Annexin-A7 protects normal prostate cells and induces distinct patterns of RB-associated cytotoxicity in androgen-sensitive and -resistant prostate cancer cells

Annexin-A7 protects normal prostate cells and induces distinct patterns of

RB-associated cytotoxicity in androgen-sensitive and -resistant prostate

cancer cells

Yelizaveta Torosyan, Olga Simakova, Shanmugam Naga, Katerina Mezhevaya, Ximena Leighton, Juan Diaz,Wei Huang, Harvey Pollard and Meera Srivastava*

Department of Anatomy, Physiology and Genetics, Institute for Molecular Medicine,Uniformed Services University of Health Sciences (USUHS) School of Medicine, Bethesda, MD

The tumor suppressor role of annexin-A7 (ANXA7) was previ-ously demonstrated by cancer susceptibility in Anxa7(1/2)-miceand by ANXA7 loss in human cancers, especially in hormone-re-sistant prostate tumors. To gain mechanistic insights into ANXA7tumor suppression, we undertook an in vitro study in whichwe compared wild-type (WT)-ANXA7 and dominant-negative(DN)-ANXA7 effects to a conventional tumor suppressor p53 inprostate cancer cells with different androgen sensitivity. Unlikep53 (which caused cell growth arrest and apoptosis to a noticeableextent in benign PrEC), WT-ANXA7 demonstrated profoundcytotoxicity in androgen-sensitive LNCaP as well as in the andro-gen-resistant DU145 and PC3 prostate cancer cells, but not inPrEC. In androgen-sensitive LNCaP, WT-ANXA7 decreased low-molecular-weight (LMW) AR protein forms and maintainedhigher retinoblastoma 1 (RB1)/phospho-RB1 ratio. In contrast,DN-ANXA7 (which lacks phosphatidylserine liposome aggregationproperties) increased LMW-AR forms and hyperphosphorylatedRB1 that was consistent with the lack of DN-ANXA7 cytotoxicity.According to the microarray-based Ingenuity Pathways Analysis,a major WT-ANXA7 effect in androgen-sensitive LNCaP consti-tuted of upregulation of the RB1-binding transcription factorE2F1 along with its downstream proapoptotic targets such asASK1 and ASPP2. These results suggested a reversal of the RB-dependent repression of the proapoptotic E2F-mediated transcrip-tion. However, DN-ANXA7 increased RB1/2 (but not E2F1)expression and induced the proliferation-promoting ERK5,thereby maintaining the RB-dependent repression of E2F-medi-ated apoptosis in LNcaP. On the other hand, in androgen-resistantcells, WT-ANXA7 tumor suppressor effects involved PTEN andNFkB pathways. Thus, ANXA7 revived the RB-associated cellsurvival control and overcame androgen resistance and dysfunc-tional status of major tumor suppressors commonly mutated inprostate cancer.Published 2009 UICC. This article is a US Government work and, assuch, is in the public domain in the United States of America.

Key words: annexin-A7 vs. p53; RB-E2F; AR; prostate cancer;Ingenuity Pathways Analysis

Annexin-VII (ANXA7 or synexin) is a ubiquitously expressedmember of the multifunctional Ca/phospholipid-binding annexinfamily. In addition to Ca21-activated GTPase activity as well asinvolvement in membrane fusion and exocytosis, ANXA7 demon-strated tumor suppressor function in murine model1 and in humanprostate and breast cancers.2,3 Our large-scale study on ANXA7protein expression in different cancers4 identified a drasticANXA7 loss in the hormone-refractory prostate cancers as wellas an abundant ANXA7 presence in the adrenal gland, a majorsource of sex hormone precursors. Although tumor suppressormechanisms of ANXA7 are not yet elucidated, these data sug-gested that ANXA7, which has androgen-receptor (AR) regula-tory elements in its promoter, may be involved in the regulationof androgen-dependent cell survival in prostate cancer cells.

Androgen-dependent proliferation is profoundly affected by ret-inoblastoma 1 (RB1) status, which was recognized as a critical de-terminant of therapeutic response in prostate cancer cells.5 RB1/p105 is a main transducer between cell cycling and gene expres-sion programming, and RB2/p130 regulates cell senescence alongwith p53.6,7 The binding partners, AR and RB1,8,9 and the AR

antagonist, PTEN,10 are considered major regulators in prostatecarcinogenesis. Functional status of these tumor suppressorsdefines cell-specific programmed cell death (PCD) responses indifferent prostate cell lines.

In a conventional androgen-sensitive prostate cancer model,LNCaP, PTEN inactivation leads to unopposed activity of mutantAR.10 Representing the androgen-resistant prostate cancer,DU145 and PC3 cells share androgen resistance, but differ in RB,PTEN and p53 status. Unlike wild-type RB1 status in PC3 cells,RB1 pocket function is abolished in DU145.11 However, the coin-troduced AR1RB8 and p53-associated regulators12 can inducePCD in DU145 despite their RB1-null/p53-heterozygous/Bax-de-ficient status. In the PTEN/p53-null PC3, the ectopically expressedp53 can intricately regulate PTEN.13

Consequently, in this study, we used these prostate cancer cellswith different androgen sensitivity and mutational status versusnormal prostate (PrEC) cells to assess cell growth and viability/cy-totoxicity responses to WT- and DN-ANXA7 vs. p53 by multipleassays that differentiated early and late apoptosis as well as nona-poptotic cell death. Next, we explored ANXA7 associations withthe major cell survival regulators, AR and RB1, and juxtaposedcorresponding gene expression profiles to tumor suppressor effectsof WT/DN-ANXA7 vs. p53 in different prostate cancer cells.These results have shown that WT-ANXA7 is capable ofmodulating key crossroads in prostate carcinogenesis and canexert a tumor suppressor role in prostate cancer cells despite path-ologic androgen signaling and dysfunctional status of major tumorsuppressors.

Material and methods

Cell culturing and infection

Normal prostate epithelial cells (PrEC) and prostate cancer cells(LNCaP, DU145 and PC3) were routinely grown at 37�C in ahumidified incubator with 5% CO2 (American Type Culture Col-lection, Manassas, VA). Cells were cultured as follows: PrEC inmedia with PrEGM Bullet Kit; LNCaP in RPMI-1640; DU145and PC3 in glutamine-containing MEM Eagle and F12K, respec-tively (Cambrex, East Rutherford, NJ; Invitrogen-Gibco, Carlsbad,CA). Adenoviral vector-based plasmids were constructed withAdEasy System (Johns Hopkins Oncology Center, Baltimore,MD). Subconfluent cells were infected with an ‘‘empty’’ vector,WT/DN-ANXA7 or p53 constructs. Cells infected with an emptyvector were used as control if not otherwise specified. All con-structs contained a green fluorescent protein (GFP) expressionmarker. Infection efficiency was monitored by GFP expression.

