ARTICLE pubs.acs.org/jpr
Comparative Proteomic Analysis of Blood Eosinophils Reveals Redox Signaling Modifications in Patients with FIP1L1-PDGFRA-Associated Chronic Eosinophilic Leukemia Jean-Emmanuel Kahn,*,†,‡ Virginie Dutoit-Lefevre,‡ Sophie Duban-Deweer,§ Philippe Chafey,|| Gwenael Pottiez,§ Didier Lefranc,‡ Olivier Fain,^ Jean-Francois Cordier,# Pierre-Yves Hatron,z Olivier Bletry,† Lionel Prin,‡ and The French Eosinophil Network †
Service de Medecine Interne, H^opital Foch, 40 rue Worth, 92151 Suresnes Cedex, France Reseau Eosinophile (French Eosinophil Network), Laboratoire d’Immunologie EA2686, CHRU de Lille, France § Laboratoire de Physiopathologie de la Barriere Hemato-Encephalique, E.A. 2465, IMPRT-IFR 114, Universite d'Artois, Rue Jean Souvraz SP 18, Lens Cedex, 62307, France Institut Cochin, INSERM U1016, Plate-forme de Proteomique, CNRS (UMR8104), Paris, France ^ Service de Medecine Interne, CHU Jean Verdier, Bondy, France # Service de Pneumologie, CHU Edouard Herriot, Lyon, France z Service de Medecine Interne, CHRU de Lille, France
)
‡
bS Supporting Information ABSTRACT: The FIP1L1-PDGFRA (F/P) fusion gene, which was identified as a recurrent molecular finding in hypereosinophilic syndrome (HES), lead to a constitutively increased tyrosine kinase activity of the fusion protein. Despite data obtained in animals or cell lines models, the mechanisms underlying the predominant eosinophil lineage targeting and the cytotoxicity of eosinophils in this leukemia remain unclear. To define more precisely intrinsic molecular events associated with F/P gene, we performed a proteomic analysis comparing F/Pþ eosinophils (F/P-Eos) and eosinophils from healthy donors (C-Eos). Using 2D-DIGE and mass spectrometry techniques, we identified 41 proteins significantly overexpressed between F/P-Eos and C-Eos. Among them, 17.8% belonged to the oxidoreductase family. We further observed a down-expression of peroxiredoxin-2 (PRX-2) and an overexpression of src-homology-2 domain containing tyrosine phosphatase (SHP-1), enzymes regulating PDGFR downstream pathways, and especially intracellular reactive oxygen species (ROS) production. This profile, confirmed in immunoblot analysis, appears specific to F/P-Eos compared to controls and patients with idiopathic HES. In this clonal disorder possibly involving a pluripotent hematopoietic stem cell, we postulate that the well documented relationships between PDGFRA downstream signals and intracellular ROS levels might influence the phenotype of this leukemia. KEYWORDS: hypereosinophilic syndrome, chronic eosinophilic leukemia, eosinophil, FIP1L1-PDGFRA, proteomic analysis, mass spectrometry, peroxiredoxin-2, SHP-1, 15 lipoxygenase-1, reactive oxygene species, hydrogen peroxide
’ INTRODUCTION Hypereosinophilic syndromes (HES) are a rare and heterogeneous group of hematologic disorders, characterized by a sustained and unexplained blood eosinophilia exceeding 1.5 109/L. Two main physiopathological mechanisms have been identified so far. First, in lymphocytic variant of HES, extrinsic eosinophilia results from the overproduction of eosinophilopoietic cytokines (mainly interleukine-5) by abnormal circulating T-cell subsets, which usually express an aberrant surface phenotype. Second, in chronic eosinophilic leukemia (CEL), intrinsic eosinophilia arises from genetic abnormalities leading to a clonal and uncontrolled expansion of the myeloid linage, including the eosinophil lineage. r 2011 American Chemical Society
The FIP1L1-PDGFRA (F/P) fusion gene, generated by a cryptic deletion at 4q12, encodes a constitutively active tyrosine kinase (TK), and has been recently identified as the most frequent genetic defect in HES.1 Before treatment with tyrosine kinase inhibitors (TKI), patients with F/P fusion gene display a more aggressive disease phenotype, especially with a higher frequency of cardiac involvement.2 Imatinib mesylate, the first widely used TKI, induces a complete hematological and molecular remission
Received: August 17, 2010 Published: February 08, 2011 1468
dx.doi.org/10.1021/pr100836p | J. Proteome Res. 2011, 10, 1468–1480
Journal of Proteome Research
ARTICLE
Table 1. Characteristics of FIP1L1-PDGFRAþ (F/P) and Idiopathic HES Patients (i-HES)a F/Pþ patients (n = 7)
a
i-HES patients (n = 8)
sex/age
Eos ( 109/L)
symptoms
sex/age
Eos ( 109/L)
symptoms
FP1
M/37
42.6
skin, heart
i-HES1
M/50
1.30
skin, lung
FP2
M/46
22.62
GI, lung
i-HES2
F/33
1.91
GI
FP3
M/60
7.05
none
i-HES3
M/55
0.83
skin, GI
FP4
M/46
2.78
lung, skin
i-HES4
M/55
5.26
skin
FP5
M/35
6.52
none
i-HES5
M/68
2.90
skin
FP6
M/43
5.80
lung
i-HES6
F/32
16.0
FP7
M/64
8.60
none
i-HES7
M/75
6.2
none
i-HES8
M/50
3.91
skin, GI
GI
Eos, blood eosinophil count at the time of eosinophils isolation; GI, gastro-intestinal tract.
