Application of Saturation Dye 2D-DIGE Proteomics to Characterize Proteins Modulated by Oxidized Low Density Lipoprotein Treatment of Human Macrophages Annabelle Dupont,†,‡,§,| Maggy Chwastyniak,†,‡,§ Olivia Beseme,†,‡,§ Anne-Laure Guihot,†,‡,§ Hervé Drobecq,‡,§,⊥ Philippe Amouyel,†,‡,§ and Florence Pinet*,†,‡,§ INSERM, U744, Lille, France, Institut Pasteur de Lille, Lille, France, University of Lille 2, IFR 141, Lille, France, and CNRS, UMR 8525, Lille, France Received October 22, 2007
Abstract: Macrophages are believed to play a crucial role in atherogenesis and atherosclerotic plaque progression, mainly through their role in the accumulation of large amounts of cholesteryl ester and foam cell formation after the uptake into the arterial intima of oxidized LDL (oxLDL) particles known to be proatherogenic. The aim of this study was to use a differential proteomic approach to identify the response of human monocyte-derived macrophages after treatment with oxLDL for 24 h. Mass spectrometry analysis (MALDI-TOF) of 2D-DIGE gels made it possible to identify 9 intracellular and 3 secreted proteins that were up-regulated, 11 intracellular and 1 secreted proteins that were down-regulated, and 2 secreted proteins that were induced. This methodological approach not only confirmed the differential expression levels of proteins known to be regulated by oxLDL in macrophages, such as catalase and pyruvate kinase, but also identified oxLDL modulation of other proteins for the first time, including heat shock proteins (HSP) and Actin cytoskeletal proteins. Semiquantitative Western blot confirmed their role. The HSPs identified included heat shock cognate 71 kDa protein (Hsc70), 75 kDa glucose-regulated protein (GRP75), heat shock 70 kDa protein (Hsp70), and 60 kDa (Hsp60) proteins. These highly conserved intracellular protein chaperones, commonly seen in atherosclerotic plaques, appear to participate in protection against cellular stress. Interestingly, oxLDL also modulated several F-Actin capping proteins involved in Actin polymerization and motility: gelsolin, CapG, and CapZ. In conclusion, we have demonstrated the effects of oxLDL in the modulation of several proteins in human macrophages and established a functional profile of the human macrophage during the atherosclerotic process. * To whom correspondence should be addressed. Dr Florence Pinet, INSERM U744-IPL, 1 rue du professeur Calmette, 59019 Lille cedex, France. Tel: (33) 3 20 87 72 15. Fax: (33) 3 20 87 78 94. E-mail: florence.pinet@ pasteur-lille.fr. † INSERM. ‡ Institut Pasteur de Lille. § University of Lille 2. | Present address: INSERM EA 2693, Université de Lille 2, Lille. ⊥ CNRS.
3572 The Journal of Proteome Research 2008, 7, 3572–3582 Published on Web 06/13/2008
Keywords: 2D-DIGE • saturation dyes • macrophages • human • oxLDL • HSP • Actin
Introduction Atherosclerosis, with its ensuing coronary heart disease, is the most common cause of death in industrialized nations.1 Plasma low density lipoprotein (LDL) is transported across the intact endothelium and becomes trapped in the extracellular matrix where it is subjected to oxidative modification.2 Oxidized LDL (oxLDL) alters various signaling pathways and biological responses, including inflammation, gene expression, cell expression, and apoptosis, and is involved in the pathogenesis of atherosclerosis. The retention and subsequent oxidation of LDL by arterial wall cells, including endothelial and smooth muscle cells and macrophages, is a central event in the early development of atherosclerotic lesions.3,4 Immunocytochemical studies of human atherosclerotic lesions show that they are composed predominantly of macrophages. In the intima, they bind and take up oxLDL and subsequently change into foam cells, which are early sites of potential atheroma development. Profiling gene expression using microarrays and the foam cell model has proven useful in identifying new genes that may contribute to atherosclerotic lesions.5 Because a change in gene expression does not necessarily lead to measurable changes in plasma protein levels,6 however, we chose to use this model by profiling proteomes rather