Proteomics-Based Strategy To Delineate the Molecular Mechanisms

May 25, 2007 - Jose´ Rivera,*,† Diego Megias,‡ and Jero´nimo Bravo*,†. Signal Transduction Group, Structural Biology and Biocomputing Programm...
1 downloads 0 Views 654KB Size
Proteomics-Based Strategy To Delineate the Molecular Mechanisms of the Metastasis Suppressor Gene BRMS1 Jose´ Rivera,*,† Diego Megias,‡ and Jero´ nimo Bravo*,† Signal Transduction Group, Structural Biology and Biocomputing Programme, and Confocal Microscopy and Cytometry Unit, Biotechnology Programme, Centro Nacional de Investigaciones Oncolo´gicas (CNIO), Melchor Ferna´ndez Almagro 3, E-28029 Madrid, Spain Received May 25, 2007

The breast cancer metastasis suppressor 1 (BRMS1) gene has been shown to suppress metastasis without affecting the growth of the primary tumor in mouse models. It has also been shown to suppress the metastasis of tumors derived from breast, melanoma, and, more recently, ovarian carcinoma (see ref 1). However, how BRMS1 exerts its metastasis suppressor function remains unknown. To shed light into its metastatic mechanism of action, the sensitive 2D-DIGE analysis coupled with MS has been used to identify proteins differentially expressed by either overexpressing (Mel-BRMS1) or silencing BRMS1 (sh635) in a melanoma cell line. After comparison of the protein profiles from WT, Mel-BRMS1, and sh635 cells, 79 spots were found to be differentially expressed. Mass spectrometry analysis allowed the unambiguous identification of 55 polypeptides, corresponding to 43 different proteins. Interestingly, more than 75% of the identified proteins were down-regulated in Mel-BRMS1 cells compared to WT. In contrast, all the identified proteins in sh635 cells extracts were up-regulated compared to WT. Most of the deregulated proteins are involved in cell growth/maintenance and signal transduction among other cell processes. Six differentially expressed proteins (Hsp27, Alpha1 protease inhibitor, Cofilin1, Cathepsin D, Bone morphogenetic protein receptor2, and Annexin2) were confirmed by immunoblot and functional assays. Excellent correlation was found between DIGE analysis and immunoblot results, indicating the reliability of the analysis. Available evidence on the reported functions of the identified proteins supports the emerging role of BRMS1 as negative regulator of the metastasis development. This work opens an avenue for the molecular mechanisms’ characterization of metastasis suppressor genes with the aim to understand their roles. Keywords: proteomics • melanoma • metastasis • metastasis suppressors • BRMS1 • DIGE

1. Introduction Metastasis, or the spreading of malignant tumor cells from primary foci at distant sites, is one of the most life-threatening aspects of cancer, representing around 90% of deaths in cancer patients. On the basis of current cancer biology knowledge, it is clear that, once malignant cells have spread and formed secondary foci, cancers are largely incurable, despite the progress in clinical medicine. Metastasis constitutes a complex process which has to fulfill a cascade of events, known as the metastatic cascade. A significant amount of evidence is acumulating favoring the notion that not only cancer cells are important, but also host cells may be useful therapeutic targets for treatment of cancer patients. To improve treatment, better molecular targets are needed to successfully distinguish be* Corresponding authors. Dr. Jose´ Rivera Torres: e-mail, [email protected]. Dr. Jero´nimo Bravo Sicilia: e-mail, [email protected]; phone, +34 91 224 69 00; Fax, + 34 912 246 976. † Signal Transduction Group, Structural Biology and Biocomputing Programme, Centro Nacional de Investigaciones Oncolo´gicas (CNIO). ‡ Confocal Microscopy and Cytometry Unit, Biotechnology Programme, Centro Nacional de Investigaciones Oncolo´gicas (CNIO).

4006

Journal of Proteome Research 2007, 6, 4006-4018

Published on Web 09/14/2007

tween cells showing high relevant metastatic potential from those that remain normal. Learning more about the cellular and molecular bases of the metastatic cascade events will help understand the mechanism and develop new therapeutic tools. Whereas a number of genes are known to inhibit both primary tumor formation and metastasis in experimental models, over the last 10 years, a new group of genes specifically involved in metastases suppression has been described. These genes, known as metastasis suppressor genes (MSGs), are capable of reducing metastasis without displaying a significant effect on tumorigenicity upon re-expression at physiologic levels in a metastatic cell line.2 Further studies have shown that MSGs regulate metastasis at many stages, with the majority of them being documented as playing a role at the last steps of the metastatic pathway affecting the metastatic colonization.2,3 At present, 114 out of 20 MSGs identified by different laboratories5 have been confirmed to display anti-metastatic effects in animal models. Moreover, the metastasis suppressors commonly have more than one validated role in metastasis that may serve to integrate various pathways. 10.1021/pr0703167 CCC: $37.00

 2007 American Chemical Society

research articles

Proteomic Analysis of BRMS1-Metastasis Associated Proteins

BRMS1, one of these MSGs, was identified by differential display upon comparison of metastasis-suppressed MDA-MB435 chromosome 11 chimeras with the parental MDA-MB-435 cell line.6 This gene has been shown to suppress human breast cancer and melanoma metastasis using in vivo models,7 and more recently, its ability to suppress human ovarian carcinoma metastasis also has been reported.1 The underlying mechanisms of action proposed for BRMS1 include facilitation of cellcell communication via gap junctions,8 interaction with histone deacetylases,9,10 shutting down of PI3K signaling,11 and also gene expression inhibition by targeting NFB.12 However, the molecular mechanism governing the metastasis suppression process remains largely unknown. Proteomic approaches, as 2D gel analysis, represent a powerful tool to identify differential protein expression pattern as well as post-translational modifications in complex biological systems. However, two-dimensional differential in-gel electrophoresis (2D-DIGE) coupled with mass-spectrometry analysis is a more powerful technique, indeed representing the proteomic equivalent of gene expression analysis by DNA microarrays. The system allows accurate detection of minor differences of protein expression across multiple samples simultaneously with statistical confidence by using the DeCyder software.13 The comparison of spot intensities using the DIGE approach and DeCyder software is more objective than the conventional approach based on the comparison of the brightness of gel images obtained by conventional staining.14 DIGE also allows sample normalization due to the presence of an internal standard during the entire experiment. The quantification of abundance changes is obtained over a linear dynamic range of 4 orders of magnitude due to the fluorescence properties of the Cye dyes.15 Besides, this method highly reduces the variability among samples because all of them are analyzed in the same gel under same conditions. This novel technology is being largely applied to the analysis of protein expression in different neoplasia for the search of markers.14 Although BRMS1 has been related to both breast and melanoma metastasis suppression, the vast majority of the reports only involves breast models. We have analyzed the differential expression profile in samples from wild-type (WT) melanoma cells, stable melanoma cells overexpressing BRMS1 (Mel-BRMS1), and stable melanoma cells where the BRMS1 endogenous levels were largely depleted by silencing (sh635). After protein extraction from the three different cultures included in the study, lysates were differentially labeled with Cy3 and Cy5 fluorescent dyes. Four replica samples were obtained from each cell type, which were randomly labeled and distributed along the six gels run. An internal pool was generated by combining equal amounts of the three extracts from all cell lines and then labeling with Cy2 fluorescent dye; this sample was included in all gel runs. Thus, 18 images were taken into account for the computational and statistical analysis by using the DeCyder software v.5.01. Seventy-nine spots were found to be differentially expressed when comparing the distinct cell lines with statistical variance within the 95th confidence level (Student’s t test; p < 0.05). Among those spots, 55 were unambiguously identified by further peptide mass fingerprint (PMF) and MS/MS analysis where applicable. Confirmation experiments by Western blot and functional activity assays were carried out. These proteins are mainly related to maintenance, cell growth regulation, and survival, as well as cell signaling. Such differentially expressed proteins

are candidates to be directly or indirectly related to the metastasis suppression exerted by BRMS1 in melanoma cells.

