Proteomic analysis of pancreatic endocrine tumor cell lines treated with the histone deacetylase...

RESEARCH ARTICLE

Proteomic analysis of pancreatic endocrine tumor cell

lines treated with the histone deacetylase inhibitor

trichostatin A

Daniela Cecconi1, 2*, Massimo Donadelli3*, Sara Rinalducci4, Lello Zolla4,Maria Teresa Scupoli5, Aldo Scarpa2, Marta Palmieri3 and Pier Giorgio Righetti6**

1 Department of Science and Technologies, Section of Biochemical Methodologies, University of Verona,Verona, Italy

2 Department of Pathology, Section of Anatomical Pathology, University of Verona, Verona, Italy3 Department of Morphological-Biomedical Sciences, Section of Biochemistry, University of Verona, Verona, Italy4 Department of Environmental Sciences, University of Tuscia, Viterbo, Italy5 Interdepartmental Laboratory for Medical Research (LURM), University of Verona, Verona, Italy6 Department of Chemistry, Materials and Engineering Chemistry “Giulio Natta”, Polytechnic of Milan, Milan, Italy

Effects of the histone-deacetylases inhibitor trichostatin A (TSA) on the growth of three differenthuman pancreatic endocrine carcinoma cell lines (CM, BON, and QGP-1) have been assessed viadosage-dependent growth inhibition curves. TSA determined strong inhibition of cell growth withsimilar IC50 values for the different cell lines: 80.5 nM (CM), 61.6 nM (BON), and 86 nM (QGP-1),by arresting the cell cycle in G2/M phase and inducing apoptosis. 2DE and nano-RP-HPLC-ESI-MS/MS analysis revealed 34, 33, and 38 unique proteins differentially expressed after TSA treat-ment in the CM, BON, and QGP-1 cell lines, respectively. The most important groups of modu-lated proteins belong to cell proliferation, cell cycle, and apoptosis classes (such as peroxiredoxins 1and 2, the diablo protein, and HSP27). Other proteins pertain to processes such as regulation ofgene expression (nucleophosmin, oncoprotein dek), signal transduction (calcium-calmodulin),chromatin, and cytoskeleton organization (calgizzarin, dynein, and lamin), RNA splicing(nucleolin, HNRPC), and protein folding (HSP70). The present data are in agreement with pre-vious proteomic analyses performed on pancreatic ductal carcinoma cell lines (Cecconi, D. et al.,Electrophoresis 2003; Cecconi, D. et al., J. Proteome Res. 2005) and place histone-deacetylases inhib-itors among the potentially most powerful drugs for the treatment of pancreatic tumors.

Received: October 24, 2006Revised: January 26, 2007

Accepted: February 14, 2007

Keywords:

Epigenetic events / Histone deacetylase inhibitor / Pancreatic endocrine tumor / Tri-chostatin A

1644 Proteomics 2007, 7, 1644–1653

1 Introduction

Pancreatic endocrine tumor (PET) is yet poorly characterizedat the molecular level, although recent studies have aimed toassess the involvement of specific genetic defects in itsdevelopment [1–3]. The most frequent genetic event identi-fied in PET is the inactivation of MEN1 gene, while muta-

Correspondence: Professor Marta Palmieri, Department of Mor-phological-Biomedical Sciences, Section of Biochemistry, Uni-versity of Verona, Verona, ItalyE-mail: [email protected]: 139-045-8027170

Abbreviations: ANOVA, analysis of variance; CALM1, calmodu-lin; GO, gene ontology; HDAC, histone deacetylase; HNRPC,Heterogeneous nuclear ribonucleoprotein C; NCL, nucleolin;NPM, nucleophosmin; PET, pancreatic endocrine tumor; PRDX,peroxiredoxin; SSP, standard spot number; TSA, trichostatin A;VDAC, voltage-dependent anion channel

* Both of the authors contributed equally.** Additional corresponding author: Professor Pier Giorgio

Righetti, E-mail: [email protected]

DOI 10.1002/pmic.200600811

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com

Proteomics 2007, 7, 1644–1653 Biomedicine 1645

tions in KRAS, P53, P16, and SMAD4 genes, typicallyinvolved in pancreatic adenocarcinoma, are uncommon [4,5]. Genome-wide screenings for detection of chromosomalanomalies in PET using microsatellite allelotyping and com-parative genomic hybridization (CGH) have identified sev-eral frequently deleted or amplified chromosomal regionsharboring putative PET-related genes [6–9].

In human cancer, the importance of epigenetic events,including DNA methylation or histone acetylation, isbecoming increasingly evident [10, 11]. A recent investiga-tion on the methylation status of DNA in PET has demon-strated the presence of aberrant methylations in severaltumor suppressor genes, such as RASSF1A (75%), INK4a/p16 (40%), O6-MGMT (40%), RAR-beta (25%), and hMLH1(23%) [12]. These alterations were associated with earlytumor recurrence and reduced survival of PET patients [12].

Although DNA methylation remains the best studiedepigenetic mechanism and appears to be tightly connected tocancer development, also histone acetylation has beendemonstrated to have a prominent role in the transcriptionalinactivation of tumor-suppressor genes [11]. However, untilnow no information is available on epigenetic modificationinvolving histone acetylation in PET.

To identify potential target genes for aberrant acetylationthat may confer growth advantage to neoplastic cells, we an-alyzed global changes in proteomic profiles of three PET celllines (CM, BON, and QGP-1) after treatment with the his-tone deacetylase (HDAC) inhibitor trichostatin A (TSA) and,in parallel, we tested the effect of TSA on cell proliferation,cell cycle distribution, and apoptosis. We show that TSAtreatment strongly inhibits proliferation of PET cells bycausing cell cycle arrest and apoptosis and regulates theexpression of proteins of particular interest. The TSA-mod-ified expression pattern of these proteins well correlates withcell growth inhibition, suggesting that at least some of themmay be involved in the neoplastic transformation of PET.

