ONCOGENOMICS Proteomic Identiﬁcation of the Wt-P53-Regulated Tumor Cell Secretome

Home , CYR61

Oncogene (2006) 25, 7650–7661 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ONCOGENOMICS Proteomic identiﬁcation of the wt-p53-regulated tumor cell secretome

FW Khwaja1,2, P Svoboda3, M Reed3, J Pohl3 , B Pyrzynska1,2 and EG Van Meir1,2

1Laboratory of Molecular Neuro-Oncology, Department of Neurosurgery, Winship Cancer Institute, Emory University School of Medicine, Atlanta, GA, USA; 2Department of Hematology/Oncology, Winship Cancer Institute, Emory University School of Medicine, Atlanta, GA, USA and 3Emory University Microchemical and Proteomics Facility, Emory University School of Medicine, Atlanta, GA, USA

Tumor–stroma interactions play a major role in tumor contributes to the neoplasm (Hanahan and Weinberg, development, maintenance and progression. Yet little is 2000) and that the tumor–stroma interactions are known on how the genetic alterations that underlie cell an active process initiated by transforming events transformation elicit cell extrinsic changes modulating (Bhowmick and Moses, 2005; Taieb et al., 2006). heterotypic cell interactions. We hypothesized that these Consequently, we need to understand these tumor– events involve a modification in the complement of stroma interactions to develop more effective therapies. secreted proteins by the cell, acting as mediators of We hypothesized that loss of tumor-suppressor function intercellular communication. To test this hypothesis, we during cell transformation may have cell extrinsic effects examined the role of wt-p53, a major tumor suppressor, on through the modulation of secreted factors. We focused the tumor microenvironment through its regulation of on p53, as it is frequently mutated in cancer and is secreted factors. Using a combination of 2-DE and cICAT a transcription factor that can directly control the proteomic techniques, we found a total of 111 secreted synthesis of a large number of proteins (Harris and proteins, 39 of which showed enhanced and 21 inhibited Levine, 2005). secretion in response to wt-p53 expression. The majority Tumor-suppressive p53 is best known for its role in of these were not direct targets of p53 transcription factor maintaining genomic integrity by controlling cell cycle activity, suggesting a novel role for wt-p53 in the control progression and cell survival in response to DNA of intracellular protein trafficking and/or secreted protein damage (Steele and Lane, 2005). Nevertheless, some stability. Evidence for p53-controlled post-translational studies have suggested that p53 can influence the modifications on nine secreted proteins was also found. tumor microenvironment through suppression of angio- These findings will enhance our understanding of wt-p53 genesis and tumor invasion (Van Meir et al., 1994; modulated interactions of the tumor with its environment. Zigrino et al., 2005). These processes might be influ- Oncogene (2006) 25, 7650–7661. doi:10.1038/sj.onc.1209969; enced by p53 through two mechanisms; the induced published online 9 October 2006 secretion of inhibitory factors (Van Meir et al., 1994) or the negative regulation of secreted protumorigenic Keywords: p53; proteomics; secretion; glioma; brain can- proteins (Chiarugi et al., 1998; Sun et al., 2000). While cer; two-dimensional electrophoresis p53-regulated intracellular proteins are well studied, the extracellular ones have not been systematically analysed. Identification of the p53 controlled secreted proteins will clarify how p53 loss in tumors may lead to the altered regulation and response of the tumor cells to their Introduction environment. To examine the regulation of p53 on the cell’s Traditionally, cancer formation is thought of as a cell secretome we identified secreted proteins by p53-null autonomous process driven by mutations in genes that tumor cells in the presence or absence of reconstituted ONCOGENOMICS increase cell proliferation and survival, where a tumor is wt-p53 expression. This is the first comprehensive study primarily composed of transformed cells. Increasing of how p53 plays a role in the process of transformation evidence suggests that the tumor microenvironment also through its manipulation of the tumor microenvironment. Our studies identified 50 new secreted proteins Correspondence: Dr EG Van Meir, Laboratory of Molecular Neuro- controlled by p53. These proteins have known roles in Oncology, Department of Neurosurgery, Winship Cancer Institute, cancer-related processes that are dependent on hetero- Emory University, 1365C Clifton Rd, NE, C5078, Atlanta, GA 30322, USA. typic cell–cell communication such as immune response, E-mail: [email protected] angiogenesis, extracellular matrix (ECM) interaction, FWK and EGVM designed and interpreted experiments and wrote the and cell survival. Many of these proteins are secreted manuscript. FWK performed experiments with the help of PS, MR through receptor-mediated nonclassical secretory path- and JP for the MS analyses. BP performed the microarrays and Northern blot. All authors read the manuscript. ways. These results are important to advance our Received 10 May 2006; revised 26 July 2006; accepted 27 July 2006; understanding on how tumor–stroma interactions con- published online 9 October 2006 tribute to cancer progression. wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7651 Results expression using two complementary proteomic techniques: two-dimensional gel electrophoresis (2-DE) and To identify p53-regulated secreted proteins involved in cleavable isotope-coded affinity tag (cICAT). the cell–cell communication events important for human cell transformation, we selected the LN-Z308 cell line as it was derived from a malignant human glioma that lost 2-DE of the tumor cell secretome both p53 alleles in vivo by well-characterized genetic The secreted proteins were separated by 2-DE analysis events suggesting selective pressure for their loss using nonlinear pH range of 3–10 and linear range of (Albertoni et al., 1998). Reactivating wt-p53 function 4–7 in triplicates to ensure reproducibility (Go¨ rg et al., in these cells would revert or restore the release of p53- 2004). The proteins were visualized by silver staining regulated secreted proteins and allow their identification and analysed using ImageMaster software. As a further in the conditioned media (CM). To this purpose, we precaution against artefacts, we profiled both 2024 and used isogenic clones of LN-Z308 with tetracycline- WT11 clones and only proteins found secreted in both inducible (2024 p53 tet-on) (Albertoni et al., 2002) and were retained. repressable (WT11 p53 tet off) (Van Meir et al., 1994) The protein spots were next excised from the gel, wt-p53 expression (Figure 1). These cells undergo subjected to in-gel digestion with trypsin, and identi- growth arrest but not apoptosis in response to p53 fied using matrix-assisted laser desorption/ionization (Van Meir et al., 1995) and show induction of the cell time-of-flight (MALDI-TOF/TOF) MS/MS analysis. cycle inhibitor p21 upon p53 induction (Figure 1a). We found on average >150 spots on each gel Using this system, we generated differential profiles of representing 68 individual proteins (Figure 1b; Tables the cell lines’ secretome with and without wt-p53 1–3). A semiquantitative analysis of this differential ac

