Proteomics 2015, 15, 3075–3086

3075

DOI 10.1002/pmic.201400489

RESEARCH ARTICLE

Proteomic analysis of secreted proteins by human bronchial epithelial cells in response to cadmium toxicity De-Ju Chen1∗ , Yan-Ming Xu1∗,∗∗ , Wei Zheng1 , Dong-Yang Huang2 , Wing-Yan Wong3 , William Chi-Shing Tai3,4∗∗ , Yong-Yeon Cho5 and Andy T. Y. Lau1∗ 1

Laboratory of Cancer Biology and Epigenetics, Department of Cell Biology and Genetics, Shantou University Medical College, Shantou, Guangdong, P. R. China 2 The Key Laboratory of Molecular Biology for High Cancer Incidence Coastal Chaoshan Area, Department of Cell Biology and Genetics, Shantou University Medical College, Shantou, Guangdong, P. R. China 3 Centre for Cancer and Inflammation Research, School of Chinese Medicine, Hong Kong Baptist University, Kowloon Tong, Hong Kong, P. R. China 4 Institute of Integrated Bioinfomedicine & Translational Science, Hong Kong Baptist University Shenzhen Research Institute and Continuing Education, Shenzhen, P. R. China 5 College of Pharmacy, The Catholic University of Korea, Bucheon, Korea

For years, many studies have been conducted to investigate the intracellular response of cells challenged with toxic metal(s), yet, the corresponding secretome responses, especially in human lung cells, are largely unexplored. Here, we provide a secretome analysis of human bronchial epithelial cells (BEAS-2B) treated with cadmium chloride (CdCl2 ), with the aim of identifying secreted proteins in response to Cd toxicity. Proteins from control and spent media were separated by two-dimensional electrophoresis and visualized by silver staining. Differentiallysecreted proteins were identified by MALDI-TOF-MS analysis and database searching. We characterized, for the first time, the extracellular proteome changes of BEAS-2B dosed with Cd. Our results unveiled that Cd treatment led to the marked upregulation of molecular chaperones, antioxidant enzymes, enzymes associated with glutathione metabolic process, proteins involved in cellular energy metabolism, as well as tumor-suppressors. Pretreatment of cells with the thiol antioxidant glutathione before Cd treatment effectively abrogated the secretion of these proteins and prevented cell death. Taken together, our results demonstrate that Cd causes oxidative stress-induced cytotoxicity; and the differentially-secreted protein signatures could be considered as targets for potential use as extracellular biomarkers upon Cd exposure.

Received: October 21, 2014 Revised: March 20, 2015 Accepted: April 30, 2015

Keywords: BEAS-2B / Biomedicine / Cadmium / Human lung cells / Mass spectrometry / Secretomics



Additional supporting information may be found in the online version of this article at the publisher’s web-site

Correspondence: Professor Andy T. Y. Lau, Laboratory of Cancer Biology and Epigenetics, Department of Cell Biology and Genetics, Shantou University Medical College, 22 Xinling Road, Shantou, Guangdong 515041, P. R. China E-mail: [email protected] Fax: +86-754-8890-0437 Abbreviations: ALDOA, fructose-bisphosphate aldolase A; ANXA, annexin; CCT5, T-complex protein 1 subunit epsilon; CFL1, cofilin-1; EIF5A, eukaryotic translation initiation factor 5A1; ENO1, alpha-enolase; FTH1, ferritin heavy chain; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GST, glutathione

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

S-transferase; HSP, heat-shock protein; LGALS1, galectin-1; MDH, malate dehydrogenase; NME, nucleoside diphosphate kinase; PRDX, peroxiredoxin; RAN, GTP-binding nuclear protein RAN; RPSA, 40S ribosomal protein SA; RUVBL2, RuvB-like 2; SERPINB5, Serpin B5; SERPINE1, plasminogen activator inhibitor 1; SFN, 14-3-3 protein sigma; TAGLN2, transgelin-2; TXN, thioredoxin; UCHL1, ubiquitin carboxyl-terminal hydrolase isozyme L1 ∗ These

authors contributed equally to this work. corresponding authors: Professor Yan-Ming Xu, E-mail: [email protected]; Dr. William Chi-Shing Tai, E-mail: [email protected]

∗∗ Additional

www.proteomics-journal.com

3076

1

D.-J. Chen et al.

Introduction

Lung cancer was the most common form of cancer contributing 13% of the total number of new cases diagnosed in 2012 worldwide (http://www.wcrf.org/int/cancer-factsfigures/worldwide-data) [1]. There is an obvious correlation of lung cancer development with inhalation of polluted air and consumption of tobacco products [2, 3]. Cd is one of the important carcinogenic heavy metals that is present in polluted air and cigarette smokes [2, 4]. In 1993, Cd and its derivatives have been classified as group 1 carcinogens by the International Agency for Research on Cancer (IARC) [5]. Studies of the last two decades have been established that the effects of Cd are multifaceted [6–8]; it has been suggested that Cd exerts cytotoxic or carcinogenic effects through an alteration in expressions of immediate early genes or tumor suppressor genes, oxidative stress, promotion of cell proliferation, and acquisition of apoptotic resistance [9–17]. In addition, epigenetic changes associated with alterations of DNA methylation levels and histone modifications have also been reported recently [14, 15, 18, 19]. In the past, many studies have been conducted to investigate the intracellular response of cells challenged with toxic metal(s) [20–26], yet, the corresponding secretome responses, especially in lung cells, are largely unexplored. As the secretomes contain a potential rich source of biomarkers [27–30], and the secretome response would be more reminiscent to human situation if human lung cells are to be used. Therefore, in this study, we resolved to use the normal human bronchial epithelial cells (BEAS-2B) and sought to conduct a secretome analysis on this cell line treated with cadmium chloride (CdCl2 ), with the aim of identifying secreted proteins in response to Cd toxicity. Proteins from spent media of sham-exposed or Cd-exposed BEAS-2B cells were separated by two-dimensional electrophoresis (2DE) and visualized by silver staining. Differentially-secreted proteins were identified by MALDI-TOF-MS analysis and database searching. To this end, we characterized for the first time, the extracellular proteome changes of BEAS-2B dosed with Cd. Our proteomic analysis identified 36 unique types of differentiallysecreted proteins between the spent media of control and Cdtreated cells. Our results unveiled that Cd treatment led to the marked upregulation of molecular chaperones, antioxidant enzymes, enzymes associated with glutathione metabolic process, proteins involved in cellular energy metabolism, as well as tumor-suppressors. Pretreatment with the thiol antioxidant glutathione before Cd treatment effectively abrogated the secretion of these proteins and prevented cell death. The implications of these findings are that Cd causes oxidative stress-induced cytotoxicity in human lung cells; and the differentially-secreted protein signatures could be considered as targets for potential use as extracellular biomarkers upon Cd exposure.

