Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 http://www.translational-medicine.com/content/11/1/260 RESEARCH Open Access Secreted proteins from carotid endarterectomy: an untargeted approach to disclose molecular clues of plaque progression Silvia Rocchiccioli1*, Gualtiero Pelosi1, Silvia Rosini1, Michele Marconi2, Federica Viglione1, Lorenzo Citti1, Mauro Ferrari2, Maria Giovanna Trivella1 and Antonella Cecchettini1,3 Abstract Background: Atherosclerosis is themain cause of morbidity and mortality inWestern countries and carotid plaque rupture is associated to acute eventsand responsible of 15-20% ofall ischemic strokes. Several proteomics approaches have been up to now used to elucidate themolecular mechanisms involved inplaque formation as well as to identify markers of pathology severity for earlydiagnosis or target of therapy. The aim of this study was to characterize the plaque secretome. The advantage ofthis approach is thatsecretome mimics the in vivo condition and implies a reduced complexity compared to thewhole tissue proteomics allowing the detection of under-represented potential biomarkers. Methods: Secretomes from carotid endarterectomy specimensof 14 patients were analyzedby a liquid chromatography approach coupled withlabelfree mass spectrometry.Differential expression of proteins released from plaques and from their downstream distal side segments were evaluatedin each specimen. Resultswere validated byWestern blot analysis and ELISA assays. Histology and immunohistochemistry were performed to characterize plaques and to localise the molecular factors highlighted by proteomics. Results: A total of 463proteins were identified and 31 proteins resulted differentially secreted from plaquesand corresponding downstream segments. Aclear-cut distinction in thedistribution of cellular- and extracellular-derived proteins, evidently related to the higher cellularity ofdistalside segments, was observed along thelongitudinalaxis ofcarotid endarterectomy samples. The expressionsof thrombospondin-1, vitamin D binding protein, and vinculin, as examplesof extracellular and intracellular proteins, were immunohistologically compared betweenadjacent segments and validatedby antibody assays. ELISA assays of plasma samples from 34 patients and 10 healthy volunteers confirmeda significantly higher concentration ofthrombospondin-1 and vitamin D binding protein in atherosclerotic subjects. Conclusions: Taking advantage of theoptimizedworkflow, a detailed protein profilerelated to carotid plaque secretome has been produced which may assist and improvebiomarkerdiscovery of molecular factors in blood. Distinctive signatures of proteins secreted by adjacent segments ofcarotid plaques were evidenced and they may helpdiscriminating markers of plaque complication from those ofplaque growth. Keywords: Atherosclerosis,Carotid plaques, Proteomics, Secretome *Correspondence:[email protected] 1NationalResearchCouncil,InstituteofClinicalPhysiology,ViaMoruzzi,Pisa, Italy Fulllistofauthorinformationisavailableattheendofthearticle ©2013Rocchicciolietal.;licenseeBioMedCentralLtd.Thisisanopenaccessarticledistributedunderthetermsofthe CreativeCommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse, distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited. Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page2of15 http://www.translational-medicine.com/content/11/1/260 Background interest for discovery of novel biomarkers and it Atherosclerosisisachronicimmune-inflammatorydisease representsawaytogainknowledgeof biologicalmech- of the wall of medium- and large-sized arteries and it is anisms[10,11]. the main cause of morbidity and mortality in Western An optimal culture set-up of arterial secretome in countries due toacute cardiovascular eventssecondaryto order to reduce plasma contamination and detect low partial or total thrombotic obstruction of vessel lumen abundance proteins is a recent achievement [12]. Also [1,2]. Rupture of carotid atherosclerotic plaques leads to recently, secretomes from thromboendartectomy speci- atherothrombosis that accounts for approximately 20% of mens were exploited to select nine secretome-specific allstrokes,aleadingcauseofdisability,andthethirdmost antibodies that allowed the immuno-purification and commoncauseofdeath[3]. successive identification of 22 proteins. Among them, Recently, a great interest has been focused on the junction plakoglobin has been suggested as a potential identification of tissue markers of atherosclerosis, at biomarkerofatherosclerosis[13]. genomics, transcriptomics [4] or proteomics levels. The A main issue in molecular studies of vascular path- majority of