2024 WT11 --LNZ308 ++- Dox --+ - + wtp53 2024 2024+wt p53 p53

p21

Actin

b MW 2024 - wt p53 2024 + wt p53

110 X-linked factor 75 KIAA0548 MMP-2 α-catenin BCL6 repressor MIF Saposin ADAM-10 KIAA0828 50 PG-M PG-M PEDF PAI-1 NF1

SPARC CYR61 33 Gal-1 Gal-1 OPN Gal-3

Transgelin 2 VEGF FGF-4 RTVP-1 25 α-2HS IGFBP6

β-2M 15 BDNF TIMP-3 14-3-3

10 Thioredoxin MT2A NTF1 TPM4-ALK oncoprotein TGF-β pH 4 pH 9 pH 4 pH 9 Figure 1 Representative 2-DE gels of secreted proteins from glioma cells with inducible wt-p53 expression. (a) Two p53-null clones (WT11 and 2024) with dox-inducible wt-p53 expression were used (See Materials and methods). Western blot shows wt-p53 induction, and downstream activation of the p21 cell cycle inhibitor, 48 h postinduction in serum-free media. (b) Secreted proteins found in the CM from uninduced (left) and wt-p53 induced (right) 2024 cells were analysed by 2-DE analysis using IEF strips pH3-10 NL and 12.5% SDS–Page. Protein spots circled indicate proteins with enhanced secretion (right) or reduced (left) in response to wt-p53. Samples were run in triplicate and location of representative proteins is indicated. Black arrow shows p53-induced post-translational modiﬁcation of Gal-1. White arrow shows the location of KIAA0828. (c) Enlargement showing acidic shift of Gal 1 in CM from wt p53 expressing cells (arrows).

Oncogene Oncogene 7652

Table 1 Secreted proteins with enhanced accumulation in CM upon p53 induction 2024 WT11

Protein name Access. no. MW pI mRNA 2-DE H:L s.d. H:L s.d. Functional category Mode of secretion Post-translational requirement

Galectin-3 gi:4504983 26.19 8.58 1.05 C 1.60 0.12 1.53 0.10 Adhesion and matrix Vesicle-mediated; Glycosylation/ interactions; apoptosis ectocytosis phosphorylation Lysyl oxidase-like pro- gi:4959425 71.10 6.32 1.05 1.65 0.07 1.65 0.00 Adhesion and matrix Classical Glycosylation tp3rgltdscee rtisi loacells glioma in proteins secreted wt-p53-regulated tein 2 interactions b-Galactosidase binding gi:12804557 14.72 5.33 C 1.36 0.18 1.27 0.12 Cell proliferation reg- Vesicle-mediated; Glycosylation lectin ulation ectocytosis Nm23 protein gi:35068 20.41 7.06 1.04 2.50 0.00 Cell proliferation/dif- Unknown non- Phosphorylation ferentiation classical pathway X-linked brain speciﬁc gi:21322252 99.19 6.00 Cell proliferation/dif- Unknown non- Phosphorylation factor ferentiation classical pathway BCL6 corepressor gi:21040336 78.85 8.28 DNA binding Vesicle-mediated; Glycosylation ectocytosis

Dickkopf-1* gi:6049604 28.67 8.80 ECM component/sig- Unknown non- Phosphorylation naling classical pathway Khwaja FW Growth arrest-speciﬁc 6 gi:7512417 43.14 5.26 1.14 2.08 0.12 1.72 0.11 ECM component Classical Glycosylation Collagen type XI alpha- gi:6165881 176.65 5.24 1.01 2.50 0.00 2.50 0.00 ECM component Classical Glycosylation 1 Proteoglycan PG-M gi:1008913 74.25 7.43 1.34 C 2.20 0.00 2.50 0.00 ECM component Non-classical; re- Glycosylation al et (V3) ceptor-mediated Galectin-1 gi:12804557 14.72 5.33 1.05 C 1.36 0.18 1.27 0.12 Cell proliferation reg- Vesicle-mediated; Glycosylation/ ulation ectocytosis phosphorylation Connective tissue gi:4503123 39.07 8.36 1.15 1.87 0.54 1.66 0.14 Cell proliferation/dif- Classical Glycosylation growth factor ferentiation CD83 antigen; activated gi:4757946 23.04 8.45 1.11 1.94 0.46 1.27 0.00 Immunity and defense Non-classical; re- Gycosylation lymphocytes ceptor-mediated KIAA0548 gi:3043620 50.08 7.61 1.17 C 2.28 0.04 2.50 0.00 Immunity and defense Unknown non- Glycosylation/ classical pathway phosphorylation b-2M gi:179318 13.73 6.06 1.12 C 1.25 0.01 1.46 0.19 Immunity and defense Classical Glycosylation Myeloid leukemia-inhibi- gi:187141 21.34 9.37 1.08 C 2.44 0.00 2.36 0.00 Immunity and defense Vesicle-mediated; Glycosylation tory factor ectocytosis Macrophage migration gi:312334 124.76 7.73 1.07 C 1.53 0.04 1.43 0.14 Immunity and defense Vesicle-mediated; Glycosylation inhibitory factor ectocytosis 2-Phosphopyruvate-hy- gi:119339 47.17 7.01 1.06 C 1.28 0.06 1.27 0.00 Metabolic enzyme Vesicle-mediated; Unknown dratase-enolase ectocytosis Saposin precursor gi:134218 58.11 5.06 1.18 C 1.43 0.03 1.39 0.00 Metabolism Unknown non- Glycosylation classical pathway Autotaxin-t gi:1160616 99.02 7.14 1.14 1.43 0.05 1.26 0.00 Metabolism Non-classical; re- Glycosylation ceptor-mediated Epididymal secretory E1 gi:48429027 16.57 7.57 1.71 0.28 2.04 0.10 Protease Ectocytosis Phosphorylation precursor u-Type plasminogen ac- gi:137112 48.53 87.80 1.09 1.58 0.11 1.40 0.11 Protease Classical Glycosylation tivator

PAI-1* gi:10835159 45.59 6.68 Serine-type protease Vesicle-mediated; Glycosylation exocytosis TIMP-3 gi:490094 23.17 8.46 1.35 2.01 0.16 1.55 0.17 Protease inhibitor Classical Glycosylatoin/ phosphorylation Table 1 (continued ) 2024 WT11

Protein name Access. no. MW pI mRNA 2-DE H:L s.d. H:L s.d. Functional category Mode of secretion Post-translational requirement