Grant sponsor: Department of Defense; Grant number: DAMD17-03-1-0107.*Correspondence to: Department of Anatomy, Physiology and Genet-

ics, USU School of Medicine, 4301 Jones Bridge Road, Bethesda, MD20814, USA. Fax:1301-295-1786. E-mail: [email protected] 11 March 2009; Accepted after revision 12 May 2009DOI 10.1002/ijc.24592Published online 26 May 2009 in Wiley InterScience (www.interscience.

wiley.com).

Int. J. Cancer: 125, 2528–2539 (2009)' 2009 UICC

Publication of the International Union Against Cancer

The amounts of virus were optimized to reach similar levels of ef-ficiency (70–80%). WT-ANXA7 construct corresponded to amore abundant short isoform of ANXA7 (NP 001147.1, NCBI),which lacks an alternate in-frame exon that is present in a tissue-specific ANXA7 transcript. Compared to WT-ANXA7, DN-ANXA7 (which is known to inhibit the WT-ANXA7-inducedphosphatidylserine liposome aggregation) contained triple muta-tions affecting COOH-terminal residues in annexin repeats 2, 3and 4 (E277 fi Q277, D360-E361 fi N360-Q361 and D435-D436fi N435-N436, respectively).

Viability, cytotoxicity and PCD detection with cell cyclingand cell size analysis

Recombinant ANXA7 protein (that was used in cytotoxicityand viability assays) was prepared and purified using histidine af-finity chromatography (Veritas, Rockville, MD). Cytotoxicity wasdetermined by Cytotoxicity Detection Kit (LDH, from RocheApplied Science, Indianapolis, IN) based on lactate dehydrogen-ase (LDH) release from damaged/dying cells. Cell proliferationwas quantified using Cell Proliferation Kit II (XTT, from RocheApplied Science, Indianapolis, IN) based on XTT metabolizationby viable cells. Untreated parental cells were used as control.

PCD detection, cell cycle and cell size analysis were performedin cells transfected with WT/DN-ANXA7 or p53 (18 hr). PCDwas detected by AnnexinV-PE and APO-BRDU Kits (BDPharmingen, San Jose, CA). Phosphatidylserine exposure andmembrane permeabilization (as early and late stages of PCD,respectively) were identified using AnnexinV-PE and flow cytom-etry (EPICs XL-MCL, Beckman Coulter, Fullerton, CA). DNAfragmentation in the end-stage PCD was detected using TUNEL(terminal deoxynucleotidedyltransferase dUTP nick end labeling)-based APO-BRDU and flow cytometry (LSRII, BD Biosciences,San Jose, CA). To avoid the overlap with GFP-marker, FITC-labeled anti-BRDU mAb from the kit was substituted with PE-conjugated anti-BRDU mAb (BD Pharmingen, San Diego, CA).Apoptosis detection and forward scatter (FSC)-based cell sizeanalysis were performed in single GFP-positive cell populationsusing WinList Software (Verity Software House, Topsham, ME).To discriminate between viable and nonviable (dead) subpopula-tions in the cell size analysis, AnnexinV-PE assay was combinedwith 7-amino-actinomycin D (7-AAD) staining (BD Pharmingen,San Diego, CA).

Cell cycle analysis was performed in cells fixed in 70% ethanolafter the removal of dsRNA by DNase-free RNase and propidiumiodide staining (Sigma-Aldrich, St. Louis, MO). Flow cytometry(EPICs XL-MCL, Beckman Coulter, Fullerton, CA) was per-formed using ModFit LT (Verity Software House, Topsham, ME).

Clonogenic growth (6, 12, 24 and 48 hr) was assessed by 0.5%crystal violet staining in PC3 cells transfected with vector aloneand WT/DN-ANXA7. PC3 cells (1 3 105) were seeded on 24-well plates and infected with designated amounts (0.5–4 U) of cor-responding adenoviral vectors (24 hr). After incubation (6–48 hr),media was aspirated and cells were washed and fixed with 2%formaldehyde. Remaining live cells were stained, whereas deadcells were washed off. After destaining, the absorbance (OD) wasestimated using FLUOstar Optima Plate Reader (540 nm).Cell survival was assessed by absorbance and quantified based oncorresponding virus amounts.

Statistical analysis was performed using a Student’s t-test forindependent samples (mean 6 SD from replicates) and a 2-tailedZ-test for proportions; p-values <0.05 (2-sided test) wereconsidered statistically significant.

Western blotting

The preparation of cell extracts was carried out using standardprocedure. Equal amounts of total protein were electrophoresed on4–20% Tris/Glycine gels using MagicMark as protein standard(Invitrogen, Carlsbad, CA). Immunoblotting was performed usingantibodies specific for AR (AR-441 from Santa Cruz Biotechnol-

ogy, CA) and total or phosphorylated RB1 (Rb-4H1 and phos-phoRb-Ser780, from Cell Signaling Technology, Danvers, MA)with housekeeping beta-actin as control. Intensity of bands (means6 SD, arbitrary units) was assessed using ImageJ (http://rsb.onfo.nih.gov/ij).

cDNA microarray

Cells were transfected with WT/DN-ANXA7 or p53 and har-vested (18 hr) as described earlier. mRNA (3 lg per experiment)was isolated and processed using RNAqueous-4PCR Kit (Ambion,An Applied Biosystems Business, Austin, TX). cDNA microarrayanalysis was performed using Atlas Human Cancer 1.2 Arrays andAtlasImage 2.01 software as recommended by the manufacturer(Clontech, Palo Alto, CA).

After filtration and global normalization, adjusted intensitieswere calculated as spot intensities minus background values andmultiplied by the normalization coefficients. Adjusted spot inten-sities were used to calculate ratios (R) in pair-comparisons.Reports with outlier genes (R � 2 based on average values fromduplicates) were generated for WT- vs. DN-ANXA7 and WT-ANXA7 vs. p53 pair-comparisons in different cell lines. Next, In-genuity Pathways Analysis (IPA) was used to validate expressionprofiling and to discern molecular paths in ANXA7 tumorsuppressor effects in prostate cancer cells.

Ingenuity Pathways Analysis

To discriminate the molecular pathways responsible for WT-ANXA7 effects versus DN-ANXA7 or p53, we used IPA software(www.ingenuity.com, Ingenuity Systems, Redwood City, CA).Average expression ratio R > 2 in WT- vs. DN-ANXA7 or p53array comparisons was used as threshold. The reports with outliergenes from cDNA microarray analysis were uploaded and mappedto corresponding objects (genes/proteins) in IPA’s database. First,core analysis was used to identify WT- vs. DN-ANXA7 signaturesin each prostate cancer cell line (LNCaP, DU145 and PC3).Subsequently, comparison core analysis was used to evaluatecanonical pathways and bio functions that were most distinctlyaffected by WT-ANXA7 in different cells.