2. Materials and Methods 2.1. Generation of Retroviral Vectors. Three different short hairpin RNA (shRNA) were designed for silencing the endogenous BRMS1 protein with the help of Web-based algorithms (https://rnaidesigner.invitrogen.com/sirna; http://www.dharmacon.com/sidesign). No complementarity was found when compared with any other human genes as determined by BLAST. The shRNA sequences for the BRMS1 silencing oligonucleotides, which encode inverted repeats of 21 nucleotides separated by a 4-nucleotides spacer, were as follow (only top strands are shown): sh203, 5′CACCGAGAAGCAGTTCTCGGAGCTAACGAATTAGCTCCGAGAACTGCTT3′ sh387, 5′CACCGATCAGGAATAAGTACGAATGCGAACATTCGTACTTATTCCTGAT3′ sh635, 5′CACCGACATCGTGTACATGCTTCAAGCGAACTTGAAGCATGTACACGATG3′ Nomenclature for the shRNAs indicates sequence localization within the full-length BRMS1 mRNA sequence (BC009834). Sense and antisense BRMS1-specific sequences are underlined. Annealed oligonucleotides were cloned into a Pol-III U6 promoter-based shRNA vector (pENTR/U6; Invitrogen). This procedure results in a pENTR/U6 plasmid containing a human U6-driven silencing cassette flanked by recombination sites from phage λ (attL1 and attL2). The cassette was then transferred to a pMSCVpuro (Clontech) retroviral modified destination vector containing a Gateway destination cassette (phage λ attR1 and attR2 recombination sites flanking a ccdB toxic gene) cloned upstream of the GFP reporter gene (directed by an IRES) by performing an LR clonase reaction (Invitrogen). GFP expression indicates the presence of the vector in the resulting melanoma cell line. To overexpress BRMS1 WT protein in the melanoma cell line, we inserted the human BRMS1 full-length coding sequence upstream of the yellow fluorescence protein (YFP) gene into the pEYFP-N1 vector (Clontech). This construct was utilized as a template to amplify the BRMS1-YFP fusion protein by PCR. The resulting amplicon was cloned into the HpaI site of pMSCVpuro (Clontech) retroviral vector. As a control vector, the PCR-amplified YFP sequence was inserted into pMSCVpuro, thus, generating pMSCVpuro/YFP construct. Fidelity of the constructs was confirmed by DNA sequencing. The exact primer sequences and cloning strategies are available from the authors upon request. 2.2. Cell Culture and Stable Cell Line Generation. The WT melanoma cell line utilized for the present study kindly provided by Dr. Carnero (CNIO, Madrid, Spain) was obtained as previously described.16 Cells were maintained at 37 °C in a 5 % CO2 atmosphere in F-10 Ham medium supplemented with 10% heat-inactivated foetal bovine serum (FBS) (GibcoBRL), 2 mM glutamine, 1% Ultroser G (Biosepra), 100 U/mL penicillin, and 100 µg/mL streptomycin. Human embryonic kidney 293T (HEK-293T) (ATCC CRL-11268) packaging cell line was maintained in DMEM medium supplemented with 10% FBS, 2 mM glutamine, 100 U/mL penicillin, and 100 µg/mL streptomycin. HEK-293Tpackaging cell line was plated, at 5 × 106 cells per 100 mm diameter dish, 12 h prior to transfection. Then, cells were transfected, using standard Lipofectamine 2000 (Invitrogen) protocols, with either 10 µg of pCL-Amphotropic vector Journal of Proteome Research • Vol. 6, No. 10, 2007 4007

research articles and 10 µg of the desired plasmid vector, or with a negative control vector. Viral supernatants, harvested 48 h post-transfection and filtered, were used as viral stocks. Melanoma recipient cells, plated in advance, were incubated with the retroviral supernatants and supplemented with 4 µg/mL Polybrene, for 4 h at 37 °C. Infections were carried out sequentially three times. Then, cells were washed once with PBS and cultured in growth medium in the presence of 0.5 µg/mL puromycin until complete selection was achieved. Use of the fluorescence-activated cell sorter was carried out to enrich for subpopulations of human melanoma cells expressing the YFP reporter. 2.3. Quantitative PCR (qPCR). Total RNA, from the three different stable cell lines generated, was extracted with the RNeasy Mini Kit (Qiagen) and was reverse-transcribed to cDNA according to the manufacturer’s instructions (SuperScript II kit; Invitrogen). Appropriate gene-specific primer sets were designed (sequences are available from the authors upon request). Real-time PCR was carried out in triplicates using SYBR Green PCR Master Mix (Applied Biosystems) on an ABI Prism 7900HT Sequence Detector (Applied Biosystems) using the following parameters: 50 °C for 2 min and 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. The primer sequences used for q-PCR, designed based on published sequences, were as follow: 5′-AGC GAG TGT GTC AGT GAG ATG-3′ and 5′-AGG TGC TGT TTG GCT CCC TGC-3′ for BRMS1. To control for the amount of cDNA loaded, expression of the housekeeping β-actin gene serves as an endogenous normalization control amplified with primers 5′-CATGTACGTTGCTATCCAGGC-3′and 5′-CTCCTTAATGTCACGCACGAT-3′. Sequence Detector software (version 2.2.2) was utilized for data analysis, and relative fold induction was determined by the comparative threshold cycle method (Ct).17,18 2.4. 2D-DIGE. Cells were harvested when 80% of confluency was attained, then washed with ice-cold PBS twice. The resulting pellets were lysed by sonication in the appropriate lysis buffer for 2D-DIGE analysis (7 M urea, 2 M thiourea, 25 mM Tris-HCl, and 4% CHAPS, pH 8.5). Protein extracts were precipitated using the 2D Clean-UP kit (GE Healthcare) and labeled according to the manufacturer’s instructions (CyDye DIGE fluor minimal labeling kit, GE Healthcare). Briefly, 50 µg of protein from WT-YFP, BRMS1-YFP, and BRMS1 sh635 extracts was minimally labeled with 400 pmol of Cy3 or Cy5 fluorescent dye on ice for 30 min, in the dark. An internal pool was generated by combining equal amounts of protein from all three samples paired cell lines included in the study. This pool was labeled with Cy2 fluorescent dye and was included in all the gels run. The labeling reaction was quenched with 0.2 mM lysine. Following the labeling reaction, two samples, along with a pool aliquot, mixed, and run in a single gel. The samples were focused using broad range IPG strips (pH 3-11 non-lineal gradient, 24 cm) on an IPGphor apparatus (GE Healthcare). The IPG strips were equilibrated for 15 min with gentle shaking in 50 mM Tris-HCl, pH 8.8, containing 6 M urea, 4% (w/v) SDS, 65 mM DTT, 30% glycerol, and a trace of bromophenol blue. Iodoacetamide (53 mM) was added to the second equilibration buffer instead of DTT, and the strips were incubated for 15 min longer in this solution. Standard continuous SDS-PAGE electrophoresis for the second dimension (12%) was carried out at 15 mA/gel for 16 h. Proteins were visualized by using a fluorescence scanner at appropriate wavelengths for Cy2, Cy3, and Cy5 dyes (Typhoon 9400 ; GE Healthcare). Image analysis was carried out with the 4008

Journal of Proteome Research • Vol. 6, No. 10, 2007

Rivera et al.

DeCyder 5.01 software (GE Healthcare). The differential in-gel analysis (DIA) module was used for pairwise comparisons of each WT and stable melanoma cells overexpressing BRMS1 (Mel-BRMS1) on one side and WT against sh635 sample to the mixed standard present in each gel, and for the calculation of normalized spot volumes/protein abundance. The spot maps corresponding to the six gels were used to calculate average abundance changes and paired Student’s t-test p-values for each protein across the six gels. This was done by using the DeCyder biological variation analysis (BVA) module and the Cy3/Cy2 and the Cy5/Cy2 ratios for each individual protein. 2.5. PMF Identification by MALDI-MS and Database Searching. A preparative gel containing 600 µg of protein (coming from the 3 extracts) was run to identify those proteins of interest. Proteins were detected by staining the gel with SYPRO Ruby dye following manufacturer’s guidelines (Molecular Probes). Briefly, the gel was first fixed in 30% methanol and 7.5% acetic acid overnight, and then incubated with SYPRO Ruby for 4 h. After three washing steps with water, proteins were visualized by fluorescence scanning. The image obtained was matched against the analytical ones by using the DeCyder software, and spots were excised from the gel automatically using an Ettan Spot Picker robot. Proteins were subjected to tryptic digestion according to a previous protocol19 with minor variations.20 Proteins were first reduced (10 mM DTT) and then alkylated (50 mM iodoacetic acid). Following vacuum-drying, the gel pieces were incubated with modified porcine trypsin (Promega) for 16 h at 37 °C, at a final concentration of 10 ng/ µL in 50 mM ammonium bicarbonate. Two additional extraction steps were carried out with the digested samples in order to achieve complete peptide recovery. Finally, the peptide extracts were vacuum-dried and dissolved in 10 µL of 0.1% TFA/ 33% ACN. Peptide extracts were deposited onto MALDI plates using 2 vol of a matrix of R-cyano-4-hydroxicinnamic acid (5 mg/mL) in a solvent consisting of ACN/methanol/water (40:50:10 v/v/ v) with 0.1% TFA. The samples were allowed to dry at room temperature. A 4800 MALDI TOF/TOF Analyzer (Applied Biosystems/MDS SCIEX) was used to analyze samples, and spectra with signalto-noise ratio >10 were obtained for the m/z range corresponding to polypeptides with molecular mass of 0.9-4 kDa, under optimized instrument settings. Each spectrum was the result of an average of 1000 laser shots, delivered in 20 sets of 50 shots to each of 20 different positions on the surface of the matrix spot. For MS/MS data collection, 2500 laser shots per spot were used, delivered in 50 sets of 50 shots to each of 50 different positions on the surface of the matrix spot. The entire irradiation program was automated using the instrument software function. Spectra were acquired in reflector mode geometry and calibrated with a mix of commercially available calibration standards (Bruker Daltonics catalog no. 206195 (1000 units) in its level of fluorescence than the cell line overexpressing BRMS1 (10-100 units) as shown in Figure 1A. A high fluorescence (100-1000 units) also was observed in the melanoma cell line where the endogenous BRMS1 was downregulated by the constitutive expression of a short hairpin RNA (sh635). Taken together, these results suggest that endogenous levels of BRMS1 might be finely balanced and that its overexpression above a delicate threshold could compromise cell survival. Conversely, BRMS1-silenced cells are still capable of culturability with no problems detected. Available commercial antibodies against BRMS1 (Abnova) allowed only detection of BRMS1 in overexpressing cell lines but were insufficient for the detection of endogenous levels of the protein either in WT or sh635 cells (data not shown). We therefore decided to analyze BRMS1 levels in the different cell lines generated by qPCR. To achieve this goal, we extracted Journal of Proteome Research • Vol. 6, No. 10, 2007 4009