Glioma pathogenesis-re- gi:5803151 30.34 8.80 1.32 2.50 0.00 2.50 0.00 Protease inhibitor Classical Glycosylation lated protein Palmitoyl-hydrolase gi:2135879 34.18 6.07 1.10 2.75 0.00 2.50 0.00 Protein synthesis Classical Glycosylation precursor Mono-ADP-ribosyl- gi:47087626 5.56 4.34 1.07 1.49 0.04 1.43 0.00 Protein synthesis Unknown non- Unknown transferase classical pathway Mitogen-activated pro- gi:66932916 41.39 6.50 1.18 4.02 0.67 2.96 0.28 Signaling Unknown non- Unknown tein kinase classical pathway a2-HS glycoprotein/Fe- gi:7106502 39.32 5.43 1.47 0.10 1.55 0.05 Signaling Classical Glycosylation tuin-A Cytosolic thyroid hor- gi:338827 58.00 7.95 1.22 2.50 0.00 Signaling Unknown non- Phosphorylation mone-bp classical pathway NF1 protein isoform gi:219940 62.30 8.04 1.09 2.82 0.00 2.58 0.00 Signaling Unknown non- Glycosylation/ classical pathway phosphorylation Importin-7 gi:5453998 119.52 4.70 1.09 1.42 0.05 1.37 0.26 Signaling Unknown non- Unknown classical pathway Osteopontin* gi:189405 33.84 4.40 Structure and support Vesicle-mediated; Glycosylation/ cells Khwaja glioma FW in proteins secreted wt-p53-regulated exocytosis phosphorylation b-5-Tubulin gi:21104420 49.67 4.78 1.05 C 1.55 0.25 1.59 0.09 Structure and motility Classical Phosphorylation a-Catenin gi:433411 100.07 5.95 1.08 C 4.02 0.00 3.55 0.31 Structure and motility Vesicle-mediated; Glycosylation/ exocytosis phosphorylation al et BDNF gi:987872 27.76 8.77 1.01 C 2.02 0.16 1.57 0.00 Survival Classical Glycosylation HP1Hs-g gi:1773227 19.72 5.03 1.77 0.24 1.70 0.08 Unknown Unknown non- Unknown classical pathway KIAA0828 gi:24308043 66.72 7.13 1.12 1.41 0.00 1.54 0.00 Unknown Unknown non- Glycosylaton classical pathway Unnamed protein pro- gi:31873592 35.25 6.51 1.89 0.22 1.65 0.01 Unknown Undetermined Unknown duct

Proteins with enhanced levels in the CM in response to wt-p53 are represented with their GenBank accession number (Access. no.), molecular weight (MW; kilodaltons), isoelectric point (pI), functional category, mode of secretion and post-translational requirements for secretion. Changes in mRNA levels of the corresponding proteins under p53-induction are indicated (mRNA). Listed ratios are an average of three separate experiments using 2024 cells, after 48 h induction of p53 in serum conditions. Ratios of >1.5 or o0.5 were considered signiﬁcantly differentially expressed. As positive controls, wt-p53 and p21 mRNA levels were on average increased 2.81- and 5.02-fold respectively in these microarray studies. The cICAT (H:L) ratio refers to the relative secretion of proteins under wt-p53 (labeled with heavy (H) reagent) compared to p53-null (labeled with light (L) reagent) conditions, that is, proteins upregulated by wt-p53 have a ratio >1 and those downregulated a ratio o1. The ratios given are an average of three separate experiments for each cell line (2024 and WT11). Only proteins with ProtScore >1.0 (>85% conﬁdence) were considered. The average ratios are followed by the standard deviation (s.d.). The 2-DE column indicates whether the proteins were also found by 2-DE analyses (combined results from 2024 and WT11 cells) and if their secretion levels were found concordant (C) with cICAT. Proteins in bold and with an * are previously known targets of p53. Proteins in italics (13 total) were found upregulated by both analyses (see Figure 3b). Empty space indicates that proteins were not found in that particular experiment or that the gene was not represented on the microarray chip. Oncogene 7653 Oncogene 7654

Table 2 Secreted proteins showing reduced accumulation in the CM upon p53-induction tp3rgltdscee rtisi loacells glioma in proteins secreted wt-p53-regulated 2024 WT11

Protein name Access. no. MW pI mRNA 2-DE H:L s. d. H:L s.d. Functional category Mode of secretion Post-translational requirement

ADAM-10* gi:4557251 84.14 8.04 Adhesion Classical Glycosylation

PEDF* gi:189778 46.33 5.84 C 0.38 0.00 0.56 0.00 Angiogenic Classical Glycosylation CYR61 protein gi:12584866 41.99 8.64 1.13 C 0.69 0.10 0.65 0.00 Angiogenic Classical Unknown VEGF* gi:181971 22.31 7.88 Angiogenic Classical Glycosylation Transforming growth gi:339558 12.79 8.59 1.18 C 0.75 0.00 Cell proliferation/ Classical Glycosylation factor b* angiogenic FGF-4 gi:4503701 22.05 9.73 Cell growth/proliferation/ Classical Glycosylation Khwaja FW angiogenic RTVP-1* gi:27735198 30.37 8.80 Cell proliferation/ Classical Glycosylation/phosphorylation differntiation TPM4-ALK fusion gi:10441386 27.53 4.77 Cell proliferation Non-classical; receptor Glycosylation al et oncoprotein mediated Granulin gi:183613 63.57 6.50 1.18 0.86 0.21 0.80 0.07 Cell proliferation Unknown non-classical Glycosylation method Interleukin 8 gi:33959 10.90 9.10 1.16 C 0.63 0.02 0.62 0.00 Immunity and defense/ Classical Unknwon angiogenic Attractin gi:3676347 175.00 4.71 1.16 C 0.39 0.16 0.27 0.09 Immunity and defense Unknown non-classical Unknown pathway ANP32A protein gi:76825059 14.79 5.27 1.21 0.44 0.33 0.78 0.01 Immunity and defense Undetermined Unknown Aldolase A gi:34577112 39.42 8.30 1.37 0.18 0.18 0.75 0.05 Metabolism Undetermined Unknown MMP-2* gi:11342666 73.88 5.26 Protease Classical Glycosylation/phosphorylation Helicase-MOI gi:5019620 218.81 5.47 0.23 0.00 0.25 0.00 RNA modiﬁcation Classical Glycosylation/phosphorylation SPARC/Osteonectin* gi:4507171 34.63 4.73 1.08 C 0.79 0.03 0.78 0.06 Structure and motility Classical Glycosylation A Chain A, Bm-40 gi:2624793 27.01 5.53 1.12 C 0.76 0.00 0.73 0.00 Structure and motility Classical None Insulin-like growth gi:183894 25.19 8.15 1.05 C 0.78 0.05 0.74 0.08 Survival Unknown non-classical Unknown factor bp6 method Metallothionein II D gi:223529 6.04 8.23 Unknown Classical Glycosylation Transgelin 2 gi:4507357 22.39 8.41 Unknown Unknown non-classical None method

Legend as in Table 1. wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7655 Table 3 Secreted proteins unchanged by wt-p53 or unclear as found by 2-DE and cICAT analyses 2024 WT11