The ‘‘Cell cycle G1/S checkpoint regulation’’ canonical path-way (which was most specifically affected by WT-ANXA7 inLNCaP) was customized by adding the outliers from pair-compar-isons as well as ANXA7-associated objects from IPA’s database.As a result, the WT-ANXA7-specific signature with genes (nodes)and relationships (edges) was integrated with a major IPA canoni-cal pathway representing RB1-associated cell death/proliferationnetwork. Path Designer (IPA), was used to create a final networkwith plausible molecular paths between tumor suppressor ANXA7and AR/RB-E2F-associated cell survival control (more details areincluded in ‘‘Results’’ section and corresponding figure legends).

Results

ANXA7 protein affected cell viability in prostate cancer cells, butnot in normal prostate cells

To assess WT-ANXA7 cytotoxicity, we first treated cells withWT-ANXA7 protein versus its DN-ANXA7 form (Fig. 1). Inbenign PrEC, both ANXA7 forms caused a hardly detectable cyto-toxicity (<2%). In androgen-sensitive LNCaP prostate cancercells, WT-ANXA7 affected cell survival in a dose- and time-dependent manner causing a noticeable cytotoxicity and eventu-ally reducing cell viability by �10% (100 nM for 72 hr, p <0.001). Androgen-resistant cells responded to ANXA7 more dras-tically. In DU145 cells, WT-ANXA7 (100 nM for 72 hr) almostdoubled the DN-ANXA7-induced levels of cytotoxicity (20%, p <0.009) and cell viability reduction (40%, p < 0.001). In PC3 cells(where endogenous ANXA7 level is higher than in LNCaP orDU145, data not shown), both ANXA7 forms induced moderatecytotoxicity, decreasing live cell populations by >20%. Thus, theoverexpressed ANXA7 protein protected normal prostate cells,

2529ANXA7 VERSUS p53 IN BENIGN AND CANCEROUS PROSTATE CELLS

but effectively killed androgen-resistant as well as -sensitiveprostate cancer cells.

The transfected WT-ANXA7 protected normal prostate cells, butprofoundly increased PCD in prostate cancer cells regardless oftheir androgen sensitivity

Next, we assessed PCD effects (DNA fragmentation by APO-BRDU as well as phosphatidylserine exposure and membranepermeabilization by AnnexinV-PE) of the transfected WT/DN-ANXA7 vs. p53. Compared to WT-ANXA7 protein (as shown inFig. 1), the transfected WT-ANXA7 maintained protective effectsin PrEC, but caused a higher cytotoxicity in cancerous cells(Fig. 2).

In androgen-sensitive LNCaP cells, WT-ANXA7 inducedDNA fragmentation (APO-BRDU, Fig. 2a) in almost half ofthe cell population, causing a 2-fold PCD increase in com-parison to DN-ANXA7 (p < 0.001). DN-ANXA7 also failed(p < 0.001 for both comparisons) to reach the WT-ANXA7-induced levels of early apoptosis with phosphatidylserineexposure and of late apoptosis with membrane permeabiliza-tion (AnnexinV-PE, Fig. 2a). In the meantime, p53 barelyincreased PCD, phosphatidylserine exposure or membranepermeabilization in LNCaP, but caused apoptosis and PCD innormal PrEC.

In androgen-resistant prostate cancer cells, ANXA7 effectswere more cell type specific compared to the p53, which hadalmost similar increase in PCD in both cell lines. In DU145, WT-

(but not DN)-ANXA7 drastically induced early apoptosis(phosphatidylserine exposure) and DNA fragmentation (�40%, p� 0.001). In PC3, both ANXA7-forms substantially increasedmembrane permeabilization and DNA fragmentation. As a result,the WT- vs. DN-ANXA7 difference in PCD rates was approxi-mately 3- to 4-fold in DU145 compared to approximately 1- to1.5-fold in PC3. The relatively effective induction of apoptosis byDN-ANXA7 in PC3 was concordant to the higher endogenousANXA7 protein expression in PC3 compared to DU145 (notshown).

To further assess the effects of transfected WT/DN-ANXA7 oncell survival in PC3, we used crystal violet staining that includedalternative (nonapoptotic) cell death responses. WT-ANXA7 abro-gated PC3 cell survival in a time- and dose-dependent manner,unequivocally exceeding the effects of DN-ANXA7 (Figs. 2band 2c).

Thus, the transfected WT(but not DN)-ANXA7 matched a ca-nonical tumor suppressor p53 in eliminating androgen-resistantDU145 and PC3 prostate cancer cells and surpassed p53 in killingandrogen-sensitive LNCaP prostate cancer cells or protectingnormal prostate PrEC cells.

The transfected WT-ANXA7 caused G1-arrest in LNCaP and G2-arrests in DU145 and PC3, but did not affect cell cycling in PrEC

Consistent with selective PCD induction, WT-ANXA7 did notchange cell cycling in normal prostate cells (PrEC), whereas p53essentially prolonged their G1-phase (Fig. 3). Among cancer cells,

FIGURE 1 – Cell viability and cytotoxicity in normal and cancerous prostate cells treated with exogenous WT- and DN-ANXA7 protein. Cellcytotoxicity (LDH, columns) and viability (XTT, lines) in cells treated with WT/DN-ANXA7 protein were determined as described in ‘‘Materialand methods’’ section. Responses to WT-ANXA7 (in black) versus DN-ANXA7 (in gray) were measured in triplicate experiments (mean 6 SD,%) and presented as levels of cytotoxicity (control as 0%, left Y-axis) and viability (control as 100%, right Y-axis) normalized tountreated control.

2530 TOROSYAN ET AL.

WT-ANXA7 caused a moderate G1-increase in LNCaP and aprofound G2-arrest in both DU145 and PC3. DN-ANXA7 didnot match the WT-ANXA7-caused cell cycle changes in anycancer cell line (p < 0.001). Unlike WT-ANXA7, p53 delayedG1-phase in DU145 and PC3 cells, but not in LNCaP. Thus,the transfected WT-ANXA7 arrested prostate cancer cellsregardless of androgen sensitivity, but protected cell cycling innormal prostate cells.

