Secretion by Fibronectin and Type I Collagen of Human Airway Smooth

Home , Collagen, Type I collagen

Multiple β1 Integrins Mediate Enhancement of Human Airway Smooth Muscle Cytokine Secretion by Fibronectin and Type I Collagen

This information is current as Qi Peng, Dilys Lai, Trang T.-B. Nguyen, Vivien Chan, of September 27, 2021. Takeshi Matsuda and Stuart J. Hirst J Immunol 2005; 174:2258-2264; ; doi: 10.4049/jimmunol.174.4.2258 http://www.jimmunol.org/content/174/4/2258 Downloaded from

References This article cites 33 articles, 6 of which you can access for free at: http://www.jimmunol.org/content/174/4/2258.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 27, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

␤ Multiple 1 Integrins Mediate Enhancement of Human Airway Smooth Muscle Cytokine Secretion by Fibronectin and Type I Collagen1

Qi Peng,2 Dilys Lai,2 Trang T.-B. Nguyen, Vivien Chan, Takeshi Matsuda, and Stuart J. Hirst3

Altered airway smooth muscle (ASM) function and enrichment of the extracellular matrix (ECM) with interstitial collagen and fibronectin are major pathological features of airway remodeling in asthma. We have previously shown that these ECM components confer enhanced ASM proliferation in vitro, but their action on its newly characterized secretory function is unknown. Here, we examined the effects of fibronectin and collagen types I, III, and V on IL-1␤-dependent secretory responses of human ASM cells, and characterized the involvement of specific integrins. Cytokine production (eotaxin, RANTES, and GM-CSF) was eval- uated by ELISA, RT-PCR, and flow cytometry. Function-blocking integrin mAbs and RGD (Arg-Gly-Asp)-blocking peptides were Downloaded from used to identify integrin involvement. IL-1␤-dependent release of eotaxin, RANTES, and GM-CSF was enhanced by fibronectin and by fibrillar and monomeric type I collagen, with similar changes in mRNA abundance. Collagen types III and V had no effect on eotaxin or RANTES release but did modulate GM-CSF. Analogous changes in intracellular cytokine accumulation were found, ␣ ␤ ␣ ␤ ␣ ␤ but in <25% of the total ASM cell population. Function-blocking Ab and RGD peptide studies revealed that 2 1, 5 1, v 1, ␣ ␤ ␤ ␣ ␤ and v 3 integrins were required for up-regulation of IL-1 -dependent ASM secretory responses by fibronectin, while 2 1 was ␤ an important transducer for type I collagen. Thus, fibronectin and type I collagen enhance IL-1 -dependent ASM secretory http://www.jimmunol.org/ ␤ responses through a 1 integrin-dependent mechanism. Enhancement of cytokine release from ASM by these ECM components may contribute to airway wall inflammation and remodeling in asthma. The Journal of Immunology, 2005, 174: 2258–2264.

hronic asthma is characterized by poorly reversible air- logic studies of asthmatic airways have revealed abnormalities in way obstruction and airway hyperresponsiveness that is both the quantity and composition of ECM proteins. Collagen C associated with intense persistent inflammation and types I, III, and V; fibronectin; tenascin; hyaluronan; versican; and ␣ ␤ structural remodeling of the airway wall (1, 2). Two prominent laminin 2/ 2 chains are increased, whereas other major ECM features of the remodeling process that may underlie or contribute components including collagen type IV and elastin are decreased to the development of airways hyperresponsiveness in asthmatics (5–9). Increased type I collagen, hyaluronan, and versican have by guest on September 27, 2021 4 include accumulation of airway smooth muscle (ASM), possibly been found localized within and surrounding ASM bundles from due to hyperplastic and/or hypertrophic changes (3, 4) and alter- asthmatics (9, 15). In the bronchoalveolar lavage fluid of asthmat- ations in the amount and composition extracellular matrix (ECM) ics, increased levels of fibronectin, hyaluronan, and laminin prod- proteins (5–9). In addition to ASM accumulation in asthma, evi- ucts are found (reflecting increased ECM turnover), which corre- dence suggests that ASM may contribute to airway wall inflam- late with asthma severity (16). These studies support marked ECM matory events by expressing cell adhesion receptors and costimu- changes in the airway wall in asthma. However, little is known of latory molecules, and by releasing multiple cytokines and the impact of enrichment of the airway wall with collagen and chemokines including those which activate eosinophils such as fibronectin on airway cell function. eotaxin, RANTES, and GM-CSF (10–13). Previously, we have reported that both fibronectin and type I ECM proteins ligating via cell surface integrin ECM receptors collagen enhance responses of cultured human ASM cells to mi- regulate diverse cellular functions including growth, migration, ␣ survival, and the maintenance of differentiated status (14). Histo- togens such as PDGF-BB and -thrombin (17). Bonacci et al. (18) report similar findings with fibroblast growth factor-2-dependent proliferation of bovine tracheal smooth muscle cultured on type I Department of Asthma, Allergy and Respiratory Science, The Guy’s, King’s and St. collagen; and Freyer et al. (19) have shown that both type I col- Thomas’ School of Medicine, King’s College London, Guy’s Hospital Campus, Lon- don, United Kingdom lagen and fibronectin increase survival of human ASM cells. More- Received for publication August 18, 2004. Accepted for publication November over, ASM cells cultured from asthmatic individuals or grown on 16, 2004. homologous ECM from asthmatic ASM cells proliferate more rapidly The costs of publication of this article were defrayed in part by the payment of page (20). Collectively, these observations affirm that ASM in situ likely is charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. subjected to the influences of signaling through ECM/integrin inter- 1 This study was supported by Asthma U.K. (no. 00/44), the Special Trustees of Guy’s actions, and have led to the suggestion that alterations in airway wall and St. Thomas’ Hospitals, and the Wellcome Trust (no. 051435). ECM microenvironment in remodeling in asthma may favor en- 2 Q.P. and D.L. contributed equally to this work. hanced survival and growth responses of ASM during inflammation 3 Address correspondence and reprint requests to Dr. Stuart J. Hirst, Department of (17–19). However, it is unknown whether the ECM has analogous Asthma, Allergy and Respiratory Science, The Guy’s, King’s and St. Thomas’ School effects on ASM secretory responses. of Medicine, Thomas Guy House, Guy’s Hospital Campus, London, U.K. SE1 9RT. E-mail address: [email protected] In the present study, we hypothesized that enrichment of the 4 Abbreviations used in this paper: ASM, airway smooth muscle; ECM, extracellular airway wall ECM environment with fibronectin and type I colla- matrix; Pl, plastic. gen, as occurs in asthma, confers increased secretory properties of

