Prostate cancer is characterized by epigenetic ...

Oncogene (2004), 1–8

& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $30.00 www.nature.com/onc

ORIGINAL PAPER

Prostate cancer is characterized by epigenetic silencing of 14-3-3r expression Dimitri Lodygin1, Joachim Diebold2 and Heiko Hermeking*,1 1 Molecular Oncology, Max-Planck-Institute of Biochemistry, Am Klopferspitz 18, D-82152 Martinsried/Munich, Germany; 2Institute of Pathology, Thalkirchner Strasse 36, Ludwig-Maximilians University, D-80337 Munich, Germany

In order to identify tumor suppressive genes silenced by CpG methylation in prostate carcinoma (PCa), we determined genome-wide expression changes after pharmacological reversal of CpG methylation-mediated transcriptional repression in three PCa cell lines using microarray analysis. Thereby, epigenetic silencing of the 14-3-3r gene was detected in the cell line LNCaP. 14-3-3r encodes a p53-regulated inhibitor of cell cycle progression. Laser microdissection was used to isolate different cell types present in diseased prostatic tissue. Subsequent methylation-specific PCR analysis showed CpG methylation of 14-3-3r in all 41 primary PCa samples analysed, which was accompanied by a decrease or loss of 14-3-3r protein expression. In contrast, normal prostate epithelial and benign prostate hyperplasia cells showed high levels of 14-3-3r expression. PCa-precursor lesions (prostatic intraepithelial neoplasia) also displayed decreased levels of 14-3-3r expression in luminal cells, which are known to contain shortened telomeres. RNA interference-mediated inactivation of 14-3-3r compromised a DNA damageinduced G2/M arrest in the PCa cell line PC3. The generality of CpG methylation and downregulation of 143-3r expression in PCa suggests that it significantly contributes to the formation of PCa, potentially by allowing the escape from a DNA damage-induced arrest elicited by telomere shortening. Oncogene advance online publication, 18 October 2004; doi:10.1038/sj.onc.1208004 Keywords: epigenetic silencing; CpG methylation; tumor suppressor; p53; 14-3-3s; prostate cancer

Introduction Prostate cancer (PCa) is the most commonly diagnosed malignancy in the male population of the Western world. For the year 2004, it is estimated that PCa will be diagnosed in 230 110 men and cause death in 29 900 cases in the US (Jemal et al., 2004). Small prostatic carcinomas are detected in 29% of men between 30 and 40 years of age and in 64% of men from 60 to 70 years *Correspondence: H Hermeking; E-mail: [email protected] Received 19 April 2004; revised 28 May 2004; accepted 17 June 2004

of age (Sakr et al., 1994). This usually indolent disease behaves aggressively in 25% of the affected men and accounts for B10% of all male cancer deaths, second only to lung cancer (Jemal et al., 2004). The molecular basis of PCa is still poorly understood, particularly due to the extreme heterogeneity of primary tumors. Silencing of tumor suppressive genes by CpG methylation is presumably of equal importance for tumor development as functional inactivation by point mutation or allelic loss (Jones and Baylin, 2002; Feinberg and Tycko, 2004). For example, in tumors with LOH at the p16/INK4A locus, the remaining allele is inactivated by either point mutation or CpG methylation. Aberrantly methylated CpG dinucleotides are bound by the methylCpG binding proteins MBD1–4 and MeCP2, which recruit histone deacetylases (HDACs) to the respective promoters and thereby generate transcriptionally inactive chromatin. How the CpG methylation of tumor suppressive genes is established during tumor progression is still unclear, but stochastic, age-associated accumulation of aberrantly methylated CpG sites may be involved in this process (Jones and Baylin, 2002). Experimental reversion of CpG methylation leads to reexpression of silenced genes in tumor cells, which may have profound consequences at the cellular level: for example, restored sensitivity to apoptotic stimuli after reactivation of caspase-8 (Hopkins-Donaldson et al., 2003) and Apaf-1 (Soengas et al., 2001) or inhibition of cell proliferation after re-expression of the CDK inhibitor p16 (Rhee et al., 2002). Consistent with a role of gene repression mediated by CpG methylation in tumorigenesis, inactivation of MBD2 suppresses intestinal tumorigenesis in Apcmin mice (Sansom et al., 2003). In PCa, CpG hypermethylation and subsequent loss of expression have been reported for a limited number of genes: for example, GSTP1 is silenced in B90%, RASSF1A in B63% and RARb2 in B79% of primary PCa (Nakayama et al., 2001; Singal et al., 2001; Kuzmin et al., 2002). CpG methylation of GSTP1 has proven useful for diagnosis of PCa cells in biopsies and body fluids (Harden et al., 2003). Silencing of hypermethylated genes depends on both, CpG methylation and subsequent recruitment of MBDassociated HDAC activity (Cameron et al., 1999). Therefore, optimal re-expression of genes silenced by CpG methylation is detectable after simultaneous inhibition of DNA methyltransferase and HDAC

