Proteomic Analysis of Clonal Interstitial Aortic Valve Cells Acquiring a Pro-calcific Profile Elisa Bertacco,|,† Renato Millioni,|,† Giorgio Arrigoni,‡ Elisabetta Faggin,† Laura Iop,§ Massimo Puato,† Lorenzo A. Pinna,‡ Paolo Tessari,† Paolo Pauletto,† and Marcello Rattazzi*,† Dipartimento di Medicina Clinica e Sperimentale, Dipartimento di Chimica Biologica and VIMM, Venetian Institute of Molecular Medicine, and Dipartimento di Scienze Cardiologiche, Toraciche e Vascolari, Universita` degli Studi di Padova, Italy Received July 2, 2010
Calcific degeneration represents the most frequent aortic valve disease observed in industrialized countries. Our aim is to study modifications in the cytosolic and membrane protein profile of aortic interstitial valve cells (VIC) acquiring a pro-calcific phenotype. We studied a clonal population of bovine VIC that expresses bone-related proteins (such as alkaline phosphatase [ALP]) and calcifies a collagen matrix in response to endotoxin (LPS) treatment. A proteomic analysis was performed on proteins extracted from cells treated for 12 days with LPS (100 ng/mL) versus control. We identified 34 unique cytosolic and 10 unique membrane-associated proteins showing significant changes after treatment. These proteins are involved in several cellular functions, such as chaperone-mediated protein folding, protein metabolism and transport, cell redox/nitric oxide homeostasis, and cytoskeletal organization. Reduced expression of proteins involved in NOS bioactivity (such as DDAH-1 and -2) suggested a role for the L-arginine/ADMA ratio in controlling VIC phenotypic profile. In accordance with this hypothesis, we observed that exposure of clonal cells to L-arginine prevented LPS-induced ALP expression and collagen calcification. In conclusion, we identified several proteins involved in structural, metabolic, and signaling functions that are significantly altered in aortic VIC acquiring a pro-calcific profile, thus giving new insights into the pathogenesis of aortic valve degeneration. Keywords: Aortic valve calcification • interstitial valve cells • differential proteomics • L-arginine • ADMA
Introduction Calcific degeneration represents the most frequent aortic valve disease observed in industrialized countries.1,2 Despite increased knowledge about risk factors and mechanisms of disease progression,3,4 no medical treatment is available to either prevent or slow calcium deposition within the valve leaflets. Nevertheless, in the past few years, a series of experimentally and clinically based evidence have shown that celldriven processes might be actively involved in vascular/valve calcification.3,5 These novel findings are challenging the classic view of the calcific aortic valve as an ineluctable, progressive phenomenon and could open a new scenario for the development of treatment strategies.6 Several in vitro and in vivo studies showed that interstitial aortic valve cells (VIC) from different animal species can acquire a pro-calcific profile following stimulation with inflammatory cytokines, modified lipids, reactive oxygen species (ROS), and other pathological factors.3,4 We and others demonstrated that the phenotypical heterogeneity observed among VIC resident in the normal valve * To whom correspondence should be addressed. Tel: +390498211867. Fax: +390498754179. E-mail:
[email protected]. † Dipartimento di Medicina Clinica e Sperimentale. ‡ Dipartimento di Chimica Biologica and VIMM. § Dipartimento di Scienze Cardiologiche. | These authors equally contributed to this study. 10.1021/pr100682g
2010 American Chemical Society
is accompanied by a different cellular propensity for the acquisition of a pro-calcific phenotype.7-9 In particular, after exposure to pathologic mediators a specific VIC subpopulation could be preferentially induced to differentiate into “osteoblastlike” cells and drive matrix mineralization.8 Hence, it appears that the description of molecular pathways and mediators controlling these complex phenomena might help to identify novel therapeutic targets. For this purpose, we performed a proteomic study on a specific clonal population of bovine VIC (BVIC), which has been previously shown to express bonerelated proteins (such as alkaline phosphatase [ALP] and osteocalcin) and to calcify a type-I collagen matrix in response to endotoxin treatment.8 The proteomic approach allowed us to identify several proteins whose expression was significantly altered in clonal BVIC acquiring a pro-calcific phenotype. Starting from the proteomic findings, we performed additional investigations that highlighted a role for cell redox/nitric oxide (NO) homeostasis in controlling VIC phenotypic profile.