ASM. We compared IL-1␤/TNF-␣-dependent eosinophil-activat- sciences) (22). After analysis, cells were confirmed microscopically to be ing cytokine release from human ASM cells that were cultured intact. either on plastic (Pl) or fibronectin or on varying forms of collagen. Measurement of cytokine levels by ELISA Moreover, we characterized the integrin heterodimer subunits expressed by ASM and their involvement in modulating secretory Levels of eotaxin, RANTES, and GM-CSF were determined in duplicate in responses. the same samples of cell-conditioned medium by specific sandwich ELISA as described previously (13). Manual cell counts were repeated after col- lection of cell-conditioned medium to confirm numbers were unchanged Materials and Methods after stimulation, and to allow cytokine levels to be expressed initially in Isolation and culture of human ASM cells nanograms per milliliter per million cells before normalization for stimulation on Pl to correct for small variations in release between cell lines. Human ASM cells were obtained in accordance with procedures approved by the Guy’s and St. Thomas’ Hospitals’ Research Ethics Committee from RT-PCR the lobar or main bronchus of 45 nonasthmatic patients (mean age, 64 Ϯ 4 years; range, 28–78 years; 24 male, 21 female) undergoing lung resection Total RNA was isolated using the RNeasy mini kit (Qiagen), and its con- for carcinoma of the bronchus using methods described previously (13). centration was determined using the RiboGreen RNA quantification kit Fluorescent immunocytochemical and flow cytometric techniques con- (Molecular Probes). Full-length first-strand cDNA was synthesized using 2 ␮ ␮ firmed that near-confluent, FBS-deprived human ASM cells (passage 2) g of total RNA in a 33- l reverse transcriptase reaction mixture with stained (Ͼ95%) for smooth muscle-specific ␣-actin and calponin (17). Ready-To-Go You-Prime beads and random hexamer primers (Amersham Cells at passages 3–5 were used in all experiments. Biosciences) as previously described (22). To confirm PCR were per- formed within the linear for amplification, the number of cycles was initially varied from 20 to 28. The size of the amplicons corresponded to Surface coating with ECM proteins Downloaded from predicted sizes (eotaxin, 182 bp; RANTES, 332 bp; GM-CSF, 277 bp), Lyophilized human plasma fibronectin (Sigma-Aldrich), rat monomeric confirmed by sequencing. Primers used in RT-PCR were: eotaxin (24 cy- type I collagen (ICN Chemicals), and human fibrillar collagen types I, III, cles), 5Ј-CCC AAC CAC CTC CTG CTT TAA C-3Ј as sense and 5Ј-CCA and V (Chemicon International) were reconstituted in sterile PBS. ECM GAT ACT TCA TGG AAT CCT GCA C-3Ј as antisense; RANTES (26 proteins (0.1–10 ␮g/ml) diluted in PBS were adsorbed to tissue culture cycles), 5Ј-TGC TTT GCC TAC ATT GCC CG-3Ј as sense and 5Ј- CTG plasticware overnight at 37°C as previously described (17). Excess un- GGG AAG GTT TTG TAA CTG C-3Ј as antisense; GM-CSF (22 cycles), bound ECM protein was removed by aspiration and washing with PBS. 5Ј-TGA GTA GAG ACA CTG CTG CTG A-3Ј as sense and 5Ј- TCA

Unoccupied protein-binding sites were blocked with 0.1% BSA for 30 min. AAG GGG ATG ACA AGC AGA A-3Ј as antisense; 18S rRNA, 5Ј-TGA http://www.jimmunol.org/ CTC AAC ACG GGA AAC CTC AC-3Ј as sense and 5Ј-GGA CAT CTA Cell stimulation and application of blocking mAbs AGG GCS TCA CAG ACC-3Ј as antisense. cDNA samples were amplified as previously described (22). All primers were purchased from MWG Near-confluent, FBS-deprived cells from flasks were seeded (5000 cells/ Biotech. cm2) in DMEM containing 1% FBS on to Pl or ECM substrate (0.1–10 ␮g/ml)-precoated plasticware and left overnight at 37°C. In some experi- Statistical analysis ments, cells in suspension were pretreated for 30 min at room temperature with integrin function-blocking mAbs (1 ␮g/ml) or isotype-matched con- Data are mean Ϯ SEM from cells cultured from n patient donors. Data trol Abs with continuous roller mixing before seeding. After attachment, were compared using one- or two-way ANOVA, where appropriate, fol- cells were washed twice with FBS-free RPMI 1640 (containing 25 mM lowed by Bonferroni’s t test post hoc to evaluate statistical differences