2 Materials and methods

2.1 Chemicals

TSA was obtained from Sigma-Aldrich, solubilized inDMSO, and stored at 2807C until use.

2.2 Cell culture

The human PETcell lines were: CM (metastatic insulinoma),BON (metastatic carcinoid), and QGP-1 (somatostatinoma).QGP-1 cells were grown in RPMI 1640 supplemented with10% FBS; CM cells were grown in RPMI 1640 supplementedwith 5% FBS; BON cells were grown in DMEM and F12K(ratio 1:1) supplemented with 10% FBS. All the culturemedia were supplemented with 20 mM glutamine and50 mg/mL of gentamicin sulfate (BioWhittaker, Italy). Cellswere incubated at 377C with 5% CO2.

2.3 Cell proliferation assay

Cells were plated in 96-well culture plates (46103 cells/well),treated with TSA for 48 h at the indicated concentrations,and then stained with crystal violet (Sigma). The dye wassolubilized in 1% SDS in PBS and measured photometrically(A595nm) to determine cell growth. Four independent experi-ments were performed for each assay condition.

2.4 Cell cycle analysis

The cell cycle distribution was analyzed by using propidiumiodide (PI)-stained cells. Briefly, 106 cells were washed withPBS, incubated with 0.1% sodium citrate dihydrate, 0.1%Triton X-100, 200 mg/mL RNase A, 50 mg/mL propidiumiodide (Roche Molecular Biochemicals), and analyzed byusing a flow cytometer (Becton Dickinson). The percentageof cells in the various stages of the cell cycle was determinedwith the ModFitLT software program. Three independentexperiments were performed for each assay condition.

2.5 Apoptosis

The percentage of apoptotic cells was evaluated by annexinV-FITC/PI staining using the Becton-Dickinson FACScan.Briefly, 106 cells were stained with annexin V-FITC (BDPharMingen) and 5 mg/mL propidium iodide in bindingbuffer (10 mM HEPES/NaOH (pH 7.4), 140 mM NaOH, and2.5 mM CaCl2) for 10 min at room temperature in the dark.The samples were analyzed by flow cytometry within 1 h todetermine the percentage of cells displaying annexin V1

(early apoptosis) or annexin V1/propidium iodide1 staining(late apoptosis). Three independent experiments were per-formed for each assay condition.

2.6 2-D gel electrophoresis

Protein extraction from cells untreated and treated with0.2 mM TSA for 16 h was performed in a 2-D solubilizing/lysing solution: 7 M urea (Sigma, Sigma-Aldrich, St. Louis,MO, USA), 2 M thiourea (Sigma), 3% CHAPS (Sigma),20 mM Tris (Sigma), 1% pH 3–10 ampholyte (Fluka, Buchs,SG, Switzerland), and one protease inhibitor cocktail tablet(Complete, Mini; Roche, Basel, Switzerland).

The samples were sonicated 5630 s on ice with 1 minrest in between times, and the sonicates were centrifuged for10 min at 10 0006g at 47C to remove the nucleic acids com-plexed with ampholytes. The supernatant cell lysates wereincubated with 5 mM tributyl phosphine and 20 mM acryl-amide for 60 min at room temperature to reduce proteindisulfide bonds and alkylate the cysteine thiolic groups. Thereaction was blocked by the addition of 10 mM DTT (Sigma)and the samples were collected and stored at 2807C.

The protein concentration was evaluated with the DCProtein assay (BioRad, Hercules, CA, USA) based on theLowry method.


1646 D. Cecconi et al. Proteomics 2007, 7, 1644–1653

Seventeen centimeters long, pH 3–10 IPG strips (IPG;BioRad) were rehydrated for 8 h with 450 mL of 2-D solubi-lizing solution (7 M urea, 2 M thiourea, 3% CHAPS, and20 mM Tris) containing 5 mg/mL of total protein from cells.IEF was carried out with a Protean IEF Cell (BioRad), with alow initial voltage and then by applying a voltage gradient upto 10 000 V with a limiting current of 50 mA/strip. The totalproduct time6voltage applied was 70 000 Vh for each stripand the temperature was set at 207C. For the second dimen-sion, the IPGs strips were equilibrated for 26 min by rockingin a solution of 6 M urea, 2% SDS, 20% glycerol, 375 mMTris-HCl, and pH 8.8. The IPG strips were then laid on an 8–18%T gradient SDS-PAGE with 0.8% agarose in Tris/gly-cine/SDS running buffer (192 mM glycine, 0.1% SDS, andTris to pH 8.3). The second dimension was performed in aProtean Plus Dodeca cell (BioRad) with Tris/glycine/SDSrunning buffer. The electrophoresis was conducted withcontinuous cooling and mixing by setting a current of 40 mAfor each gel for 3 min, then 2 mA/gel for 1 h, and 20 mA/geluntil the track dye, bromophenol blue, reached the anodicend of the gels. The protein zones were finally revealed withSYPRO Ruby stain. Gels were incubated in a fixing solutioncontaining 40% ethanol and 10% acetic acid for 30 min fol-lowed by overnight staining in a ready-to-use SYPRO Rubysolution. Destaining was performed in 10% methanol and7% acetic acid for 1 h, followed by a rinse of at least 3 h inpure water.

2.7 Protein pattern differential analysis

2-D gel analysis was performed by PDQuest software(BioRad), version 7.3. Each gel was analyzed for spot detec-tion, background subtraction, and protein spot OD intensityquantification (spot quantity definition). The gel imageshowing the higher number of spots and the best proteinpattern was chosen as a reference template, and spots in astandard gel were then matched across all the gels. Spotquantity values were normalized in each gel dividing the rawquantity of each spot by the total quantity of all the spotsincluded in the standard gel. Gels were divided into twoseparated groups (control and TSA-treated samples) and, foreach protein spot, the average spot quantity value and itsvariance coefficient in each group were determined. A Stu-dent’s t-test was performed in order to compare the twogroups and identify sets of proteins that showed a statisticallysignificant difference with a confidence level of 0.05.

