Regulation of Protein Kinase B activity by PTEN and SHIP2 in human prostate-derived cell lines

Cellular Signalling 19 (2007) 129 – 138 www.elsevier.com/locate/cellsig

Regulation of Protein Kinase B activity by PTEN and SHIP2 in human prostate-derived cell lines R. Michael Sharrard ⁎, Norman J. Maitland YCR Cancer Research Unit, Department of Biology, University of York, Heslington, York YO10 5DD, UK Received 17 May 2006; accepted 29 May 2006 Available online 7 June 2006

Abstract Protein Kinase B (PKB/Akt) is a key regulator of cell proliferation, motility and survival. The activation status of PKB is regulated by phosphatidylinositol 3-kinase (PI3K) via the synthesis of phosphatidylinositol-3,4,5-trisphosphate (PI(3,4,5)P3, PIP3). PTEN antagonises PI3K by degrading PIP3 to phosphatidylinositol-4,5-bisphosphate (PI(4,5)P2). Deregulation of PKB through loss of functional PTEN has frequently been implicated in the progression of tumours, including prostate cancer, and the PTEN-negative prostate cancer cell lines LNCaP and PC3 have been widely used as models for this mechanism of constitutive PKB activation. However, other enzymes in addition to PTEN can antagonise PI3K, including SHIP2, which degrades PIP3 to phosphatidylinositol-3,4-bisphosphate (PI(3,4)P2). We investigated the role of PTEN and SHIP2 in the regulation of PKB phosphorylation in a panel of human prostate-derived epithelial cell lines. In the PTEN-positive prostate-derived cell lines PNT2, PNT1a and P4E6, PI3K inhibition by LY294002 caused rapid dephosphorylation of PKB at ser473 (T1/2 < 2 min), leading to its inactivation. In the PTEN-null line LNCaP, LY294002-induced PKB dephosphorylation was much slower (T1/2 > 20 min), but in PC3 cells (also PTEN-null) it was only slightly slower than in PTEN-positive cells (T1/2 = 3 min). PKB dephosphorylation paralleled loss of plasma membrane PIP3. PNT1a, P4E6 and PC3, but not PNT2 or LNCaP, expressed SHIP2. SiRNA-mediated knockdown of SHIP2 expression markedly slowed PKB inactivation in response to LY294002 in PC3 but not in other SHIP2-positive cells, whereas knockdown of PTEN expression in PNT2, PNT1a and P4E6 resulted in higher steady-state levels of PKB phosphorylation and slowed, but did not prevent, LY294002-induced PKB inactivation. Thus SHIP2 substitutes for PTEN in the acute regulation of PKB in PC3 cells but not other prostate cell lines, where PTEN may share this role with further PIP3-degrading mechanisms. © 2006 Elsevier Inc. All rights reserved. Keywords: Protein Kinase B; PTEN; SHIP2; Phosphatidylinositol 3-kinase; Prostate; Cancer; Cell line

1. Introduction Cell growth and survival in normal epithelial cells is regulated by signals from growth factors and from cell–cell and cell–matrix contact. Cell dependence upon these signals allows control of tissue growth and maintenance of epithelial tissue architecture. Tumour progression to invasiveness and metastasis characteristically involves loss of dependence upon such signals. The PI3K–PKB pathway, activated in response to growth factors and adhesion to matrix or other cells, is central to

⁎ Corresponding author. Tel.: +44 1904 328708; fax: +44 1904 328710. E-mail addresses: [email protected] (R.M. Sharrard), [email protected] (N.J. Maitland). 0898-6568/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2006.05.029

signalling pathways that regulate these processes (see Ref. [1,2] for reviews). Activated PI3K converts PI(4,5)P2 to PI(3,4,5)P3 (PIP3), which recruits PKB to the cell membrane and allows phosphatidylinositol-dependent kinase 1 (PDK1) and a second kinase (termed PDK2, though not yet conclusively identified) to phosphorylate and activate PKB at thr308 and ser473 respectively [3,4]. Activation of PKB in turn leads to the phosphorylation of a number of downstream targets which control proliferation and survival, including glycogen synthase kinase 3 (GSK3), the proapoptotic protein BAD, mTOR (via TSC2) and MDM2/HDM2. PKB activation also regulates p27Kip1 and the forkhead family of transcription factors [5]. The PI3K–PKB pathway is deregulated in many tumour types, and constitutively raised levels of activated PKB have been implicated in enhanced proliferation and inappropriate cell