than mRNA. Proteomics is a unique tool for analyzing the genome expression during atherosclerotic processes and provides important clues to the mechanisms involved. To investigate them, we looked for direct monitoring proteins that are specifically up- or down-regulated in human monocyte-derived macrophages (MDM) that have been activated by oxLDL. Proteomic analysis with silver-stained 2D gel and mass spectrometry is a powerful technique that allows 2D mapping of human macrophages.7 It also uses substantial amounts of protein samples, however, at least 100 µg protein per experiment. It thus presents practical problems, especially for clinical studies, in which it can be difficult to obtain large amount of blood to prepare MDM from each patient. To overcome these technical limitations imposed by the need for relatively high quantities of protein, we tested fluorescent dyes to find a technique suitable for analyzing scarce samples. The use of fluorescent dyes8 has resolved some of the limitation of 2D-PAGE, such as low sensitivity, narrow dynamic 10.1021/pr700683s CCC: $40.75
2008 American Chemical Society
HSP Proteins in Human Macrophages Treated by oxLDL range, and laborious image analysis. The method of choice for analyzing scarce samples with a proteomic approach appears to be the DIGE saturation labeling technique. This method enables complete 2D analysis and quantification of changes in protein abundance. The new generation of dyes reacting with cysteine residues used with standard proteins has increased detection sensitivity by a factor of 100, compared with the “minimal labeling” at lysine residues.9 The use of saturation dyes can profile the protein expression of samples that would otherwise be too small to analyze, as Greengauz-Roberts et al. demonstrated.10 Fujii et al.11 found a yield of 1500 protein spots detected with 12.5 µg of labeled protein, and most were reproducible. This suggests that saturation cysteine dyes may be very useful for proteomic studies of materials from which only a small amount of protein is available, such as cells recovered from frozen tissue by laser microdissection and biopsy specimens. We chose as our cell model primary cultures of MDM instead of established cell lines such as THP-1 for human cells or J774 for murine cells. This deliberate choice allows us to determine the conditions for proteomic analysis of MDM from blood samples of patients carefully phenotyped for atherosclerosisrelated diseases, such as abdominal aortic aneurysms. We also decided to focus not only on secreted or excreted proteins found in cell culture media as Fach et al. did12 but to analyze in the same MDM culture intracellular and secreted proteins whose expression is modulated by oxLDL. In this study, we demonstrate the usefulness of DIGE saturation labeling for analyzing scarce samples. We used 2DDIGE with Western blotting to confirm that the proteins selected were modulated by oxLDL in human macrophages. Interestingly, we found both heat shock proteins (HSPs) and proteins involved in Actin polymerization in oxLDL-treated macrophages.
Experimental Procedures Isolation and Culture of Human MDM. Primary cultures of human MDM were prepared as previously described, with a technique adapted from Boyum.13 Briefly, peripheral blood mononuclear cells from healthy human donors were isolated from buffy coats obtained from the Regional Blood Transfusion Center (Lille, France). Buffy coat (30 mL) was diluted with PBS (1/1 v/v) and carefully loaded onto a Ficoll gradient. After an initial centrifugation step (370 × g for 20 min at room temperature), monocytes were collected at the interface and then washed three times in PBS containing 0.1% EDTA (at 370 × g for 10-min periods), and then once in PBS alone at 370 × g for 10 min. Finally, the cell pellet was resuspended in an RPMI-1640 medium (Bio Whittaker, Verviers, Belgium) containing penicillin (10 000 U/mL), streptomycin (10 000 mg/mL) (0.4%), glutamine (1%), and sodium pyruvate (2%). The cells were then seeded in 90 mm Primaria dishes at a density of 6 × 106 cells. After sedimentation for 90 min, the supernatant containing the nonadhering cells was discarded. The adherent cells, consisting of monocytes, were washed twice with PBS, and fresh medium containing 10% v/v heat-inactivated human serum (Promocell) was added. Cells were incubated at 37 °C in humidified air containing 5% CO2. The culture medium was then changed every two days. LDL Isolation and Modification. LDL was isolated from fresh human EDTA-plasma obtained from healthy donors (density 1.019–1.063 g/Ll) by sequential ultracentrifugation.14 For oxidation, LDL was diluted to 1 mg/mL with EDTA-free PBS and