Protein name Access. no. MW pI mRNA 2-DE H:L s.d. H:L s.d. Functional category

Thrombospondin-1* gi:40317626 129.38 4.71 1.13 C 1.05 0.04 1.09 0.10 Anti-angiogenic Endothelin 1 gi:556203 24.43 9.52 1.06 C 1.01 0.00 0.90 0.11 Angiogenic MAC25 gi:307151 28.75 8.40 1.15 1.03 0.01 1.02 0.00 Cell proliferation regulation GDNF family receptor a 1 isoform gi:20141405 51.46 8.30 1.05 1.63 0.00 Cell differentiation Neurotrophin gi:4505469 29.35 9.34 Differentiation and survival Stage-specific S antigen homolog gi:51466832 68.80 11.82 C 1.09 0.06 1.12 0.03 DNA-binding protein Hyp. Zinc-finger protein KIAA0296 gi:40788207 201.84 7.05 1.18 C 1.08 0.02 0.91 0.03 DNA-binding protein Ah receptor-interacting protein gi:6226814 37.66 6.04 1.19 U 1.46 0.34 1.43 0.07 DNA binding protein Fibulin 1A gi:19743813 138.97 5.41 ECM component; cell signaling Prion protein gi:220016 26.82 9.04 1.23 0.14 0.20 0.96 0.00 ECM protein Procollagen C-endopeptidase enhancer gi:4505643 47.95 7.41 1.11 1.18 0.02 1.45 0.08 Enzyme Phosphodiesterase I a gi:662290 99.04 7.49 1.25 1.22 0.00 1.09 0.08 Enzyme Follistatin-like 1 gi:2498390 34.99 5.39 1.06 C 1.11 0.02 1.03 0.03 Immunity and defense Immunoglobulin heavy chain variable gi:553428 16.16 9.52 1.03 U 1.45 0.00 1.45 0.09 Immunity and defense MAC-2 binding protein precursor gi:41150551 52.47 8.91 C 0.98 0.16 1.02 0.06 Immunity and defense T cell receptor a chain gi:416065 11.39 6.10 1.04 C 0.86 0.00 1.10 0.15 Immunity and defense Immunoglobulin kappa chain variable gi:185974 12.50 5.14 1.07 C 0.94 0.00 0.94 0.05 Immunity and defense Glycosylase I gi:62821794 66.22 8.31 1.06 0.02 1.04 0.04 Metabolism T-cell receptor delta chain gi:540457 12.98 5.50 1.03 U 1.00 0.00 1.00 0.00 Immunity and defense Isomerase, triosephosphate gi:223374 26.63 7.09 1.04 C 1.00 0.05 1.09 0.10 Metabolic glycolytic enzyme HSP 70/71; isoform2 gi:24234686 53.52 5.62 1.11 C 1.02 0.10 1.04 0.00 Oxidative damage repair enzyme Rnase H gi:52000844 91.95 9.21 RNA degradation Cathepsin B preproprotein gi:22538437 37.82 5.88 1.21 C 0.98 0.05 1.00 0.06 Protease heterogeneous nuclear ribonucleoprotein gi:14165439 51.03 5.19 1.08 U 0.75 0.14 0.78 0.09 Protein synthesis RNA binding protein regulatory subunit gi:14198257 19.89 6.33 1.08 1.00 0.00 Protein synthesis Ribosomal protein S12 gi:14277700 14.51 6.81 1.18 C 1.37 0.00 Protein synthesis CAMKII gi:25952118 54.09 6.61 Kinase inhibitor Calmodulin related protein-A11 gi:47458820 83.13 6.70 U 1.01 0.06 1.12 0.03 Protein kinase Fibrillin-2 pre gi:66346695 314.77 4.73 1.06 C 0.83 0.05 0.93 0.07 Structure and motility 14-3-3 protein tau gi:112690 27.76 4.68 1.11 U 0.80 0.12 0.87 0.00 Structure and motility Cofilin, non-muscle isoform gi:5031635 18.52 8.22 1.07 C 0.93 0.02 0.97 0.05 Structure and motility Filamin gi:8100574 278.20 5.49 1.15 U 0.00 0.00 2.03 0.16 Structure and motility Vimentin gi:62414289 53.65 5.06 1.06 0.87 0.04 0.96 0.00 Structure and motility Myosin light chain 3 gi:4557777 21.93 5.03 1.06 C 1.06 0.08 1.13 0.01 Structure and motility Stanniocalcin 2 precursor gi:4507267 33.25 6.93 1.21 C 1.07 0.05 1.10 0.11 Signaling Tyrosine 3-monooxygenase gi:136574 58.52 5.90 1.16 1.04 0.13 1.12 0.04 Signaling Peptidylprolyl isomerase A gi:10863927 18.01 7.68 1.02 1.24 0.02 1.21 0.00 Signaling YWHAZ protein gi:68085909] 27.75 4.73 1.18 1.31 0.21 1.24 0.00 Signaling Insulin-like growth factor bp5 precursor gi:184820 30.57 8.58 1.25 U 1.92 0.10 1.30 0.09 Survival; immunity and defense Thioredoxin/ NKEF-B gi:50592994 11.74 4.82 1.08 U 1.64 0.37 1.63 0.00 Survival/immunity Amyloid A4 protein gi:28721 84.82 4.71 1.28 1.12 0.07 1.06 0.08 Transport Apolipoprotein-E gi:671882 11.84 6.57 Transport Human serum albumin gi:178344 69.37 5.92 Transport GDI-a gi:4757768 23.21 5.03 Vesicle-mediated transport CDw44 antigen precursor gi:180197 39.56 5.13 1.22 0.85 0.16 0.82 0.05 Adhesion Unnamed protein product; phosphatase gi:19701027 77.58 5.05 1.15 0.06 1.31 0.02 Unknown Similar to HSPC280 gi:6841210 15.80 7.91 0.90 0.09 0.97 0.00 Unknown Hypothetical FGF-like protein gi:4557679 133.10 5.71 U 1.15 0.04 1.33 0.17 Unknown KIAA0012 gi:40789057 90.55 6.64 U 1.31 0.21 1.17 0.00 Unknown MGC:71022 gi:38303909 10.83 4.69 1.23 0.00 0.99 0.00 Unknown Unnamed protein product; zinc-finger gi:21751981 82.65 6.78 1.69 0.08 0.98 0.00 Unknown

Legend as in Table 1. Proteins with secretion levels found concordant (C) or unclear (U) between the two methods are indicated. expression was performed by comparing spot intensity Secretome analysis by cICAT and volume using ImageMaster (Figure 2). The levels of Given that the total number of secreted proteins 34 proteins in the CM were found to be largely identified by 2-DE was smaller than we had anticipated, invariable regardless of p53 expression, whereas 32 and to avoid potential bias of using a single identifica- individual proteins showed differential expression levels tion method, we sought a second complementary in the CM in response to p53 (Tables 1–3). Among the approach. Recently, internally standardized gel-free differentially expressed proteins, 18 had increased levels quantitative proteomic methods have been developed and 16 decreased levels in the CM in response to wt-p53 to alleviate limitations of 2-DE. One of these methods is expression in the cells (Figure 3b). The 68 secreted isotope-coded affinity tag (ICAT) reagent labeling and proteins identified in the 2-DE screen belonged to 15 tandem mass spectrometry (MS/MS) (Gygi et al., 2002). functional categories (Figure 3c; Tables 1 and 2). Secreted proteins from 2024 and WT11 cells were very

Oncogene wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7656 - +Wt-p53 - +Wt-p53

TSP1 Gal-3

SPARC Gal-1

FGF-4 β-2M

TGF-β Pre-alb.