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Proteomics 2015, 15, 3075–3086

2

Materials and methods

2.1 Materials CdCl2 was purchased from Sigma Aldrich (St. Louis, MO). PlusOne 2D Clean-Up kit and Silver Staining kit were purchased from GE Healthcare (Uppsala, Sweden). All other general chemicals were purchased from GE Healthcare and Sigma Aldrich. Antibodies used for Western blot were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), GeneTex (Irvine, CA), and Sigma Aldrich, with the following dilutions: EIF5A (sc-390202; Santa Cruz), 1:500; GSTP1 (GTX112953; GeneTex), 1:1000; PRDX1 (GTX113793; GeneTex), 1:2000; HSP90B1 (GTX103232; GeneTex), 1:1000; GAPDH (GTX100118; GeneTex), 1:5000; and ␤-actin (A5441; Sigma Aldrich), 1:5000.

2.2 Cell culture The BEAS-2B cell line was purchased from the American Type Culture Collection (ATCC) (Rockville, MD). BEAS-2B cells were isolated from normal human bronchial epithelium obtained from autopsy of a non-cancerous individual. Cells were routinely grown in LHC-9 medium (Gibco, Grand Island, NY) at 37⬚C in an atmosphere of 5% CO2 /95% air as recommended by ATCC. LHC-9 is a defined, serum-free medium which is prepared by mixing LHC basal medium with growth factors, cytokines, and supplements; and has been described previously [31].

2.3 Cd treatment Cells were grown to 75% confluence in 60 mm cell culture dishes and then rinsed with Hank’s balanced salt solution (HBSS) before they were either sham-exposed or treated with different concentrations of CdCl2 (2, 20, and 30 ␮M) in LHC-9 medium for 12, 24, and 36 h. Cells were pretreated with GSH for 1 h before the addition of Cd. Cell viability was measured by naphthol blue black (NBB) staining assay as described previously [21].

2.4 Protein sample preparation and conditions of two-dimensional PAGE For isolation of proteins in spent medium, equivalent volume of cell-free supernatant (normalized by cell number) was collected by centrifugation and filtration, then mixed with acetone at 1:4 ratio (v/v) and centrifuged at 12 000 g for 10 min at 4⬚C. Precipitated proteins were redissolved in rehydration buffer (8 mol/l urea, 2% CHAPS). For 2D PAGE analysis, protein samples were further purified by using the PlusOne

www.proteomics-journal.com

3077

Proteomics 2015, 15, 3075–3086

2D Clean-Up kit. 2DE was done with Ettan IPGphor 3 IEF system (GE Healthcare) and Hoefer SE 600 electrophoresis units in accordance with the manufacturer’s instructions. Briefly, equal amounts of cleaned-up protein samples were mixed with rehydration buffer containing 8 M urea, 2% CHAPS, 0.28% (w/v) dithiothreitol and 0.5% (v/v) Pharmalyte in a volume of 250 ␮l (GE Healthcare). Rehydration was performed with precasted 13 cm pI 3–11 NL IPG strips for 12 h under a low voltage of 30 V, 50 ␮A per strip at 20⬚C. IEF was run following a stepwise voltage increase procedure: 500 and 1000 V for 1 h each and then 64 000 Vh. After IEF, the strips were subjected to two-step equilibration of 15 min each in the equilibration buffer [6 M urea, 30% (v/v) glycerol, 2% (w/v) SDS and 50 mM Tris/HCl, pH 8.8] with 1% dithiothreitol for the first step and 2.5% (w/v) iodoacetamide for the second step. The strips were then loaded onto the second dimension 12.5% SDS-polyacrylamide gel, positioned into the gel with an overlay of warm agarose sealing solution [0.5% (w/v) agarose and 0.002% (w/v) Bromophenol Blue in 1×SDS running buffer] and resolved at 15 mA/gel for 15 min. Separation continued at 30 mA/gel until the dye front nearly reached the bottom. All gels were visualized by silver staining using the PlusOne Silver Staining kit in accordance with the manufacturer.

2.5 Western blot analysis For 1DE Western blot analysis, equal amounts of proteins were fractionated on appropriate percentage of SDS– polyacrylamide gel and transferred onto PVDF membranes. For 2DE Western blot analysis, protein samples were resolved by 2D PAGE with conditions mentioned above and then transferred onto PVDF membranes. After the transfer, the membranes were blocked with 5% non-fat dry milk in PBS containing 0.05% Tween 20 and probed with various primary antibodies. After incubation with secondary antibodies, immunoblots were visualized with the enhanced chemiluminescence detection kit (GE Healthcare).

drated with acetonitrile, and then rehydrated in trypsin solution (10 ␮g/ml in 25 mM NH4 HCO3 ) at 37⬚C overnight.

2.8 MALDI-TOF-MS analysis and protein identification Tryptic peptide mass spectra were obtained using a Bruker Autoflex III MALDI-TOF/TOF mass spectrometer equipped with a 200 Hz N2 laser operating at 337 nm. Data were acquired in the positive ion reflector mode over a mass range of 800–4000 m/z using Bruker calibration mixture as an external standard. Bruker calibration mixture consists of the following peptides (monoisotopic mass of the singly protonated ion is given in parenthesis in Da); bradykinin (757.3992), angiotensin II (1046.5420), angiotensin I (1296.6853), substance P (1347.7361), bombesin (1619.8230), renin substrate (1758.9326), ACTH clip 1–17 (2093.0868), ACTH clip 18–39 (2465.1990), somatostatin 28 (3147.4714). ␣-Cyano-4hydroxycinnamic acid (␣-CHCA) was used as matrix for thin layer sample preparation. Briefly, 40 mg/ml ␣-CHCA in 98% acetone was spread onto 384-spot 600 ␮m AnchorChip. One microliter of peptide samples was deposited onto the matrix spot and allowed to dry. The dried sample spot was washed with 1 ␮l of 10 mM ammonium citrate for 10 s. Keratin contamination peaks, matrix ion peaks, and trypsin ion peaks were excluded from spectra. Typically 400 shots were accumulated per spectrum in MS mode and 2000 shots in MS/MS mode. The spectra were processed using the Bruker FlexAnalysis 3.0 and BioTools 3.1 software tools. Protein identification was performed using Mascot 2.2.04 (http://www.matrixscience.com). Peptide masses were matched with the theoretical peptides of all proteins in the database using the Mascot search program. The following parameters were used for database searches: monoisotopic mass accuracy < 100 ppm, missed cleavages 1, carbamidomethylation of cysteine as fixed modification, oxidation of methionine as variable modifications. In MS/MS mode, the fragment ion mass accuracy was set to < 0.5 Da.

