DOI:10.1111/micc.12173

Original Article

Human Microvascular Pericyte Basement Membrane Remodeling Regulates Neutrophil Recruitment PARID SAVA, IAN O. COOK, RAJWANT S. MAHAL, AND ANJELICA L. GONZALEZ Department of Biomedical Engineering, Yale University, New Haven, Connecticut, USA Address for correspondence: Anjelica L. Gonzalez, Yale University, 55 Prospect Street, Malone Engineering Center 314, New Haven, CT 06520, USA. E-mail: [email protected] Received 25 April 2014; accepted 8 September 2014.

ABSTRACT Objective: Neutrophil extravasation at post-capillary venules, consisting of EC, PC, and the shared ECM, increases following fibrotic remodeling in the lung, liver, and skin. The role of fibrotic pericyte-derived ECM in regulating EC activation and neutrophil recruitment remains unexplored. Methods: To elucidate the role of human pericyte-derived ECM in EC activation, we characterized PC-derived ECM following transforming growth factor-b1, IL-1b, CCL2, or bleomycin activation, and examined surface adhesion molecule expression and neutrophil recruitment by EC cultured on PC-ECM. Results: Pro-inflammatory activation of PC-induced deposition of compositionally distinct ECM compared with non-activated control. Bleomycin activation induced fibronectin-rich and collagenpoor ECM remodeling by PC, facilitating increased neutrophil transendothelial migration when compared with non-activated pericyte ECM (49.9  3.4% versus 29.7  1.4%). Increases in fibronectin compared to collagen I, are largely responsible for ECMregulated neutrophil recruitment, as EC cultured on fibronectin

supported increased neutrophil transmigration compared to collagen I (51.6  6.2% versus 28.0  4.8%). We attribute this difference to increased expression of ICAM-1 and its redistribution to EC borders. Conclusions: This is the first demonstration of human pericyte sensitivity to inflammatory stimuli, inducing fibrotic matrix deposition that regulates EC adhesion molecule expression and neutrophil recruitment. WORDS: transmigration, fibrosis, fibronectin, ICAM-1, extracellular matrix

KEY

Abbreviations used: Bleo, bleomycin sulfate; BSA, bovine serum albumin; CCL2, chemokine (C-C motif) ligand 2; EC, endothelial cells; ECM, extracellular matrix; HUVEC, human umbilical vein endothelial cells; ICAM-1, intracellular adhesion molecule-1; IL-1b, interleukin-1b; MFI, mean fluorescence intensity; PBS, phosphatebuffered saline; PC, pericytes; PECAM-1, platelet endothelial cell adhesion molecule-1; PFA, paraformaldehyde; TGF-b1, transforming growth factor b1; TNF-a, tumor necrosis factor-a.

Please cite this paper as: Sava P , Cook IO, Mahal RS , Gonzalez AL. Human microvascular pericyte basement membrane remodeling regulates neutrophil recruitment. Microcirculation 22: 54–67, 2015.

INTRODUCTION Approximately, 45% of US mortalities are related to tissue fibrosis, therefore, a more complete understanding of the underlying mechanisms regulating fibrotic progression, including inflammation, is needed [48]. Early fibrosis coincides with a marked increase in TGF-b1, IL-1b, and CCL2, resulting in altered matrix deposition by myofibroblasts and formation of mechanically stiff ECM that impedes the function of the affected organ and may cause organ failure [46,50]. Fibrotic disease progression can be positively correlated with inflammation, as neutrophils in the extravascular space release cytokines, including TNF-a and IL-1b, neutrophil elastases, and reactive oxygen species, resulting in additional leukocyte recruitment, prolonged survival, and remodeling of the affected tissue [40,49,50]. Neutrophil transmigration into the

54

extravascular space occurs at the post-capillary venules, comprised of an EC lumen and circumferentially oriented PC embedded within a vascular basement membrane composed of laminins, collagen IV, and fibronectin [38]. Neutrophil recruitment is initiated by pro-inflammatory activation of EC, leading to EC upregulation of L-selectins and ICAM-1 [2,8,23,47]. Rolling neutrophils respond to EC-bound chemokines, including IL-8, facilitating EC selectin and adhesion molecule capture and guidance of neutrophils to regions permissive to transmigration across the EC-lined lumen of the vessel [25,29]. The role of ICAM-1 as an adhesion ligand and guidance molecule is evident in all cell layers of the microvessel; it was recently demonstrated that PC are incapable of supporting high levels of neutrophil adhesion and transmigration due to the relatively low levels of ICAM-1 and inefficient ICAM-1 localization, compared to IL-1b-activated EC [2,18].

ª 2014 John Wiley & Sons Ltd

PC ECM Regulates Neutrophil Recruitment

In response to pro-fibrotic activation, as during fibrosis of the lung, liver, heart, and skin, PC remodel the vascular basement membrane and extravascular matrix, demonstrating that fibrosis is indeed a microcirculatory disease [15]. Indicative of scar tissue formation, the vascular basement membrane and interstitial matrix become high in fibronectin and collagen I, altering EC function through integrin-specific signaling [46]. EC-presented integrin a5b1 and avb3 bind to the Arg-Gly-Asp (RGD) sequence of fibronectin in the provisional, remodeled ECM. This ligand–receptor interaction induces an angiogenic state that increases vascular cell survival and migration [3,42]. Furthermore, direct interactions between EC integrins and ECM proteins, particularly fibronectin, induce increased sensitivity to circulating cytokines and activation, suggesting that ECM modifications can be extremely influential in promoting inflammation [39]. Beyond direct regulation of EC activity, similarities between PC and myofibroblast matrix deposition in response to proinflammatory molecules point to an indirect role of PC as microvascular regulators of inflammation through matrix deposition. In fact, neutrophil transmigration across EC layers is aligned with regions of low expression of basement membrane proteins, as well as gaps in PC coverage due to PC retraction, suggesting that the process of neutrophil transmigration is concurrently regulated by PC contractile abilities and the PC remodeled basement membrane [43,45]. These findings demonstrate that ECM composition, particularly that deposited by PC, can directly alter EC sensitivity to circulating cytokines, and their ability to recruit neutrophils [39,43,45]. While PC are capable of increasing total ECM in the pro-fibrotic environment, the detailed proportional changes in fibronectin, laminin, and collagen I, among other ECM proteins, has not been investigated. Likewise, the ability of pro-inflammatory cytokines and chemokines found in the pro-fibrotic environment (i.e., TGF-b1, IL-1b and CCL2) to induce PC matrix deposition remains unknown [1,14,15,30,35]. In addition, the effect of bleomycin activation on PC ECM deposition has yet to be explored. Bleomycin is of particular interest as a pro-fibrotic reagent that induces matrix deposition and increased inflammatory cell infiltration characteristic of pulmonary fibrosis and scleroderma [11,15,24,28]. While the leukocyte adhesion paradigm defining EC/ neutrophil interactions has been described, the role of PC regulation in EC-mediated leukocyte recruitment remains poorly defined. This is likely due to the previous unavailability of human microvascular PC for study, limiting evaluations to isolated human EC interactions with leukocytes. These limitations have inadvertently prevented the investigation of PC and PC-derived protein regulation of inflammation, and have caused reliance upon inaccurate mouse models with species variations in inflammatory molecules, including IL-8 [13]. Here, we advance the

ª 2014 John Wiley & Sons Ltd

leukocyte adhesion cascade by elaborating on the influence of microvascular PC; accomplished through long-term culture of PC under pro-inflammatory conditions, decellularization, and preservation of PC-deposited matrix proteins. Furthermore, we use the deposited PC-ECM as an EC substrate, utilizing HUVEC to evaluate the influence of human cell-derived matrix proteins on EC adhesion molecule expression and resulting regulation of neutrophil transmigration. The novel methods and results described in this study emphasize the importance of a composite microvascular structure, inclusive of EC, PC, and the shared basement membrane, in advancing our understanding of inflammation and related disorders.