CpG methylation of 14-3-3r in prostate cancer D Lodygin et al

2

activity by 5-aza-20 -deoxycytidine (5Aza-20 dC) and trichostatin A (TSA), respectively. The genes re-expressed after this treatment can be identified in a genome-wide manner by microarray analysis (Suzuki et al., 2002; Yamashita et al., 2002). By using this combined approach, we identified 14-33s as a gene that shows silencing by CpG methylation in PCa. The requirement of 14-3-3s for maintenance of a stable DNA damage-induced cell cycle arrest suggests that silencing of 14-3-3s expression contributes to the escape of hyperproliferative prostate epithelial cells from a cell cycle arrest elicited by telomere erosion or other DNA-damaging insults that may occur during PCa progression.

a

CpG 1 2

ATG

3 4

5

6 7

8 9 +170

+64

non-methylation specific amplicon (U)

+ −

+ +

M U

M U

b

TSA 5Aza-2‘dC

- GSTP1 - 14-3-3

− −

Results

+ −

+ +

TSA 5Aza-2‘dC

- GSTP1

Identification of 14-3-3s in a genome-wide screen for CpG methylation

Oncogene

+169

+65

c

In order to identify silenced tumor suppressive genes involved in PCa formation, we optimized the conditions for maximal re-expression of CpG-methylated genes in several PCa cell lines after a combined treatment with 5Aza-20 dC and TSA (data not shown). The three PCa cell lines PC3, LNCaP and Du-145, which are derived from metastatic prostate carcinomas, were selected for further analysis since they showed the most effective demethylation as detected by methylation-specific PCR (MSP) analysis. For subsequent microarray analysis, RNA was isolated from these cell lines after exposure to 5Aza-20 dC for 72 h and TSA for the last 24 h or, as a control, to TSA for 24 h. In parallel, genomic DNA (gDNA) was isolated from all states, to confirm efficient demethylation of CpG dinucleotides (data not shown). RNA was converted to biotinylated cRNA and hybridized to oligonucleotide arrays representing B22 000 individual transcripts. A large number of genes were found to be induced after treatment with 5Aza-20 dC plus TSA vs TSA alone (Lodygin et al., 2002). GSTP1, a gene known to be hypermethylated in the majority of PCa (Singal et al., 2001), was induced 1.87-fold in LNCaP cells after combined treatment with 5Aza-20 dC and TSA. The demethylation and induction of GSTP1 were confirmed by MSP and RT–PCR analysis (Figure 1b and c). Among the re-expressed genes was 14-3-3s, which was induced 4.3-fold as determined by microarray analysis. As shown in Figure 1a–c, the demethylation and induction of the 14-3-3s gene were confirmed by independent methods. The CpG-methylation status of 14-3-3s was examined by MSP in primary prostate epithelial cells (PrECs), in a panel of six PCa cell lines and in the BPH-1 cell line established from a benign prostate hyperplasia (BPH) by introduction of SV40 large T antigen (Figure 2a). This analysis revealed that 14-3-3s is also CpG methylated in the PCa cell lines PPC1 and LAPC4, but not in the BPH-1 cell line and in PrECs. By Western blot analysis, an inverse correlation between the degree of CpG methylation and protein expression was

methylation specific amplicon (M)

- 14-3-3 - EF1 Figure 1 Detection of reversible silencing of 14-3-3s in the PCa cell line LNCaP. (a) Location of the PCR primers used for MSP analysis of 14-3-3s. (b) MSP analysis after pharmacological reversion of epigenetic silencing in the PCa cell line LNCaP. Cells were treated with 1 mM 5Aza-20 dC for 72 h and with 300 nM TSA for the last 24 h or with TSA for 24 h. ‘M’-labelled lanes correspond to CpG-methylated and ‘U’-labelled lanes to unmethylated 14-3-3s or GSTP1 alleles. (c) mRNA expression of 14-3-3s and GSTP1 was analysed by semiquantitative RT–PCR after removal of CpG methylation. As a loading control and expression standard, the RT–PCR products of EF1a mRNA are shown. For details, see Materials and methods

a

PrEC M

U

Du-145

LNCaP

M

M

U

U

PC3 M

PPC1 U

M

U

TSU-Pr1 M

U

LAPC.4 M

U

BPH-1 M

U

14-3-3 σ

b

PrEC

Du-145

LNCaP

PC3

PPC1 TSU-Pr1

- 14-3-3σ - α -tubulin

Figure 2 CpG-methylation status and protein expression of 14-33s in prostatic cell lines. (a) 14-3-3s-specific MSP analysis of PrECs and PCa and BPH-1 cell lines. gDNA was isolated from exponentially proliferating cells, treated with sodium bisulfite and used as a template for MSP analysis with primers specific for the methylated (M) and unmethylated (U) 14-3-3s allele. BPH-1: benign prostate hyperplasia cells immortalized with SV40 large T antigen; PrECs: human primary prostate epithelial cells. (b) Detection of 14-3-3s protein expression. Western blot analysis was performed with extracts from exponentially growing cells using an affinity-purified rabbit 14-3-3s antiserum. Detection of atubulin was employed as a loading control. For details, see Materials and methods

identified (Figure 2b): LNCaP cells, which display complete CpG methylation of 14-3-3s, were devoid of 14-3-3s expression; PPC1 cells, which have