2.6 Image acquisition and analysis

2.9 Bioinformatic analyses

The stained gels were scanned using an ImageScanner III (GE Healthcare) operated by the LabScan 6.0 software. Image analysis was carried out by using the PDQuest software, version 8.0 (Bio-Rad). Three biological replicates were applied to each protein sample. Only spots changed in expression for more than 2-fold or spots that either appeared/disappeared were selected for analysis with MS.

Differentially-secreted proteins were classified from Gene Ontology (GO) database (http://www.geneontology. org/). Proteins were further analyzed by three different online softwares for secretory protein prediction (http://www. cbs.dtu.dk/services/SignalP/, http://www.cbs.dtu.dk/ services/TMHMM/, and http://www.cbs.dtu.dk/services/Sec retomeP/); as well as by two databases that provide the experimental evidence of secretory proteins validated in human plasma or exosome (http://www.plasma proteomedatabase.org and http://www.exocarta.org/). Furthermore, differentially-secreted proteins were listed and the data were analyzed through the use of QIAGEN’s R R Pathway Analysis (IPA , QIAGEN Redwood Ingenuity

2.7 Tryptic in-gel digestion Protein spots were excised and transferred into siliconized 1.5 ml Eppendorf tubes. Gel chips were destained, dehy C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.proteomics-journal.com

3078

D.-J. Chen et al.

Figure 1. Experimental workflow in this study.

City, www.qiagen.com/ingenuity), to highlight the relationships among candidate proteins using networks and canonical pathways. Ingenuity Knowledge Base, the core R , provides a wide range of high-quality detailed behind IPA information, including direct and indirect protein interaction networks, thus aiding the generation of hypotheses for a comprehensive analysis of a large number of data.

2.10 Statistical analysis Statistical analysis was performed by using two-tailed Student’s t-test, and P < 0.05 was considered significant. Data are expressed as the mean ± SD of triplicate samples, and the reproducibility was confirmed in three separate experiments.

3

Results

3.1 Comparison of the proteome profiles between the spent medium of BEAS-2B cells treated with different dosages of Cd From our previous study, we determined the cytotoxicity of CdCl2 in BEAS-2B cells [32]. In this study, we therefore wish to examine the secretome changes of BEAS-2B cells dosed with low to high concentrations of CdCl2 [2 (non-cytotoxic), 20 (sub-lethal), and 30 ␮M (lethal)] (Fig. 1). After treatment, purified spent medium proteins were isolated and applied to 2DE and proteins visualized by silver staining. 2D gels were run three times for each sample (three individual sets of cell culture experiment) to minimize gel-to-gel variation. Spot volume comparison was made between the samples with the PDQuest software. Overall, BEAS-2B cells treated with 30 ␮M CdCl2 showed the greatest changes of proteome profiles. Figure 2A shows the representative 2D gel image of spent medium proteins from BEAS-2B cells treated with 30 ␮M of Cd for 36 h. A total of 38 protein spots were found to be significantly-induced by Cd treatment, all these protein spots were cut out and subjected to trypsin digestion, MS analysis and database searching. Table 1, Supplementary Table 1, and Supporting Information summarize the identified proteins and their alterations between control and 30 ␮M  C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Proteomics 2015, 15, 3075–3086

Cd-treated cells. The upregulated proteins included molecular chaperones (such as CCT5 and HSP90B1); antioxidant enzymes (such as PRDX1, PRDX2, and TXN); enzymes associated with glutathione metabolic process (such as GSTO1 and GSTP1); proteins involved in cellular energy metabolism (such as ENO1 and GAPDH); as well as tumor-suppressors (such as EIF5A, SERPINB5, and UCHL1). In parallel, we also examined the spent medium proteome profiles of BEAS-2B cells treated with 2, 20, and 30 ␮M Cd at 12, 24, and 36 h time points, from the results, the induction kinetics of all these differentially-expressed proteins were revealed (Supplementary Table 2). Figure 2B and C showed the induction kinetics of several secreted proteins (EIF5A, GSTP1, PRDX1, HSP90B1, and GAPDH) upon different dosage of Cd treatment. It can be seen that there were little differences in the secretome upon 2 ␮M Cd treatment as compared with control, while 20 ␮M Cd treatment altered the secretome profiles in much the same trend as those of 30 ␮M Cd treatment. However, the induction of these proteins by 20 ␮M Cd treatment was somewhat weaker when compared with 30 ␮M Cd treatment (Fig. 2B and Supplementary Table 2). Nevertheless, the majority of these proteins began to be elevated starting at 20 ␮M Cd treatment and 24 h interval. We also found that half of the identified proteins were normally-secreted at unstressed condition (including TXN, LGALS1, PRDX2, NME1, SFN, TPM4, CAPZA1, LDHB, SCG3, RUVBL2, ENO1, ANXA1, MDH1, ALDOA, ANXA2, GAPDH, CFL1, and HSP90B1) as they were present at detectable levels in spent medium of control cells (Table 1). At the same time, we also checked on cell lysates to see whether these proteins were also similarly induced inside the cells upon Cd treatment. Therefore, the expression levels of several selected proteins (EIF5A, GSTP1, PRDX1, HSP90B1, and GAPDH) in basal and Cd-treated cells were further confirmed by Western blot analysis, it can be seen that the majority of these proteins showed a similar trend of induction as their secreted counterparts (Fig. 2D); suggesting that the increased levels of these proteins in spent media upon Cd treatment were likely caused by enhanced cellular secretion, not due to intracellular protein contamination caused by cell lysis. Furthermore, it has been known that some of the proteins we found are known to be multi-spot proteins (e.g. PRDX1). Thus, finding one spot regulated may not indicate the whole protein (all other spots) is regulated or regulated in the same way. Therefore, we performed 2DE blots of spent medium on several of these proteins (PRDX1, GAPDH, and EIF5A), our results showed that indeed PRDX1 and GAPDH exist as multi-spot proteins (Fig. 2E). Based on the Western blot results, it can be seen that both PRDX1 and GAPDH exist as three obvious spots in control media and become 6 spots upon Cd treatment [Please note that spot number 5 and 1 on the 2DE blot of GAPDH correspond to spots 32 and 33, respectively on 2D gel; while spot number 2 on the 2DE blot of PRDX1 corresponds to spot 34 on 2D gel] (Fig. 2E). For GAPDH, all the six spots (including the original three spots www.proteomics-journal.com

