MS for a Comparative Proteomic Analysis of BALf

Mar 20, 2009 - of Rheumatology, University of Pavia, Via Taramelli 5, 27100 Pavia, Italy, ... Rheumatology, Immunology and Allergy, Georgetown Univers...
2 downloads 0 Views 987KB Size
2-DE and LC-MS/MS for a Comparative Proteomic Analysis of BALf from Subjects with Different Subsets of Inflammatory Myopathies Ileana Passadore,†,| Paolo Iadarola,*,† Cristina Di Poto,†,⊥ Serena Giuliano,† Carlomaurizio Montecucco,‡,| Lorenzo Cavagna,‡,| Claudia Bonino,‡,| Federica Meloni,§,| Anna Maria Fietta,§,| Antonella Lisa,∇ Roberta Salvini,† and Anna Maria Bardoni† Department of Biochemistry “A. Castellani”, University of Pavia, Via Taramelli 3/B, 27100 Pavia, Italy, Division of Rheumatology, University of Pavia, Via Taramelli 5, 27100 Pavia, Italy, Department of Haematological, Pneumological and Cardiovascular Sciences, University of Pavia, Via Taramelli 5, 27100 Pavia, Italy, IRCCS San Matteo Foundation, Piazzale Golgi 2, 27100 Pavia, Italy, Division of Rheumatology, Immunology and Allergy, Georgetown University Proteomics Laboratory, Washington, D.C. 20057, and Institute of Molecular Genetics, CNR, 27100 Pavia, Italy Received November 3, 2008

The protein profiles of bronchoalveolar lavage fluid (BALf) of patients belonging to three selected subsets of Polymyositis/Dermatomyositis (PM/DM) have been compared by using a combination of 2-DE and MALDI-TOF/MS or LC-MS/MS. Our study examined the hypothesis that there were distinct differences in protein expression profiles that were related to the phenotype. From among the 323 ( 51 protein spots that may represent the most highly expressed proteins in BALf of these patients, 24 unique spots were isolated and proteins identified. In particular, 9 spots were present in BALf of PM/DM patients only; 12 spots were exclusive of Overlap patients and 3 spots of AS patients. From among the proteins identified, a few were classified as cytoskeletal proteins, others were involved in oxidative stress and a number of proteins were associated with general metabolic activity or immunological response and inflammation. This is the first study in which evidence is provided that a number of different proteins are expressed in different subsets of PM/DM and supports our contention that the proteomic approach would be beneficial in discovering molecules which could represent possible prognostic factors of these rare pathologies. Keywords: 2-DE • LC-MS/MS • PM/DM • AS • Overlap Syndrome

1. Introduction Polymyositis/Dermatomyositis (PM/DM) are systemic inflammatory pathologies which impair skeletal muscles and result in proximal muscle weakness.1 Although proximal striated muscles are the main targets of these disorders, other organs, including the lung, may be involved by severe clinical manifestations. The reported prevalence of pulmonary complications, mostly interstitial lung disease (ILD), varies between 5 and 46%, depending on whether clinical, radiological, functional, or pathological criteria have been used in cross-sectional studies.2-7 Remarkable differences can also be observed between different serological and clinical subsets of PM/DM patients in the evolution of their pulmonary interstitial picture. A subset of PM/DM, in which patients develop ILD more likely * Corresponding author: Prof. Paolo Iadarola, Department of Biochemistry, University of Pavia, Via Taramelli 3/B, 27100, Pavia, Italy. Phone: +39-0382987264. Fax: +39-0382-423108. E-mail: [email protected]. † Department of Biochemistry “A. Castellani”, University of Pavia. | IRCCS San Matteo Foundation. ⊥ Georgetown University Proteomics Laboratory. ‡ Department of Internal Medicine, University of Pavia. § Department of Haematological, Pneumological and Cardiovascular Sciences, University of Pavia. ∇ Institute of Molecular Genetics. 10.1021/pr800943t CCC: $40.75

 2009 American Chemical Society

than others, has been found positive for anti-synthetase antibodies, in particular anti-histidyl-tRNA synthetase (Jo-1).8 This subset, called anti-synthetase syndrome (AS) shows an interstitial pulmonary involvement in 80% of cases. Whether the ILD associated with AD is different from that found in other PM/DM subsets is still a matter of debate.9 Thus, studies on bronchoalveolar lavage fluid (BALf) might be helpful to answer this question. Results of previous investigations have shown that protein composition of BALf in subjects with sarcoidosis, idiopathic pulmonary fibrosis, allergic asthma and chronic obstructive pulmonary disease is altered if compared with that of healthy subjects.10-12 Rottoli and co-workers13 have shown that the proteomic profile of BALf from patients affected by systemic sclerosis with pulmonary fibrosis (SScFib+) was “intermediate” between that of sarcoidosis and of idiopatic pulmonary fibrosis patients. Likewise, Fietta et al.14 demonstrated both qualitative and quantitative differences between the 2D gel electrophoresis (2-DE) profiles of six SSc patients without signs and symptoms of ILD and those of nine SSc patients with documented clinical-radiological fibrotic ILD. On the basis of these encouraging data, the hypothesis was drawn that the analysis of BALf of PD/DM patients could potentially Journal of Proteome Research 2009, 8, 2331–2340 2331 Published on Web 03/20/2009