in almost all F/Pþ patients, due to its high affinity for the ATPbinding site of PDGFRR.1,3 The oncogenic properties of the F/P gene have been largely studied in various models but remain unclear. Transfection of F/P in BaF3 cells leads to IL-3 independent growing of this murine hematopoietic cell line.4 In vitro, the TK activity generated by F/P transfection in CD34þ hematopoietic progenitors stimulates eosinophil differentiation and activates different signaling pathways as signal transducers and activators of transcription (STAT)-5, extracellular signal-regulated kinase (ERK) 1/2, p38 mitogen-activated protein kinase, phosphoinositol-3 kinase (PI3-K), which in turn regulate cell proliferation and survival.5 However, contradictory results have been obtained using in vivo mouse model, where transplantation of F/P-transfected hematopoietic stem cells (HSC)/progenitors, in the absence of IL-5, is unable to spontaneously induce generation of eosinophils.6 Additionally, in F/Pþ patients, the nearly exclusive expansion of eosinophils remains unexpected while F/P fusion gene has been identified in almost all human circulating hematopoietic cells types (both from myeloid and lymphoid lineages).7 Nevertheless, Fukushima et al. recently showed that F/P gene is able to promote ex vivo eosinophil development from various murine hematopoietic progenitors, including c-KithighSca-1þLineagecells, common myeloid progenitors, megakaryocytic/erythrocyte progenitors, and common lymphoid progenitors, by up-regulation of C/EPBR, GATA-1, GATA-2 expression and downregulation of PU.1, majors transcription factors involved in eosinophilopoiesis. Due to the rarity of the disease and to the difficulty to collect sufficient amounts of blood eosinophils, these data were not confirmed on human F/Pþ eosinophils, which remain the key cell of this leukemia. In the present work, we used a global proteomic approach by 2-D Fluorescence Difference Gel Electrophoresis (2D-DIGE) to identify the modification of proteome induced by F/P fusion gene, using for the first time human circulating eosinophils purified from F/Pþ patients (F/P-Eos) and healthy volunteers (C-Eos). We identify in F/P-Eos several proteins involved in redox signaling and oxidative stress which were either upregulated as the phosphotyrosine phosphatase (PTP) src-homology-2 domain containing tyrosine phosphatase 1 (SHP-1), or down-regulated as peroxiredonin-2 (PRX-2) or 15 lipoxygenase1 (15LOX-1) compared to C-Eos. Their possible implications toward the promotion of the F/Pþ eosinophil lineage, the leukemic process and the cytotoxic status of F/P-Eos are discussed.