HEPES, 2 mM L-glutamine, 100 U/ml penicillin/100 ␮g/ml streptomycin) between treatment groups (SigmaStat; SPSS). A p value of Ͻ0.05 was by guest on September 27, 2021 and then stimulated for 24 h in RPMI 1640 with 1 ng/ml recombinant considered significant. human IL-1␤ or 10 ng/ml recombinant human TNF-␣. Soluble RGD integrin-blocking peptides (10 ␮M) or a negative control peptide were added Results directly to adherent cells 30 min before stimulation with cytokines. Integrin function-blocking anti-human mouse mAbs (Chemicon International, un- Regulation of ASM secretory responses by ECM proteins ␣ ␣ less stated otherwise) were anti- 1 (clone FB12), anti- 2 (clone P1E6), ␤ ␣ ␣ ␣ IL-1 (1 ng/ml) induced marked release of eotaxin, RANTES, or anti- 3 (clone P1B5), anti- 5 (clone P1D6), anti- v (clone L230, gift from ␤ GM-CSF from human ASM cells when cultured on Pl (Fig. 1), Dr. J. F. Marshall (John Vane Science Center, London, U.K.)(21)), anti- 1 ␣ ␤ ␤ confirming previous observations (13). After culture in plates pre- (clone 6S6), anti- v 3 (clone LM609), and anti- 3 (clone 25E11). Mouse- purified or isotype-matched IgG (Chemicon International) was used as a coated with fibronectin or fibrillar type I collagen, IL-1␤-stimu- nonimmune control. lated cells were significantly more secretory compared with culture on Pl ( p Ͻ 0.05–0.01; Figs. 1 and 2). A similar concentration- Flow cytometric labeling of surface integrins and intracellular dependent increase in IL-1␤-dependent secretory capacity was cytokines found with monomeric type I collagen ( p Ͻ 0.05–0.01; Fig. 2), Cell surface integrins were localized by binding of the above anti-human which did not differ from the up-regulation found with fibrillar integrin subunit mAbs. Near-confluent growth-arrested cells on Pl were Ͼ ϫ type I collagen ( p 0.05, two-way ANOVA; Fig. 2). Up-regu- harvested using trypsin/EDTA, washed twice in PBS (200 g for 5 min), ␤ resuspended, and fixed with 200 ␮l of 4% formaldehyde (methanol-free, lation of IL-1 -dependent eotaxin or RANTES release was not EM-grade; Polysciences) for 30 min on ice in round-bottom FACS tubes. found with culture on either type III or V collagen ( p Ͼ 0.05, n ϭ Fixed cells (20,000/tube) were resuspended in PBS containing 3% FBS for 5; data not shown), although in the same samples there was a 30 min to prevent nonspecific mAb binding. After centrifugation, cells concentration-dependent decrease (52 Ϯ 3% at 10 ␮g/ml) in GM- were incubated on ice with the anti-␣ or anti-␤ integrin subunit unconju- gated Abs (1 ␮g/ml for 30 min), washed in PBS, then incubated in the dark CSF with type V collagen and an increase in GM-CSF release with an anti-mouse PE-conjugated secondary Ab (Sigma-Aldrich) for 30 (195 Ϯ 13% at 10 ␮g/ml) with type III collagen compared with min and washed twice in PBS. IL-1␤-stimulated cells on Pl ( p Ͻ 0.05–0.01, n ϭ 5–6; data not Cells on Pl or ECM substrates (10 ␮g/ml) for 24 h in the absence or shown). Additionally, fibronectin and type I collagen (10 ␮g/ml) combined presence of cytokine stimulation were harvested and fixed as each potentiated GM-CSF release by 156 Ϯ 5% and 137 Ϯ 7%, above. The protein transport inhibitor, brefeldin A, was not used in these Ͻ studies. Cells (20,000/tube) were permeablized with 100 ␮l of 0.5% sapo- respectively ( p 0.05–0.01 compared with Pl) when cells were nin (catalog no. S-4521; Sigma-Aldrich)/0.1% BSA in PBS containing Abs costimulated with maximally effective concentrations of IL-1␤ (1 (1 ␮g/ml). Abs used (BD Pharmingen, unless stated otherwise): PE-con- ng/ml) and TNF-␣ (10 ng/ml). In all experiments, release of jugated anti-GM-CSF (clone no. BVD2-21C11), PE-conjugated anti- eotaxin, RANTES, or GM-CSF from cells cultured on these ECM RANTES (clone no. 2D5), PE-conjugated anti-eotaxin (clone no. 43915; ␮ ␤ ␣ R&D Systems), or PE-conjugated isotype-matched control Abs (IgG1/ substrates (1 or 10 g/ml) in the absence of IL-1 /TNF- did not IgG2a). Numbers of PE-labeled cells were determined on a FACSCalibur differ from release by unstimulated cells on Pl ( p Ͼ 0.05, n ϭ 6–7 flow cytometer (BD Biosciences) using CellQuest Pro software (BD Bio- for each ECM; Figs. 1 and 2). 2260 ECM BINDING INTEGRINS AND ASM SECRETION Downloaded from http://www.jimmunol.org/

FIGURE 1. Enhancement of IL-1␤-dependent cytokine release by fi- FIGURE 2. Enhancement of IL-1␤-dependent cytokine release by type bronectin. Human ASM cells on Pl or fibronectin were either unstimulated I collagen. Human ASM cells on Pl or fibrillar or monomeric type I col- (open symbols) or stimulated (filled symbols) with 1 ng/ml IL-1␤ for 24 h. lagen were either unstimulated (open symbols) or stimulated (filled sym- by guest on September 27, 2021 Levels of eotaxin (A), RANTES (B), or GM-CSF (C) were determined by bols) with 1 ng/ml IL-1␤ for 24 h. Levels of eotaxin (A), RANTES (B), or p Ͻ ,ءء ;pϽ 0.05,ء .(p Ͻ 0.01 compared with stimulated GM-CSF (C) were determined by ELISA (n ϭ 6–7 ,ءء ;pϽ 0.05,ء .(ELISA (n ϭ 6–7 cells on Pl. 0.01 compared with stimulated cells on Pl.