MATERIALS AND METHODS Microvascular Cell Isolation Human microvascular PC were isolated from human placental microvascular segments by explant outgrowth as previously described [2,18,21]. Briefly, a human placenta was segmented, digested with collagenase, and the resulting cells were cultured in M199 (Gibco, Grand Island, NY, USA) medium supplemented with 20% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT, USA), and 1% penicillinstreptomycin (Gibco). We verify that cultured PC express Thy-1, NG2, and smooth muscle contractile proteins, whereas lacking markers for EC (CD31), leukocytes (CD45), or smooth muscle cells (SM22-a, SMMHC), as detected by flow cytometry [2,21]. While NG2 positive PC may be derived from the post-capillary or arterial vessels, their use in examining leukocyte extravasation has been validated in several studies [26,27,33]. Human umbilical vein EC were obtained from the Yale Vascular Biology and Therapeutics Core Facility. EC were isolated from human umbilical veins by collagenase digestion and cultured on 0.1% gelatin coated flasks with M199 medium supplemented with 20% FBS, 1% penicillin-streptomycin, 0.1 mg/mL heparin, and 50 mg/mL endothelial cell growth supplement (Collaborative Biomedical Products, Bedford, MA, USA) as described [2]. Cultured EC express CD31, but lack CD45, as characterized by flow cytometry.

Pericyte Matrix Deposition PC (passage 4–9) were cultured on uncoated 25 mm circular glass coverslips (VWR, Radnor, PA, USA) or 3.0 lm pore polycarbonate Transwells (Corning, Corning, NY, USA) until confluent. The cell media were then supplemented with 100 lM ascorbic acid (Spectrum Chemicals, New Brunswick, NJ, USA) and activated with control media (non-activated control), 1 ng/mL TGF-b1 (PeproTech, Rocky Hill, NJ, USA), 10 U/mL IL-1b (PeproTech), or 12.5 ng/mL CCL2 (PeproTech). The media were replaced three times a week for four consecutive weeks (28 days) to mimic the chronic

55

P. Sava et al.

pro-inflammatory conditions present during fibrosis [24]. For bleomycin activation, the cells were activated with 1 lg/ mL bleomycin sulfate (Cayman Chemical, Ann Arbor, MI) for 1 day only as previously described [37]. The following day, the media were replaced with non-activated media containing 100 lM ascorbic acid and were replaced three times a week for four consecutive weeks (28 days) (see Figure S1a–g for schematic explanation). TGF-b1 and CCL2 concentrations were based on levels found in the fibrotic lung, whereas bleomycin concentrations were based on described fibrotic activation in fibroblasts. Lastly, IL-1b dose was based on IL-1b activation studies [2,17,22,37]. PC viability and functionality after treatment was monitored by staining for a-smooth muscle actin (SMA) using mouse anti-human a-SMA (Clone 1A4; Abcam, Cambridge, MA, USA) and secondary anti-mouse fluorescein isothiocyanate (FITC) antibody (Sigma, St Louis, MO, USA) at 7 days following activation (Figure S2a–b).

Neutrophil Isolation

Matrix Decellularization, Use, and Characterization

Neutrophil Transmigration

After 28 days of activation, the PC cellular layer was removed using 0.5% Triton-X (Sigma), and 10 mM ammonium hydroxide (JT Baker, Center Valley, PA, USA) in PBS for five minutes with shaking [4]. For EC cell culture, surfaces with PC-ECM were thoroughly rinsed in PBS and immediately seeded. For matrix composition analysis, the coverslips were fixed in 4% PFA/PBS solution, blocked against non-specific binding with 2% BSA, and stained with mouse antihuman collagen I (Clone ab233446; Abcam), rabbit antihuman laminin (Clone ab11575; Abcam), mouse anti-human collagen IV (Clone C1926; Sigma), or rabbit anti-human fibronectin (Clone F3648; Sigma) antibodies. The samples were then stained with secondary IRDye 680LT donkey antimouse IgG (Li-Cor, Lincoln, NE, USA), or IRDye 800CW goat anti-rabbit IgG (Li-Cor) antibodies, rinsed with PBS, and immediately read on Li-Cor Odyssey Infrared Scanner at 1.1X magnification. MFI measurements were made by selecting areas with PC matrix deposition and subtracting background signal using Li-Cor Image Studio software. Excitation intensity was optimized to ensure that signal was in the linear range for each experiment. MFI was used as a measure of the matrix protein concentrations and normalized to non-activated controls within each experiment.

Neutrophil transmigration assays were performed as previously described [2,18]. Briefly, EC (passage 2–5) were cultured on 3.0 lm pore polycarbonate Transwells containing decellularized PC-deposited matrix, 0.1% gelatin, 100 lg/mL collagen I (Sigma), or 10 lg/mL fibronectin (Millipore, Billerica, MA, USA). Collagen I and fibronectin concentrations are based on quantification studies performed on PC-deposited ECM. EC were cultured until confluence (four to five days) and activated with 12.5 ng/mL CCL2, or 10 U/mL IL-1b for four hours. Non-activated EC cultured on the same matrix proteins were used as negative controls. Following activation, EC-cultured Transwell inserts were transferred to six-well plates coated with a layer of 1% agarose, which facilitated the removal of transmigrated neutrophils for further analysis. Approximately 3 million na€ıve neutrophils were placed on top of the insert and allowed to transmigrate through the activated monolayers for one hour. Transmigrated neutrophils were collected from the bottom of the insert, counted by hemocytometer, and expressed as percent transmigration of total. For b2 integrin blocking, neutrophils were pre-incubated with 10 lg/mL mouse anti-human b2 integrin antibody (Clone TS1/18; BioLegend, San Diego, CA, USA) for 15 minutes prior to introduction to EC transwells.

Neutrophils were isolated from healthy human volunteers under a protocol approved by the Yale Human Investigation Committee, with written consent given by each volunteer as described [2,18]. Venous blood was drawn into a syringe containing citrate-phosphate-dextrose (Sigma) and 6% HMW Dextran (Polysciences INC, Warrington, PA, USA). The suspension was allowed to stand for 45 minutes, and the top layer consisting of plasma was removed and centrifuged at 1200 RPM. Following centrifugation and removal of the supernatant, the pellet was resuspended in PBS and gently added to Histopaque 1077 (Sigma). The suspension was centrifuged at 1200 RPM, the supernatant was removed, and the neutrophil-rich pellet was resuspended in PBS. The remaining Histopaque was removed with another centrifugation step, the resulting pellet was resuspended in PBS containing calcium, magnesium and 6% glucose, and neutrophils were counted by hemocytometer.

EC Adhesion and Growth EC (passage 2–5) were cultured on 25 mm circular coverslips containing decellularized PC-deposited matrix. Images were taken at four hours, one day, two days, and five days following seeding, and cell counts were performed. Adhesion percentage of EC was determined by calculating the percentage of the total cells seeded that adhered on the PC matrix at four hours. EC population doublings were calculated compared to the initial count at four hours.