research articles RNA from each cell line, followed by cDNA synthesis and further real time PCR amplification, using β-actin as an internal normalizing control for data analysis as determined by the Ct method.17,18 As shown in Figure 1B, the levels of BRMS1 in the Mel-BRMS1 cell line are slightly over 5-fold higher than the endogenous levels. To examine the potential role of loosing BRMS1 expression in a cell line, we designed retroviral vectors encoding for short hairpin RNAs (shRNAs). The Brms1 gene was targeted at three different locations, precisely from nucleotides 203 to 223 (construct named sh203); nucleotides 387-407 (sh387), and from nucleotide 635 to 655 (sh635), as described above in Material and Methods. These vectors also contain an EGFP reporter gene downstream from the shRNA cassettes driven by an internal ribosome entry site (IRES) sequence as an internal control for transfection/transduction. Melanoma cells were transduced as the recipient cell line with either control YFP retroviral vector or the three different BRMS1 silencing constructs. After puromycin selection of cells expressing the retroviral cassette, they were further enriched by passing them through a cell sorter for EGFP detection. Figure 1B shows that sh203 construct had no effect on the endogenous BRMS1 mRNA levels compared to the parental WT cell line. However, the sh387 and sh635 constructs showed a pronounced effect on BRMS1 levels, and the sh635 construct achieved the highest level of down-regulation of endogenous BRMS1. We therefore selected the sh635 cell line for further characterization. 3.2. Morphological Changes in Silenced BRMS1 and YFP Transformed Melanoma Cells. Control WT melanoma cells overexpressing YFP (WT) did not show any detectable alteration in cell shape when compared to WT melanoma cells showing the typical flat morphology of epithelial cells accompanied by a high number of protrusions with lamellipodia as the predominant structure at the cell periphery. Similarly, the MelBRMS1 cell line was morphologically indistinguishable from the WT cells (data not shown). Conversely, those cells where the cellular BRMS1 levels had been highly depleted because of endogenous production of a powerful small interfering RNA (sh635) showed a high decrease in lamellipodia formation along with a more elongated morphology compared to the WT melanoma cells (Figure 1C). An intermediate state was observed for cells expressing the sh387 RNA. In all cases, the presence of vector vehicle and silenced short hairpin was confirmed by detecting the presence of GFP under a fluorescence microscope (data not shown). No morphological changes were found in cells overexpressing a nonfunctional short hairpin RNA (data not shown). Thus, the reduction of the endogenous levels of BRMS1 correlates with a significant alteration in cell morphology which might play a role in melanoma cell behavior. This alteration might be important for the metastatic process, and it is currently under study. 3.3. Identification of Differentially Expressed Proteins by 2D-DIGE Analysis Coupled with MS. Differential protein expression patterns for WT, Mel-BRMS1, and sh635 melanoma cell lines were analyzed by 2D-DIGE. After protein cell extraction, lysates were differentially labeled with Cy3 and Cy5 dyes, and a pool fraction from the three of them was labeled with Cy2 dye as an internal standard control. Four independent samples were prepared for the six different replica gels run and considered for the quantitative and statistical analysis using the DeCyder software. This procedure allowed all possible pairwise comparisons between the different cell lines (WT vs 4010

Journal of Proteome Research • Vol. 6, No. 10, 2007

Rivera et al.

Figure 2. Image of a representative 2D-DIGE gel. Protein extracts from each cell line were differentially labeled with Cy3 and Cy5 and pairwise-analyzed in replicates. A pooled internal standard combining all the proteins from both extracts, labeled with an additional Cy2 dye, was included in all gels. IPG strips, ranging from pH 3-11 in a non-lineal gradient, were used for the IEF in the first dimension and 12% SDS-PAGE in the second dimension. The image shows the spot map corresponding to the mixed internal standard, which is common to all gels analyzed. Arrows indicates protein spots identified in the screening as shown in Tables 1 and 2 with abundance changes within a 95th confidence level. Asterisks represent those spots that could not be identified after PMF analysis.

Mel-BRMS1, WT vs sh635, and Mel-BRMS1 vs sh635) to a pooled standard for the calculation of normalized spot volume ratios. Replica images were used to calculate average changes and Student’s t-test p-values for the detection of significant abundance changes in the protein expression among the samples. After image analysis of the Mel-BRMS1 versus WT and sh635 versus WT, the software showed changes in protein abundance of 79 spots, with a statistical variance within the 95th confidence level (Student’s t test: p < 0.05). Spots not showing such statistical significance values were discarded for further studies. All the spots shown in Figure 2 were excised from a preparative gel, followed by in-gel digestion and further identification by MALDI-TOF MS. Figure 2 shows the 2D-map of a representative gel indicating those proteins which expression levels changed consistently between the three biological samples. Out of the 79 proteins analyzed by MS, 55 were unambiguously identified by PMF and MS/MS (70% identification rate). These 55 identifications corresponded to 43 different proteins. Proteins found to be differentially regulated (up or down) appear in the 2D gel image (Figure 2) with their identified code name. These code names are also used in Tables 1 and 2. Spots marked with an asterisk in Figure 2 indicate those proteins that we were not able to identify. As it can be seen, most of the non-identified proteins show weak staining, suggesting that those proteins are expressed at low levels. We plan to perform different fractionation strategies to enrich the extracts in proteins at the low level of expression. In some cases, the same protein was identified in different spots over the gel, suggesting the occurrence of posttranslational modifications. In a few other cases, several

research articles

Proteomic Analysis of BRMS1-Metastasis Associated Proteins

Table 1. Differentially Expressed Proteins between Stable Mel-BRMS1 and WT Melanoma Cell Samples Identified by 2D-DIGE and MS Analysis prot. acc. no.a

name

Mr (Da)

pI

Mascot score

Coverage (%)

P06733 P04406

ENOA G3P

47037 35922

7.0 8.6

224 103

35 19

P09211 P18669 P14618 P32119 P30041

GSTP1 PGAM1 PKM2b PRDX2 PRDX6

23224 28672 57937 21760 25035

5.4 6.7 7.9 5.7 6.0

140 164 88 54 182

29 28 27 9 50

P31946 P07355 P08758 Q13873

14-3-3R ANXA2 ANXA5 BMPR2

27951 38472 35805 115201

4.8 7.6 4.9 5.8

135 219 500 35

36 36 48 1

O43852 P37235 P30626 P31948

CALUb HPCL1 SORCN STIP1

37106 22182 21676 62639

4.5 5.2 5.3 6.4

103 136 136 140

20 26 31 29

P01009 P04792 P07339

A1AT HSPB1 CATD

46736 22782 44552

5.4 5.9 6.1

105 282 58

P67809

YBOX1

35793

9.9

119

P04632 P10809

CPNS1 HSP-60b

28315 61055

5.0 5.7

84 56

P61163 P12814 P35609 P23528 P08670 P02545 P52565 P04264 P35527 P35527 P07437 P68371

ACTZ ACTN1b ACTN2b COF1 VIMb LMNA GDIR K2C1b K1C9b KRT10b TBB5 TBB2C

42613 103057 103853 18371 53652 71439 23075 65886 62129 59519 49670 49831

6.2 5.2 5.3 8.3 5.0 6.5 5.0 8.2 5.2 5.1 4.8 4.8

86 86 109 93 79 176 171 88 107 66 218 217

Q9Y5Z4

HEBP2

22875

4.6

53

a

peptides matched

average ratiod

Enzyme Activity 13 5 18 5 8 + 2 (MS/MS) 2 (MS/MS) 7 (MS) + 4 (MS/MS) Signal Transduction 9 12 11 2 (MS/MS)

5 5 5 18 Protein Metabolism 18 7 39 8 13 5 (MS/MS) Nucleic Acid Metabolism 15 5 Protein Folding 13 3 27 10 Cell Growth and Maintenance 22 6 5 4 7 7 6 1 (MS/MS) 22 7 + 1 (MS/MS) 28 15 (MS) + 2(MS/MS) 25 4 29 14 17 6 18 8(MS)+1(MS/MS) 38 14 38 14 Unknown Activity 13 2 (MS/MS)

-2.19 -1.87 -2.07 -1.39 3.27 -1.62 -1.43 -2.82 -1.49 -1.64 -1.94

common name and function

Alpha-enolase Glyceraldehyde-3-phosphate dehydrogenase Glutathione S-transfer P Phosphoglycerate mutase 1 Pyruvate kinase 2/3 Peroxiredoxin-2 Peroxiredoxin-6

-2.69 -1.9 -4.44 -1.69

14-3-3 protein beta/alpha Annexin A2 Annexin A5 Bone morphogenetic protein receptor type-2 [Precursor] Calumenin [Precursor] Hippocalcin-like protein 1 Sorcin Stress-induced- phosphoprot. 1 (Hop)

3.18 -1.98 -2.82

Alpha-1 protease inhibitor Heat shock protein 27c Cathepsin D [Precursor]

2.3

Y-box transcription factor

-2.21 3.27

Calpain small subunit 1 Heat Shock Protein 60

-2.37 -2.69 -2.69 1.95 3.27 -2.04 -2.93 -2.59 -2.59 3.27 -2.91 -2

Alpha-centractin Alpha-actinin-1 F-actin cross-linking prot Cofilin1 Vimentin LamininA-C Rho GDP-dissociation inhibitor 1 Keratin 1 Keratin 9 Keratin 10 Tubulin beta-5 chain Tubulin beta-2C chain

-2.31

Heme-binding protein 2

b

Swiss-Prot database accession number are referenced. More than one protein were identified in this spot; thus, the abundance ratio corresponds to the combination of all these proteins. c It has been shown to play a role in actin cytoskeleton polymerization and reorganization. d Average ratios were calculated considering 6 replica gels. Proteins grouped by function according to the Human Protein Reference Database (http://www.hprd.org/).

individual proteins were clearly identified within the same spot, as noted by footnote b in Tables 1 and 2, probably due to the broad pH range used for the IEF. In these particular cases, the ratio obtained resulted from the combination of all the proteins present in the spot, and further experiments are necessary to confirm changes in individual proteins. Tables 1 and 2 summarize the 43 unique distinctly deregulated proteins that were unambiguously identified by MS, taking into account the MASCOT score, the sequence coverage, and the number of matching peptides over all the searched ones. Intriguingly, we identified twice as many deregulated proteins (34 against 17) upon BRMS1 overexpression than when BRMS1 was silenced. Besides, a vast majority (80%) of identified proteins in MelBRMS1 cells (Table 1) were down-regulated. Conversely, 100% of the proteins identified in sh635 cell cultures (Table 2) were up-regulated. Eight proteins are present in both Tables 1 and 2 showing an inverse correlation between their expression levels and that of BRMS1.

