Binding of myeloperoxidase to the extracellular matrix of smooth muscle cells and subsequent matrix modification

The extracellular matrix (ECM) of tissues is susceptible to modification by inflammation-associated oxidants. Considerable data support a role for hypochlorous acid (HOCl), generated by the leukocyte-derived heme-protein myeloperoxidase (MPO) in these changes. HOCl can modify isolated ECM proteins and cell-derived matrix, with this resulting in decreased cell adhesion, modulated proliferation and gene expression, and phenotypic changes. Whether this arises from free HOCl, or via site-specific reactions is unresolved. Here we examine the mechanisms of MPO-mediated changes to human coronary smooth muscle cell ECM. MPO is shown to co-localize with matrix fibronectin as detected by confocal microscopy, and bound active MPO can initiate ECM modification, as detected by decreased antibody recognition of fibronectin, versican and type IV collagen, and formation of protein carbonyls and HOCl-mediated damage. These changes are recapitulated by a glucose/glucose oxidase/MPO system where low continuous fluxes of H 2 o 2 are generated. HOCl-induced modifications enhance Mpo binding to ecM proteins as detected by eLiSA and Mpo activity measurements. these data demonstrate that MPO-generated HOCl induces ECM modification by interacting with ECM proteins in a site-specific manner, and generates alterations that increase MPO adhesion. This is proposed to give rise to an increasing cycle of alterations that contribute to tissue damage. The 23 Considerable evidence supports a role for MPO in the pathogenesis of atherosclerosis. MPO mRNA, pro- tein and enzymatic activity are present in lesions 24 , elevated levels of HOCl-damaged proteins are present 18,25 , and increased concentrations of the HOCl-generated biomarkers, 3-chlorotyrosine and 5-chlorouracil have been reported 17,21 . There is a strong association between the levels of HOCl-damaged proteins and the severity of disease (as judged by intimal-medial thickness) 18,19 , and the plasma levels of MPO have been shown, in multiple epidemiological and clinical studies, to be both diagnostic and prognostic of poor cardiovascular (and other disease) outcomes MPO is released from intracellular storage granules of activated neutrophils, monocytes and some tissue mac- rophages, and shows a high affinity for both ECM proteins and negatively-charged glycosaminoglycan (GAG) chains due to the high abundance of (positively-charged) Lys and Arg residues on the protein surface . These observations suggest that the HOCl formation by MPO/H 2 O 2 /Cl − may occur in a localized and site-specific man- ner. This is consistent with binding of (highly-cationic) MPO to ECM materials such as GAG chains 30,32 including those of the glycocalyx 33 , proteoglycans such as perlecan 31 , and lipoproteins such as high- and low-density lipoproteins (HDL and LDL, respectively) 17,22,34 . MPO protein has been reported to be particularly localized to the shoulder regions of lesions, which are especially prone to rupture and thrombus formation 35,36 , and high numbers of MPO-expressing cells have been reported at such sites out to some non-specific to used a matrix proteins, including fibronectin, type IV collagen and versican, appear as fibre-like structures, indicating that these ECM proteins interact with each other to form a 3-dimensional scaffold to which the HCASMCs adhere and proliferate. We have previously prepared decellularized HCASMC-ECM using NH 4 OH 43 , which allows efficient removal of adherent cells, but this protocol has been shown here to result in loss and disruption of the ECM materials and structure. However, DOC, which has been successfully used to decellularize other tissues 42 , allows the harvesting of matrix components laid down by HCASMC, which appear to be structurally-intact. This intact decellularized HCASMC-ECM provides a better model to investigate matrix-cell and matrix-MPO interactions in vitro . The current data indicate that MPO binds to matrix proteins, as evidenced by the colocalization of fibronectin and MPO detected via double immunofluorescence staining, confirming the data of Kubala et al . 30 . This suggests that MPO-induced damage to matrix proteins may be dependent on the specificity of binding to matrix proteins or sites. This is supported by the immunofluorescence staining, which shows that the HOCl-oxidized epitopes were primarily detected in the same regions as MPO binding to the ECM materials. The mechanism(s) that underlie the increased MPO binding to HOCl-modified ECM have not been determined in detail, but the results obtained, and literature data provide a potential explanation. It is known that HOCl preferentially targets Cys, Met, cystine, His, α -amino, Trp, Lys and Tyr residues, in order of decreasing reactivity 51,52 . The resulting products have been elucidated 53 . For some residues, reaction results in conversion of a neutral species to a negatively-charged product (e.g. conversion of Cys and cystine to oxy acids, oxidation