research articles provide important information about changes in protein expression and secretion during the course of interstitial lung disease. For this aim, we have compared the proteins expressed in BALf of (i) PM/DM patients, (ii) AS patients with ILD, and (iii) subjects with polymyositis and dermatomyositis in association with various autoimmune and connective tissue diseases, in particular with systemic sclerosis (Overlap syndrome). 2-DE and matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS) and/or liquid chromatography tandem mass spectrometry (LC-MS/MS) have been used as tools to produce protein maps from each of these three selected subgroups of PM/DM patients. Qualitative and/or quantitative protein differences evidenced from the comparison of proteomic profiles could hopefully represent the first step toward the identification of possible prognostic factors and markers of clinical progression.

2. Materials and Methods 2.1. Reagents. Carrier ampholytes and immobilized pH gradient gel strips were from GE Healthcare (Uppsala, Sweden). Protease inhibitor cocktail, 1,4-Dithioerythritol (DTE), 3-[3Cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), glycerol, iodoacetamide, agarose, urea, sodium dodecyl sulfate (SDS), tris(hydroxymethyl)aminomethane (Tris) were purchased from Sigma-Aldrich (St. Louis, MO). Bicinchoninic Acid (BCA) was purchased from Pierce (Rockford, IL). Carrier ampholytes and immobilized pH gradient gel strips were from GE Healthcare (Uppsala, Sweden). Antibodies for detection of gelsolin, cofilin 1 and vimentin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). All other reagents were of analytical grade and used without further purification. 2.2. Patients. All patients (n ) 11) were nonsmokers who attended the outpatient Rheumatology Unit of the IRCCS San Matteo Hospital, Pavia, Italy. Their median age was 51 years (range 45-73 years) and the median duration of disease was 3 months (range 1-48). BALf samples were collected between 2004 and 2005 from 3 DM patients, 4 AS patients anti-Jo-1 positive and 4 patients with myositis overlap syndrome. Ten patients had active disease and were not treated at the time of BAL collection. One patient was treated with low dose prednisone and had persistently active disease despite treatment. All of them signed an informed consent to undergo bronchoscopy. The female/male ratio was 4:1. All patients presented functional signs of interstitial fibrosis and satisfied Bohan and Peter’s criteria for PM/DM.15 AS patients showed serological reactivity to anti-histidyl t-RNA synthetase by ELISA. 2.3. BALf Collection and Phenotype Analysis. BALf collection was performed as previously described.16 In synthesis, the distal tip of the bronchoscope was wedged into the middle lobe or lingular bronchus; a total of 150 mL of warm sterile saline solution was instilled in 30 mL aliquots and sequentially retrieved by gentle aspiration. The first aliquot collected (20 mL) was used for a series of analyses including microscopic and cultural examination of common bacteria and fungi; direct acid fast bacilli smears (Kinyoun method) and cultures for mycobacteria; microscopic examination of Pneumocystis jiroveci (Gomorri Grocott silver stain) and direct and cultural investigations for respiratory viruses. All other (50 mL aliquots, n ) 5) were used (i) to assess the total and differential cell counts (performed on cyto-centrifuged preparations using ¨´nwald-Giemsa plus Papanicolaou stainings), (ii) to May-Gru assess the percentages of CD4+ and CD8+ T cells (assessed 2332