2.8 Statistical analysis

The data set generated from 2-DE analysis was also analyzedby hierarchical cluster analysis using the Ludesi 2-DE Inter-preter software package available at www.ludesi.com. Thequantitative data derived from PDQuest were exported to theLudesi 2-DE Interpreter. The clustering procedure includesthe generation of a distance matrix, which in our case wascalculated by the Pearson correlation distance on the pro-

teins with lowest p-value selected from an ANOVA (analysisof variance) procedure, summarizing all the pairwise simila-rities between expression profiles and generating a dendro-gram (hierarchical tree). The clustering was based on a sig-nificant number of proteins that appear coexpressed withinthe different groups (controls or TSA-treated).

2.9 In-Gel digestion

Spots were carefully cut out from 2-D SYPRO Ruby stainedgels and subjected to in-gel trypsin digestion according toShevchenko et al. [13], with minor modifications. The gelpieces were swollen in a digestion buffer containing 50 mMNH4HCO3 and 12.5 ng/mL of trypsin (modified porcinetrypsin, sequencing grade, Promega, Madison, WI) in an icebath. After 30 min, the supernatant was removed and dis-carded, 20 mL of 50 mM NH4HCO3 were added to the gelpieces and digestion allowed to proceed at 377C overnight.The supernatant containing tryptic peptides was dried byvacuum centrifugation. Prior to mass spectrometric analysis,the peptide mixtures were redissolved in 10 mL of 5% formicacid.

2.10 Peptide sequencing by nano-RP-HPLC-ESI-MS/

MS

Peptide mixtures were separated by using a nanoflow-HPLCsystem (Ultimate; Switchos; Famos; LC Packings, Amster-dam, The Netherlands). A sample volume of 10 mL was loa-ded by the autosampler onto a homemade 2 cm fused silicaprecolumn (75 mm id; 375 mm od; Resprosil C18-AQ, 3 mm(Ammerbuch-Entringen, Germany) at a flow rate of 2 mL/min. Sequential elution of peptides was accomplished usinga flow rate of 200 nL/min and a linear gradient from SolutionA (2% ACN; 0.1% formic acid) to 50% of Solution B (98%ACN; 0.1% formic acid) in 40 min over the precolumn inlinewith a homemade 10–15 cm resolving column (75 mm id;375 mm od; Resprosil C18-AQ, 3 mm (Ammerbuch-Entrin-gen).

Peptides were eluted directly into an IT Esquire 3000plus (Bruker-Daltonik, Germany). Capillary voltage was 1.5–2 kV and a dry gas flow rate of 3 L/min was used with atemperature of 2307C. The scan range used was from 300 to1800 m/z. Protein identification was performed by searchingin the National Center for Biotechnology Information non-redundant (NCBInr) database using the MASCOT program(http://www.matrixscience.com). The following parameterswere adopted for database searches: complete carbamido-methylation of cysteines and partial oxidation of methio-nines, peptide mass tolerance 6 1.2 Da, fragment mass tol-erance 6 0.9 Da, missed cleavages 2. For positive identifica-tion, the score of the result of (2106log(P)) had to be overthe significance threshold level (P,0.05).

Even though high MASCOT scores are obtained withvalues greater than 60, when proteins were identified withonly one peptide a combination of automated database



search and manual interpretation of peptide fragmentationspectra was used to validate protein assignments. In thismanual verification, the mass error, the presence of frag-ment ion series, and the expected prevalence of C-terminuscontaining (Y-type ions) in the high mass range, were alltaken into account. Moreover, replicate measurements haveconfirmed the identity of these protein hits.

2.11 Protein categorization

Gene ontology (GO) lists were downloaded from AMIGO(http://www.godatabase.org/cgi-bin/amigo/go.cgi): eachprotein was classified with respect to its cellular component,biological process, and molecular function using GO anno-tation. When no GO annotation was available, proteins wereannotated manually based on literature searches and closelyrelated homologs.

3 Results

3.1 Inhibition of proliferation of pancreatic endocrine

cancer cell lines by TSA

The effect of the HDAC inhibitor TSA on the proliferation ofthree pancreatic endocrine cancer cell lines was examined bymeasuring cell growth after incubation with increasing con-centrations of TSA for 48 h. All of the cell lines showedstrong inhibition of cell growth after TSA treatment (Fig. 1),with similar IC50 values: 80.5 nM (CM), 61.6 nM (BON), and86 nM (QGP-1).

Figure 1. TSA inhibits growth of three pancreatic endocrine can-cer cell lines. TSA was added at the indicated concentrations for48 h. Cell proliferation was determined using the crystal violetcolorimetric assay as described in Section 2.3. Values are themeans of triplicate wells from three independent experiments.

3.2 Cell cycle arrest of pancreatic endocrine cancer

cell lines by TSA

In order to assess whether cell cycle arrest was involved incell growth inhibition by TSA we measured by flow cy-tometry the percentage of cells in the three phases of the cell

Figure 2. (a) Alterations of cell cycle distribution after TSA treat-ment. CM, BON, and QGP-1 cell lines were treated with 0.2 mMTSA for 48 h. Cell cycle distribution was analyzed by a flowcytometer after DNA staining with propidium iodide. The per-centages were calculated by subtracting the untreated controlvalues from the treated values. Values are the means of three in-dependent experiments. (b) Induction of apoptosis by TSA. CM,BON, and QGP-1 cell lines were exposed to 0.2 mM TSA for 48 h.26105 cells were analyzed by flow cytometry to determine thepercentage of cells displaying annexin V1 (early apoptosis) orannexin V1/propidium iodide1 staining (late apoptosis). A repre-sentative experiment is shown. Three independent experimentsshowed similar results.

cycle after treatment with TSA for 48 h. As shown in Fig. 2a,TSA increased the percentage of cells in G2/M phase by 6,37, and 15% in CM, BON, and QGP-1 cell lines, respectively.