of His to give the elevated levels of oxidized materials in advanced human atherosclerotic lesions compared to healthy tissues, and an accumulation with age and disease 7 . The low turnover rate and poor antioxidant activity of the ECM 5 would also be expected to favor such accumulation. our at in vitro level, by drives increasing modification ECM MPO appears to have a high affinity for at least some matrix proteins, with the matrix-bound MPO able to catalyse HOCl generation, and oxidative modifications to the ECM; this appears to be localized at, or near, sites of MPO-binding. These modifications then enhance further binding of MPO and damage, with this accumulating in a time- and HOCl flux-dependent manner. These data increase our under-standing of the mechanisms that underlie MPO-matrix interactions and MPO-induced matrix oxidation, and provide potential insights into the prevention or amelioration of inflammation-associated diseases, including atherosclerosis. casein peroxidase (HRP)-conjugated sheep anti-mouse or anti-rabbit IgG secondary antibodies (1:1000) followed by two more washes with PBST. The were subsequently developed with 50 mM sodium citrate (pH 4.6) containing 2 mM ABTS [2,2 ′ -azinobis-(3-ethylbenzothiazoline-6-sulfonic acid] and 3 mM H 2 O 2 for 30 min. For the experiments using MPO, the plates were incubated with alkaline phosphatase-conjugated rabbit anti-mouse (Abcam) or goat anti-rabbit IgG (Abcam) secondary antibodies at 21 °C for 1 h, washed twice with PBST then developed with p -nitrophenylphosphate solution (Merck) at 21 °C for 20 min. The optical absorbance of the resulting chromophores was determined using a Spectra Max i3x microplate reader (Molecular Devices). The alkaline phosphatase system was employed for MPO experiments, rather than the HRP/ABTS/H 2 O 2 system, as both MPO and HRP react with H 2 O 2 , which causes erroneous absorbance data. ELISA readings. μ NuPAGE and L of 10x NuPAGE reducing denatured for 5 min at 90 as per the CA, USA). Seventeen μ L of were loaded per lane, on to 3–8% Tris-acetate gels and electrophoresed for 65 min at 160 V. The separated were then transferred onto a PVDF membrane using an iBlot 2 system (Invitrogen, CA, USA) for 7 min at 25 V. The membrane was then blocked using 1% (w/v) BSA/TBST (0.1% Tween-20) blocking buffer for 1 h, at 21 °C, then incubated for 1 h with anti-FN pAb (1:2000 dilution; ab2413; Abcam) on a shaker at 21 °C. The membrane was washed twice with TBST prior to incubating with donkey anti-rabbit IgG HRP conjugated (1:5000 dilution) secondary antibody for 1 h at 21 °C followed by 4 × 15 min TBST washes. The membrane was rinsed twice with TBS and subsequently imaged using a Sapphire Azure biomolecular imager (Azure Biosystems, CA, USA) to examine the presence of MPO-FN protein complexes immunoprecipitated using anti-MPO mAb (2C7).


Results
Decellularization of primary HCASMC cultures using sodium deoxycholate and immunolocalization of ECM components. Previous studies have commonly used NH 4 OH to remove cells from cultures to give isolated ECM preparations 41 . This method was used in initial experiments, but such treatment results in a loss of some ECM materials as judged by a decreased detection of fibronectin (determined by immunofluorescence using an anti-fibronectin pAb) in NH 4 OH-treated samples compared to controls (Supplementary Fig. 1).
Recent studies have employed sodium deoxycholate (DOC, 1%) to decellularize murine tissues, with minimal reported damage to native ECM components, allowing high-resolution imaging and proteomic analyses 42 . Consequently, this treatment was examined as a potential method for obtaining native ECM from cell cultures. HCASMCs were cultured on chamber culture slides for 1 week to produce HCASMC-ECM. The slides were then examined either as intact structures, or incubated with 1% DOC to remove cells. Double immunofluorescence staining using primary antibodies against cell-derived fibronectin-extra domain A (EDA) (mAb 3E2) and type IV collagen (pAb) was then employed to examine both the presence and levels of ECM materials. The intact HCASMC-containing wells stained positive for type IV collagen both intracellularly and extracellularly (Fig. 1,   Figure 1. Immunolocalization of native and decellularized ECM components synthesized by primary HCASMCs. HCASMCs were cultured for 1 week to allow the synthesis and assembly of native ECM, then either left untreated (control) or treated with 1% sodium deoxycholate (DOC; bottom panel) to lyse and solubilize the attached cells. Matrix proteins were detected via double immunofluorescence staining using primary antibodies against type IV collagen (pAb) and the cell-derived fibronectin EDA region (FN-3E2). Alexa Fluor 488-conjugated anti-rabbit secondary antibody (green) and an Alexa Fluor 594-conjugated antimouse secondary antibody (red) were used to visualize matrix protein-primary antibody complexes, with nuclei counterstained using DAPI (blue). Representative images from n = 3 independent experiments are shown. Scale bars: 50 μm.