proteomics studies were performed on ology is the cellular and extracellular heterogeneity of plaque extracts analyzed by two-dimensional electro- the plaque and of the adjacent wall where multiple com- phoresis followed by MALDI TOF mass spectrometry ponents (calcium, lipids, collagens and others) and cell [5,6], but, with this approach, mainly constitutive pro- types (vascular smooth muscle cells (VSMCs), endothe- teins are identified and potential, usually low-expressed lial cells (EC), macrophages and other inflammatory biomarkers may not be detected. To overcome these cells) are present, all contributing to plaque progression limitations, Olson et al. [7] applied two-dimensional dif- and/orcomplication. ferential gel electrophoresis (2D-DIGE) in combination Several previous and recent papers have highlighted a with tandem mass spectrometry (MS/MS) to compare close link between the longitudinal distribution of mech- protein distribution in an intra-individual study of se- anical forces (flow shear stress and plaque wall stress) verely atherosclerotic segments of internal carotid artery and corresponding morphological features (cell distribu- compared to partially preserved segments. In this way, tion and type) along plaque and its distal side [14-18]. In they identified 19 proteins with a differential distribution particular, the low flow shear stress in downstream side along the artery. Another interesting approach was is associated to atherosclerosis progression with in- exploited by Martinet et al. [8] who screened human ca- creased VSMCs and macrophages, whereas the high rotid endarterectomy (CEA) specimens using 823 mono- plaque wall stress in the upstream area is associated to clonal antibodies with Western array technology and cap rupture of vulnerable lesions and increased expres- were able to identify 7 differentially expressed proteins. sion of proteolysis and apoptosis markers [14]. These re- The potential use of tissue proteomics as a tool for ports support the opinion that carotid plaque and its clinically useful biomarker discovery has been recently corresponding adjacent distal side may retain distinctive prospected by de Kleijn et al. [9] who found that protein signatures: therefore, differential expression of osteopontin expression in carotid plaques and in blood proteins released by plaque-containing upstream seg- is predictive for atherothrombotic events but it does not ment(P)andbyitsdownstream distalside(DS)segment correlatewithvulnerability features. hasbeen evaluatedineachCEA specimen. An alternative proteomics approach consists of the Aims of the study were: (a) to characterize the overall analysis of proteins secreted by explanted arteries and atherosclerotic carotid secretion with an untargeted ap- wasfirstsuggestedbyDuranetal.[10].Thegreatadvan- proach able to reconstruct a complete protein map; (b) tage of tissue secretome studies is that secretome to evaluate the differences in secretomes from central mimics the in vivo condition and implies a reduced plaque and corresponding distal side segments, as puta- complexity compared to serum/plasma or entire tissue tive areas of plaque complication and stable growth re- proteomics as well as a much narrower protein dynamic spectively; (c) to conduct a secretome-assisted plasma range, thus allowing the detection of under-represented analysis of some differentially expressed proteins to be potentialbiomarkers. exploited as markers of disease for non-invasive risk In fact, in biomarker discovery, plasma represents the assessment. sample of choice since it shows traces of all biological Secretomes were analyzed by a liquid chromatography events and, moreover, it can be easily and non-invasively approach coupled with mass spectrometry (LC-MS/MS). collected.Onthecontrary,inproteomicsstudies,plasma Label-free MS/MS based quantification was performed proteome is hampered by major limits such as the high and validated by Western blot analysis and ELISA assay. dynamic range of plasma proteins and a great biological Histology and immunohistochemistry were performed to variability. For all these reasons, the analysis of proteins characterise CEA specimens and to localize the molecu- that are secreted by tissues into circulation has attained lar factors highlighted by proteomics. ELISA assays of Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page3of15 http://www.translational-medicine.com/content/11/1/260 thrombospondin-1 and vitamin D binding protein were Symptomatic patients were classified as CEA candi- performed in plasma samples of 34 CEA candidates and dateswho presentedwithTIA orstrokewithin 6months 10healthycontrols. before surgery. No