Figure 2 Semiquantitative analysis of differentially expressed proteins found by 2-DE and identiﬁed by MS/MS analysis. 3D representation of differential expression for representative proteins found upregulated (Gal-3, Gal-1 and b-2M), downregulated (SPARC, FGF-4 and TGF-b), and unchanged (TSP-1 and Pre-alb) by 2-DE as analysed by ImageMaster software.

similar in their expression patterns and differed sig- as U in Table 3) were either secreted at a very low level nificantly in expression levels for only 10 of the 91 or not differentially expressed to a high degree. When proteins identified by this analysis. Through cICAT looking at the concordance between p53-regulated alone, we found 34 proteins with increased levels, and 13 proteins, we found 13 proteins upregulated, eight with decreased levels by at least 20% while 44 remained downregulated and 16 unchanged by both 2-DE and unchanged in response to wt-p53 expression (Tables 1–3; cICAT analysis while the remaining 11 (listed as U) were Figure 3b). These proteins were found differentially found differentially expressed only by one of the two expressed in CM (Po0.05) in at least two of the three indicated methods (Figure 3b Tables 1–3). experiments for both cell lines. The quantification from cICAT was found to be consistent between experiments as seen by small standard deviation values for each Verification of proteomic results tested cell line (Tables 1–3). Similar to 2-DE results, the Some of the p53-regulated secreted proteins found in 91 proteins found secreted in the media by cICAT our analysis had been previously reported and validated, experiments belonged to 15 functional groups including vascular endothelial growth factor (VEGF), (Figure 3c; Tables 1 and 2). osteopontin (Opn) (Morimoto et al., 2002), secreted protein with acidic and cysteine rich domains (SPARC) and dickkopf (Wang et al., 2000). To confirm our results Comparison of 2-DE and cICAT results we picked three proteins whose levels were increased Combining both techniques, we were able to identify (galectin-1 (Gal-1), galectin-3 (Gal-3) and beta 2 111 separate secreted proteins; 68 by 2-DE and 91 by microglobulin (b-2M)) and three decreased (SPARC, cICAT analysis (Figure 3a). It is noteworthy that 48 of fibroblast growth factor-4 (FGF-4) and transforming the 91 (B50%) secreted proteins identified by cICAT growth factor beta (TGF-b)) in the CM in response to analysis, were identical to the ones identified by 2-DE wt-p53 expression for validation by Western analysis. analysis, showing concordant results between the The levels of Gal-1, Gal-3 and b-2M in the CM were techniques in identifying the complement of secreted clearly increased by p53 in 2024 cells (Figure 4, compare proteins (Tables 1–3; Figure 3). Thirty-seven of the 48 lanes 2 and 4). In contrast, secreted levels of SPARC, (77%) proteins commonly identified by the techniques TGF-b and FGF-4 were decreased. The downregulation showed similar responses to p53 activation in the cells. of SPARC and TGF-b levels in the CM by p53 was The majority of the remaining 11 (24%) proteins (listed particularly strong as it was able to antagonize their

Oncogene wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7657 ab 50 cICAT (91) 44 Up-regulated 2-DE (68) 45 Down-regulated 40 Unchanged/ Unclear 35 34 34 30 25 18 11 2048 43 20 16 13 15 13 10 8 16 5

# of proteins identified # of proteins 0 2-DE cICAT 2-DE+cICAT

c 14 12 2-DE 10 cICAT 8 6 4 2

# of proteins identified # of proteins 0

Survival Adhesion Protease Signaling TransportUnknown Metabolism Proliferation AngiogenesisDifferentiationDNA binding Protein synthesis Anti-angiogenesis Immune response Protease inhibitor Structure & motility Figure 3 Comparative analysis of wt-p53-regulated secreted proteins using both proteomic analyses: (a) Ven Diagram showing the total number (111) of secreted proteins found by 2-DE (white), and cICAT (gray). (b) Number of proteins found unchanged (white), up- (dark gray) or downregulated (light gray) by wt-p53 expression using 2-DE and c-ICAT analysis alone and those common to both techniques. (c) Distribution of identified secreted proteins by 2-DE (white) and c-ICAT (gray) analyses according to their general functional categories. Each protein is seen in a single category only even though some might play multiple functions. increase by doxycycline (dox) as seen in the LNZ308- 2003). Preliminary indications of such post-trans- C16 control cells that lack p53. Thrombospondin-1 criptional modifications were noted for a subset of (TSP-1) was used as a loading control since its levels are the identified proteins through 2-DE analysis (Table 4), not found to be increased by wt-p53 in our glioma as seen for example by the horizontal and vertical system (Tenan et al., 2000). The data show that our shifts of Gal-1 protein spots from their original pI proteomic analysis with 2-DE and cICAT can be used to and MW positions (Figure 1b, c; black arrows). This reliably identify differential expression of secreted suggests a potential novel function of the p53 tumor proteins in the CM (Tables 1–3). suppressor, the modulation of post-transcriptional modifications. Alternatively, p53 may also be involved in the regulation of a specific secretory pathway (Yu Investigation of the mechanism underlying p53 control et al., 2006). Indeed, most proteins whose levels were over protein secretion positively regulated by p53 were found secreted through To examine whether the CM levels of the secreted nonclassical mechanisms including vesicle-mediated proteins identified were regulated by p53 at the gene pathways like exocytosis, ectocytosis as well as through expression level, we examined the differential expression transporter-mediated pathways. In contrast, most pro- of their mRNAs by microarrays in the 2024 cell line in teins released through classical pathways were down- three independent experiments. None of the mRNAs regulated (Tables 1 and 2). corresponding to the secreted products found in our analysis appeared to have levels significantly modulated by p53 (Tables 1–3; column 5). These findings suggest a role of wt-p53 in the modulation of Discussion the extracellular levels of secreted proteins through either enhanced stability or secretion. One way that This is the first comprehensive analysis of the tumor cell p53 could potentially affect protein stability and/or secretome to identify secreted targets of wt-p53. The secretion is through regulation of post-translational term ‘secretome’ refers to proteins released through modifications, for example, phosphorylation, glycosyla- classical as well as nonclassical secretion pathways tion, acetylation and hydroxylation of proteins, (Volmer et al., 2005). In addition, it also includes events that may mark certain proteins either for intracellular proteins and protein fragments that might degradation or for localization (Kamemura and Hart, be released in exosomes as a result of wt-p53 expression.

Oncogene wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7658 In our analysis, p53 expression led to increased levels of any of the secreted proteins found in our analysis to 39 and decreased levels of 21 proteins in the CM of be signiﬁcantly regulated by p53 at the transcriptional glioma cells (Tables 1–3). level. Instead, our microarray and Western analyses The mechanism through which p53 might regulate the showed that most p53-regulated secreted proteins secretion of proteins is currently unknown. A number of were not direct p53 targets and may have accumulated secreted proteins regulated at the transcriptional in the CM indirectly through different mechanisms. level have been reported. However, we did not ﬁnd One possibility is enhanced stability, which could result through multiple means including changes in protein stability and localization or downregulation of proteases like matrix metalloproteinase (MMPs) thus leading to the accumulation of the affected proteins in

Western C16 2024 the media. In fact, MMP-1 and MMP-13 have already -+ -+Dox been shown to be downregulated by wt-p53 (Sun et al., -- -+p53 2000). Alternatively, p53 could alter the secretion rate of intracellular proteins through either augmented release of speciﬁc proteins or through upregulation Gal-1 of a particular secretory pathway, thus leading to enhanced levels of all proteins secreted through that pathway. There is a precedence in the literature for at Gal-3 least one p53-regulated protein (TSAP6) that can facilitate the secretion of another protein (TCTP) β-2M through ectocytosis (Amzallag et al., 2004). Recent evidence suggests that p53 may act as a general regulator of this nonclassical secretory pathway (Yu et al., Pre-alb. 2006).