top row; green fluorescence), cell-derived fibronectin localized extracellularly as fibre-like structures (Fig. 1, top row; red fluorescence), and cell nuclei (Fig. 1, top row; blue fluorescence from DAPI). Control samples showed no fluorescence indicating a high specificity of these antibodies ( Supplementary Fig. 2). Examination of DOC-treated wells showed the presence of extracellular type IV collagen as fibre-like structures and cell-derived fibronectin but no blue fluorescence from DAPI (Fig. 1, bottom row). These data indicate that DOC treatment is an efficient and gentle decellularizing agent for monolayers of HCASMCs in vitro. Versican, another key matrix component, was also detected in the DOC treated cultures by double immunofluorescence staining using an anti-fibronectin pAb and anti-versican mAb (12C5, which recognizes the hyaluronate-binding G1 domain) ( Supplementary Fig. 3).
MPO binds to matrix proteins in HCASMC-ECM. As MPO binds with high affinity to some biological materials (see Introduction) we hypothesized that MPO may bind to and colocalize with HCASMC-ECM materials and induce selective ECM damage. Consequently, we investigated the affinity of MPO for matrix proteins present in decellularized HCASMC-ECM. HCASMC-ECM samples were prepared using DOC, blocked with BSA, then incubated with MPO (20 nM) for 30 min at 37 °C. After washing to remove non-adherent materials, the presence of fibronectin and MPO was examined via double immunofluorescence using an anti-fibronectin pAb and an anti-MPO mAb (2C7). Fibronectin was detected within the decellularized HCASMC-ECM as fibre-like structures, with the fluorescence from MPO strongly co-localized with this material, as evidenced by the merged images ( Fig. 2; see also Supplementary Fig. 2).
Further evidence for a direct MPO-FN interaction was obtained by immunoprecipitation of an MPO-FN complex using an anti-MPO mAb (2C7). The material pulled-down by mAb 2C7 showed positive staining at the expected mass of FN monomers, after separation of the proteins on SDS-PAGE and probing with an anti-FN pAb in immunoblotting experiments (Fig. 3).

Modification of HCASMC-ECM proteins by MPO/H 2 o 2 /Cl − .
We have shown previously that MPO-derived HOCl modifies HCASMC-ECM preparations decellularized using NH 4 OH 43 . Here, we investigated the extent of HOCl-induced modification generated on (more intact) DOC-decellularized ECM by a bound MPO/H 2 O 2 /Cl − system, with alterations to the ECM detected by ELISA using primary antibodies that recognize HCASMC-ECM proteins. Native HCASMC-ECM was prepared using DOC (as above) and exposed to MPO (20 nM), Cl − (200 mM) and varying concentrations of H 2 O 2 (5-100 μM). This system gives near quantitative conversion of H 2 O 2 to HOCl 40,43 . Exposure of the decellularized HCASMC-ECM to the MPO system with 100 μM H 2 O 2 resulted in significant losses in epitope recognition by both the anti-fibronectin pAb (Fig. 4A) and anti-versican mAb (12C5), with ≥20 μM or higher H 2 O 2 (Fig. 4B). With the type IV collagen pAb, a significant increase in epitope recognition was detected with 10 μM H 2 O 2 (Fig. 4C), possibly as a result of increased exposure of cryptic epitopes as a result of modification to other ECM components. However, with 100 μM H 2 O 2 , a significant decrease in epitope recognition by this antibody was apparent. Concurrently, a significant increase in the presence of HOCl-generated epitopes was detected, as evidenced by recognition by mAb 2D10G9 (which recognises HOCl-modified proteins), with this being dose-dependent, and significant with ≥20 μM H 2 O 2 (Fig. 4D).
Increased recognition of ECM materials by the 2D10G9 mAb upon exposure to MPO/H 2 O 2 /Cl − system was also examined using immunofluorescence staining. The fibronectin pAb, and mAb 2D10G9 were used to detect fibronectin and HOCl-generated epitopes in the decellularized HCASMC-ECM exposed to the MPO system, in Increased affinity of MPO for oxidized HCASMC-ECM. The above data are consistent with binding of MPO to HCASMC-ECM, with the MPO remaining catalytically-active and able to induce modifications to the ECM. However, it is unclear whether oxidized ECM also has this capacity. We therefore investigated the affinity of MPO for native, or HOCl-exposed HCASMC-ECM, via ELISA. Concurrently, the activity of matrix-bound MPO was examined in a quantitative manner using the TMB assay which relies on the conversion of taurine to taurine chloramine by HOCl, and subsequent assay of this chloramine 44 .