Materials and Methods BVIC Isolation and Characterization. Primary aortic BVIC were obtained from explants of intact aortic bovine valve leaflets as previously described.8 Bovine hearts were obtained from a local slaughterhouse within 15 min from slaughter, the aortic leaflet was isolated, minced in fragments, and digested Journal of Proteome Research 2010, 9, 5913–5921 5913 Published on Web 09/08/2010
research articles with type-I collagenase, elastase, and soybean trypsin inhibitor. Fragments were collected and cultured in DMEM containing 4.5 g/L glucose plus FBS 20%, 100 U/mL penicillin, and 100 µg/mL streptomycin (all reagents were obtained from SigmaAldrich if not specified). BVIC spread from predigested fragments after 7-10 days. Collected cells were subjected to a procedure of cloning by using a limited dilution technique. Clonal cell expansion and subcultures were made using the same medium described above. Cloned BVIC Culture and Treatment. Selected clonal cells (passages 6-9) were seeded at density of 104 cells/cm2 in 100 mm Petri dishes (Falcon BD Biosciences). At confluence cells were treated with DMEM, 5% FBS, 100 U/mL penicillin, and 100 µg/mL streptomycin supplemented with LPS (100 ng/mL; E. coli) or control. The media was changed every third day, and after 12 days of treatment proteins were extracted to conduct two-dimensional gel electrophoresis (2-DE) analysis (see below). The same clonal cells were seeded on six-well plates (Falcon BD Biosciences) and treated upon reaching confluence with different combinations of LPS (100 ng/mL), L-arginine (5 mM, 50 mM, 100 mM), L-NAME (Nω-nitro-L-arginine methyl ester, 1 mM), ADMA (asymmetric dimethylarginine, 1 µM, 10 µM, 100 µM), cPTIO [2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide, 100 µM, Alexis Biochemicals]. Protein Extraction and 2-DE Separation. Before protein extraction, cells were washed three times with PBS. For each plate, 0.250 mL of lysis buffer, composed of TRIS-HCl 12 mM, DTT 1 mM, and a cocktail of protease inhibitors, was used. The lysate was collected and submitted to three cycles of freeze-thawing in liquid nitrogen and sonication in ice. A sample prefractionation was accomplished by ultracentrifugation of protein lysates (at 100000 RCF for 1 h at 10 °C). The pellet, constituting a membrane-enriched protein fraction, was solubilized in a specific buffer for hydrophobic proteins (urea 6 M, tiourea 2 M, Triton X-100 1% v/v, CHAPS 2% p/v, IPG buffer 0.5%). The supernatant (cytosol-enriched protein fraction) was concentrated through ultrafiltration (MicroconAmicon YM-3, Millipore Corporation). At the end of the procedure proteins were quantified by Bradford method. A constant amount of solubilized proteins (200 µg) was diluted to 450 µL using a solution of urea 8 M, CHAPS 2%, IPG buffer 0.5%, and DTT 1% for the cytosolic fraction (CF) and a solution of urea 6 M, thiourea 2 M, CHAPS 2%, IPG buffer 0.5%, and DTT 1% for the membrane fraction (MF). Isoelectric focusing was carried out on 24 cm long immobilized pH gradient strips providing a linear 4-7 pH range (GE Healthcare Life Sciences) for a total of about 55,000 Vh. The second dimension was run on 12% polyacrylamide gels, after strip equilibration in a buffer containing urea 6 M, glycerol 30%, SDS 2%, Tris-HCl 50 mM pH 8.8, and DTT 1% and then in the same buffer containing iodoacetamide 2.5% instead of DTT. Gel Image Analysis. 2-DE gels were stained with Coomassie Brilliant Blue G-250, and images were recorded using an Epson Expression 1680 Pro scanner (Seiko-EPSON Corp.) with 16 bit dynamic range and 300 dpi resolution. Gel image analysis was performed on a total of 20 2-DE gels (6 technical replicates for both control and LPS-treated CF, 4 technical replicates for both control and LPS-treated MF) by using the Proteomweaver software (Bio-Rad Laboratories). After completion of spot matching in 2-DE gels, the normalized intensity values of individual protein spots, expressed as a numeric value of optical density, were used to compare the protein quantitative levels between the two groups (treated vs untreated) in both the 5914
Journal of Proteome Research • Vol. 9, No. 11, 2010