3

CpG-methylated and unmethylated 14-3-3s alleles, showed a significant downregulation of 14-3-3s protein expression. The cell lines Du-145, PC3 and TSU-Pr1 did not reveal any CpG methylation in the 14-3-3s gene and showed relatively high levels of 14-3-3s protein expression. The highest level of 14-3-3s expression was detected in the PrECs, which lack CpG methylation of the 14-3-3s gene. Interestingly, the cell lines Du-145 and PC3 harbor p53 mutations, whereas LNCaP cells express wild-type p53. This correlation suggests that silencing of 14-3-3s may potentially alleviate the requirement to inactivate p53 in PCa. Analysis of 14-3-3s silencing in prostatic tissue sections The in vivo CpG-methylation status of 14-3-3s was determined in primary PCa samples obtained after radical (37 cases) and transurethral (four cases) prostatectomy. Resected prostate tissue containing PCa also includes a number of other cell types, such as stromal cells, infiltrating T cells and prostate epithelial cells of normal glands. Since these cells are in close proximity to or overlap with areas of PCa, they would obscure the analysis of PCa-specific CpG methylation in case DNA is extracted from larger areas surrounding the cancer tissue. Therefore, laser microdissection was employed to specifically isolate primary PCa cells from paraffinembedded sections obtained from 41 patients. gDNA obtained from the isolated cells was subjected to MSP analysis. In the PCa cells, 14-3-3s showed CpG methylation at medium to high degrees (Figure 3a): in three cases, only the methylated allele was detected and, with the exception of one case, the PCR product representing the methylated allele was more prominent than the PCR product specific for the unmethylated allele. From prostatic paraffin sections of 36 PCa patients, enough adjacent epithelial cells with normal appearing morphology could be isolated by laser microdissection to allow MSP analysis. The four transurethral and one radical prostatectomy specimens did not yield enough gDNA for MSP analysis. The adjacent, normal prostate epithelial cells did not show CpG methylation of 14-3-3s in 31 cases (86.1% (31 of 36 cases); Figure 3b). Only five cases (pn1, pn19, pn23, pn25, pn30) showed CpG methylation of 14-3-3s in normal prostate epithelial, which may be caused by isolation of PCa-precursor cells displaying a normal morphology. In three cases, PCR products with aberrant sizes not corresponding to CpG-methylated 14-3-3s were detected (pn6, pn14, pn26). CpG methylation of 14-3-3s was previously detected in histologically normal epithelial cells adjacent to cancer cells in other types of cancer and may represent precursor lesions (Umbricht et al., 2001 ; Lodygin et al., 2003). In stromal cells of the prostate, we found a prominent CpG methylation of 14-3-3s, which underscores the necessity of laser microdissection for these analyses (Figure 3c). Furthermore, hyperproliferative prostate epithelial cells obtained from five patients with BPH, which did not present PCa and were in the same age group as the 41 PCa patients, were subjected to MSP analysis. The 14-3-

3s promoter was not CpG methylated in prostate epithelial cells in four cases of BPH (Figure 3d). In a fifth BPH specimen, a weak signal representing a CpGmethylated 14-3-3s allele was detected (Figure 3d). Loss of 14-3-3s protein expression in PCa and its precursor, PIN In order to determine whether CpG methylation of the 14-3-3s gene affects the expression of the corresponding gene product in vivo, the level of 14-3-3s protein expression was determined by immunohistochemistry in tissue sections of the prostate. PCa samples from 41 different patients were analysed with an affinity-purified antibody specific for the 14-3-3s protein. In normal basal and luminal prostate epithelial cells and in prostate epithelial cells representing BPH and atrophic lesions, an intense, cytoplasmic staining for 14-3-3s protein was detected (Figure 4a–f). The expression of 14-3-3s protein was downregulated markedly (450%) in neoplastic cells and glands of 26 PCa samples. Representative examples are shown in Figure 4a and b. A total of 12 specimens showed a moderate reduction (examples in Figure 4d and e) and four cases showed a minor decrease in 14-3-3s protein-specific staining (data not shown). We also detected downregulation of 14-3-3s protein expression in PIN (prostatic intraepithelial neoplasia) lesions, which represent precursors of PCa (Figure 4e and f). Loss of 14-3-3s expression was restricted to luminal cells in the PIN lesions, whereas basal cells stained positive for 14-3-3s. At present, it is unclear whether the downregulation of 14-3-3s expression in PIN is due to CpG methylation, since the amount of gDNA obtained after microdissection of PIN lesions was not sufficient to perform MSP analyses (data not shown). From other types of cancer, it is known that CpG methylation of 14-3-3s is often an early event during tumor progression. For example, in breast cancer it has been shown that CpG methylation of 14-3-3s occurs during the transition from atypical hyperplastic lesions to carcinoma in situ (Umbricht et al., 2001). The loss of 14-3-3s protein expression in PIN lesions indicates that 14-3-3s inactivation is an early event during PCa progression. Downregulation of 14-3-3s by RNA interference In order to assess the functional consequences of downregulation of 14-3-3s expression in prostate epithelial cells, a retroviral vector expressing a short RNA hairpin specifically directed against the open reading frame of 14-3-3s was generated (see Materials and methods for details). Two independent PC3 cell lines (PC3-sh-s1 and PC3-sh-s4) stably expressing this construct showed significant downregulation of 14-3-3s protein (Figure 5a). After treatment with DNA-damaging agents, PC3-sh-s1 and PC3-sh-s4 cells were unable to stably arrest, whereas PC3 cells transfected with the empty vector showed a stable arrest in the G2/M phase (Figure 5b). The PC3 cell lines showed a significant amount of cells with a 42N DNA content. The peak at Oncogene