’ MATERIALS AND METHODS Chemical and Reagents
Unless otherwise specified, generic chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) and GE Healthcare (Uppsala, Sweden). Reagents for 2D-DIGE, scanner and software for images analysis were purchased from GE Healthcare. All buffers were prepared with Milli-Q water. All the chemicals and biochemicals used in this study were of analytical grade. Cell Culture
EOL-1 cell line was kindly provided by Pr M. Capron (U995 Institut National de la Sante et de la Recherche Medicale, Lille France). Cells were grown in RPMI 1640 supplemented with 1% Glutamine, 2.5% Hepes, 0.04% Gentamycin and 10% fetal calf serum. Imatinib was purchased from Santa Cruz, stored in water and diluted in RPMI. Cell cultures were initiated at 105 cells/mL, with or without imatinib. Cell counts and viability (trypan blue exclusion) were determinated using Countess Cell Counter (Invitrogen). Mean viability of EOL-1 were 94% at H0, 93, 93 and 89% at H12 and 92, 81 and 81% at H24 for 1, 10, and 100 nM concentrations of imatinib, respectively. Patients
Inclusion criteria for HES were documentation of a persistent peripheral eosinophils count of g1.5 109/L, with or without signs or symptoms of end organ involvement, for which another etiology could not be found.8 F/P mutation analysis was assessed both by nested polymerase chain reaction (PCR) and fluorescence in situ hybridization (FISH).9 Patients lacking F/P gene, other chromosomal translocations or peripheral T cell abnormalities (e.g., CD3-CD4þ, CD3þCD4-CD8-Rβþ, and CD3þCD7or clonal TCR rearrangement by PCR) were classified as idiopathic-HES (i-HES). Some main characteristics of F/Pþ and i-HES patients are detailed in Table 1. All peripheral blood samples from patients were obtained at diagnosis, before initiation of any HES-related treatment (e.g., imatinib, steroid, interferon alpha, hydroxyurea, histamine antagonists...). Healthy volunteers did not have any history of atopic disease and were free from any treatment. Peripheral blood samples (50 mL in F/Pþ and i-HES patients, and 100 mL in healthy volunteers) were collected after informed consent. Ethical approval was obtained from Ethics committee. Purification of Eosinophils
Eosinophils were purified as previously described.10 Briefly, after dilution of sample 1:1 in phosphate-buffered saline (PBS), 1469
dx.doi.org/10.1021/pr100836p |J. Proteome Res. 2011, 10, 1468–1480
Journal of Proteome Research granulocytes and erythrocytes were separated from peripheralblood mononuclear cells by centrifugation (700 g for 20 min at room temperature, without break) on Ficoll Plaque Plus. Directly onto the remaining pellet of polymorphonuclear cells, a two stephypotonic lysis was performed to remove the erythrocytes. Granulocytes were counted in with PBS, and then resuspended in ice-cold filtered PBS-BSA 0.5% for an incubation with antiCD16-coated magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany) at 4 C. Eosinophils were eluted in CD16-negative fraction following the passage in MACS system (Miltenyi Biotec). Purity of eosinophils was assessed by MayGrunwald-Giemsa staining, and was greater than 98% in all cases. Eosinophils were washed in PBS and then immediately stored at -80 C in aliquots containing 10 millions of cells. Sample Preparation for Proteomic Analysis
Purified eosinophils were lysed in buffer [8 M urea, 2 M thiourea, 50 mM DTT, 4% (w/v) CHAPS, 1 mg/mL Perfabloc], incubated for 1H min at 4 C on a stirring wheel and centrifuged at 20 000 g for 1 h at 10 C. All samples were treated with 2D Clean-Up kit. The resultant dry pellets were resuspended into lysis buffer without DTT, and the pH was set to 8.5 with 1.5 M Tris-base. Protein concentration of samples was measured using by the method of Bradford (Bio-Rad Laboratories, Hercules, CA), and were further adjusted to 2,5 mg/mL. Finally, protein quantification and the quality of sample were controlled by 1D SDS-PAGE. 2D-DIGE