Journal of Proteome Research • Vol. 8, No. 5, 2009

Passadore et al. by cytofluorimetry), and (iii) for proteomic studies. The BALf samples were immediately filtered twice through gauze, then centrifuged at 1500 rpm for 10 min before being promptly stored frozen at -80 °C until use. 2.4. Sample Preparation for Proteomic Studies. BALf samples were dialyzed (tubing cutoff:1 kDa) for 2 days against several changes of distilled water, in the presence of protease inhibitors. The BCA protein assay kit (Pierce) with BSA as standard was used to determine the protein concentration of each desalted sample. After lyophilization, the protein pellet was resuspended in a swelling solution containing 8 M urea, 4% (w/v) CHAPS, 65 mM DTE and bromophenol blue (2 µL). A total of 70 µg of protein was typically used for each analytical experiment; a higher amount (1 mg) was used when preparative experiments with mass spectrometry identification of spots were performed. 2.5. Two-Dimensional Gel Electrophoresis. A total of 0.7-1 mg of proteins was loaded on IPG gel strips (length 18 cm) from GE Healthcare with a nonlinear (NL) pH 3-10 gradient range. When proteins could not be resolved by applying this pH range, separations were obtained by using narrow range pH gradient 6-9 or 4-7. Gel strips were rehydrated at 16 °C in a swelling solution (350 µL of a buffer containing 8 M urea, 4% (w/v) CHAPS, 65 mM DTE, 0.8% (v/v) carrier ampholytes and 5 µL bromophenol blue), for 8 h using a voltage of 30 V. Each step of isoelectrofocusing (IEF) was carried out according to a program driven by the Ettan IPGphor system: 1 h at 120 V; 30 min at 300 V; linear ramping from 300 to 3500 V in 3 h; 10 min at 5000 V and then 7950 V to reach a total of 62 kV/h. The focused IPG strips were incubated for 12 min at room temperature in a first equilibration buffer containing 6 M urea, 2% (w/v) SDS, 50 mM Tris, pH 6.8, 30% glycerol, and for 5 min in a second equilibration buffer containing 2.5% (w/v) iodoacetamide and 2% (w/v) DTE.17 The strips were held in place with 0.4% low melting temperature agarose and loaded onto 20 × 18 cm, 9-16% SDS-polyacrylamide gels. Electrophoresis was carried out at a constant current of 40 mA in a PROTEAN II xi 2-D Cell (Bio-Rad, Richmond, CA) apparatus. The twodimensional gels were stained with ammoniacal silver nitrate18 or “Blue silver” according to the manufacturer’s instructions.19 2.6. Immunoblotting. The proteins separated as indicated above were transferred onto nitrocellulose (NC) membranes by using a Trans-Blot Electrophoretic Transfer Cell (Bio-Rad, Hercules, CA) to which a current of 200 mA for 2 h was applied. Individual membranes were incubated with monoclonal antigelsolin, anti-cofilin 1 and anti-Vimentin antibodies. Membranes were then washed and proteins visualized using an enhanced chemiluminescence (ECL) kit. 2.7. Gel Evaluation. Gels were scanned using the Versadoc Imaging Model 3000 System (Bio-Rad, Richmond, CA) and the total number of spots and spot quantities was measured using the PD QUEST 7.1 software (Bio-Rad). To detect protein spots differentially expressed among the three subsets of myopathies, the software was instructed to create one reference map (master gel) for each patient group. These maps were established by considering only the spots which were constantly expressed in each patient of the same group. Spot intensity (optical density) was measured by summing pixels within each spot boundary (spot volume) and recorded as a percentage of the total spot intensity on the gel: %V ) (spot volume/∑ volumes of all spots resolved in the gel). At least three gels for each patient were evaluated. To assess differences in spot quantity among the three study groups,

research articles

2-DE and LC-MS/MS for a Comparative Proteomic Analysis of BALf initially, we compared the three master gels. These differences were then confirmed by analyzing spot quantities observed in gels from each single patient. The following formula was used for determining the spot quantity: Spot height · π · σx · σy. This formula results in image unit (IU2) multiplied by optical density (OD). The spot height represents the peak of the spot’s Gaussian representation, while σx and σy are the SD of the spot’s Gaussian distribution in the direction of the x and y axis, respectively. On the basis of the recent guidelines for proteomics research,20 unmatched spots and spots that significantly differed in median quantity (p < 0.05) were regarded as being differentially expressed in PM/DM patients compared to AS and Overlap. 2.8. Reproducibility of the Study. 2-DE maps of proteins expressed in BALf of PM/DM, AS and Overlap patients were produced in triplicate. Gels presented in this report are the best representatives from among all generated. Experimental steps (sample preparation, electrophoretic run, gel staining) were performed “in parallel” on all samples. 2.9. Protein Identification. Proteins were identified by comparing the maps obtained in this study with the Swiss-2D PAGE (http://www.expasy.org/sprot) human plasma map and with previously published BALf maps,12,13 or by using immunoblotting, or matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS), and/or liquid chromatographytandem mass spectrometry (LC- MS/MS). For identification of protein spots with MALDI-MS, the Voyager DE-PRO (Applied Biosystem, Foster City, CA) mass spectrometer was used. The LC-MS/MS system was a Q-TOF Ultima hybrid (Micromass, Toronto, Canada) mass spectrometer. The MASCOT (Matrix Science, www.matrixscience.com) peptide fingerprinting search program and the ProteinLynx software were used, respectively, for peptide searching. These analyses were performed at the CEINGE Biotecnologie Avanzate s.c.a.r.l. center, Federico II University, Naples (Italy). 2.10. In Situ Digestion and MALDI-MS Analysis. When 2-DE runs (3-10 NL gradient gels; 9-16% T) were performed for preparative purposes, selected protein spots were excised from the gel and washed in 50 mM ammonium bicarbonate, pH 8.0, followed by a 50% aqueous acetonitrile (ACN) solution until complete destaining. Proteins were recovered, dried under vacuum, resuspended in 50 mM ammonium bicarbonate, pH 8.0, and digested at 37 °C with 5 µg/mL sequencing grade trypsin (Promega, Madison, WI). After overnight incubation, peptides were extracted sequentially three times with 50% ACN, 5% trifluoroacetic acid (TFA) in water. Each extraction involved 10 min of stirring followed by centrifugation and removal of the supernatant. The original supernatant and those obtained from three sequential extractions were combined and dried. The peptide mixture was then solubilized with 0.5% TFA for MS analysis. MALDI-TOF analysis of the trypsin digest was performed using the system indicated above equipped with a pulsed N2 laser operating at 337 nm. Positive ions spectra were acquired in reflector mode over a m/z range of 500-5000 using an extraction delay time of 40 ns. The analysis was performed by spotting 1 µL of sample mixed with an equal volume of the matrix solution (R-cyano-4-hydroxycynnamic acid 10 mg/mL in 1:1 ACN/water containing 0.1% TFA) onto the target plate. External mass calibration was performed using the peptide mass standard kit provided by manufacturer. Database searches were performed against the NCBI nonredundant database using MASCOT search engine. Mascot scores greater than 65 were