3.3 Apoptotic cell death of pancreatic endocrine

cancer cell lines by TSA

To determine whether TSA treatment resulted in apoptoticcell death, cells were treated for 48 h with TSA and analyzedby flow cytometry with annexin V-FITC and propidium



iodide staining. As shown in Fig. 2b, TSA gave rise to anevident apoptotic effect. The enhancement of the percentageof cells in the early phase of apoptosis (annexin V1) was 17,35, and 6% for CM, BON, and QGP-1 cell lines, respectively.The enhancement of the percentage of cells in the late phaseof apoptosis (annexin V1/propidium iodide1) was 9, 37, and68% for the CM, BON, and QGP-1 cell lines, respectively. Itis worth noting that, after TSA treatment, BON and QGP-1cells showed both early and late phases of apoptosis, whileCM cells showed mainly the apoptotic features of the earlyphase of apoptosis, as demonstrated by the shift of all cells tothe lower right area of the graph. These data suggest thatTSA-induced apoptosis of CM cells is a retarded event withrespect to that observed in BON and QGP-1 cells.

3.4 2-DE protein pattern analysis of pancreatic

endocrine cancer cell lines

We attempted to address the molecular basis of the biologicaleffect of TSA by the study of protein expression profiles ofpancreatic endocrine cancer cells before and after TSA treat-ment. To assess overall differences in 2-DE pattern betweendifferent groups, we applied a hierarchical cluster analysis.In the resulting tree diagram (dendrogram) (Fig. 3), the dif-ferent samples (CM, BON, and QGP-1) formed two clusters:those taken from the control versus those taken after TSAtreatment, on the basis of the protein expression patternswithin each group (p,0.05, ANOVA). The dendrogramsobtained demonstrate a real division into two groups for thethree different cell lines. By PDQuest analysis, we measureddifferential protein expression between control and TSA-treated cells by analyzing the five replicas of 2-D mapsobtained for each group.

A total of 30, 39, and 29 different spots (matched acrossall the replica 2-D maps) were found to be differentiallyexpressed after TSA treatment in CM, BON, and QGP-1 celllines, respectively. In particular, 19 spots were found to bedownregulated and 11 upregulated in the CM cell line; 26downregulated and 13 upregulated in the BON cell line; and19 downregulated and 10 upregulated in the QGP1 cells.

The typical high resolution 2-DE protein patternsobtained from CM, BON, and QGP-1 cell lines are reportedin Fig. 4a (upper three panels); whereas Fig. 4b (lowerpanels) shows representative differentially expressed spots(p,0.05) between control and TSA-treated samples. Spotsselected from the differential analysis of three different celllines (CM, BON, and QGP-1) were subjected to RP-HPLC-ESI-MS/MS analysis for protein identification. The uniquedifferentially expressed proteins identified were 34 for CM,33 for BON, and 38 for QGP-1 cell lines, respectively.

In Table 1 of the Supporting Information, the identity ofthe successfully identified proteins corresponding to up- ordownregulated spots are shown, together with the standardspot number (SSP), the identification parameters, and theindication of their GO annotation (cellular component, bio-logical process, and molecular function). Not all the differ-

Figure 3. Dendrograms obtained by hierarchical cluster analysisdemonstrating that control and TSA-treated cell lines form twoseparate clusters for each of the three cell lines and each of thefive replica maps. Spots for hierarchical clustering analysis wereselected by ANOVA (p,0.05). Each column is a single proteinexpression across each sample. Light grey, downregulated; darkgrey, upregulated.

entially expressed spots were identified because of their rela-tive low concentrations.

The overlap between the identified differentially ex-pressed proteins of the three cell lines is shown in Fig. 5. Thediagram obtained demonstrates a poor overlap; nevertheless,the identified proteins are involved in similar biological pro-cesses and so they belong to the same functional categories.

Figure 6 shows as a pie chart the distribution of theidentified proteins catalogued according to the biologicalprocesses in which they are involved.



Figure 4. (a) Example of SYPRORuby stained 2-D gels of proteinsextracted from CM, BON, and QGP-1cell lines, respectively. (b) PDQuest(version 7.3) output showing arepresentative differentially expres-sed spots (p,0.05) between controland TSA-treated samples. Spot no.3203, PRDX2, expressed in lowerlevels in TSA-treated CM cell line(n = 5) than in control (n = 5) cells;spot no. 2507, NPM1, expressed inlower levels in TSA-treated BON cellline (n = 5) than in control (n = 5)cells; spot no. 1002, calcium-CALM1,expressed in lower levels in TSA-treated QGP-1 cell line (n = 5) than incontrol (n = 5). For each spot, anenlarged region of the respective 2-DE map is shown which contains thereferred spot (highlighted by anarrow) and the corresponding spotvolume. Each spot was identified byMS/MS analysis.

Figure 5. Venn Diagram show-ing the overlapping between thedifferentially expressed proteinsof the three cell lines identifiedby MS/MS analysis. The identi-fied proteins are indicated bytheir HUGO gene name.

4 Discussion

In this study, we report that TSA, a strong HDAC inhibitor,determined cell cycle arrest, and apoptosis of three pancrea-tic endocrine carcinoma cell lines. For analyzing the molec-ular mechanisms involved in cell growth inhibition and the

potential target genes for aberrant acetylation, we performeda proteomic analysis of these three PET cell lines after treat-ment with TSA. To the best of our knowledge, this is the firstfunctional classification of differentially expressed proteinsfollowing an HDAC inhibitor treatment in pancreatic endo-crine carcinoma cell lines.