differences between symptomatic and asymptom- atic patients were found for any of the reported parame- Methods ters. Smokers were classified as individuals who smoked Ethicsstatement at least three cigarettes per day at the time of analysis, Clinical data were derived from medical records, follow- past smokers had quit smoking for at least 6 months, ing data security guidelines and declaration of Helsinki. and no-smokers were individuals who had never All subjects gave written informed consent to participate smoked. Smoker patients were the combined group of to the study in accordance with the Ethics Committee thepast-andthe current-smokers. requirements. The ethical approval for this study was Stenosis on ICA was ≥ 70% at carotid Echocolor granted by Ethics Commission (Clinical Trials Ethics Doppler in all patients, whose CEA specimen secretomes Commission)ofPisaUniversityHospital. were analysed (mean 80±9%), without significant differences between asymptomatic and symptomatic (79±8vs82±10%). Patientclinicalcharacterization Human internal carotid artery (ICA) specimens and pre Tissueprocessingandsecretomepreparation surgery plasma samples from 10 males and 4 females The arterial tissue was cultured according to de la undergoingCEAforsymptomatic orasymptomatic sten- Cuesta et al. [12] with modifications. CEA specimens osis ≥70% were obtained from theVascular Surgery Unit (n=14) from surgery were transported in PBS on ice to of Pisa University Hospital and immediately processed at the laboratory. Each specimen was crosscut under theInstituteofClinicalPhysiology.Additional20plasma macroscopic examination to separate P from DS seg- samples of CEA candidates and 10 plasma samples of ment. After repeated washing in PBS to remove blood healthyvolunteerswere collectedattheVascularSurgery traces, samples were incubated in 6-well plates in 2 ml Unit of Pisa University Hospital and at the Institute of of Eagle's Minimal Essential Medium (Sigma-Aldrich) ClinicalPhysiology respectively. supplemented with penicillin and streptomycin, without Patient characteristics and database information on FBS and phenol red at 37°C in a humidified atmosphere smoking habit, hypercholesterolemia and diabetes are of 5% CO . After 24 h the culture medium was 2 summarizedinTable1. harvested, centrifuged at 300× g for 10 min and stored Table1Clinicalcharacteristicsofpopulation Feature Value(±SD) Value(±SD) Value(±SD) CEApatientsfor CEApatientsforplasma Healthyvolunteers secretome(N=14) validation(N=34) (N=10) Gender Females(n) 4 10 4 Males(n) 10 24 6 Age(years±SD) 72±9 74±7 70±5 Symptomatic(n) 6 11 - Typeofevent Stroke(n) 2 3 - TIA(n) 4 8 - Asymptomatic(n) 8 23 - StenosisofICA(%±SD) 80±9* 79±8* - Diabetes(n) 6 9 - Hypertension(n) 10 27 1 Smoking(n) 4 13 1 Statintreatment(n) 10 21 - AntiaggregantTherapy(n) 14 32 - Hypercholesterolemia(n) 4 18 2 Categoricalvaluesaregivenasnumberofpatients(n). Continuousvariablesaregivenasmean±SD. TIA:TransientIschemicAttack.ICA:InternalCarotidArtery. *StenosiswasmeasuredbycarotidEchoColorDoppler. Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page4of15 http://www.translational-medicine.com/content/11/1/260 at −80°C until analysis. Samples were concentrated by (Olympus D70 camera) interfaced to Olympus Cell Sens centrifugal devices Amicon Ultra-3 (Millipore) following Dimension software for image acquisition and morpho- the manufacturer's recommendations. Protein concen- metric analysis. Quantitative analysis of antibody stain- trationwasdeterminedbybicinchoninic acid(Pierce). ing within the lesional area of each sample was carried out onseveralmicroscopic fieldsofconsecutivesections, Histologyandimmunohistochemicalcharacterization digitized at the same light source settings, by double After 5 to 7 day fixation in 5% buffered formalin, P and observer colour thresholding: reproducibility and statis- DSsamples were incubated indecalcifying solution (Bio- tical significance of results (expressed as% positive, dark optica 05-M0300 ), washed and dehydrated in ascending brown pixels of the entire plaque area) as well as alcohols series and embedded in paraffin: transverse localization of different antibodies on the same region of 5 μm thick consecutive sections were cut from each adjacent sections wasthusaccomplished. paraffin block by a rotary microtome (Microm HM Carotid plaques were classified according to Stary’s 300, Bio-optica). Haematoxylin and Eosin (H&E) and stages for atherosclerosis, American Heart Association Masson’s trichrome stains were used for histologic ana- CommitteeonVascular Lesions [19,20]. lysis. For immunohistochemistry, sections were placed on positively charged slides, de-paraffinized, rehydrated