Cell size alterations in response to WT/DN-ANXA7 and p53reflected their effects on cell survival

Homeostatic cell size control is regulated by checkpoints that

couple cell size to cell cycle progression and cell death. Tumor

suppressors such as RB and PTEN can function as cell size check-

points, while their mutations result in carcinogenesis and aberrant

cell size regulation. For instance, deficiency of RB1 reduces cell

FIGURE 2 – PCD in normal and cancerous prostate cells transfected with WT/DN-ANXA7 and p53. As described in ‘‘Material and methods’’section, (a) PCD responses to WT/DN-ANXA7 and p53 (18 hr) were assessed by Annexin V-PE (phosphatidylserine exposure as early apoptosisand membrane permeabilization as late apoptosis) and APO-BRDU (DNA fragmentation). Stacked bars represented average WT/DN-ANXA7-and p53-induced PCD rates (%) from replicates after the subtraction of corresponding vector values. *In DU145, neither WT/DN-ANXA7 norp53 caused essential changes in membrane permeabilization (late apoptosis by Annexin V-PE, <3%). (b and c) PC3 cells were transfected withdesignated amounts of WT/DN-ANXA7 or ‘‘empty’’ vector. Clonogenic survival (6–48 hr) was assessed using crystal violet staining (cellimages, b) and quantified (18 hr) based on absorbance (means 6 SD from triplicates, c). [Color figure can be viewed in the online issue, whichis available at www.interscience.wiley.com.]


size, whereas PTEN increases cell size.14–16 Growth anomalies in

cancer-prone Anxa7(1/2) mice included organomegaly and beta-

cell hypertrophy,1,17 suggesting that ANXA7 can also be involved

in cell size regulation.

Analysis of total cell populations transfected with WT-ANXA7,DN-ANXA7 or p53 (Fig. 4, top) revealed that the cell size wasaltered in PTEN-deficient LNCaP and PC3, but not in RB1-nullDU145. Further cell size analysis in viable versus nonviable sub-populations in LNCaP and PC3 cells (Fig. 4, bottom) identifieddistinct trends in responses to WT-ANXA7 and p53. Unlike theirviable counterparts, the WT(but not DN)-ANXA7-transfectednonviable LNCaP cells were smaller than p53-transfected. On thecontrary, the WT(but not DN)-ANXA7-transfected nonviable PC3cells were larger than p53-transfected. Juxtaposed to PCDresponses (as shown in Fig. 2), a smaller size of nonviable cellscorresponded to apoptotic shrinkage in WT-ANXA7-transfected

LNCaP and p53-transfected PC3, whereas the enlargement mostlikely reflected nonapoptotic cell death in WT-ANXA7-trans-fected PC3 and p53-transfected LNCaP. In addition, a left shift inp53-transfected PrEC cells (top) was consistent with an apoptoticresponse to p53 in normal prostate cells. Thus, cell sizeanalysis reflected cell type-specific cell death responses to bothWT-ANXA7 and p53, whereas less discernible cell sizealterations under DN-ANXA7 were concordant with the reducedDN-ANXA7 cytotoxicity.

WT-ANXA7 distinctly affected AR-RB1 protein expression andRB1 phosphorylation in prostate cancer cells with differentandrogen sensitivity and RB1 status

Based on the aforementioned results, WT-ANXA7 acted as atumor suppressor gene affecting cell proliferation, cell cycling andcell death in hormone-sensitive as well as -resistant prostate

FIGURE 3 – Cell cycling in normal and cancerous prostate cells transfected with WT/DN-ANXA7 and p53. WT/DN-ANXA7 and p53 (18 hr)cell cycling responses were presented as cell numbers in different phases in comparison with corresponding controls (%, mean values fromreplicates after the subtraction of vector levels).

FIGURE 4 – Cell size analysis in normal and cancerous prostate cells transfected with WT/DN-ANXA7 and p53. Cell size was assessed by for-ward scatter-based analysis as described in ‘‘Material and methods’’ section. Cell size assessment was performed in total cell populations (top)in all cell lines as well as in viable versus nonviable subpopulations in LNCaP and PC3 (bottom). A left shift represented cell size reduction anda right shift represented enlargement. Responses to vector alone, WT/DN-ANXA7 and p53 in total populations were marked as described in thefigure; the area under WT-ANXA7 curve (dotted line) was additionally highlighted in gray. Responses in viable and nonviable cell subpopula-tions included WT-ANXA7 and p53 patterns only.


cancer cells. AR is a master regulator of G1-S phase progressionin prostate cancer cells that is able to induce RB1 phosphoryla-

tion/inactivation and thereby govern the androgen-dependent pro-liferation.9 The major cell cycle driver RB1 is growth stimulatoryin hyperphosphorylated and growth inhibitory in hypophosphory-

lated status.6 Hence, we assessed AR profile in association withRB1 protein expression and phosphorylation status in response toANXA7 vs. p53.

In LNCaP with the highest AR expression, WT-ANXA7 andp53 specifically reduced the 20–50 kDa AR products (p < 0.001

compared to DN-ANXA7, Fig. 5a). It is known that AR degradesrapidly in the absence of the ligand, and its breakdown productscan suppress endogenous AR signaling thereby affecting the AR-dependent cell growth.18 The formation of widely ranging frag-ments (26–116 kDa) was associated with the G1-S phaseprogression in LNCaP,19 and AR45-variant, in particular, wasshown to inhibit full-length AR signaling.20 Correspondingly,both tumor suppressors (WT-ANXA7 and p53) reduced, but DN-ANXA7 (which resulted in the cell growth progression) enhancedthe expression of LMW-AR products including �45 kDa-AR inLNCaP.

FIGURE 5 – AR-RB1 synexpression in prostate cancer cells transfected with WT/DN-ANXA7 and p53. (a) Western images for full-length(�110 kDa) and LMW-AR as well as total and phosphorylated (Ser780) RB1 (beta-actin as control). (b) Protein expression levels for total andphosphorylated RB1 in response to WT/DN-ANXA7 or p53 were calculated as ratios to corresponding vector controls in each cell line. Normal-ized protein expression levels were presented on logarithmic scale in juxtaposition to PCD rates.


