Biochemical Journal

Research article

Phosphoinositide phosphatase SHIP-1 regulates apoptosis induced by edelfosine, Fas ligation and DNA damage in mouse lymphoma cells

Maaike C. Alderliesten , Jeffrey B. Klarenbeek , Arnold H. van der Luit , Menno van Lummel , David R. Jones , Shuraila Zerp , Nullin Divecha , Marcel Verheij , Wim J. van Blitterswijk

Abstract

S49 mouse lymphoma cells undergo apoptosis in response to the ALP (alkyl-lysophospholipid) edelfosine (1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine), FasL (Fas ligand) and DNA damage. S49 cells made resistant to ALP (S49AR) are defective in sphingomyelin synthesis and ALP uptake, and also have acquired resistance to FasL and DNA damage. However, these cells can be re-sensitized following prolonged culturing in the absence of ALP. The resistant cells show sustained ERK (extracellular-signal-regulated kinase)/Akt activity, consistent with enhanced survival signalling. In search of a common mediator of the observed cross-resistance, we found that S49AR cells lacked the PtdIns(3,4,5)P3 phosphatase SHIP-1 [SH2 (Src homology 2)-domain-containing inositol phosphatase 1], a known regulator of the Akt survival pathway. Re-sensitization of the S49AR cells restored SHIP-1 expression as well as phosphoinositide and sphingomyelin levels. Knockdown of SHIP-1 mimicked the S49AR phenotype in terms of apoptosis cross-resistance, sphingomyelin deficiency and altered phosphoinositide levels. Collectively, the results of the present study suggest that SHIP-1 collaborates with sphingomyelin synthase to regulate lymphoma cell death irrespective of the nature of the apoptotic stimulus.

  • alkyl-lysophospholipid
  • apoptosis resistance
  • DNA damage
  • Fas/CD95
  • SH2 (Src homology 2)-domain-containing inositol phosphatase (SHIP)

INTRODUCTION

Synthetic alkylphospholipids such as edelfosine [1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine; an ALP (alkyl-lysophospholipid)] are promising new anticancer agents that act on lipid-dependent signal-transducing enzymes in tumour cell membranes rather than the DNA [14]. They easily insert into the lipid bilayer, accumulate in lipid rafts and persist due to poor degradation [57]. Interference with lipid metabolism and signal transduction leads to apoptosis, which is selective for tumour cells [1,3,8]. The therapeutic value of these compounds appeared most prominent in combination with conventional DNA-damaging regimens, such as γ-radiation [1,3,911].

We previously reported that, in S49 mouse lymphoma cells, ALP is internalized by raft-dependent endocytosis and inhibits the biosynthesis of major phospholipids, which induces apoptosis [6,7]. A variant cell line, S49AR, made resistant to ALP, showed impaired uptake of ALP and related alkylphospholipids [6,7]. As a result, they are not only resistant to apoptosis by these lipid compounds but, strikingly, also by ligation of the FasL (Fas ligand) death-receptor [12]. Intriguingly, S49AR cells were deficient in the major raft lipid SM (sphingomyelin), owing to complete down-regulation of SMS1 (SM synthase-1) [13]. Although siRNA (small interfering RNA)-induced down-regulation of SMS1 in S49 cells mimicked apoptosis resistance, reconstitution of SMS1 in S49AR cells failed to restore apoptosis sensitivity, which led to the conclusion that SMS1 deficiency is important, but not sufficient, to explain the apoptosis resistance to ALP and FasL [12]. The present paper reports on a second determinant of resistance and on the observation that S49AR cells are also cross-resistant to DNA damage.

Resistance to apoptosis is often accompanied by up-regulation of survival signalling via ERK (extracellular-signal-regulated kinase) and Akt [1]. This notion was also confirmed in the present study in our resistant S49AR cells. Using pharmacological inhibitors, we checked whether activity of these kinases was essential for apoptosis resistance, but found that this was only the case for resistance to DNA damage, not to ALP or FasL. Since Akt is therefore not the common and sole determinant of cross-resistance to the various apoptotic stimuli, we looked upstream of this protein kinase and focused on phosphoinositides. SHIP-1 [SH2 (Src homology 2)-domain-containing inositol phosphatase 1] is an important regulator of survival signalling [1416]. It is mainly expressed in haemopoietic cells where it converts PtdIns(3,4,5)P3 into PtdIns(3,4)P2, and is as such a negative regulator of the PI3K (phosphoinositide 3-kinase)/Akt signalling pathway [15]. Next to the catalytic (phosphatase) domain, SHIP-1 contains a number of structural motifs that allow physical interaction with other signalling proteins, such as a SH2 domain, proline-rich motifs that can bind SH3-containing proteins and NPXY sites that, when phosphorylated on the tyrosine residue, bind proteins containing a PTB (phosphotyrosine-binding) domain, for example Shc [14,15].