considered significant (p < 0.05) and were subsequently blasted against the Swiss-Prot database. 2.11. LC-MS/MS Analysis. When the identity of proteins could not be established by peptide mass fingerprinting, the peptide mixtures were further analyzed by LC-MS/MS. The peptide mixture (10 µL) was first loaded onto a reverse-phase trap-column (Waters) using 0.2% formic acid as eluent at a flow rate of 10 µL/min. The sample was then transferred to a C18 reverse-phase capillary column (75 µm × 20 mm) at a flow rate of 280 nL/min and separated using a linear gradient from 7% eluent A (0.2% formic acid in 5% ACN) to 60% eluent B (0.2% formic acid in 95% ACN) in 50 min. The mass spectrometer was set up in a data-dependent MS/MS mode to alternatively aquire a full scan (m/z acquisition range from 400 to 1600 Da/ e) and a tandem mass spectrum (m/z acquisition range from 100 to 2000 Da/e). The three most intense peaks in any full scan were selected as precursor ions and fragmented by collision energy. For the peptide sequence searching, the mass spectra were processed and analyzed using the MASCOT MS/ MS ion search program using the Swiss-Prot 49.1 database. The mass accuracy was within 50 ppm for the peptide mass tolerance and within 0.25 Da for fragment mass tolerance. Protein identification was repeated at least once using matching spots from different gels. 2.12. Statistical Analysis. Proteomic expression data were expressed as median spot quantities, and differences were assessed with the nonparametric Wilcoxon test. For all tests, a p-value e0.05 was regarded as significant. Statistical prcedures were run with JMP 5.1.2 from SAS Institute, Inc.; p < 0.05 was regarded as statistically significant. Data are expressed as mean ( SD.

3. Results and Discussion To compare the complex network of proteins expressed in BALf of PM/DM, AS and Overlap syndrome patients, proteomic profiles of the above patients were produced and analyzed “in parallel”. 3.1. Interpretation of 2-DE Gels. 2-DE analysis was performed in triplicate on each BALf to produce 33 gels in total. All gels within a group of patients were scanned and interpreted with the PDQuest 7.1.1 software. Spot detection was achieved using the spot detection wizard tool after defining and saving a set of detection parameters. After spot detection, the original gel scans were filtered and smoothed to clarify spots, remove vertical and horizontal streaks and remove speckles. Threedimensional Gaussian spots were then created from Filtered images. Three images were created from the process: the original raw 2-D scan, the Filtered image and the Gaussian image. A match set for each group of disease was then created for comparison after the gel images had been aligned and automatically overlaid. If a spot was saturated, irregularly shaped, or otherwise of poor quality, then the Gaussian modeling was unable to accurately determine quantity. In these cases, the spot was defined in the Filtered image using the spot boundary tools. Thus, for each group, a master gel was produced which included protein spots only if present at least in two out of the three gels. The mean spot numbers in Coomassie stained gels were 258 ( 48 in PM/DM, 284 ( 32 in AS and 323 ( 51 in Overlap Syndrome. The master gels from each group showed patterns of proteins similar such that they could be matched to each other. This facilitated the correlation of gels and the creation of a virtual image, indicated as high Journal of Proteome Research • Vol. 8, No. 5, 2009 2333

research articles

Passadore et al.