4

a

b

c

d

pc01 M U

pc02 M U

pc03 M U

pc04 M U

pc05 M U

pc06 M U

pc07 M U

pc08 M U

pc09 M U

pc10 M U

pc11 M U

pc12 M U

pc13 M U

pc14 M U

pc15 M U

pc16 M U

pc17 M U

pc18 M U

pc19 M U

pc20 M U

pc29 M U

pc30 M U

pc21 M U

pc22 M U

pc23 M U

pc24 M U

pc25 M U

pc26 M U

pc27 pc28 M U M U

pc31 M U

pc32 M U

pc33 M U

pc34 M U

pc35 M U

pc36 M U

pc37 M U

pn03 M U

pn04 M U

pn05 M U

pn06 M U

pn07 M U

pn16 M U

pn17 M U

pn26 M U

pn27 M U

pn01 M U

pn02 M U

pn11 M U

pn12 M U

pn13 M U

pn14 M U

pn21 M U

pn22 M U

pn23 M U

pn24 M U

pn15 M U

pn25 M U

pc38 M U

pn08 M U

pn19 M U

pn29 M U

pn31

pn33

pn34

pn35

pn38

pn39

pn40

pn41

M U

M U

M U

M U

M U

M U

M U

M U

str3 M U

str4 M U

str1 M U

bph1 M

U

str2 M U

bph2 M

U

bph3 M

U

bph4 M

U

pc39 M U

pc40 M U

pn09 M U

pn10 M U

pc41 M U

pn20 M U

pn30 M U

bph5 M

U

Figure 3 14-3-3s-specific MSP analysis of in vivo CpG methylation in prostatic tissue after laser microdissection. Different neoplastic and non-neoplastic cell types were isolated from 5 mm sections of archival formalin-fixed, paraffin-embedded samples using laserpressure catapulting. MSP analysis was performed on (a) PCa cells (pc01–pc41) and (b) normal prostate epithelial cells (pn01–pn41). (c) Prostatic stromal cells (str1–str4) and (d) prostate epithelial cells representing BPH (bph1–5) obtained from five BPH patients. For details, see Materials and methods

8N observed after DNA damage therefore corresponds to tetraploid cells in the G2/M phase (Figure 5b). The inability of the PC3-sh-s1 and PC3-sh-s4 cells to stably arrest was accompanied by an increased rate of apoptosis as evidenced by an elevated portion of cells with a sub-G1 DNA content (Figure 5c). Discussion Of all 14-3-3 proteins, the 14-3-3s isoform is most evidently implicated in human cancer (reviewed in Oncogene

Hermeking, 2003). After DNA damage, 14-3-3s is transcriptionally induced by p53 (Hermeking et al., 1997). Experimental inactivation of 14-3-3s in colorectal carcinoma cells leads to impairment of the G2/M cell cycle checkpoint (Chan et al., 1999) and increased genomic instability (Dhar et al., 2000). CpG hypermethylation of 14-3-3s and loss of 14-3-3s expression have been detected in a number of different types of carcinomas (reviewed in Hermeking, 2003). Recently, we found 14-3-3s inactivation by CpG methylation in 28 of 41 (68%) cases of basal cell carcinoma of the skin (Lodygin et al., 2003). The high frequency of CpG