Cytokine gene regulation by ECM proteins creased to 4–5% after culture on 10 ␮g/ml fibronectin or fibrillar To examine whether up-regulation of IL-1␤-dependent secretory type I collagen (n ϭ 2). Given these low levels of labeling, cells in responses by ECM substrates involved increased gene activation, subsequent studies were costimulated with maximally effective cytokine mRNA levels were examined in human ASM cells cul- concentrations of IL-1␤ (1 ng/ml) and TNF-␣ (10 ng/ml), shown tured on fibronectin. Consistent with previous reports, treatment of previously in ELISA studies to act synergistically to induce release cells growing on Pl with IL-1␤ (1 ng/ml) for 16 h increased the of GM-CSF and RANTES (10, 13). Optimal numbers of GM-CSF mRNA abundance for eotaxin, RANTES, and GM-CSF ( p Ͻ (12.2 Ϯ 3.7%)-, RANTES (19.03 Ϯ 3.72%)-, or eotaxin (11.4 Ϯ 0.05–0.01; Fig. 3) (10, 11). A further increase in eotaxin and GM- 3.41%)-labeled cells occurred with stimulation for 24 h (data not CSF mRNA occurred in IL-1␤-stimulated cells after culture on 10 shown). Kinetics for GM-SCF and RANTES accumulation were ␮g/ml fibronectin ( p Ͻ 0.05; Fig. 3). An apparent increase in similar, consistent with our previous ELISA findings in IL-1␤- IL-1␤-dependent RANTES mRNA was also found, but this failed stimulated cells (13). In keeping with the current ELISA (Figs. 1 to reach significance. No change in 18S rRNA between the treat- and 2) and mRNA findings (Fig. 3), culture on fibronectin or fibril- ments was found (Fig. 3). In two separate experiments, a similar lar type I collagen further increased the number of IL-1␤/TNF-␣- enhancement of IL-1␤-dependent eotaxin, RANTES, and GM- stimulated eotaxin-, RANTES-, or GM-CSF-positive cells by CSF mRNA abundance at 16 h was found after culture on fibrillar ϳ1.5-fold ( p Ͻ 0.05–0.01 compared with stimulated cells on Pl; type I collagen (10 ␮g/ml), compared with stimulation on Pl (not Fig. 4). Likewise, culture on type III collagen increased IL-1␤/ shown). TNF-␣-stimulated GM-CSF ( p Ͻ 0.05) but not RANTES labeling, which overall resembled closely the ELISA findings with collagen Modulation of intracellular cytokine labeling by ECM proteins type III (described above). To investigate whether enhancement of cytokine production by ECM substrates occurred in all ASM cells or in a subset, intracel- Attenuation of fibronectin or collagen I-enhanced eotaxin lular cytokine levels were examined by flow cytometry in fixed release from cultured human ASM cells by integrin specific saponin-permeablized cells. In preliminary studies, positive label- blocking mAbs ing for eotaxin, RANTES, or GM-CSF was detected in Ͻ2% of the Having defined a probable role for ECM substrates in the regula- total population after stimulation with IL-1␤ for 24 h, which in- tion of cytokine production from human ASM cells, subsequent The Journal of Immunology 2261 Downloaded from

FIGURE 3. Enhancement of IL-1␤-dependent cytokine mRNA levels by fibronectin. Human ASM cells on Pl or fibronectin (FN; 10 ␮g/ml) were either unstimulated or stimulated with 1 ng/ml IL-1␤ for 16 h. A, Repre- http://www.jimmunol.org/ sentative gels. B, Band intensities for cytokines were quantitated using ImageQuant software, expressed as the cytokine:18S rRNA ratio (n ϭ .pϽ 0.05 compared with stimulated cells on Pl,ء .(4–3 experiments examined possible integrin receptors that could mediate this response. Flow cytometry of cells on Pl confirmed that ␣ ␤ 5 and 1 integrin subunits were universally expressed and that ϳ ␣ ␣ ␣ ␣ by guest on September 27, 2021 50–60% of cells had detectable surface 1, 2, 3, and v in- ϳ ␣ ␤ tegrin subunits and 50% expressed the v 3 heterodimer (Fig. 5A). To exclude possible underestimation of the number of labeled cells due to cleavage of Ab recognition sites by the trypsin used to FIGURE 4. Enhancement of human ASM cell intracellular cytokine ex- remove the cells from the culture flasks, we compared this profile pression by ECM substrates. A, Typical experiment illustrating high of integrin subunit expression in cells harvested using PBS-con- autofluorescence levels in the absence of Ab (no Ab). Numbers in gates are taining EDTA (0.5 mM for 20 min) and found no differences, eotaxin-positive cells compared with an isotype-matched IgG control (IMC). Data in B are mean Ϯ SEM of duplicate values (n ϭ 3–4). Cells on which was further supported by examination of cultured human Pl or in plates coated with 10 ␮g/ml fibronectin (FN) or fibrillar collagen ASM cells in situ by reflected UV light immunofluorescence mi- (Col) types I or III were cultured in the absence (Ⅺ) or combined presence pϽ 0.05,ء .(croscopy (T. T.-B. Nguyen and S. J. Hirst, unpublished of 1 ng/ml IL-1␤ and 10 ng/ml TNF-␣ for 24 h (f observations). compared with stimulated cells on Pl. §,pϽ 0.05; §§§,pϽ 0.001 The panel of blocking mAbs was also used to investigate the compared with unstimulated cells on Pl. For all other comparisons, p Ͼ specific involvement of integrin subunits in the enhanced ASM 0.05. secretory responses elicited by fibronectin and type I collagen. Preliminary flow cytometry studies showed enhancement of IL- 1␤/TNF-␣-dependent intracellular eotaxin and RANTES labeling To further support a role for integrins in enhanced secretory ␮ ␤ Ͻ ␮ was reduced by 1 g/ml 1 integrin subunit-blocking mAb ( p signals by the ECM, preincubation of cells with 10 M soluble 0.05, Table I). Likewise, in ELISA studies, blocking the binding of RGD-blocking peptides (Gly-Arg-Gly-Asp-Ser (GRGDS) or Gly- ␣ ␤ matrix factors to 2 or 1 integrin subunits abolished enhancement Arg-Gly-Asp-Thr-Pro (GRGDTP)), but not the negative control of IL-1␤-dependent eotaxin release by either fibronectin or fibrillar Gly-Arg-Ala-Asp-Ser-Pro (GRADSP), was found to abolish the en- type I collagen ( p Ͻ 0.05 compared with control IgG1, Fig. 5, B hancement of IL-1␤-dependent eotaxin release by fibronectin ( p Ͻ ␮ ␣ ␣ ␣ ␤ ␣ ␤ and C). Function-blocking mAbs (1 g/ml) to 5 or v subunits 0.05, Fig. 5D). This result is in keeping with a role for 5 1, v 1, and ␣ ␤ ␣ ␤ ␣ ␤ ␣ ␤ had a similar effect, but only against enhancement by fibronectin v 3 integrin-mediated secretory signals, though 3 1, 8 1, IIb 3, Ͻ ␣ ␤ ␣ ␤ ␣ ␤ ( p 0.05 compared with control IgG1). Abs against 3 or 3 v 5, and v 6 are also known to be RGD sensitive (23). subunits were without effect ( p Ͼ 0.05 compared with control ␣ ␤ IgG1). However, the v 3 heterodimer-specific blocking mAb (1 Discussion ␮g/ml) abolished enhancement of eotaxin release by fibronectin In this present study, we demonstrate selected airway wall ECM ( p Ͻ 0.05 compared with control IgG1). A similar profile of neu- components that are increased in asthma, such as collagen and tralization was obtained when IL-1␤-dependent RANTES release fibronectin (5–9), have the capacity to up-regulate ASM cell se- by either fibronectin or type I collagen was examined in superna- cretory responses. Although absolute values for the amounts of tants in place of eotaxin (data not shown, n ϭ 4). ECM substrates in contact with ASM in the asthmatic airway are 2262 ECM BINDING INTEGRINS AND ASM SECRETION