Figure 1. Upper panel: 2-DE High Master Gel (HMG) created from the correlation of master gels of three different subsets of PM/DM. Solid circles indicate spots unmatched among the three groups. Dashed circles indicate spots common to all groups but differentially expressed among them. Lower panel: Representative 2-DE gels of protein spots obtained from BALf of PM/DM, AS, and Overlap patients (left to right). Solid and dashed circles refer to unmatched and common spots, as indicated above.

master gel (HMG, Figure 1, upper panel), comprehensive of all matched spots derived from master gels. 3.2. Quantitation of Gel Spots. Spot quantities of all gels were normalized to remove nonexpression-related variations in spot intensity and data were exported as clipboard for further statistical analysis. The raw amount of each protein in a gel was divided by the total quantity of all proteins (spots) that were included in that gel. The results were evaluated in terms of spot optical density (OD). Statistical analysis of PDQuest data (Smith’s Statistical Package) by nonparametric Wilcoxon test allowed to assess differences in protein abundance on a protein-by-protein basis. According to recent guidelines for differential proteomic research,20 only unmatched spots and spots that significantly differed in density (p < 0.05, by nonparametric Wilcoxon test) were considered “differentially expressed” in the three groups of patients. 3.3. MALDI-TOF-MS and LC-MS/MS Analysis of Unmatched Protein Spots. The comparative analysis of profiles evidenced that 24 protein spots were apparently differentially expressed among groups. The distribution of these 24 spots in HMG is indicated by numbers and solid circles in Figure 1. The lower panel of the same figure shows (left to right) the position in individual real gels of the nine spots (i.e., spots 1-9), present in PM/DM patients only, of the 3 spots (i.e., spots 10-12) exclusive of AS patients and of the 12 spots (i.e., spots 13-24) exclusively present in Overlap patients. Given the aim of our study, these 24 unmatched spots were selected as the spots on which our attention should be focused for further analysis. They were carefully excised from the gel, destained, digested with trypsin and peptides submitted to MALDI-TOF/MS fol2334

Journal of Proteome Research • Vol. 8, No. 5, 2009

lowing the procedure detailed in Materials and Methods. The MS fragmentation data were searched against the NCBI database and the database queries were performed using the Mascot search engine. A confident identification by MALDITOF was obtained only for 7 spots (i.e., spots 1-4 and 18-20). Because of the bad quality of MS signals, most likely determined by the low abundance of material, identification by MALDI of other 17 spots (i.e., spots 5-17 and 21-24) failed. Fortunately, confident identification also of these proteins could be achieved by applying LC-MS/MS. Detailed identification data, including accession number, theoretical pI, molecular mass, percent of sequence coverage, number of peptides identified and MOWSE score of each of the 24 proteins identified by MALDI or LC-MS are shown in Table 1. Additional information concerning the primary sequence of all peptides identified for each protein analyzed has been included in Table 1S of Supporting Information. 3.4. Functional Classification of Differentially Expressed Proteins. As shown in Figure 2, the function of proteins identified was rather heterogeneous. According to the literature, a significant proportion of these proteins was associated with general metabolism (50%); others were related to immunological response and inflammation (15%), tissue repair and proliferation (15%), lipid metabolism (4%), antioxidant (10%), cytoskeleton (3%), and a series of proteins (3%) whose role is still unknown. 3.5. Cytoskeleton/Tissue Architecture Proteins. Interestingly, seven (i.e., spots 1-6 and spot 9) of the nine spots exclusively detectable in PM/DM contained homogeneous proteins. Instead, spots 7 and 8 contained three and two proteins, respectively. Spots 9, 4, and 6 corresponded to

a

b

gi/73535278 gi/6013427 gi/20127450 gi/3912938 gi/6013427 gi/6013427 gi/6013427 gi/119586024 gi/119626065 gi/66473265 gi/4506883 gi/6013427 gi/189724 gi/183817 gi/4506773 gi/6013427 gi/6013427 gi/6013427 gi/6013427 gi/110590597 gi/6013427 gi/4557871 gi/21620055 gi/6013427 gi/113584 gi/10334611 gi/4557871 gi/6013427 gi/26892090 gi/113584

gi/4505591 gi/4505591 gi/4503571 gi/62414289 gi/20521934 gi/181603 gi/4557581 gi/119568841 gi/13376737 gi/5031635 gi/63100756 gi/63252913 gi/122920512

accession number (NCBI)

nd ) not detected.

24

23

22

21

18 19 20

17

16

15

14

12 13

11

10

9

8

7

1 2 3 4 5 6

spot

Aminoacid sequences of peptides are shown in Table 1S.

Peroxiredoxin 1 Peroxiredoxin 1 Enolase 1 Vimentin KIAA1538 protein Myotonic dystrophy kinase Fatty acid binding protein 5 hCG2030818, isoform CRA_b Coenzyme Q10 homologue B Cofilin 1 3-hydroxybutyrate dehydrogenase, type 2 Gelsolin-like capping protein ChainA, human serum albumin complexed with myristate and Aspirin Chain A, human pyruvate kinase Serum albumin precursor Protein kinase C, β isoform 2 MRE11 homologue hMRE11 Serum albumin precursor Serum albumin precursor Serum albumin precursor GATA binding protein 4, isoform CRA_a Albumin, isoform CRA_b β globin chain Semenogelin I isoform a preprotein Serum albumin precursor Prostate Specific Antigen precursor β globin S100 calcium-binding protein A9 Serum albumin precursor Serum albumin precursor Serum albumin precursor Serum albumin precursor Chain A, Apo human Transferrin (Non-Glycosylated) Serum albumin precursor Transferrin Ig H G3 protein Serum albumin precursor Ig R 1 chain C region Immunoglobulin heavy chain Transferrin Serum albumin precursor β globin chain variant Ig R 1 chain C region