Reagent HOCl modified the decellularized HCASMC-ECM, with significantly increased mAb 2D10G9 recognition detected via ELISA, with this increasing in a dose-dependent manner; this increase was significant at ≥20 μM HOCl (Fig. 7A, black bars). Interestingly, an increased absorbance at 405 nm, arising from recognition by the MPO mAb, was also observed for the samples treated with ≥20 μM HOCl, and in a dose-dependent manner, consistent with MPO binding to the modified ECM, and having a higher affinity for the modified, compared to native, HCASMC-ECM (Fig. 7A, white bars). To confirm the increased binding of MPO to the oxidized ECM, the enzymatic activity of the matrix-bound MPO was quantified using the TMB assay, with a fixed concentration of H 2 O 2 (50 μM) added to the MPO adherent on the ECM in the presence of taurine to trap the resulting HOCl. A significantly increased absorbance at 645 nm was observed for MPO bound to the ECM treated with 20 μM HOCl, when compared to the MPO bound to native ECM, consistent with an increased concentration of active MPO bound to the oxidized, compared to native, HCASMC-ECM (Fig. 7B). The yield of taurine chloramine quantified using TMB reached an apparent plateau when the initial ECM was treated with higher concentrations of HOCl, suggesting that the H 2 O 2 added to the system can be depleted by the activity of the bound MPO.
The potential role of ECM oxidation products in the increased affinity of MPO for the oxidized matrix, was examined by quantifying the level of carbonyls formed on the HCASMC-ECM treated with different concentrations of reagent HOCl (10-200 μM). The carbonyl yields were measured by derivatization with www.nature.com/scientificreports www.nature.com/scientificreports/ 2,4-dinitrophenylhydrazine (DNPH) and detection of the resulting hydrazone products using an anti-DNPH antibody (Fig. 8). Treatment with HOCl resulted in a marked increase in HCASMC-ECM-derived carbonyl levels even at the lowest oxidant levels tested (10 μM), though the levels of these products did not increase in a linear manner with increasing HOCl concentrations, consistent with further reaction of these (reactive) products. Such behavior is consistent with previous reports 45,46 .

MPO-induced HCASMC-ECM oxidation is dependent on the flux of H 2 o 2 . The data presented
above indicate that a MPO-mediated oxidation system can modify decellularized HCASMC-ECM, as detected by ELISA or immunofluorescence staining with mAb 2D10G9. This occurs in a concentration-dependent manner with increasing concentrations of H 2 O 2 ≥ 20 μM. Decreased ECM epitope recognition was also detected by antibodies raised against fibronectin, type IV collagen and versican mAb (12C5). However, the H 2 O 2 used in these experiments was added as a bolus dose rather than continuously, with the latter more likely to reflect the biological situation. The overall concentrations are however close to those reported as a steady-state level (cf. reports of H 2 O 2 concentrations in plasma of 1-10 μM 47 ). As a consequence, we examined modification of the HCASMC-ECM by associated MPO where the H 2 O 2 was generated in situ using a glucose/glucose oxidase (GO) system, that generates H 2 O 2 over a long time period, with potential damage to the HCASMC-ECM detected using immunofluorescence staining and ELISA, as described above.
The GO/glucose/MPO/Cl − system generated oxidative damage to HCASMC-ECM in a GO concentrationdependent manner as determined by ELISA, with GO levels ≥100 mU mL −1 giving a statistically-significant increase in damage (Fig. 9A, black bars) relative to the corresponding controls that did not contain MPO (Fig. 9A, white bars). Exposure of HCASMC-ECM to a fixed concentration of the oxidant system (400 mU mL −1 GO, 5 mM glucose, 20 nM MPO, 200 mM Cl − ) for increasing incubation times gave a time-dependent increase in HCASMC-ECM damage, as evidenced by increasing ELISA signals arising from mAb 2D10G9, for exposure times ≥2 h (Fig. 9B, black bars). In the absence of MPO this increase in absorbance was not detected (Fig. 9B, white bars). Immunofluorescence staining using the pAb against fibronectin, and mAb 2D10G9 against HOCl-mediated damage, supported this conclusion. Thus, significant green fluorescence arising from the presence of fibronectin was observed in each group (Fig. 9C, left column), whereas red fluorescence arising from the presence of HOCl-modified epitopes was absent in the controls, but increased in extent with longer incubation times (Fig. 9C, middle column), such that a marked increase in fluorescence was detected after 20 h. The fluorescence from these two antibodies was strongly colocalized (Fig. 9C, right column). ). Data are presented as means ± SEM from n = 3 independent experiments, and are expressed as a % of the control data (fibronectin, versican and type IV collagen), or the % of the maximal ELISA signal detected with mAb 2D10G9, with statistical analysis carried out using one-way ANOVA with Tukey's post-hoc tests. * indicates statistical significance from the control (0 μM added H 2 O 2 ) at the p < 0.05 level, and ** the p < 0.01 level.

