is a pro-inflammatory cytokine implicated in numerous diseases and is responsible for many actions including cell proliferation, survival, and differentiation [30]. Furthermore, soluble TNFα is known to induce NETosis in in-vitro studies [31], which may modulate nearby cells in vitro or in vivo.
In the tumor microenvironment where neutrophil polarization has been most studied, upregulated TNF-α, ICAM-1, and increased NETosis, are factors associated with N1 polarization towards a more inflammatory phenotype (anti-tumor TANs). The presence of N1 TANs in a tumor microenvironment results in reduced local vascularization and increased cytotoxic mediators are known to prevent tumor growth [7]. Both TNF-α and ICAM-1 were upregulated in HL60 incubated on HMD films in addition to increased absolute quantities of citH3, suggesting that HMD films (and proteins that adsorb to them) enhance the inflammatory response compared to MAA films and MM films. Although ICAM-1 and TNF-α are pro-inflammatory indicators and suggest that the cells are polarizing toward an N1 phenotype, timepoints would need to be extended to determine the biological impact of these expression profiles.

Polymer film compositions modulate BMDN adhesion and NETosis
BMDN were isolated as a primary cell source that does not require PMA to adhere to materials. The magnitude of adhered cells was similar between BMDN and HL60, with the live cell ratio in BMDN incubated with MAA films noticeably lower than with MM films and HMD, and significantly lower in HL60s at the 24 h time point. The higher live cell ratio observed in BMDN incubated with LPS is consistent with the literature as LPS is known to increase neutrophil survival, which is known to also increase the secretion of inflammatory proteins [32]. Decreased neutrophil survival on MAA films at 24 h points to a decreased acute inflammatory response promoted by MAA.
CitH3 was increased by BMDN incubated on HMD films relative to MM films and MAA films, indicating that surface chemistry plays a role in neutrophil activation. This response is particularly interesting, considering there is only a modest number of amine groups on HMD films (58 ± 18 nmol cm −2 ). Furthermore, there was an increase in overall citH3 from 2 to 16 h from cells on all samples, and normalized citH3 ratios after 2 h show that -NH 2 functionalized surfaces induced a more immediate and potent response than PMA or LPS-treated cells. This implies that either the rate or the mechanism or the rate of histone citrullination is altered by the presence of amine groups or proteins adsorbed to the surface.
Elucidating the mechanism by which BMDN incubated with HMD have increased NETosis is simplified because no PMA is required to promote adhesion. PMA is a strong activator of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase via the Nox2 complex [33] to produce reactive oxygen species (ROS), activate peptidyl arginine deiminase 4 (PAD4) and release NETs [34][35][36]. Including an NADPHoxidase inhibitor such as diphenyleneiodonium may elucidate the dependency of histone citrullination by HMD films on Nox2, as is seen with PMA [33][34][35][36]. NETosis may also be NADPH oxidase-independent, as seen with neutrophil activation with calcium ionophores, GM-CSF or TNF-α [36]. Alternatively, the binding of different surface receptors may be responsible for the increased rate and quantity of citH3. For example, PMA is associated with the activation of complement receptor 1 (CR1) whereas LPS with tolllike receptor 4 (TLR4) [37,38]. Activation of TLR4 promotes non-lytic forms of NETosis, where NETs are released through vesicles and the cell can continue to function (phagocytose, etc.). Interestingly, the viability of BMDN did not change between BMDN incubated with the different coatings, possibly indicating non-lytic NETosis is occurring. However, while cells on HMD films had increased NETosis, they had a modest release of cytokines from BMDN relative to LPS, indicating that other pathways may also be involved.

Proteins released by BMDN on films indicates the films may be low TLR4 agonists
The release of a select panel of cytokines from BMDN incubated with the copolymers were measured to elucidate additional mechanisms by which the copolymers modulate neutrophil cell behavior. No large trends were observed between the copolymers, but observations help elucidate the effects of the copolymers in general. IL-1β, MCP-1, and IL-10 were only detected in the medium from BMDN incubated with PMA or LPS indicating that BMDN require strong stimuli to produce these cytokines in vitro and are likely upregulated strongly through TLR4 signaling, as LPS is a TLR4 agonist [37]. MIP-2 and MIP-1α release were observed with BMDN incubated on all three types of films (in the absence of PMA or LPS). Both MIP-2 and MIP1-α are regulated to some extent through TLR4, as demonstrated with notably increased expression from LPS-treated cells (control groups), but the low expression of these proteins in cells incubated on coatings demonstrates that the materials are modest TLR4 agonists and NETosis may proceed through non-lytic NETosis. Whether another pathway is contributing to MIP-2 and MIP-1α production remains to be elucidated. The presence of MIP-1α and MIP-2 at early time-points in all treatment groups is consistent with the literature regarding their presence in and release by neutrophils. MIP-2 is constitutively expressed in a subset of bone marrow neutrophils for immediate release upon activation, while MIP-1α is not [39,40]. MIP-2 then, may be released during neutrophil degranulation, while MIP-1α may be upregulated by inflammatory stimuli.