protein identified

Table 1. List of Proteins Identified by MALDI-TOF and LC-MS/MS Analysisa

nd

nd nd +

+ +

nd

nd

nd

nd

nd nd nd

nd

nd

nd

nd

nd nd

nd

nd

nd

nd

+

nd

nd

nd

nd

nd nd nd

nd

nd

nd

+

+

+

+

+ + +

+

+

+

+

nd +

+ nd nd

nd

+

nd

nd

nd nd nd nd nd nd

nd nd nd nd nd nd

+ + + + + +

Overlap

AS

relative expression PM/DM

7.01 5.91 6.59 5.56 5.91 5.91 5.91 9.49 6.96 5.90 9.30 5.91 7.70 6.28 5.71 5.91 5.91 5.91 5.91 6.58 5.91 6.81 7.95 5.91 6.08 8.30 6.81 5.91 7.86 6.08

8.27 8.27 7.01 5.06 8.68 4.92 6.60 6.33 9.67 8.22 7.56 5.82 5.62

theoretical pI

62.57 71.18 78.16 77.91 71.18 71.18 71.18 42.45 61.12 11.54 52.16 71.18 28.78 19.20 13.29 71.18 71.18 71.18 71.18 76.81 71.18 79.28 58.37 71.18 38.49 40.58 79.28 71.18 16.10 38.49

22.32 22.32 47.48 53.68 107.8 64.48 15.50 17.66 27.50 18.72 27.08 85.00 68.41

theoretical Mr (kDa)

3 6 2 2 8 24 8 2 14 2 27 7 3 6 8 3 11 10 14 7 4 7 3 43 10 3 14 19 3 3

9 9 11 26 2 2 3 2 2 8 3 3 8

number of peptides matchedb

10% 7% 5% 3% 13% 21% 13% 8% 24% 24% 39% 10% 8% 25% 46% 7% 15% 21% 21% 10% 7% 12% 3% 46% 26% 9% 23% 20% 23% 8%

44% 44% 33% 57% 2% 5% 13% 11% 9% 49% 13% 13% 12%

sequence coverage

124 266 28 26 340 530 36 36 465 53 871 324 94 122 110 88 114 122 131 254 110 288 46 1348 300 86 591 461 145 117

122 133 158 336 39 29 73 33 29 340 33 138 281

MOWSE score

2-DE and LC-MS/MS for a Comparative Proteomic Analysis of BALf

research articles

Journal of Proteome Research • Vol. 8, No. 5, 2009 2335

research articles

Passadore et al. that regulate cell size and shape in a variety of organisms. DMPK is mainly expressed in muscle, which is a key targettissue for insulin-dependent regulation of glucose metabolism. Functions and modes of regulation in vivo of this novel kinase are currently unknown, but its localization suggests that it is likely to function as a component of a previously uncharacterized signal transduction pathway in these tissues.23

Figure 2. Classification of the identified proteins by their function. Proteins are associated with general metabolism (50%), immunological response and inflammation (15%), tissue repair and proliferation (15%), antioxidant (10%), lipid metabolism (4%), cytoskeleton (3%) and unknown (3%).

homogeneous tissue architecture proteins, while spot 8 included a cytoskeleton protein and a protein involved in oxidative stress. Because of their pivotal role in remodeling of the actin cytoskeleton and in degenerative and regenerative changes in muscle, the tissue architecture proteins were considered the most promising candidates in terms of future work on inflammatory pathologies which impair skeletal muscles. Their properties will be discussed below. Gelsolin (spot 9) is one of the most abundant proteins secreted by human bronchial epithelial cells into the airway surface liquid. The previous finding21 that this protein was 3-fold increased in the liquid of epithelia treated (“in vitro”) with IL-4 suggested that it was released by epithelial cells into the airways and that its secretion was increased by IL-4. These data also indicated that gelsolin plays a possible role in improving fluidity of airway surface liquid through the degradation of filamentous actin released in large amounts by dying cells during inflammation. That this protein was able not only to control mucus viscosity, but also to preserve the innate antimicrobial activity in the airway surface was demonstrated by the observation that gelsolin fully reversed the inhibition of antimicrobial activity determined by some peptides,22 Given the stoichiometric nature of the interaction between gelsolin and actin, it is possible to speculate that the massive actin release observed during severe inflammatory processes may overwhelm the counteracting capacity of endogenous gelsolin. This means that gelsolin might also have an important role in the airway surface liquid (ASL) barrier and makes exogenous gelsolin an excellent candidate for therapeutic use in diseases characterized by chronic inflammation with significant levels of filamentous actin release. Vimentin (spot 4), a major intermediate filament protein in airway and vascular smooth muscle, links dense bodies and desmosomes in smooth muscle cells, thus, being a marker of degenerative and regenerative changes in muscle. In polymyositis, T-cell mediated myocytotoxicity is directed against strongly human leukocyte antigen class I positive (HLA-I+) muscle fibers and fiber regeneration is, at least in part, responsible for HLA-I up-regulation. Distinctive patterns of HLA-I, nerve cell adhesion molecule (NCAM), and vimentin expression accompany denervation and regeneration. An overexpression of vimentin could suggest a more rigid structure of cells, making them less sensitive to environmental stress. Human myotonic dystrophy protein kinase (DMPK, spot 6) is a member of a novel class of multidomain protein kinases 2336