FIGURE 6 – Tumor suppressor ANXA7 in the AR and RB-E2F-associated cell survival regulation in prostate cancer cells (Ingenuity PathwaysAnalysis, IPA). (a) Canonical pathway of cell cycle: G1/S check point regulation (IPA) with WT- vs. DN-ANXA7 expression overlay that wasbased on WT/DN-ANXA7 expression ratios (R > 2) in LNCaP; p21/Cip1 and RB were downregulated (green) and E2F upregulated (red) byWT-ANXA7. (b) In the next step, the original pathway was used as a template to explore the relationships with other WT- vs. DN-ANXA7 out-liers, which were able to connect to the G1/S checkpoint regulation. As a result, the same pathway was enriched with other outliers distinctlyaffected by WT-ANXA7 and, thereby represented possible connections (as additional edges) between G1/S regulatory cascade and the ANXA7-associated network in our study. The customized pathway outlook (Path Designer, IPA) included cell image with different compartments:nucleus was depicted as a big round shape (brown) with DNA strand (red), ribosome was observed as a small oval shape (marine), and cell mem-brane as a strand on the top (blue). The ANXA7-affected genes that were placed around the nucleus represented possible transcriptional changesin response to ANXA7 (which can be found in the nucleus in addition to its conventional membrane localization, as depicted). The expressionoverlay represented ratios from WT- vs. DN-ANXA7 array in comparison with LNCaP; downregulated by WT-ANXA7 genes are shown ingreen and upregulated - in red (more intense color represented a higher significance of regulation). To avoid further complexity, the WT-ANXA7 vs. p53 expression overlay was omitted. However, some of the genes that were distinctly affected by WT-ANXA7 vs. p53 [blank geneshapes with no color coding among the genes that were added to conventional pathway in (a)] were shown due to their close connections to theG1/S checkpoint. Through the binding relations to Ca21/sorcin and galectin-3, LGALS3 (black edges in bold), ANXA7 was directly tied to theG1/S checkpoint network, where RB-E2F, AR and p53 as well as the p53-associated cell survival regulators were the major hubs with multipleintertwined connections (as edges in gray). (c) Major ANXA7-induced alterations that were discerned from WT- vs. DN-ANXA7 synexpressionanalysis as plausible consecutive molecular paths that could be responsible for WT- vs. DN-ANXA7 effects on LNCaP cell survival. ANXA7effects in LNCaP cells were represented by background cell images showing dead and viable cells in response to WT- and DN-ANXA7, respec-tively. (d) Major results from IPA comparison core analysis on WT- vs. DN-ANXA7 effects in different prostate cancer cell lines. Significancescores were calculated by IPA using WT/DN-ANXA7 expression ratios (R > 2) from WT- vs. DN-ANXA7 array comparison in each cell line.Bars represented Fisher-based significance scores (2log[p-value], no bars when �0) in canonical pathways and bio functions (IPA) that weremost distinctly changed in response to WT- vs. DN-ANXA7 in different cell lines. Yellow horizontal line in all graphs designated significancethresholds (p < 0.05). A score for each network was computed according to the fit of outliers’ sets and derived from ap-value indicating the likelihood of the outliers in a network being found together due to random chance.


Although androgen-insensitive DU145 and PC3 similarly lackfunctional AR, they have different AR status. Methylated AR pro-moter and mutant status of the RB1, which is a selective ARbinder and coactivator,21 virtually prevent AR expression inDU145. In contrast, PC3 have a normal AR gene and retain thecoregulators required for full-length AR tumor suppressor abil-ity.22 Remarkably, in the PC3 cells with a residual AR expression,WT-(but not DN)-ANXA7 decreased the LMW-AR products(especially �45 and 30 kDa), but not full-length AR (Fig. 5a).

RB1 expression was also distinctly modulated by WT-ANXA7in different cells (Fig. 5a). In LNCaP, WT-ANXA7 maintained

higher levels of total RB1 in comparison to p53, whereas DN-ANXA7 hyperphosphorylated RB1. In DU145 with the lowestbaseline RB1 level, WT-ANXA7 increased total RB1 expression,but decreased its phosphorylation similar to p53. DN-ANXA7(which did not suppress DU145 cell growth) decreased the phos-pho-RB1, but failed to upregulate total RB1. In PC3 with theinitially higher RB1 expression, p53 (but not ANXA7) greatlyreduced the phospho-RB1.

As a result, total RB1 levels most closely corresponded to PCDrates induced by WT-ANXA7 and p53 in the RB1-null DU145(Fig. 5b). In the PC3 cells with highly expressed RB1 (as well as

FIGURE 6 – CONTINUED.


endogenous ANXA7, not shown), both ANXA7 forms resulted insimilar RB1 expression and phosphorylation status that adequatelyreflected their comparable PCD rates. As expected, high PCD ratesand G1-arrests induced by the cell cycle regulator p53 in DU145and PC3 coincided with RB1 hypophosphorylation, whereas theinadequate tumor suppressor effects of p53 (as well as DN-ANXA7) in LNCaP coincided with RB1 hyperphosphorylation.

Overall AR-RB1 synexpression demonstrated that WT-ANXA7effects in prostate cancer cells could involve the reinstatement oftumor suppressor function of RB1 as well as AR, especially in theandrogen-sensitive cells.

ANXA7-induced AR-RB synexpression was linked to theRB-associated G1/S checkpoint regulation and E2F-mediatedapoptosis in androgen-sensitive LNCaP prostatecancer cells (IPA)

To further explore AR-RB associations in ANXA7 network, weanalyzed cDNA microarray-derived WT/DN-ANXA7 and p53gene expression profiles using IPA (Fig. 6).

Consistent with the experimental evidence that WT-ANXA7induced G1-arrest, while DN-ANXA7 preserved the cell growthprogression in LNCaP, the RB-associated G1/S checkpoint reg-

ulation pathway (Fig. 6a) was identified by IPA core analysisas one of the most distinctly affected by WT- vs. DN-ANXA7in LNCaP (2log[p-value] 5 4). Active/unphosphorylated RB1preferentially binds to its sequence-specific transcription factorE2F1 and forms the self-regulating RB-E2F1 complex, whosecapability of promoting or restraining cell cycle progressionand inducing apoptosis is dependent on cellular context.6,23

While DN-ANXA7 upregulated both RB1 and RB2 andupstream p21, WT-ANXA7 significantly increased E2F1 (R >2.4 and 4 compared to DN-ANXA7 and p53, respectively) inaddition to maintaining unphosphorylated/active RB1 protein(as shown in Fig. 5).

The WT-ANXA7-associated network demonstrated multiplelinks to the RB-E2F-mediated G1/S checkpoint regulation path-way (Fig. 6b). First of all, ANXA7 binding partners, galectin-3,LGALS324 and sorcin, SRI25 are directly linked to the AR-RBcrosstalk. Galectin-3 was shown to affect RB1 profile26,27 similarto ANXA7, implying that WT-ANXA7 could employ its bindingpartner for regulating RB1 expression and phosphorylation inprostate cancer cells.

Unlike DN-ANXA7 and p53, which had inadequate PCDeffects in LNCaP, WT-ANXA7 greatly reduced N-myc expression.Direct connection between the loss of functional RB and

FIGURE 6 – CONTINUED.


amplification of the RB-binding N-myc oncogene was shown inhuman retinoblastoma and its murine model.28

Indicating transcriptional changes, WT-ANXA7 signatureincluded multiple paths leading to the nucleus. The E2F1transcriptional target29 and Bcl-2/p53 binding tumor suppressorTP53BP2 (or apoptosis stimulating protein 2, ASPP2) washighly upregulated by WT-ANXA7, but not by DN-ANXA7or p53 (R > 18 and 130, respectively). WT-ANXA7 alsoupregulated the apoptosis signal-regulating kinase ASK1 (R >3 and 2 compared to DN-ANXA7 and p53, respectively). Adirect E2F1 target and RB1 binding partner, ASK1 (orMAP3K5/MEKK5), is a MAPK kinase that can link RB-asso-ciated PCD and cell cycling through a positive feedback regu-lation of the E2F1 apoptotic activity via RB inhibition.30,31

Lastly, by upregulating proapoptotic BAK, WT-ANXA7appeared to reinstate the BAK-dependent AR-RB control.8 Incontrast, DN-ANXA7 could lack proapoptotic effects due tothe upregulation of prosurvival ERK5 (or MAPK7/BMP1).