Figure 6. Distribution of the identified proteins according to thebiological process in which they are involved. Assignments weremade on the basis of information provided by GO lists down-loaded from AMIGO (http://www.godatabase.org/cgi-bin/amigo/go.cgi).

Our data support the concept that TSA does not haveglobal effect on protein expression, but rather alters theexpression of a relatively low number of proteins [14, 15].Among those with known or deduced cellular function, thedifferentially expressed proteins were found to be involved ina wide variety of cellular processes, including cell prolifera-tion, regulation of gene expression, signaling, chromatin andcytoskeleton organization, RNA splicing, protein folding,and electron and mitochondrial transport.

Here, we discuss briefly the most interesting differen-tially expressed proteins after TSA treatment related to cellgrowth inhibition and neoplastic transformation of PET.

4.1 Cell proliferation, cell cycle, and apoptosis

The antitumoral effect of the TSA could be explained by reg-ulation of expression of proteins related to cell proliferation(such as calcyclin; cytokeratin 19; diablo; enhancer of rudi-mentary homolog; GST; HSP27; microtubule-associatedprotein RP/EB family, member 1; peroxiredoxin (PRDX) 1and 2, porin; septin 8, and thioredoxin). Among these pro-teins some are of particular interest and are discussed here.PRDX1 was upregulated 3- and 2.7-fold by TSA in CM andBON cell lines, respectively. PRDX1 is an antioxidant en-zyme with cell growth suppression activity. Recently char-acterized families of thiol proteins, such as PRDXs, mayprove to have a fundamental role in cancer biology, in cellularredox state by modulation of intracellular signaling cascadesthat apply hydrogen peroxide as a second messenger mole-cule, and in regulation of cell proliferation [16]. It has beendemonstrated that PRDX1 is significantly activated by theHDAC inhibitor FK228 and chromatin immunoprecipita-tion analysis revealed that FK228 induced the accumulationof acetylated histones H3 and H4 in PRDX1 promoter,

including the Sp1-binding site [17]. Our data are in agree-ment with the observation that Prdx1 activation by modulat-ing acetylation of histones resulted in tumor cell growthinhibition. We previously demonstrated that TSA treatmentinduced cell death and PRDX1 expression also in pancreaticadenocarcinoma cell lines [18, 19].

PRDX2 was downregulated 117-fold by TSA in CM cellline. PRDX2 has an antiapoptotic function protecting cellsagainst oxidants, radiation, and chemotherapeutic agents[20]. The strong inhibition of expression of PRDX2 correlateswith the observed cell growth inhibition by TSA.

Moreover, we found that TSA treatment strongly down-regulated HSP27 (by a factor of 95-fold) in the CM cell line.HSPs are known to protect cells against various aggressionsand to assist in the correct folding of nascent proteins. Inparticular HSP27 is known to protect cells against apoptosis[21].

We reported also that TSA treatment induces upregula-tion of diablo (DIABLO) by 6.3-fold in QGP-1 cell line. Diablo(direct inhibitor of apoptosis-binding protein with a low pI)is a mitochondrial protein released into the cytosol duringthe progression of apoptosis [22]. This protein promotesapoptosis by neutralizing the antiapoptotic effect of the pro-teins inhibiting apoptosis (IAPs) [22]. It has been demon-strated that many human cancers do not express this protein,suggesting that the lack of diablo expression in cancer cellsmay inhibit apoptosis, thereby promoting their survival [23].It has also been proved that overexpression of DIABLO geneenhances apoptosis-inducing effects of several chemother-apeutic treatments [24–27]. Moreover, small-moleculeagents, resembling diablo in function, were shown topotentiate cytotoxicity of chemotherapy in different malig-nancies. Our results demonstrated that the antitumoraleffect of TSA treatment correlates with the upregulation ofthe proapoptotic diablo.

Porin (voltage-dependent anion channel (VDAC1)) wasdownregulated 2.9-fold in BON cell line. The VDAC1 is amajor mitochondrial outer-membrane transporter and hasan important role in energy production by controlling me-tabolite traffic [28]. The role of VDAC1 in regulating celldeath was previously investigated by silencing VDAC1 geneexpression using an shRNA-expressing vector. Decreasedexpression of VDAC1 strongly slows cell proliferation whilenormal growth was restored upon expression of VDAC1 in atetracycline-regulated manner [28]. Cells with low levels ofVDAC1 showed 4-fold-lower ATP-synthesis capacity andcontained low ATP and ADP levels, with a strong correlationbetween ATP levels and cell growth [26]. Our result demon-strated that the downregulation of VDAC1 correlates withinhibition of cell proliferation by TSA.

4.2 Regulation of gene expression

The action of TSA in contrasting tumor cell growth could beexplained by the regulation of proteins involved in geneexpression (such as calreticulin; electron transfer flavopro-



tein; nucleophosmin 1 (NPM1); oncoprotein dek; PRDX3;prohibitin 2, and ras-related protein rab-11B).

Of particular interest is NPM1, which was stronglydownregulated (150-fold) in the CM cell line and 6.4-fold inthe BON cell line after TSA treatment. NPM1 is a multi-functional protein frequently overexpressed in actively pro-liferating cells including tumor and stem cells [29]. It hasbeen demonstrated that overexpression of NPM1 inhibitsapoptosis, while knockdown of NPM1 induces cell death [30].This protein acts as a cellular p53 negative regulator to pro-tect cells from apoptosis [29]. Overexpression of the mutantNPMDeltaC, which lacks the p53-interacting domain, fails toconfer cellular resistance to stress-induced apoptosis, sug-gesting that NPM1 protects cells from apoptotic cell deaththrough a mechanism involving p53 [29]. The proapoptoticeffect of NPM1 downregulation is in agreement with theantitumoral effect of TSA treatment suggested by our data.Moreover, we previously demonstrated that TSA treatmentinduced cell death and NPM1 downregulation in pancreaticadenocarcinoma cell lines [18].