130

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

survival independently of growth factor stimulation and normal requirements for cell–cell and cell–matrix contact. The PTEN tumour-suppressor gene is frequently inactivated in a wide variety of tumour types; PTEN activity is characteristically lost (or reduced by haploinsufficiency) in late-stage tumours and has been associated with the development of invasiveness and metastatic capacity ([6]; see also review, [7]). The PTEN gene product is a multiple-specificity phosphatase that antagonises PI3K by degrading PI(3,4,5)P3 back to PI(4,5)P2 [8] in addition to its ability to dephosphorylate protein targets such as focal adhesion kinase (FAK) [9]. Many reports have implicated reduction or loss of PTEN activity as causal in the constitutive activation of PKB in tumour cells. However, a second group of enzymes, the SHIP (SH2-containing inositolphosphatase) family, has been identified as also being potentially important in regulating PKB through degradation of PI(3,4,5)P3 to PI(3,4)P2 [10,11]. SHIP1 expression is largely confined to the haematopoietic system and sSHIP to stem cells, but SHIP2 is ubiquitously expressed [12,13] and may act as an alternative or additional mechanism for antagonising PI3K both in the presence and the absence of PTEN. Although the PH domain of PKB can bind to PI (3,4)P2, the demonstration that SHIP2 as well as PTEN can downregulate PKB activation indicates that PI(3,4,5)P3 rather than either PI(3,4)P2 or PI(4,5)P2 is the critical phospholipid involved in assembling the PKB activation complex at the plasma membrane [14]. PTEN inactivation has been widely implicated in progression to metastasis of prostate cancer, with approximately 60% of latestage prostate tumours showing loss of PTEN function [15,16]. Haploinsufficiency of the PTEN gene has been shown to promote prostate cancer progression in a transgenic mouse model [17]. In one of the first papers describing PTEN, two of the model cell lines for metastatic prostate disease, LNCaP and PC3, were shown to lack expression of the protein: LNCaP has a two-base pair deletion in codon 6, while PC3 cells have a deletion at the 3′ end of the gene [18,19]. These cell lines have been widely used as models for the behaviour of PTEN-null cells; in particular, the high level of activated PKB observed in LNCaP and PC3 cells has been ascribed to PIP3 accumulation resulting from failure to degrade this compound due to absence of the lipid phosphatase activity of PTEN [20,21]. Deregulation of the PI3K–PKB signalling axis can clearly contribute to survival and deregulated proliferation in tumorigenesis and cancer metastasis [1,2]. However, the significance of loss of PTEN in promoting this deregulation will depend on whether PTEN is the only lipid phosphatase responsible for the negative control of this pathway in a given cell type, or whether additional PIP3-degrading enzymes function to regulate PKB activity alongside PTEN or in its absence. The aim of the present study is to determine whether PTEN shares the role of antagonist to PI3K in the regulation of PKB phosphorylation with other PIP3-degrading enzymes, including SHIP2, in cell lines derived from human prostatic epithelium. We also investigate the extent to which alternative mechanisms for the breakdown of PIP3 can substitute for PTEN in regulating PKB in the PTEN-null model prostate cancer cell lines LNCaP and PC3.

2. Materials and methods 2.1. Cell lines PNT1a and PNT2 [22,23] are non-tumorigenic epithelial cell lines derived by SV40 immortalisation of normal prostate epithelial outgrowths. P4E6 is a cell line derived from an early stage carcinoma of prostate by immortalisation with the E6 gene of human papillomavirus 16 [24]. LNCaP and PC3 are established prostatic carcinoma cell lines with androgen-dependent and androgenindependent phenotypes respectively, obtained from American Type Culture Collection. PNT1a, PNT2, P4E6 and LNCaP cells were grown in RPMI1640 containing 2 mM glutamine, 10 mM HEPES and 10% foetal calf serum (FCS). PC3 cells were grown in Ham's F12 medium containing 2 mM glutamine and 7% FCS.

2.2. Reagents and expression constructs LY294002 (Calbiochem) was prepared at 20 mM in dimethylsulfoxide (DMSO). The plasmid GRP1PH(ΔNLS)EGFP, encoding the pleckstrin homology (PH) domain of GRP1 (with the nuclear localisation signal deleted) fused to Enhanced Green Fluorescent Protein (EGFP), was a generous gift of Professor Peter Downes, Dundee University.

2.3. Western blotting Cells grown in 24-well plates were harvested into 200 μl buffer containing 1% SDS, 1% dithiothreitol, 62.5 mM tris HCl pH 6.8, 10% glycerol, 0.1 M sodium fluoride, 10 mM sodium pyrophosphate and 1mM sodium orthovanadate, heated to 100 °C for 5 min, analysed by SDS-PAGE, and blotted to PVDF membranes (Roche). After blocking in 1% Blocking Agent (Roche) in TBS containing 0.05% Tween20, the membranes were probed overnight with primary antibodies as follows: rabbit antibodies against Akt, phospho-Akt (ser473), phospho-GSK3α/β (ser21/9) and PTEN (Cell Signaling Technologies), each at 1:5000; goat anti-SHIP2 (I 20) (Santa Cruz) at 1:2000; mouse anti-pan-cytokeratin (Sigma) at 1:100,000. After washing, membranes were incubated for 60 min with 1:5000 anti-rabbit IgGperoxidase, 1:5000 anti-mouse IgG-peroxidase, or 1:20,000 anti-goat IgG -peroxidase (Sigma) as appropriate. The membranes were washed in TBSTween and detected using BM-chemiluminescence substrate (Roche) and pre-flashed Hyperfilm ECL (Amersham Biosciences). The images were scanned using a Hewlett Packard ScanJet 5370C and quantified using NIHimage software.

2.4. Polymerase chain reaction (PCR) 8 ng of DNA from each cell line was amplified in 20 μl reaction mixture containing 200 μM dNTPs, 1.5 mM MgCl2, 0.7 μM of each primer, 1 × Expand buffer and 1 unit of Expand High Fidelity DNA polymerase (Roche). The primers were ACCTGTTAAGTTTGTATGCAAC (PTEN exon 5 fwd) and TCCAGGAAGAGGAAAGGAAA (PTEN exon 5 rev). Amplification was as follows: 2 min at 94 °C; 35 cycles of 20 s 94 °C, 15 s at 58 °C and 75 s + 2 s/cycle at 72 °C; 5 min at 72 °C. Products were analysed by agarose gel electrophoresis in the presence of ethidium bromide.

2.5. Reverse transcriptase-PCR (RT-PCR) 5 μg total RNA was reverse-transcribed as described previously [19]. Equal volumes of each of the cDNA preparations were PCR-amplified as above, using the primers 5′-CGCTCTGGCTCCACCAGCATT-3′ and 5′-TGCGGTCGTGCG TGAAGGTGG-3′ to target the core sequence (bases 430–1090) of the SHIP2 open reading frame. RNA quality and efficiency of cDNA synthesis were tested by amplifying equal volumes of each of the cDNA preparations for 16 cycles using the primers 5′-AAGGTGAAGGTCGGAGTCAA-3′ and 5′-GGACACGGAAGGC CATGCCA-3′, which target a 700 bp product from the cDNA of the GAPDH gene. Gel analysis confirmed that equal amounts of GAPDH product were amplified from all cDNAs.