In the present paper we report that SHIP-1 is down-regulated in apoptosis-resistant S49AR cells and that, accordingly, the PtdIns(3,4,5)P3 level is increased, consistent with increased phosphorylation and thus activation of Akt and ERK1/2. Spontaneous re-sensitization of S49AR cells (yielding S49ARS cells) leads to regained expression of SHIP-1 and normalized phosphoinositide levels, comparable with S49 cells. Knocking-down SHIP-1 by siRNA mimics the apoptosis resistance to DNA damage and FasL and, partly, the resistance to ALP. The results of the present study indicate that absence of SHIP-1 negatively regulates apoptosis induced by multiple stimuli, leading to multi-stress-resistant lymphoma cells.

EXPERIMENTAL

Materials

The ALP edelfosine (Et-18-OCH3) was purchased from BioMol. [3H]Edelfosine (58 Ci/mmol) was synthesized by Moravek Biochemicals. [1-3H]Sphingosine was synthesized by Piet Weber (DSM, Delft, The Netherlands). Etoposide was from Sigma–Aldrich. Tween-20 and Silica 60 TLC plates were from Merck. Anti-Fas monoclonal antibody 7C10 was from CAMPRO Scientific BV. Protein G–Sepharose fast-flow beads were from Amersham Biosciences; soluble recombinant human FASL (APO-1L) was from Alexis Biochemicals. Rabbit anti-β-actin, anti-pSer473-Akt, anti-pThr308-Akt, anti-Akt, anti-MAPK (mitogen-activated protein kinase), anti-PTEN (phosphatase and tensin homologue deleted on chromosome 10), LY294002, rapamycin and U0126 were from Cell Signaling Technology. AKTi1/2 and z-VAD-FMK (benzyloxycarbonyl-Val-Ala-DL-Asp-fluoromethylketone) were from Calbiochem. Anti-Fas monoclonal antibody Jo2 (hamster IgG2) was from BD Biosciences Pharmingen. Rabbit anti-SHIP-1 and rabbit anti-p85 PI3K were from Upstate Biotechnology. Rabbit anti-SHIP-2 was from Stem Cell Technology and mouse anti-pMAPK and mouse anti-tubulin were from Sigma. Rabbit anti-mouse immunoglobulin conjugated to HRP (horseradish peroxidase) and swine anti-(rabbit immunoglobulin)–HRP were from Dako A/S.

Cells and culture conditions

Mouse S49 (S49.1 lymphoma) cells were cultivated in DMEM (Dulbecco's modified Eagle's medium; Gibco-Invitrogen), containing high levels of glucose and pyruvate, supplemented with 8% fetal bovine serum, 2 mM L-glutamine and antibiotics. ALP-resistant variants (S49AR) were isolated in two selection rounds of growth in 15 μM ALP (edelfosine) for 72 h, followed by plating in semi-solid medium and isolation of colonies of surviving cells [17]. S49AR cells could be grown continuously in 15 μM ALP. All experiments with S49AR cells were performed with cells grown without the selection agent for at least 1 week. Spontaneous resensitization of S49AR cells towards S49ARS cells occurred gradually after 4–5 weeks of culturing in the absence of ALP [12].

Plasmids, quantitative PCR and RNAi (RNA interference)

SMS1 was down-regulated in S49 cells by retroviral transduction of siRNAs, yielding S49siSMS1 cells, as described previously [13]. Control cells (S49mock) were transduced with scrambled siRNA. Retroviral transduction of HA (haemagglutinin)-tagged SMS1 in S49AR cells, yielding AR-SMS1 cells, has been described previously [12].

To knock down SHIP-1, S49 cells were retrovirally transduced by siRNA, yielding S49siSHIP cells. SiRNA directed against SHIP-1, 5′-TAAGTTCTACAGCCACAAA-3′, was inserted into the retroviral vector pRETRO-SUPER. The following siRNA primers were used: sense 5′-GATCCCCCAAGTTCTACAGCCACAAATTCAAGAGATTTGTGGCTGTAGAACTTGTTTTTGGAAA-3′; and anti-sense 5′-AGCTTTTCCAAAAACAAGTTCTACAGCCACAAATCTCTTGAATTTGTGGCTGTAGAACTTGGGG-3′. Retroviral supernatants were obtained from Phoenix cells and used to transduce S49 cells. Stable S49siSHIP cells were selected with puromycin.