E2F activity can be integrated with DNA damage response by14-3-3 family.32 The loss of 14-3-3-sigma contributes to tumori-genesis in different tissues including prostate.33 Compared to DN-ANXA7, WT-ANXA7 upregulated 14-3-3-sigma that was shownto be induced by p53.34 WT-ANXA7 also upregulated the cellcycling associated with CDC25B, which can accumulate afterDNA damage35 and act as AR coactivator.36

The elevated E2F1 could potentially affect AR transcriptionby binding to AR promoter along with pocket family mem-bers.37 Consistently, gene synexpression of NCOR2/SMRT andSTAT5(A/B) was in line with the ANXA7-mediated AR proteinexpression (as shown in Fig. 5a). The AR-interacting nuclearreceptor and corepressor NCOR238 can negatively regulate theSTAT5-dependent transcription39 in a direct functional coopera-tion between AR and STAT5.40 WT-ANXA7 increasedNCOR2, but decreased STAT5(A/B) gene expression suggestinga subsequent AR downregulation (AR was not available onarrays). On the other hand, an opposite NCOR2-STAT5 profilewas consistent with a higher expression of AR protein productsin response to DN-ANXA7.

In summary, we propose that in addition to AR regulation, theWT-ANXA7 tumor suppressor mechanism in androgen-sensitivecells involved the reciprocal regulation of RB1 and E2F1 gov-erned by p53/p21 (Fig. 6c). Overexpressed E2F1 is capable ofinducing apoptosis, but its activity is repressed by RB1. The p53/p21-initiated RB activation can impede the propensity of E2F1-mediated apoptosis, and p21 can bind to as well as increaseexpression and decrease phosphorylation of RB1.23 Hence, theWT-ANXA7-induced E2F1 upregulation with a maintainedactive/unphosphorylated RB1 status can constitute a reversal ofthe RB-dependent repression of E2F-mediated proapoptotictranscription. By upregulating the downstream targets ASPP2/TP53BP2, ASK1/MAP3K5, and BAK, WT-ANXA7 reinstated theRB1-associated control on cell survival. Instead, by maintaining ahigher expression of LMW-AR products, p21, and RB1/2 alongwith RB1 hyperphosphorylation, DN-ANXA7 continued torepress the E2F-mediated apoptosis and upregulated the prolifera-tion-promoting ERK5/MAPK7.

Thus, IPA identified the reversed RB repression of the E2F-mediated proapoptotic transcription as a major WT-ANXA7 effectin androgen-sensitive prostate cancer cells.

In androgen-resistant prostate cancer cells, IPA identifiedadditional ANXA7-associated pathways includingPTEN and NFkB

Different AR-RB profiles with no E2F upregulation suggestedalternative pathways in the WT-ANXA7-induced PCD and G2-arrests in androgen-resistant cells. To further explore ANXA7 reg-ulatory pathways, we compared WT- vs. DN-ANXA7 profilesusing comparison core analysis (IPA) in the 3 prostate cancer celllines (Fig. 6d). In addition to G1/S checkpoint regulation, the G2/

M checkpoint regulation as well as the p53 and 14-3-3-mediatedsignaling (Canonical Pathways, IPA) comprised a majordifference between WT- and DN-ANXA7 in LNCaP. In andro-gen-resistant cells, most distinct WT- vs. DN-ANXA7 profileswere associated with NF-kB signaling in DU145, and with deathreceptor and PTEN signaling in PC3. Remarkably, the most signif-icant difference in proapoptosis (that presumably involved NF-kB) corresponded to the most drastic difference between WT- andDN-ANXA7-induced apoptotic rates that was found in DU145 (asshown in Fig. 2). In contrast, antiapoptosis remained insignificant,whereas cell death and cell cycle categories were similarly signifi-cant in all cell lines. In PC3 cells which completely lack 2 majortumor suppressors, p53 and PTEN, cancer signaling constitutedthe most significant difference in WT- vs. DN-ANXA7 profiles(2log [p-value] 5 17.5).

Thus, by positioning ANXA7 within a large-scale regulatorynetwork in different cancer cells, IPA revealed potential involve-ment of NF-kB and PTEN in tumor suppressor effects of ANXA7in androgen-resistant prostate cancer cells.

Discussion

WT-ANXA7 demonstrated profound cytotoxicity in andro-gen-sensitive (LNCaP) as well as androgen-resistant (RB/Bax-deficient DU145 and p53/PTEN-null PC3) prostate cancer cells,but not in normal prostate (PrEC) cells. In contrast, DN-ANXA7 did not match the WT-ANXA7-induced cell growtharrest or cell death. WT- vs. DN-ANXA7 effects correspondedto distinct LMW-AR protein formation as well as RB1 expres-sion and phosphorylation status. According to Western- andcDNA microarray-derived AR-RB-associated synexpression,WT-ANXA7 tumor suppressor effects in androgen-sensitiveLNCaP cells were based on the reversed RB-dependent repres-sion of E2F1 followed by the transcriptional upregulation ofproapoptotic E2F-targets (IPA).

By affecting the RB-E2F network, WT-ANXA7 displayed simi-larity to the upstream RB1 regulator p53. Among E2F transcrip-tion factors, E2F1 is the only specific inducer of apoptosis that cantarget p53 in a reciprocal self-regulation of RB/E2F pathway.41

AR can regulate the cell cycle-dependent transcription in prostatecancer cells by directly interacting with E2F1,42 whereas E2F1can deregulate androgen-dependent growth and enhance apoptosisin LNCaP.43 By binding to RB1 in a cell cycle-dependent manner,E2F1 can specifically mediate the p53-dependent cell proliferationand apoptosis via upregulation of ASPP2.29 Unlike p53, ASPPproteins do not affect the E2F-induced apoptosis.44 Hence, theE2F-mediated upregulation of the p53-binding ASPP2/TP53BP2by WT-ANXA7, but not p53 itself, could hold a clue to moreadvanced tumor suppressor effects of WT-ANXA7 vs. p53 inLNCaP.

ANXA7 was reduced in hormone-resistant prostate cancers.4

However, exogenous WT-ANXA7 was capable of overcomingandrogen-resistant status in vitro in both DU145 and PC3. Bytargeting AR, WT-ANXA7 represented a potential alternativeto antiandrogens which may promote tumor invasiveness.45

Because of the gain of function, AR becomes an oncogene intype-II androgen depletion insensitive cells, but not in thetype-I DU145 and PC3. In fact, RB1 can activate AR-depend-ent apoptosis,8,21 and the reinforced AR can still function as atumor suppressor in PC3 cells.22 As a result, the WT-ANXA7-induced downregulation of LMW-AR products, but not of full-length AR in LNCaP and PC3 suggested that ANXA7 canpotentially revive the RB-associated AR tumor suppressorcapacity in prostate cancer cells.