Proteomics 2015, 15, 3075–3086

3079

Figure 2. Analyses of spent media from BEAS-2B cells dosed with CdCl2 . (A) A representative gel image of spent medium from BEAS-2B cells dosed with 30 ␮M CdCl2 for 36 h visualized by 2D gel (12.5%) and silver staining. Differentially-secreted proteins were indicated by arrows and identified by MALDI-TOF-MS. (B) Protein expression profile of the six proteins in spent media from BEAS-2B cells with low to high levels (2–30 ␮M) of CdCl2 treatments for 0 to 36 h were assessed by 2DE analyses as shown in montage view. Graphic presentations of the normalized spot volume (%) were also shown on the right. The results are representative of three independent experiments. *, a significant difference (P < 0.05) as compared with control. (C) Time- and dose-dependent effects of CdCl2 on several secreted proteins from BEAS-2B cells as confirmed by Western blot analyses, using antibodies against EIF5A, GSTP1, PRDX1, HSP90B1, and GAPDH. (D) Western blot analyses were performed on cell lysates to detect the expression levels of several selected proteins in control and 30 ␮M CdCl2 -treated BEAS-2B cells. The same blot was stripped and reprobed with the monoclonal ␤-actin antibody to monitor the loading difference. (E) 2DE blots of spent medium on several secreted proteins. Protein samples were resolved by 2D PAGE and then transferred onto polyvinylidene difluoride membranes. Western blot analyses for the detection of PRDX1, GAPDH, and EIF5A were performed, using antibodies against PRDX1, GAPDH, and EIF5A. Please note that spot number 5 and 1 on the 2DE blot of GAPDH correspond to spots 32 and 33, respectively, on 2D gel; while spot number 2 on the 2DE blot of PRDX1 corresponds to spot 34 on 2D gel. The results are representative of three independent experiments.

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.proteomics-journal.com

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

P09211 P15531 P09936

P31947

6. Glutathione S-transferase P

7. Nucleoside diphosphate kinase A 8. Ubiquitin carboxyl-terminal hydrolase isozyme L1

9. 14-3-3 protein sigma

SCG3 RUVBL2 CCT5 ENO1 SERPINE1

Q8WXD2 Q9Y230 P48643 P06733 P05121

LDHB RPSA GLRX3 SERPINB5

P07195 P08865 O76003 P36952

ANXA3

P12429 CAPZA1

CAPZB

P47756

P52907

TPM4 PSME1

P67936 Q06323

SFN

NME1 UCHL1

47.2 (7.01) 45.1 (6.68)

53 (4.94) 51.2 (5.49) 59.7 (5.44)

36.6 (5.71) 32.8 (4.79) 37.4 (5.31) 42.1 (5.72)

32.9 (5.45)

36.4 (5.62)

31.4 (5.36)

28.5 (4.67) 28.7 (5.78)

27.8 (4.68)

17.1 (5.81) 24.8 (5.33)

23.4 (5.43)

21.2 (5.31) 21.9 (5.66)

11.8 (4.82) 14.7 (5.3) 16.8 (5.07)

Mass, kD (pI)

11.6 ± 1.5 ± ± ± ±

3.4 ± 1.3 4.4 ± 1.4 ND ND ND

ND ND

3.1 ± 0.9 3.2 ± 0.9 ND

+2.39 +2.88 N/A N/A N/A

50.2 ± 2.9 10.3 ± 3.1

+3.43 N/A N/A N/A

+3.41

7.4 ± 1.9 9.2 ± 1.7 10.4 ± 3.4

3.5 3.5 1.2 3.2

N/A

10.9 ± 1.0

ND

15.1 12.3 8.6 14.2

N/A

9.3 ± 1.4

ND

+2.33 N/A

57.6 ± 8.4 7.5 ± 1.2

24.7 ± 4.3 ND

+4.44

54.2 ± 9.4

12.2 ± 3.4

N/A

8.5 ± 1.1 +4.76 N/A

N/A +4.74

9.8 ± 3.2 16.2 ± 3.5

29.5 ± 4.4 6.8 ± 0.9

+6.67 +3.77 N/Ac)

42.7 ± 6.4 16.6 ± 2.3 32.2 ± 3.3

Cd

Fold differencea)

6.2 ± 1.4 ND

ND

ND 3.4 ± 1.3

6.4 ± 2.9 4.4 ± 1.9 NDb)

Control

Volume, × 10−2 % (mean ± S.D.)

Cellular metabolic process RNA metabolic process Cell redox homeostasis Extracellular matrix organization/Tumor suppressor Protein metabolism ATP catabolic process ‘de novo’ posttranslational protein folding Glycolysis Extracellular matrix organization

Calcium-dependent phospholipid binding Actin cytoskeleton organization

Cell redox homeostasis Apoptotic process Apoptosis/Protein biosynthesis/Tumor suppressor Cellular iron ion homeostasis Cellular response to oxidative stress Glutathione metabolic process/Response to ROS DNA catabolic process Cell death/Protein deubiquitination/Tumor suppressor Adaptor protein/Signal transduction Cellular component movement Implicated in immunoproteasome assembly Actin cytoskeleton organization

Cellular function

D.-J. Chen et al.

19. Secretogranin-3 20. RuvB-like 2 21. T-complex protein 1 subunit epsilon 22. Alpha-enolase 23. Plasminogen activator inhibitor 1

14. F-actin-capping protein subunit alpha-1 15. L-lactate dehydrogenase B chain 16. 40S ribosomal protein SA 17. Glutaredoxin-3 18. Serpin B5

10. Tropomyosin alpha-4 chain 11. Proteasome activator complex subunit 1 12. F-actin-capping protein subunit beta 13. Annexin A3

FTH1 PRDX2

P02794 P32119 GSTP1

TXN LGALS1 EIF5A

Gene name

P10599 P09382 P63241

UniProtKB accession ID

1. Thioredoxin 2. Galectin-1 3. Eukaryotic translation initiation factor 5A-1 4. Ferritin heavy chain 5. Peroxiredoxin-2

Spot no./identified protein

Table 1. MS-results of secreted proteins that were significantly increased in the spent medium of BEAS-2B treated with 30 ␮M CdCl2 for 36 h as compared with sham-exposed cells

3080 Proteomics 2015, 15, 3075–3086

www.proteomics-journal.com

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

ANXA1 MDH1 GSTO1 RAN ALDOA ANXA2 GAPDH GAPDH PRDX1

P04083 P40925 P78417 P62826 P04075 P07355 P04406 P04406 Q06830 P37802 P23528 P22392 P14625

35. Transgelin-2 36. Cofilin-1 37. Nucleoside diphosphate kinase B 38. Endoplasmin

22.4 (8.41) 18.5 (8.22) 17.3 (8.53) 92.5 (4.76)

22.1 (8.27)

36.1 (8.57)

24.4 (7.01) 39.4 (8.3) 38.6 (7.57) 36.1 (8.57)

27.6 (6.24)

38.7 (6.57) 36.4 (6.91)

47.2 (7.01) 41.4 (6.46)

Mass, kD (pI)

ND 3.6 ± 1.1 ND 3.7 ± 0.3

ND

N/A +3.31 N/A +2.92

1.1 1.5 1.4 1.9

± ± ± ± 4.3 11.9 6.5 10.8

N/A

9.8 ± 1.8

49.7 ± 8.2

4.7 ± 1.1

+10.57

N/A +3.86 +6.95 +8.56

± ± ± ±

21.8 52.5 45.2 46.2

ND 13.6 ± 1.4 6.5 ± 1.0 5.4 ± 1.2

2.1 6.9 11.2 10.4

N/A

13.2 ± 2.1

ND

+6.16 +3.46

39.4 ± 7.1 14.2 ± 2.4

6.4 ± 1.8 4.1 ± 1.8

+3.42 N/A

57.1 ± 12.8 12.7 ± 3.2

Cd

Fold differencea)

16.7 ± 4.2 ND

Control

Volume, × 10−2 % (mean ± S.D.)

a) Average expression level in Cd-treated cells compared to BEAS-2B cells from 3 independent analyses (+, increase). b) ND, non-detectable. c) N/A, not applicable.