Immunoblotting and immunoprecipitation

S49 cells were washed once with PBS and lysed in lysis buffer [10 mM Tris/HCl, pH 7.8, 150 mM NaCl, 5 mM EDTA, 1% (v/v) Nonidet P40, protease inhibitor cocktail (Roche), 1 mM sodium vanadate and 20 mM NaF]. Lysates were incubated for 30 min at 4°C and centrifuged to remove cellular debris (12000 g for 20 min) before being normalized for protein content. Prior to Western blot analysis, samples were heated for 10 min at 70°C in reducing SDS sample buffer from Invitrogen Life Technologies containing 1 mM dithiothreitol and run on a Novex mini-gel in NuPage Mes/SDS running buffer (Invitrogen Life Technologies). Separated proteins were transferred on to nitrocellulose membranes and blocked with 3% (w/v) BSA for 1 h in TBS-T buffer (10 mM Tris/HCl, pH 7.5, 150 mM NaCl, and 0.05% Tween 20) containing a 1:50 dilution of Roche Blocking reagent (Roche). Blots were incubated overnight at 4°C with primary antibody (1:1000 dilution), followed by incubation with HRP-conjugated rabbit anti-(rat immunoglobulin) (1:2000 dilution), and immunoreactive proteins were visualized by ECL.

Apoptosis assay and flow cytometry

Cells were seeded at 1×106 cells/ml, cultured overnight and incubated for 6 h with ALP (10 μM), FasL (500 ng/ml) or etoposide (5 μg/ml). Cells were stained with propidium iodide and the percentage of sub-diploid apoptotic nuclei was determined on a FACScan (Becton Dickinson) [6,18]. Fas surface expression was measured using flow cytometry. Cells were washed three times with 0.5% BSA in PBS containing 0.01% sodium azide. Cells were incubated with FITC-conjugated Jo2 monoclonal antibody and FACScan analysis was performed using FCS express 2.0.

Cellular uptake of ALP

Cells were grown to a density of 2.5×106/ml and [3H]edelfosine (0.2 μCi, 15 μM) was added. At given time points, samples were taken, incubated for 2 min on ice, and washed with ice-cold PBS. Samples were lysed in 0.1 M NaOH for liquid scintillation counting.

Lipid analysis

Phosphoinositides were radiolabelled in vivo. Cells were first washed three times with phosphate-free DMEM before being radiolabelled with [32P]Pi (500 μCi/ml, HCl-free, PerkinElmer) in phosphate-free DMEM for 40 min. Total cellular lipids were extracted in the presence of dilute HCl, deacylated using methylamine and separated by SAX (strong anion exchange)-HPLC with on-line radiochemical detection (Lablogic), as described in detail previously [19,20]. A portion of the deacylated lipid extract was subjected to phosphate analysis using the molybdenum blue assay spectrophotometrically at 630 nm. The amount (nmol) of phosphate was derived from an appropriate standard curve with linearity in the range 0–170 nmol. Radioactivity (c.p.m.) within peaks eluting from the HPLC was divided by the amount of lipid phosphate for normalization purposes.

Sphingolipids were labelled with [1-3H]sphingosine (1 μCi/ml, 1×106 cells/ml) for 4 h. Cells were washed and resuspended in 200 μl of PBS. Lipids were extracted with chloroform/methanol (1:2, v/v) and phase-separated using 1 M NaCl. The organic phase was washed in methanol/H2O/chloroform, 47:49:3 (by vol.), and separated by silica TLC, using chloroform/methanol/0.2% CaCl2, 60:40:9 (by vol.). Tritiated lipids were visualized after dipping the TLC plate in 12.5% diphenyloxazole dissolved in diethylether, drying and subsequent autoradiography. Lipids were identified using iodine-stained standards.

RESULTS

ALP-resistant S49 lymphoma cells (S49AR) are cross-resistant to other apoptotic stimuli

The mouse T-cell lymphoma cell line S49 was made resistant to various synthetic alkylphospholipids by continuous culture in the presence of 8 μg/ml ALP (edelfosine) for two rounds of 72 h [17]. These ALP-resistant cells (named S49AR) did not go into apoptotic cell death when exposed to ALP or structurally related ALP analogues for 6 h, whereas 70% of the wild-type S49 cells died (Figure 1A) [6,7]. ALP-induced cell death could be inhibited by z-VAD-FMK, a well-known inhibitor of caspase 3, indicating that ALP causes cell death via apoptosis (Figure 1A).