Discussion
It is well established that elevated levels of oxidized DNA, RNA, proteins and lipids are present in atherosclerotic lesions, but the nature of the oxidants that generate these changes and the specific pathways involved remain to be fully elucidated 48 . We have reported that exposure of HCASMC-ECM to either reagent or MPO-derived HOCl can give rise to oxidative modifications of the HCASMC-ECM and its components, and that these modified materials affect the behavior of HCASMCs, including loss of adhesion, accelerated proliferation and modulated gene expression 43 . While MPO is vital for innate immunity 49 , aberrant expression of MPO and subsequent oxidant production have been implicated in many inflammatory diseases, including atherosclerosis (ref. 36 , see also Introduction).
Here we have examined, at an in vitro level, how MPO modifies matrix proteins. MPO is shown to bind to ECM proteins, and to retain its catalytic activity, resulting in the generation of HOCl in the presence of Cl − and H 2 O 2 , resulting in modifications to HCASMC-ECM materials. These modifications are localized to the vicinity of MPO, though the exact targets and specific binding sites require confirmation. These changes were recapitulated by exposure of HCASMC-ECM to a GO/glucose/MPO/Cl − system, as evidenced by increased antibody recognition of HOCl-mediated damage (via ELISA and immunofluorescence assay) by 2D10G9 mAb. This occurred in a time-dependent manner, with increasing levels detected with higher fluxes of H 2 O 2 as judged by a dependence on the GO concentration. A previous study has also reported direct modification of isolated fibronectin by H 2 O 2 , but this was dependent on trace metal ions which generate hydroxyl radicals (HO . ) and required high levels of H 2 O 2 50 . Whether such HO . mediated damage is of importance in vivo remains to be established, as redox active metal ions would be expected to be tightly-bound to endogenous metal-ion binding proteins.
Interestingly, the extent of MPO-binding appears to be dependent on the level of HOCl-induced modifications, with greater binding apparent (when compared to the non-treated HCASMC-ECM) after pre-treatment of the ECM with HOCl, as indicated by the ELISA data where increased epitope recognition of MPO by anti-MPO mAb (2C7) was observed with HOCl-pre-treated HCASMC-ECM. Our data are therefore consistent with a new paradigm in which MPO participates in the development and progression of atherosclerosis through the generation of damage to matrix proteins, which then bind increasing concentrations of MPO as a result of an enhanced affinity of the enzyme for modified ECM. This bound MPO has been shown to retain a high level of enzymatic activity. This would result in a vicious cycle of increasing ECM alteration by MPO-derived oxidants, greater MPO binding to the damaged materials, and increased site-dependent oxidant formation. This has been associated with aberrant cell behavior including decreased cell adhesion, modulated proliferation, phenotypic switching of smooth muscle cells, altered gene expression, release of pro-inflammatory species, and matrix remodeling 43 .
The cultured primary HCASMC employed in this study produce abundant ECM materials in vitro, which in turn support the adhesion and survival of adherent cells. Immunofluorescence staining shows that different www.nature.com/scientificreports www.nature.com/scientificreports/ matrix proteins, including fibronectin, type IV collagen and versican, appear as fibre-like structures, indicating that these ECM proteins interact with each other to form a 3-dimensional scaffold to which the HCASMCs adhere and proliferate. We have previously prepared decellularized HCASMC-ECM using NH 4 OH 43 , which allows efficient removal of adherent cells, but this protocol has been shown here to result in loss and disruption of the ECM materials and structure. However, DOC, which has been successfully used to decellularize other tissues 42 , allows the harvesting of matrix components laid down by HCASMC, which appear to be structurally-intact. This intact decellularized HCASMC-ECM provides a better model to investigate matrix-cell and matrix-MPO interactions in vitro.
The current data indicate that MPO binds to matrix proteins, as evidenced by the colocalization of fibronectin and MPO detected via double immunofluorescence staining, confirming the data of Kubala et al. 30 . This suggests that MPO-induced damage to matrix proteins may be dependent on the specificity of binding to matrix proteins or sites. This is supported by the immunofluorescence staining, which shows that the HOCl-oxidized epitopes were primarily detected in the same regions as MPO binding to the ECM materials.