unknown, we report that culture on fibronectin, fibrillar, or monomeric type I collagen at concentrations similar to those found to enhance ASM cell proliferative and survival responses (17–19), provided a transcriptionally regulated enhancement of the IL-1␤- dependent cytokine secretory signal. ELISA studies indicated that ␤ enhancement by fibronectin was mediated through multiple 1 in- ␣ ␤ ␣ ␤ ␣ ␤ tegrins expressed by ASM including 2 1, 5 1, and v 1, but ␣ ␤ also by v 3, as function-blocking mAbs directed against these ␣ ␤ integrin subunits or to the v 3 heterodimer prevented enhancement of IL-1␤-dependent eotaxin release by fibronectin, as did blocking a fibronectin recognition site with soluble RGD peptides. Similarly, enhancement by fibrillar collagen type I appeared to ␣ ␤ involve the 2 1 integrin. Furthermore, flow cytometric analysis indicated these events may involve a minority of cultured ASM cells because intracellular cytokine expression elicited by IL-1␤/TNF-␣ and its potentiation by the ECM could only be detected in Ͻ25% of the total population; even under conditions considered to induce max- imal cytokine release when measured by ELISA (10, 13).

Although the involvement of speciﬁc ECM components in up- Downloaded from regulation of airway and vascular smooth muscle responses such as attachment, proliferation, migration, and survival is well estab- lished (17, 18, 24, 25), up-regulation of cytokine secretory capacity by ﬁbronectin or collagen ECM substrates to our knowledge has not previously been reported in mesenchymal cells from the

lung or elsewhere. In ELISA studies, we found that overnight cul- http://www.jimmunol.org/ ture of human ASM cells in plates precoated with fibronectin, fibrillar, or monomeric denatured type I collagen increased IL-1␤- dependent cytokine release compared with similarly stimulated cells cultured on Pl. Collagen types III and V had no effect on IL-1␤-dependent eotaxin or RANTES release, but differentially affected GM-CSF release. Thus, different ECM proteins could elaborate varying patterns of cytokine secretion compared with Pl, suggesting these differences when present did not result from a nonspecific protein interaction, rather they reflected a specific re- by guest on September 27, 2021 sponse to the ECM. Enhancement by both fibrillar and monomeric type I collagen suggests both the native helical configuration (fibrillar) and proteolytic denatured (monomeric) forms of type I collagen, occurring after cleavage of fibrillar collagen in inflammation by matrix metalloproteinases (26), are sufficient to elicit secretory signals in ASM. Moreover, the addition of soluble forms of fibronectin or type I collagen directly to the cell cultures does FIGURE 5. Integrin-dependent enhancement of IL-1␤-stimulated eotaxin not result in potentiation of IL-1␤-dependent eotaxin or RANTES release by fibronectin and fibrillar type I collagen. A, Flow cytometric analysis of integrin expression by human ASM cells cultured on Pl (n ϭ 6). Also release, suggesting enhancement of cytokine release requires these shown is the effect of integrin function-blocking mAbs on enhancement of ECM substrates to be in the polymerized form present after coating IL-1␤-dependent eotaxin release (ELISA) by either fibronectin (B; FN) or (Q. Peng and J. S. Hirst, unpublished observations). collagen (C; Col) type I (10 ␮g/ml). D, The effect of soluble RGD integrin- The extent of enhancement of cytokine release was consistent blocking peptides (10 ␮M) compared with a negative control RAD peptide across each of the substrates examined with GM-CSF being the -pϽ 0.05 denotes significant reduction in eotaxin release com- most increased (ϳ200%) followed by eotaxin (ϳ150%) and RAN,ء .(n ϭ 3–4) pared with IgG1 or absence of blocking peptide. TES being the least susceptible (ϳ130%). In keeping with this