Figure 1 S49AR cells made resistant to ALP-induced apoptosis are cross-resistant to Fas ligation and DNA damage

(A) ALP-sensitive S49 cells (open bars) and ALP-resistant S49AR cells (closed bars) were left untreated (control) or were treated with ALP (edelfosine; 10 μM) with or without the caspase inhibitor z-VAD-FMK (indicated) for 6 h. (B) Cells were left untreated (control) or were treated with ALP (10 μM), FasL (500 ng/ml) or etoposide (5 μg/ml) for 6 h, or were γ-radiated (8 Gy) for 6 h. Apoptotic nuclear fragmentation was measured by FACScan analysis (see the Experimental section). Results are the means of four experiments±S.D.

Previously, we have shown that S49 cells internalize ALP by lipid raft-mediated endocytosis and that S49AR cells are deficient to do so, which would explain the resistance to ALP-induced apoptosis [6,7]. However, we found that S49AR cells are also resistant to other cell death-inducing agents, such as FasL [12] and to DNA damage by etoposide and γ-radiation (Figure 1B). Apoptosis induced by these agents, unlike ALP, was not blocked by disruption of lipid rafts [12] (results not shown), which implies that mechanism(s) other than impaired raft-dependent endocytosis should explain the cross-resistance of S49AR cells. We previously identified that down-regulation of SMS1 and the consequent lack of SM as only part of the reason why S49AR cells are resistant to ALP and FasL [12]. Our search for another common mechanism or (lack of) mediator that would fully explain this multi-stress resistance to cell death-inducing agents is described below.

Resistant S49AR cells show constitutively increased Akt and ERK1/2 activity

The protein kinases Akt and ERK1/2 are key members of the canonical survival signalling pathways. Activation of these pathways is expected to interfere with apoptosis induced by ALP and other cell-death inducers. We found that both ERK1/2 and Akt show constitutively increased phosphorylation in the S49AR cells compared with the S49 cells (Figure 2A), which is indicative of activation of these survival pathways. The increase in phosphorylation of Akt Ser473 in the resistant cells is more pronounced than that of Akt Thr308, explained by the fact that these regulatory phosphorylation sites are controlled by different kinases and phosphatases (see Discussion). To study further the role of these pathways in apoptosis resistance, we used the pharmacological inhibitors LY294002 (for PI3K), rapamycin [mTOR (mammalian target of rapamycin)], U0126 [MEK1/2 (MAPK/ERK kinase 1/2)] and AKTi (Akt1/2) in S49 and S49AR cells. The first three inhibitors did not relieve the resistance of S49AR cells to ALP- and FasL-induced cell death, whereas some minor sensitizing effect of AKTi was found when control values were subtracted (note that AKTi by itself showed 10% apoptotic toxicity in these cells). However, the PI3K inhibitor LY294002, the Akt1/2 inhibitor AKTi and, to a lesser extent, the mTOR and MEK1/2 inhibitors re-sensitized S49AR cells to the DNA- damaging agent etoposide (Figure 2B). These divergent results further suggest that no common mediator of the cross-resistance to ALP and the other cell-death inducers is to be found at, or downstream of, Akt and ERK1/2. We therefore focussed our effort to more upstream signalling molecules.

Figure 2 S49AR cells show increased Akt and ERK activity, inhibition of which restores apoptosis induced by DNA damage, but not by ALP or FasL

(A) Western blots of S49 and S49AR cell lysates using monoclonal antibodies against pSer473 and pThr308 sites of Akt, and polyclonal rabbit serum against pERK1/2. (B) Cells were treated with inhibitors of the PI3K/Akt/mTOR pathway (LY294002, 5 μM), AKTi (8 μM), rapamycin (10 nM)) and the ERK pathway (U0126, 10 μM) in the absence (control) or presence of ALP (10 μM), FasL (500 ng/ml) or etoposide (5 μg/ml) for 6 h. Apoptotic nuclear fragmentation was measured by FACScan analysis. Results are the means of four experiments±S.D.

SHIP-1 is down-regulated in S49AR cells

Since Akt and (indirectly) ERK1/2 phosphorylation and activation are regulated by PtdIns(3,4,5)P3 formation [21,22], we next measured the levels of the various phosphoinositides in the cells after [32P]Pi in vivo labelling of the ATP pool. The results in Figures 3(A)–3(C) show an increase (approximately 2-fold) in PtdIns(3,4,5)P3 levels in S49AR cells compared with S49 cells. On the other hand, PtdIns(3,4)P2 levels decreased, consistent with the presence of a 5-phosphatase activity (Figures 3B and 3C). SHIP-1 is an SH2 domain-containing PtdIns 5-phosphatase that is expressed only in haemopoietic cells. It removes the phosphate at the 5′ position of PtdIns(3,4,5)P3, thereby down-regulating the pool of PtdIns(3,4,5)P3 formed by PI3K (Figure 3D). We found that SHIP-1 is down-regulated in S49AR cells compared with S49 cells at both the protein and gene expression levels (Figures 3E and 3F). The related isoform SHIP-2 is ubiquitously expressed in virtually all tissues. Its protein expression was similar in either of the S49 cell lines. The level of the p85 subunit of PI3K was also similar in both S49 cell lines. PTEN, a PtdIns(3,4,5)P3 3-phosphatase, was not expressed in either of the S49 cell lines (Figure 3E).