Reduction,alkylationanddigestion and washed in distilled water; after incubation in H O 100 μL of 40 mM ammonium hydrogen carbonate 2 2 at room temperature, antigen retrieval was accomplished (pH=8) were added to 100 μL of secretome (1 μg/μL (citrate buffer pH6 in microwave for 10 min at 500W) concentration). Reduction was obtained by adding 1 μL and then sections incubated with diluted normal of 1 M dithiothreitol to each sample, with an incubation blocking serum. Primary antibodies anti-αSMA (alpha of 20 min at 80°C. For alkylation, 20 μL of 100 mM smooth muscle actin, clone 1A4 ADB, Serotec) as a iodoacetamide were added to the samples and incubated SMC phenotype marker, anti-CD68 (mouse, clone PG- for 30 min at 37°C. Digestion was performed incubating M1, Thermo Scientific, diluted 1:150) as a macrophage overnight with 8 μL of trypsin solution (0.25 mg/mL) phenotype marker, anti-vinculin (goat polyclonal antibody at37°C. diluted 1:100, Santa Cruz Biotechnology, Inc.) and anti- thrombospondin-1 (mouse polyclonal antibody diluted LC-MS/MSanalysis 1:100, Santa Cruz Biotechnology, Inc.) were applied over- Chromatographic separation of peptides was performed nighttotheslidesina4°Chumidchamber. using an Ultimate 3000 nano-HPLC system (LC Pack- Following30minbiotinylatedsecondaryantibodyand30 ings, DIONEX, USA). 100 μL of filtrate were added to a minVectastainEliteABCreagentincubationinperoxidase solution composed by 2% ACN and 0.1% formic acid up substrate solution (DAB), slides were counterstained to 200 μL of final volume. The loading pump pre- with Mayer’s haematoxylin for 1 min. and mounted concentrated the sample in a pre-column cartridge (Neo-Entellan Merk). Antibody binding to cells and/or to (PepMap-100 C18 5 mm 100 A, 30 mm id × 5 mm). extracellularcomponentsisvisibleasbrownordarkbrown Chromatographic separation of peptides was performed stain (DAB), negative cells are stained blue (haematoxylin using a C18 PepMap-100 column (15 cm × 75 mm id, counterstain). LC Packings DIONEX) equilibrated at 45°C with a solv- The same steps had been previously applied to a pool ent A (water/acetonitrile 98/2 vol/vol, 0.1% formic acid) of CEA specimens (from symptomatic and asymptom- at a flow rate of 300 nL min-1. Runs were performed atic patients) which were processed en bloc and longitu- under 60 min linear gradient from 10 to 45% of solvent dinally, instead of transversely, sectioned by microtome: B (water/acetonitrile 2/98 vol/vol, 0.1% formic acid) microscopic examination of the obtained serial sections followed by 10 min of a purge step at 95% of B before a helped to guide and standardise the splitting of fresh 20 min re-equilibration step to the starting conditions. CEA samples, subsequently collected for secretome, into The column was directly coupled to TripleTOF™ 5600 PandDSsegments. System (AB SCIEX, Toronto, Canada), equipped with a Thrombospondin-1 and vinculin double immunostain- DuoSpray™ionsource(ABSCIEX,Toronto, Canada). ing was performed using Vectastain Elite ABC reagent Peptides eluted from chromatography were directly incubation in phosphatase (AP) substrate solution and processed using TripleTOF™ 5600 mass spectrometer Permared-AP stain for thrombospondin-1 and DAKO (AB SCIEX, Toronto, Canada). The mass spectrometer LSAB system HRP and subsequent peroxidase substrate was controlled by Analyst® 1.6.1 software (AB SCIEX, solution (DAB)incubation for vinculin. Toronto, Canada). For positive ionization,ionsource pa- All sections from each carotid segment were examined rameters were the following: the spray voltage was 3 kV, under a light microscope (Olympus BX43) at 4× to 40× source temperature 150°C with curtain gas set at 25, original magnification and digitized by a video system GS1 10 and GS2 0 psi nitrogen flow. For information Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page5of15 http://www.translational-medicine.com/content/11/1/260 dependent acquisition (IDA) analysis, survey scans were (Bio-Rad). Photographic films were scanned and the acquired in 250 ms and 25 product ion scans were col- pixel intensities of bands was measured subtracting the lected if exceeding a threshold of 125 counts per second pixel intensity of the background and all the signals (counts/s). Total cycle time was fixed to either 1.25 s. were recorded as Optical Density (O.D.). All the O.D. Four time bins were summed for each scan at a pulser were reported in graphs and the comparative analyses frequency value of 11 kHz through monitoring of the 40 of the different levels of expression of vinculin and GHz multichannel TDC detector with four-anode/chan- thrombospondin-1 between P and DS samples was done nel