56

EC Expression of Leukocyte Adhesion Molecules EC were cultured on 25 mm circular coverslips containing decellularized PC-deposited matrix, 0.1% gelatin, 100 lg/ mL collagen I, or 10 lg/mL fibronectin. The EC were cultured to confluence and activated with 12.5 ng/mL CCL2, or 10 U/mL IL-1b for four hours. Non-activated EC cultured on the same matrix proteins were used as negative controls. After four hours of activation, the cells

ª 2014 John Wiley & Sons Ltd

PC ECM Regulates Neutrophil Recruitment

were trypsinized, immediately fixed in 4% PFA, and then blocked with 2% BSA. ICAM-1 expression was determined by labeling the cells with mouse anti-human ICAM-1 antibody (Clone BBIG-I1; R&D Systems, Minneapolis, MN, USA), followed by anti-mouse FITC antibody (Sigma). PECAM-1 expression was determined by labeling the cells with goat anti-human PECAM-1 antibody (Clone sc-1505; Santa Cruz Biotechnology, Santa Cruz, CA, USA), followed by anti-goat Alexa Fluor 647 antibody (Invitrogen, Grand Island, NY, USA). After incubation, the samples were rinsed with PBS and run on a BD Accuri C6 Flow Cytometer (BD, Franklin Lakes, NJ, USA). As negative control, EC were incubated with isotype control IgG and all samples were normalized to negative control. FlowJo software (Tree Star, Ashland, OR, USA) was used to analyze histograms and the MFI was recorded. For localization of ICAM-1 and PECAM-1, EC were cultured on 0.1% gelatin, 100 lg/mL collagen I, or 10 lg/mL fibronectin coated coverslips to confluence and activated for four hours with 10 U/mL IL-1b. After activation, the EC monolayers were rinsed with PBS, fixed with 4% PFA, and blocked with 2% BSA. Immunofluorescence images of ICAM-1 and PECAM-1 were determined by labeling the cells with mouse anti-human ICAM-1 antibody followed by anti-mouse FITC antibody, and goat anti-human PECAM-1 antibody followed by anti-goat Alexa Fluor 647 antibody. Fluorescence images were taken at 40X magnification with ZEISS Axiovert 200M fluorescent microscope. As control for non-specific antibody signal, EC were incubated with isotype control IgG and evaluated as described above. Localization analysis was performed on randomly selected EC using a MatLab (Mathworks, Natick, MA, USA) algorithm that generates curves of pixel intensity as a function of distance from the cell nucleus, as previously described [9]. Microsoft Excel was used to determine the location on the cell where the intensity peak occurred, and the corresponding distance along the cell was determined. The cell border was determined by labeling the cell membrane with PKH26 (Sigma). A threshold for the cell membrane was empirically determined by calculating the intensity difference between the cell edge and background. Both ICAM-1 and PECAM-1 localization were normalized to the cell borders as determined by PKH26 labeling. The analysis was performed on n ≥ 30 cells from n ≥ 3 samples collected from n ≥ 3 individual experiments.

Interleukin-8 Expression Interleukin-8 (IL-8 or CXCL8) secretion was measured using a human IL-8 ELISA Ready-SET!-Go kit (eBioscience, San Diego, CA, USA). EC were cultured on 25 mm circular coverslips containing decellularized PC-deposited matrix, gelatin, collagen I, or fibronectin to confluence. The monolayers were activated with IL-1b for four hours and the media

ª 2014 John Wiley & Sons Ltd

were assayed with the IL-8 ELISA kit. For bound IL-8 imaging, EC were cultured on 25 mm circular coverslips containing decellularized PC-deposited matrix, gelatin, collagen I, or fibronectin to confluence and activated with IL-1b for four hours. EC were then fixed in 4% PFA, blocked with 2% BSA, and labeled with primary anti-human IL-8 antibody (eBioscience), followed by secondary anti-mouse FITC antibody (Sigma). Samples were imaged using a ZEISS Axiovert 200M and overall IL-8 intensity was analyzed using ImageJ (NIH) [32]. As control for non-specific antibody signal, EC were incubated with isotype control IgG and evaluated as described above.

Growth Factor Binding to the ECM The presence of bound growth factors or cytokines on the PC-deposited ECM was examined. Following 28 days of activation, the PC-deposited ECM was decellularized and labeled with FITC-conjugated mouse anti-CCL2 antibody (eBioscience), mouse anti-TGF-b1 antibody (R&D Systems) followed by anti-mouse FITC antibody (Sigma), or goat antiIL-1b antibody (Santa Cruz Biotechnology) followed by antigoat Alexa Fluor 647 antibody (Invitrogen). Glass coverslips coated with 12.5 ng/mL CCL2, 1 ng/mL TGF-b1, or 10 U/ml IL-1b for 24 hours were used as positive controls and noncoated glass coverslips were used as negative controls. Immunofluorescence images of CCL2, TGF-b1, or IL-1b were taken and representative images are shown in Figure S3a–c.

Statistical Analysis Data are expressed as mean  standard error and were analyzed for significance by One-way ANOVA with Tukey’s post hoc test, with significance defined at p ≤ 0.05, indicated by *, #, or $, using Microsoft Excel and GraphPad Prism 6. Correlation analysis using the Pearson R coefficient was performed on GraphPad Prism 6. An R2 value closer to 1 or 1 indicates higher linear correlation. A minimum of three individuals were used as donors for each neutrophil transmigration experiment.

Ethics Statement The use and attainment of human cells was approved by Yale University Human Investigation Committee (HIC) of the Internal Review Board (IRB) as part of the Human Research Protection Program. Human tissues were acquired with compliance to HIPPA standards. This research is conducted in accordance with approved protocols and with the standards set by the Helsinki Declaration. All advertisements for volunteers and written informed consent are approved by the Yale University HIC IRB and were obtained from all human volunteers prior to blood collection. Data collection and analyses were performed anonymously.

57

P. Sava et al.

RESULTS Pro-Inflammatory Activation of PC Alters Deposited ECM In vivo, microvascular PC deposit matrix proteins contributing to the formation of microvascular basement membrane and the extravascular stromal matrix, whereas additionally remodeling the matrix under pro-inflammatory conditions [38,43,44]. Here, PC matrix deposition was characterized following activation with pro-inflammatory chemokines or cytokines commonly upregulated in the fibrotic environment, CCL2, IL-1b, and TGF-b1, and bleomycin, a commonly used experimental inducer of fibrotic matrix formation (Figure S1) [1,15,28,48–50]. PC were activated for 28 days and cell survival was monitored throughout culture, demonstrating no differences in cell survival between activation strategies (Figure S2). After 28 days of activation, the matrix was decellularized and stained for basement membrane proteins collagen I, laminin, collagen IV, and fibronectin. Non-activated PC-deposited matrix (PC-ECM) was used as a negative control to which activation conditions were compared. TGF-b1 activation of the PC (TGF-b1PC-ECM) induced a 91.2  40.3% increase in collagen I (p ≤ 0.05) and a 25.4  6.2% increase in laminin (p ≤ 0.05) production when compared with non-activated PC-ECM (Figure 1A,B). IL-1b activation of PC (IL-1b-PC-ECM) induced a 49.7  18.5% increase in collagen I (p ≤ 0.05), 34.4  12.8% increase in collagen IV (p ≤ 0.05), 67.4  24.7% increase in laminin (p ≤ 0.05), and a 43.2  16.6% increase in fibronectin (p ≤ 0.05) production when compared with non-activated PC-ECM (Figure 1A,B). CCL2 activation of PC (CCL2-PC-ECM) did not affect the deposition of collagen I, laminin, collagen IV or fibronectin, when compared with non-activated PC-ECM (Figure 1A,B). Interestingly, bleomycin activation (bleo-PC-ECM) did not significantly alter PC fibronectin or laminin deposition, though it did reduce the production of collagen I by 49.3  14.2% (p ≤ 0.05), and collagen IV by 37.1  9.0% (p ≤ 0.05) when compared with non-activated PC-ECM (Figure 1A,B). These data demonstrate that the composition of ECM deposited by PC is altered in response to proinflammatory cytokines and bleomycin [1]. Microvascular EC come in direct contact with matrix proteins deposited by PC as part of the basement membrane, influencing EC function through interactions with ECpresented matrix binding integrins. Our results demonstrate that altered PC-deposited matrices differentially support early adhesion of EC (Figure S4a–b). After four hours of contact, bleo-PC-ECM supported lower EC adhesion than all other PC-ECM. However, EC proliferative abilities were not affected, as all groups reached confluence by day 5, with daily growth rates of 33.7  12.3% on PC-ECM, 36.4  14.9% on bleo-PC-ECM, 18.3  6.4% on TGF-b1-PC-ECM,

58

35.2  8.9% on IL-1b-PC-ECM, and 47.7  11.6% on CCL2-PC-ECM. No distinct changes in EC morphology were noted on different PC matrices once the EC reached confluence (Figure S4c).