Figure 3 Resistant S49AR cells are SHIP-1 deficient and show an increased ratio of PtdIns(3,4,5)P3 against PtdIns(3,4)P2

(A) HPLC profiles of 32P-labelled phosphoinositides in a typical experiment showing an increased PtdIns(3,4,5)P3 and decreased PtdIns(3,4)P2 peak from S49AR against S49 cells. Lipids were extracted using chloroform/methanol, deacylated and subjected to HPLC (see the Experimental section). Quantification of integrated peaks (percentage of total counts) from S49 and S49AR respectively: PtdIns(3,5)P2, 0.023 and 0.019; PtdIns(3,4)P2, 0.061 and 0.042; PtdIns(4,5)P2, 9.67 and 9.09; and PtdIns(3,4,5)P3, 0.0031 and 0.0057. (B) 32P-labelled phosphoinositide levels quantified relative to [32P]PtdIns (mean values±S.D., three experiments); and (C) as mean PtdIns(3,4,5)P3/PtdIns(3,4)P2 ratio from these three experiments. (D) Schematic diagram of the enzymatic interconversion of phosphoinositides. (E) Western blots of cell lysates using polyclonal rabbit sera against SHIP-1/2, PTEN and p85-PI3K, demonstrating that S49AR cells are deficient in SHIP-1. PTEN is lacking in both S49 and S49AR cells, A431 cells served as a positive control. (F) Quantitative PCR (in arbitrary units) and image of agarose gel (inset) showing decreased levels of SHIP-1 mRNA in S49AR cells.

Spontaneous re-sensitization of S49AR cells results in regained SHIP-1 expression and SM synthesis

Previously, we have shown that S49AR cells can regain their sensitivity to ALP-induced apoptosis gradually after 4–5 weeks of culturing in the absence of ALP [12]. These re-sensitized S49AR cells (S49ARS) also show regained apoptosis sensitivity towards other related alkylphospholipids (results not shown), to FasL [12] and to etoposide (Figure 4A). Interestingly, this complete re-sensitization of cells is associated with regained SM biosynthesis by SMS1 [12], as well as a return of SHIP-1 protein expression to the levels seen in parental S49 cells (Figure 4B). Accordingly, the PtdIns(3,4,5)P3 level decreased and the PtdIns(3,4)P2 level increased in these S49ARS cells, close to the original levels in S49 cells (Figures 4C and 4D). Phosphorylation of Akt at Ser473, which was increased in the S49AR cells, was lowered again in the re-sensitized S49ARS cells (Figure 4B). Thus both SHIP-1 expression and SMS1 activity are positively correlated with apoptosis sensitivity towards multiple stress stimuli and are down-regulated upon induction of cross-resistance to these stimuli.

Figure 4 Spontaneous resensitization of S49AR in time (S49ARS cells) towards multiple apoptotic stimuli is accompanied by re-expression of SHIP-1 and decreased pAkt

(A) Apoptosis sensitivity of S49 and S49ARS cells towards ALP (10 μM), FasL (500 ng/ml) or etoposide (5 μg/ml) after 6 h in comparison with S49AR cells. Apoptotic nuclear fragmentation was measured by FACScan analysis. (B) Western blot showing re-expression of SHIP-1 and decreased phosphorylation of Akt at Ser473 in S49ARS cells, in comparison with S49AR cells. Actin served as a loading control. (C) 32P-labelled phosphoinositide levels, showing regain of PtdIns(3,4,5)P3 and concomitantly decreased PtdIns(3,4)P2 in S49ARS cells. Lipids were quantified relative to [32P]PtdIns and the results are means± S.D. for three experiments; (D) results from (C) alternatively expressed as the ratio of radiolabelled PtdIns(3,4,5)P3 to PtdIns(3,4)P2.