Functional implications for tumorigenesis SPARC Wild-type p53 has been shown to inhibit many processes required for tumor growth including migration, angiogenesis, survival and cell proliferation (Fulci and Van FGF-4 Meir, 1999). It has also been implicated in eliciting an immune response against neoplastic cells (Bueter et al., TGF-β 2006). The results of our screen found wt-p53 regulating the secretion of many proteins that are candidate mediators for the above biological effects. TSP1 p53 and metastasis/invasion Figure 4 Veriﬁcation of selected 2-DE and cICAT results: Our analysis found several ECM components (growth Western blot analysis on TCA-precipitated serum-free CM arrest-speciﬁc 6, collagen type XI a-1, proteoglycan collected after 48 h from LNZ308-C16 (control for dox) and 2024 PG-M) or proteins involved in adhesion and cell–matrix cells with tet-on wt-p53 expression. Differential expression of SPARC, FGF-4, TGF-b, Gal-1, Gal-3, and b-2M in response to interactions (galectin-3, lysyl oxidase-like protein 2, wt-p53 expression was examined. TSP1 and Pre-alb were loading Opn, a-catenin and b-5 tubulin) as well as protease controls and remained unchanged. inhibitors (TIMP-3 and glioma pathogenesis-related

Table 4 Secreted proteins with potential post-translational modiﬁcations induced by wt-p53 Protein name Access. no. Th. MW Th. pI Obs. MW Obs. pI Potential post-translational modiﬁcation

Galectin-3 gi:4504983 26.19 8.58 25–35 7.5–8.7 Phosphorylation and/or glycosylation Galectin-1 gi:12804557 14.72 5.33 17–33 5.0–5.5 Phosphorylation and/or glycosylation b-2 microglobulin gi:179318 13.73 6.06 15 5.0–5.5 Phosphorylation Proteoglycan PG-M (V3) gi:1008913 74.25 7.43 1.34 7.5–7.75 Dephosphorylation and/or deglycosylation Cytosolic thyroid hormone-bp gi:338827 58.00 7.95 57–62 6.4–6.5 Dephosphorylation and/or deglycosylation ADAM-10 gi:4557251 84.14 8.04 70 8.7–9.0 Dephosphorylation or proteolysis MMP-2 gi:11342666 73.88 5.26 70–75 4.1 Not determined Thrombospondin-1 gi:40317626 129.38 4.71 125 4.5–4.7 Phosphorylation 14-3-3 protein tau gi:112690 27.76 4.68 13 4.0–5.0 Dephosphorylation and/or proteolysis

Proteins whose spot pattern on 2-DE indicated possible wt-p53 induced post-translational modiﬁcations are listed along with their accession number (Acess. no.), theoretical molecular weight (Th. MW), theoretical isoelectric point (Th. pI), observed molecular weight (Obs. MW), observed isoelectric point (Obs. pI) and the possible post-translational modiﬁcation observed. A vertical shift (changes in MW) indicates possible glycosylations while a horizontal shift towards lower pI indicates possible phosphorylations.

Oncogene wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7659 protein) upregulated in the CM from the glioma cells p53 and tumor proliferation and survival after wt-p53 induction. The induction of these structural We found several proteins regulating tumor prolifera- and proadhesion proteins would be expected to improve tion and survival to be regulated by wt-p53. Brain cell–cell and cell–matrix interactions, thus resulting in derived neurotrophic factor (BDNF) exhibited en- reduced migratory potential of tumor cells. hanced secretion in response to wt-p53. The secreted In addition to upregulation of antimigratory factors, form of BDNF mediates apoptosis of cells containing other proteins directly involved in induction of migra- neurotrophin receptors (Lee et al., 2001). Other pro- tion and invasion in multiple tumor types, were found teins, including FGF-4, RTVP1, TPM-ALK fusion downregulated by wt-p53. These included SPARC, oncoprotein fragment, TGF-b, PEDF, IGFBPs, and MMP-2, TGF-b, ADAM-10, and Tau (Framson granulin were all found to have inhibited secretion to and Sage, 2004; Mazzocca et al., 2005; Stuelten et al., varying degrees in response to wt-p53. RTVP1, TGF-b 2005). These findings point to a potential new facet of (Tsuzuki et al., 1998), and PEDF (Pietras et al., 2002) p53’s multimodal function as a tumor-suppressor are previously known targets of wt-p53. gene, the downregulation of tumor invasion and metastasis. Concluding remarks In this study we have identified secreted proteins whose p53 and the immune response extracellular levels are regulated by p53. We found 39 In recent years various studies have suggested that proteins with enhanced and 21 with inhibited levels in wt-p53 could stimulate immune responses against tumor response to wt-p53 expression out of a total of 111 cells. For example, secreted Opn found upregulated in proteins identified to be secreted by the cells. None of our analysis is one of the key cytokines for type 1 the tested proteins were found to be transcriptional immune responses mediated by macrophages. It has targets indicating that wt-p53 may have an indirect role already been reported as a direct target of wt-p53 and in their stability or secretion. These secreted targets will has been implicated in suppressing tumor growth in vivo be helpful in better understanding of how wt-p53 may (Morimoto et al., 2002). Our screen identified increased modulate interactions of tumor cells with their environ- secretion of immune response-related proteins b-2M and ment and establishes p53 loss in tumors as an originator macrophage migration and myeloid leukemia inhibitory of changes in tumor–stroma interactions. They may also factors in response to wt-p53. b-2M is a MHC class I help explain some of the ‘bystander effects’ observed in molecule and several studies have shown that tumor p53-mediated cancer gene therapy or with radio- and development might be inhibited by immune responses chemotherapies that activate p53. stimulated by this class of proteins (Bueter et al., 2006). Other immune-related proteins like interleukin-8 (IL-8), attractin, and ANP32A were found downregulated by Materials and methods wt-p53. IL-8 is known to be upregulated in glioma, possi- bly in response to immune cell infiltration (Desbaillets Cell lines and culturing conditions et al., 1997). Attractin is upregulated in glioma patient LN-Z308 (p53 null) human glioblastoma cell line (Albertoni cerebrospinal fluid (CSF) and can modulate T-cell et al., 1998), and its isogenic clones LNZ308-C16 (contains a reverse tetracycline transactivator (rtTA)), 2024 (tet-inducible motility (Khwaja et al., 2006). These results encourage wt-p53) (Albertoni et al., 2002) and WT11 (tet-off for wt-p53) further research into how p53 may modulate the tumor (Van Meir et al., 1994) were grown in Dulbecco’s modified immune response. Eagle’s medium supplemented with 5% tet-tested fetal calf serum. Cells were grown in serum-free media and wt-p53 p53 and angiogenesis expression was induced by modulation with 2 mg/ml of dox. CM from the cells was collected after 48 h induction and Our results show repression of at least five proangio- frozen at À201C after removal of floating cells and cell debris genic proteins, VEGF, IL-8, TGF-b, PEDF, and by centrifugation at 1000 g. CYR61 by wt-p53 in glioma cells. VEGF has been shown to be downregulated by wt-p53 in many systems Two-dimensional polyacrylamide gel electrophoresis (2-DE) (Qin et al., 2006) while CYR61 has not been reported Samples were analysed in triplicates using 2-DE as described as a p53 target before. CYR61 is a secreted ECM- (Goldman et al., 1980). The first dimension was performed on associated signaling molecule that has been shown to IPGphor system (Amersham Biosciences, NJ, USA). Iso- promote the adhesion and proliferation of endothelial electric focusing of 200 mg of trichloroacetic acid (TCA) cells (Babic et al., 1998). CYR61 has been shown to be precipitated protein was performed on 13 cm or 17 cm overexpressed in several cancers including breast and Immobiline dry strips (IPG strips) using either pH range of brain tumors, where it promotes angiogenesis and 3–10NL or 4–7L (total run ¼ 130 000 Vh). Strips were then increased tumor growth (Tsai et al., 2000; Xie et al., equilibrated sequentially in equilibration buffer (6 M urea, 2% 2004). Similarly, IL-8 is expressed and secreted at high sodium dodecyl sulfate (SDS), 0.05 M Tris base pH 8.8, 20% glycerol) first containing 10 mg/ml dithiothreitol (DTT) and levels in human gliomas and involved in glial tumor neo- then 25 mg/ml iodoacetamide followed by separation in the vascularity and progression (Brat et al., 2005). Overall second dimension on 12.5% polyacrylamide gels with 2% SDS our findings suggest a model in which p53 loss in tumors using the Protean II xi system (BioRad, CA, USA). Silver activates angiogenesis by an increase in secretion of Stain Plus kit (BioRad) was used to visualize protein spots and proangiogenic factors and decrease of inhibitors. the gels were analysed using Melanie (http://au.expasy.org/