PC-Deposited Protein Composition Regulates ICAM-1-Dependent Neutrophil Recruitment Neutrophils in the extravascular space contribute to the progression of fibrosis through the release of cytokines, elastases, and reactive oxygen species that damage the affected tissue [49]. Neutrophil transmigration across EC layers corresponds to areas of altered expression of ECM proteins, potentially modified by perivascular PC [43,45]. To demonstrate any influence of PC-ECM on EC ability to recruit neutrophils, EC were grown to confluence on PC-ECM and subsequently activated with pro-inflammatory chemokine CCL2 or pro-inflammatory cytokine IL-1b. Human neutrophils were introduced to the activated EC monolayers for assessment of neutrophil transmigration. IL-1b-activated EC cultured on non-activated, bleo-, TGFb1-, IL-1b-, and CCL2-PC-ECM had a significantly increased ability to support neutrophil transmigration (p ≤ 0.05) when compared with non-activated EC on the same matrix (Figure 2A). Overall, IL-1b-activated EC cultured on bleo-PC-ECM had the highest neutrophil transendothelial migration compared to IL-1b-activated EC cultured on other PC-deposited ECMs. Inversely, IL-1b-activated EC on TGFb1-PC-ECM supported the lowest levels of neutrophil transendothelial migration. These trends were consistent, though not significant, when neutrophils were applied to non-activated (quiescent) and CCL2-activated EC. These data suggest that bleo-PC-ECM elicits a pro-inflammatory response from the cultured EC that is exacerbated when the EC are activated with IL-1b, and to a lesser extent CCL2. Our own work, and that of others, has demonstrated that ICAM-1 and PECAM-1 are important mediators in neutrophil transendothelial migration in microvessels [2,10,47]. Functional inhibition of ICAM-1, or ICAM-1 binding neutrophil b2 integrin, significantly diminishes the ability of neutrophils to adhere and transmigrate across cytokineactivated EC [16,20]. Therefore, we examined whether the altered EC recruitment of neutrophils on PC-deposited matrices of different composition was due to changes in ICAM-1 or PECAM-1 expression. In response to IL-1b activation, EC cultured on IL-1b-PC-ECM or CCL2-PCECM showed no increase in ICAM-1 expression when compared with EC cultured on PC-ECM. EC cultured on TGF-b1-PC-ECM had a 29.6  8.3% decrease in ICAM-1 expression when compared with EC culture on PC-ECM (Figure 2B). Interestingly, IL-1b-activated EC cultured on bleo-PC-ECM showed the largest increase in ICAM-1 levels amongst all matrix conditions, with ICAM-1 expression 131.1  49.4% higher than the EC ICAM-1 levels expressed

ª 2014 John Wiley & Sons Ltd

PC ECM Regulates Neutrophil Recruitment

A

B

Figure 1. Pro-inflammatory activation of PC alters matrix deposition. PC were cultured on coverslips and activated with control media (non-activated control), 1 ng/mL TGF-b1, 10 U/mL IL-1b, or 12.5 ng/mL CCL2, three times per week for four weeks. For the bleo-activated group, PC were activated with 1 lg/mL bleomycin sulfate for one day, and the media were replaced with non-activated media for four weeks. (A) Representative fluorescent images of PC-deposited matrix show collagen I (i–v) and collagen IV (xi–xv) in red, and laminin (vi–x) and fibronectin (xvi–xx) in green. (B) MFI measurements of (i) collagen I, (ii) laminin, (iii) collagen IV, and (iv) fibronectin, were performed and normalized as fold increase over non-activated PC ECM of n ≥ 5 experiments (*p ≤ 0.05 versus non-activated PC ECM).

on non-activated PC-ECM. PECAM-1 levels remained constant and were not affected by culture on matrices of different compositions (Figure 2B). Our results indicate that PC-deposited ECM can induce changes in EC recruitment of neutrophils as an effect of altered ICAM-1 expression.

Fibronectin Increases EC ICAM-1 Expression and Localization to EC–EC Borders To determine the relative effects of individual proteins that comprise the PC-deposited matrix, we cultured EC on gelatin, collagen I and fibronectin, matrix proteins demonstrated to be alternatively expressed in our PC-activated ECM and commonly upregulated in the fibrotic microenvironment [14,30,35,46]. Gelatin is used as an experimental control, commonly used for EC culture. Because PC-deposited ECM contributed to altered EC expression of ICAM-1, presumably the effect of altered collagen I and fibronectin deposition, here, we isolate the contribution of collagen I and fibronectin to altered EC expression of ICAM-1. IL-1bactivated EC grown on fibronectin had 33.4  13.8% higher ICAM-1 expression compared to EC cultured on collagen I, as shown in Figure 3A. As seen on EC cultured on PC-deposited matrix, regardless of activation strategy, EC expression of PECAM-1 did not change when EC were cultured on gelatin, collagen I or fibronectin. Immunofluorescence images of ICAM-1 and PECAM-1, presented on randomly selected IL-1b-activated EC, were analyzed to determine differences in distribution on the EC

ª 2014 John Wiley & Sons Ltd

surface. A MatLab algorithm was used that determines the pixel intensity of ICAM-1 or PECAM-1 as a function of the distance from the cell nucleus. By analyzing the ICAM-1 expression intensity at its peak, we discovered that EC cultured on fibronectin have increased ICAM-1 expression compared to EC cultured on collagen I, verifying the results from flow cytometry analysis. In fact, in-depth analysis revealed that the difference in ICAM-1 expression intensity for ECs cultured on fibronectin versus EC cultured on collagen I was significant for all points past 74.5% distance from the nucleus (Figure 3B). In addition, culture of EC on fibronectin significantly increased ICAM-1 redistribution to the EC periphery, with an intensity peak at 93.2  2.6% away from the nucleus compared to 81.8  2.5% for EC cultured on collagen I (Figure 3C). However, neither PECAM-1 localization nor peak expression intensity were significantly different across all three substrates (Figure 3B, C), supporting the hypothesis that ICAM-1, and not PECAM-1, distribution to EC periphery is regulated in part by matrix protein substrates.

Increases in Neutrophil Recruitment are Due to Increased ICAM-1 Expression, and not IL-8 IL-8 has been identified as a key effector in neutrophil extravasation, creating a chemotactic gradient that directs leukocytes toward the site of inflammation by binding directly to EC and the ECM [10]. Therefore, we examined whether differences in IL-8 expression and/or

59

P. Sava et al.

A i

B

ii

i

iii

ii

Figure 2. PC-deposited matrix affects EC recruitment of neutrophils due to altered EC ICAM-1 expression. (A) Confluent EC, cultured on Transwells containing non-activated, bleomycin, TGF-b1, IL-1b, or CCL2-PC-ECM, were activated with (i) non-activated control media, (ii) IL-1b, or (iii) CCL2 and neutrophil transmigration was examined at one hour (*p ≤ 0.05 versus non-activated-PC ECM of the corresponding EC activation, #p ≤ 0.05 versus nonactivated EC of the corresponding PC-ECM). Results from n ≥ 5 experiments with n ≥ 3 donors. (B) ICAM-1 and PECAM-1 expression of IL-1b-activated EC were determined using flow cytometry with anti-ICAM-1 followed by secondary FITC antibodies, or anti-PECAM-1 followed by secondary AlexaFluor647 antibodies from n ≥ 3 individual experiments. (i) ICAM-1 and (ii) PECAM-1 expression were normalized to EC cultured on non-activated PC-ECM within each experiment and the average expression are shown (*p ≤ 0.05 versus EC cultured on non-activated PC-ECM).

presentation on the IL-1b-activated EC surface were partially responsible for the altered levels of neutrophil recruitment on the different matrices. Immunostaining of IL-8 demonstrated no significant difference in presentation of surface bound IL-8 on EC cultured on PC-deposited ECM or isolated matrix proteins (Figure 4A,B). In addition, ELISA analysis of the condition media from IL1b-activated EC cultured on the different substrates demonstrated no significant differences in IL-8 expression (Figure 4C). These results suggest that the altered levels of EC-mediated neutrophil recruitment on different matrices were not likely due to increased IL-8 production or its surface presentation. To determine the functional effects of altered EC ICAM-1 expression and redistribution due to culture on different matrix proteins, we performed neutrophil transmigration assays. IL-1b activation increased neutrophil transmigration on EC cultured on gelatin and fibronectin when compared