detection. Dynamic exclusion was set for 1/2 of peak integratingtheresultingareaundercurvebyusingOrigin width (∼8 s), and then the precursor was refreshed off 7.0(Originlab). theexclusionlist. ELISAassays Proteinidentificationandlabel-freecomparativeanalysis Dosage by double-antibody sandwich enzyme-linked im- MS/MS data were processed with ProteinPilot™ Software munosorbent assay (ELISA) was performed on vitamin (AB SCIEX, Toronto, Canada), using the Paragon™ and Dbindingprotein andthrombospondin-1. Pro Group™ Algorithms and SwissProt 2011 as protein ELISA kits were used and reagents were prepared fol- database for Homo Sapiens. The false discovery rate lowing the manufacturer's manual. Briefly, for vitamin D (FDR) analysis was done using the integrated tools in binding protein (Uscn Life Science Inc, Wuhan, China) ProteinPilot software and a confidence level of 95% was the calibration curve ranged between 10 and 0.156 ng/ set. mL (considering a dilution for plasma samples of The statistical comparative analysis was performed 200000-fold) and the calibrator diluent was used as the using MarkerView™ Software 1.2.1 (AB SCIEX). The ion zero standard. All standards and samples were assayed chromatograms of high confidence peptides identifiedby induplicates. ProteinPilot were extracted using PeakView™ Software For thrombospondin-1 assay (Bio Medical Assay, and then MS peak areas and identifications were Beijing, China) the calibration curve ranged between 20 imported into MarkerView™ Software. Normalization of and0.312ng/μL.Plasma sampleswerediluted100fold. the total plaque size was obtained using a global The ELISA assay for Vitamin D binding protein normalization of profiles (total protein content) using (Quantikine® ELISA Human Immunoassay- USA & MarkerView1.2software. Canada R&D Systems, Inc.) was performed also on Principal Component Analysis (PCA) was performed secretome samples without dilution. The calibration in order to evidence groupings among the data set. All curverangedbetween1000ng/mland15.5ng/ml.Theab- profile areas were normalized. The two groups (P and sorbance was read at 450 nm with a Fluorstar (Omega) DS) were compared with t-test using a threshold of 95% microplatereader(MolecularDevices,Sunnyvale,CA). (P value=0.05)andfoldchange>2. Statisticalanalysis Westernblotanalysis Principal component analysis (PCA) was conducted on Secretomes were run on a 10% SDS-PAGE, separated mass spectrometric data of samples using Marker View proteins transferred onto a nitrocellulose membrane 1.2software. (Amersham) using a wet transfer system (Biorad). Mem- Student’stTestwasusedasstatisticalparameterbetween branes were blocked with 3% BSA in TBST for 1 h at the means of continuous variables to determine significant room temperature. Primary and secondary antibodies differences between categories of mass spectrometric data. were diluted in 3% BSA in TBST. All primary antibodies Pvalue<0.05wereconsideredsignificanttovalidatediffer- were incubated overnight at 4°C. HRP-conjugated encesbetweencategoriesandfoldchange>2. secondary antibodies were incubated for 1 h at Statistical analyses of other data were conducted using room temperature. The following antibodies were used: Origin 7.0 software. Data are expressed as the mean ± anti-vinculin goat polyclonal antibody (1:500) (catalog SD. Differences between the means of the 2 continuous number sc-7648, Santa Cruz Biotechnology, Inc), anti- variables were evaluated by the Student’s t Test and re- thrombospondin-1 mouse monoclonal antibody (1:200) sults accepted when t Test>95% and P value<0.05. (catalog number sc-73158, Santa Cruz Biotechnology, Paired t test was used for quantitative immunohisto- Inc), and anti-α tubulin mouse polyclonal antibody chemistryandELISA assay. (1:5000) (catalog number T 6074, Sigma). Chemi- luminescence was detected with ECL™ detection kit Ingenuitypathwayanalysis(IPA) (AmershamBiosciences,Uppsala,Sweden). IPA(http://www.ingenuity.com/products/pathways_analysis. Densitometric quantification of photographic films was html) was performed on a restricted and selected number performed using Quantity One 1-D Analysis Software of identified proteins (n=56) from the secretomes of CEA Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page6of15 http://www.translational-medicine.com/content/11/1/260 specimens. The subset of chosen proteins was constituted haemorrhage/thrombi. The DS segments, although free of of 34 extracellular matrix proteins, 9 plasma membrane grossly evident changes, showed small lesions at histologic proteins and 13 intracellular proteins. This subset of pro- examination, with atherosclerotic features similar to those teins was