5

a

b

PCa

PCa

N

100x

c

N

400x

d

BPH N

A 200x

e

PCa

400x

f

LC N

PIN

PCa

BC

N PIN

200x

400x

Figure 4 Analysis of 14-3-3s protein expression in primary PCa, PIN, BPH and normal prostate epithelium. Depicted are immunohistochemical detections of 14-3-3s protein (red color) in 3 mm sections of paraffin-embedded prostate tissue obtained from PCa and BPH patients. (A: atrophic prostate epithelial cells; BC: basal PIN cells; BPH: benign prostate hyperplasia; LC: luninal PIN cells; N: non-neoplastic prostate epithelial cells; PCa: prostate cancer cells; PIN: prostatic intraepithelial neoplasia). (a, b, d) Sections containing normal prostate glands and PCa; (c) section with BPH and an atrophic gland; (e, f) section with PIN, PCa and normal prostate glands. The respective magnifications are indicated in the frames ( fold). For details, see Materials and methods

methylation (100%) at the 14-3-3s locus observed in primary PCa shows that the CpG methylation initially detected in PCa cell lines did not result from prolonged in vitro passaging of PCa cell lines, but reflects a carcinoma-specific event. This notion is further supported by the finding that benign but hyperproliferative BPH shows no significant 14-3-3s silencing. The presence of CpG methylation in all PCa samples analysed may indicate an absolute requirement of 143-3s downregulation for PCa formation. The strategy used to identify 14-3-3s as a gene specifically silenced in PCa, that is, reversion of CpG methylation-mediated gene repression in PCa cell lines, clearly argues for a role of CpG methylation in the downregulation/loss of 14-33s at the level of mRNA and protein expression. The CpG methylation of 14-3-3s may be used to detect PCa in the future. In general, detection of

aberrant CpG methylation by MSP-based assays has several significant advantages for detection of cancer when compared to protein- or RNA-based tumor markers (Laird, 2003). Since 14-3-3s is also hypermethylated in a number of other carcinomas, a combination of this marker with PCa-specific CpG methylation markers may allow the development of a highly sensitive and specific diagnostic assay in the future. As changes in DNA methylation seem to occur early in carcinogenesis, this assay may be suited to detect cells or free DNA released from early PCa lesions in body fluids. A potential obstacle to the use of 14-3-3s CpG methylation as a diagnostic marker may be the presence of CpG-methylated 14-3-3s alleles in stromal cells and lymphocytes (Umbricht et al., 2001; Bhatia et al., 2003; Lodygin et al., 2003). However, preliminary studies with other CpG marker genes show that these Oncogene


6 4 σ 1 h- σ s hs m 3- PC3 3 PC PC oc

a

k

- 14-3-3σ - α -tubulin

b

PC3-sh- σ4

PC3 mock

Etoposide for: 0 hours

telomeres in PIN was restricted to luminal cells, which also show downregulation of 14-3-3s expression as detected here. It is therefore possible that telomere shortening promotes the selection of cells that have lost 14-3-3s expression due to CpG methylation. The silencing of 14-3-3s expression observed here may contribute to PCa formation by promoting the escape from the telomere checkpoint. Subsequent cell proliferation with unprotected telomeres would contribute to genetic instability and accelerate the inactivation of additional tumor suppressive genes or activation of oncogenes (Artandi et al., 2000). Materials and methods

24 hours

Cell lines and cell culture 72 hours

96 hours

2N

% of cells in sub-G1

c

4N

18 16 14 12 10 8 6 4 2 0

8N

2N

4N

8N

PC3 mock PC3-sh-sigma1 PC3-sh-sigma4

Patient material 0h

24h 48h 72h 96h tim e after addition of etoposide

120h

Figure 5 Effects of 14-3-3s downregulation on the DNA damage response. (a) Western blot analysis of 14-3-3s protein expression in PC3 cell lines stably expressing short RNA hairpins directed against 14-3-3s (PC3-sh-s1 and PC3-sh-s4) and PC3 cells harboring a pRetroSUPER vector not encoding an shRNA (PC3 mock). (b) Cell cycle distribution of PC3 cell lines differing in 14-33s expression. PC3 cells treated with the DNA-damaging agent etoposide (20 mM) for the indicated periods of time were fixed, stained with propidium iodide and DNA content was determined by FACS. (c) Representation of the sub-G1 ( ¼ apoptotic) fraction of PC3 cell lines after addition of etoposide for the indicated periods as detected by FACS. For details, see Materials and methods

cell types may not interfere with CpG methylation-based diagnostic approaches (reviewed in Laird, 2003). Previously, ectopic expression of 14-3-3s by adenoviral infection was shown to cause a G2/M arrest in PrECs (Hermeking et al., 1997). In line with an involvement of 14-3-3s in the regulation of G2/M progression, RNA interference-mediated downregulation of 14-3-3s in the PCa cell line PC3 facilitated an escape from a stable G2/M cell cycle arrest after DNA damage. During the early phases of PCa progression, a DNA damage signal is presumably generated, as telomere shortening has been detected in PIN lesions (Meeker et al., 2002). Interestingly, the shortening of Oncogene

The PCa cell lines Du-145, LNCaP, PC3, PPC1 and TSU-PrI were cultured in RPMI-1640 supplemented with 10% fetal bovine serum (FBS) and antibiotics (Invitrogen, Carlsbad, CA, USA). LAPC4 cells were cultured in RPMI-1640 with 20% FBS. Human BPH cells immortalized with SV40 large T antigen (BPH-1) were obtained from the German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany) and passaged in RPMI-1640 medium supplemented with 20% FBS, 20 ng/ml testosterone, 50 mg/ml transferrin, 50 ng/ml sodium selenite, 50 mg/ml insulin and a mixture of trace elements (Invitrogen). Human PrECs from an 18-year-old donor (Clonetics, Walkersville, MD, USA) were cultured in PrEGM (Clonetics) on collagen type I-coated dishes (BioCoat, BD Falcon).