Oncogene wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7660 melanie/) and the ImageMaster softwares (Amersham Bio- 15% TCA for 2 h at 41C, washed twice with ice-cold acetone, sciences, NJ, USA). and then resuspended in lysis buffer (8 M urea, 4% SDS, 100 mM protease inhibitor cocktail (Roche, Germany), in In-gel digestion of proteins and MALDI-TOF/TOF-MS analysis 10 mM Tris, pH 7.4). Antibodies used were: a-TSP1 (Ab-4 Protein spots of interest were excised from the gel and destained NeoMarkers, Freemont, CA, USA; 1:1000), a-FGF-4 (sc- using SilverOUT kit (GenoTech, MO, USA). The proteins were 16812, Santa Cruz, CA, USA; 1:500), a-SPARC (sc-13324, digested overnight with 150 ng trypsin (Promega, WI, USA) and Santa Cruz; 1:500), a-VEGF (Santa Cruz; 1:500), a-b-2M the resulting peptides extracted using Montage In-gel peptide (Clone B2M-01; Abcam, MA, USA; 1:250), a-TGFb extraction kit (Millipore, MA, USA), spotted onto target plates (AE1109.1, Immunodiagnostik, Germany; 1:100), a-galectin-3 and overlaid with a-cyano-4-hydroxycinnamic acid (Agilent, (Santa Cruz, CA, USA; sc-14364; 1:500), a-galectin-1 (Santa DE, USA). The plates were analysed using a 4700 Proteomics Cruz, CA, USA; 1: 500). Pre-albumin (Pre-alb) (sc-13098; Analyzer (Applied Biosystems, CA, USA). The combined MS Santa Cruz; 1:1000) and actin (sc-1615; Santa Cruz, CA, USA; and MS/MS spectra from each spot were processed using GPS 1:1000) were used as a loading controls. Explorer V2.0 (Applied Biosystems, CA, USA) with MASCOT (Matrix Science, MA, USA) as the database search engine. Only proteins that generated multiple peptides with ion scores above 30 were considered positively identiﬁed. Abbreviations

cICAT analysis b-2M, beta-2-microglobulin; 2-DE, two-dimensional gel cICAT technology uses stable isotope tags in combination with electrophoresis; cICAT, cleavable isotope-coded affinity tag two-dimensional (2D) chromatography of complex peptide technology; CM, conditioned media; ECM, extracellular mixtures (Applied Biosystems) (Gygi et al., 2002). Hundred matrix; FGF-4, fibroblast growth factor-4; Gal-1, Galectin-1; micrograms each of precipitated secreted protein from the CM Gal-3, galectin-3; Pre-alb, pre-albumin; SPARC, secreted were treated with denaturing (50 mM Tris; 0.1% SDS) and protein with acidic and cysteine-rich domains; TGF-b, reducing (50 mM TCEP (Tris(2-carboxyethyl)phosphine hy- transforming growth factor beta; TSP1, thrombospondin-1. drochloride)) reagents. Next, the control and wt-p53 induced samples were respectively labeled with light (9 12C atoms) and heavy (9 13C atoms) reagents for 2 h at 371C. After trypsin Acknowledgements digestion and purification, the peptides were analysed using an Ultimate nanoHPLC LC-MS/MS (Dionex/LC Packings, CA, We thank Drs JC Lucchesi, D Pallas, I Matsumura and USA) interfaced to a QSTAR XL mass spectrometer (Applied P Vertino for their support. This work was supported by Biosystems). The MS/MS data were processed using ProICAT National Institutes of Health (NIH) grants CA 86335 (to software for protein identification and quantification. Only EGVM); NCRR 02878, 12878, 13948 (to Microchemical and proteins with ProtScore >1.0 (>85% confidence) were con- Proteomics Facility), the Pediatric Brain Tumor Foundation sidered. Also, the heavy to light ratios were tested for signifi- of the US (to EGVM) and the American Brain Tumor Asso- cance using Student t-test and Po0.05 was considered significant. ciation (to BP), the Genetics and Molecular Biology (GMB) program of the Graduate Division of Biological and Western blot analysis Biomedical Sciences (GDBBS) of Emory University, Immunoblots were performed on cell lysates (lysed in 8 M urea, and the National Science Foundation (NSF) (PRISM; 4% SDS, in 10 mM Tris, pH 7.4). The CM was precipitated by DGE0231900).