Oncoprotein dek (DEK) was downregulated 2.7-fold inthe BON cell line after TSA treatment. The DEK proto-onco-gene is a nucleic acid binding protein involved in carcino-genesis, autoimmune disease, and viral infection [31]. It hasbeen reported that cell death in response to DEK depletionwas accompanied by increased protein stability and tran-scriptional activity of the p53 tumor suppressor and con-sequent upregulation of known p53-target genes such asP21(CIP) and BAX [30]. The inhibition of expression ofoncoprotein DEK suggested by our data correlates with thereported antitumoral effect of TSA.

PRDX3 was downregulated 163- and 12.4-fold by TSA inthe CM cell line. PRDX3 is a mitochondrial member of theantioxidant family of thioredoxin peroxidases that usesmitochondrial thioredoxin 2 as a source of reducing equiva-lents to scavenge hydrogen peroxide (H2O2) [32]. It has beendemonstrated that depletion of PRDX3 resulted in increasedintracellular levels of H2O2 and sensitized cells to inductionof apoptosis by staurosporine (an anticancer drug) or TNF-alpha [33]. In PRDX3-depleted cells, rates of mitochondrialmembrane potential collapsed, cytochrome c was released,and caspases were activated; these effects were reversed byectopic expression of PRDX3 [33]. Our data are in agreementwith the observation that PRDX3 depletion resulted in tumorcell growth inhibition.

4.3 Signal transduction

The importance of TSA as a cancer treatment is highlightedby the regulation of expression of proteins related to signaltransduction (such as 14-3-3 protein epsilon and zeta iso-forms; 90 kDa HSP; calcium-calmodulin (CALM1), andchloride intracellular channel 1 p64).

Of particular interest is CALM1 which was down-regulated 7.2-fold by TSA in QGP-1 cell line. CALM1 is anubiquitous, calcium-binding protein that is responsible for

many of the intracellular actions of calcium. Because of theimportance of Ca21 in progression through the cell cycle,CALM1 also plays a critical role in the regulation of cell pro-liferation [34, 35]. It has been reported that diseases char-acterized by pathological, unregulated cell growth, such ascancer, are associated with elevated levels of CALM1 [36, 37].In addition, CALM1 antagonists have been shown to inhibittumor cell invasion in vitro [38] and metastasis in vivo [39],suggesting that the CALM1 antagonists are promisingchemotherapeutic agents for malignancies. Moreover,CALM1 inhibition with either KN-93 or specific interferingRNA (siRNA) caused cell growth inhibition in breast cancercell lines [40]. The CALM1 depletion that we reported corre-lates with the observed cell growth inhibition by TSA.

4.4 Chromatin and cytoskeleton organization

The antitumoral effect of the TSA could be explained byenhanced regulation of proteins related to cytoskeleton organ-ization (such as annexin II; beta-actin; cytokeratin 8, 18, and19; dynein light chain 2A; gamma-actin; lamin A/C; profilin I;tropomyosin alpha-4 chain, and tubulin beta) and regulationof proteins involved in chromatin organization (such as cal-gizzarin; heterochromatin-like protein 1; histone 1 H2AJ,H2B, 1 H4i, H3, and H4/o; parathymosin and SET protein).Among these proteins some were of particular interest.

For example, calgizzarin (S100A11) was downregulated3.5-fold in QGP1 cell line. Calgizzarin is a Ca21-binding pro-tein of the S100 family that has been implicated in the reg-ulation of a variety of cellular processes. Several diseases,including cancers, are related to the abnormal expression ofS100 proteins [41]. It has been demonstrated that suppressionof S100A11 by small interfering RNA (siRNA) led to apopto-sis, and that the overexpression of S100A11 inhibited apopto-sis in human uterine smooth muscle tumor cells [41]. Theproapoptotic effect of calgizzarin downregulation is in agree-ment with the TSA antitumoral effect suggested by our data.

Moreover, we found that TSA treatment increased theexpression of SETprotein (SET) of 4.4-fold in QGP-1 cell line.SET protein is a cyclin B-interacting protein that recently hasbeen shown to associate with p21Cip1 [42]. It has been demon-strated that overexpression of SETprotein blocks the cell cycleat the G(2)/M transition in COS and colon cancer cells bymodulating cyclin B-cyclin-dependent kinase 1 (B-CDK1) ac-tivity [42] and this inhibitory capacity is additive to that ofp21Cip1. SET is also a potent inhibitor of protein phosphatase2A, which is involved in different cellular functions, includingcell proliferation [43]. Our data suggest that SETupregulationplays a role in tumor cell growth inhibition.

4.5 RNA splicing

The ability of TSA to induce cell growth arrest could beexplained also by regulation of proteins related to RNAsplicing (such as heterogeneous nuclear ribonucleoprotein2H9B, A2/B1, C and H1, nucleolin (NCL); splicing factor ar-



ginine/serine-rich 1, 3, and 7; TAR DNA binding protein andU6 snRNA-associated Sm-like protein).

NCL was downregulated 7.4-fold in CM cell line. NCL isan abundant, nonribosomal protein of the nucleolusinvolved in the organization of the nuclear chromatin, rDNAtranscription, packaging of the pre-RNA, ribosome assembly,nucleocytoplasmic transport, cytokinesis, nucleogenesis,and apoptosis [44]. This protein is expressed at high levels inmalignant cells [44]. It has been demonstrated that UV irra-diation or anticancer treatment (with camptothecin) of lym-phoma cell lines induced apoptosis and caused a significantdecrease in the levels of NCL [44]. It has also been reportedthat NCL downregulation promotes p53 expression; in factNCL binds to the 50 untranslated region (UTR) of p53 mRNAand controls p53 translation and induction after DNA dam-age [45]. The NCL decrease shown by our data correlates withthe antitumoral effect of TSA.