The mechanism(s) that underlie the increased MPO binding to HOCl-modified ECM have not been determined in detail, but the results obtained, and literature data provide a potential explanation. It is known that HOCl preferentially targets Cys, Met, cystine, His, α-amino, Trp, Lys and Tyr residues, in order of decreasing reactivity 51,52 . The resulting products have been elucidated 53 . For some residues, reaction results in conversion of a neutral species to a negatively-charged product (e.g. conversion of Cys and cystine to oxy acids, oxidation of His to give the carboxylic-acid containing residue Asp) 53 . Reaction at other residues (e.g. at His, α-amino and Lys) converts positively-charged (protonated nitrogen) groups to neutral species (e.g. Asn from His, carbonyls and nitriles from α-amino and Lys species) 53 . This increased negative charge, and loss of positive charge, may result in increased binding of (highly-positively charged) MPO via electrostatic interactions. Covalent bonding may also contribute, as elevated levels of carbonyls (a known product of amine (Lys) oxidation by HOCl 45 ), were detected on HOCl-treated HCASMC-ECM. Carbonyls can undergo Schiff base reactions with Lys and Arg residues, that are present in abundance in the MPO structure (55 Arg and 14 Lys in the mature protein, ~12% abundance: UniProt entry: P05164). Such reactions have been reported previously to contribute to inter-molecular protein cross-linking (e.g. 54 ).
The MPO bound to oxidized HCASMC-ECM is enzymatically active, as shown by the detection and quantification of taurine-derived chloramines. Whether this bound material is more or less active than free MPO, is unclear and difficult to assess due to difficulties in quantifying the amount of MPO bound to the ECM. Recent mass spectrometry data suggests that some ECM proteins are less extensively modified by the MPO system relative to the same concentration of reagent HOCl (e.g. fibronectin 40 ), and others such as laminins more modified 55 , www.nature.com/scientificreports www.nature.com/scientificreports/ as judged by both the number, and level of modifications. These differences have been rationalized in terms of MPO binding to these species and possible alterations in activity. In HCASMC-EMC both fibronectin and laminins are present, therefore it is not possible to decipher the which effect may be more important. Although the nature of the specific MPO binding sites remain to be elucidated, it is nonetheless clear that higher levels of MPO are bound to the ECM after prior matrix oxidation.  www.nature.com/scientificreports www.nature.com/scientificreports/ These data imply that in vivo, increased inflammation, enhanced neutrophil/monocyte/macrophage activation and release of MPO, may result in increasing levels of MPO being associated with the ECM, which could promote further matrix modification. This positive feedback loop could potentially increase the rate of disease development and result in the accumulation of oxidized ECM proteins. This is consistent with the detection of The extent of ECM damage was then quantified using the mAb 2D10G9 in an ELISA assay. (C) HCASMC-ECM was exposed to a GO-MPO system (400 mU mL −1 glucose oxidase, 5 mM glucose, 20 nM MPO) for 0-20 h, with parent ECM fibronectin and oxidative modifications to the HCASMC-ECM detected using immunofluorescence. Fibronectin and HOCl-induced modifications were detected using fibronectin pAb (left column) and mAb 2D10G9 (middle column), the corresponding images are merged in the right column. An Alexa Fluor 488-conjugated anti-rabbit secondary antibody (green) and an Alexa Fluor 594-conjugated anti-mouse secondary antibody (red) were used to visualize matrix protein-primary antibody complexes. In (A and B), absorbance data at 405 nm are presented as means ± SEM from n = 3 independent experiments. * Indicates statistical significance from the control (0 μM added GO or time zero data) at the p < 0.05 level, and ** at the p < 0.01 level by ANOVA with Tukey's post-hoc tests. In (C), representative images from n = 3 independent experiments are shown. Scale bars: 30 μm. www.nature.com/scientificreports www.nature.com/scientificreports/ elevated levels of oxidized materials in advanced human atherosclerotic lesions compared to healthy tissues, and an accumulation with age and disease 7 . The low turnover rate and poor antioxidant activity of the ECM 5 would also be expected to favor such accumulation.
Overall our data demonstrates, at an in vitro level, a mechanism by which MPO induces and drives increasing modification of the ECM (Fig. 10). MPO appears to have a high affinity for at least some matrix proteins, with the matrix-bound MPO able to catalyse HOCl generation, and oxidative modifications to the ECM; this appears to be localized at, or near, sites of MPO-binding. These modifications then enhance further binding of MPO and damage, with this accumulating in a time-and HOCl flux-dependent manner. These data increase our understanding of the mechanisms that underlie MPO-matrix interactions and MPO-induced matrix oxidation, and provide potential insights into the prevention or amelioration of inflammation-associated diseases, including atherosclerosis.