We performed four gels, each containing whole eosinophil extract from one F/Pþ patient (50 μg), one healthy volunteer (50 μg), labeled either with Cy3 or Cy5 fluorescent dyes and from internal standard (IS) labeled with Cy2 fluorescent dye (50 μg). A reciprocal labeling of F/P-Eos and C-Eos samples was performed in order to have half sample of each group labeled with Cy3 and the other half with Cy5. The IS resulted from the pool of an equal amount of aliquots prepared from 4 F/Pþ patients and 4 healthy donors samples used in this experiment. All samples were separately incubated with the corresponding CyDye 30 min on ice, in the dark, and the reaction was stopped by an excess of lysine for 10 min. These labeled samples were then combined for 2D-DIGE analysis. Samples containing 150 μg of labeled proteins were mixed with rehydratation buffer [8 M urea, 2 M thiourea, 10 mM DTT, 2% (w/v) CHAPS, 1.0% (v/v) pH 3-11 IPG buffer and trace of bromophenol blue] in a final volume of 350 μL and were left overnight to rehydrate into 18 cm 3-11 nonlinear immobilized pH gradient DryStrips. Isoelectric focusing was performed at 20 C using IPGphor3, succeedingly for 1 h at 150 V, 5 h at 200 V, 2 h at 500 V, then a gradient voltage was applied to reach 8000 V in 8 h, and focusing was continued at 8000 V until a total of 50 kVh was reached. After this first dimension separation, the strips were incubated at room temperature in equilibrium buffer [6 M Urea, 2% (w/v) SDS, 30% (v/v) glycerol, 50 mM Tris-HCl pH 8.8] with 1% (w/v) DTT for 15 min and subsequently in the same buffer containing 4.7% (w/v) iodoacetamide instead of DTT for another 15 min. Equilibrated strips were rinsed with running buffer [25 mM Tris, 192 mM glycine and 0.1% (w/v) SDS], placed onto homemade 8-18% gradient polyacrylamide gels and overlaid with agarose solution [0.5% low-melting agarose and trace of bromophenol blue in running buffer]. Electrophoresis was performed at 20 C, in the dark at 4W/gel for 30 min then 18 W/gel for approximately 6 h (until the
ARTICLE
bromophenol blue dye front reached the bottom of the gel). The gel images were collected using a Typhoon Variable Mode Imager 9400 as described elsewhere,11 and imported in DeCyder software 6.5. A spot was considered as differentially expressed between F/Pþ-Eos compared to C-Eos if fold change of protein spot abundance was larger than þ1.5 or smaller than -1.5 and if t-test p-value were less than 0.05. Protein Identification by MS and Database Searching
Two semipreparative 2D-gels (one containing F/P-Eos extract and one containing C-Eos extract, which were both used in the first part of the experiment) were performed using the protocols described above, except that 400 μg of proteins were loaded on strips. After the second dimension, gels were stained with Coomassie Brillant Blue G250 as previously described,11 and scanned with ImageScanner. Spots identified as significantly up or down-regulated on analytical gels were reported on preparative gels. Protein identifications from 2-DE gels were performed using a Proteineer workflow from Bruker Daltonik GmbH (Bremen, Germany). Colloidal Coomassie blue stained spots were excised from gels with a spot picker (PROTEINEER spII) equipped with a 2 mm needle and placed into 96-well microtiter plates. In-gel digestion and sample preparation for MALDI analysis were performed according to the manufacter’s instructions using a digester/spotter robot (PROTEINEER dp) and tryptic digest kits (DP 384 standard kit, Bruker Daltonik). The MALDI target plate (AnchorChip, Bruker Daltonics) was covered with extracted peptides mixed with R-cyano-4-hydroxycinnamic acid matrix (0.3 mg/mL in acetone:ethanol, 3:6 v/v). The molecular mass measurements were performed in automatic mode using FlexControl 2.2 software on an Ultraflex II TOF/ TOF instrument and in the reflectron mode for MALDI-TOF peptide mass fingerprint (PMF) or LIFT mode for MALDITOF/TOF peptide fragmentation fingerprint (PFF). External calibration over a 1000-3200 mass range was performed using the [M þ H]þ monoisotopic ions of bradikinin 1-7, angiotensin I, angiotensin II, substance P, bombesin, and adrenocorticotropic hormone (clip 1-17 and clip 18-39) from a peptide calibration standard kit (Bruker Daltonik). Briefly, an accelerating voltage of 25 kV, a reflector voltage of 26.3 kV and a pulsed ion extraction of 160 ns were used to obtain the MS spectrum. Each spectrum was produced by accumulating data from 800 laser shots. A maximum of five precursor ions per sample were chosen for LIFT-TOF/ TOF MS/MS analysis. Precursor ions were accelerated to 8 kV and selected in a timed ion gate. Metastable ions generated by laser-induced decomposition (LID) were further accelerated by 19 kV in the LIFT cell and their masses measured in reflectron mode. Peak lists were generated from MS and MS/MS spectra using Flexanalysis 2.4 software (Bruker Daltonik). Database searches, through Mascot (Matrix Science Ltd., London, U.K.), using combined PMF and PFF data sets were performed via ProteinScape 1.3 (Bruker Daltonik). A mass tolerance of 75 ppm and 1 missing cleavage site for PMF and MS/MS tolerance of 0.5 Da and 1 missing cleavage site for MS/MS search were allowed. Carbamidomethylation of cysteine and oxidation of methionine residues were also considered. Relevance of protein identities was judged according to the probability based Mowse score calculated with a p-value of 0.05 (P < 0.05). Immunoblotting
Immunoblotting was used to confirm expression changes detected by 2D-DIGE of SHP-1, PRX-2 and 15LOX-1. EOL-1 and eosinophils purified from F/Pþ patients (n = 7, including 1470
dx.doi.org/10.1021/pr100836p |J. Proteome Res. 2011, 10, 1468–1480
Journal of Proteome Research
ARTICLE
Figure 1. Representative images of 2-D DIGE analytical gels of eosinophil proteome in F/Pþ patients and healthy donors. Protein extracts from human eosinophils were prepared, labeled and separated by 2D-DIGE, as described in Materials and Methods. Proteome maps from (A) healthy volunteer, labeled with Cy3, and from (B) FIP1L1-PDGFRA positive patient, labeled with Cy5. (C) Color overlay of 2D-DIGE images of internal standard (labeled with Cy2, blue), C-Eos (labeled with Cy3, green) and F/P-Eos (labeled with Cy5, red) samples. All spots colored in white represent equally expressed proteins, while spots in green correspond to spots overexpressed in healthy donors and spots in red those overexpressed in eosinophils from F/Pþ patient. (D) Principal component analysis of all spots detected on analytical gels from F/P-Eos and C-Eos.
the four used in DIGE experiment), healthy donors (n = 8, including the four used in DIGE experiment) and i-HES (n = 8) were lysed in RIPA buffer [Tris 50 mM pH 7.8, 150 mM NaCl, 1% Nonidet P40, 0,1% SDS, 0,5% sodium deoxycholate] with a cocktail of protease inhibitors (Complete Mini, Roche, B^ale, Switzerland) and centrifugated to remove cell debris. Proteins concentration was determined using a Bradford assay (BioRad). Samples were boiled for 10 min at 70 C. Lysates were resolved by 1D-SDS-PAGE after loading 30 μg (for detection of SHP-1 and PRX-2) or 20 μg (for 15LOX-1) of proteins per lane on 412% mini-gels, and subsequently transferred onto nitrocellulose membrane. Membranes were incubated in TBS-T blocking buffer [15 mM Tris pH 8.0, 150 mM NaCl, 0,1% Tween] containing 5% nonfat milk for 1 h at room temperature, rinsed with TBS-T, and probed overnight at 4 C while rocking either with anti-SHP-1 and anti-PRX-2 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), anti-15LOX-1 (Cayman Chemical, Ann Arbor, MI) and anti-GAPDH (Abcam, Cambridge, MA) primary antibodies, respectively used at 1/100, 1/100, 1/15 000 and 1/ 2500 in 5% nonfat milk TBS-T. Blots were washed three times with TBS-T, and probed for 1 h with horseradish peroxidase (HRP)-conjugated antimouse IgG antibody (Sigma Aldrich, Saint-Louis, MO) used at 1/1000 for SHP-1 and PRX-2, HRPconjugated antisheep IgG antibody (Santa Cruz Biotechnology) at 1/10 000 for 15LOX-1, and HRP-conjugated antirabbit IgG antibody (Sigma) at 1/4000 or Alexa 647-conjugated antirabbit IgG antibody (Invitrogen) at 7 μg/mL for GAPDH. After washing the membranes three times in TBS-T, antibody binding was quantified by measuring chemifluorescence on Storm PhosphoImager (GE Healthcare) using ECL Plus Western Blotting detection Reagents (GE Healthcare) for HRP-conjugated antibodies. Signal intensity of SHP-1, PRX-2 and 15LOX-1 and GAPDH was quantified by densitometry using the 1D gel analysis of ImageQuantTL software (GE Healthcare) and normalized for GAPDH expression. Data on SHP-1, PRX-2 and 15LOX-1 expression in F/Pþ or i-HES patients, and healthy donors were compared using Mann-Whitney test. Immunoblots on EOL-1 were performed with the same procedures except that anti-PRX-2 primary antibody was purchased from Abcam and used at 1/500. Otherwise, HRP-conjugated antimouse IgG
antibody was used at 1/2500 for SHP-1 and PRX-2 detection. Data were analyzed by the Kruskall-Wallis test using the bootstrapping method of Monte Carlo (SPSS 12.0) with a random iteration of 30 data.