References

Albertoni MD, Daub M, Arden KC, Viars CS, Powell C, Bueter M, Gasser M, Lebedeva T, Benichou G, Waaga- Van Meir EG. (1998). Genetic instability leads to loss Gasser AM. (2006). Inﬂuence of p53 on anti-tumor of both p53 alleles in a human glioblastoma. Oncogene 16: immunity (review). Int J Oncol 28: 519–525. 321–326. Chiarugi V, Magnelli L, Gallo O. (1998). Cox-2, iNOS and p53 Albertoni M, Shaw PH, Nozaki M, Godard S, Tenan M, as play-makers of tumor angiogenesis (review). Int J Mol Hamou M-F et al. (2002). Anoxia induces macrophage Med 2: 715–719. inhibitory cytokine-1 (MIC-1) in glioblastoma cells inde- Desbaillets I, Diserens AC, Tribolet N, Hamou MF, pendently of p53 and HIF-1. Oncogene 21: 4212–4219. Van Meir EG. (1997). Upregulation of interleukin 8 by Amzallag N, Passer BJ, Allanic D, Segura E, Thery C, Goud B oxygen-deprived cells in glioblastoma suggests a role in et al. (2004). TSAP6 facilitates the secretion of translation- leukocyte activation, chemotaxis, and angiogenesis. J Exp ally controlled tumor protein/histamine-releasing factor via Med 186: 1201–1212. a nonclassical pathway. J Biol Chem 279: 46104–46112. Framson PE, Sage EH. (2004). SPARC and tumor growth: Babic AM, Kireeva ML, Kolesnikova TV, Lau LF. (1998). where the seed meets the soil? J Cell Biochem 92: 679–690. CYR61, a product of a growth factor-inducible immediate Fulci G, Van Meir EG. (1999). p53 and the CNS: tumors and early gene, promotes angiogenesis and tumor growth. Proc developmental abnormalities. Mol Neurobiol 19: 61–77. Natl Acad Sci USA 95: 6355–6360. Goldman D, Merril CR, Ebert MH. (1980). Two-dimensional Bhowmick NA, Moses HL. (2005). Tumor–stroma interac- gel electrophoresis of cerebrospinal ﬂuid proteins. Clin Chem tions. Curr Opin Genet Dev 15: 97–101. 26: 1317–1322. Brat DJ, Bellail AC, Van Meir EG. (2005). The role of Go¨ rg A, Weiss W, Dunn MJ. (2004). Current two-dimensional interleukin-8 and its receptors in gliomagenesis and tumoral electrophoresis technology for proteomics. Proteomics 4: angiogenesis. Neuro-oncol 7: 122–133. 3665–3685.

Oncogene wt-p53-regulated secreted proteins in glioma cells FW Khwaja et al 7661 Gygi SP, Rist B, Griffin TJ, Eng J, Aebersold R. (2002). differentially regulate the gene expression of human col- Proteome analysis of low-abundance proteins using multi- lagenase-3 (hMMP-13). J Biol Chem 275: 11327–11332. dimensional chromatography and isotope-coded affinity Taieb J, Chaput N, Menard C, Apetoh L, Ullrich E, Bonmort tags. J Proteome Res 1: 47–54. M et al. (2006). A novel dendritic cell subset involved in Hanahan D, Weinberg RA. (2000). The hallmarks of cancer. tumor immunosurveillance. Nat Med 12: 214–219. Cell 100: 57–70. Tenan M, Fulci G, Albertoni M, Diserens AC, Hamou MF, Harris SL, Levine AJ. (2005). The p53 pathway: positive and El Atifi-Borel M et al. (2000). Thrombospondin-1 is negative feedback loops. Oncogene 24: 2899–2908. downregulated by anoxia and suppresses tumorigenicity of Kamemura K, Hart GW. (2003). Dynamic interplay between human glioblastoma cells. J Exp Med 191: 1789–1798. O-glycosylation and O-phosphorylation of nucleocytoplas- Tsai MS, Hornby AE, Lakins J, Lupu R. (2000). Expression mic proteins: a new paradigm for metabolic control of signal and function of CYR61, an angiogenic factor, in breast transduction and transcription. Prog Nucleic Acid Res Mol cancer cell lines and tumor biopsies. Cancer Res 60: Biol 73: 107–136. 5603–5607. Khwaja FW, Duke-Cohan JS, Brat DJ, Van Meir EG. (2006). Tsuzuki T, Izumoto S, Ohnishi T, Hiraga S, Arita N, Hayakawa T. Attractin is elevated in the cerebrospinal fluid (CSF) of (1998). Neural cell adhesion molecule L1 in gliomas: patients with malignant astrocytoma and mediates glioma correlation with TGF-beta and p53. J Clin Pathol 51: cell migration. Clinical Cancer Research, in press. 13–17. Lee R, Kermani P, Teng KK, Hempstead BL. (2001). Van Meir EG, Polverini PJ, Chazin VR, Su Huang HJ, de Regulation of cell survival by secreted proneurotrophins. Tribolet N, Cavenee WK. (1994). Release of an inhibitor of Science 294: 1945–1948. angiogenesis upon induction of wild type p53 expression in Mazzocca A, Coppari R, De Franco R, Cho JY, Libermann glioblastoma cells. Nat Genet 8: 171–176. TA, Pinzani M et al. (2005). A secreted form of ADAM9 Van Meir EG, Roemer K, Diserens A-C, Kikuchi T, Rempel promotes carcinoma invasion through tumor–stromal inter- SA, Haas M et al. (1995). Single-cell monitoring of growth actions. Cancer Res 65: 4728–4738. arrest and morphological changes induced by transfer of Morimoto I, Sasaki Y, Ishida S, Imai K, Tokino T. (2002). wild type p53 alleles to glioblastoma cells. Proc Natl Acad Identification of the osteopontin gene as a direct target of Sci USA 92: 1008–1012. TP53. Genes Chromosomes Cancer 33: 270–278. Volmer MW, Stuhler K, Zapatka M, Schoneck A, Klein- Pietras K, Rubin K, Sjoblom T, Buchdunger E, Sjoquist M, Scory S, Schmiegel W et al. (2005). Differential proteome Heldin CH et al. (2002). Inhibition of PDGF receptor analysis of conditioned media to detect Smad4 regu- signaling in tumor stroma enhances antitumor effect of lated secreted biomarkers in colon cancer. Proteomics 5: chemotherapy. Cancer Res 62: 5476–5484. 2587–2601. Qin G, Kishore R, Dolan CM, Silver M, Wecker A, Lue- Wang J, Shou J, Chen X. (2000). Dickkopf-1, an inhibitor of demann CN et al. (2006). Cell cycle regulator E2F1 modu- the Wnt signaling pathway, is induced by p53. Oncogene 19: lates angiogenesis via p53-dependent transcriptional control 1843–1848. of VEGF. Proc Natl Acad Sci USA 103: 11015–11020. Xie D, Yin D, Wang HJ, Liu GT, Elashoff R, Black K et al. Steele RJ, Lane DP. (2005). P53 in cancer: a paradigm for (2004). Levels of expression of CYR61 and CTGF are modern management of cancer. Surgeon 3: 197–205. prognostic for tumor progression and survival of individuals Stuelten CH, DaCosta Byfield S, Arany PR, Karpova TS, with gliomas. Clin Cancer Res 10: 2072–2081. Stetler-Stevenson WG, Roberts AB. (2005). Breast cancer Yu X, Harris SL, Levine AJ. (2006). The regulation of cells induce stromal fibroblasts to express MMP-9 via exosome secretion: a novel function of the p53 protein. secretion of TNF-alpha and TGF-beta. J Cell Sci Cancer Res 66: 4795–4801. 118(Part 10): 2143–2153. Zigrino P, Loffek S, Mauch C. (2005). Tumor-stroma Sun Y, Cheung JM, Martel-Pelletier J, Pelletier JP, Wenger L, interactions: their role in the control of tumor cell invasion. Altman RD et al. (2000). Wild type and mutant p53 Biochimie 87: 321–328.

Oncogene