We grouped the identified proteins according to the biological processes in which they are likely to play a role (Figure 3) such as signaling (26.5% or 23.5% in Mel-BRMS1 and sh635, respectively), cell growth and maintenance (35.3%; 23.5%), protein (8.8%; 17.6%) and nucleic acid metabolism (2.9%; 5.9%), protein folding (2.9%; 5.9%), and enzyme activity (20.6%; 23.5%) following the Human Protein Reference Database21 (http:// www.hprd.org/) recommendations. Most of the proteins deregulated by either BRMS1 overexpression or silencing show a primary cytosolic localization and are involved in cell growth and signal transduction activities. The assigned protein functions were similar in both groups, though the number of proteins involved in protein folding and protein metabolism was almost twice in BRMS1-silenced cells compared to cells overexpressing it (see Figure 3A). This result could suggest that in BRMS1-depleted cells, as it occurs in metastatic cells, protein folding, and metabolism activities, might have an important role for a normal cell to acquire a metastatic phenotype. Journal of Proteome Research • Vol. 6, No. 10, 2007 4011

research articles

Rivera et al.

Table 2. Differentially Expressed Proteins between Stable sh635 and WT Melanoma Cell Samples Identified by 2D-DIGE and MS Analysis prot. acc. no.a

name

Mr (Da)

pI

Mascot score

P06733 P18669 P15259 P30041

SERA PGAM1b PGAM2b PRDX6

56651 28804 28766 25035

6.3 6.7 9.0 6.0

81 138 56 182

P31946 P61981 O43852 P37235

14-3-3 14-3-3γ CALUb HPCL1

29174 28303 37106 22182

4.6 4.8 4.5 5.2

86 63 97 136

P01009 P04792 P07339

EF1G HSPB1 CATD

50119 22782 44552

6.2 5.9 6.1

106 288 75

P07910

HNRPC

33670

4.9

87

P62938

PPIA

18012

7.7

135

P13645 P35908 P07437 P68371

K1C10b K22Eb TBB5b TBB2Cb

59519 65865 49670 49831

5.1 8.1 4.8 4.8

165 124 333 302

coverage (%)

peptides matched

Enzyme Activity 2 (MS) + 2 (MS/MS) 3 (MS) + 3 (MS/MS) 2 (MS/MS) 7 (MS) + 4 (MS/MS) Signal Transduction 28 3 (MS) + 2 (MS/MS) 20 2 (MS) + 2 (MS/MS) 13 2 (MS) + 2 (MS/MS) 26 5 Protein Metabolism 15 2 (MS) + 4 (MS/MS) 52 4 (MS) + 5 (MS/MS) 17 4 (MS) + 2 (MS/MS) Nucleic Acid Metabolism 15 2 (MS) + 2 (MS/MS) Protein Folding 31 4 (MS) + 1 (MS/MS) Cell Growth and Maintenance 25 9 (MS) + 3 (MS/MS) 17 8 (MS) + 2 (MS/MS) 50 13 + 4 (MS/MS) 48 13 + 3 (MS/MS) 10 38 8 50

average ratiod

common name and function

1.77 1.9 1.9 1.45

D-3-phosphoglycerate dehydrogenase Phosphoglycerate mutase 1 Phosphoglycerate mutase 2 Peroxiredoxin-6

1.44 1.72 1.85 2.34

14-3-3 protein epsilon Protein kinase C inhibitor protein 1 Calumenin [Precursor] Hippocalcin-like protein 1

1.51 1.72 2.55

Elongation factor 1-gamma Heat shock protein 27c Cathepsin D [Precursor]

2.2

Heterogeneous nuclear ribonucleoprotein

2.26

Cyclophilin A

1.87 1.41 1.16 1.16

Keratin 10 Cytokeratin 2e Tubulin beta-5 chain Tubulin beta-2C chain

a Swiss-Prot database accession number are referenced. b More than one protein were identified in this spot; thus, the abundance ratio corresponds to the combination of all these proteins. c It has been shown to play a role in actin cytoskeleton polymerization and reorganization. d Average ratios were calculated considering 6 replica gels. Proteins grouped by function according to the Human Protein Reference Database (http://www.hprd.org/).

Interestingly, more than 50% of the deregulated proteins reported to play a role in signal transduction were involved in calcium binding. This number still could increase considering that CPNS1 (calpain small subunit1) has also been described as a regulatory subunit of the calcium-regulated nonlysosomal thiol-protease which catalyzes limited proteolysis of substrates involved in cytoskeletal remodelling and signal transduction.22 A representative network showing direct interactions, protein partners binding, and targets of some of the identified proteins in the 2D-DIGE screening is depicted in Figure 3B. Figure 4 illustrates the quantitative analysis of four, out of the eight identified proteins, that showed an inverse correlation with BRMS1 levels. The graph displays the average ratio of each protein shown as the standardized Log abundance, referred to the internal standard. The p-value representing the Student’s t test (p < 0.05) is also indicated for each pairwise comparison. In all cases, the t-test p-values were below this threshold. The insets show representative 3D views of the abundance of each protein, as revealed by DIGE analysis of one of the 2D replica gel used for the statistical analysis. 3.4. Validation of Differential Expression Data by Western Blot Analysis. To validate the differential expression observed in our 2D-DIGE screening, we selected four candidates, namely, R1 anti-trypsin (A1AT),23 Heat shock protein 27 (HSPB1),24,25 Annexin2 (ANXA2),26 and Cofilin1 (COF1),27 due to their relationship to the tumoral processes, for further study by immunoblot analysis. As shown in Figure 5A, A1AT was strongly up-regulated in Mel-BRMS1 cells as compared to WT cells, where the endogenous levels were almost undetectable. Despite no statistically significant changes observed for A1AT expression in sh635 cells after 2D-DIGE analysis, Western blot proved that the level of A1AT expression is at least similar to the WT levels. 4012

Journal of Proteome Research • Vol. 6, No. 10, 2007

Quantification of three independent analyses by densitometry showed that HSPB1 was down-regulated in Mel-BRMS1 cells compared to WT cells as shown in Figure 5A,B. We were also able to confirm the differential expression of HSPB1 in sh635 cell population, where the protein showed an increase when compared to WT melanoma cell line. After Western blot analysis, we could see the down-regulation of Annexin 2 in cell lysates from Mel-BRMS1 cultures when compared to the WT cells. This protein did not show significant changes in sh635 melanoma cells after the 2D screening; however, immunoblot analysis shows that silencing of endogenous BRMS1 correlates with an increase in the ANXA2 protein levels. COF1 is capable of regulating actin dynamics when phosphorylated by LIM kinase1 on its serine 3, rendering it inactive, as phospho-cofilin is unable to bind actin and mediates its depolymerization.28 To validate the results of 2D-DIGE, where we found almost 2-fold increase of COF1 in Mel-BRMS1 compared to WT melanoma cells, and analyze the activity state of COF1, we carried out an immunoblot assay, utilizing an antihuman phospho-cofilin1 which recognizes cofilin only when phosphorylated at serine 3. As shown in Figure 5B,C, MelBRMS1 cells significantly up-regulated the F-actin stabilizing form of COF1. These phospho-protein levels show a similar level to basal in sh635 cell lysates. Taken together, Western blot results demonstrate a convincing proteomic analysis of stable melanoma cell lines. 3.5. Cathepsin D Heavy Chain (34 kDa) Is Overexpressed in BRMS1-Silenced Melanoma Cells. Cathepsin D (CATD) is an aspartic endoprotease that is ubiquitously distributed in lysosomes and it is synthesized on the rough endoplasmic reticulum as an inactive 52 kDa pro-cathepsin D form. In the early lysosomes, the inactive form is converted to an active single chain intermediate of 48 kDa. Then, the protein is further

Proteomic Analysis of BRMS1-Metastasis Associated Proteins

research articles

Figure 3. Functional categories and identification of signaling pathways of differentially modified proteins upon BRMS1 deregulation in melanoma cell lines. (A) Analysis of identified protein reveals proteins from diverse functional categories within the different subcellular compartments. Functional classification was performed according to the “The Human Protein Reference Database”. (B) The protein list described in Tables 1 and 2 was queried against the high-confidence and manually curated ResNet database using the Pathway Studio software analysis (Ariadne Genomics, Inc.). The software allowed the identification and visualization of pathways, gene interaction networks, and protein interaction maps. The figure displays some of the identified proteins after 2D-DIGE screening of BRMS1 deregulated melanoma cell lines along with some of their multiple interactions: binding (purple line); direct regulation (gray line), protein modification (green line) between some of the identified proteins (*). ARHGDIA and YWHAB in the picture correlates to GDIR and 14-3-3 R proteins, respectively, as shown in Table 1.