technical notes incubated with 5 µmol/L CuSO4 for 24 h at 37 °C. This procedure produced high LDL oxidation. The resulting preparations are hereafter referred to as oxLDL. The degree of LDL oxidation was estimated according to the levels of thiobarbituric acid-reactive substances (0.8 ( 0.7 and 39.5 ( 3.3 nmol/mg protein for native and oxLDL (n ) 3))15 and lipid hydroperoxides (10.8 ( 3.4 and 293.4 ( 86.2 nmol/ mg LDL protein, for native and oxLDL (n ) 3)).16 LDL concentrations were determined by the Peterson technique17 with BSA as the standard. Cell Treatment. On day 12 of the primary culture, MDM was washed three times with PBS and then incubated in serumfree culture medium with oxLDL (100 µg/mL) or native LDL (100 µg/mL). The group of control cells comprised the MDM not treated with either oxidized or native LDL. The quality of MDM cultures was evaluated as previously described18 by characterization of ECE-1 mRNA, an intracellular marker, and MMP-9, a marker secreted into the culture medium. ECE-1 mRNA was detected by RT-PCR on total RNA isolated from MDM and MMP-9 activity was determined in culture medium with gelatin zymography.18 Intracellular and secreted proteins were extracted as previously described7 to perform 2D gel electrophoresis. MDM was washed three times with 25 mmol/L Tris, pH 7.4, and scraped in lysis buffer containing 30 mmol/L Tris pH 8, 4% CHAPS, 2 mol/L thiourea, 7 mol/L urea (Bio-Rad, Hercules, CA) to yield a concentration of approximately 60 000 cells/mL. Cells were then lysed in ice with a mixer suitable for 1.5 mL Eppendorf tubes. Protein samples were stored at -20 °C until the protein quantity was estimated with the commercial Bradford reagent (Biorad) and BSA as the protein standard. One-hundred milliliter aliquots of the proteins were then prepared and stored at -20 °C until use. 2D-DIGE. Five micrograms of each sample (control and oxLDL-treated macrophages) adjusted to 9 µL with lysis buffer were then reduced by incubation with 2 mmol/mL Tris-(2carboxethyl)phosphate hydrochloride (TCEP; Sigma, St. Louis, MO, USA) for 1 h at 37 °C. Cy3 or Cy5 sulfhydryl-reactive dye was added (0.8 nmol/µg protein; GE Healthcare, Buckinghamshire, UK) and incubation was continued for 30 min at 37 °C. The final ratio of 0.4 nmol TCEP and 0.8 nmol Cy3 or Cy5 dyes per microgram of protein provided the best resolution. The reaction was stopped by the addition of an equal volume of sample buffer containing 7 mol/L urea, 2 mol/L thiourea, 4% CHAPS, 130 mmol/L DTT, and 2% Pharmalyte (GE Healthcare). Samples were stored either briefly on ice before use or stored at -80 °C for future use. All labeling procedures were performed in the dark. Cy3-labeled samples were mixed with Cy5-labeled samples before 2D gel electrophoresis. For each 2D gel, one control and one treated sample were mixed together either labeled with Cy3 or Cy5. For the isoelectrofocusing (IEF) step, the IPG strip gels (length 18 cm, pI range 3–10 linear gradient, GE Healthcare) were rehydrated with 345 µL of Cy3- and Cy5-labeled protein mixture diluted in rehydration buffer containing 1% Pharmalytes, 1300 mmol/L DTT, 4% CHAPS, 7 mol/L urea, and 2 mol/L thiourea on a Protean IEF cell system (Biorad) for 24 h without applying any current. After rehydration, IEF was performed at 300 V for 3 h, and then a gradient to 1000 V for 6 h and a gradient to 8000 V for 3 h; it was finally completed at 8000 V for 3 h. The temperature was maintained at 20 °C. After IEF, the IPG strip gels were removed and equilibrated with buffer containing 6 mol/L urea, 0.1 mmol/L Tris-HCl pH 8, 30% The Journal of Proteome Research • Vol. 7, No. 8, 2008 3573