Bertacco et al. cytosolic and the membrane fractions. Normalization was run automatically by the Proteomweaver software (Bio-Rad Laboratories), which computes relative spot volumes by setting total spot volume on a gel to 100%, thus making normalized spot intensities comparable among different gels without the need of an internal standard. Individual values obtained from technical replicates were then used for the quantitative analyses. A paired t test analysis was performed, and significance was accepted at p < 0.05. To provide high statistical confidence in the subsequent identification of proteins affected by the treatment, MS analysis was conducted only on spots whose difference in intensity between groups reached both statistical significance and at least 1.5 fold-change. Protein Identification by Mass Spectrometry. The spots of interest were excised from the gels, and protein digestion was performed in gel using sequencing grade modified trypsin (Promega, Madison, WI). Briefly, gel pieces were repeatedly washed with 50% acentonitrile/50 mM NH4HCO3 and then dried under vacuum. Then 8-10 µL of trypsin (12.5 ng/µL in 50 mM NH4HCO3) was added to each gel spot, and samples were incubated for 30 min at 4 °C. Digestion was carried out at 37 °C overnight. The peptides obtained after trypsin digestion were extracted from the gel with 3 changes (20-30 µL each time) of 75% acetonitrile/0.1% trifloroacetic acid (TFA). The final peptide mixtures were then dried under vacuum and finally resuspended with 4 µL of 20% acetonitrile/0.1% TFA. One microliter of sample solution was mixed with 1 µL of matrix solution (R-cyano-4-hydroxycinnamic acid, 5 mg/mL in 70% acetonitrile/0.1% TFA), and 0.8 µL of the final mixture was spotted onto a stainless steel MALDI target plate. Samples were analyzed using a MALDI-TOF/TOF 4800 mass spectrometer (Applied Biosystems, Toronto, Canada) operating in a data dependent mode: a full MS scan was followed by MS/MS scans on the 10 most intense peaks for each sample. MS/MS data were converted into MGF (Mascot Generic Format) files and searched using Mascot (Matrix Science, London, U.K.) against the entire Swissprot database (version 55.4; 385,721 entries). Enzyme specificity was set to trypsin with 1 missed cleavage using carbamidomethylcysteine as fixed modification. The tolerance of the precursor ion was set to 50 ppm while the tolerance of fragment ions was set to 0.3 Da. Proteins were considered as positively identified if at least 2 peptides with individual significant score (p < 0.05) were sequenced. The search was done also against the corresponding randomized database that did not return any positive identification under the same conditions (i.e., at least 2 peptides sequenced with individual significant score). A few proteins could not be identified because of the poor quality of the MS/MS data, therefore for these samples peptide mass fingerprint (PMF) identification was obtained. In these cases, Mascot search was done against the same database and using the same search parameters, with a peptide tolerance of 50 ppm. All the proteins and peptides identified are reported as Supporting Information (Supplemental Table 1 and 2). In the majority of the cases, the proteins were correctly identified as bovine proteins, although for few samples the search engine returned the identification of human proteins. For these cases, a manual inspection of the MS/MS spectra was carried out to confirm the good quality of the data and the correct identification. Western Blotting Analysis. To confirm some 2-DE/MS analysis results Western blotting was performed on different cell preparations obtained following the experimental design described above. Protein extracts were collected with the same
research articles
Proteome of Calcifying Valve Cells
Figure 1. ALP activity and calcium deposition in cloned BVIC treated with endotoxin. (A) A selected clonal BVIC population was treated with LPS (100 ng/mL) or control. After 12 days we observed a progressive increase of ALP activity. (*p < 0.01 vs control). (B) The same clonal cells were seeded on microfibrillar type-I collagen sponges and treated for 12 days with LPS (100 ng/mL), phosphate (Pi, 2.4 mmol/L) or LPS plus Pi. At the end of the treatment an extensive matrix mineralization could be documented by Von Kossa staining in the sponges treated with a combination of LPS and Pi [200X].