processed into a light (14 kDa) amino-terminal chain and a heavy (34 kDa) carboxyl-terminal chain.29 In the 2D-DIGE screening, we detected CATD in three different spots (Figures 2 and 6A; C1, C2, and C3). All of them display similar apparent molecular masses in the 2D gel, corresponding to the 34 kDa heavy form and showed reduced intensities in WT and Mel-BRMS1 cells compared to sh635 melanoma cell lines (Figure 6B). These multiple isoforms of CATD are present in the three stable melanoma cell lines generated and are down-regulated in WT and Mel-BRMS1 cell lines compared to sh635 cell line. Immunoblots of 2D gels probed with anti-human CATD antibody directed against the carboxy-terminal end also revealed different forms, each with different pI values. All detected isoforms have a similar molecular mass, and the acidic form is the most abundant in melanoma cell lines (Figure 6C). These data suggest the existence of additional post-translational modifications on the heavy chain of CATD to those previously described. In-depth analyses should be performed to characterize these modifications and their relevance on CATD functionality. 3.6. BMPR2 Is Down-Regulated in Stable Melanoma Cell Line Overexpressing BRMS1. Bone morphogenetic protein receptors (BMPRs) are transmembrane serine/threonine protein kinases that belong to the TGF-β family of receptors. Ligand binding leads to phosphorylation of Smad proteins causing their activation and translocation to the nucleus, where they regulate transcription of BMP-responsive genes.30 Our 2D-DIGE analysis identified a single spot corresponding to the receptor, BMPR type II (BMPR2), which was reduced

almost 2-fold in the Mel-BRMS1 cells compared to WT cells (Table 1). No significant change for the level of this protein was found in sh635 cells. To assess whether the down-regulation of BMPR2 in MelBRMS1 cell line correlates with a decreased Smad 1-5-8 phosphorylation activity, we incubated the cells in normal growth medium in the presence (100 ng mL-1) or absence of its cognate BMP2 ligand for different exposure times. The ability of BMP2 to phosphorylate Smad 1-5-8 was examined by Western blot analysis using an antibody which specifically recognizes the phosphorylated form. As shown in Figure 7, untreated Mel-BRMS1 cells showed an approximate 85% reduction in Smad 1-5-8 phosphorylation levels when compared to the untreated WT melanoma cells. BMP2 induced phosphorylation of Smad 1-5-8 in MCF7 cells used as a positive control for BMPR2 signaling through the Smad protein phosphorylation mechanism (data not shown). The presence of the ligand in the growth medium induced the phosphorylation of Smad 1-5-8 either in WT and BRMS1 cells with a medium short-term effect achieving the maximum after 90 min of treatment. At these times, the phosphorylation levels were consistently lower in BRMS1 cells compared to WT cells in all the experiments carried out as depicted in Figure 7A where a typical blot result is shown. As expected, this effect is reverted in BMP2-treated sh635 cells. Actually, these cells show increased phosphorylation levels of Smad 1-5-8 compared to WT cells, at all tested time points. For quantification of the Western blot results, phospho-protein levels were normalized against the total Smad protein levels. Then, phospho-protein levels, obtained after treatment with Journal of Proteome Research • Vol. 6, No. 10, 2007 4013

research articles

Rivera et al.

Figure 4. Expression abundance and 3D views of four selected proteins differentially regulated in Mel-BRMS1 (() and sh635 (9) when independently compared to WT (b) melanoma cells. The graph view represents the average ratio of expression for each selected protein, as obtained by computational analysis with DeCyder 5.01 software which allows the detection of significant abundance changes (95th confidence level) based on the variance of the mean change within the cohort; p-values are indicated. The insets shows a representative 3D view for each protein in a particular replica gel from the three different cell lines compared along the study.

Figure 5. Validation of 2D-DIGE results by Western blot analysis. (A and C) Total cell lysates from WT, Mel-BRMS1, and sh635 melanoma cell lines were analyzed by Western blotting. The expression of different metastasis-related proteins identified by 2D-DIGE screening were analyzed with the indicated antibodies. Anti-human actin was utilized as loading control. (B) Relative protein expression of HSPB1, ANXA2, and phospho-Cofilin (pCofilin) was normalized to the signal intensity of β-actin as an internal standard. Values for Mel-BRMS1 and sh635 are referred to WT levels as fold induction. Data represent the mean ( SD of triplicate determinations. Images were captured by a documentation system and analyzed by ImageJ version 1.37. A1AT, R1antitrypsin; HSPB1, heat shock protein 27; ANXA 2, annexin2.

BMP2 ligand, of both Mel-BRMS1 and sh635 cell lines were referred to WT melanoma cells values. Results are shown in Figure 7B. Taking these results into account, an obvious question that arises is whether the differences in phosphorylation levels that 4014

Journal of Proteome Research • Vol. 6, No. 10, 2007

Figure 6. Increased Cathepsin D expression in sh635 melanoma cells. (A) A 2D-DIGE gel of cell extracts showing three different (C1, C2, C3) spots identified as Cathepsin D after PMF coupled with MS/MS analysis. All of them correlated with the 34 kDa form, but showed different pI values. (B) Immunoblotting of 1D gel probing cell extract from WT, Mel-BRMS1, and sh635 melanoma cell lines with an antibody raised against Cathepsin D. A total of 25 µg of total cell lysates was loaded onto a 12% SDS PAGE gel per lane. Sh635 cell extracts showed higher accumulations of Cathepsin D than the WT melanoma cells. (C) Immunoblotting the 2D gel with anti-human Cathepsin D raised against its C-terminal end detected several peptides (C1, C2, C3) indicating post-translational modification.

we identified have a biological role. Further work is in progress to address this and other related questions.

Proteomic Analysis of BRMS1-Metastasis Associated Proteins

Figure 7. Expression and functional activation of BMP type II receptor (BMPR2) in human melanoma cell lines. (A) Melanoma cell lines as described in the figure were untreated (0 min) or treated with human recombinant BMP2 ligand (100 ng mL-1) for 30, 60, and 90 min, and then phosphorylation of Smad 1-5-8 protein (pSmad 1-5-8) was analyzed by immunoblots. Data show a significant decrease in the phosphorylation levels upon exposure of the Mel-BRMS1 cell line to the BMP2 ligand (100 ng mL-1) compared to WT at different time points. These differences were even higher when compared to sh635 cell line. MCF7 breast cancer cell line was used as a positive control for BMP2-induced response (data not shown). (B) Relative protein expression of phospho-Smad 1-5-8 in the three different melanoma cells lines were normalized to the signal intensity of Smad 1 (Smad) as internal control, and then Mel-BRMS1 and sh635 Smad phosphorylation levels referred to the WT values at indicated time points. For this, membranes were incubated in stripping buffer for 45 min at 55 °C and then probed with Smad antibody. Error bars represent the mean value of three different experiments ( SD.

4. Discussion Metastasis is the ultimate cause of cancer-related death, yet it remains poorly understood perhaps due to its complexity. The use of novel technologies is much needed in order to progressively reduce this lack of understanding. 2D-DIGE can contribute to the characterization of some of the molecular aspects involved in the complex process of metastasis by comparison of highly with poorly metastatic cells. In the past few years, an increasing number of genes have been included within a protein family known as metastasis suppressors genes. MSGs have the ability to negatively regulate metastasis by suppressing the metastatic dissemination of cancer cells to secondary sites without affecting tumorigenicity at the primary site. Recent work by several groups indicates that BRMS1 serves as one of such metastasis suppressor genes, although the mechanistic insight into how BRMS1 suppresses metastasis remains largely unknown. The main goal of this study was to reveal downstream targets of BRMS1-mediated metastasis suppression in a melanoma cell model in order to shed light onto the molecular mechanism underlying it. Our results extend to melanoma cells, recent reports on the