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

131

3. Results 3.1. Sensitivity of prostatic epithelial cell lines to serum deprivation and PI3K inhibition

Fig. 1. PKB phosphorylation at ser473 in five prostate cell lines in response to PI3K inhibition or serum withdrawal and replacement. Cells were treated as follows: maintained continuously in medium plus serum (see Materials and methods) (lane 1); treated with DMSO vehicle only for 3 h (lane 2) or with 10 μM LY294002 (PI3K inhibitor) for 15 min (lane 3), 1 h (lane 4) or 3 h (lane 5); incubated overnight in medium without serum, and then harvested immediately (lane 6) or at 15 min (Lane 7), 1 h (Lane 8), or 3 h (lane 9) after refeeding with medium plus serum. Analysis was by Western blotting and probing for PKB phosphorylated at ser473. The results shown are typical of at least three independent experiments.

PNT2 and PNT1a (non-tumour prostate epithelium) and P4E6 (early prostate tumour) cells express wild-type PTEN [19]. The prostate tumour cell lines LNCaP (from lymph node metastasis) and PC3 (from bone metastasis) lack wild-type PTEN [18,19]. When grown in medium containing serum, all of these cells showed PKB activation as evidenced by ser473 phosphorylation (Fig. 1, lane 1). In PNT2, PNT1a, P4E6 and PC3, treatment with the PI3K inhibitor LY294002 (10 μM) for 15 min resulted in almost complete dephosphorylation of PKB at

2.6. 3′-rapid amplification of cDNA ends (3′-RACE) for alternativelyspliced PTEN transcripts 10 μg of RNA from PC3 cells was primed with 0.5 μg L1-dT primer (5′GACTCGAGTCGACATCGATTTTTTTTTTTTTTT-3′) and reverse-transcribed as above. The cDNA was PCR-amplified for 35 cycles using L1 primer (5′-GACTCGAGTCGACATCGA-3′) and a primer from within exon 1 (5′GAGGATGGATCGACTTAGA-3′), exon 2 (5′-CCAAACATTATTGCTATGGGA-3′) or exon 5 (5′-ACAAGAGGCCCTAGATTTCT-3′) of PTEN. The products were cloned into pGEM T-easy (Promega) and the inserts sequenced from flanking T7 and SP6 primer sites.

2.7. Transfection 5 × 104 cells were plated into 1 cm2 tissue culture wells on a glass coverslip. After adhesion, the cells were transfected with 0.5 μg/well of GRP1PH(ΔNLS) EGFP using Fugene 6 (Roche) according to the manufacturer's instructions.

2.8. Confocal microscopy of transfected cells 24–48 h after the start of transfection, the cells were observed using a Zeiss LSM 510 meta inverting confocal microscope fitted with a stage-mounted tissue culture chamber maintained at 37 °C with a humidified atmosphere of 5% CO2 – 95% air. After equilibration, images of EGFP-expressing cells were captured using the argon laser and a 512 nm detection filter. The medium was then removed from individual wells and replaced with CO2-equilibrated, pre-warmed medium containing 10 μM LY294002. Further images were then captured at timed intervals.

2.9. Small inhibitory RNA (siRNA) treatment of cells SmartPool siRNAs (Dharmacon) were used to knock down expression of PTEN and SHIP2 (INPPL1). Cells plated at 105/well in 48-wells were washed with serum-free medium, and 125 μl serum-free medium containing 1 μl of Oligofectamine (InVitrogen) mixed with 25 pmol of PTEN-directed siRNA, 25 pmol of SHIP2-directed siRNA, or 25 pmol of each siRNA was added to each well. Mock-transfected cells received Oligofectamine only. After 4 h, 1 ml of medium containing serum was added and the cells returned to the incubator for 68 h. The cells were then treated for 5 min with medium containing either 10 μM LY294002 or an equivalent amount of DMSO before harvesting and analysis by Western blotting.

Fig. 2. Kinetics of dephosphorylation of PKB and GSK3α following inhibition of PI3K. Cells were grown in medium containing serum. Triplicate wells were treated for 0, 2, 5, 10, 15 or 20 min with 10 μM LY294002 before harvesting and analysis by SDS-polyacrylamide gel electrophoresis, Western blotting, and probing for ser473-phosphorylated PKB and for ser21-phosphorylated GSK3α. The blots were then stripped and reprobed for total PKB protein. The band intensities were quantified (see Materials and methods) and the ratio of ser473phosphorylated PKB (A) and ser21-phosphorylated GSK3α (B) to total PKB for each lane was determined and expressed as a percentage of the ratio at time = 0. The mean and s.d. (n = 3) for each time point are shown. □, PNT2; ΔPNT1a; P4E6; ○, LNCaP; , PC3.

▪

132

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

ser473, while a similar level of PKB dephosphorylation was achieved in LNCaP cells only after 3 h (Fig. 1, lanes 3–5). While complete dephosphorylation of PKB in the presence of LY294002 was maintained in P4E6 and PC3 cells, a low level of PKB rephosphorylation was observed in PNT2 and PNT1a cells over 3 h. This phenomenon has been shown to result from the activation of a form of PI3K that has a higher IC50 for LY294002 (Sharrard RM, Spalton JC and Maitland NJ, manuscript in preparation). Serum deprivation resulted in complete loss of PKB ser473-phosphorylation in PNT2, PNT1a and P4E6; in contrast, PC3 cells showed only partial loss of PKB phosphorylation in the absence of serum, and LNCaP were unaffected (Fig. 1, lane 6). Refeeding with medium containing serum resulted in rapid rephosphorylation of PKB (Fig. 1, lanes 7–9). We investigated the time-course of PKB dephosphorylation at ser473 in the five cell lines in response to PI3K inhibition by LY294002 (i.e. in the absence of synthesis of PI(3,4,5)P3 from PI(4,5)P2). PNT2, PNT1a and P4E6 showed rapid loss of PKB phosphorylation following exposure to the drug (T1/2 < 2 min) while the response of LNCaP cells was considerably slower (T1/2 > 20 min)(Fig. 2A). This is consistent with LNCaP harbouring a defective mechanism for degrading PIP3, as indicated by the results in Fig. 1. In PC3 cells, however, the rate of PKB dephosphorylation in the presence of LY294002 (T1/2 = 3 min) was