Down-regulation of SHIP-1 using siRNA induces multi-stress resistance

To determine the involvement of SHIP-1 in ALP-induced resistance of S49AR cells, we made a SHIP-1-knockdown cell line using siRNA technology. S49 cells were stably transduced with a retroviral vector expressing a siSHIP-1 targeting sequence. Figures 5(A) and 5(B) show that SHIP-1 knockdown in the S49siSHIP cells was almost complete, both at the gene and protein expression levels, whereas S49 cells showed clear SHIP-1 expression. S49AR cells showed decreased SHIP-1 expression, but not as much decreased as in the S49siSHIP cells. Exposure of the S49siSHIP cells to 10 μM ALP (6 h) led to an approximately 50% decrease in ALP-induced apoptosis compared with parental S49 and mock-transduced cells, but did not reach the complete ALP resistance found in the S49AR cells. On the other hand, S49siSHIP cells were indeed completely resistant to etoposide and FasL-induced apoptosis (Figure 5C).

Figure 5 Silencing of SHIP-1 by siRNA mimics S49AR cells in the shift of phosphoinositide levels and enhanced activity of Akt and ERK1/2, and makes cells (cross-)resistant to multiple apoptotic stimuli

(A) Quantitative PCR and (B) Western blot showing down-regulation of SHIP-1 mRNA and protein in S49AR, S49siSHIP and S49siSMS1 cells compared with parental and mock-transduced S49 cells. (C) Apoptosis induced by ALP (10 μM), FasL (500 ng/ml) or etoposide (5 μg/ml) for 6 h in S49siSHIP cells, in comparison with fully sensitive (mock-transduced) S49 cells and fully resistant S49AR cells. Apoptotic nuclear fragmentation was measured by FACScan analysis. Results are the means±S.D. for four experiments. (D) 32P-labelled phosphoinositide levels showing enhanced PtdIns(3,4,5)P3 and concomitantly decreased PtdIns(3,4)P2 in S49siSHIP cells. Lipids were quantified relative to [32P]PtdIns; and (E) expressed as the ratio of radiolabelled PtdInsP3 to PtdInsP2. (F) Western blots of S49siSHIP cell lysate showing increased phosphorylation of Akt at its Ser473 site and pERK1/2, both indicative of their activation.

To strengthen the relevance of these findings, we measured the vulnerability of S49siSHIP, S49AR and parental S49 cells towards DNA damage in long-term colony assays. To this end, cells were treated with increasing doses (Gys) of ionizing radiation and seeded in methylcellulose matrix. Surviving cells were allowed to form colonies that were scored after 2 weeks. The SHIP-1-deficient S49siSHIP and S49AR cells survived this DNA damage better than the parental and mock-transfected S49 control cells (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/440/bj4400127add.htm), which is indicative of radioresistance.

We checked whether the down-regulation of SHIP-1 altered the levels of its substrate PtdIns(3,4,5)P3 and its direct product PtdIns(3,4)P2 accordingly. Using 32P-radiolabelling and HPLC, we found that the ratio of PtdIns(3,4,5)P3/PtdIns(3,4)P2 was increased in the S49siSHIP cells (Figures 5D and 5E), suggesting that SHIP-1 was indeed a regulator of phosphoinositide levels in these cells. Since PtdIns(3,4,5)P3 activates Akt and (indirectly) ERK1/2, we measured the activation status of these enzymes by their degree of phosphorylation. The immunoblots in Figure 5(F) show that down-regulation of SHIP-1 in the S49siSHIP cells results in increased phosphorylation of ERK1/2 and of Akt at its Ser473 site, indicative of their activation.

SHIP-1 down-regulation leads to SM and Fas deficiency and to reduced cellular uptake of ALP

Intriguingly, the biosynthesis of SM, a major constituent of lipid rafts, which was virtually absent in resistant S49AR cells [12,13], was also completely down-regulated in the apoptosis-resistant S49siSHIP cells (Figure 6A). Conversely, siRNA-induced down-regulation of SMS1 in S49 cells and consequent SM deficiency (S49siSMS1 cells), described previously [13], resulted in apoptosis resistance [12,13] and concomitant down-regulation of SHIP-1 expression, both at the mRNA and the protein level (Figures 5A and 6A respectively). Spontaneous re-sensitization (S49ARS cells), which led to re-expression of SHIP-1 (Figure 4B), also led to regained SM synthesis [12]. However, the mere re-expression of SMS1 in the resistant cells (AR-SMS1) which rescued SM synthesis (Figure 6B) did not re-sensitize the cells, as reported previously [12], and also did not lead to full re-expression of SHIP-1 (Figure 6B). Unfortunately, we were unable to re-express SHIP-1 in the resistant S49AR cells, despite numerous attempts. For unknown reasons, the transfected cells did not survive.