research articles molecular mechanisms of action of BRMS1 such as the reduction of phosphoinositide signaling,11 and the NF-kB mediated osteopontin transcriptional inhibition in breast cancer cells.12 To simulate two different metastatic cell stages, we engineered two stable melanoma cell lines. In the first, the BRMS1 gene was overexpressed, thus, imitating a poor metastatic phenotype. In the second melanoma cell line, the endogenous levels of BRMS1 are constitutively down-regulated by expression of a highly specific short hairpin RNA (sh 635 cell line). This down-regulation leads to a loss of expression such as that reported for highly metastatic cell lines.1,31,32 Seventy-nine differentially expressed spot proteins were found to be statistically significant after 2D-DIGE screening between WT, BRMS1, and sh635 melanoma cell lines. Following PMF and MS/MS analysis, 43 different proteins were unambiguously identified. Six proteins, previously reported to play a role in metastasis, have been selected, namely, Cofilin1, A1AT, BMPRII, Hsp27, Annexin II, and Cathepsin D. The differential expression of these proteins was confirmed by Western blot (Figures 5 and 6) and functional analysis (Figure 7), therefore validating the results attained from the 2D-DIGE analysis. 4.1. Cofilin and 14-3-3 Proteins. Activity regulation of Cofilin 1, a regulator of actin dynamics, is mediated by its phosphorylation on Ser 3. In living cells, this phosphorylation is regulated by at least two distinct types of protein kinases, LIM kinases (LIMK) and TESK, of which TESK1 appears to play a role in integrin-mediated actin cytoskeletal reorganization.33 The 14-3-3 family of proteins consists of at least seven highly conserved mammalian isoforms (β, , γ, η, σ, τ, ζ). These proteins interact with diverse signaling proteins and thus are involved in a variety of cellular functions, including cell proliferation, cell cycle progression, differentiation, and apoptosis.34 Precisely, the β isoform has been reported to bind TESK1 and thus regulates its kinase activity.35 It postulates that binding of 14-3-3 β may have a role to sequester TESK1 as an inactive complex to prevent accession of activator proteins. BRMS1 overexpressing cells shows almost 3-fold downregulation of 14-3-3 protein (Table 1). A reduced level of 143-3 β protein might translate into an increase of free TESK1, that could promote inactivation of Cofilin 1 by phosphorylation. Interestingly, these results are in complete agreement with the increased levels in phospho-Cofilin that we observed after immunoblot of total cell extracts probed with an antibody which specifically recognize COF1 only when phosphorylated at its Ser3 (Figure 5B,C). Integrins are a family of cell surface heterodimer proteins that bind to extracellular matrix proteins and transduce the signal to control cell survival, proliferation, and migration.36 TESK1 functions downstream of integrins playing a key role in integrin-mediated actin reorganization.33 When TESK1 was overexpressed in a human cell line, it stimulated the formation of actin stress fibers and focal adhesions via Cofilin phosphorylation.33 We report here a down-regulation of 14-3-3 β protein as well as an up-regulation of Cofilin phosphorylation upon BRMS1 overexpression in melanoma cells. Our results suggest that BRMS1 might play an important role in integrin-mediated Cofilin phosphorylation, actin reorganization, and focal adhesion formation as described.33 Further studies are devoted to address this important question. 4.2. Bone Morphogenetic Protein Receptors (BMPRs). These are transmembrane ser/threonine protein kinases that belong to the TGF-β family of receptors, which consist of a type I signaling receptor and a type II ligand-binding receptor. Upon Journal of Proteome Research • Vol. 6, No. 10, 2007 4015

research articles ligand binding, type I and type II receptors are brought together on the cell surface. This event allows BMPR type II to phosphorylate and activate the BMPR type I receptors, which in turn phosphorylate Smad proteins. Phosphorylated Smad are thus activated and translocated to the nucleus, where they regulate the transcription of BMP-responsive genes.30 As shown in Table 1, we observe a 2-fold decrease on the expression levels of BMPR type II in cells overexpressing BRMS1, compared to the WT cell line. Despite their lower protein levels, these receptors were still capable of been stimulated by one of their very wellknown ligands (BMP2). This was demonstrated by Smad protein phosphorylation, detected by Western blot (Figure 7), after stimulation with 100 ng mL-1 of BMP2 ligand at different time points. Conversely, the levels of these receptors, also capable of stimulation, were higher in sh635 cells than in the WT cells. BMPR-II negatively regulates the activity of the second type of kinase (LIMK1) that phosphorylates Cofilin1.The tail of BMPRII binds LIMK1 reducing its level in the cytosol.37 Our findings in Table 1 and Figure 5 that, respectively, show the down-regulation of BMPRII and the increased levels of phosphorylation state of Cofilin1 in Mel-BRMS1 cells support this concept. It is possible that lower levels of BMPRII in this cell line increase the availability of LIMK1 in the cytoplasm, enhancing its interaction with other proteins, such as its activators, eventually leading to COF1 phosphorylation, among other effects. BRMS1 might therefore be affecting actin dynamics by controlling the phosphorylation state of Cofilin via two different ways. On the one hand, BRMS1 would contribute to keep low levels of 14-3-3β, as found in Mel-BRMS1 cells, thus, increasing the availability of TESK1 Cofilin kinase. On the other hand, BRMS1 would also maintain low levels of BMPRII enhancing the availability of a second COF1 kinase (LIMK1). Both ways would lead to an increase in the inactive phosphorylated COF1. Phospho-COF1 is largely related to a decrease in cell motility indicating its possible role in metastasis. In such a way, our results are in agreement with previous published results38 where BRMS1-transfected MDA-MB-435 cells showed a modest but significant reduction in motility compared to vehicle control as measured in a monolayer wound healing and a modified Boyden chamber assay. A different but not less interesting point is the fact that BMPs have been related with invasiveness in melanoma cells, suggesting that these proteins promote melanoma cell invasion and migration and therefore have an important role in the progression of malignant melanoma.39 It is known that malignant melanoma cells have altered expression of growth factors compared with normal human melanocytes favoring tumor growth and progression. In this sense, our results indicate that the overexpression of BRMS1 could have a protective effect against invasiveness capability by decreasing the type II receptors levels which render the cells less responsive toward the high growth factors levels within the tumor environment. 4.3. Cathepsin D (CATD). CATD, a lysosomal aspartic proteinase, is secreted in the form of an enzymatically inactive precursor, procathepsin D, in some human cancer cells.40 This precursor is significantly up-regulated in highly metastatic MDA-MB-435 cells playing an important role in the development and metastasis of several types of cancer including breast cancer.25 Several in vitro studies have shown that human 4016

Journal of Proteome Research • Vol. 6, No. 10, 2007

Rivera et al.

melanoma cells may release CATD cooperating in extracellular matrix degradation, thus, facilitating the invasion of the tumor.41,42 Our preliminary proteomic analysis identified CATD as a highly down-regulated protein (almost 3-fold) in Mel-BRMS1 cells compared to WT cells (Figure 3). Three different peptides containing the 34 kDa carboxyl-terminal were detected but showed different pI values, likely due to post-translational modifications (Figure 6A,C). Western blot analysis corroborated the identity of this active isoform after probing total cell extracts from the three different melanoma cell lines with an antibody raised against the COOH terminal end of the protein. Our data (Figure 6B) revealed also that sh635 cell extracts increased CATD levels higher than 2.5-fold (quantification data not shown) compared to the WT cells. Thus, the levels of BRMS1 are inversely correlated to CATD levels, suggesting that the capability of BRMS1 to negatively regulate metastasis might be due to down-regulation of CATD among other prometastatic proteins. Our results are consistent with the published observations supporting a role for CATD in melanoma metastasis transformation. 4.4. Annexins. Two annexins (ANXA2 and ANXA5) were found to be down-regulated in Mel-BRMS1 cell line (1.5- and 1.6-fold, respectively, relative to WT). These proteins belong to a family of cytosolic Ca2+-binding proteins. Calcium binding allows annexins to interact with membranes containing acidic phospholipids. These proteins are implicated in many membrane-related events, such as the regulated organization of membrane domains, regulation of phagocytosis, cell signaling, and proliferation.43 Little is known about ANXA5 expression in human cancer and metastasis. Overexpression of ANXA2 has been reported in pancreatic cancer and correlated with invasive potential of tumors.44 Annexin II was also identified as a metastasis-associated protein in a proteomic analysis of head and neck squamous cell carcinoma comparing primary and metastatic cell lines derived from the same patient.26 Furthermore, ANXA2 is highly expressed on the surface of metastatic sublines of murine large-cell lymphoma RAW117 cells and human colorectal carcinoma KM12 cells, whereas expression is less intense on the poorly metastatic sublines RAW117-P and KM-12C.45 In this sense, our results are in agreement, since less invasive melanoma cells (Mel-BRMS1) showed an inverse correlation with the annexin levels and support a role for ANXA2 and ANXA5 in metastasis promotion. Interestingly, annexins were not the only Ca2+-binding proteins identified to be down-regulated by BRMS1 in our study. We have also identified three other proteins described in the literature such as HPCL1,46 Sorcin,47 and Calumenin.48,49 The latter was identified using a proteomic approach as a metastasis-associated protein.26 However, as far as we know, the mechanism of its involvement in tumor metastasis remains unknown. Precisely, Calumenin and HPCL1 were strongly upregulated (1.8- and 2.3-fold, respectively) in sh635 cells compared to the parental counterpart. A previous report described the reduced capability of BRMS1-transfected MDA-MB-435 cell line to respond by intracellular calcium mobilization after growth factor (PDGF) stimulation, a known activator of phosphoinositide signaling.11 4.5. Drug- and Chemoresistance. Interestingly, some of the identified proteins such as sorcin, 14-3-3γ, and HSPB1 have been involved in drug and chemoresistance in different tumor types. 14-3-3γ was found to be overexpressed in chemoresistant melanoma cell lines.50 Overexpression of sorcin upon gene

research articles

Proteomic Analysis of BRMS1-Metastasis Associated Proteins

transfection resulted in increased drug resistance in human leukaemia cells.51 Finally, HSPB1 is a cytoprotective chaperone that is phosphoactivated during cell stress preventing aggregation and/or regulating the activity and degradation of certain client proteins. Recent evidence suggests that HSPB1 may be involved in tumor progression and in the development of treatment resistance in various tumors, including bladder,52 melanoma,53 and breast cancer.54 In our present analysis, Sorcin and 14-3-3γ, as well as HSPB1, were detected at higher levels in Mel-BRMS1 cells compared to the WT as shown in Tables 1 and 2. Thus, it is tempting to speculate that therapeutic treatments leading to an increase in the endogenous levels of BRMS1 could improve the sensitivity to chemo-therapy in cancer patients.