only slightly slower than in the PTEN-positive PNT2 and P4E6 cells. The phosphorylation status of the PKB substrate GSK3α was found to follow PKB phosphorylation for all five cell lines, indicating that loss of PKB ser473 phosphorylation resulted in loss of its catalytic activity (Fig. 2B). 3.2. Kinetics of loss of PIP3 from the plasma membrane in response to LY294002 The slow kinetics of dephosphorylation of PKB in LY294002-treated LNCaP cells could be due to slow degradation of PIP3 or could result from a low activity of protein phosphatases that dephosphorylate ser473 of PKB in these cells. We therefore investigated whether the loss of PKB phosphorylation paralleled loss of PIP3 in cells treated with LY294002. Cells were transfected with a plasmid encoding the PH-domain of GRP1 fused to EGFP. The GRP1 PH-domain shows high specificity of binding for PIP3 and this construct is thus a useful tool for tracking the location of PIP3 in live cells [25]. Confocal microscopy of transfected cells showed an accumulation of EGFP at the plasma membrane, indicating the presence of PIP3 (Fig. 3). Cells were treated with 10 μM LY294002 and monitored for redistribution of EGFP. In PNT2, PNT1a, P4E6 and PC3 cells (Fig. 3a–c and e), EGFP signal was clearly depleted from the membrane within 5 min of

Fig. 3. Loss of PI(3,4,5)P3 from the cell membrane following inhibition of PI3K. Cells transfected with GRP1PH(ΔNLS)-EGFP expression plasmid were examined by confocal microscopy. For each cell line, a complete Z-series of a field showing representative cells was captured and used to create a three-dimensional projection to demonstrate the overall distribution of signal in the different cell types (labelled ‘3D’). Strong EGFP signal was seen at areas of cell–cell contact in PNT2, PNT1a, P4E6 and LNCaP cells, but in PC3 cells the highest concentration of EGFP was in areas of motility-associated ruffled membrane and was independent of cell contact. Individual transfected cells from each line were observed before (0′) and at the indicated times (in minutes) after addition of 10 μM LY294002 to the medium. EGFP fluorescence at the cell surface indicates the presence of PIP3. A complete Z-series of images was acquired at each time point to ensure that alterations in the intensity of membrane-associated fluorescence were not due to changes in the shape of the cells or redistribution of the signal to restricted areas within the plane of the membrane. In each case the image at the centre of the Z-series is shown. The results are typical of at least three independent experiments. (a) PNT2. (b) PNT1a. (c) P4E6. (d) LNCaP; note the persistence of membrane-associated signal for between 60 and 90 min (arrowheads) and an area of relocation of signal within the plane of the membrane (0–30 min) before loss of signal from the cell surface between 30 and 90 min (thin arrows). (e) PC3; note the accumulation of signal at the ruffled membranes (arrowed) seen in the 3D projection of the PC3 cell.

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

133

treatment with 10 μM LY294002, indicating rapid loss of PIP3. However, in LNCaP cells a similar degree of loss was only achieved after 90 min (Fig. 3d), consistent with much slower kinetics of loss of membrane PIP3 after PI3K inhibition. For each time-point, a Z-series of the entire cell under examination was captured to confirm loss of the signal from the plasma membrane rather than redistribution within the plane of the membrane. The time-course of loss of PIP3 from the plasma membrane thus parallels the kinetics of loss of PKB phosphorylation in response to LY294002 treatment (see Fig. 2), and the slow decrease in PKB phosphorylation in LNCaP cells results from slow PIP3 degradation rather than lack of protein phosphatases. 3.3. Levels of PI3K activity in the prostate cell lines The slower kinetics of loss of PIP3 in LY294002-treated LNCaP cells compared to the other cell lines might result from low levels of enzymes that degrade PIP3 or from higher rates of PIP3 synthesis in LNCaP cells. We tested the five cell lines for the rate of PKB rephosphorylation after treatment with LY294002 and its washout (Fig. 4). The differences in rephosphorylation rates between the cell lines were relatively small, with the highest rate (in PNT1a) being only 1.66 times the lowest (PNT2); LNCaP cells showed a rate intermediate between these (1.29 time the rate of PNT2). Thus the slow rate of loss of PIP3 in LY294002-treated LNCaP is not due to higher rates of PIP3 synthesis in these cells.

Fig. 5. Expression of SHIP1, SHIP2 and PTEN in prostate cell lines. (a) Western blotting for the expression of SHIP1. The first lane contains an extract of THP1 cells as a positive control for detection of SHIP1. (b) Western blotting for expression of SHIP2 in the five prostate cell lines. (c) Western blotting for PTEN expression. Samples are the same as in (b). (d) RT-PCR for the core sequence of SHIP2 (bases 430–1090 of the open reading frame). The lane at the right contains a DNA ladder, demonstrating that the major PCR product has the expected size of 660 bp. The sizes of protein molecular weight markers are as indicated to the right of panels (a), (b) and (c).

3.4. PC3 cells contain exons 1 and 2, but not exon 5, of PTEN

Fig. 4. Kinetics of rephosphorylation of PKB after removal of the PI3K inhibitor LY294002. Cells from each line were plated into 24-wells at 105 cells per well and treated with 10 μM LY294002 in serum-free medium for 165 min. The cells were then placed in medium containing serum and 10 μM LY294002 and the incubation continued for a further 15 min. Three wells of each cell line were then harvested immediately and three were washed twice with medium plus serum without LY294002 before being incubated in drug-free medium for 5 min followed by harvesting. Analysis of PKB phosphorylation at ser473 was by Western blotting and densitometry as before. Values for levels of phosphorylated PKB/105 cells for the five cell lines (mean ± s.d., n = 3) before (−) or 5 min after (+) LY294002 washout were calculated relative to the quantity found in PNT2 cells after washout of the drug.