TAGLN2 CFL1 NME2 HSP90B1

ENO1 ACAT2

P06733 Q9BWD1

24. Alpha-enolase 25. Acetyl-CoA acetyltransferase, cytosolic 26. Annexin A1 27. Malate dehydrogenase, cytoplasmic 28. Glutathione S-transferase omega-1 29. GTP-binding nuclear protein Ran 30. Fructose-bisphosphate aldolase A 31. Annexin A2 32. Glyceraldehyde-3-phosphate dehydrogenase 33. Glyceraldehyde-3-phosphate dehydrogenase 34. Peroxiredoxin-1

Gene name

UniProtKB accession ID

Spot no./identified protein

Table 1. Continued

Hydrogen peroxide catabolic process Epithelial cell differentiation Actin cytoskeleton organization Nucleotide metabolism Molecular chaperone

Apoptosis/Glycolysis

Calcium/phospholipid-binding Cellular carbohydrate metabolic process Glutathione derivative biosynthetic process Nucleocytoplasmic transport Glycolysis Calcium/phospholipid-binding Apoptosis/Glycolysis

Glycolysis Energy pathways

Cellular function

Proteomics 2015, 15, 3075–3086

3081

www.proteomics-journal.com

3082

D.-J. Chen et al.

Figure 3. Schematic representation of the molecular function and biological process of the identified proteins classified from Gene Ontology (GO) database.

1 to 3) are upregulated as the spot intensity are all increased, indicating that the whole GAPDH species are upregulated. Similarly, for PRDX1, except for spot number 1 which showed diminished spot intensity upon Cd treatment, all the other spots 2 to 6 are upregulated, indicating that the majority of PRDX1 species are upregulated as a whole in the spent media (Fig. 2E). For EIF5A, it appears as one spot on 2DE blot and is elevated in expression upon Cd treatment (Fig. 2E). All these results are in agreement and comparable with the results as they appeared in silver-stained 2D gels.

3.2 Bioinformatic analyses of differentially-secreted proteins from BEAS-2B cells dosed with Cd From GO database, differentially-secreted proteins were classified, as shown in Fig. 3. GO molecular function analysis showed that 41.6% of the proteins possess catalytic activity, followed by binding (22.2%), structural molecule activity (19.4%), enzyme regulator activity (5.6%), translation regulator activity (5.6%), and antioxidant activity (5.6%). GO biological process analysis showed that 40.0% of the proteins are related to metabolic process, followed by cellular process (18.3%) and cellular component organization or biogenesis (8.3%). Other involved proteins were found to be in C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Proteomics 2015, 15, 3075–3086

volved in response to stimulus (6.7%), biological regulation (6.7%), developmental process (6.7%), immune system process (5.0%), localization (5.0%), and multicellular organismal process (3.3%). From the above, we can see that Cd-treated cells secreted a panel of proteins with diverse functions, and may play an important role in antagonizing/detoxifying the harmful effect of Cd. Next, the proteins detected in the spent media were analyzed based on various parameters, such as signal peptide predominance, membrane protein shedding, exocytosis, or exosome delivery. According to the GO Database, 97% (35/36) of the proteins were classified as extracellular and/or membrane-related proteins (Supplementary Table 3). In addition, the bioinformatic softwares, SignalP 4.1 (which characterizes the presence of a signal peptide) [33], SecretomeP 2.0 (which predicts proteins secreted by a non-signal peptide trigger) [34], and TMHMM (which predicts the transmembrane helices in proteins) [35], were used to predict proteins released into the medium through classical secretion pathway, non-classical secretion pathway, and/or membrane protein shedding, respectively. By SignalP 4.1 analysis, it can be seen that out of 36 identified proteins, three of them are likely secreted by signal peptide. From the TMHMM analysis, it can be seen that none of the proteins contain transmembrane domains (suggesting that the proteins are not likely released into the medium by membrane protein shedding). While SecretomeP 2.0 analysis showed that out of 36 identified proteins, 18 (50%) of them are likely released into the medium by a non-signal peptide trigger. Next, the Exocarta database was used to investigate the proportion of the identified proteins that could be secreted by the cells via exosomes [36, 37]. Of the 36 identified proteins, 92% (33/36) of the identified proteins existed in the human exosome database. In addition, among the 36 identified proteins, 100% (36/36) were found in the plasma proteome database [38, 39]. Collectively, these analyses suggest that the majority of the identified proteins (33/36) were released into the medium via non-classical secretion pathway/exosome secretion, and all of them have been validated to be existed in human plasma. For the sake of providing a general picture of the differentially-secreted proteins, the subcellular localization and their interactions were R analysis, as shown in Supplementary generated by the IPA R results showed that the majority of Fig. 1. In general, IPA proteins (26 out of 36) reside in the cytoplasm; while four in the extracellular space, two in the plasma membrane, and four in the nucleus. However, we want to emphasize that this subcellular map is over-simplified as one particular protein is shown in only one subcellular location. From GO database, we know that indeed the majority of the identified proteins can reside in multiple subcellular locations, such as in the nucleus, cytoplasm, and secretory vesicles (Supplementary Table 3). As just mentioned above, 92% (33/36) of our identified proteins existed in the human exosome database and 100% (36/36) existed in the plasma proteome database, further suggesting that these proteins are very likely to be secreted outside the cells, via different mechanisms. Nevertheless, this www.proteomics-journal.com

3083

Proteomics 2015, 15, 3075–3086

subcellular map (Supplementary Fig. 1) gives us a brief picture of the possible subcelluar locations of the differentiallysecreted proteins as well as their inter-relationships. R By Diseases and Functions Annotation from Ingenuity Pathway Analysis, the proteins identified in the spent medium of BEAS-2B cells upon Cd treatment were found mostly related to the categories of cell death and survival. This showed the dynamics of protein expression in cells batR analysis tling for survival from apoptosis. In particular, IPA results indicated that ALDOA, ANXA2, CCT5, CFL1, FTH1, GSTP1, HSP90B1, MDH1, NME2, PRDX1, PRDX2, SERPINE1, SFN, TAGLN2, and TXN, are likely to play a role in promoting cell survival; while EIF5A, ENO1, GAPDH, LGALS1, SERPINB5, and UCHL1, are likely to promote cell death (Supplementary Fig. 2). Moreover, by further annotation analysis, several of the differentially-secreted proteins (including GSTP1, PRDX1, PRDX2, TXN, FTH1, and SERPINB5) are closely-related to free radical scavenging and inhibition of reactive oxygen species (ROS) (Supplementary Fig. 3), suggesting that BEAS-2B cells are undergoing oxidative stress upon Cd exposure.