Tumor suppressor effects of WT-ANXA7 (10q21) in LNCaPand PC3 cells, in particular, implied an ability to surpass PTENdeficiency. Located within the same TSG region (10q23.3), PTENis antagonistic to AR in cell survival control in prostate cancercells10 and is directly linked to the cell cycle controlling


machinery whose final target is RB1.46 Reintroduction of func-tional RB1 to RB1-deficient cells restores their sensitivity toPTEN, which inhibits RB1 hyperphosphorylation and induces cellcycle arrest. By similarly inhibiting RB1 hyperhosphorylation,WT-ANXA7 could act as a PTEN substitute, forcing the RB1-and PTEN-deficient cells to undergo cell growth arrest and apo-ptosis. Consistent with a presumably restored PTEN function byWT-ANXA7 in our study, ANXA7 deficiency in Anxa7(1/2)murine model resulted in PTEN inhibition.1 Remarkably, the bind-ing partner of ANXA7, galectin-3, was recently identified as anovel PTEN regulator.47 These data suggested that WT-ANXA7could intervene with the p53/PTEN control over RB-mediated cellsurvival, thereby opposing pathologic AR signaling in prostatecancer cells.

On the other hand, distinct effects of DN-ANXA7 (that abol-ishes WT-ANXA7-mediated Ca/phosphatidylserine liposomeaggregation) unequivocally implicated the membrane-associatedannexin properties in ANXA7 tumor suppression. By distinctlyaffecting LMW-AR in LNCaP, the Ca-binding ANXA7-formscould, in particular, differentially regulate the Ca-dependent ARbreakdown that is relevant to cell cycling.19

In summary, ANXA7 acted as a tumor suppressor that over-came pathologic androgen signaling and dysfunctional status ofRB1, PTEN and p53, which are commonly mutated in humanprostate cancer. Our ongoing studies are expected to furtherelucidate ANXA7 diagnostic and therapeutic potential that can

eventually be translated into the clinical setting in prostate cancer.

References

1. Srivastava M, Montagna C, Leighton X, Glasman M, Naga S,Eidelman O, Ried T, Pollard HB. Haploinsufficiency of Anx7tumor suppressor gene and consequent genomic instability pro-motes tumorigenesis in the Anx7(1/2) mouse. Proc Natl AcadSci USA 2003;100:14287–92.

2. Srivastava M, Bubendorf L, Srikantan V, Fossom L, Nolan L, Glas-man M, Leighton X, Fehrle W, Pittaluga S, Raffeld M, Koivisto P,Willi N, et al. ANX7, a candidate tumor suppressor gene for prostatecancer. Proc Natl Acad Sci USA 2001;98: 4575–80.

3. Leighton X, Srikantan V, Pollard HB, Sukumar S, Srivastava M.Significant allelic loss of ANX7 region (10q21) in hormone receptornegative breast carcinomas. Cancer Lett 2004;210:239–44.

4. Srivastava M, Torosyan Y, Raffeld M, Eidelman O, Pollard HB,Bubendorf L. ANXA7 expression represents hormone-relevant tumorsuppression in different cancers. Int J Cancer 2007;121:2628–36.

5. Sharma A, Comstock CE, Knudsen ES, Cao KH, Hess-Wilson JK,Morey LM, Barrera J, Knudsen KE. Retinoblastoma tumor suppressorstatus is a critical determinant of therapeutic response in prostatecancer cells. Cancer Res 2007;67:6192–203.

6. Goodrich DW. The retinoblastoma tumor-suppressor gene, the excep-tion that proves the rule. Oncogene 2006;25:5233–43.

7. Helmbold H, Deppert W, Bohn W. Regulation of cellular senescenceby Rb2/p130. Oncogene 2006;25:5257–62.

8. Wang X, Deng H, Basu I, Zhu L. Induction of androgen receptor-dependent apoptosis in prostate cancer cells by the retinoblastomaprotein. Cancer Res 2004;64:1377–85.

9. Balk SP, Knudsen KE. AR, the cell cycle, and prostate cancer. NuclRecept Signal 2008;6:e001.

10. Li P, Nicosia SV, Bai W. Antagonism between PTEN/MMAC1/TEP-1 and androgen receptor in growth and apoptosis of prostatic cancercells. J Biol Chem 2001;276:20444–50.

11. Bookstein R, Shew JY, Chen PL, Scully P, Lee WH. Suppression oftumorigenicity of human prostate carcinoma cells by replacing amutated RB gene. Science 1990;247:712–15.

12. Hemmati PG, G€uner D, Gillissen B, Wendt J, von Haefen C, Chinna-durai G, D€orken B, Daniel PT. Bak functionally complements for lossof Bax during p14ARF-induced mitochondrial apoptosis in humancancer cells. Oncogene 2006; 25:6582–94.

13. Tang Y, Eng C. p53 down-regulates phosphatase and tensin homo-logue deleted on chromosome 10 protein stability partially throughcaspase-mediated degradation in cells with proteasome dysfunction.Cancer Res 2006;66:6139–48.

14. Jorgensen P, Tyers M. How cells coordinate growth and division.Curr Biol 2004;14:R1014–R1027.

15. Lee C, Kim JS, Waldman T. PTEN gene targeting reveals a radiation-induced size checkpoint in human cancer cells. Cancer Res 2004;64:6906–14.

16. Fang SC, de los Reyes C, Umen JG. Cell size checkpoint control bythe retinoblastoma tumor suppressor pathway. PLoS Genet 2006;2:e167.

17. Srivastava M, Eidelman O, Leighton X, Glasman M, Goping G,Pollard HB. Anx7 is required for nutritional control of gene expres-sion in mouse pancreatic islets of Langerhans. Mol Med 2002;8:781–97.

18. Butler LM, Centenera MM, Neufing PJ, Buchanan G, Choong CS,Ricciardelli C, Saint K, Lee M, Ochnik A, Yang M, Brown MP, TilleyWD. Suppression of androgen receptor signaling in prostate cancercells by an inhibitory receptor variant. Mol Endocrinol 2006;20:1009–24.

19. Pelley RP, Chinnakannu K, Murthy S, Strickland FM, Menon M,Dou QP, Barrack ER, Reddy GP. Calmodulin-androgen receptor(AR) interaction: calcium-dependent, calpain-mediated breakdown

of AR in LNCaP prostate cancer cells. Cancer Res 2006;66:11754–62.

20. Ahrens-Fath I, Politz O, Geserick C, Haendler B. Androgen receptorfunction is modulated by the tissue-specific AR45 variant. FEBS J2005;272:74–84.