PC3 cells possess a mechanism for degrading PIP3 that is almost as efficient as that of PTEN-expressing cells (Fig. 2A). We previously showed that PC3 cells contain exon 1, but not exon 9, of PTEN and that human cells can express two alternatively-spliced forms of PTEN, designated PTEN-Δ and PTEN-B, which contain exon 5 (encoding the phosphatase domain) but not exon 9 [26]. PC3 cells might thus express a cryptic form of PTEN activity encoded by the partially-deleted PTEN gene. We carried out 3′-RACE analysis of PC3 RNA using primers from exons 1, 2 and 5. 3′-RACE downstream of exon 1 showed that PC3 cells express a sequence that contains exons 1 and 2 of PTEN fused to material encoded by DNA from chromosome 1 (results not shown). No clones could be derived by 3′-RACE using an upstream primer from exon 5, and genomic PCR using intron-directed primers flanking exon 5 failed to detect this exon in PC3, whereas positive results were obtained from PNT2, PNT1a, P4E6 and LNCaP cells. We thus concluded that PC3 cells do not express an active phosphatase derived from residual segments of the PTEN gene. 3.5. PTEN, SHIP1 and SHIP2 expression in prostate cell lines We investigated whether one or more of the PIP3-degrading SHIP enzymes, which convert PI(3,4,5)P3 to PI(3,4)P2, might

134

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

contribute to inactivation of PKB following inhibition of PI3K in the prostate cell lines. SHIP1 protein was not detectable by Western blotting in any of the prostate cell lines, although it is

clearly detected in the positive control cell line THP1 (Fig. 5a). In contrast, Western blotting for SHIP2 showed that the 160 kD SHIP2 protein is expressed in PNT1a, P4E6 and PC3 cells, but

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

135

Fig. 6. Effects of siRNA-mediated knockdown of expression of PTEN and SHIP2 on the rate of loss of PKB phosphorylation following PI3K inhibition. Cells were mock-transfected or transfected with siRNA directed against PTEN, SHIP2 or both molecules as described in Materials and methods. After 72 h, the cells were treated for 5 min with either 10 μM LY294002 or an equivalent amount of vehicle (DMSO) control. The cells were then harvested, analysed by Western blotting, and probed sequentially for ser473-phosphorylated PKB, PTEN, SHIP2, and pan-cytokeratin. Each experiment was repeated on three separate occasions and each time gave the same pattern of results. The graphs shown here are representative of these data. For each cell line, the results of probing the Western blots for PTEN and for SHIP2 are shown in the panels at the upper left. Band intensities for phosphorylated PKB and cytokeratin were quantified as described in Materials and methods. The ratio of ser473-phosphorylated PKB to cytokeratin (S473/CK) for each lane was calculated and normalised against the S473/CK ratio for mock-transfected cells treated with DMSO to give the values for ‘relative PKB ser473 phosphorylation’ shown in histogram (a) for each cell line. Then, for each siRNA treatment (mock, PTEN, SHIP2, and PTEN + SHIP2), the ratio of S473/CK after treatment with LY294002 to S473/CK after treatment with DMSO was calculated to obtain the ‘PKB ser473 phosphorylation ratio (+LY294002/control)’ shown in histogram (b) for each cell line.

not in PNT2 or LNCaP cells (Fig. 5b). We also confirmed that PTEN protein expression was detectable in PNT2, PNT1a and P4E6 cells but not LNCaP or PC3 cells (Fig. 5c). Interestingly, RT-PCR for the core sequence of SHIP2 demonstrated its expression as mRNA in all five cell lines (Fig. 5d), suggesting that the coding sequence of the expressed mRNA is either defective or subject to stringent translational control in PNT2 and LNCaP. In RT-PCR using primers for the complete open reading frame of SHIP2, LNCaP cells gave an abnormally short

band (results not shown), suggesting that the gene is either partially deleted or abnormally spliced. 3.6. Effects of siRNA-mediated knockdown of expression of PTEN and SHIP2 In order to study the function of PTEN and SHIP2 in regulating the PI3K–PKB pathway in the different prostate cell lines, we used siRNA directed against these proteins to knock

136

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

down their endogenous expression, either singly or together (Fig. 6). Western blot analysis of mock-transfected and siRNAtransfected cells after 72 h demonstrated the effectiveness of this treatment (Fig. 6, inset panels). Densitometry of the blots showed that the average knockdown of PTEN expression was 88.4% in PNT2, 96.3% in PNT1a and 89.6% in P4E6; the average knockdown of SHIP2 expression was 92.5% in PNT1a, 85.2% in P4E6 and 85.9% in PC3. The effects of PTEN and SHIP2 expression knockdown upon the extent of dephosphorylation of PKB ser473 in response to a 5-minute treatment with 10 μM LY294002 were measured by Western blot analysis and probing for ser473phosphorylated PKB. To account for variations in growth between cells exposed to the different siRNA treatments over 72 h, the blots were also probed for cytokeratin, which we have found to be a reliable indicator of cell number in all the cell lines used in these experiments, and the values for the ratio of phosphorylated PKB to cytokeratin were calculated for each sample. In PNT2, PNT1a and P4E6 cells, knockdown of PTEN resulted in an increase in PKB phosphorylation before treatment with LY294002 (Fig. 6, histograms Aa, Ba and Ca). This represents a steady-state level influenced by the rate of both synthesis and degradation of PIP3, so the result demonstrates that PTEN acts as a negative regulator of PKB activation in these cells. The extent of the effect of PTEN knockdown on the steady-state level of PKB phosphorylation varies between these three lines in the rank order PNT2 > PNT1a > P4E6, reflecting their relative expression levels of PTEN (see Fig. 5). In these cell lines, PTEN knockdown also results in increased amounts of ser473phosphorylated PKB remaining after 5 min of PI3K inhibition with LY294002 compared to LY294002-treated mock-transfected cells. However, the proportion of PKB remaining phosphorylated after LY294002 treatment of PTEN-depleted cells is still quite low (Fig. 6, histograms Ab, Bb and Cb: 12% in PNT2, 20% in PNT1a, 52% in P4E6), indicating that these cells still retain the capacity to degrade PIP3 rapidly under conditions of reduced PTEN expression. In contrast, treatment of LNCaP and PC3 cells with siRNA directed against PTEN showed little effect upon either the steady-state level of PKB phosphorylation or the rate of dephosphorylation in the presence of LY294002 (Fig. 6D and E), providing further evidence that PTEN is not involved in PKB regulation in these cells. SiRNA treatment directed against SHIP2 expression showed different effects on PKB phosphorylation between the cell lines. In PNT2, siRNA against SHIP2 alone resulted in a decrease in the steady-state level of PKB phosphorylation (resulting in an anomalously high ratio of phosphorylated PKB in LY294002treated cells relative to controls in histogram Ab); on the other hand, SHIP2-directed siRNA did not affect steady-state levels of PKB or rates of LY294002-induced PKB dephosphorylation when PTEN expression was also inhibited. PNT2 cells may express a modified form of SHIP2, not detected as a 160 kD band on Western blots (see Fig. 5), that has little or no PIP3degrading activity but which interferes with the action of PTEN. In PNT1a cells, knockdown of SHIP2 alone had little effect on the steady-state or LY294002-inhibited levels of PKB