3.3 Induction of differentially-secreted proteins by Cd is modulated by intracellular GSH in BEAS-2B cells Since intracellular GSH level is vital for the redox homeostasis of cells and it has been shown that GSH is a first line of defense against Cd toxicity [40], we investigated the protective role of intracellular GSH in the induction of differentiallysecreted proteins by Cd in BEAS-2B cells. We use 30 ␮M Cd to treat the cells, as this concentration induced all secreted proteins and produced obvious proteome changes. GSH was added 1 h before the addition of CdCl2 . Pretreatment with 20 mM GSH before Cd treatment effectively inhibited the induction of the majority of secreted proteins as determined by 2D PAGE and Western blot analyses (Fig. 4A, B and Supplemetary Fig. 4).

3.4 Cd-induced cytotoxicity is correlated with oxidative stress in BEAS-2B cells To further confirm that Cd-induced cytotoxicity in BEAS-2B cells is due to oxidative stress, BEAS-2B cells were exposed to 30 ␮M CdCl2 for 36 h in the absence or presence of GSH (20 mM). GSH was added 1 h before the addition of CdCl2 . NBB staining showed that pretreatment with GSH before Cd treatment effectively protected the cells against oxidative stress-induced cytotoxicity by Cd. In the absence of GSH pretreatment, massive cytotoxicity can be observed by Cd treatment (Fig. 5A and B), indicating the important role of oxidative stress in Cd-induced cytotoxicity.  C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Figure 4. Induction of differentially-secreted proteins by Cd is inhibited by antioxidant GSH. (A) BEAS-2B were exposed to CdCl2 (30 ␮M) in the absence or presence of 20 mM GSH (pH adjusted to 7.6). GSH was added 1 h before the addition of CdCl2 . BEAS-2B cells were also treated with GSH alone. After treatment for 36 h, the spent media were collected. The effects of GSH on the protein expression profiles of secreted proteins in basal and Cd-treated cells were assessed by 2DE analyses and shown in montage view (6 were shown here among the 38 spots). The results are representative of three independent experiments. Graphic presentations of the normalized spot volume (%) were also shown below. *, P < 0.05 versus Cd-treated cells only. (B) The corresponding protein expression levels of (A) as determined by Western blot analyses.

4

Discussion

In the past, many proteomic studies have been conducted to investigate the intracellular response of cells challenged with toxic metal(s), such as in yeast [41], zebrafish [42], algae [43], or mammalian cells [20,22–24,26]. As the secretomes contain www.proteomics-journal.com

3084

D.-J. Chen et al.

Figure 5. Cd-induced cytotoxicity is correlated with oxidative stress and is countered by GSH pretreatment. BEAS-2B cells were exposed to 30 ␮M CdCl2 for 36 h in the absence or presence of GSH (20 mM). (A) NBB staining assay for the determination of cell viability. The percentage of viability was plotted as 100% for control (no treatment of Cd). Columns, means of triplicate samples and reproducibility was confirmed in three separate experiments; bars, S.D. *, P < 0.05 versus Cd-treated cells only. (B) The corresponding cell morphology of BEAS-2B cells under light microscope. The results are representative of three independent experiments.

a potential rich source of biomarkers [27–30], therefore, investigating the secretome upon the exposure to a tested agent by proteomic approach will allow simultaneously characterizing a large number of relevant proteins. However, to our knowledge, no proteome analysis of Cd has yet been conducted in the secretome of Cd-exposed lung cells to improve our understanding of the extracellular proteome dynamics in response to Cd exposure. Moreover, the secretome response would be more reminiscent to human situation if human lung cells are to be used. For these reasons, in this study, we resolved to use the normal BEAS-2B cells to examine the secretome response to environmentally-relevant concentrations of Cd. It is documented that Cd alters immediate early gene expressions and activates various signaling pathways [9–11, 44, 45]. In addition, experiments using cDNA microarrays have demonstrated that treatment of cells with low levels of Cd altered their gene expression profile [20]. However, we could not detect any significant changes in the secretome when cells  C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Proteomics 2015, 15, 3075–3086

were treated with 2 ␮M Cd up to 36 h. This was probably because changes in gene expression level do not necessarily correlate to alterations in protein expression level. Most kinase signaling pathways are regulated rapidly by posttranslational modifications, such as phosphorylation of target proteins and effectors. It may be that 2DE was not sensitive enough to pick up minor changes induced by low levels of Cd. In addition, we cannot also rule out the possibility that different cell lines may respond differently to the treatment of Cd. Furthermore, while a low level of Cd exposure has been shown to induce cell transformation, this occurred after long-term exposure [44–46]. It is believed that chronic low levels of Cd exposure to cells induce sustained oxidative stress and may gradually promote the carcinogenesis process over a more prolonged period of time [44, 45]. It is therefore not surprising that our secretomic analyses were unable to detect significant proteome changes in BEAS-2B cells treated with 2 ␮M Cd for a relatively short period of time. On the other hand, we were able to identify major proteins that are involved in the secretome response to higher levels of Cd exposure in BEAS-2B cells. Comparative analysis of proteome profiles between the secretome of sham-exposed and Cd-treated BEAS-2B cells allowed the identification of proteins whose levels were altered upon 20 to 30 ␮M Cd treatment, identifying them as extracellular and primary Cdresponsive proteins against Cd. We show that treatment of BEAS-2B cells with 20 to 30 ␮M Cd produced significant changes in at least 38 areas (36 protein types) on 2D gels. Significantly, we show that the proteome profile with 20 ␮M Cd treatment is similar to that for 30 ␮M Cd. However, the induction of differentially-secreted proteins with 20 ␮M Cd treatment was somewhat less pronounced when compared with 30 ␮M Cd treatment. The stress exerted by 20 ␮M Cd was also less catastrophic when compared with 30 ␮M Cd insults. R Pathway Analysis, the proteins identified in By Ingenuity the spent medium of BEAS-2B cells upon Cd treatment were found mostly related to the categories of cell death and survival. This showed the dynamics of protein expression in cells R analybattling for survival from apoptosis. Moreover, IPA sis results indicated that several of the differentially-secreted proteins (including GSTP1, PRDX1, PRDX2, TXN, FTH1, and SERPINB5) are closely-related to free radical scavenging and inhibition of ROS, suggesting that BEAS-2B cells are undergoing oxidative stress upon Cd exposure. Indeed, our results demonstrated that Cd-induced cytotoxicity in BEAS-2B cells is mainly due to oxidative stress, as pretreatment with thiol antioxidant GSH before Cd treatment effectively abrogated the induction of differentially-secreted proteins and sustained cell viability. Our results are in agreement with others which demonstrated oxidative stress-induced cytotoxicity by Cd treatment in various cell types [12, 13, 47–51]. Although more work is needed to be done in dissecting the functional roles of Cd-responsive proteins in the secretome, we believe that part of these differentially-secreted proteins may actually serve as an extracellular line of defense against the toxic effects exerted by Cd. By doing so, at least part of www.proteomics-journal.com