Table I. Blocking of ﬁbronectin-enhanced and type I collagen-enhanced intracellular cytokine labeling by ␤ a anti- 1 integrin

Type I Collagen (% stimulated Fibronectin (% stimulated on PL) on PL)

Treatment Eotaxin RANTES Eotaxin RANTES

Unstimulated/PL 21 Ϯ 7.4 14 Ϯ 7.7 21 Ϯ 7.4 14 Ϯ 7.7 IL-1␤/TNF-␣/PL 100 100 100 100 ␤ ␣ Ϯ Ϯ Ϯ Ϯ IL-1 /TNF- /ECM/IgG1 141 21 179 50 124 6 157 29 ء Ϯ ء Ϯ ء Ϯ ء Ϯ ␤ ␣ ␤ IL-1 /TNF- /ECM/anti- 1 116 14.2 71 14 81 20 91 9 a ␮ ␤ Effect of blocking mAb (1 g/ml) directed against the 1 integrin subunit on flow cytometric labeling of cotaxin and RANTES in human ASM cells cultured on fibronectin (10 ␮g/ml) or type I collagen (10 ␮g/ml) ECM following combined treatment with IL-1␤ (1 ng/ml) and TNF-␣ (10 ng/ml) for 24 h. Ͻ Ϯ ϭ ء , p 0.05 compared with enhancement in presence of IgG1 control by one-way ANOVA. Means SEM are shown (n 4). The Journal of Immunology 2263 enhancement of eosinophil-activating cytokine protein release by ϳ50% of human ASM cells express the promiscuous vitronectin ␣ ␤ ␣ ECM substrates, analogous increases in these cytokines were receptor, v 3, and 65–70% of cells express 2 integrin required found with techniques examining intracellular cytokine expression for type I collagen binding. These data confirm an earlier report ␤ (flow cytometry) and mRNA abundance (RT-PCR), though the showing similar expression patterns of multiple 1 integrin family ␣ ␣ ␣ increase for RANTES mRNA failed significance, which may re- subunits including 1, 3, and v (19). In keeping with this earlier ␣ ␤ flect either insufficient powering or saturation of the PCR. En- report, 5 and 1 subunits were found to be universally expressed. hancement of cytokine mRNA levels by fibronectin or type I col- ELISA studies with integrin function-blocking mAbs suggest that lagen implies the increase in eotaxin, RANTES, and GM-CSF enhancement of IL-1␤-stimulated eotaxin release by fibronectin ␣ ␤ ␣ ␤ ␣ ␤ ␣ ␤ protein detected by the ELISA and flow cytometry studies involves involved multiple integrins including 2 1, 5 1, v 1, and v 3; ␣ ␤ increased cytokine gene transcription or stabilization of mRNA while enhancement on type I collagen involved 2 1. The findings transcripts. Although not examined in the present study, it was with fibronectin support the consensus that a single ECM substrate ␣ ␣ ␤ ␣ ␤ recently shown that GM-CSF release from TNF- -stimulated eo- may ligate several integrin heterodimers. For example, 3 1, 5 1, ␣ ␤ ␣ ␤ sinophils in the presence of fibronectin involved mainly GM-CSF v 1, and v 3, which are expressed on ASM cells are all known mRNA stabilization, which was dependent on ERK phosphoryla- to bind fibronectin with varying affinities (31). Furthermore, the tion and the RNA binding proteins, YB-1 and HuR (27). Else- finding that in each case blocking Abs abolished enhancement of where, evidence suggests that growth factors can bind ECM com- cytokine release by fibronectin or type I collagen, irrespective ␣ ␤ ponents to form complexes that enhance subsequent growth factor whether the target integrins were universally ( 5 and 1)orpar- ␣ ␣ activity through integrin collaboration with growth factor recep- tially expressed ( 2 and v), supports our hypothesis from the tors. In the case of VEGF interacting with fibronectin to enhance intracellular cytokine flow cytometry findings that the enhance- Downloaded from endothelial cell migration, the amplified response was attributed to ment of eotaxin by fibronectin or type I collagen in the ELISA sustained ERK activation requiring activation of both the Flk-1 study likely involved only a subpopulation of human ASM cells ␣ ␤ ␣ ␣ (VEGF receptor) and 5 1 (fibronectin receptor) (28). Whether that also express 2 and v integrin subunits. similar mechanisms operate in human ASM cells remains an open Block of fibronectin-induced potentiation of IL-1␤-dependent ␣ question. eotaxin release, as well as that by type I collagen, by 2 subunit