5. Concluding Remarks In the present work, we performed a large-scale identification of BRMS1 metastasis-associated proteins comparing two different stable melanoma cell lines generated against a WT melanoma cells. These cell lines imitate two distinct metastatic phenotypes: melanoma cell line overexpressing BRMS1 gene represents a poor metastatic phenotype, while its endogenous down-regulation, acquired by small interfering RNAs, represents an increase in metastatic capability. Combining general and specific proteomic methods, 43 differentially expressed proteins were reliably identified, constituting the largest available data set of effector proteins under BRMS1 regulation. The reliability of DIGE results were confirmed by Western blot and functional activity assays. Most of the differentially expressed proteins have clear relationships with the process of tumor metastasis, and their changes are consistent with the literature. Complementary experimental procedures are in progress to assess the biological functions of these proteins in a metastasis melanoma model. The present work enumerates for the first time a list of effectors for the BRMS1 gene, utilizing a 2D-DIGE approach suggesting that this strategy, which opens a large avenue to delineate the molecular mechanisms underlying the functions of other MSGs, could represent a rapid way to decipher the role where these metastasis suppressor genes might be involved in the metastatic disease. Abbreviations: BRMS1, Breast cancer metastasis suppressor 1; Mel-BRMS1, Melanoma cell overexpressing BRMS1; sh635, Melanoma cell with silenced endogenous BRMS1; MSGs, Metastasis suppressor genes; DIGE, digest-in-gel electrophoresis; PMF, peptide mass fingerprint; shRNA, Short hairpin RNA; q-PCR, Quantitative real time PCR; EGFP; Enhanced green fluorescent protein; ANXA, Annexin; A1AT, R1 anti-trypsin; HSPB1, Heat shock protein 27; COF1, Cofilin1; CATD, Cathepsin D; BMPR2: Bone morphogenetic protein receptor type II; BMP2, Bone morphogenetic protein 2.

Acknowledgment. We gratefully acknowledge the assistance of Patricia Alfonso, Francisco Fernandez, and Antonio Nun ˜ ez (Protein Technology Unit at CNIO, Madrid, Spain) for proteomic support and helpful discussions. We also thank Dr. Amancio Carnero and Victoria Moneo (CNIO, Madrid, Spain) for providing us the melanoma cell line. This work has been partially supported by Grant SAF2006-10269 of the “Ministerio de Ciencia y Tecnologı´a”, Spain, and “Fundacio´n Mutua Madrilen ˜ a”, Spain.

References (1) Zhang, S.; Lin, Q. D.; Di, W. Suppression of human ovarian carcinoma metastasis by the metastasis-suppressor gene, BRMS1. Int. J. Gynecol. Cancer 2006, 16 (2), 522-31. (2) Steeg, P. S.; Ouatas, T.; Halverson, D.; Palmieri, D.; Salerno, M. Metastasis suppressor genes: basic biology and potential clinical use. Clin. Breast Cancer 2003, 4 (1), 51-62. (3) Steeg, P. S. Tumor metastasis: mechanistic insights and clinical challenges. Nat. Med. 2006, 12 (8), 895-904. (4) Berger, J. C.; Vander Griend, D. J.; Robinson, V. L.; Hickson, J. A.; Rinker-Schaeffer, C. W. Metastasis suppressor genes: from gene identification to protein function and regulation. Cancer Biol. Ther. 2005, 4 (8), 805-12. (5) Eccles, S. A.; Welch, D. R. Metastasis: recent discoveries and novel treatment strategies. Lancet 2007, 369 (9574), 1742-57. (6) Seraj, M. J.; Samant, R. S.; Verderame, M. F.; Welch, D. R. Functional evidence for a novel human breast carcinoma metastasis suppressor, BRMS1, encoded at chromosome 11q13. Cancer Res. 2000, 60 (11), 2764-9. (7) Meehan, W. J.; Welch, D. R. Breast cancer metastasis suppressor 1: update. Clin. Exp. Metastasis 2003, 20 (1), 45-50. (8) Saunders, M. M.; Seraj, M. J.; Li, Z.; Zhou, Z.; Winter, C. R.; Welch, D. R.; Donahue, H. J. Breast cancer metastatic potential correlates with a breakdown in homospecific and heterospecific gap junctional intercellular communication. Cancer Res. 2001, 61 (5), 1765-7. (9) Liu, Y.; Smith, P. W.; Jones, D. R. Breast cancer metastasis suppressor 1 functions as a corepressor by enhancing histone deacetylase 1-mediated deacetylation of RelA/p65 and promoting apoptosis. Mol. Cell. Biol. 2006, 26 (23), 8683-96. (10) Meehan, W. J.; Samant, R. S.; Hopper, J. E.; Carrozza, M. J.; Shevde, L. A.; Workman, J. L.; Eckert, K. A.; Verderame, M. F.; Welch, D. R. Breast cancer metastasis suppressor 1 (BRMS1) forms complexes with retinoblastoma-binding protein 1 (RBP1) and the mSin3 histone deacetylase complex and represses transcription. J. Biol. Chem. 2004, 279 (2), 1562-9. (11) DeWald, D. B.; Torabinejad, J.; Samant, R. S.; Johnston, D.; Erin, N.; Shope, J. C.; Xie, Y.; Welch, D. R. Metastasis suppression by breast cancer metastasis suppressor 1 involves reduction of phosphoinositide signaling in MDA-MB-435 breast carcinoma cells. Cancer Res. 2005, 65 (3), 713-7. (12) Cicek, M.; Fukuyama, R.; Welch, D. R.; Sizemore, N.; Casey, G. Breast cancer metastasis suppressor 1 inhibits gene expression by targeting nuclear factor-kappaB activity. Cancer Res. 2005, 65 (9), 3586-95. (13) Van den Bergh, G.; Arckens, L. Fluorescent two-dimensional difference gel electrophoresis unveils the potential of gel-based proteomics. Curr. Opin. Biotechnol. 2004, 15 (1), 38-43. (14) Zhou, G.; Li, H.; DeCamp, D.; Chen, S.; Shu, H.; Gong, Y.; Flaig, M.; Gillespie, J. W.; Hu, N.; Taylor, P. R.; Emmert-Buck, M. R.; Liotta, L. A.; Petricoin, E. F., III; Zhao, Y. 2D differential in-gel electrophoresis for the identification of esophageal scans cell cancer-specific protein markers. Mol. Cell. Proteomics 2002, 1 (2), 117-24. (15) Tonge, R.; Shaw, J.; Middleton, B.; Rowlinson, R.; Rayner, S.; Young, J.; Pognan, F.; Hawkins, E.; Currie, I.; Davison, M. Validation and development of fluorescence two-dimensional differential gel electrophoresis proteomics technology. Proteomics 2001, 1 (3), 377-96. (16) Blanco-Aparicio, C.; Pequeno, B.; Moneo, V.; Romero, L.; Leal, J. F.; Velasco, J.; Fominaya, J.; Carnero, A. Inhibition of phosphatidylinositol-3-kinase synergizes with gemcitabine in low-passage tumor cell lines correlating with Bax translocation to the mitochondria. Anticancer Drugs 2005, 16 (9), 977-87. (17) Yuan, J. S.; Reed, A.; Chen, F.; Stewart, C. N., Jr. Statistical analysis of real-time PCR data. BMC Bioinf. 2006, 7, 85. (18) Livak, K. J.; Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001, 25 (4), 402-8. (19) Shevchenko, A.; Wilm, M.; Vorm, O.; Mann, M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 1996, 68 (5), 850-8. (20) Alfonso, P.; Rivera, J.; Hernaez, B.; Alonso, C.; Escribano, J. M. Identification of cellular proteins modified in response to African swine fever virus infection by proteomics. Proteomics 2004, 4 (7), 2037-46. (21) Peri, S.; Navarro, J. D.; Amanchy, R.; Kristiansen, T. Z.; Jonnalagadda, C. K.; Surendranath, V.; Niranjan, V.; Muthusamy, B.; Gandhi, T. K.; Gronborg, M.; Ibarrola, N.; Deshpande, N.; Shanker, K.; Shivashankar, H. N.; Rashmi, B. P.; Ramya, M. A.; Zhao, Z.;

Journal of Proteome Research • Vol. 6, No. 10, 2007 4017

research articles

(22)

(23)

(24)

(25)

(26)

(27) (28)

(29)

(30) (31)

(32)

(33)

(34) (35)

(36)