related to five function/disease pathways: migra- classified as Stary’s types III and IV. Proximal P segments, tion of cells (n=39), proliferation of smooth muscle cells ontheotherhand,showedgrosslyevidentfibrolipidicand/ (n=9), vascular disease (n=11), atherosclerotic lesion (n=8) or fibrocalcific, mostly complicated plaques classified as and angiogenesis (n=22). Colored points were used to typesVandVI(Table2)[19,20]. evidencedifferentiallyreleasedproteinsbetweenPandDS. Masson’s trichrome stain, αSMA and CD68 immuno- staining representative of P and DS segments are shown Results in Figure 2; the quantitative analyses for αSMA and Morphologicalcharacterization CD68 are also reported, demonstrating a significantly Microscopic examination of longitudinal sections of un- higher content of VSMCs and CD68 positive macro- divided CEA specimens evidences a plaque core phages in the distal side as compared to central plaque containing area (P) comprised of typical components region(Figure 2). (lipid-rich necrotic core, fibrous cap, cholesterol clefts, large calcium deposits, intraplaque haemorrage/ fresh Secretomeanalysisofendarterectomyspecimens thrombi) and the distal side segment (DS) with P and DS segments from the 14 enrolled patients were fibromuscular tissue, small extracellular lipid and cal- incubated for 24 h in serum free medium, secretomes cium deposits. Fresh CEA specimens for secretome were were collected, protein digested and peptides fraction- subsequently divided to separate central plaque P from ated by reverse phase chromatography and analyzed by itsdistalsideDS,asdepictedinFigure 1. 5600TripleTOF massspectrometer. SectionsfromallCEAspecimens,eachcutintwo(Pand Using this LC-MS/MS approach, 463 proteins were DS)wereexaminedandclassifiedforthepresenceofhisto- identified with a Protein Score (Confidence) > 95% and logic hallmarks of atherosclerosis, including lipid-rich using local false discovery rate analysis >1% as stringent necrotic core, cholesterol clefts, fibrosis, calcium, thin criterion to avoid false positives (Additional file 1: Table fibrous caps, macrophage infiltrate, neovessels, intraplaque S4 and Additional file 2: Table S5). The entire list of Figure1LongitudinalsectionofanundividedCEAspecimenatarterycenterlinelevelshowingthecuttinglineforPandDS segmentsseparation.PrevalentcomplicatedplaquefeaturesofP(lipid-necroticcore,calciumdeposits,fibrosisandhaemorrage)andmilder changesofitsdownstreamsideDS(prevalentVSMCsandcollagencomponent,smalllipidandcalciumdeposits)areevidencedbyMasson’s trichromeandα-SMAimmunostaining(fromtoptobottom).Originalmagnification2×,insets10×. Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page7of15 http://www.translational-medicine.com/content/11/1/260 Table2HistologiccharacterizationofPandDSsamples considering their localization in cytosol, membrane, organ- P(%specimens) DS(%specimens) elle,cytoskeletonandothers(Figure3A). Thesecretion potentialoftheseidentified proteins was TFCA(1) 33 0 computed by submittingthem to SecretomePserver that Ca(2) 65 10 predicts protein secretion route on the basis of the NC-Chol(3) 75 20 presence of the signal peptide, responsible for endoplas- He(4) 35 0 mic reticulum addressing. The non-classical secretion (1)TFCA=thinfibrouscap,(2)Ca=calcification,(3)NC-Chol=necroticcore through multi vesicular bodies was also determined by cholesterolclefts. SecretomeP usingspecific databanks[21](Figure 3B). (4)He=intraplaquehemorrhage/thrombus. proteins reported in supplementary tables was analyzed in the context of the published literature and firstly Comparativeanalysisbetweenplaqueanditsdistal grouped as extracellular and intracellular proteins. Using counterpartsecretomes Uniprotdatabase(www.uniprot.com),extracellularproteins In order to evaluate differentially released proteins be- were further subdivided in ECM proteins and ECM- tween P and DS specimens, MarkerView 1.2 software associated proteins. The intracellular ones were grouped was used and 31 proteins were found differentially Figure2Left:PandDSsectionsstainedwith(a)Masson’strichromestain,(b)α-SMAand(c)CD68immunostain(original magnification2×,insets10×).PisatypeVIplaquewithanecroticcore,calcifications,intraplaquethrombusandathinfibrouscapcomposed ofcollagenandαSM-actinpositivecells;minorintraplaqueCD68positivestainingispresentandα-SMA-actinpositivecontractileVSMCsarealso visibleontheouterborderofthespecimen(oppositetothelumen)duetosurgicalcleavageofplaquefromtheintacttunicamedia.InDS,a muchthinnerlesion(equivalenttoStarytypeIII)withextensiveα-SMAandCD68cellularpositivityispresent.Right:Fromtoptobottom. AveragevaluesandSDoflesionalarea(mm2),intralesionalα-SMAandCD68stainpositivity(%oflesionalarea)ofPandDSsegmentsofall specimens’sections:asmallerlesionalareaandamarkedlyhigherpositivityofbothcellularmarkersinthedistalsideoftheplaqueisevident (*p<0.05,**p<0.02pairedt-test). Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page8of15 http://www.translational-medicine.com/content/11/1/260 Figure3PiechartsofthetotalidentifiedproteinsfromPandDS.A)Identifiedproteinsaredividedonthebasisoftheirlocalizationin intracellularorextracellularspacebasedonGeneOntology.B)IdentifiedproteinsareevaluatedwithSecretomePsoftwaretocomputetheir secretionpotential.Theyweredividedin1)classicallysecreted2)not-classicallysecretedand3)not-predicted. secreted. These proteins were classified into two cat- stain is confined to the plaque lipid-necrotic core and egories: cellular (n=17) and extracellular proteins (n=14) surroundingfibrouscap. (Table 3). The putative role of each factor in atheroscler- Quantitative results derived from single immunostain- osis and cardiovascular disease was reported in the table ing of both markers on sections of all CEA specimens as suggested by the literature, in order to underline confirm the significantly higher (6.4±5.0 vs 3.4±3.0 a possible correlation of their dis-regulation to the P<0.02) vinculin positivity and the much lower (0.8±0.8 pathology. vs 7.3±6.0 P<0.02) thrombospondin-1 positivity in the Nine out of the 14 extracellular proteins were found lesion area of DS as compared to corresponding P over-expressed in P secretome, whilst cellular proteins segments. were found over-expressed in DS secretome, with the exception of smoothelin, which is a recognized marker ELISAassayofplasmasamples of VSMC contractile phenotype [23]. Of these proteins Dosage by double-antibody sandwich ELISA assays of eighteen have been already reported in serum/plasma thrombospondin-1 and vitamin D binding protein were and listed in Human Protein Reference database (www. performed in plasma samples of 34 CEA patients and 10 hprd.org). Origin 7.0 was used to represent the modula- age-matched healthy controls. As reported in Table 1, 11 tion of the differentially released proteins and the most patients were symptomatic while 23 were asymptomatic. interesting are reported as representatives of the results Plasma levels of thrombospondin-1 and vitamin D bind- (Figure 4). ing protein were significantly higher in atherosclerotic patients than in healthy subjects (thrombospondin-1: ValidationbyWesternBlotandELISAassay 71±20 vs 48±9 ng/mL P value= 0.02; Vitamin D binding Western Blot analysis and also ELISA assay were applied protein: 205±133 vs 76±8 ng/mL P value= 0.05), while to secretomes in order to validate the differential release no significant differences were observed between symp- observed by mass spectrometric analysis. Vinculin and tomatic and asymptomatic patients (thrombospondin-1: thrombospondin-1 were selected for Western blot valid- 67±14 vs 72±20 ng/mL P value=0.72; vitamin D binding ation (Figure 5A), while vitamin D binding protein was protein: 234±137 vs 192±122 ng/mL P value=0.69) selected forELISA assay(Figure5B). (Figure 7). Immunohistochemistryquantification Discussion Double immunostaining of anti-vinculin and anti- Several proteomics approaches have been carried out on thrombospondin-1 antibodies on P and DS sections of a in order to clarify the mechanisms of atherogenesis as CEAspecimenisshowninFigure6.Anti-vinculinantibody well as to search for plaque presence and severity. The selectively stains lesional cells whereas thrombospondin-1 majority ofthesestudiesconcerncarotid atherosclerosis. Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page9of15 http://www.translational-medicine.com/content/11/1/260 Table3DifferentiallyreleasedproteinsbyCEAspecimens Localization Proteinname(1) Gene Roleinatherosclerosis(2) P/DS Pvalue Plasma name (3) (4) (5) Cellular Alpha-actinin1 ACTN1 Focaladhesion ▼ 6.70E-04 ☑ Integrin-linkedproteinkinase ILK Focaladhesion ▼ 1.10E-04 ☑ Actin-relatedprotein2/3 ARPC3 Focaladhesion ▼ 1.20E-04 ☑ FilaminA FLNA Focaladhesion ▼ 1.35E-05 ☑ PDZandLIMdomainprotein1 PDLI1 Focaladhesion ▼ 1.34E-03 ☑ PDZandLIMdomainprotein7 PDLI7 Focaladhesion ▼ 5.70E-04 ☑ Vimentin VIM Focaladhesion ▼ 2.02E-03 ☑ Vinculin VCL Focaladhesion[12] ▼ 2.80E-04 ☑ Cysteine-richprotein2 CRIP2 SMCdifferentiation[22] ▼ 1.10E-04 Calponin1 CNN1 SMCdifferentiation ▼ 1.90E-04 Calponin2 CNN2 SMCdifferentiation ▼ 2.57E-03 ☑ Calponin3 CNN3 SMCdifferentiation ▼ 1.10E-04 Transgelin TAGL SMCdifferentiation ▼ 3.00E-03 Tropomyosin-1alpha TPM1 SMCdifferentiation ▼ 2.63E-03 ☑ Tropomyosinbeta TPM2 SMCdifferentiation ▼ 1.14E-03 ☑ Smoothelin SMOO SMCdifferentiationandplaque ▲ 3.40E-04 ☑ stability[23] Vascularcelladhesionprotein1 