phosphorylation; however, the rate of loss of PKB phosphorylation in the presence of LY294002 was slower when both proteins were inhibited compared to when only PTEN was knocked down, suggesting that SHIP2 may partially substitute for PTEN in its absence in these cells. A third pattern was seen in P4E6 cells; here knockdown of SHIP2 alone reduced PKB phosphorylation and increased its rate of loss on treatment with LY294002, but also had the same effect on PTEN-reduced cells (i.e. when both siRNAs were used, compared to PTEN siRNA alone). This indicates a negative effect of SHIP2 on PIP3 degradation that may operate via other proteins than PTEN. SiRNA against SHIP2 appeared to have a small effect on LNCaP cells (Fig. 6D), raising the possibility that these cells also express a modified, undetected form of the protein with low-level PIP3-degrading activity; however, SHIP2 siRNA treatment did not enhance PKB dephosphorylation when LNCaP were treated with LY294002. In contrast, knockdown of SHIP2 in PC3 cells not only doubled the steady-state levels of PKB phosphorylation but dramatically reduced the extent of PKB dephosphorylation after 5 min exposure to LY294002 from 89% to 34% (Fig. 6E). These results indicate that SHIP2 can substitute for absent PTEN in the regulation of PKB phosphorylation in PC3 cells, though it cannot do so in LNCaP cells due to lack of expression. Despite efficient suppression of PTEN and (in PNT1a and P4E6) SHIP2 expression by dual siRNA treatment, PNT1a cells still lost 63%, P4E6 cells lost 66% and PNT2 lost 90% of their PKB phosphorylation within 5 min of PI3K inhibition by LY294002. In contrast, LNCaP cells lost only 4%, and PC3 cells 34%, of their PKB phosphorylation under parallel conditions (Fig. 6D, E). Although the residual levels of PTEN and SHIP2 after knockdown will undoubtedly continue to participate in PIP3 degradation, the results suggest that PTEN and SHIP2 may not be the only enzymes which antagonise PI3K in the cell lines PNT2, PNT1a and P4E6. 4. Discussion The cell lines LNCaP and PC3 have been extensively used as model systems of metastatic prostate cancer and as examples of PTEN-null cells. Indeed, it is often assumed that PKB in these cell lines is constitutively activated as a result of PTEN abrogation and concomitant effects of deregulated synthesis of PIP3. In this paper we show clear differences between LNCaP and PC3 cells in terms of their regulation of the PI3K/PKB pathway. While LNCaP cells show a genuine reduction in ability to degrade PIP3 compared to PTEN-expressing prostatic lines, SHIP2 activity in PC3 cells at least partially compensates for the absence of PTEN, resulting in the potential for regulated rather than constitutive PKB activation in these cells. Identification of SHIP2 as a regulator of the PI3K–PKB pathway in PC3 cells calls for a reassessment of this cell line as a model for PTENnull prostate cancer. It is well established that the mechanisms that activate this pathway upstream of PI3K differ between LNCaP and PC3 [21,27]. The present report underlines the need to consider these two widely-used model cell lines for metastatic prostate cancer as separate, and very different, examples of

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

the disruption of signalling pathways regulating proliferation and survival that may accompany loss of the PTEN gene. In the PTEN-positive prostate cell lines investigated, SHIP2 appears to play a minor role in regulating PI3K activation of PKB when cells are maintained in the continuous presence of serum. This is consistent with the findings of Blero et al. [28] who showed that SHIP2 regulates PIP3 levels and PKB phosphorylation in mouse embryonic fibroblasts only after short-term (5–10 min) stimulation with serum. They suggested that SHIP2 might only be able to degrade PIP3 while PTEN is inactivated by serum-induced reactive oxygen species in these cells. In the PTEN-positive prostatic epithelial cells studied here, siRNA-mediated knockdown of PTEN was not compensated to any great extent by SHIP2, suggesting that activation of additional factors, possibly controlling recruitment of SHIP2 to the plasma membrane [14], may be required for SHIP2 to replace PTEN as the main PIP3-degrading enzyme. However, we also found evidence that a third activity, possibly but not necessarily also a PIP3-degrading phosphatase, shares the ability to downregulate PKB when PI3K is inhibited. This is especially clear in PNT2 cells, which lack expression of SHIP2. These cells lose almost 90% of their PKB phosphorylation after 5 min of PI3K inhibition even when PTEN expression is reduced to less than 12% of its normal level in these cells (Fig. 6A; compare the much smaller loss of PKB phosphorylation in PI3K-inhibited LNCaP cells, Fig. 6D). If multiple PTEN-like activities can degrade PIP3 in epithelial cells, SHIP2 may have limited opportunities to be the predominant regulator of the PI3K–PKB pathway in these cells. In PC3 cells, both PTEN and the putative additional PIP3-degrading enzyme(s) may have been lost, leaving SHIP2 as the major antagonist to PI3K, while LNCaP must have lost all three PIP3-degrading activities. The identity of the additional enzyme(s) is not yet known, although various candidates exist amongst known PTEN-like phosphatases, including TPIP [29], PTEN2 [30], PLIP [31], and SKIP [32]. The expression of these proteins shows restricted tissue distribution and functional analysis, indicating that they are unlikely to represent widespread functional homologues of PTEN in prostatic epithelium [7]. The recently-described PTEN homologue C1-TEN has also been shown to downregulate PKB through its phosphatase activity [33]. What physiological role may SHIP2 play in prostatic epithelial cells? Sasaoka et al. [14] showed that SHIP2 is important in adipocytes in the regulation of the PKB isoform Akt2, while PTEN is associated with regulation of Akt1. If this mechanism were to operate in prostatic cells, we would expect increased levels of phosphorylated Akt2 in cells that lack SHIP2 or when SHIP2 expression is suppressed by siRNA. However, although all five cell lines used in this study express Akt2, only Akt1 (clearly resolved from Akt2 in our Western blotting system; Sharrard RM and Maitland NJ, manuscript in preparation) showed detectable levels of ser473 phosphorylation throughout the study. Targeting of SHIP2 to the plasma membrane is required for its efficient negative regulation of PKB [14]. Translocation to the membrane may be mediated via Shc or c-Cbl [34,35]; SHIP2 also associates with filamin and p130Cas to regulate actin-based