Heterogeneous nuclear ribonucleoprotein C (HNRPC)was downregulated 150-fold in CM cells, 811.5- and 3.1-fold inthe BON cell lines, respectively. HNRNP C is a nuclearrestricted pre-mRNA-binding protein and functions in vivo topackage nascent transcripts and nucleate assembly of the 40 SHNRNP complex [46]. It has been demonstrated thatHNRNP C belongs to a critical set of protein substrates thatare specifically cleaved by proteases during apoptosis inducedby a variety of stimuli, including ionizing radiation, etoposide,and ceramide [47]. No cleavage was observed in cells that areresistant to apoptosis induced by ionizing radiation.

4.6 Protein folding

The antitumoral effect of the TSA could be explained by regula-tion of expression of proteins related to protein folding (such ascyclophilin B; ER protein 29; heat shock 70 kDa protein 9B(HSPA9B); PEST-containing nuclear protein; TNF receptor-associated protein 1, and tumor rejection antigen-1 gp96).

For example, HSPA9B, also known as mortalin-2, wasdownregulated 2.1-fold in CM cell line. This protein isinvolved in cell cycle regulation with important roles in cel-lular senescence and immortalization pathways. It is over-expressed in many human tumors [48] and its upregulationcontributes significantly to tumorigenesis [49]. It is known tobind to and inactivate p53 and its apoptotic functions and toinfluence the ras-raf-MAPK pathway. It has also been reportedthat reduction in mortalin level by its antisense expressionpromotes growth arrest in lung fibroblast immortalized cells[50] and induces apoptosis in human gastric cancer cells [51].Our data are in agreement with the observation that mortalin-2 depletion resulted in tumor cell growth inhibition.

5 Conclusion

Previously, we demonstrated that TSA treatment was able tostrongly inhibit proliferation of pancreatic ductal carcinomacells and to sensitize them to treatment with a panel of

chemotherapeutic agents (such as gemcitabine, 5-fluoro-uracil, CPT11, oxaliplatin, and gefitinib) [52]. Here, wedemonstrate that TSA is able to act also on cell lines of pan-creatic endocrine carcinomas derived from metastatic pan-creatic insulinoma, carcinoid, and somatostatinoma. Thedosage dependent growth inhibition curves follow essen-tially identical profiles suggesting that TSA acts in a similarmanner on all three cell lines. Additionally, the present dataon up- and downregulation of some of the principal proteinsgoverning cell proliferation and apoptosis are in good agree-ment with previous data obtained by our group on pancreaticductal carcinoma [18, 19]. It is noteworthy that TSA is able toregulate the expression of some proteins (such as NPM1,oncoprotein dek, and NCL) involved in the p53 stabilizationat both post-transcriptional and post-translational levels,suggesting a consistent role of p53, which is rarely mutatedin PET [53–55], and p53-target genes on cell growth inhibi-tion by TSA. Our study demonstrates that TSA is an efficientagent for treatment of pancreatic endocrine carcinoma andsuggests that epigenetic modifications of the p53 pathwaycould be involved in carcinogenesis of the endocrine pan-creas.

We thank Dr. Elisa Dalla Pozza for technical support and forthe contribution in the analysis of the cellular function of the dif-ferentially expressed proteins. This study was supported by grantsfrom AIRC (to PGR and AS), PRIN 2005 and 2006 (MIUR,Rome), by Fondazione Cassa di Risparmio di Verona (Bando2004 to MP and 2005 to AS), European Community grant FP6program, Fondazione Giorgio Zanotto per il pancreas, ConsorzioInteruniversitario Biotecnologie (to LZ).

6 References

[1] Oberg, K., Eriksson, B., Best Practice & Res. Clin. Gastro-enterol. 2005, 19, 753–781.

[2] Solcia, E., Klöppel, G., Sobin, L.Histological Typing of Endo-crine Tumors. W.H.O. Classification of Endocrine Tumors.,Springer-Verlag, Berlin, Heidelberg, New York 2000.

[3] House, M. G., Schulick, R. D., Curr. Opin. Oncol. 2006, 18, 23–29.

[4] Moore, P. S., Missiaglia, E., Antonello, D., Zamo, A. et al.,Genes Chromosom. Cancer 2001, 32, 177–181.

[5] Moore, P. S., Orlandini, S., Zamboni, G., Capelli, P. et al., Br.J. Cancer 2001, 84, 253–262.

[6] Chung, D. C., Brown, S. B., Graeme-Cook, F., Tillotson, L. G.et al., Cancer Res. 1998, 58, 3706–3711.

[7] Rigaud, G., Missiaglia, E., Moore, P. S., Zamboni, G. et al.,Cancer Res. 2001, 61, 285–292.

[8] Missiaglia, E., Moore, P. S., Williamson, J., Lemoine, N. R. etal., Int. J. Cancer 2002, 98, 532–538.

[9] Speel, E. J., Scheidweiler, A. F., Zhao, J., Matter, C. et al.,Cancer Res. 2001, 61, 5186–5192.

[10] Baylin, S. B., Ohm, J. E., Nat. Rev. Cancer. 2006, 6, 107–116.



[11] Liu, T., Kuljaca, S., Tee, A., Marshall, G. M., Cancer Treat Rev.2006, 32, 157–165.

[12] Dammann, R., Schagdarsurengin, U., Liu, L., Otto, N. et al.,Oncogene. 2003, 22, 3806–3812.

[13] Shevchenko, A., Wilm, M., Vorm, O., Mann, M., Anal. Chem.1996, 68, 850–858.

[14] Van Lint, C., Gene Expr. 1996, 5, 245–253.

[15] Moore, P. S., Barbi, S., Donadelli, M., Costanzo, C. et al.,Biochim. Biophys. Acta. 2004, 1693, 167–176.