Materials and Methods
Reagents. All  Immunofluorescence staining of ECM materials and modification products. HCASMCs (passage 4) were seeded onto eight-well chamber glass culture slides (Becton Dickinson) at an initial density of 2 × 10 4 cells per well and cultured for 1 week to allow ECM synthesis. HCASMCs on the slides were treated twice with or without 1% (w/v) sodium deoxycholate (DOC) for 10 min at 21 °C with gentle rocking, followed by 5 washes with PBS. The slides were then fixed with 4% (v/v) formaldehyde at 37 °C for 15 min, then permeabilized using 0.5% (v/v) Triton X-100 in PBS on ice for 5 min. The slides were blocked with 1% (v/v) BSA in PBS at 21 °C for 1 h and then incubated with primary antibodies at 4 °C overnight. The primary antibodies were used at the following dilutions: anti-cell derived fibronectin EDA epitope (Sigma, 3E2 mAb; 1:1000), anti-fibronectin pAb (Abcam, ab2413; 1:1000), anti-type IV collagen (Abcam, ab6586; 1:1000), anti-MPO (Abcam, 2C7 mAb; 1:1000), anti-versican G1 domain (Developmental Studies Hybridoma Bank, Univ. of Iowa, USA, 12C5 mAb; 1:1000) and anti-HOCl generated epitopes (kindly provided by Prof. Ernst Malle, Medical University of Graz, Austria; 2D10G9 mAb, 1:200). The slides were rinsed twice with PBS then incubated with anti-mouse IgG conjugated with Alexa Fluor 488 antibody and anti-rabbit IgG conjugated with Alexa Fluor 594 antibody (both Molecular Probes; 1:1000 in 1% (v/v) BSA in PBS) at 21 °C for 1 h. The slides were rinsed again with PBS and then counterstained with 1 μg mL −1 DAPI (4′,6-diamidino-2-phenylindole, Molecular Probes) in PBS in the dark at 4 °C for 10 min, followed by three rinses with PBS, before being air dried and the cover slips added, with imaging using a fluorescence microscopy (Olympus, Japan) equipped with cellSense Entry v1.5 software. www.nature.com/scientificreports www.nature.com/scientificreports/ ELISA of ECM materials and modification products. HCASMCs were cultured in 96-well plates at an initial density of 1 × 10 4 cells per well in 200 μL of growth medium (as above) for 1 week to allow synthesis of native HCASMC-ECM. The plates were rinsed twice with PBS followed by the incubation with 1% sodium deoxycholate (DOC) twice for 10 min at 21 °C. The plates were rinsed 5 times with PBS and the prepared native HCASMC-ECM was then treated with 50 μL of HOCl (0-100 μM), or an MPO-H 2 O 2 -Cl − system (20 nM MPO, 0-100 μM H 2 O 2 , 200 mM Cl − ) at 37 °C for 2 h. Alternatively, a glucose oxidase (GO)-MPO system was used to treat HCASMC-ECM. Briefly, the HCASMC-ECM in 96-well plates were prepared as described above and then incubated with either 50 μL of GO-MPO system (0-400 mU mL −1 GO and 20 nM MPO) for 20 h at 37 °C, or with 50 μL of GO-MPO system (400 mU mL −1 GO, 20 nM MPO) for 0-20 h at 37 °C. The plates were rinsed twice with PBS followed by blocking with 0.1% (w/v) casein in PBS. Plates were then incubated with the following primary antibodies, at the indicated dilutions, at 4 °C overnight: anti-fibronectin pAb (1:1000), anti-versican G1 domain mAb (12C5; 1:1000), mouse anti-MPO mAb (2C7, 1:1000) and anti-HOCl generated epitope mAb (2D10G9; 1:200). For the experiments using HOCl, the plates were rinsed twice with PBST (0.1% Tween-20 in PBS) and incubated with either horseradish peroxidase (HRP)-conjugated sheep anti-mouse or anti-rabbit IgG secondary antibodies (1:1000) followed by two more washes with PBST. The plates were subsequently developed with 50 mM sodium citrate (pH 4.6) containing 2 mM ABTS [2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid] and 3 mM H 2 O 2 for 30 min. For the experiments using MPO, the plates were incubated with alkaline phosphatase-conjugated rabbit anti-mouse (Abcam) or goat anti-rabbit IgG (Abcam) secondary antibodies at 21 °C for 1 h, washed twice with PBST then developed with p-nitrophenylphosphate solution (Merck) at 21 °C for 20 min. The optical absorbance of the resulting chromophores was determined using a Spectra Max i3x microplate reader (Molecular Devices). The alkaline phosphatase system was employed for MPO experiments, rather than the HRP/ABTS/H 2 O 2 system, as both MPO and HRP react with H 2 O 2 , which causes erroneous absorbance data. Values for 1% BSA/PBS only wells were used as background to normalize the ELISA readings.