technical notes glycerol, and 2% SDS for 10 min at room temperature. The equilibrated IPG gels were applied to the top of a 12.5% polyacrylamide gel and sealed with low melting temperature agarose (GE Healthcare). SDS-PAGE was performed with the Ettan Dalt II system (GE Healthcare) at a constant 17 W per 6 gels for 18 h. All electrophoresis procedures were performed in the dark. Image Acquisition and Bioinformatic Analysis. Gels between two low-fluorescence glass plates were scanned with a Typhoon 9400 fluorescence scanner (GE Healthcare) and saved in.gel format with ImageQuant software (GE Healthcare). The excitation wavelength for Cy3 is 532 nm and for Cy5 633 nm, and the emission wavelengths are 580 and 670 nm, respectively. Platinum 6.0 gel image analysis software was used to analyze the images (GE Healthcare). Briefly, after automatic spot detection and background subtraction from each gel image (Cy3 and Cy5), spots were edited manually, such as adding, splitting, and removal of artifacts, for example, dust particles or streaks detected as protein spots. One gel was chosen as the master gel and used for the automatic matching of spots in the other 2-D gels images corresponding to either the Cy3 or Cy5 labeled sample. Total spot volume was calculated for each image and each spot assigned a normalized spot volume as a proportion of this total value. After editing and manual matching, the images were analyzed for protein spot differences. Polypeptide spots were considered to have significantly different normalized spot volumes between control (6 different image gels) and treated cells (6 different image gels) according to threshold (1.2-fold) (p < 0.05) and three criteria: (1) presence of the spot on all gel images used for the bioinformatic analysis; (2) reproducible modulation of the spot on 2 gels performed with the same dye; and (3) modulation of the spot detected independently of the dye used. Protein Identification. Five micrograms of sample was labeled with Cy3 or Cy5 saturation dyes as described above and 495 µg of unlabeled sample was added to Cy-labeled sample, yielding 500 µg of sample (intracellular and secreted proteins), which underwent 2D electrophoresis, exactly as described above. Samples were run on several 2D gels to identify all the spots selected by bioinformatic analysis. The 2D gels were then scanned while still between two lowfluorescence glass plates with Typhoon 9400 fluorescence scanner (GE Healthcare) at appropriate excitation and emission wavelengths. The 2D gels were then Coomassie blue-stained according to the protocol previously described by Neuhoff et al.19 Gels were fixed in 50% v/v ethanol containing 2% w/v orthophosphoric acid for at least 2 h, rinsed three times in distilled water, incubated for 1 h in 34% v/v methanol containing 17% ammonium sulfate, 2% w/v orthophosphoric acid, and 1 g CBB G-250 and then incubated overnight. Coomassie bluestained gels were digitized at 200 density per inch resolution with an Imagescanner scanner (GE Healthcare). The two gel images were then imported into Platinum software and the spots of interest previously selected in analytical gels were matched to those on the preparative gels and excised in 2D gel. Proteins were then identified by an in-gel digestion method. Briefly, the gel plugs excised were washed with water and destained with 50 mmol/L Tris buffer pH9/50% ACN. They were then dried completely in a Speedvac evaporator and rehydrated with 25 mmol/L Tris pH 9 containing 50 ng of trypsin (Promega, Madison, WI). After digestion overnight at 37 °C, the supernatant was removed, the gel pieces were washed twice with 0.1% TFA/50% ACN and the supernatant was removed again. The 3574
The Journal of Proteome Research • Vol. 7, No. 8, 2008