Archival formalin-fixed, paraffin-embedded samples of primary prostate carcinoma (Gleason Sum 5–10) and cancer-free samples of prostate (BPH) were obtained from the Institute of Pathology, Ludwig-Maximilians University, Munich. The specimens were taken from consecutive cases of a single year (2001). All patients had undergone surgery at the same institution. In all cases, two board-certified pathologists agreed on the carcinoma diagnosis. Microarray analysis LNCaP, Du-145 and PC3 cells were seeded at low density 24 h before demethylation. Total RNA was isolated from cell lines using the RNAgent kit (Promega) and assessed photometrically and by agarose gel electrophoresis. Oligonucleotide microarray analyses were performed on the GeneChip Human Genome U133A according to the manufacturer’s protocol (Affymetrix). In brief, 25 mg of total RNA was reverse transcribed using an anchored oligo-dT-T7 primer and the double-stranded cDNA synthesis kit (Invitrogen). Biotinlabelled cRNA was generated by in vitro transcription using the RiboMax T7 kit (Promega), fragmented, hybridized to U133A oligonucleotide arrays (15 mg of the RNA probe per chip) and analysed with a GeneChips Scanner 3000. Genes upregulated by combined 5Aza-20 dC plus TSA treatment vs TSA alone were identified using the Microarray Suite 4.0 software (Affymetrix). RT–PCR analysis For RT–PCR analysis, 5 mg total RNA was reverse transcribed using an oligo-(dT)18 primer and SuperScriptt double-


7 stranded cDNA synthesis kit (Invitrogen) at 501C for 60 min in a total volume of 20 ml. cDNA was diluted twofold and quantified by real-time PCR with EF1a-specific primers using FastStart-DNA Master SYBR Green 1 on a LightCycler (Roche Diagnostics, GmbH, Mannheim, Germany). The cDNA was diluted to equal concentrations and used for PCR with the respective gene-specific primers. Two units of Platinum Taq polymerase (Invitrogen) was used per 20 ml reaction with 2 ml cDNA. The total reaction was analysed by agarose gel electrophoresis. The primer sequences and PCR cycle numbers for each analysed gene were as follows (at 601C annealing temperature (Ta) for each gene): 31 cycles for 14-3-3s with forward primer 50 -CCAGGCTACTT CTCCCCTC-30 and reverse primer 50 -CTGTCCAGTTCT CAGCCACA-30 ; 33 cycles for GSTP1 with forward primer 50 -TCACTCAAAGCCTCCTGCCTAT-30 and reverse primer 50 -CAGTGCCTTCACATAGTCATCC-30 ; 18 cycles for EF1a with forward primer 50 -CACACGGCTCACATTGCAT-30 and reverse primer 50 -CACGAACAGCAAAGCGACC-30 . Bisulfite treatment of gDNA gDNA was isolated by overnight incubation in a solution containing 100 mg/ml proteinase K (Sigma) and 0.1% SDS (Sigma) at 551C with subsequent phenol/chloroform extraction and isopropanol precipitation. DNA was denatured in 0.2 M NaOH for 10 min at 371C in 50 ml total volume. After addition of 30 ml of 10 mM hydroquinone (Sigma) and 520 ml of 3.5 M sodium bisulfite pH 5.0 (Sigma), the mixture was incubated for 16 h at 501C. After column purification (Qiagen, Valencia, CA, USA), DNA was incubated in 0.3 M NaOH for 5 min at room temperature. Converted DNA was ethanol precipitated and dissolved in 40 ml TE. A 2 ml portion was used for a single MSP. MSP analysis MSP analysis (Herman et al., 1996) was performed in a total volume of 20 ml using 2 units Platinum Taq per reaction and gene-specific primer sets, discriminating between methylated (M) and unmethylated (U) DNA. After 5 min denaturation at 951C, 40 PCR cycles were performed for gDNA obtained from cell lines and 45 for microdissected gDNA. Amplified fragments were separated by 8% polyacrylamide gel electrophoresis and visualized by ethidium bromide staining. Sequences of MSP primers used were as follows: for 14-3-3s M-forward 50 -TGGTAGTTTTTATGAAAGGCGTC-30 , Mreverse 50 -CCTCTAACCGCCCACCACG-30 at Ta 651C, Uforward 50 -TTATTAGAGGGTGGGGTGGATTGT-30 and U-reverse 50 -CAACCCCAAACCACAACCATAA-30 at Ta 651C; for GSTP1 M-forward 50 -GGTTTTTTCGGTTAG TTGCGCGGCG-30 , M-reverse 50 -CCAACGAAAACCTCG CGACCTCCG-30 at Ta 601C, U-forward 50 -AAAGAGG GAAAGGTTTTTTTGGTTAGTTGTGTGGTG-30 and Ureverse 50 -AAACTCCAACAAAAACCTCACAACCTCCA30 at Ta 641C (Ta ¼ annealing temperature). Laser-assisted tissue microdissection Archival specimens of primary PCa and tumor-free prostate tissue were deparaffinized in xylene and briefly stained with hematoxylin and eosin. One section was covered and used as a reference slide. Microdissection and laser-pressure catapulting were performed using a MicroBeam system (PALM, Bernried, Germany). Material obtained from 2–3 parallel sections (B103 cells) was pooled and gDNA was isolated by the proteinase K/ SDS method. Before bisulfite treatment, 1 mg of herring sperm