Figure 6 SHIP-1 deficiency correlates with lack of SM synthesis

(A) SHIP-1 deficient S49AR, S49siSHIP and S49siSMS1 cells were compared with parental and mock-transduced S49 cells. Cells were labelled with [1-3H]sphingosine for 4 h. Lipids were extracted and separated by TLC (see the Experimental section). The location of SM and other sphingolipids, glucosylceramide (GlcCer), lactosylceramide (LacCer) and sphingosine (So) is indicated. Phosphatidylethanolamine (PE) is a catabolic end-product of sphingosine degradation. (B) Re-expression of SMS1 in S49AR cells (AR-SMS1 cells) recovers SM synthesis, but does not lead to full re-expression of SHIP-1.

Multi-stress resistance induced by the down-regulation of SHIP-1 or SMS1 includes the resistance to ALP and to Fas ligation (Figure 5C). The resistance to ALP was previously shown to be due to decreased ALP uptake by raft-dependent endocytosis [6,7,23]. In agreement with this notion, siRNA-induced down-regulation of SHIP-1 or SMS1 in S49 cells reduced the uptake of [3H]ALP in a time-dependent fashion by approximately 50%, down to the level of the ALP-resistant S49AR cells (Figure 7). As regards FasL resistance, both SM and the Fas receptor are located in lipid raft microdomains at the cell surface ([12], and references therein). The first step in Fas signalling is ligand-induced, raft-dependent formation of Fas microaggregates at the cell surface [24]. Previously, we have shown that resistant S49AR cells have decreased expression of Fas and less Fas on the cell surface [12]. We therefore compared the surface expression of Fas in the FasL-resistant, SM-deficient S49AR, S49siSHIP and S49siSMS1 cells with the FasL-sensitive S49 and S49mock cells, using a fluorescent anti-Fas antibody and FACS analysis. Figure 8(A) shows that the expression of Fas on the surface of S49siSHIP and S49siSMS1 cells was significantly and approximately equally reduced in comparison with the parental cells (S49 and S49mock), but not as much as in the S49AR cells. Furthermore, immunoblots of the total lysates of these cells confirm that Fas expression was much reduced in S49AR, S49siSHIP and S49siSMS1 cells (Figure 8B).

Figure 7 Down-regulation of SHIP-1 or SMS1 leads to reduced ALP uptake

Uptake of [3H]ALP by resistant S49siSHIP, S49siSMS1 and S49AR cells in comparison with ALP-sensitive S49 cells. At various time-points, samples of cells were taken, put on ice, and washed three times with ice-cold PBS. Radioactivity in the cells was expressed as d.p.m./mg of cellular protein.

Figure 8 Reduced Fas expression in S49siSHIP, S49siSMS1 and S49AR cells

(A) Cell-surface expression of Fas analysed by flow cytometry using Jo2 antibody (against mouse Fas) conjugated to FITC (Jo2-FITC). Shown are the fluorescence distributions of control cells (no antibody; trace 1), S49 cells (trace 2), S49AR cells (trace 3), S49mock (trace 4), S49siSMS1 cells (trace 5) and S49siSHIP (trace 6). Traces 5 and 6, representing S49siSHIP and S49siSMS1, as well as traces 2 and 4, representing S49 and S49mock, are overlapping. (B) Fas protein expression in lysates from the cells indicated, determined by Western blotting using anti-Fas antibodies. Tubulin served as loading control.

An important and intriguing conclusion to be drawn from our studies is that the expression of SHIP-1 [PtdIns(3,4,5)P3 dephosphorylation] and SMS1 (SM synthesis) in parallel decrease with multi-stress resistance to apoptosis, and increase with spontaneous re-sensitization. This correlation also holds when either of these enzymes is down-regulated by RNAi. How this tight interplay between synthesis of SM and degradation of PtdIns(3,4,5)P3 is regulated remains unknown.

Altered gene expression profile in apoptosis-resistant cells

Comparison of global gene expression in apoptosis-resistant S49AR, S49siSMS1 and S49siSHIP cells, as well as in S49mock cells, with the parental apoptosis-sensitive S49 cells results in a set of 1889 genes that are differently expressed. Hierarchical clustering analysis showed that within this set of genes there was a prominent clustering of the expression profiles of the resistant compared with the sensitive (parental) cell lines (see Supplementary Figure S2 at http://www.BiochemJ.org/bj/440/bj4400127add.htm). The profiles of S49siSHIP and S49siSMS1 cells are very similar and at the same time resemble the profile of S49AR cells, but differ from the parental S49 and S49mock cells. A list of the 56 most down-regulated genes and 30 most up-regulated genes is provided in Supplementary Tables S1 and S2 respectively (at http://www.BiochemJ.org/bj/440/bj4400127add.htm). Although Ingenuity pathway analysis suggests that some of these genes relate to cell survival or lipid signalling, there is no apparent (direct) functional relationship to SHIP-1 or SMS1 in our cell system (see Results and Discussion in the Supplementary Online Data at http://www.BiochemJ.org/bj/440/bj4400127add.htm).