Dupont et al. collected supernatants were pooled, concentrated in SpeedVac evaporator, desalted with ZipTip C18 (Millipore, Bedford, MA), and eluted with 0.1% TFA/50% ACN. One microliter of this solution was then mixed with 1 µL of matrix solution (50 mg/ mL of R-cyano-4-hydroxycinnamic acid dissolved in 0.1% TFA/ 50% CAN) and spotted onto the MALDI-TOF MS target. MALDITOF MS was then performed with a Voyager DE STR mass spectrometer (PerSeptive Biosystems, Framingham, MA, USA) equipped with a 337.1 nm nitrogen laser and the delayed extraction facility (125 msec). All spectra were acquired in a positive ion reflector mode under 20 KV voltage, 61% grid. Typically, 200 laser shots were recorded per sample. DataExplorer software version 4.0 (PerSeptive Biosystems) was used to calibrate the resultant spectra internally with trypsin autolysis products ([M + H]+ 842.51; [M + H]+ 2211.10; [M + H]+ 2383.95) and externally with lysozyme and then to pick peaks, with a threshold that depended on the background, a resolution >10 000, and contaminant ions not excluded. Tryptic monoisotopic peptide masses were identified by peptide mass fingerprinting with Profound (http://Prowl.rockfeller.edu/prowlcgi/profound.exe) software against the NCBI nr (2007/08/04) (118 570 sequences) and MS-Fit (http://prospector.uscf. edu/prospector/4.08) software against Swiss-Prot (SwissProt.2007.04.19) (16 150 entries) with the following parameters: Human species, one missed cleavage site and a mass tolerance setting of 25 ppm. Partial chemical modifications such as oxidation of methionine and carbamidomethylation of cysteine were considered for the queries. The criteria used to accept identifications included the extent of sequence coverage, the number of peptides matched (minimum of four), the Mowse probability score (minimum of 70), the mass accuracy and whether human protein appeared as the top candidate in the first-pass search with no species restriction. Identifications were accepted when peptides matched multiple members of a protein family only when the top candidates were obtained from the fractions of at least 2 mass spectra of trypsin digest of spots from two 2D gels, and theoretical and experimental Mr and pI were expected to be similar. Otherwise, the identification was not considered valid. Western Blotting Analysis. After incubation in serum-free culture medium with or without oxidized LDL (100 µg/mL) or native LDL (100 µg/mL), MDM was washed three times with PBS and scrapped into lysis buffer. Protein concentrations were measured with the commercial Bradford reagent (Bio-Rad, Hercules, CA, USA) with BSA as a protein standard. Denatured protein samples (20 µg) of cell lysates were separated by SDSPAGE (9% polyacrylamide). Molecular weight markers were purchased from Bio-Rad. The proteins were transferred to Hybond C membranes (GE Healthcare) for 60 min at 125 mA and equivalent total protein loads were confirmed visually by Ponceau red staining of the membranes. Membranes were then blotted with antibodies against Hsp70 (SPA-810, Stressgen, 1:1000), Grp 75 (SPA-825, Stressgen, 1:2000), Hsc-70 (SPA-815, Stressgen, 1:2000), Hsp60 (SC-1052, SantaCruz, 1:500), gelsolin (G4896, Sigma Aldrich, 1:5000), CapG (15–28822527F, Genoway, 1:20 000), CapZ (A21950, Interchim, 1:15 000), β-Actin (A431602, Sigma, 1:5000), catalase (C0979, Sigma, 1:2000), pyruvate kinase (ab6191, Abcam, 1:10 000), vimentin (ab15248, Abcam, 1:3000), and carbonic anhydrase II (ab6621, Abcam, 1:5000). The blots were subsequently washed in PBS-Tween 20 and incubated with the appropriate secondary antibodies (peroxidase-conjugated antimouse and antirat IgG, GE Healthcare) at a dilution of 1: 5000 for 1.5 h. The membranes were washed three times
technical notes
HSP Proteins in Human Macrophages Treated by oxLDL
Figure 1. Morphology of human macrophages used in this study. (A) Time-dependent effect of native LDL and oxLDL on primary culture of MDM at day 12. (B) Morphology of human macrophages treated with 100 µg/mL of oxLDL for 24 h. Example of 2 primary cultures of MDM.
in PBS-Tween 20 for 15 min, incubated with enhanced chemiluminescence (ECL) reagents (GE Healthcare) according to the instruction manual, exposed to radiographic film and then digitized with an Imagescanner scanner. The intensity of the bands was quantified with Quantity One image analyzer software (Biorad). Data are presented as means ( SEM. Differences between means were considered significant when p < 0.05, according to Student’s t test.