NETs and materials
Interesting is the observed robustness with NETosis modulated similarly in both HL60 and BMDN cells over different timeframes, with MAA films (carboxyl groups) resulting in lower adhesion and NETs and HMD films (amine groups) generally higher in figures 3 and 5. How the cell interacts with the proteins that adsorb to these materials is critical to the observed responses. Interestingly, different materials with similar functional groups were observed to similarly modulate HL60 cells. For example, a previous study incubating HL60 with PMMA similarly show an increase in cell adhesion and the accumulation of NETs at the surface of aminated PMMA in comparison to HL60 incubated with PMMA or carboxylated PMMA [27]. In general, hydrophobic materials result in the adsorption of proteins with higher amounts of denaturation than hydrophilic materials, and charged groups can modulate which proteins interact at the surface, with negative groups attracting positively charged or positively charged side-groups of proteins (and vice versa for positively charged surfaces).
Importantly, Clarke et al isolated NETs from HL60 cells and pretreated culture wells with them. Subsequent incubation of macrophage-like THP1 cells, with and without additional activation with PMA, adhered to the NET-treated well plates with reduced viability compared to no NETS, as measured with alamarBlue and immunofluorescence (live/dead assays) [27]. NETs at the surface of polystyrene modulated HL60 cells. It is expected that the local microenvironment surrounding materials would introduce similar changes potentially resulting in different downstream cell responses after materials of different chemistries are implanted.
Importantly, this research focused on 2D scenarios relevant to intraocular lenses. After the removal of a cataractous lens, a prosthetic intraocular lens is inserted to support eyesight. During this surgery, immune cells will interact within a 2D environment within the capsular sac that holds the lens. This would also be similar to vascular implants, stents, sensors, and some other non-degradable implanted medical devices. An important next step is to evaluate similar studies in a 3D environment to recapitulate more complex systems.

Conclusions
The behavior of both HL60 and BMDN vary significantly when incubated on copolymer films of different chemistry clearly indicating the importance of neutrophil-material interactions on the microenvironment surrounding implants. MAA and HMD copolymer films had varying effects on both HL60 and BMDN cells. In general, MAA was associated with indicators of low inflammation and HMD high inflammation, including NETosis.
The composition of the copolymers modulated the amount of NETs that accumulated at the surface in both scenarios that were explored; HL60 with PMA and freshly isolated BMDN. There was an increase in NET accumulation in cells incubated with HMD films, which had a modest number of amine groups at the surface. The different chemistry of the copolymers varied the number of HL60 cells that adhered and therefore the overall amount of NETosis. In the absence of PMA, BMDN appear to possibly undergo non-lytic NETosis as their viability is retained over 16 h while NETosis increased and because TLR-4 appears to not be stimulated, as shown with low levels if MIP-2 and MIP1-α in comparison to LPS controls. With and without PMA, copolymer composition is modulating overall amounts of NETs either through differences in adhesion (HL60 with PMA) or differences in rates of NETosis (BMDN). The increased presence of NETs will modulate other immune cells including macrophages, as demonstrated in vitro with THP1 cells with NETs causing adhesion in the absence of other stimulators/differentiators (i.e. PMA) [27].
In addition to the overall increase in NETs when cells were incubated with copolymers with amine groups (HMD), HL60 cells on HMD films also had increased expression of TNF-α and ICAM-1 genes indicating they were activated towards a more inflammatory phenotype. Taken together, MAA appears to have a low inflammatory effect, which may result in reduced levels of M1 macrophages and increased M2 macrophages in vivo, explaining the increased vascularization surrounding MAA copolymer implants in past studies. When BMDN were incubated on the different films, chemotactic cytokines MIP-1α and MIP-2 were secreted at low levels in the absence of added stimuli, while the production of pro-inflammatory factors was not induced by any of the functionalized films. These pro-inflammatory cytokines were increased by LPS but not the films, indicating they are somewhat regulated via TLR4 and that the materials are modest TLR4 agonists. These observations give some insight into the level of activation caused by these coatings, though the mechanisms of which require further studies.
Importantly, this research demonstrates the importance of understanding the roles neutrophils play when a host reacts to polymer implants. While HL60 cells and BMDN demonstrated consistent, inflammatory behaviors in response to the aminefunctionalized copolymer films (HMD). Future investigations on detailed timecourse and in-vivo studies will further determine whether this response is helpful or harmful in an implanted biomaterial context.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.