Journal of Proteome Research • Vol. 8, No. 5, 2009

As indicated above, spot 8 contained, in addition to D-3hydroxybutyrate dehydrogenase, also cofilin 1, a protein that is essential in a number of cell systems for cell viability and for actin-based motility.24 Cofilin severs actin filaments, thus, enhancing the dynamics of filament assembly. Many cellular processes involve rapid remodeling of the actin cytoskeleton and changes in the pools of assembled and monomeric actin. The cofilin/actin depolymerizing factor (ADF) family of proteins appears to be a major factor contributing to actin depolymerization in cells, which is essential for recycling actin subunits to support the growth of new filaments. By severing actin filaments, cofilin creates free barbed and pointed ends available for polymerization or depolymerization, depending on the local concentration of actin monomer.25 Since we are particularly interested in the evaluation of the expression level of cofilin, this protein would have to be separated from the other incorporated in spot 8. For this aim, 2-DE with a pH gradient 6-9 was performed on all BALf samples. The good resolution of spots enabled us to identify cofilin 1 (spot 8 in Figure 3, panel A) in homogeneous form and to observe that its amount (in terms of optical density) in PM/ DM was comparable to that of other proteins discussed above. By contrast, cofilin spot in other two disorders (Figure 3, panel A) was so faint that it could not be detected. Thus, as cofilin and other cytoskeleton proteins were not detected on the AS and Overlap gels, we were not able to define any alteration in abundance among groups. Although these observations remain primarily qualitative, our findings raise the question of the significance of this drastic differences among subsets of myopathies. Obviously, given the limited body of known information, it remains a speculation whether these molecules can be reliable biomarkers of the disorder. However, such variations are expected to trigger different physiological responses which could partly explain the possible different mechanisms among subsets of myopathies here considered. 3.6. Detection of Gelsolin, Cofilin 1, and Vimentin in Plasma Samples. The finding of gelsolin, cofilin 1, and vimentin in BALf raised the question whether these proteins could actually be secreted into the airways. To obtain this information, their expression in paired BALf and plasma samples of all patients investigated was compared. Although spots of gelsolin, cofilin 1, and vimentin were found to be relatively more intense in 2-D gels of BALf than in plasma of PM/DM (not shown), our Western blots were successful in positively identifying only gelsolin in both fluids. As shown in Figure 4, panels a-d, the use of a specific anti-human gelsolin antibody allowed to evidence the gelsolin spots (indicated by arrows and solid circles) both in BALf and in plasma of PM/DM. Efforts in developing the spots of other two proteins in AS and Overlap failed, most likely due to interference of cross-reacting contaminants (not shown). However, the finding in BALf and plasma of PM/DM of comparable concentrations of gelsolin (panel d) allowed to speculate that, at least for these patients, gelsolin is produced also in the airways, although additional

2-DE and LC-MS/MS for a Comparative Proteomic Analysis of BALf

research articles

Figure 3. Magnification of 2-DE gels perfomed, with a pH gradient 6-9 (panel A) or 4-7 (panel B), on BALf of PM/DM; AS and Overlap patients (left to right), to separate proteins clustered in a single spot. Solid circles indicate the main components of the cluster. Arrows point to spots originated from separation.