Results Effect of oxLDL Loading on Intracellular and Secreted Proteins by Human MDM. To determine whether cellular cholesterol accumulation influences protein synthesis, human MDM were exposed to 100 µg/mL oxLDL protein for 6, 12, 24 and 48 h in serum-free medium (Figure 1A). As previously shown, these conditions cause a substantial elevation in the The Journal of Proteome Research • Vol. 7, No. 8, 2008 3575
technical notes
Dupont et al.
Figure 2. (A) Representative 2D-DIGE gel analysis of intracellular proteome of human macrophages. An average gel was established from the six gels performed for each group (untreated and oxLDL-treated macrophages), with Platinum 6.0 software. Gels were loaded with 5 µg protein. The position of molecular weight (Mr) standards are indicated on the left and the pI are indicated on the bottom of the gel. Protein names in squares and circles indicate those respectively up- and down-regulated by oxLDL treatment. (B) Representative 2D-DIGE gel analysis of the secretome of human macrophages. An average gel was established from the four gels performed for each group (untreated and oxLDL-treated macrophages), with Platinum 6.0 software. Gels were loaded with 5 µg protein. The position of molecular weight (Mr) standards are indicated on the left and the pI are indicated on the bottom of the gel. Protein names in squares and circles indicate those respectively up- and down-regulated by oxLDL treatment. Spots 3 and 5 represented in triangles were present only in oxLDL-treated macrophages.
cellular content of both free cholesterol and cholesteryl ester.20 Figure 1B shows the morphology of human MDM treated or not with 100 µg/mL oxLDL for 24 h and used for the proteomic differential analysis. Six 2D gels, loaded with 5 µg intracellular proteins of each condition (untreated and oxLDL-treated) were run from 3 different primary cultures of MDM treated or not with 100 µg/ mL oxLDL for 24 h. Cy3- or Cy5-labeled samples were able to resolve as many as 1102 polypeptidic spots on average for untreated macrophages and 1248 for oxLDL-treated macrophages. An average image was established from the six scanned images for each condition (Figure 2A). This differential analysis found 20 proteins with differential abundance levels and a 3576
The Journal of Proteome Research • Vol. 7, No. 8, 2008
statistically reproducible difference over the series of gels. Nine were up-regulated (spots 1, 7, 10, 11, 12, 14, 17, 18, and 20) and 11 down-regulated (spots 2, 3, 4, 5, 6, 8, 9, 13, 15, 16, and 20) in oxLDL-treated macrophages. The intensity of each spot was calculated as the mean ( SD and expressed as a percentage of normalized volume. The same approach was used for the proteins secreted by control and oxLDL-treated macrophages. Four 2D-DIGE gels resolved an average of 854 spots for untreated macrophages and 884 for the oxLDL-treated macrophages. An average image was established from the four scanned images for each condition (Figure 2B). This differential analysis found 6 proteins with differential abundance levels and a statistically reproducible difference over the series of gels.