␣ ␤ http://www.jimmunol.org/ This study is among the first with ASM cells to demonstrate neutralization was an unexpected finding as 2 1 is the major type intracellular cytokine expression by flow cytometric techniques. I collagen receptor. The response to fibronectin may depend on The synthesis and release of proinflammatory mediators by ASM autocrine type I collagen production, possibly via an IL-1␤-depen- is a relatively new finding (29) and has almost without exception dent GM-CSF-mediated amplification loop, and GM-CSF has re- been estimated by ELISA- or RT-PCR-based methods in experi- cently been shown to induce the production of type I collagen by ments that have not been designed to distinguish between signals ASM cells (32). Alternatively, collagen-binding regions are generated from a subset of cells rather than the total population. present in purified fibronectin (33) and other recent data in a cell- ␣ ␤ Although the single labeling flow cytometry method used here free system has revealed that reconstituted 2 1 heterodimers li- demonstrated that culture on fibronectin or type I collagen poten- gate both type I collagen and fibronectin (34). tiated intracellular cytokine expression (also type III collagen in In conclusion, we have demonstrated that interstitial ECM pro- by guest on September 27, 2021 the case of GM-CSF elaboration) in human ASM cells stimulated teins such as fibronectin and type I collagen, among the most for 24 h with maximally effective concentrations of IL-1␤ and widely expressed proteins in the lung, enhance release of IL-1␤- TNF-␣ in combination, overall numbers of positive-labeled cells stimulated eotaxin, RANTES, or GM-CSF from human ASM generally did not exceed 25% of the total population indicating cells. The underlying mechanism involves increased cytokine ␤ that only a subpopulation of human ASM cells could be induced to mRNA levels and appears to require multiple 1 integrins as well ␣ ␤ express eosinophil-activating cytokines under the current experi- as v 3 in the case of fibronectin. Although the relevance of these mental conditions. It is possible that further increases in the fre- events has not been demonstrated in vivo, their significance is quency of positive cells may have been detected at time points unlikely to be restricted to a culture environment. The fact that beyond 24 h, though inclusion of the protein transport inhibitor fibronectin and collagen are increased in asthmatic airways (5–9) brefeldin A either reduced labeling or had no effect and was there- adds weight to the hypothesis that such changes in the ECM en- fore omitted (D. Lai and S. J. Hirst, unpublished observations). vironment surrounding ASM cell may favor not only enhanced Technical difficulties associated with the high autofluorescence of survival (19) and growth (17, 18) responses, but also enhanced human cultured ASM cells (Fig. 4A) precluded use of double- or secretion during inflammation and remodeling. triple-labeling strategies to examine whether those cells expressing eotaxin also expressed RANTES or GM-CSF, and so it remains Acknowledgments possible that distinct multiple populations of cells exist with each We thank the thoracic surgeons, operating theater staff, and pathologists of capable of expressing one or more cytokine. Likewise, the current Guy’s and St. Thomas’ Hospitals, London, for the supply of human lung study did not ascertain whether culture on various ECM substrates tissue, and Dr. Maria B. Sukkar for invaluable initial advice in optimizing increased expression levels in a single subpopulation of cells al- the flow cytometry protocol for intracellular cytokine staining in human ready expressing or recruited additional cells. Nevertheless, the ASM cells. data raise an important question concerning the possible hetero- geneity of ASM cell cytokine expression, which has not previously References been addressed. 1. Davies, D. E., J. Wicks, R. M. Powell, S. M. Puddicombe, and S. T. Holgate. 2003. Airway remodeling in asthma: new insights. J. Allergy Clin. Immunol. Integrins, which comprise ␣␤ heterodimers are the principal re- 111:215. ceptors mediating multiple cell responses to ECM substrates (30). 2. Bousquet, J., A. M. Vignola, P. Chanez, A. Campbell, G. Bonsignore, and F.-B. Michel. 1995. Airways remodelling in asthma: no doubt, no more? Int. Arch. Their extracellular domains recognize short peptide sequences Allergy Immunol. 107:211. (e.g., Arg-Gly-Asp (RGD) found on some ECM proteins (fi- 3. Heard, B. E., and S. Hossain. 1973. Hyperplasia of bronchial smooth muscle in bronectin and vitronectin)), while the intracellular domains are in- asthma. J. Pathol. 110:319. 4. Ebina, M., T. Takahashi, T. Chiba, and M. Motomiya. 1993. Cellular hypertrophy volved in the formation of focal adhesion complexes and down- and hyperplasia of airway smooth muscles underlying bronchial asthma: 3-D stream signaling events. Our data provide new information that morphometric study. Am. Rev. Respir. Dis. 148:720. 2264 ECM BINDING INTEGRINS AND ASM SECRETION