VCAM1 Monocyterecruitment ▼ 1.39E-02 ☑ (VCAM-1) Extracellularmatrix Collagenalpha1(III) COL3A1 Plaquestability ▲ 1.70E-04 (ECM) Collagenalpha1(VI) COL6A1 Plaquestability ▲ 2.40E-04 Collagenalpha1(XIV) COL14A1 Atherogenesis ▲ 1.05E-03 Collagenalpha1(XVI) COL16A1 Atherogenesis ▲ 2.07E-03 Versican VCAN Atherogenesis[24] ▲ 1.45E-05 Lumican LUM Plaquerupture[24] ▲ 4.07E-02 ☑ Fibulin2 FBLN2 Cellmigration[25] ▲ 9.00E-03 Emilin1 EMILIN1 Elasticlamellaeintegrity[26] ▲ 1.82E-03 ☑ Periostin POSTN Angiogenesis ▲ 2.07E-03 Thrombospondin-1 THBS1 Angiogenesis[27] ▲ 1.71E-02 ☑ VitaminDbindingprotein GC Acuteevents[28,29] ▲ 1.93E-02 ☑ ApolipoproteinA-II APOA2 Lipidtransport ▲ 1.28E-03 ☑ ApolipoproteinD APOD Lipidtransport ▲ 3.11E-05 ☑ Aggrecan ACAN Lipoproteinretention ▲ 1.51E-05 (1)ProteinnameisreportedaccordingwithSwissProt2011database.(2)Roleinatherosclerosisisreferredtodatareportedontherecentliterature(3)▼:down releasedinPcomparedwithDS;▲:upreleasedinPvsDS.(4)PvaluesweregeneratedusingMarkerViewsoftware.(5)Theirpresenceinplasmawasshowed accordingtoHumanProteinReferencedatabase(www.hprd.org). All these studies, due to technical limits as well as to high-throughput western blot analyses and microarray low sensitive and low-throughput methods, identified technologywereperformedallowingthefindingof7differ- only a small number of proteins, restricting considerably entially expressed proteins [8] and 11 proteins distributed thedetectionofnovellowabundantbiomarkers. between stable and unstable plaques [30]. A tissue proteo- Lepedda et al. [6] identified 9 differentially expressed mics study has recently identified and demonstrated that proteins in unstable compared to stable plaques while, osteopontin tissue expression in carotid plaques and its only one year later,Olsonet al. [7]suggested 19 proteins plasma levels are predictive of adverse cardiovascular with a differential distribution in stable compared to eventsinallatheroscleroticvasculardistricts[9]. more complicated carotid segments of the same patients. Oneofthefirstpioneeringsecretomestudies,usingtwo- To increase sensitivity in the search of new biomarkers, dimensional gel electrophoresis, allowed the identification Rocchicciolietal.JournalofTranslationalMedicine2013,11:260 Page10of15 http://www.translational-medicine.com/content/11/1/260 Figure4DifferentialproteinexpressionobtainedbymassspectrometricanalysisofPandDS.Fourofthe31proteinsreportedintable3 arechosenandcomparativeexpressionanalysisispresentedasBoxplots.Boxplotsshowmedianvalue,standarddeviation(SD),50%,minimum andmaximumintensityvaluesandwereexportedfromMarkerViewsoftware. of 14 proteins secreted by autoptic human coronary artery majority of the identified proteins (70%) are predicted as segments [12], but only recently, in order to avoid high secretory proteins by SecretomeP software. 40% of these, abundant plasma protein contamination, Cooksley- due to the presence of the characteristic signal peptide, Decasper and colleagues [13] studied the secretomes from are likely to follow the classical exocytosis pathway carotid plaques combining antibody phage display with across the Golgi apparatus. The other proteins are non- mass spectrometry. This last approach couples sensitivity classicallysecretedthroughmicrovesiclesandmultivesicu- and accuracy providing a direct detection of secreted pro- lar bodies-derived exosomes, asevidenced by the inclusion teinsinblood.Themethodisquitecomplexrespecttothe inExoCartadatabank(http://www.exocarta.org). full-screen untargeted approach chosen in our study and In our procedure, CEA specimens were cut in P and the number of identified proteins is rather limited, thus downstream DS segments in order to compare the restrictingthe“discoverypotential”ofthatapproach. secretomes of the upstream area of maximum stenosis to In our study the secretome of human carotid plaques that of partially preserved downstream area, subjected to was tackled with a high performance, very sensitive, gel- distinct flow shear stress and wall strain regimen. P seg- and label-free LC-MS/MS workflow. The unbiased ap- ments were classified as Stary’s types V and VI [19,20], proach, able to give a high sensitive full screen overview while DS segments displayed less severe wall changes. of secreted proteins from human carotid artery, is We considered this approach useful to differentiate presented here for the first time. This method allowed specific molecular factors of the central plaque area, the identification of 463 proteins thus producing a wide where complication features are present, from those of shot on proteins released by atherosclerotic arteries. The its distal side.
Description: