Proteomic Analysis of Chemonaı¨ve Pediatric Osteosarcomas and Corresponding Normal Bone Reveals Multiple Altered Molecular Targets Cecilia Folio,†,# Marı´a I. Mora,‡,# Marta Zalacain,† Fernando J. Corrales,‡ Victor Segura,‡ Luis Sierrasesu ´ maga,† Gemma Toledo,§ Mikel San-Julia´n,| and Ana Patin ˜ o-Garcı´a*,† Laboratory and Department of Pediatrics, University Clinic of Navarra, Pamplona, Spain, Division of Hepatology and Gene Therapy, Proteomics and Bioinformatics Unit, Centre for Applied Medical Research (CIMA), Pamplona, Spain, Department of Pathology, M.D. Anderson International Espan ˜ a, Madrid, Spain, and Department of Traumatology and Orthopaedic Surgery, University Clinic of Navarra, Pamplona, Spain Received February 10, 2009
With a view to identify the proteins involved in transformation, metastasis or chemoresistance in pediatric osteosarcoma, we carried out a new experimental approach based on comparison of the proteomic profile of paired samples of osteosarcoma and normal bone tissues from the same patient. The proteomic profiles of five pairs of cell lines (normal vs tumoral) were obtained by two-dimensional difference gel electrophoresis. We detected 56 differential protein spots (t test, p < 0.05). Subsequent protein characterization by nano-LC-ESI-MS/MS enabled us to identify some of these proteins, 16 of which were chosen on the basis of the change of their relative abundance between osteosarcomas and paired normal bones and also because their involvement was supported by the genomic analysis. Two of the 16 proteins, Alpha-crystallin B chain (CRYAB) and ezrin (EZR1), were selected for further studies: an immunohistochemical analysis of a TMA (tissue microarray) and real-time PCR for a set of 14 osteosarcoma/normal-bone pairs. The results of this second tier of studies confirmed that there were significant increases in the amounts of CRYAB and ezrin, especially in advanced stages of the disease. Our overall conclusion is that proteomic profiling of paired samples of osteosarcoma and normal bone tissues from the same patient is a practicable and potentially powerful way of initiating and proceeding with a search for proteins and genes involved in pediatric osteosarcoma. Keywords: pediatric osteosarcoma • proteomics • CRYAB • ezrin
Introduction Osteosarcoma is the most common primary malignant tumor of bone in children and adolescents and is characterized by production of osteoid. It most frequently occurs in the second decade of life; about 60% of patients are under 25 yearsold, whereas only 30% are over 40.1,2 At present, the standard treatment for high-grade osteosarcoma includes neoadjuvant chemotherapy followed by surgical resection and postoperative chemotherapy. Several clinical factors have been found to be related to survival, and these include the presence of metastatic disease and the histological response to preoperative chemotherapy. Despite new therapeutic modalities, the survival rate is around 65-70%.3,4 * Corresponding author: Ana Patin ˜ o-Garcı´a, Ph.D., Laboratory of Pediatrics, University of Navarra/University Clinic, Irunlarrea SN, Los Castan ˜ os Building, 31080 Pamplona, Spain. T. +34948425600/6304/F. +34948425649. E-mail:
[email protected]. † Laboratory and Department of Pediatrics, University Clinic of Navarra. # These authors have contributed equally to the work. ‡ Centre for Applied Medical Research (CIMA). § M.D. Anderson International Espan ˜ a. | Department of Traumatology and Orthopaedic Surgery, University Clinic of Navarra.
3882 Journal of Proteome Research 2009, 8, 3882–3888 Published on Web 06/03/2009
In studies of classical high-grade osteosarcomas, few, if any, genetic alterations have been found to be common to a substantial proportion of tumors. For alterations which have been identified, different studies usually find different frequencies depending on the study design and the criteria used to determine which specimens to include. About 70% of osteosarcomas have altered karyotypes: the most frequently encountered chromosomal aberrations are an extra chromosome 1; losses of chromosomes 9, 13 (including the RB1 gene), and 17 (including the TP53 gene); and structural alterations to chromosomes 11, 19, and 20.3,5-7 At present, global proteomic profiling of human osteosarcomas has scarcely begun. To our knowledge, there are only two papers in the literature about proteomic profiling of primary osteosarcoma samples. One is by Li and co-workers,8 who use SELDI-TOF-MS to identify secreted proteins in 29 osteosarcoma specimens and thereby establish a way to differentiate between osteosarcoma and benign osteochondroma. The other paper is by Kawai and co-workers,9 who use proteomic profiling of 23 osteosarcoma biopsies to identify 10 electrophoretic protein spots potentially involved in osteosarcoma chemosensitivity. Neither of these studies included profiling the biomarkers in a matched control group. In the 10.1021/pr900113w CCC: $40.75
2009 American Chemical Society
research articles
Proteomic Analysis of Chemo-Naïve Pediatric Osteosarcomas
Table 1. Clinical Description of the Pediatric Osteosarcoma Patients from Whom Cell Lines Were Derived
c
case
location
sexa
473 475 486 491 500 524 531 554 588 595 598 604 628 631 636 648 652
Tibia Femur Tibia Femur Femur Femur Femur Femur Tibia Femur Femur Pelvis Femur Femur Femur Tibia Tibia
M F M M M F F M M F F F M M M F M
age
15 23 22 16 11 11 22 16 16 16 14 27 13 12 15 15 10
y y y y y y y y y y y y y y y y y
1 mo 6 mo 10 mo 8 mo
8 mo 9 mo 3 mo 2 mo 8 mo 9 mo
responseb
metastasis
alive
survivalc
Good Good N/Ad Good Good Poor Poor Good Poor Good N/Ad N/Ad Good Good Good Good Good
No No Yes Yes No No Yes No Yes Yes Yes Yes No No No No No
Yes Yes No Yes Yes Yes No Yes No No No No Yes Yes Yes Yes Yes
71 71 14 66 65 70 33 53 50 8 7 15 22 21 20 14 11
a Sex: M ) male, F ) female. b Response: good means that >90% necrosis was induced by the neoadjuvant chemotherapy; Poor, e90% necrosis. Survival is expressed in months since diagnosis. d N/A, not available.
work reported here, focused on the analysis of classical highgrade osteogenic sarcoma cases, we adopted a different approach. We isolated and characterized several human chemona¨ıve cell populations derived from primary osteosarcomas and their normal primary osteoblastic counterparts. Once wellcharacterized pairs of cell cultures had been established, we used two-dimensional difference gel electrophoresis (2D-DIGE) to compare their proteomic profiles. The long-term objective of such an approach is to understand the molecular mechanisms involved in the genetic etiology of osteogenic sarcoma and to unveil conserved molecular pathways.
Materials and Methods Patients and Clinical Samples. Primary tumor samples were obtained by needle biopsy from 17 high-grade osteosarcomas prior to induction chemotherapy. Corresponding normal osteoblasts were isolated when patients underwent surgery (Table 1) for tumor resection. All patients were treated and followed up at the Paediatric Oncology Unit and the Department of Orthopedic Surgery of the University Clinic (Pamplona, Spain). All the samples were obtained with written informed consent from patients, their parents, or both. Ethical approval of the study was granted by the Ethics Committee of the University Clinic. Human osteoblasts were isolated according to procedures we have published previously.10,11 Cultures of normal osteoblastic cells were passaged 1-3 times; those of tumoral cells over 20 times. In all cases, cells were 90% confluent and in logarithmic growth phase at the time of harvesting for proteomic analysis. All normal and tumor-derived cell lines were tested for the expression of alkaline phosphatase using a commercial histochemical assay (Sigma-Aldrich Quimica SA, Madrid, Spain). The presence of other markers of osteoblastic lineage, collagen R1 type I, osteopontin, bone sialoprotein and osteocalcin, was tested for by semiquantitative RT-PCR using previously described primers and conditions.12 Five pairs of cell lines were selected for comparative proteomic analysis by 2-D DIGE. The characteristics of these five cases, labeled 473, 486, 500, 595 and 604, are given in Table 1.
The remaining cases included in Table 1 were used for validation purposes. Two-Dimensional Difference Gel Electrophoresis (2DDIGE) and Imaging. All proteomic determinations were done at the Proteomics Core Facility of the Center for Medical Applied Research (CIMA), an affiliate of ProteoRed, the Spanish National Institute of Proteomic Facilities. Culture medium was removed and cells were washed three times with ice-cold PBS. Cellular pellets were solubilized in 2-D DIGE sample buffer: 7 M urea, 2 M thiourea, 4% CHAPS, and 30 mM Tris, buffered to pH 8. Protein concentration was determined using Bradford’s assay (Bio-Rad, Hercules, CA). Protein was precipitated using the Clean Up kit (Bio-Rad, Hercules, CA). Then, 50 µg of protein was taken, labeled with 400 pmol of CyDye DIGE Fluor minimal dyes (GE Healthcare, Madrid, Spain), and incubated on ice in the dark for 30 min according to the manufacturer’s instructions. Cy3 and Cy5 were used with samples; Cy2 was used with an internal control mixture composed of equal amounts of protein from all samples. Paired samples were reverse-labeled in order to prevent potential dye labeling bias. The reaction was stopped by addition of 1 µL of 10 mM lysine and incubated on ice for 10 min. Samples were up-loaded onto IPG strips (24 cm long, pH 3-11 NL, from GE Healthcare, Madrid, Spain), and subjected to isoelectrofocusing in the IPGphor IEF System (GE Healthcare, Madrid, Spain) according to the manufacturer’s recommendations. Upon completion of IEF, strips were incubated for 15 min in equilibration buffer (50 mM Tris-HCl of pH 8.8, 6 M urea, 30% glycerol, 2% SDS, and a trace of bromophenol blue) containing 0.5% DTT, and subsequently for 15 min in equilibration buffer with 4.5% iodoacetamide. For the second dimension, strips were loaded onto 12.5% polyacrylamide gels and run, at 1 W/gel, for 12-14 h., until the bromophenol blue reached the bottom end of the gel. Subsequently, 2D gels were scanned using a Typhoon Trio Imager (GE Healthcare, Madrid, Spain) at 100 λ resolution with λex/λem of 488/520 nm for Cy2, 532/580 nm for Cy3, and 633/ 670 nm for Cy5. The photomultiplier tube was set to ensure that the maximum pixel intensity was between 90 000 and 99 000 pixels. Image analysis was performed with DeCyder 6.5 software (GE Healthcare, Madrid, Spain) as specified in the Journal of Proteome Research • Vol. 8, No. 8, 2009 3883
research articles user’s manual and briefly summarized below. The differential in-gel analysis (DIA) module was used for spot detection, spot volume quantification, and volume ratio normalization of different samples in the same gel. The Biological Variation Analysis (BVA) module was then used to match protein spots on different gels and to identify protein spots with substantial differences. Manual editing was performed in the BVA module to ensure that spots were correctly matched on different gels and to remove streaks and speckles. Differentially expressed spots were considered for mass spectrometry (MS) analysis if the corresponding t test p-value was