Proteomics 2015, 15, 3075–3086

the Cd toxicity is able to be handled immediately or detoxified outside the cells, thereby preventing further intracellular damages. Overall, this extracellular line of defense, in conjunction with the intracellular detoxification system inside the cells [52], could fortify together for more efficient protection of cells against the toxic effects of Cd. And we suggest that the differentially-secreted protein signatures in this study could be considered as targets for potential use as extracellular biomarkers upon Cd exposure. In summary, our work documented the first secretome response of BEAS-2B cells challenged with environmentallyrelevant concentrations of Cd. The results clearly showed that human lung cells can adjust the production of secreted proteins in response to Cd exposure. Finally, the work we reported here might also implicate the need for ongoing research on the understanding of the extracellular microenvironment in response to toxic metal (such as Cd, As, Cr, and Pb) exposure in human cells, and hopefully lead to the development of agents that can antagonize the cytotoxic/carcinogenic effects exerted by these toxic metals. This work was supported by National Natural Science Foundation of China Grants 31170785, 81101785, 30870497 and 31271445, Fund for University Talents of Guangdong Province, and Guangdong Natural Science Foundation of China Grant S2012030006289. We would like to thank Yuan Zhou from the University of Hong Kong for technical assistance as well as members of the Lau and Xu laboratory for critical reading of this manuscript. The authors have declared no conflict of interest.

5

References

[1] Ferlay, J., Soerjomataram, I., Ervik, M., Dikshit, R. et al., GLOBOCAN 2012 v1.0, cancer incidence and mortality worldwide. IARC CancerBase 2013, 11. [2] Kampa, M., Castanas, E., Human health effects of air pollution. Environ. Pollut. 2008, 151, 362–367. [3] Gibbons, D. L., Byers, L. A., Kurie, J. M., Smoking, p53 mutation, and lung cancer. Mol. Cancer Res. 2014, 12, 3–13. ¨ [4] Oberdorster, G., Airborne cadmium and carcinogenesis of the respiratory tract. Scand. J. Work Environ. Health 1986, 12, 523–537. [5] International Agency for Research on Cancer, IARC monographs on the evaluation of the carcinogenic risks to humans. Beryllium, cadmium, mercury, and exposures in the glass manufacturing industry. IARC 1993, 58, 119–238. [6] Hartwig A., Mechanisms in cadmium-induced carcinogenicity: recent insights. Biometals 2010, 23, 951–960.

3085 [9] Jin, P., Ringertz, N. R., Cadmium induces transcription of proto-oncogenes c-jun and c-myc in rat L6 myoblasts. J. Biol. Chem. 1990, 265, 14061–14064. [10] Wang, Z., Templeton, D. M., Induction of c-fos protooncogene in mesangial cells by cadmium. J. Biol. Chem. 1998, 273, 73–79. [11] Xu, G., Zhou, G., Jin, T., Zhou, T., et al., Apoptosis and p53 gene expression in male reproductive tissues of cadmium exposed rats. Biometals 1999, 12, 131–139. [12] Yang, C. F., Shen, H. M., Shen, Y., Zhuang, Z. X., Ong, C. N., Cadmium-induced oxidative cellular damage in human fetal lung fibroblasts (MRC-5 cells). Environ. Health Perspect. 1997, 105, 712–716. [13] Almenara, C. C., Broseghini-Filho, G. B., Vescovi, M. V., Angeli, J. K., et al., Chronic cadmium treatment promotes oxidative stress and endothelial damage in isolated rat aorta. PLoS One 2013, 8, e68418. [14] Wang, B., Li, Y., Tan, Y., Miao, X., et al., Low-dose Cd induces hepatic gene hypermethylation, along with the persistent reduction of cell death and increase of cell proliferation in rats and mice. PLoS One 2012, 7, e33853. [15] Yuan, D., Ye, S., Pan, Y., Bao, Y., et al., Long-term cadmium exposure leads to the enhancement of lymphocyte proliferation via down-regulating p16 by DNA hypermethylation. Mutat. Res. 2013, 757, 125–131. [16] Lau, A. T., Zhang, J., Chiu, J. F., Acquired tolerance in cadmium-adapted lung epithelial cells: roles of the c-Jun N-terminal kinase signaling pathway and basal level of metallothionein. Toxicol. Appl. Pharmacol. 2006, 215, 1–8. [17] Lau, A. T., Chiu, J. F., The possible role of cytokeratin 8 in cadmium-induced adaptation and carcinogenesis. Cancer Res. 2007, 67, 2107–2113. [18] Somji, S., Garrett, S. H., Toni, C., Zhou, X. D., et al., Differences in the epigenetic regulation of MT-3 gene expression between parental and Cd+2 or As+3 transformed human urothelial cells. Cancer Cell Int. 2011, 11, 2. [19] Cheng, T. F., Choudhuri, S., Muldoon-Jacobs, K., Epigenetic targets of some toxicologically relevant metals: a review of the literature. J. Appl. Toxicol. 2012, 32, 643–653. [20] Andrew, A. S., Warren, A. J., Barchowsky, A., Temple, K. A., et al., Genomic and proteomic profiling of responses to toxic metals in human lung cells. Environ. Health Perspect. 2003, 111, 825–835. [21] Lau, A. T., Li, M., Xie, R., He, Q. Y., Chiu, J. F., Opposed arsenite-induced signaling pathways promote cell proliferation or apoptosis in cultured lung cells. Carcinogenesis 2004, 25, 21–28. [22] Lau, A. T., Chiu, J. F., Proteomic and biochemical analyses of in vitro carcinogen-induced lung cell transformation: synergism between arsenic and benzo[a]pyrene. Proteomics 2006, 6, 1619–1630.

[7] Hartwig A., Cadmium and cancer. Met. Ions Life Sci. 2013, 11, 491–507.

[23] Lei, T., He, Q. Y., Cai, Z., Zhou, Y., et al., Proteomic analysis of chromium cytotoxicity in cultured rat lung epithelial cells. Proteomics 2008, 8, 2420–2429.

[8] Hartwig, A., Metal interaction with redox regulation: an integrating concept in metal carcinogenesis? Free Radic. Biol. Med. 2013, 55, 63–72.