60

with non-activated EC on the same matrices, whereas increases in neutrophil transmigration across EC cultured on collagen I were not significant (Figure 5A–C). In addition, non-activated, CCL2 , or IL-1b-activated EC cultured on fibronectin supported significant increases (p ≤ 0.05) in neutrophil transmigration when compared with EC cultured on collagen I. These findings indicate that high concentrations of fibronectin can significantly contribute to the creation of pro-inflammatory microenvironments that facilitate EC-mediated leukocyte recruitment, exacerbating an IL1b and CCL2-induced response. A number of studies have demonstrated the importance of the neutrophil b2 integrin interaction with ICAM-1, required for directed migration of the neutrophil to reach regions of permissible transendothelial migration [34]. To confirm that the increased transmigration we see on matrix cultured EC is b2 dependent, we blocked the b2 integrin on neutrophils and allowed the cells to transmigrate across IL-1b-activated EC

ª 2014 John Wiley & Sons Ltd

PC ECM Regulates Neutrophil Recruitment

i

A

ii

B i

C

ii

iii

v

iv

i

ii

Figure 3. Fibronectin-rich microenvironment increases EC ICAM-1, but not PECAM-1, expression and localization to cell borders. (A) ICAM-1 and PECAM-1 expression of IL-1b-activated EC were determined using flow cytometry with anti-ICAM-1 followed by secondary FITC antibodies, or antiPECAM-1 followed by secondary Alexa Fluor 647 antibodies from n ≥ 5 individual experiments. (i) ICAM-1 and (ii) PECAM-1 expression were normalized to EC cultured on gelatin within each experiment and the average expression is shown (*p ≤ 0.05 versus EC cultured on collagen I). (B) Immunofluorescent images of (i) ICAM-1, (ii) PECAM-1, and (iii) PKH26 were taken of n ≥ 30 IL-1b-activated ECs cultured on collagen I, fibronectin, or gelatin. Distribution was performed using a MatLab algorithm that examines cell cross sections and generates curves of intensity of (iv) ICAM-1 and (v) PECAM-1 versus distance along the cell, with the cell nucleus at 0% and the cell border at 100% (*p ≤ 0.05 versus EC cultured on collagen I). PKH26 was used to determine EC-EC junctions and ICAM-1 and PECAM-1 localization was normalized to PKH26 borders. (C) The locations of the (i) ICAM-1 and (ii) PECAM-1 expression peak, in relation to the cell nucleus, were determined and graphed for each condition (*p ≤ 0.05 versus EC cultured on collagen I).

layers in which we saw the most robust migration patterns. b2 integrin blocking reduced neutrophil transmigration across EC cultured on all three matrix substrates, with none of the b2 integrin blocked groups significantly different from one another (Figure 5D). In this system, ICAM-1 interactions with neutrophil b2 integrins are largely responsible for neutrophil diapedesis across EC cultured on fibronectin.

ª 2014 John Wiley & Sons Ltd

Fibronectin to Collagen I Ratio in PC-derived ECM Corresponds to Increased EC ICAM-1 Expression and Neutrophil Recruitment Our findings indicate that a fibronectin-rich microenvironment increases EC recruitment of neutrophils by increasing ICAM-1 expression on the EC surface. In addition, the composition of the PC-deposited matrix may significantly

61

P. Sava et al.

alter the ability of EC to respond to IL-1b, subsequently supporting more neutrophil transmigration. Interestingly, an increase in the ratio of fibronectin to collagen I in the PCdeposited ECM corresponds with an increase in the EC recruitment of neutrophils as well as EC ICAM-1 expression, as evaluated using Pearson Correlation using the results from Figures 1, 2, and 6). The bleo-PC-ECM had the highest fibronectin to collagen I ratio of 2.22  0.47, followed by IL-1b activated, non-activated and CCL2-activated PC ECM, with ratios of 1.02  0.18, 1.00  0.09, and 0.96  0.12, respectively. TGF-b1-PC ECM had the lowest fibronectin to collagen I ratio of 0.72  0.13 (Figure 6A). We discovered that an increase in the fibronectin to collagen I ratio in the PC-derived ECM positively correlated with an increase in the neutrophil transendothelial migration across IL-1b-activated EC cultured on the different PC-ECM, with an R2 value of 0.960 (p < 0.01) (Figure 6B). Furthermore, an increase in the fibronectin to collagen I ratio positively correlated with an increase in ICAM-1 expression, with an R2 value of 0.969 (p < 0.01), but showed no correlation with the PECAM-1 expression, with an R2 value of 0.062 (not significant) (Figure 6C,D). These results demonstrate that the composition of the PC-derived ECM, particularly the fibronectin to A i

ii

iii

B i

ii

iii

collagen I ratio, are partially responsible for the differences in EC-mediated recruitment of neutrophils.

DISCUSSION Microvascular PC and EC deposit matrix proteins that contribute to the formation of the microvascular basement membrane as well as the extravascular stromal matrix that are necessary for proper microvascular stabilization during angiogenesis [38]. However, the role of PC-derived ECM as it relates to the inflammatory response has been largely ignored. While the role of PC in the progression of tissue fibrosis is well documented, the composition of the ECM deposited by cytokine, or bleomycin, activated PC and the effect that this ECM has on EC recruitment of neutrophils remains incompletely characterized [5,14,15,30]. Therefore, the main objective of this study was to advance the leukocyte adhesion cascade by elaborating on the influence of microvascular PC-derived matrix, accomplished through longterm culture of PC under pro-inflammatory conditions, decellularization, and preservation of pericyte-deposited matrix proteins. Furthermore, we used the deposited pericyte matrix as an EC substrate, evaluating the influence of human

iv

vi

iv

i

C

62

v

ii

Figure 4. Endothelial cell substrate does not affect expression of IL-8. (A) Confluent EC cultured on (i) collagen I, (ii) fibronectin, or (iii) gelatin were activated with IL-1b and immunofluorescence images of IL-8 binding to the EC layer were performed using primary anti-IL8 followed by secondary-FITC antibodies. (iv) Fluorescence intensity was used to determine the amount of bound IL-8. (B) Similarly, bound IL-8 was examined on IL-1b-activated EC cultured on (i) non-activated, (ii) bleo, (iii) TGF-b1, (iv) IL-1b, or (v) CCL2 -PC-ECM, and (vi) fluorescence intensity was used to determine the amount of bound IL-8. (C) EC, cultured on (i) collagen I, fibronectin, or gelatin or (ii) decellularized PC-deposited matrices, were activated with IL-1b for four hours and the media were assayed with a human IL-8 ELISA kit. Results from n ≥ 5 experiments.

ª 2014 John Wiley & Sons Ltd

PC ECM Regulates Neutrophil Recruitment

Figure 5. Fibronectin-rich microenvironment increases endothelial cell recruitment of neutrophils due to b2 integrin-ICAM-1 interactions. EC cultured on Transwells containing collagen I, fibronectin, or gelatin were activated with (A) non-activated control media, (B) IL-1b, or (C) CCL2 for four hours and neutrophil transmigration was examined (*p ≤ 0.05 versus collagen I of the corresponding EC activation, #p ≤ 0.05 versus non-activated EC of the corresponding ECM protein). (D) To demonstrate the role of b2 integrin interactions with ICAM-1, b2 integrin blocking on neutrophils was achieved by incubating the cells with anti-b2 antibody prior to neutrophil transmigration ($p ≤ 0.05 versus IL-1b-activated EC of the corresponding ECM protein). Results from n ≥ 5 experiments with n ≥ 3 donors.