technical notes
HSP Proteins in Human Macrophages Treated by oxLDL Three were upregulated (spots 2, 4, and 6), one was downregulated (spot 1) and two were induced (spots 3 and 5) in oxLDL-treated macrophages. The intensity of each spot was calculated as mean ( SD and expressed as percentage of normalized volume. To obtain information about the identity of the differentially expressed proteins, we used MALDI-TOF mass spectrometry of a preparative gel of intracellular and secreted proteins for tryptic peptide fingerprinting, in accordance with recent guidelines.21 Table 1 summarizes the identity and factor of variation of the intracellular proteins differentially expressed by oxLDLtreated macrophages. Criteria for identification acceptance were: same top candidate proteins obtained from fractions of at least 2 mass spectra of two 2D gels with 4 peptides matched, a MOWSE score of at least 70 and correspondence between the theoretical and experimental Mr and pI of each identified protein. Of the 20 proteins differentially expressed, 3 were not identified by mass spectrometry (spots 1, 2 and 6) and one of these, spot 2, was identified by comparison to the reference macrophage map.7 Three proteins were identified as enzymes involved in metabolism: aconitase hydratase (spot 2), R-enolase (spot 9), and aldolase 1-epimerase (spot 14). The function of one protein (protein 71–7A, spot 10) when expressed in macrophages is unknown. Another protein is specific to this cell type: monocyte protein 5 (spot 4). Four proteins were identified as involved in cytoskeletal stabilization: merlin (spot 3), fascin (spot 8), γ-Actin (spot 17) and β-Actin (spot 18). Interestingly, one protein identified (catalase, spot 5) is involved in oxidative stress and two belongs to the HSP family: Grp75 (spot 7) and Hsp7c (spot 11). Table 2 summarizes the identity and factor of variation of the proteins secreted differentially by oxLDL-treated macrophages. The six spots were identified by mass spectrometry, by application of the same criteria as for intracellular proteins. Two of the proteins identified are enzymes involved in metabolism: carbonic anhydrase II (spot 5) and proteasome subunit beta type 2 (spot 6). Two more are involved in cytoskeletal structures (moesin, spot 3 and vimentin, spot 4) and another is involved in proteolyzis of extracellular matrix (MMP-9, spot 1). Spot 2 was identified as a protein belonging to the HSP family: Hsp70. Differential Regulation of HSP Members in Native LDL and oxLDL-Treated Macrophages. First, we tested macrophage treatment by native LDL and oxLDL over time. Figure 3A shows that the most pronounced effect of oxLDL on intracellular expression of Hsp70 was found at 24 and 48 h. Second, dosedependent treatment of macrophages by oxLDL showed that the most pronounced effect was obtained with 100 µg/mL (Figure 3B). The time of treatment and concentration of oxLDL used for the proteomic analysis is consistent with the more pronounced effect of oxLDL on macrophage expression of Hsp70. The up-regulation of Hsp70 in oxLDL-treated macrophages observed with the 2D-DIGE technique was confirmed by Western blotting, which also showed that native LDL had no effect on Hsp70 expression (Figure 3C). Given the known role of HSPs in atherosclerosis, we used Western blotting to analyze the expression of several members of the HSPs family: Hsp27, Hsp60, Hsc70, and Grp75. We did not detect the presence of Hsp27 in any human macrophages, regardless of whether they had or had not been treated by either native LDL or oxLDL (not shown). Figure 4A shows that Hsp60 expression was up-regulated in oxLDL-treated macrophages but not those treated with native
LDL. The latter, as shown in Figure 4B, modulated Hsc70 slightly, whereas oxLDL-treated macrophages had no effect on it. Figure 4C shows that Grp75 expression was down-regulated in ox-LDL-treated macrophages but not in their native LDLtreated counterparts. Modulation of the Gelsolin Protein Family and Metabolic Enzymes in oxLDL-Treated Macrophages. Several proteins involved in cytoskeletal stabilization, such as gelsolin and β-Actin, were modulated in oxLDL-treated macrophages. Figure 5 shows the Western blot results, which demonstrate the modulation of these two proteins and of CapG and CapZ proteins by oxLDL treatment of human MDM. Only the upregulation of β-Actin was statistically significant (p < 0.001), although trends were observed for gelsolin and CapZ (p ) 0.07). In contrast, oxLDL treatment appeared to be down-regulated CapG (p ) 0.09). Gelsolin and β-Actin were also found to be up-regulated by 2D gel analysis. We observed a statistically significant down-regulation by oxLDL treatment of the expression of the two metabolic enzymes tested, catalase (p < 0.05) and pyruvate kinase (p < 0.001). This finding is consistent with the 2D gel analysis. Modulation by oxLDL Treatment of Macrophage Secretion of Proteins. Of the six secreted proteins found to be modulated by oxLDL, three were tested by Western blot (Hsp70, carbonic anhydrase II, and vimentin) and one was tested by zymography (MMP-9). OxLDL treatment of human macrophages induced up-regulation of Hsp70 (p < 0.05), carbonic anhydrase III (p < 0.05) and vimentin (p < 0.05) and downregulation of MMP-9 activity (p < 0.01) (Figure 6), consistent with the 2D gel analysis.