lysis buffer described for 2-DE samples, and the protein content was measured by using BCA protein assay kit (Pierce). Equal amounts of protein were separated by 12.5% SDS-PAGE followed by electrophoretic transfer to Immun-Blot PVDF membranes (Bio-Rad Laboratories). Blotted membranes were blocked for 1 h at room temperature with 5% nonfat dry milk in TBS (Bio-Rad Laboratories) and subjected to overnight incubation at +4 °C with antisuperoxide dismutase [Cu-Zn] (SOD1) antibody (1:1500, Abcam), anti-N(G),N(G)-dimethylarginine dimethylaminohydrolase 1 (DDAH1) antibody (1:2000, LifeSpan Biosciences), or anti-β-tubulin antibody (1:1000, Sigma-Aldrich). After incubation for 1 h at room temperature with HRP conjugated IgG (Dako), proteins were visualized by using a chemiluminescence reagent kit (SuperSignal West Pico Chemiluminescent Substrate, Pierce), according to manufacturer’s instructions. ALP Activity Assay. ALP activity was determined, as previously described,8 using a kinetic assay (Chema Diagnostica) that measure p-nitrophenol production. ALP activity data were normalized to the cells protein content (Pierce BCA Protein Assay). For determination of ALP activity in clonal cells seeded in collagen scaffolds (see below) we directly incubated sponge cryosections with the working solution used for detecting ALPconjugated antibodies (Red Alkaline Phosphatase Substrate Kit I, Vector Laboratories).
Figure 2. Representative 2-DE maps of clonal cell cytosolic and membrane proteins after treatment with LPS or control. (A) Clonal cells were treated for 12 days with LPS (100 ng/mL) or control. A total of 170 spots in the cytosolic fraction (CF) and 60 spots in the membrane fraction (MF) were separated by 2-DE and statistically analyzed. Significant modification (considered as both p < 0.05 and at least 1.5-fold change) was observed in 70 spots in CF and 12 spots in the MF. Spot numbers reported in the figure correspond to proteins significantly altered among groups and subsequently identified by MS analysis. (B) Enlargement of selected spots. (AnxA2: Annexin A2. DDAH1: dimethylarginine dimethylaminohydrolase 1. Hsp70: Heat Shock Protein 70. SOD1: superoxide dismutase [Cu-Zn]).
Calcium Deposition Assay. Clonal cells (2 × 106 cells/cm2) were statically seeded in bovine porous microfibrillar type-I collagen sponges (Davol) and treated with a combination of LPS (100 ng/mL) plus phosphate (Pi; final concentration 2.4 mmol/L) in the absence or presence of L-arginine (100 mM). After 12 days of treatment, with media changed every third day, collagen sponges were decalcified overnight in 0.6 M HCl. Calcium content was quantified colorimetrically using the o-cresolphtalein complexone method (Chema Diagnostica) and values were normalized for the dry weight of the sponges. Calcification of the collagen scaffold was also confirmed by histochemical analysis of matrix cryosections with von Kossa staining. Statistical Analysis. For the in vitro experiments data were expressed as mean ( SD. The statistical analysis was performed by using one way ANOVA followed by posthoc Fisher’s LSD test. Significance was accepted at p < 0.05. Statistical methods used for 2-DE gels data analysis are reported in the gel image analysis paragraph.
Results Isolation and Characterization of Calcifying BVIC Clonal Cells. Starting from the primary culture, several clones of BVIC were obtained using a limited dilution technique. For the Journal of Proteome Research • Vol. 9, No. 11, 2010 5915
research articles
Bertacco et al.
Table 1. Proteins Showing Significant Changes in BVIC Clonal Cells Treated for 12 Days with Endotoxin (C ) cytosol, M ) membrane)
spot
quantitative analysis (treated/control ratio ( SD)
Bos taurus Bos taurus Bos taurus Homo sapiens Bos taurus Bos taurus Bos taurus Bos taurus Homo sapiens Bos taurus Bos taurus Bos taurus Homo sapiens Bos taurus Bos taurus Bos taurus
C22 M13 C96 C12 C21 C120 C75 C32 C35 C48 C49 C58 C60 C78 C79 C155 C17 C144 M20 M26 M27 M32 M33 M22 C131 C1 M19 C29 C64 C110 C127 C40 C125 C55 C80 M46 C36 M40 C59
0.64 ( 0.16 0.62 ( 0.07 1.92 ( 0.075 0.27 ( 0.14 2.41 ( 0.7 0.40 ( 0.1 2.21 ( 0.8 1.83 ( 0.7 2.80 ( 1.3 3.72 ( 1.2 2.97 ( 1.5 2.05 ( 1.1 1.96 ( 0.7 9.53 ( 3.5 2.36 ( 1.2 2.60 ( 1.3 3.43 ( 0.8 10.12 ( 3.8 2.88 ( 1.74 2.75 ( 1.3 1.95 ( 1.2 0.53 ( 0.06 0.52 ( 0.01 0.63 ( 0.14 0.22 ( 0.06 0.61 ( 0.25 2.08 ( 1.1 2.52 ( 1.4 0.57 ( 0.16