A

B

C

D

cell-derived matrix proteins on EC adhesion molecule expression and resulting regulation of neutrophil recruitment. Here, we developed a novel method by which microvascular cell matrix deposition can be induced, preserved, and characterized. Furthermore, we are the first to demonstrate, in vitro, the ability to control PC-derived matrix composition through treatment with pro-inflammatory and pro-fibrotic reagents, resulting in distinct matrix compositions (Figure 1). Collagen I deposition was upregulated by TGF-b1 (p ≤ 0.05), and IL-1b (p ≤ 0.05) activation, whereas bleomycin (p ≤ 0.05) activation downregulated deposition when compared with non-activated PC. Collagen IV deposition was upregulated by IL-1b activation (p ≤ 0.05), whereas being downregulated by bleomycin (p ≤ 0.05) activation. Laminin deposition was upregulated by TGF-b1 (p ≤ 0.05), and IL-1b (p ≤ 0.05) activation. Lastly, fibronectin deposition was unaffected by bleomycin, CCL2, or TGF-b1 activation, and increased by IL-1b activation (p ≤ 0.05). These findings are consistent with in vivo studies indicating that PC matrix deposition in microvasculature is altered in several disease states, including fibrosis of the lung, skin, heart, and liver [15,38,46,50]. Notably, TGF-b1 activation has been shown to upregulate matrix deposition in multiple cell types including fibroblasts and hepatic stellate cells [30,50], and was confirmed to have a similar effect on pericyte matrix deposition in this study. However, the effect of IL-1b activation on PC appears opposite to the effect on fibroblasts, as we show that PC increase their matrix

ª 2014 John Wiley & Sons Ltd

deposition whereas fibroblasts reduce their matrix deposition in response to IL-1b activation [36]. This difference possibly points to the unique role of PC as regulators of microvascular specific matrix modifiers, rather than interstitial ECM regulators. Our study has the distinction of being the first to examine the ability of human microvascular PC to alter their matrix deposition in response to pro-fibrotic activation, as similar studies have been performed on cells of non-human origin [5]. Utilizing the PC-deposited ECM as an EC substrate, we demonstrated that altered basement membrane composition resulted in a functional response in EC recruitment of neutrophils. Of note, bleomycin activation of PC-induced matrix composition that sensitized EC to pro-inflammatory activation and resulted in the most pronounced neutrophil transmigration (Figure 2). These results are consistent with in vivo findings that bleomycin induction of pulmonary fibrosis is exacerbated by increased neutrophil recruitment and persistence in the fibrotic tissue [40]. In addition, the effect of bleomycin activation on microvasculature includes activation of EC and PC, leading to the recruitment of leukocytes through the release of IL-1b, CCL2, and the increased expression of surface adhesion molecules [7,15,19,41]. We explored whether the altered recruitment of neutrophils was due to changes in EC expression of IL-8, or surface adhesion molecules ICAM-1 or PECAM-1, factors that are particularly important in neutrophil transendothelial migration [10]. We demonstrated that the enhanced recruitment of neutrophils on the Bleo-PC-ECM was not due to

63

P. Sava et al.

A

B

C

D

cytokine binding on the PC-ECM (Figure S3) or due to altered EC production of IL-8 (Figure 4). Analysis of surface adhesion molecule expression demonstrated that PECAM-1 levels were not affected by culture on different PC-ECM (Figure 2). However, the composition of the PC-deposited matrix altered EC ICAM-1 expression, exacerbating the EC response to IL-1b activation and subsequently enhancing neutrophil transmigration. The observed response of EC to cell-derived ECM can be replicated by culturing the EC on fibronectin substrates alone, consistent with findings indicating a pro-inflammatory role of fibronectin in other cell types [6,12,31]. Similar to EC cultured on bleo-PC-ECM, EC cultured on fibronectin substrates facilitated a relatively high level of neutrophil transmigration (Figure 5), partially attributed to increased ICAM-1 expression as well as redistribution to the cell periphery. This was supported by immunofluorescent single cell analysis of ICAM-1 localization demonstrating that the outermost 25.5% of the EC cell area expressed significantly more ICAM-1 when cultured on fibronectin versus collagen I. Single cell analysis also revealed that ICAM-1 was preferentially redistributed to the cell periphery on ECs cultured on fibronectin compared to EC cultured on collagen I (Figure 3). These are particularly important findings as the vast majority of neutrophil transendothelial migration occurs at tri- and bi-cellular borders that are rich in ICAM-1 [10,20]. Conversely, IL-1b-activated EC cultured on collagen I did not exhibit significantly increased neutrophil transmigration when compared with non-activated EC cultured on collagen I, providing evidence that high collagen I deposition in the basement membrane serves an immunoprotective role.

64

Figure 6. Fibronectin to collagen I ratio in PCdeposited ECM corresponds to increased neutrophil recruitment and ICAM-1 expression by EC. Decellularized PC matrix, activated with non-activated control media, bleomycin, TGFb1, IL-1b, or CCL2, was characterized for collagen I, and fibronectin. (A) Normalized ratio of fibronectin to collagen I was determined using MFI from matrix deposition (#p ≤ 0.05 versus bleo-PC ECM). (B) Neutrophil transmigration across EC cultured on PCdeposited ECM is plotted against fibronectin to collagen I ratio of the ECM. The R2 value of neutrophil transmigration is 0.960 (p ≤ 0.01). (C) ICAM-1 and (D) PECAM-1 expression of EC cultured on PC-deposited ECM is plotted against fibronectin to collagen I ratio of the ECM. The R2 value of ICAM-1 is 0.969 (p ≤ 0.01) and for PECAM-1 is 0.062 (n.s.). Standard error bars are not shown because compositional analysis was demonstrated in Figure 1.

When our matrix deposition analysis is taken into account, we discovered that the ratio of fibronectin to collagen I in the PC-derived ECM correlated with an increase in ICAM-1 and subsequent increase in neutrophil recruitment (Figure 6). Particularly, the bleo-PC ECM demonstrated the highest ratio of fibronectin to collagen I, as well as the highest ICAM-1 expression and neutrophil recruitment. Bleomycin activation exacerbates the fibrotic condition by increasing the ratio of pro-inflammatory fibronectin to quiescent collagen I in the vascular basement membrane, subsequently increasing the recruitment of neutrophils. Our findings provide evidence that bleomycin induction of fibrosis is exacerbated by the altered function of microvasculature, particularly PC. Together, the results of our study suggest that increases in neutrophil transmigration in response to EC cultured on bleomycin-activated PC matrices are the result of increased ICAM-1 expression due to the increase in fibronectin to collagen I in the basement matrix. In summary, our findings have demonstrated the effects of pro-fibrotic and pro-inflammatory activation on pericyte matrix deposition, and the subsequent effect on endothelial cell sensitization to pro-inflammatory factors, as measured by leukocyte recruitment and adhesion molecule expression. PC modify their matrix deposition, particularly the ratios of fibronectin to collagen I, in response to pro-inflammatory and pro-fibrotic stimuli. A high fibronectin to collagen I ratio induces increased ICAM-1 expression on the EC surface and increases leukocyte recruitment. These findings indicate that the microvascular ECM deposited by PC is affected by pro-inflammatory and pro-fibrotic stimuli and further exacerbates fibrosis by recruiting neutrophils to the injured tissue. Overall, our results lend support to the idea that in

ª 2014 John Wiley & Sons Ltd

PC ECM Regulates Neutrophil Recruitment

vitro studies of microvascular function should include PC, EC, and basement membrane together to more accurately replicate the microvasculature. While the scope of our study was limited to examining the role of the PC-derived ECM on EC sensitization to pro-inflammatory cytokines, the methods developed here can also be applied to examining the role of the PC-ECM on EC angiogenesis, vessel stability, or for tissue engineering applications.

in the microvascular EC that facilitate neutrophil recruitment. We demonstrate that altered matrix composition by PC under pro-fibrotic activation, such as increased expression of fibronectin compared to collagen I, increases EC ICAM-1 expression and leukocyte recruitment, further exacerbating fibrosis. Our results support the idea that in vitro studies of microvascular function should include PC, EC, and basement membrane presented together to accurately replicate native microvasculature.