[24] Prins, J. M., Fu, L., Guo, L., Wang, Y., Cd²+ -induced alteration of the global proteome of human skin fibroblast cells. J. Proteome Res. 2014, 13, 1677–1687.

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.proteomics-journal.com

3086

D.-J. Chen et al.

´ [25] L’Azou, B., Passagne, I., Mounicou, S., Treguer-Delapierre, M., et al., Comparative cytotoxicity of cadmium forms (CdCl2, CdO, CdS micro- and nanoparticles) in renal cells. Toxicol. Res. 2014, 3, 32–41. [26] Galano, E., Arciello, A., Piccoli, R., Monti, D. M., Amoresano, A., A proteomic approach to investigate the effects of cadmium and lead on human primary renal cells. Metallomics 2014, 6, 587–597. [27] Hathout, Y., Approaches to the study of the cell secretome. Expert Rev. Proteomics 2007, 4, 239–248. [28] Lou, X., Xiao, T., Zhao, K., Wang, H., et al., Cathepsin D is secreted from M-BE cells: its potential role as a biomarker of lung cancer. J. Proteome Res. 2007, 6, 1083–1092. [29] Makridakis, M., Vlahou, A., Secretome proteomics for discovery of cancer biomarkers. J. Proteomics 2010, 73, 2291– 2305. [30] Stastna, M., Van Eyk, J. E., Secreted proteins as a fundamental source for biomarker discovery. Proteomics 2012, 12, 722–735. [31] Reddel, R. R., Ke, Y., Gerwin, B. I., McMenamin, M. G., et al., Transformation of human bronchial epithelial cells by infection with SV40 or adenovirus-12 SV40 hybrid virus, or transfection via strontium phosphate coprecipitation with a plasmid containing SV40 early region genes. Cancer Res. 1988, 48, 1904–1909. [32] Chen, D. J., Xu, Y. M., Du, J. Y., Huang, D. Y., Lau, A. T., Cadmium induces cytotoxicity in human bronchial epithelial cells through upregulation of eIF5A1 and NFkappaB. Biochem. Biophys. Res. Commun. 2014, 445, 95– 99. [33] Petersen, T. N., Brunak, S., von Heijne, G., Nielsen, H., SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat. Methods 2011, 8, 785–786. [34] Bendtsen, J. D., Jensen, L. J., Blom, N., Von Heijne, G., Brunak, S., Feature-based prediction of non-classical and leaderless protein secretion. Protein Eng. Des. Sel. 2004, 17, 349–356. ¨ [35] Moller, S., Croning, M. D., Apweiler, R., Evaluation of methods for the prediction of membrane spanning regions. Bioinformatics 2001, 17, 646–653. [36] Mathivanan, S., Simpson, R. J., ExoCarta: a compendium of exosomal proteins and RNA. Proteomics 2009, 9, 4997–5000. [37] Mathivanan, S., Ji, H., Simpson, R. J., Exosomes: extracellular organelles important in intercellular communication. J. Proteomics 2010, 73, 1907–1920. [38] Anderson, N. L., Polanski, M., Pieper, R., Gatlin, T., et al., The human plasma proteome: a nonredundant list developed by combination of four separate sources. Mol. Cell. Proteomics 2004, 3, 311–326. [39] Muthusamy, B., Hanumanthu, G., Suresh, S., Rekha, B., et al., Plasma Proteome Database as a resource for proteomics research. Proteomics 2005, 5, 3531–3536.

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Proteomics 2015, 15, 3075–3086 [40] Singhal, R. K., Anderson, M. E., Meister, A., Glutathione, a first line of defense against cadmium toxicity. FASEB J. 1987, 1, 220–223. [41] Vido, K., Spector, D., Lagniel, G., Lopez, S., et al., A proteome analysis of the cadmium response in Saccharomyces cerevisiae. J. Biol. Chem. 2001, 276, 8469–8474. [42] Zhu, J. Y., Chan, K. M., Mechanism of cadmium-induced cytotoxicity on the ZFL zebrafish liver cell line. Metallomics 2012, 4, 1064–1076. ´ [43] Gillet, S., Decottignies, P., Chardonnet, S., Le Marechal, P., Cadmium response and redoxin targets in Chlamydomonas reinhardtii: a proteomic approach. Photosynth. Res. 2006, 89, 201–211. [44] Son, Y. O., Wang, L., Poyil, P., Budhraja, A., et al., Cadmium induces carcinogenesis in BEAS-2B cells through ROSdependent activation of PI3K/AKT/GSK-3␤/␤-catenin signaling. Toxicol. Appl. Pharmacol. 2012, 264, 153–160. [45] Jing, Y., Liu, L. Z., Jiang, Y., Zhu, Y., et al., Cadmium increases HIF-1 and VEGF expression through ROS, ERK, and AKT signaling pathways and induces malignant transformation of human bronchial epithelial cells. Toxicol. Sci. 2012, 125, 10– 19. [46] Person, R. J., Tokar, E. J., Xu, Y., Orihuela, R., et al., Chronic cadmium exposure in vitro induces cancer cell characteristics in human lung cells. Toxicol. Appl. Pharmacol. 2013, 273, 281–288. [47] Hart, B. A., Lee, C. H., Shukla, G. S., Shukla, A., et al., Characterization of cadmium-induced apoptosis in rat lung epithelial cells: evidence for the participation of oxidant stress. Toxicology 1999, 133, 43–58. [48] Shukla, G. S., Shukla, A., Potts, R. J., Osier, M., et al., Cadmium-mediated oxidative stress in alveolar epithelial cells induces the expression of gamma-glutamylcysteine synthetase catalytic subunit and glutathione S-transferase alpha and pi isoforms: potential role of activator protein-1. Cell Biol. Toxicol. 2000, 16, 347–362. [49] Liu, J., Kadiiska, M. B., Corton, J. C., Qu, W., et al., Acute cadmium exposure induces stress-related gene expression in wild-type and metallothionein-I/II-null mice. Free Radic. Biol. Med. 2002, 32, 525–535. [50] Huang, Y. H., Shih, C. M., Huang, C. J., Lin, C. M., et al., Effects of cadmium on structure and enzymatic activity of Cu,ZnSOD and oxidative status in neural cells. J. Cell. Biochem. 2006, 98, 577–589. [51] Chang, K. C., Hsu, C. C., Liu, S. H., Su, C. C., et al., Cadmium induces apoptosis in pancreatic ␤-cells through a mitochondria-dependent pathway: the role of oxidative stress-mediated c-Jun N-terminal kinase activation. PLoS One 2013, 8, e54374. [52] Lau, A. T., Wang, Y., Chiu, J. F., Reactive oxygen species: current knowledge and applications in cancer research and therapeutic. J. Cell. Biochem. 2008, 104, 657–667.

www.proteomics-journal.com