experiments are needed to establish whether this is the protein that must be pursued for further investigations on these pathologies. 3.7. Kinases and Calgranulin B. Pyruvate kinase (PK) and pyruvate kinase C (PKC) (spots 10 and 11) were exclusive of AS and calgranulin B (spot 17) was present exclusively in Overlap. As shown in Table 1, all of these spots contained two or more proteins. In an attempt to purify the proteins of interest from other contaminants (mostly albumin fragments), electrophoretic runs with a narrow range pH gradient 4-7 were performed. These optimized conditions enabled us to obtain all these proteins in homogeneous form. PK, PKC and Cal B are indicated by solid circles and numbers in Figure 3, panel B; spots corresponding to other proteins are indicated by arrows. In our opinion, special interest for future work should be focused on pyruvate kinase in view of its important role in PM previously shown by Kawachi et al.26 They induced an experimental allergic myositis (EAM) in BALB/c mice by inoculating syngeneic dendritic cells (DC) presenting peptides that were expected to match the binding anchor motif of H-2K(d). Although peptides highly expressed in skeletal muscle were selected, only when syngeneic bone marrow-derived DC presenting pyruvate kinase M1/M2 peptide 464-472 was inoculated in BALB/c mice; 41.7% of the EAM mice developed pathological changes in skeletal muscle compatible to human Polymyositis. The conclusion was drawn that pyruvate kinase M1/M2 peptide was indicated as a candidate autoantigen not only in EAM BALB/c, but also in human-PM with the HLA A*2402 allele. S100 calcium-binding protein A9 or Calgranulin B was an intense spot (spot 17) detected in Overlap patients only. Although this finding cannot readily be connected to pathogenesis, Cal B, together with Cal A, was shown by other authors27 to be unique in its myeloid-specific expression profile and in its abundance in neutrophils. Increased levels of Cal B have been found in the bronchial secretion of patients with

chronic inflammatory disease, and these levels may induce the production of IL-8 by airway epithelial cells.28 The expression of Cal B and Cal A has been found to correlate with the inflammatory activity in systemic vasculitis, thereby confirming the role of these proteins in inflammatory reactions of endothelia.29 Although the general consensus is that the Cal B function depends mainly on heterodimer formation (Cal B-Cal A), several studies have demonstrated that the monomer is also a potent stimulator of neutrophils and is involved in the recruitment of leukocytes to the site of inflammation through the regulation of adhesion and the extravasion of neutrophils.27,30 At this time, the role of PKC (spot 11) in myopathies remains unclear. It is known that PKC is a family of serine-threonine protein kinase isoenzymes and represents one of the major signal transduction systems in inflammation. A group of these (classical isoenzymes) is activated by phosphatidylserine, Ca2+ and diacylglycerol (DAG) and may regulate the expression of inflammatory genes by regulating the activation of transcription factors NF-kB and AP-1. Some authors have shown that cPKCs, most likely PKC beta II, are involved in the upregulation of Tristetraprolin (TTP) expression, a 3′-UTR-binding protein known to destabilize mRNAs of TNFa and some other cytokines, and to act as an anti-inflammatory factor in activated macrophages. This regulation is mediated through the activation of transcription factor AP-2 and serves as an additional mechanism how PKC beta regulates the inflammatory process.31 3.8. Proteins Involved in Oxidative Stress. A few proteins also found to be unique to PM/DM were Peroxiredoxin 1, Coenzyme Q10, D-3-Hydroxybutyrate dehydrogenase and β-globin. All of them have a role in the response to the oxidative stress. Whether and how these proteins are involved in the pathogenesis of the systemic diseases needs to be further clarified, although it is well-known that a variety of cellular consequences, including loss of protein function, interference Journal of Proteome Research • Vol. 8, No. 5, 2009 2337

research articles

Passadore et al.

Figure 4. (Panels a-c) Immunoblotting of Gelsolin expressed in BALf and plasma samples (left to right) of PM/DM, AS and Overlap Syndrome patients (top to bottom). Samples were probed with 1:2000 mouse monoclonal anti-gelsolin antibody, after separation by 2-DE on which 180 µg of proteins was loaded. Gelsolin isoforms are indicated by a solid circle and arrow. The small spot at the bottom of panel a was identified as a fragment with Mr ) 56 000 Da. (Panel d) Quantification of gelsolin using scanning densitometry. Each bar (mean ( SD of three independent experiments) represents the relative density (INT/mm2) of gelsolin expressed in BALf and plasma (left to right) from patients of three study groups. Table 2. List of Proteins Found to Vary Significantly between the Three Study Groupsa total quantity (unit of OD*IU2)

spot

accession number (NCBI)

protein identified

PM/DM

AS

Overlap

theoretical pI

theoretical Mr (kDa)

P-value Wilcoxon test

25 26 27 28 29

gi/1703025 gi/20178280 gi/10334611 gi/1657327 gi/4960066

R1-antitrypsin fragment Fibrinogen γ fragment Immunoglobulin heavy chain Immunoglobulin light chain Apolipoprotein A1 fragment

0.5 ( 0.1 0.5 ( 0.1 4.40 ( 3.81 8.20 ( 3.81 0.5 ( 0.1

2.13 ( 1.67 6.63 ( 1.92 0.5 ( 0.1 10.63 ( 8.18 4.30 ( 0.17

2.43 ( 1.79 5.27 ( 0.75 9.23 ( 6.04 0.5 ( 0.1 2.52 ( 1.44

4.89-4.96 5.32-5.61 6.2-8.2 4.93-9.32 5.05

130.3-120.2 49.5-49.2 55.8-51 27.9-23.2 22.3