Discussion The objective of this study was to determine the effect of oxLDL on intracellular and secreted proteins of human macrophages and at the same time to test the relevance of differential proteomic analysis by a sensitive technique. A new generation of Cys-reactive cyanine dyes allows the identification of more features than do Lys-reactive dyes. This 2D-DIGE technique increases sensitivity when sample quantities are scarce.10 We used 2D-DIGE with quantitative image analysis and sequencing mass spectrometry to investigate changes in the protein expression profiles of human MDM treated with oxLDL. This sensitive 2D-DIGE technique allowed us to assess the differential expression of peptide spots by macrophages treated for 24 h with oxLDL, compared with no treatment, and showed 20 polypeptide spots in intracellular proteins and 6 in secreted proteins. DIGE proteomic analysis also showed that treatment of human macrophages by oxLDL upregulated expression of 9 intracellular and 3 secreted proteins, downregulated 11 intracellular proteins and 1 secreted protein, and induced two secreted proteins. To perform mass spectrometry, we used 5 µg of proteins labeled with one Cys-cyanine dye, as we did for analytical 2D gel that we mixed with 495 µg of unlabeled proteins. Fuji et al.,11 on the other hand, used 500 µg of labeled proteins. We preferred to avoid the expense of 500 µg of labeled proteins as well as the problems it might present in preparing samples for mass spectrometry. We were able to identify 17 proteins from 20 spots of intracellular proteins and one more with the 2D map;7 2 could not be identified. Good resolution of 2D gel for mass spectrometry of the proteins secreted into the medium required The Journal of Proteome Research • Vol. 7, No. 8, 2008 3577
3578
P35240
Q9Y3Z3
P04040
3
4
5
The Journal of Proteome Research • Vol. 7, No. 8, 2008
R-enolase
Q16658
P06733
O75665
P11142
P52209
Q567T7
Q96C23
O14773
8
9
10
11
12
13
14
15
P07355
38.57
46.35
7.57
8.47
27
50
8.66
7.75
6.3
6.45
6.4
6.25
6.6
7.9
7.9
7.55
7.3
7.55
7.7
6.5
9
8
8
6.55
6.65
7
exp. pI
21/36
8/27
6/20
8/41
4/20
5/33
6/22
6/18
11/30
7/35
4/31
7/20
8/22
10/42
7/33
5/17
7/22
number of matched peptides/ total peptides
52.5
21.4
23.7
26.4
9.6
15.1
13.2
18.5
19.8
12.7
5.3
20.3
14.4
18.3
19.5
8.5
10.6
sequence coverage %
1.08 108
16841
586
28879
364
1848
1045
229
108443
6839
64.8
1286
490
6597
2132
148
887
MOWSE score
257.81 ( 34.77 324.18 ( 54.69 (LDL)
99.50 ( 33.15 132.33 ( 37.42 (LDL) 125.37 ( 25.73 85.74 ( 8.47 (LDL) 382.89 ( 70.68 299.54 ( 45 (LDL) 20.65 ( 12.31 37.24 ( 9.2 (LDL) 42.28 ( 24.12 109.18 ( 6 (LDL) 85.22 ( 7.53 63.46 ( 7.77 (LDL)
8.13 ( 2.74 10.12 ( 1.68 (LDL) 12.58 ( 2.57 7.32 ( 2.7 (LDL) 16.22 ( 5.51 12.26 ( 5.49 (LDL) 24.88 ( 5.69 18.12 ( 6.6 (LDL) 38.46 ( 7.04 29.22 ( 4.61 (LDL) 109.66 ( 27.23 82.72 ( 25.54 (LDL) 11.48 ( 2.91 16.59 ( 6.47 (LDL) 59.70 ( 11.04 45.14 ( 4.56 (LDL) 52.09 ( 10.80 44.66 ( 10.9 (LDL) 125.26 ( 27.54 153.03 ( 29.70 (LDL) 93.51 ( 17.70 112.47 ( 16.06 (LDL) 8.46 ( 1.65 11.74 ( 2.51 (LDL) 36.55 ( 10.55 29.61 ( 6.23 (LDL)
mean of normalized volume of spotb ( × 103)
>1.26
2.58
>1.81
1.20
>1.22