137

cytoskeletal activity [36,37] and modulates hepatocyte growth factor-mediated lamellipodium formation, cell scattering and cell spreading through direct interaction with c-Met protein [38]. SHIP2 may thus mediate both the PI3K–PKB signalling pathway and cell motility, and may function as an interface between these cellular processes. It will be interesting to determine whether SHIP2 has an enhanced role in regulating motility in PC3 cells, and if so whether it does so via interactions with one or more of the proteins mentioned above. Further work also is required to elucidate the role of SHIP2 and the factors involved in its regulation in normal and diseased prostate tissue, and to investigate whether experimental modulation of Shc or c-Cbl expression and activity affects the capability of SHIP2 to regulate the PI3K/PKB pathway. PTEN is also known to act as an interface between PI3K signalling pathways and cell motility. Significantly, neither LNCaP nor PC3 cells survive long-term reintroduction of PTEN expression [19]. LNCaP and PC3 have thus undergone alterations in the pathways surrounding the PI3K–PKB axis such that PTEN cannot be tolerated. These may include loss of mechanisms, including phosphorylation/dephosphorylation cycles, that control the lipid or protein phosphatase activities of PTEN protein. Davies et al. [20] demonstrated that PTEN induced growth inhibition but only moderately increased apoptosis in LNCaP cells. We found that PTEN re-expression in LNCaP and PC3 induced changes in motility, adhesion and spreading, leading to loss of cells from culture through detachment from the growth surface, with apoptosis (where it occurred) being a secondary event [19]. These findings indicate a major effect of re-expressed PTEN in these cells on cytoskeletal functions, possibly involving both protein and lipid phosphatase activities of PTEN as well as structural functions residing in its C-terminal segment or tensin homology domains. The data here presented suggest that loss of PTEN may have less effect on the PI3K–PKB pathway than originally proposed [8] if the cells maintain alternative PIP3degrading mechanisms. Conversely, changes in the regulation of adhesion and motility which LNCaP and PC3 cells have acquired to compensate for the absence of PTEN protein-phosphatase activity (or protein– plus lipid–phosphatase activity) may be responsible for the inability of these cells to tolerate reintroduction of PTEN (discussed in Ref. [19]). The role of multiple or redundant pathways for the degradation of PIP3 must be considered in assessing the role of PTEN loss in deregulating pathways downstream of PI3K in tumorigenesis. Different enzymes that share this function with PTEN may link regulation of the PI3K–PKB pathway with other cellular functions, in the same way that PTEN may act to integrate this pathway with cytoskeletal activity, motility, and cell adhesion. Alterations in the activity of the different PIP3degrading enzymes would then result in different cellular phenotypes combined with altered kinetics of PKB activation. These differences could profoundly affect the course of tumour progression. The presence of multiple PIP3-regulating enzymes also provides multiple targets for therapeutic targeting based upon the specific phenotype of each tumour. Analysis of the pattern of PIP3-regulating enzymes, in combination with mapping of pathways regulated by PIP3 availability, in individual

138

R.M. Sharrard, N.J. Maitland / Cellular Signalling 19 (2007) 129–138

tumours would thus supply valuable information for disease prognosis and provide a basis for improved design and selective administration of therapeutic agents. In conclusion, our results indicate that PTEN is not the only PIP3-degrading mechanism that antagonises PI3K in the regulation of PKB in non-tumour and early tumour-derived human prostatic cells. Knockdown of PTEN expression in these cells resulted in elevated steady-state levels of PKB phosphorylation, but did not prevent LY294002-induced PKB inactivation, indicating the presence of alternative enzymes for removal of PIP3. SHIP2 protein, which also degrades PIP3, can substitute for PTEN as an acute regulator of the PI3K–PKB pathway in PTEN-null cells. The cell lines LNCaP and PC3, which have been widely used for over two decades both as benchmark models of metastatic prostate cancer and as examples of PTENdeficient cells, exhibit entirely different regulation of the PI3K– PKB pathway due to the functional replacement of PTEN by SHIP2 in PC3 cells. In contrast, in cells that normally co-express PTEN and SHIP2, the latter has a more limited role, suggesting that its function in these cells is circumscribed by opportunities for interaction with location-directing or activity-modulating proteins. Acknowledgements This work was supported by Yorkshire Cancer Research (Harrogate, United Kingdom) and the United Kingdom National Cancer Research Institute. We thank Professor Peter Downes of the University of Dundee for his generous gift of the GRP1PH(ΔNLS)EGFP expression plasmid. We are also indebted to Dr. Martin Rumsby of the University of York for discussion of the results and critical reading of the manuscript. References [1] S. Brader, S.A. Eccles, Tumori 90 (2004) 2. [2] R. Parsons, Semin. Cell Dev. Biol. 15 (2004) 171. [3] D.R. Alessi, M. Andjelkovic, B. Caudwell, P. Cron, N. Morrice, P. Cohen, B.A. Hemmings, EMBO J. 15 (1996) 6541. [4] J.H. Feng, J. Park, P. Cron, D. Hess, B.A. Hemmings, J. Biol. Chem. 279 (2004) 41189. [5] M.A. Lawlor, D.R. Alessi, J. Cell Sci. 114 (2001) 2903. [6] H. Suzuki, D. Freije, D.R. Nusskern, K. Okami, P. Cairns, D. Sidransky, W.B. Isaacs, G.S. Bova, Cancer Res. 58 (1998) 204. [7] N.R. Leslie, C.P. Downes, Biochem. J. 382 (2004) 1. [8] T. Maehama, J.E. Dixon, J. Biol. Chem. 273 (1998) 13375.