PERSPECTIVE Current strategies for investigating tissue fibrosis and related inflammation are focused on the use of artificial exogenous matrices or mouse models that inaccurately reflect vascular recruitment of leukocytes due to lack of key molecules, including IL-8. Therefore, development of a strategy utilizing human microvascular cells to deposit matrix in response to pro-fibrotic activators advances our ability to detect changes

REFERENCES 1. Agostini C, Gurrieri C. Chemokine/cytokine cocktail in idiopathic pulmonary fibrosis. Proc Am Thorac Soc 3: 357–363, 2006. 2. Ayres-Sander CE, Lauridsen H, Maier CL, Sava P, Pober JS, Gonzalez AL. Transendothelial migration enables subsequent transmigration of neutrophils through underlying pericytes. PLoS ONE 8: e60025, 2013. 3. Barczyk M, Carracedo S, Gullberg D. Integrins. Cell Tissue Res 339: 269–280, 2010. 4. Burton VJ, Butler LM, McGettrick HM, Stone PC, Jeffery HC, Savage CO, Rainger GE, Nash GB. Delay of migrating leukocytes by the basement membrane deposited by endothelial cells in long-term culture. Exp Cell Res 317: 276–292, 2011. 5. Canfield AE, Allen TD, Grant ME, Schor SL, Schor AM. Modulation of extracellular matrix biosynthesis by bovine retinal pericytes in vitro: effects of the substratum and cell density. J Cell Sci 96(Pt 1): 159–169, 1990. 6. Chiang HY, Korshunov VA, Serour A, Shi F, Sottile J. Fibronectin is an important regulator of flow-induced vascular remodeling. Arterioscler Thromb Vasc Biol 29: 1074– 1079, 2009. 7. Chua F, Dunsmore SE, Clingen PH, Mutsaers SE, Shapiro SD, Segal AW, Roes J, Laurent GJ. Mice lacking neutrophil elastase are resistant to bleomycin-induced pulmonary fibrosis. Am J Pathol 170: 65– 74, 2007. 8. Evani SJ, Prabhu RG, Gnanaruban V, Finol EA, Ramasubramanian AK. Monocytes mediate metastatic breast tumor cell adhe-

ª 2014 John Wiley & Sons Ltd

9.

10.

11.

12.

13.

14.

15.

ACKNOWLEDGMENTS We thank Dr. Laura E. Niklason (Yale University) for critical edits and Dr. Kathryn Miller-Jensen (Yale University) for use of laboratory equipment. This work was supported by The Hartwell Foundation. The authors have no conflicting financial interests.

sion to endothelium under flow. FASEB J 27: 3017–3029, 2013. Furcht CM, Munoz Rojas AR, Nihalani D, Lazzara MJ. Diminished functional role and altered localization of SHP2 in non-small cell lung cancer cells with EGFR-activating mutations. Oncogene 32: 2346–2355, 2355 e2341-2310, 2013. Gane J, Stockley R. Mechanisms of neutrophil transmigration across the vascular endothelium in COPD. Thorax 67: 553– 561, 2012. Harrison JH, Jr, Lazo JS. High dose continuous infusion of bleomycin in mice: a new model for drug-induced pulmonary fibrosis. J Pharmacol Exp Ther 243: 1185–1194, 1987. Higashimoto I, Chihara J, Kawabata M, Nakajima S, Osame M. Adhesion to fibronectin regulates expression of intercellular adhesion molecule-1 on eosinophilic cells. Int Arch Allergy Immunol 120(Suppl. 1): 34–37, 1999. Hol J, Wilhelmsen L, Haraldsen G. The murine IL-8 homologues KC, MIP-2, and LIX are found in endothelial cytoplasmic granules but not in Weibel-Palade bodies. J Leukoc Biol 87: 501–508, 2010. Humphreys BD. Targeting pericyte differentiation as a strategy to modulate kidney fibrosis in diabetic nephropathy. Semin Nephrol 32: 463–470, 2012. Hung C, Linn G, Chow YH, Kobayashi A, Mittelsteadt K, Altemeier WA, Gharib SA, Schnapp LM, Duffield JS. Role of lung pericytes and resident fibroblasts in the pathogenesis of pulmonary fibrosis. Am J Respir Crit Care Med 188: 820–830, 2013.

16. Issekutz AC, Rowter D, Springer TA. Role of ICAM-1 and ICAM-2 and alternate CD11/CD18 ligands in neutrophil transendothelial migration. J Leukoc Biol 65: 117– 126, 1999. 17. Kawai Y, Kaidoh M, Yokoyama Y, Sano K, Ohhashi T. Chemokine CCL2 facilitates ICAM-1-mediated interactions of cancer cells and lymphatic endothelial cells in sentinel lymph nodes. Cancer Sci 100: 419–428, 2009. 18. Lauridsen HM, Pober JS, Gonzalez AL. A composite model of the human postcapillary venule for investigation of microvascular leukocyte recruitment. FASEB J 28: 1166–1180, 2014. 19. Leach HG, Chrobak I, Han R, Trojanowska M. Endothelial cells recruit macrophages and contribute to a fibrotic milieu in Bleomycin lung injury. Am J Respir Cell Mol Biol 49: 1093–1101, 2013. 20. Ley K, Laudanna C, Cybulsky MI, Nourshargh S. Getting to the site of inflammation: the leukocyte adhesion cascade updated. Nat Rev Immunol 7: 678–689, 2007. 21. Maier CL, Shepherd BR, Yi T, Pober JS. Explant outgrowth, propagation and characterization of human pericytes. Microcirculation 17: 367–380, 2010. 22. Maroni D, Davis JS. Transforming growth factor Beta 1 stimulates profibrotic activities of luteal fibroblasts in cows. Biol Reprod 87: 127, 2012. 23. Millan J, Hewlett L, Glyn M, Toomre D, Clark P, Ridley AJ. Lymphocyte transcellular migration occurs through recruitment of endothelial ICAM-1 to caveola- and Factin-rich domains. Nat Cell Biol 8: 113– 123, 2006.

65

P. Sava et al.

24. Moeller A, Ask K, Warburton D, Gauldie J, Kolb M. The bleomycin animal model: a useful tool to investigate treatment options for idiopathic pulmonary fibrosis? Int J Biochem Cell Biol 40: 362–382, 2008. 25. Muller WA. Leukocyte-endothelial-cell interactions in leukocyte transmigration and the inflammatory response. Trends Immunol 24: 327–334, 2003. 26. Murfee WL, Rehorn MR, Peirce SM, Skalak TC. Perivascular cells along venules upregulate NG2 expression during microvascular remodeling. Microcirculation 13: 261– 273, 2006. 27. Murfee WL, Skalak TC, Peirce SM. Differential arterial/venous expression of NG2 proteoglycan in perivascular cells along microvessels: identifying a venule-specific phenotype. Microcirculation 12: 151–160, 2005. 28. Phan SH, Varani J, Smith D. Rat lung fibroblast collagen metabolism in bleomycin-induced pulmonary fibrosis. J Clin Investig 76: 241–247, 1985. 29. Phillipson M, Heit B, Colarusso P, Liu L, Ballantyne CM, Kubes P. Intraluminal crawling of neutrophils to emigration sites: a molecularly distinct process from adhesion in the recruitment cascade. J Exp Med 203: 2569–2575, 2006. 30. Rajkumar VS, Howell K, Csiszar K, Denton CP, Black CM, Abraham DJ. Shared expression of phenotypic markers in systemic sclerosis indicates a convergence of pericytes and fibroblasts to a myofibroblast lineage in fibrosis. Arthritis Res Ther 7: R1113–R1123, 2005. 31. Ramanathan A, Karuri N. Fibronectin alters the rate of formation and structure of the fibrin matrix. Biochem Biophys Res Commun 443: 395–399, 2014. 32. Rasband WS. ImageJ. U. S. National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997–2005. 33. Schallek J, Geng Y, Nguyen H, Williams DR. Morphology and topography of retinal pericytes in the living mouse retina using in vivo adaptive optics imaging and ex vivo characterization. Invest Ophthalmol Vis Sci 54: 8237–8250, 2013. 34. Schenkel AR, Mamdouh Z, Muller WA. Locomotion of monocytes on endothelium is a critical step during extravasation. Nat Immunol 5: 393–400, 2004. 35. Schrimpf C, Duffield JS. Mechanisms of fibrosis: the role of the pericyte. Curr Opin Nephrol Hypertens 20: 297–305, 2011. 36. Siwik DA, Chang DL, Colucci WS. Interleukin-1beta and tumor necrosis factor-alpha decrease collagen synthesis and increase matrix metalloproteinase activity in cardiac fibroblasts in vitro. Circ Res 86: 1259– 1265, 2000.