21. Yeh S, Miyamoto H, Nishimura K, Kang H, Ludlow J, HsiaoP, Wang C, Su C, Chang C. Retinoblastoma, a tumor suppres-sor, is a coactivator for the androgen receptor in human prostatecancer DU145 cells. Biochem Biophys Res Commun 1998;248:361–7.

22. Litvinov IV, Antony L, Dalrymple SL, Becker R, Cheng L, Isaacs JT.PC3, but not DU145, human prostate cancer cells retain the coregula-tors required for tumor suppressor ability of androgen receptor.Prostate 2006;66:1329–38.

23. Garner E, Raj K. Protective mechanisms of p53-p21-pRb proteinsagainst DNA damage-induced cell death. Cell Cycle 2008;7:277–82.

24. Yu F, Finley RL, Raz A, Kim HR. Galectin-3 translocates to the peri-nuclear membranes and inhibits cytochrome c release from the mito-chondria. A role for synexin in galectin-3 translocation. J Biol Chem2002;277:15819–27.

25. Brownawell AM, Creutz CE. Calcium-dependent binding of sorcin tothe N-terminal domain of synexin (annexin VII). J Biol Chem1997;272:22182–90.

26. Mazurek N, Sun YJ, Price JE, Ramdas L, Schober W, Nangia-Makker P, Byrd JC, Raz A, Bresalier RS. Phosphorylation ofgalectin-3 contributes to malignant transformation of human epi-thelial cells via modulation of unique sets of genes. Cancer Res2005;65:10767–75.

27. Kim HR, Lin HM, Biliran H, Raz A. Cell cycle arrest and inhibitionof anoikis by galectin-3 in human breast epithelial cells. Cancer Res1999;59:4148–54.

28. MacPherson D, Conkrite K, Tam M, Mukai S, Mu D, Jacks T. Murinebilateral retinoblastoma exhibiting rapid-onset, metastatic progressionand N-myc gene amplification. EMBO J 2007;26:784–94.

29. Sullivan A, Lu X. ASPP: a new family of oncogenes and tumour sup-pressor genes. Br J Cancer 2007;96:196–200.

30. Dasgupta P, Betts V, Rastogi S, Joshi B, Morris M, Brennan B,Ordonez-Ercan D, Chellappan S. Direct binding of apoptosis signal-regulating kinase 1 to retinoblastoma protein: novel links betweenapoptotic signaling and cell cycle machinery. J Biol Chem2004;279:38762–9.

31. Tan J, Zhuang L, Jiang X, Yang KK, Karuturi KM, Yu Q. Apoptosissignal-regulating kinase 1 is a direct target of E2F1 and contributes tohistone deacetylase inhibitor-induced apoptosis through positive feed-back regulation of E2F1 apoptotic activity. J Biol Chem2006;281:10508–15.

32. Milton AH, Khaire N, Ingram L, O’Donnell AJ, La Thangue NB. 14-3-3 proteins integrate E2F activity with the DNA damage response.EMBO J 2006;25:1046–57.

33. Henrique R, Jer�onimo C, Hoque MO, Carvalho AL, Oliveira J, Teix-eira MR, Lopes C, Sidransky D. Frequent 14-3-3 sigma promotermethylation in benign and malignant prostate lesions. DNA Cell Biol2005;24:264–9.

34. Lee MH, Lozano G. Regulation of the p53-MDM2 pathway by 14-3-3sigma and other proteins. Semin Cancer Biol 2006;16:225–34.

35. Bansal P, Lazo JS. Induction of Cdc25B regulates cell cycle resump-tion after genotoxic stress. Cancer Res 2007;67:3356–63.

36. Ngan ES, Hashimoto Y, Ma ZQ, Tsai MJ, Tsai SY. Overexpression ofCdc25B, an androgen receptor coactivator, in prostate cancer. Onco-gene 2003;22:734–9.

37. Davis JN, Wojno KJ, Daignault S, Hofer MD, Kuefer R, Rubin MA,Day ML. Elevated E2F1 inhibits transcription of the androgen recep-


tor in metastatic hormone-resistant prostate cancer. Cancer Res2006;66:11897–906.

38. Liao G, Chen LY, Zhang A, Godavarthy A, Xia F, Ghosh JC, Li H,Chen JD. Regulation of androgen receptor activity by the nuclearreceptor corepressor SMRT. J Biol Chem 2003;278:5052–61.

39. Nakajima H, Brindle PK, Handa M, Ihle JN. Functional interaction ofSTAT5 and nuclear receptor co-repressor SMRT: implications innegative regulation of STAT5-dependent transcription. EMBO J2001;20:6836–44.

40. Carsol JL, Gingras S, Simard J. Synergistic action of prolactin (PRL)and androgen on PRL-inducible protein gene expression in humanbreast cancer cells: a unique model for functional cooperationbetween signal transducer and activator of transcription-5 and andro-gen receptor. Mol Endocrinol 2002;16:1696–710.

41. Calzone L, Gelay A, Zinovyev A, Radvanyl F, Barillot E. A compre-hensive modular map of molecular interactions in RB/E2F pathway.Mol Syst Biol 2008;4:173.

42. Mallik I, Davila M, Tapia T, Schanen B, Chakrabarti R. Androgenregulates Cdc6 transcription through interactions between androgen

receptor and E2F transcription factor in prostate cancer cells. BiochimBiophys Acta 2008;1783:1737–44.

43. Libertini SJ, Tepper CG, Guadalupe M, Lu Y, Asmuth DM, MudryjM. E2F1 expression in LNCaP prostate cancer cells deregulatesandrogen dependent growth, suppresses differentiation, and enhancesapoptosis. Prostate 2006;66:70–81.

44. Slee EA, Lu X. The ASPP family: deciding between life and death af-ter DNA damage. Toxicol Lett 2003;139:81–7.

45. Lee EC, Zhan P, Schallhom R, Packman K, Tenniswood M. Antian-drogen-induced cell death in LNCaP human prostate cancer cells.Cell Death Differ 2003;10:761–71.

46. Paramio JM, Navarro M, Segrelles C, Gomez-Casero E, Jorcano JL.PTEN tumour suppressor is linked to the cell cycle control throughthe retinoblastoma protein. Oncogene 1999;18:7462–8.

47. Mazurek N, Sun YJ, Liu KF, Gilcrease MZ, Schober W, Nangia-Mak-ker P, Raz A, Bresalier RS. Phosphorylated galectin-3 mediates tumornecrosis factor-related apoptosis-inducing ligand signaling by regulat-ing phosphatase and tensin homologue deleted on chromosome 10 inhuman breast carcinoma cells. J Biol Chem 2007;282:21337–48.


Annexin-A7 protects normal prostate cells and induces distinct patterns of RB-associated cytotoxicity in androgen-sensitive and -resistant prostate cancer cells

Documents