Immunoprecipitation (IP) of MPO-FN protein complexes. MPO (20 nM) was incubated with human
plasma-derived FN (20 nM) in PBS for 30 min at 37 °C to generate possible MPO-FN complexes. Protein A/G magnetic beads (25 μL per IP sample; Pierce, IL, USA) were prepared as per the manufacturer's instructions, and incubated with 2 μg of anti-MPO mAb (2C7) for 1 h at 21 °C on a rotating mixer. The beads were then washed twice with the wash buffer provided by manufacturer, and in accordance with the provided instructions, prior to incubating with the MPO-FN mixture overnight at 4 °C on the rotating mixer. Negative controls using beads only with MPO/anti-MPO mAb (2C7) or MPO-FN complexes alone were also included. Next day, the beads were washed twice with wash buffer and once with MilliQ water prior to the addition of 20 μL elution buffer and 2 μL of neutralization buffer for 10 min at 21 °C. The eluted samples were transferred to clean tubes and 8.5 μL of 4x NuPAGE sample buffer, and 3.4 μL of 10x NuPAGE reducing agent, were added then heat denatured for 5 min at 90 °C as per the manufacturer's protocol (Invitrogen, CA, USA). Seventeen μL of sample were loaded per lane, on to 3-8% Tris-acetate gels and electrophoresed for 65 min at 160 V. The separated proteins were then transferred onto a PVDF membrane using an iBlot 2 system (Invitrogen, CA, USA) for 7 min at 25 V. The membrane was then blocked using 1% (w/v) BSA/TBST (0.1% Tween-20) blocking buffer for 1 h, at 21 °C, then incubated for 1 h with anti-FN pAb (1:2000 dilution; ab2413; Abcam) on a shaker at 21 °C. The membrane was washed twice with TBST prior to incubating with donkey anti-rabbit IgG HRP conjugated (1:5000 dilution) secondary antibody for 1 h at 21 °C followed by 4 × 15 min TBST washes. The membrane was rinsed twice with TBS and subsequently imaged using a Sapphire Azure biomolecular imager (Azure Biosystems, CA, USA) to examine the presence of MPO-FN protein complexes immunoprecipitated using anti-MPO mAb (2C7). Quantification of HOCl production by reaction with taurine and assay with 3,3′,5,5′tetramethylbenzidine (TMB). Decellularized HCASMC-ECM in 96-well plates was prepared as described above. The plates were incubated with 50 μL of HOCl (0-100 μM) at 37 °C for 2 h followed by two rinses with PBS. Plates were then blocked with 1% BSA/PBS at 21 °C for 1 h and then incubated with 20 nM of MPO at 37 °C for 30 min. Residual materials were then removed by washing twice with PBS, followed by incubation with 50 μM H 2 O 2 , 200 mM Cl − and 10 mM taurine at 37 °C for 2 h. The HOCl generated by matrix-bound MPO, and trapped with taurine as the corresponding chloramine, was subsequently quantified 44 using a developing reagent consisting of 20 mM TMB in dimethylformamide and 2 mM sodium iodide in sodium acetate buffer (0.44 M, pH 5.4). This was prepared immediately prior to addition to all samples, with 25 μL added to each well and incubated at 21 °C for 5 min. The absorbance at 645 nm was subsequently determined using a Spectra Max i3x microplate reader (Molecular Devices). carbonyl assay. The plates for carbonyl detection were prepared as described above for the ELISA experiments. The resulting native HCASMC-ECM was exposed to 50 μL of HOCl (0-200 μM) for 2 h at 37 °C and the plates were then rinsed twice with PBS. The carbonyl levels were quantified using an OxyBlot Protein Oxidation Detection Kit (S7150, Merck, USA) according to the manufacturer's instructions with minor modifications. Briefly, the ECM on plates were derivatized by incubating with 50 μL of 1 × DNPH for 15 min at 21 °C, then rinsed twice with PBS followed by blocking with 0.1% (w/v) casein in PBS. Plates were then incubated with rabbit anti-DNP antibody (1:500) at 4 °C overnight, rinsed twice with PBST (0.1% Tween-20 in PBS), incubated with horseradish peroxidase (HRP)-conjugated sheep anti-rabbit IgG secondary antibodies (1:1000) for 1 h at 21 °C, then washed twice with PBST. The plates were subsequently developed in 50 mM sodium citrate (pH 4.6) buffer containing 2 mM ABTS and 3 mM H 2 O 2 for 30 min, with the resulting absorbance at 405 nm measured using a Spectra Max i3x microplate reader (Molecular Devices).