carrier DNA (Promega, Madison, WI, USA) was added to each sample of microdissected DNA. Immunohistochemical staining The use of an affinity-purified rabbit polyclonal, 14-3-3sspecific antiserum on paraffin-embedded sections was described previously (Lodygin et al., 2003). A 1 : 200 dilution of 14-3-3s antiserum was used in combination with the APAAP detection system (DAKO, Copenhagen, Denmark). After counterstaining with hematoxylin, the images were acquired on an Axiovert 200M microscope (Carl Zeiss Co., Oberkochen, Germany) using a DXC-390P CCD camera (Sony, Tokyo, Japan) and PALMRobo V2.1.1 software (PALM). Western blot analysis Exponentially growing cells were lysed in 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM EGTA, 10% glycerol, 1% Triton X100, 100 mM NaF, 10 mM Na4P2O7, 1 mM PMSF and 1 mM Na3VO4 (all chemicals from Sigma). Protein concentrations were determined using a Bradford assay (Bio-Rad Laboratories, Hercules, CA, USA). 80 mg of protein was diluted in 2 Laemmli buffer, separated on SDS–polyacrylamide gels, and transferred to Immobilon-P membranes (Millipore, Bedford, MA, USA). Antibodies against a-tubulin (TU-02; Santa Cruz, CA, USA; diluted 1 : 1000) and affinity-purified rabbit 14-3-3s antiserum (diluted 1 : 500) were used in combination with HRP-conjugated secondary antibodies (Promega) diluted 1 : 10 000. Signals were visualized after enhanced chemiluminescence treatment of membranes and subsequent exposure to Hyperfilm (Amersham Pharmacia Biotech, Little Chalfont, UK). Inactivation of 14-3-3s by stable RNA interference For stable expression of siRNAs, pSUPER-based constructs were used (Brummelkamp et al., 2002b). Targeting of three different 19 bp sequences was tested in transient transfection experiments by assessing the ability to downregulate expression of a 14-3-3s-eYFP fusion protein (data not shown). All target sequences did not show homology to other cDNAs in public databases. The most efficient construct targeted the sequence GGAGGCCGGGGACGCCGAG of the 14-3-3s transcript and was used here. For vector construction, two oligonucleotides (50 -CGGGATCCCCGGAGGCCGGGGACGCCGAGTTCAAGAGTCTCG-30 ; 50 -CGCTCGAGCCAAAAAGGAGGCCGGGGACGCCGAGACTCTTGA30 ) with a 12 bp overlap were annealed at 371C and subjected to five PCR cycles resulting in a double-stranded fragment, which was digested with BamHI and XhoI and ligated into the BglII and XhoI sites of pSUPER. Then entire H1-promoter-sh14-33s cassette was released from pSUPER by restriction with EcoRI and XhoI and inserted into the same restriction sites of the self-inactivating retroviral vector pRetroSUPER (Brummelkamp et al., 2002a). The resulting construct was transfected into a Phoenix A packaging cell line. pRetroSUPER was used as a control. PC3 cells were infected by incubation with viruscontaining supernatant and subjected to puromycin selection for 7 days. Single-cell clones were established by limiting dilution. DNA content analysis by flow cytometry Cells were plated in six-well plates at identical density 1 day before treatment with 20 mM etoposide in standard culture medium. At the indicated time points, cells (both adherent and floating) were harvested by trypsinization, washed once with Oncogene


8 PBS and fixed in 70% ethanol for 2 h on ice, washed once with PBS and incubated for 30 min at room temperature in PBS solution containing 50 mg/ml propidium iodide, 0.2 mg/ml RNase A and 0.1% Triton X-100 v/v (Lodygin et al., 2002). For each sample, 10 000 events were analysed with a FACScan unit (Becton Dickinson).