’ RESULTS Differential Proteomic Analysis of Whole Eosinophil Extract between F/Pþ Patients and Controls
We performed one 2D-DIGE experiment using whole eosinophil lysates from the first four consecutive F/Pþ patients included in our study between 2006 and 2007 (FP1 to FP4, Table 1), which were compared to four healthy donors. Representative images of analytical gels from one F/P-Eos and one C-Eos are shown respectively in Figure 1A and B, and revealed highly resolved protein maps, allowing the detection 1344 ( 57 spots per gels, and 1046 ( 84 spots matched for the four DIGE gels of our experiment. The proteome of each sample appear globally comparable. However, an overlay image of IS, F/P-Eos and C-Eos clearly identifies many differences in protein expression between F/Pþ patients and controls (Figure 1C). Several data report a great functional and phenotypic heterogeneity of blood eosinophils from healthy donors as well as from HES patients.12 Furthermore, in F/Pþ patients, it is suspected that a variable percentage of blood eosinophils carry the F/P fusion gene, suggesting that a proportion of eosinophils are reactive,and thus phenotypically different from the clonal subgroup. However, using principal components analysis, we confirmed that our eight samples could be correctly separated in 2 groups corresponding to affected and healthy patients (Figure 1D), with a greater spot maps heterogeneity observed in F/Pþ patients. Using the DeCyder Biological Variation Analysis (BVA) module, we further identified 113 spots differentially expressed between F/P-Eos and C-Eos, using a threshold of 1.5 fold change, and a p-value, calculated with Student’s t-test, lower than 0.05. Among these 113 spots, 60 were up-regulated and 53 were down-regulated in F/P-Eos. The rate of change in protein expression varied from -2.77 (overexpressed in F/Pþ patients) to þ6.25 (overexpressed in healthy donors). We further performed preparative gels with eosinophils whole cells extracts 1471
dx.doi.org/10.1021/pr100836p |J. Proteome Res. 2011, 10, 1468–1480
Journal of Proteome Research
ARTICLE
controls samples. 15LOX-1, a key enzyme implicated in arachidonic acid metabolism whose expression was reduced in F/PEos, is also a classical component of the eosinophil machinery. Focus on Proteins Related to FIP1L1-PDGFRA Signaling and Oxidative Stress
Figure 2. Differentially expressed protein spots identified by DeCyder software and characterized by mass spectrometry. On this gel are surrounded all differentially expressed spots (defined by abundance ratio upper than þ1.5 or smaller than -1.5 and a student’s t test p-value less than 0.05 between F/Pþ patients and controls) identified by Decyder software 6.5-BVA module (GE Healthcare) and which have further been identified by mass spectrometry (detailed in Table 2).
from one healthy donor and one F/Pþ patient. Only 65 spots present on analytical gels and differentially expressed between the 2 groups could be localized on preparative gels and were analyzed by MS and MS/MS. Identification of Differentially Expressed Spots by Mass Spectrometry
Out of these 65 spots excised and analyzed by MS and/or MS/ MS, 47 were identified (Figure 2, Table 2 and Supporting Information for protein’s identifications based on a single peptide in MS/MS), leading finally to 41 nonredundant proteins (a total of 51 proteins were identified, because some spots contained more than one protein). Based on the 2D-DIGE quantification approach using DeCyder software, 23 of these proteins were significantly increased in F/P-Eos, while the expression of the other 28 proteins was decreased in F/P-Eos (Table 2). The molecular functions and the biological process in which these proteins are involved were analyzed with the PANTHER classification system. Seventy five different biological process were allotted to these 41 proteins, involving lipid, carbohydrate or protein metabolism (31%), immunity or defense (13%), and cell proliferation, differentiation or cycle (10%) (Figure 3). From the 45 different molecular functions identified, 44% could be referred as enzymes, of which the oxidoreductase represent the most important group (17.8% of all molecular function hits). Among them, catalase and PRX-2 are two major enzymes with catalytic activity involved in H2O2 detoxification. Catalase, a dismutase which reduce H2O2 to H2O and O2, was identified in four isoforms (spots 12, 13, 15 and 16, Figure 2), all down-regulated between -1.53 and -1.6 fold in F/P-Eos. Only a small proportion of the identified proteins (4/41) were known to be preferentially expressed in the eosinophil lineage, such as eosinophil lisophospholipase, also named galectin 10 or Charcot-Leyden crystal protein, which is a major constituent of the human eosinophils. Among the four classical eosinophil cationic granule proteins, only the bone marrow proteoglycan, a proform of Major Basic Protein (MBP), was overexpressed in