[16] Kinnula, V. L., Paakko, P., Soini, Y., FEBS Lett. 2004, 569, 1–6.

[17] Hoshino, I., Matsubara, H., Hanari, N., Mori, M. et al., Clin.Cancer Res. 2005, 11, 7945–7952.

[18] Cecconi, D., Scarpa, A., Donadelli, M., Palmieri, M. et al.,Electrophoresis 2003, 24, 1871–1878.

[19] Cecconi, D., Donadelli, M., Scarpa, A., Milli, A. et al., J. Pro-teome Res. 2005, 4, 1909–1916.

[20] Immenschuh, S., Baumgart-Vogt, E., Antioxid Redox Signal2005, 7, 768–777.

[21] Eickhoff B, Germeroth L, Stahl C, Kohler G et al., J. Biol.Chem. 2000, 381, 1127–1132.

[22] Wrzesien-Kus, A., Smolewski, P., Sobczak-Pluta, A., Wierz-bowska, A. et al., Apoptosis 2004, 9, 705–715.

[23] Yoo, N. J., Kim, H. S., Kim, S. Y., Park, W. S. et al., APMIS2003, 111, 382–388.

[24] Zheng, L. D., Tong, Q. S., Wang, L., Liu, J. et al., World J.Gastroenterol. 2005, 11, 79–83.

[25] Zheng, L. D., Tong, Q. S., Tao, K. X., Wang, L. et al., Ai Zheng2004, 23, 361–366.

[26] Zheng, L. D., Xiong, Z. F., Zhu, J. W., Wang, Z. H., Chin. Med.J. (Engl.) 2005, 118, 226–230.

[27] Mizukawa, K., Kawamura, A., Sasayama, T., Tanaka, K. et al.,J. Neurooncol. 2006, 77, 247–255.

[28] Abu-Hamad, S., Sivan, S., Shoshan-Barmatz, V., Proc. Natl.Acad. Sci. USA 2006, 103, 5787–5792.

[29] Li, J., Zhang, X., Sejas, D. P., Pang, Q., Leuk Res. 2005, 29,1415–1423.

[30] Ye, K., Cancer Biol. Ther. 2005, 4, 918–923.

[31] Wise-Draper, T. M., Allen, H. V., Jones, E. E., Habash, K. B. etal., Mol. Cell. Biol. 2006, 26, 7506–7519.

[32] Nonn, L., Berggren, M., Powis, G., Mol. Cancer Res. 2003, 1,682–689.

[33] Chang, T. S., Cho, C. S., Park, S., Yu, S. et al., J. Biol. Chem.2004, 279, 41975–41984.

[34] Rasmussen, C. D., Lu, K. P., Means, R. L., Means, A. R., J.Physiol.Paris 1992, 86, 83–88.

[35] Bachs, O., Agell, N., Carafoli, E., Biochim. Biophys. Acta1992, 1113, 259–270.

[36] Wei, J. W., Morris, H. P., Hickie, R. A., Cancer Res. 1982, 42,2571–2574.

[37] Hait, W. N., Lazo, J. S., J. Clin. Oncol. 1986, 4, 994–1012.

[38] Dewhurst, L. O., Gee, J. W., Rennie, I. G., MacNeil, S., Br. J.Cancer 1997, 75, 860–868.

[39] Ito, H., Wang, J. Z., Shimura, K., Anticancer Res. 1991, 11,249–252.

[40] Rodriguez-Mora, O. G., LaHair, M. M., McCubrey, J. A.,Franklin, R. A., Cancer Res. 2005, 65, 5408–5416.

[41] Kanamori, T., Takakura, K., Mandai, M., Kariya, M. et al., Mol.Hum. Reprod. 2004, 10, 735–742.

[42] Canela, N., Rodriguez-Vilarrupla, A., Estanyol, J. M., Diaz, C.et al., J. Biol. Chem. 2003, 278, 1158–1164.

[43] Li, M., Makkinje, A., Damuni, Z., J. Biol. Chem. 1996, 271,11059–11062.

[44] Mi, Y., Thomas, S. D., Xu, X., Casson, L. K. et al., J. Biol.Chem. 2003, 278, 8572–8579.

[45] Takagi, M., Absalon, M. J., McLure, K. G., Kastan, M. B., Cell2005, 123, 49–63.

[46] Rodriguez-Mora, O. G., LaHair, M. M., McCubrey, J. A.,Franklin, R. A., Oncogene 2002, 21, 580–583.

[47] Waterhouse, N., Kumar, S., Song, Q., Strike, P. et al., J. Biol.Chem. 1996, 271, 29335–29341.

[48] Dundas, S. R., Lawrie, L. C., Rooney, P. H., Murray, G. I., J.Pathol. 2005, 205, 74–81.

[49] Wadhwa, R., Takano, S., Kaur, K., Deocaris, C. C. et al., Int. J.Cancer. 2006, 118, 2973–2980.

[50] Wadhwa, R., Takano, S., Taira, K., Kaul, S. C., J. Gene Med.2004, 6, 439–444.

[51] Zhao, Z. G., Shen, W. L., World J. Gastroenterol. 2005, 11,73–78.

[52] Piacentini, P., Donadelli, M., Costanzo, C., Moore, PS. et al.,Virchows Arch. 2006, 448, 797–804.

[53] Wang, D. G., Clin. Endocrinol. 1999, 51, 1–9.

[54] Rindi, G., Candusso, M. E., Solcia, E., Ital. J. Gastroenterol.Hepatol. 1999, 31, 135–138.

[55] Yashiro, T., Fulton, N., Hara, H., Yasuda, K. et al., Surgery1993, 114, 758–763.


Proteomic analysis of pancreatic endocrine tumor cell lines treated with the histone deacetylase...

Documents

Transcript of Proteomic analysis of pancreatic endocrine tumor cell lines treated with the histone deacetylase...