66

37. Sterling KM, Jr, Harris MJ, Mitchell JJ, Cutroneo KR. Bleomycin treatment of chick fibroblasts causes an increase of polysomal type I procollagen mRNAs. Reversal of the bleomycin effect by dexamethasone. J Biol Chem 258: 14438– 14444, 1983. 38. Stratman AN, Davis GE. Endothelial cellpericyte interactions stimulate basement membrane matrix assembly: influence on vascular tube remodeling, maturation, and stabilization. Microsc Microanal 18: 68–80, 2012. 39. Summers L, Kangwantas K, RodriguezGrande B, Denes A, Penny J, Kielty C, Pinteaux E. Activation of brain endothelial cells by interleukin-1 is regulated by the extracellular matrix after acute brain injury. Mol Cell Neurosci 57C: 93–103, 2013. 40. Takahashi T, Hato F, Yamane T, Fukumasu H, Suzuki K, Ogita S, Nishizawa Y, Kitagawa S. Activation of human neutrophil by cytokine-activated endothelial cells. Circ Res 88: 422–429, 2001. 41. Takemasa A, Ishii Y, Fukuda T. A neutrophil elastase inhibitor prevents bleomycininduced pulmonary fibrosis in mice. Eur Respir J 40: 1475–1482, 2012. 42. Turturro MV, Christenson MC, Larson JC, Young DA, Brey EM, Papavasiliou G. MMPsensitive PEG diacrylate hydrogels with spatial variations in matrix properties stimulate directional vascular sprout formation. PLoS ONE 8: e58897, 2013. 43. Voisin MB, Probstl D, Nourshargh S. Venular basement membranes ubiquitously express matrix protein low-expression regions: characterization in multiple tissues and remodeling during inflammation. Am J Pathol 176: 482–495, 2010. 44. Voisin MB, Woodfin A, Nourshargh S. Monocytes and neutrophils exhibit both distinct and common mechanisms in penetrating the vascular basement membrane in vivo. Arterioscler Thromb Vasc Biol 29: 1193–1199, 2009. 45. Wang S, Cao C, Chen Z, Bankaitis V, Tzima E, Sheibani N, Burridge K. Pericytes regulate vascular basement membrane remodeling and govern neutrophil extravasation during inflammation. PLoS ONE 7: e45499, 2012. 46. Wight TN, Potter-Perigo S. The extracellular matrix: an active or passive player in fibrosis?. Am J Physiol Gastrointest Liver Physiol 301: G950–G955, 2011. 47. Woodfin A, Voisin MB, Imhof BA, Dejana E, Engelhardt B, Nourshargh S. Endothelial cell activation leads to neutrophil transmigration as supported by the sequential roles of ICAM-2, JAM-A, and PECAM-1. Blood 113: 6246–6257, 2009.

48. Wynn TA. Fibrotic disease and the T(H)1/T (H)2 paradigm. Nat Rev Immunol 4: 583– 594, 2004. 49. Wynn TA. Integrating mechanisms of pulmonary fibrosis. J Exp Med 208: 1339– 1350, 2011. 50. Wynn TA, Ramalingam TR. Mechanisms of fibrosis: therapeutic translation for fibrotic disease. Nat Med 18: 1028–1040, 2012.

SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Figure S1. PC were cultured on glass coverslips or Transwells and (a) activated with (i) control media (non-activated control), (ii) 1 lg/mL bleomycin sulfate, (iii) 1 ng/mL TGF-b1, (iv) 10 U/mL IL-1b, or (v) 12.5 ng/mL CCL2. Bleomycin activation was only applied for one day and the media was replaced with non-activated control media for the remaining time. The other groups were activated three times a week for four weeks. (b) Following activation, the PC were decellularized, the resulting matrix was fixed in 4% PFA, and (c) stained with mouse anti-human collagen I, rabbit anti-human laminin, mouse anti-human collagen IV, or rabbit anti-human fibronectin antibodies. The samples were then stained with IRDye 680LT donkey antimouse, or IRDye 800CW goat anti-rabbit antibodies and read on a Li-Cor Odyssey Infrared Scanner. (d) EC were cultured until confluence (four to five days) on decellularized PC-deposited matrix and (e) activated with (i) control media (nonactivated control), (ii) 10 U/mL IL-1b, or (iii) 12.5 ng/mL CCL2. (f) After four hours of activation, neutrophil recruitment was evaluated by examining neutrophil transendothelial migration. To determine surface adhesion molecule expression, (g) EC were trypsinized, fixed in 4% PFA, and labeled with mouse antihuman ICAM-1 antibody, followed by anti-mouse FITC antibody, or goat antihuman PECAM-1 antibody, followed by anti-goat Alexa Fluor 647 antibody. The

ª 2014 John Wiley & Sons Ltd

PC ECM Regulates Neutrophil Recruitment

samples were then run on a BD Accuri C6 Flow Cytometer. Figure S2. PC maintained growth for 28 days. (a) PC, activated with control media (non-activated control), bleomycin, TGF-b1, IL-1b, or CCL2, were cultured for 28 days and images were taken on days 1, 7, 14, and 28 following activation to confirm that cell growth was unaffected by activation strategy. (b) At seven days following activation, immunofluorescence images of a-SMA were taken to confirm cells maintained PC phenotype. Representative images of a minimum n ≥ 5 experiments are shown. Figure S3. The presence of bound growth factors or cytokines on the PCdeposited ECM was examined. Following 28 days of activation, the PCdeposited ECM was decellularized. (a)

ª 2014 John Wiley & Sons Ltd

Binding of CCL2 was examined on (i) uncoated glass, (ii) CCL2-PC-ECM, or (iii) glass coverslip coated with 12.5 ng/mL CCL2 for 24 hours, using a FITC-conjugated mouse anti-CCL2 antibody. (b) Binding of IL-1b was examined on (i) uncoated glass, (ii) IL1b-PC-ECM, or (iii) glass coverslip coated with 10 U/mL IL-1b using a goat anti IL-1b antibody followed by secondary anti-goat Alexa Fluor 647 antibody. (c) Binding of TGF-b1 was examined on (i) uncoated glass, (ii) TGF-b1-PC-ECM, or (iii) glass coverslip coated with 1 ng/mL TGF-b1, using a mouse anti-TGF-b1 antibody followed by anti-mouse FITC antibody. Representative immunofluorescence images of n ≥ 3 experiments are shown.

Figure S4. Altered PC-deposited matrix affects EC adhesion and growth. EC were cultured on non-activated, bleo, TGF-b1, IL-1b, and CCL2-activated PC-ECM and images were taken at four hours, one day, two days, and five days. (a) EC adhesion on decellularized PC-ECM was examined after four hours (#p ≤ 0.05 versus Bleo–PC-ECM) and shown as percent of total cells seeded that adhered at four hours. No significant differences were seen on EC adhesion on activated PC-ECMs compared to the non-activated control. Additionally, (b) EC growth was monitored by taking images daily and counting cells until EC reached confluence with n ≥ 5 samples. (c) Representative images of EC at day 4 are shown to demonstrate the morphology of the cells.

67