We further investigated proteins known or suspected to regulate the downstream pathways of PDGFRA, whose TK hyperactivity is considered as the primary molecular mechanism involved in leukemogenesis of F/Pþ eosinophils. Based on recent data on eosinophils and PDGFRA-associated signaling molecules, three proteins identified in our primary analysis, PRX2, SHP-1 (also known as tyrosine-protein phosphatase nonreceptor type 6 or PTP-1C) and 15LOX-1 appeared of particular interest (Figure 4). PRX-2, which eliminates H2O2 generated following PDGF stimulation, was found at a lower ratio of -1.69 in F/P-Eos based on the proteomic analysis. Two isoforms of SHP-1, a cytosolic PTP expressed in hematopoietic cells, also known to regulate negatively regulate PDGF receptor-mediated signaling, were up-regulated in F/P-Eos, with a protein ratio of respectively þ1.55 and þ1.65. Finally, 15LOX-1 was downregulated in F/P-Eos, at a ratio of -2.43. To confirm the modification of eosinophil’s protein expression between F/Pþ patients and controls observed in the 2DDIGE experiment, we performed quantitative immunoblots using commercially available primary antibodies and chemifluorescent secondary antibodies. Western blots confirmed the increased expression of SHP-1 in F/P-Eos and the decreased expression of PRX-2 and 15LOX-1 in F/P-Eos, normalized to GAPDH expression (Figure 5). Extension to i-HES Patients
To further investigate whether the expression changes of SHP1, 15LOX-1 or PRX-2 could be associated to the F/P fusion gene or only related to the increased eosinophil number, we analyzed, by a quantitative immunoblotting approach, the expression of these three proteins in purified eosinophils from patients classified as idiopathic HES. These patients, without F/P rearrangement detected by nested RT-PCR or FISH, nor T cell abnormalities, displayed higher frequency of skin and gastrointestinal involvement (Table 1), and a reduced median eosinophils count (3.4 vs 7.1 109/L). Compared to F/P-Eos, SHP-1 expression was down-regulated and PRX-2-up-regulated in both i-HES-Eos and C-Eos (Figure 5). By contrast, as for F/P-Eos, levels of 15LOX-1 in i-HES-Eos displayed a similar decreased expression compared to controls samples. Effects of Imatinib on PRX-2 and SHP-1 Synthesis by EOL-1 Cell Line
We further evaluated, by immunoblotting, the expression of PRX-2, SHP-1 and 15LOX-1 in EOL-1, a cell line derived from the blood of a HES patient which spontaneously expresses FIP1L1PDGFRA rearrangement. The EOL-1 cells were tested in presence of various concentrations of imatinib. First, it has to be noted that in EOL-1, 15LOX-1 could not be detected at protein level (data not shown). As shown in Figure 6, imatinib decreased SHP-1 levels in imatinib-treated EOL-1, either at 1, 10, or 100 nM, after 12 or 24 h of culture. By contrast, PRX-2 levels were increased after 12 or 24 h, only at a concentration of 100 nM of imatinib.
’ DISCUSSION Using a proteomic approach with 2D-DIGE and mass spectrometry, the present work identified 41 different proteins whose 1472
dx.doi.org/10.1021/pr100836p |J. Proteome Res. 2011, 10, 1468–1480
Journal of Proteome Research
ARTICLE
Table 2. List of Identified Proteins That Were Differentially Expressed (> 1.5 Fold Ratio Increased or Decreased, with a p-value e0.05*) between Healthy Controls and Patients Carrying FIP1L1-PDGFRA Fusion Gene pIc
DeCyder
MW (kD)c seq
spot
gene
number
protein name
namea
accession
change
numbera t testb
foldb
matched
cov T
E
T
E
1 2
Major vault protein Neutral alpha-glucosidase AB
MVP GANAB
Q14764 Q14697
0.0026 -2.27 0.0035 -1.63
5.3 5.74
5.3 5.8
99.2 106.8
110 105
3
Alpha-enolase
ENO1
P06733