5. Roche, W. R., R. Beasley, J. H. Williams, and S. T. Holgate. 1989. Subepithelial 20. Johnson, P. R., J. K. Burgess, P. A. Underwood, W. Au, M. H. Poniris, M. Tamm, fibrosis in bronchi of asthmatics. Lancet i:520. Q. Ge, M. Roth, and J. L. Black. 2003. Extracellular matrix proteins modulate 6. Roche, W. R. 1991. Fibroblasts and asthma. Clin. Exp. Allergy 21:545. asthmatic airway smooth muscle cell proliferation via an autocrine mechanism. 7. Laitinen, L. A., and A. Laitinen. 1995. Inhaled corticosteroid treatment and ex- J. Allergy Clin. Immunol. 113:690. tracellular matrix in the airways in asthma. Int. Arch. Allergy Immunol. 107:215. 21. Thomas, G. J., M. P. Lewis, S. A. Whawell, A. Russell, D. Sheppard, I. R. Hart, ␣ ␤ 8. Altraja, A., A. Laitinen, I. Vitenen, M. Ka¨mpe, B. G. Simonsson, S.-E. Karlson, P. M. Speight, and J. F. Marshall. 2001. Expression of the v 6 integrin promotes L. Håkansson, P. Venge, H. Sillastu, and L. A. Laitinen. 1996. Expression of migration and invasion in squamous carcinoma cells. J. Invest. Dermatol. 117:67. laminins in the airways in various types of asthmatic patients: a morphometric 22. Peng, Q., T. Matsuda, and S. J. Hirst. Signaling pathways regulating interleukin- study. Am. J. Resp. Cell Mol Biol. 15:482. 13-stimulated chemokine release from airway smooth muscle. Am. J. Respir. 9. Roberts, C. R., D. C. Walker, and R. R. Schellenberg. 2002. Extracellular matrix. Crit. Care Med. 169:596. Clin. Allergy Immunol. 16:143. 23. Plow, E. F., T. A. Haas, L. Zhang, J. Loftus, and J. W. Smith. 2000. Ligand 10. John, M., S. J. Hirst, P. J. Jose, A. Robichaud, N. Berkman, C. Witt, C. H. Twort, binding to integrins. J. Biol. Chem. 275:21785. P. J. Barnes, and K. F. Chung. 1997. Human airway smooth muscle cells express 24. Thyberg, J., and A. Hultgardh-Nilsson. 1994. Fibronectin and the basement mem- and release RANTES in response to T helper 1 cytokines. J. Immunol. 158:1841. brane components laminin and collagen type IV influence the phenotypic prop- 11. Chung, K. F., H. J. Patel, E. J. Faldon, J. Rousell, E.-B. Haddad, P. J. Jose, erties of subcultured rat aortic smooth muscle cells differently. Cell Tissue Res. J. A. Mitchell, and M. G. Belvisi. 1999. Induction of eotaxin expression and 276:263. release from human airway smooth muscle cells by IL-1␤ and TNF␣: effects of 25. Yamamoto, M., K. Yamamoto, and T. Noumura. 1993. Type I collagen promotes IL-10 and corticosteroids. Br. J. Pharmacol. 127:1145. modulation of cultured rat arterial smooth muscle cells from a contractile to a 12. Ghaffar, O., Q. Hamid, P. M. Renzi, Z. Allakhverdi, S. Molet, J. C. Hogg, synthetic phenotype. Exp. Cell Res. 204:121. S. A. Shore, A. D. Luster, and B. Lamkhioued. 1999. Constitutive and cytokine- 26. Seandel, M., K. Noack-Kunnmann, D. Zhu, R. T. Aimes, and J. P. Quigley. 2001. stimulated expression of eotaxin by human airway smooth muscle. Am. J. Respir. Growth factor-induced angiogenesis in vivo requires specific cleavage of fibrillar Crit. Care Med. 159:1933. type I collagen. Blood 97:2323. 13. Hallsworth, M. P., L. M. Moir, D. Lai, and S. J. Hirst. 2001. Inhibitors of MAP 27. Esnault, S., and J. S. Malter. 2003. Hyaluronic acid or TNF-␣ plus fibronectin kinases differentially regulate eosinophil-activating cytokine release from human triggers granulocyte macrophage-colony-stimulating factor mRNA stabilization airway smooth muscle. Am. J. Respir. Crit. Care Med. 64:688. in eosinophils yet engages differential intracellular pathways and mRNA binding 14. Ingber, D. E., L. Dike, L. Hansen, S. Karp, H. Liley, A. Maniotis, H. McNamee, proteins. J. Immunol. 171:6780. Downloaded from D. Mooney, G. Plopper, J. Sims, et al. 1994. Cellular tensegrity: exploring how 28. Wijelath, E. S., J. Murray, S. Rahman, Y. Patel, A. Ishida, K. Strand, S. Aziz, mechanical changes in the cytoskeleton regulate cell growth, migration, and tis- C. Cardona, W. P. Hammond, G. F. Savidge, et al. 2002. Novel vascular endo- sue pattern during morphogenesis. Int. Rev. Cytol. 150:173. thelial growth factor binding domains of fibronectin enhance vascular endothelial 15. Wilson, J. W., and X. Li. 1997. The measurement of reticular basement mem- growth factor biological activity. Circ. Res. 91:25. brane and submucosal collagen in the asthmatic airway. Clin. Exp. Allergy 29. Hirst, S. J. 2003. Regulation of airway smooth muscle cell immunomodulatory 27:363. function: role in asthma. Respir. Physiol. Neurobiol. 137:309. 16. Bousquet, J., P. Chanez, J. Y. Lacoste, I. Enander, P. Venge, C. Peterson, 30. Giancotti, F. G., and E. Ruoslahti. 1999. Integrin signaling. Science 285:1028.

S. Ahlstedt, F. B. Michel, and P. Godard. 1991. Indirect evidence of bronchial 31. Wehrle-Haller, B., and B. A. Imhof. 2003. Integrin-dependent pathologies. http://www.jimmunol.org/ inflammation assessed by titration of inflammatory mediators in BAL fluid of J. Pathol. 200:481. patients with asthma. J. Allergy Clin. Immunol. 88:649. 32. Chen, G., G. Grotendorst, T. Eichholtz, and N. Khalil. 2003. GM-CSF increases 17. Hirst, S. J., C. H. Twort, and T. H. Lee. 2000. Differential effects of extracellular airway smooth muscle cell connective tissue expression by inducing TGF-␤ re- matrix proteins on human airway smooth muscle cell proliferation and pheno- ceptors. Am. J. Physiol. 284:L548. type. Am. J. Respir. Cell Mol. Biol. 23:335. 33. Obara, M., and K. Yoshizato. 1997. A novel domain of fibronectin revealed by 18. Bonacci, J. V., T. Harris, and A. G. Stewart. 2003. Impact of extracellular matrix epitope mapping of a monoclonal antibody which inhibits fibroblasts-mediated and strain on proliferation of bovine airway smooth muscle. Clin. Exp. Pharma- collagen gel contraction. FEBS Lett. 412:48. col. Physiol. 30:324. 34. Weinreb, P. H., W. J. Yang, S. M. Violette, M. Couture, K. Kimball, 19. Freyer, A. M., S. R. Johnson, and I. P. Hall. 2001. Effects of growth factors and R. B. Pepinsky, R. R. Lobb, and S. Josiah. 2002. A cell-free electrochemilumi- ␤ extracellular matrix on survival of human airway smooth muscle cells. nescence assay for measuring 1-integrin-ligand interactions. Anal. Biochem. Am. J. Respir. Cell Mol. Biol. 25:569. 306:305. by guest on September 27, 2021