Acknowledgements We thank Rene Bernards and Reuven Agami for providing the pSUPER constructs, Antje Menssen and Axel Ullrich for cell lines. Work in Heiko Hermeking’s lab is supported by the Max-Planck-Society, the Rudolf-Bartling-Stiftung and the Deutsche Krebshilfe/Dr Mildred-Scheel-Stiftung.

References Artandi SE, Chang S, Lee SL, Alson S, Gottlieb GJ, Chin L and DePinho RA. (2000). Nature, 406, 641–645. Bhatia K, Siraj AK, Hussain A, Bu R and Gutierrez MI. (2003). Cancer Epidemiol. Biomarkers Prev., 12, 165–169. Brummelkamp TR, Bernards R and Agami R. (2002a). Cancer Cell, 2, 243–247. Brummelkamp TR, Bernards R and Agami R. (2002b). Science, 296, 550–553. Cameron EE, Bachman KE, Myohanen S, Herman JG and Baylin SB. (1999). Nat. Genet., 21, 103–107. Chan TA, Hermeking H, Lengauer C, Kinzler KW and Vogelstein B. (1999). Nature, 401, 616–620. Dhar S, Squire JA, Hande MP, Wellinger RJ and Pandita TK. (2000). Mol. Cell. Biol., 20, 7764–7772. Feinberg AP and Tycko B. (2004). Nat. Rev. Cancer, 4, 143– 153. Harden SV, Sanderson H, Goodman SN, Partin AA, Walsh PC, Epstein JI and Sidransky D. (2003). J. Natl. Cancer Inst., 95, 1634–1637. Herman JG, Graff JR, Myohanen S, Nelkin BD and Baylin SB. (1996). Proc. Natl. Acad. Sci. USA, 93, 9821–9826. Hermeking H. (2003). Nat. Rev. Cancer, 3, 931–943. Hermeking H, Lengauer C, Polyak K, He TC, Zhang L, Thiagalingam S, Kinzler KW and Vogelstein B. (1997). Mol. Cell, 1, 3–11. Hopkins-Donaldson S, Ziegler A, Kurtz S, Bigosch C, Kandioler D, Ludwig C, Zangemeister-Wittke U and Stahel R. (2003). Cell Death Differ., 10, 356–364. Jemal A, Tiwari RC, Murray T, Ghafoor A, Samuels A, Ward E, Feuer EJ and Thun MJ. (2004). CA Cancer J. Clin., 54, 8–29. Jones PA and Baylin SB. (2002). Nat. Rev. Genet., 3, 415–428. Kuzmin I, Gillespie JW, Protopopov A, Geil L, Dreijerink K, Yang Y, Vocke CD, Duh FM, Zabarovsky E, Minna JD,

Oncogene

Rhim JS, Emmert-Buck MR, Linehan WM and Lerman MI. (2002). Cancer Res., 62, 3498–3502. Laird PW. (2003). Nat. Rev. Cancer, 3, 253–266. Lodygin D, Menssen A and Hermeking H. (2002). J. Clin. Invest., 110, 1717–1727. Lodygin D, Yazdi AS, Sander CA, Herzinger T and Hermeking H. (2003). Oncogene, 22, 5519–5524. Meeker AK, Hicks JL, Platz EA, March GE, Bennett CJ, Delannoy MJ and De Marzo AM. (2002). Cancer Res., 62, 6405–6409. Nakayama T, Watanabe M, Yamanaka M, Hirokawa Y, Suzuki H, Ito H, Yatani R and Shiraishi T. (2001). Lab. Invest., 81, 1049–1057. Rhee I, Bachman KE, Park BH, Jair KW, Yen RW, Schuebel KE, Cui H, Feinberg AP, Lengauer C, Kinzler KW, Baylin SB and Vogelstein B. (2002). Nature, 416, 552–556. Sakr WA, Grignon DJ, Crissman JD, Heilbrun LK, Cassin BJ, Pontes JJ and Haas GP. (1994). In Vivo, 8, 439–443. Sansom OJ, Berger J, Bishop SM, Hendrich B, Bird A and Clarke AR. (2003). Nat. Genet., 34, 145–147. Singal R, van Wert J and Bashambu M. (2001). Cancer Res., 61, 4820–4826. Soengas MS, Capodieci P, Polsky D, Mora J, Esteller M, Opitz-Araya X, McCombie R, Herman JG, Gerald WL, Lazebnik YA, Cordon-Cardo C and Lowe SW. (2001). Nature, 409, 207–211. Suzuki H, Gabrielson E, Chen W, Anbazhagan R, van Engeland M, Weijenberg MP, Herman JG and Baylin SB. (2002). Nat. Genet., 31, 141–149. Umbricht CB, Evron E, Gabrielson E, Ferguson A, Marks J and Sukumar S. (2001). Oncogene, 20, 3348–3353. Yamashita K, Upadhyay S, Osada M, Hoque MO, Xiao Y, Mori M, Sato F, Meltzer SJ and Sidransky D. (2002). Cancer Cell, 2, 485–495.