[9] L.N. Zhang, Q. Yu, J.Y. He, X.L. Zha, Mol. Cell. Biochem. 262 (2004) 25. [10] A.L. Drayer, X. Pesesse, F. DeSmedt, R. Woscholski, P. Parker, C. Erneux, Biochem. Biophys. Res. Commun. 225 (1996) 243. [11] X. Pesesse, S. Deleu, F. DeSmedt, L. Drayer, C. Erneux, Biochem. Biophys. Res. Commun. 239 (1997) 697. [12] E. Muraille, X. Pesesse, C. Kuntz, C. Erneux, Biochem. J. 342 (1999) 697. [13] L.M. Sly, M.J. Rauh, J. Kalesnikoff, T. Buchse, G. Krystal, Exp. Hematol. 31 (2003) 1170. [14] T. Sasaoka, T. Wada, K. Fukui, S. Murakami, H. Ishihara, R. Suzuki, K. Tobe, T. Kadowaki, M. Kobayashi, J. Biol. Chem. 279 (2004) 14835. [15] M.E. McMenamin, P. Soung, S. Perera, I. Kaplan, M. Loda, W.R. Sellers, Cancer Res. 59 (1999) 4291. [16] K.P. Porkka, T. Visakorpi, Eur. Urol. 45 (2004) 683. [17] B. Kwabi-Addo, D. Giri, K. Schmidt, K. Podsypanina, R. Parsons, N. Greenberg, M. Ittmann, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 11563. [18] J. Li, C. Yen, D. Liaw, K. Podsypanina, S. Bose, S.I. Wang, J. Puc, C. Miliaresis, L. Rodgers, R. McCombie, S.H. Bigner, B.C. Giovanella, M. Ittmann, B. Tycko, H. Hibshoosh, M.H. Wigler, R. Parsons, Science 275 (1997) 1943. [19] R.M. Sharrard, N.J. Maitland, Br. J. Cancer 83 (2000) 1102. [20] M.A. Davies, D. Koul, H. Dhesi, R. Berman, T.J. McDonnell, D. McConkey, W.K.A. Yung, P.A. Steck, Cancer Res. 59 (1999) 2551. [21] M.A. Rochester, J. Riedemann, G.O. Hellawell, S.F. Brewster, V.M. Macaulay, Cancer Gene Ther. 12 (2005) 90. [22] O. Cussenot, P. Berthon, A. Faille, R. Berger, I. Mowszowicz, P. Teillac, A. LeDuc, F. Calvo, J. Urol. 143 (1991) 881. [23] P. Berthon, O. Cussenot, L. Hopwood, A. Le Duc, N.J. Maitland, Int. J. Oncol. 6 (1995) 333. [24] N.J. Maitland, C.A. Macintosh, J. Hall, M. Sharrard, G. Quinn, S. Lang, Radiat. Res. 155 (2001) 133. [25] A. Gray, J. Van der Kaay, C.P. Downes, Biochem. J. 344 (1999) 929. [26] R.M. Sharrard, N.J. Maitland, Biochim. Biophys. Acta N, Gene Struct. Expr. 1494 (2000) 282. [27] D.Q. Zheng, A.S. Woodard, G. Tallini, L.R. Languino, J. Biol. Chem. 275 (2000) 24565. [28] D. Blero, J. Zhang, X. Pesesse, B. Payrastre, J.E. Dumont, S. Schurmans, C. Erneux, FEBS J. 272 (2005) 2512. [29] S.M. Walker, C.P. Downes, N.R. Leslie, Biochem. J. 360 (2001) 277. [30] Y. Wu, D. Dowbenko, M.T. Pisabarro, L. Dillard-Telm, H. Koeppen, L.A. Lasky, J. Biol. Chem. 276 (2001) 21745. [31] D.J. Pagliarini, C.A. Worby, J.E. Dixon, J. Biol. Chem. 279 (2004) 38590. [32] T. Ijuin, T. Takenawa, Mol. Cell. Biol. 23 (2003) 1209. [33] S. Hafizi, F. Ibraimi, B. Dahlbäck, FASEB J. 19 (2005) NIL433. [34] H. Ishihara, T. Sasaoka, M. Ishiki, T. Wada, H. Hori, S. Kagawa, M. Kobayashi, Mol. Endocrinol. 16 (2002) 2371. [35] I. Vandenbroere, N. Paternotte, J.E. Dumont, C. Erneux, I. Pirson, Biochem. Biophys. Res. Commun. 300 (2003) 494. [36] J.M. Dyson, C.J. O'Malley, J. Becanovic, A.D. Munday, M.C. Berndt, I.D. Coghill, H.H. Nandurkar, L.M. Ooms, C.A. Mitchell, J. Cell Biol. 155 (2001) 1065. [37] N.K. Prasad, S.J. Decker, J. Biol. Chem. 280 (2005) 13129. [38] A. Koch, A. Mancini, O. El Bounkari, T. Tamura, Oncogene 24 (2005) 3436.