4018

Chandrika, K. N.; Padma, N.; Harsha, H. C.; Yatish, A. J.; Kavitha, M. P.; Menezes, M.; Choudhury, D. R.; Suresh, S.; Ghosh, N.; Saravana, R.; Chandran, S.; Krishna, S.; Joy, M.; Anand, S. K.; Madavan, V.; Joseph, A.; Wong, G. W.; Schiemann, W. P.; Constantinescu, S. N.; Huang, L.; Khosravi-Far, R.; Steen, H.; Tewari, M.; Ghaffari, S.; Blobe, G. C.; Dang, C. V.; Garcia, J. G.; Pevsner, J.; Jensen, O. N.; Roepstorff, P.; Deshpande, K. S.; Chinnaiyan, A. M.; Hamosh, A.; Chakravarti, A.; Pandey, A. Development of human protein reference database as an initial platform for approaching systems biology in humans. Genome Res. 2003, 13 (10), 2363-71. Strobl, S.; Fernandez-Catalan, C.; Braun, M.; Huber, R.; Masumoto, H.; Nakagawa, K.; Irie, A.; Sorimachi, H.; Bourenkow, G.; Bartunik, H.; Suzuki, K.; Bode, W. The crystal structure of calciumfree human m-calpain suggests an electrostatic switch mechanism for activation by calcium. Proc. Natl. Acad. Sci. U.S.A. 2000, 97 (2), 588-92. Somiari, R. I.; Sullivan, A.; Russell, S.; Somiari, S.; Hu, H.; Jordan, R.; George, A.; Katenhusen, R.; Buchowiecka, A.; Arciero, C.; Brzeski, H.; Hooke, J.; Shriver, C. High-throughput proteomic analysis of human infiltrating ductal carcinoma of the breast. Proteomics 2003, 3 (10), 1863-73. Garrido, C.; Brunet, M.; Didelot, C.; Zermati, Y.; Schmitt, E.; Kroemer, G. Heat shock proteins 27 and 70: anti-apoptotic proteins with tumorigenic properties. Cell Cycle 2006, 5 (22), 2592-601. Li, D. Q.; Wang, L.; Fei, F.; Hou, Y. F.; Luo, J. M.; Zeng, R.; Wu, J.; Lu, J. S.; Di, G. H.; Ou, Z. L.; Xia, Q. C.; Shen, Z. Z.; Shao, Z. M. Identification of breast cancer metastasis-associated proteins in an isogenic tumor metastasis model using two-dimensional gel electrophoresis and liquid chromatography-ion trap-mass spectrometry. Proteomics 2006, 6 (11), 3352-68. Wu, W.; Tang, X.; Hu, W.; Lotan, R.; Hong, W. K.; Mao, L. Identification and validation of metastasis-associated proteins in head and neck cancer cell lines by two-dimensional electrophoresis and mass spectrometry. Clin. Exp. Metastasis 2002, 19 (4), 319-26. Bamburg, J. R.; Wiggan, O. P. ADF/cofilin and actin dynamics in disease. Trends Cell Biol. 2002, 12 (12), 598-605. Arber, S.; Barbayannis, F. A.; Hanser, H.; Schneider, C.; Stanyon, C. A.; Bernard, O.; Caroni, P. Regulation of actin dynamics through phosphorylation of cofilin by LIM-kinase. Nature 1998, 393 (6687), 805-9. Liaudet-Coopman, E.; Beaujouin, M.; Derocq, D.; Garcia, M.; Glondu-Lassis, M.; Laurent-Matha, V.; Prebois, C.; Rochefort, H.; Vignon, F. Cathepsin D: newly discovered functions of a longstanding aspartic protease in cancer and apoptosis. Cancer Lett. 2006, 237 (2), 167-79. Shi, Y.; Massague, J. Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell 2003, 113 (6), 685-700. Zhang, Z.; Yamashita, H.; Toyama, T.; Yamamoto, Y.; Kawasoe, T.; Iwase, H. Reduced expression of the breast cancer metastasis suppressor 1 mRNA is correlated with poor progress in breast cancer. Clin. Cancer Res. 2006, 12 (21), 6410-4. Ohta, S.; Lai, E. W.; Pang, A. L.; Brouwers, F. M.; Chan, W. Y.; Eisenhofer, G.; de Krijger, R.; Ksinantova, L.; Breza, J.; Blazicek, P.; Kvetnansky, R.; Wesley, R. A.; Pacak, K. Downregulation of metastasis suppressor genes in malignant pheochromocytoma. Int. J. Cancer 2005, 114 (1), 139-43. Toshima, J.; Toshima, J. Y.; Amano, T.; Yang, N.; Narumiya, S.; Mizuno, K. Cofilin phosphorylation by protein kinase testicular protein kinase 1 and its role in integrin-mediated actin reorganization and focal adhesion formation. Mol. Biol. Cell 2001, 12 (4), 1131-45. Tzivion, G.; Gupta, V. S.; Kaplun, L.; Balan, V. 14-3-3 proteins as potential oncogenes. Semin. Cancer Biol. 2006, 16 (3), 203-13. Toshima, J. Y.; Toshima, J.; Watanabe, T.; Mizuno, K. Binding of 14-3-3beta regulates the kinase activity and subcellular localization of testicular protein kinase 1. J. Biol. Chem. 2001, 276 (46), 43471-81. Hehlgans, S.; Haase, M.; Cordes, N. Signalling via integrins: implications for cell survival and anticancer strategies. Biochim. Biophys. Acta 2007, 1775 (1), 163-80.

Journal of Proteome Research • Vol. 6, No. 10, 2007

Rivera et al. (37) Foletta, V. C.; Lim, M. A.; Soosairajah, J.; Kelly, A. P.; Stanley, E. G.; Shannon, M.; He, W.; Das, S.; Massague, J.; Bernard, O. Direct signaling by the BMP type II receptor via the cytoskeletal regulator LIMK1. J. Cell Biol. 2003, 162 (6), 1089-98. (38) Samant, R. S.; Seraj, M. J.; Saunders, M. M.; Sakamaki, T. S.; Shevde, L. A.; Harms, J. F.; Leonard, T. O.; Goldberg, S. F.; Budgeon, L.; Meehan, W. J.; Winter, C. R.; Christensen, N. D.; Verderame, M. F.; Donahue, H. J.; Welch, D. R. Analysis of mechanisms underlying BRMS1 suppression of metastasis. Clin. Exp. Metastasis 2000, 18 (8), 683-93. (39) Rothhammer, T.; Poser, I.; Soncin, F.; Bataille, F.; Moser, M.; Bosserhoff, A. K. Bone morphogenic proteins are overexpressed in malignant melanoma and promote cell invasion and migration. Cancer Res. 2005, 65 (2), 448-56. (40) Vetvicka, V.; Vetvickova, J.; Benes, P. Role of enzymatically inactive procathepsin D in lung cancer. Anticancer Res. 2004, 24 (5A), 2739-43. (41) Cox, G.; O’Byrne, K. J. Matrix metalloproteinases and cancer. Anticancer Res. 2001, 21 (6B), 4207-19. (42) Leto, G.; Tumminello, F. M.; Crescimanno, M.; Flandina, C.; Gebbia, N. Cathepsin D expression levels in nongynecological solid tumors: clinical and therapeutic implications. Clin. Exp. Metastasis 2004, 21 (2), 91-106. (43) Gerke, V.; Creutz, C. E.; Moss, S. E. Annexins: linking Ca2+ signalling to membrane dynamics. Nat. Rev. Mol. Cell Biol. 2005, 6 (6), 449-61. (44) Vishwanatha, J. K.; Chiang, Y.; Kumble, K. D.; Hollingsworth, M. A.; Pour, P. M. Enhanced expression of annexin II in human pancreatic carcinoma cells and primary pancreatic cancers. Carcinogenesis 1993, 14 (12), 2575-9. (45) Yeatman, T. J.; Updyke, T. V.; Kaetzel, M. A.; Dedman, J. R.; Nicolson, G. L. Expression of annexins on the surfaces of nonmetastatic and metastatic human and rodent tumor cells. Clin. Exp. Metastasis 1993, 11 (1), 37-44. (46) Takamatsu, K.; Noguchi, T. Hippocalcin: a calcium-binding protein of the EF-hand superfamily dominantly expressed in the hippocampus. Neurosci. Res. 1993, 17 (4), 291-5. (47) Maki, M.; Kitaura, Y.; Satoh, H.; Ohkouchi, S.; Shibata, H. Structures, functions and molecular evolution of the penta-EFhand Ca2+-binding proteins. Biochim. Biophys. Acta 2002, 1600 (1-2), 51-60. (48) Honore, B.; Vorum, H. The CREC family, a novel family of multiple EF-hand, low-affinity Ca(2+)-binding proteins localised to the secretory pathway of mammalian cells. FEBS Lett. 2000, 466 (1), 11-8. (49) Yabe, D.; Nakamura, T.; Kanazawa, N.; Tashiro, K.; Honjo, T. Calumenin, a Ca2+-binding protein retained in the endoplasmic reticulum with a novel carboxyl-terminal sequence, HDEF. J. Biol. Chem. 1997, 272 (29), 18232-9. (50) Sinha, P.; Kohl, S.; Fischer, J.; Hutter, G.; Kern, M.; Kottgen, E.; Dietel, M.; Lage, H.; Schnolzer, M.; Schadendorf, D. Identification of novel proteins associated with the development of chemoresistance in malignant melanoma using two-dimensional electrophoresis. Electrophoresis 2000, 21 (14), 3048-57. (51) Zhou, Y.; Xu, Y.; Tan, Y.; Qi, J.; Xiao, Y.; Yang, C.; Zhu, Z.; Xiong, D. Sorcin, an important gene associated with multidrugresistance in human leukemia cells. Leuk. Res. 2006, 30 (4), 46976. (52) Kamada, M.; So, A.; Muramaki, M.; Rocchi, P.; Beraldi, E.; Gleave, M. Hsp27 knockdown using nucleotide-based therapies inhibit tumor growth and enhance chemotherapy in human bladder cancer cells. Mol. Cancer Ther. 2007, 6 (1), 299-308. (53) Wachsberger, P. R.; Landry, J.; Storck, C.; Davis, K.; O’Hara, M. D.; Owen, C. S.; Leeper, D. B.; Coss, R. A. Mammalian cells adapted to growth at pH 6.7 have elevated HSP27 levels and are resistant to cisplatin. Int. J. Hyperthermia 1997, 13 (3), 251-5; discussion 257-9. (54) Vargas-Roig, L. M.; Gago, F. E.; Tello, O.; Aznar, J. C.; Ciocca, D. R. Heat shock protein expression and drug resistance in breast cancer patients treated with induction chemotherapy. Int. J. Cancer 1998, 79 (5), 468-75.

PR0703167