DISCUSSION

In the present study we have shown that S49 mouse lymphoma cells that were made resistant to ALP (S49AR cells) are cross-resistant to Fas ligation [12] and to DNA damage induced by etoposide or γ-radiation. We discovered that this multi-stress resistance was dependent on down-regulation of SHIP-1, a PtdIns(3,4,5)P3 5-phosphatase (see summarizing scheme in Figure 9). Spontaneous re-sensitization of the S49AR cells could occur upon culturing without ALP for several weeks, resulting in regained SHIP-1 expression. In the absence of PTEN, the direct consequence of SHIP-1 down-regulation in the resistant cells was a constitutively elevated level of PtdIns(3,4,5)P3 and increased phosphorylation (activation) of Akt, a key protein kinase of the canonical survival pathway [21]. However, pharmacological inhibitor studies revealed that this amplification of the PI3K/Akt pathway was essential only for the resistance to DNA damage, not to ALP or FasL.

Figure 9 SHIP-1, in unresolved interplay with SMS1, regulates multi-stress resistance via divergent pathways

Down-regulation of SHIP-1 results in an increase in PtdIns(3,4,5)P3, and consequent increase of survival signalling through Akt phosphorylation (activation), which causes etoposide (DNA damage) resistance. The down-regulation of SHIP-1, in conjunction with SMS1 down-regulation, independent of Akt, also determines resistance to ALP and to Fas ligation. Decreased SMS1-mediated SM synthesis leads to diminished raft-dependent endocytosis of ALP, and decreased levels of Fas, resulting in ALP and FasL resistance respectively. An unexplained interaction between SHIP-1 and SMS1, possibly by an unknown factor (X), may lead to down-regulation of both enzymes, resulting in the cross-resistance. Note that a decrease in either SHIP-1 or SMS1 expression leads to down-regulation of the other, but that up-regulation of SMS1 in resistant cells neither leads to full up-regulation of SHIP-1 (hence dotted arrow) nor to re-sensitization of the cells.

In previous studies, we found that the acquired resistance to ALP was due to impaired uptake of this lipid via raft-dependent endocytosis [6,7]. We also reported that the apoptosis-resistant S49AR cells were deficient in the major raft constituent SM, owing to the complete down-regulation of SMS1, the only SM synthase isotype present in S49 cells [12,13]. RNAi studies revealed that the lack of SMS1 was partly responsible for the apoptosis resistance [12]. Since re-expression of SMS1, and thus regained SM synthesis (AR-SMS1 cells), failed to re-sensitize the cells, there had to be (an)other factor(s) a co-determine apoptosis sensitivity. The present study identified SHIP-1 as a common regulator of S49 cell sensitivity to all inducers of apoptosis studied. Furthermore, we found a co-expression and interdependent co-regulation of SHIP-1 and SMS1, that together regulate apoptosis sensitivity and resistance. The mechanism behind this tight interplay between these two lipid-metabolizing enzymes remains unknown. Given the observation that simple re-expression of SMS1 did not lead to re-sensitization of the cells, we postulate that ALP- or siRNA-induced multi-stress resistance in S49AR, S49siSHIP and S49siSMS1 cells should also involve (an)other regulator(s) (X in Figure 9) that remain(s) unaffected by viral transduction/re-expression of SMS1. The presence of (an) additional regulator(s) X is consistent with our observation that resistance is associated with up- or down-regulated expression of many genes (see Supplementary Figure S2 and Tables S1 and S2). However, among these genes, we did not find possible candidates that could relate to factor X. Most striking is our observation that the gene expression profiles of the apoptosis-resistant S49siSHIP and S49siSMS1 cells are very similar and at the same time resemble the profile of resistant S49AR cells, but differ from the parental (apoptosis-sensitive) S49 and S49mock cells (Supplementary Figure S2).

SHIP-1 may exert different roles in haemopoietic cell biology depending on the cell context [25,26]. Its catalytic (5-phosphatase) activity opposes PI3K signalling, and in this way SHIP-1 has been shown to counteract cell proliferation [27,28], acting as a tumour suppressor [26], to control cell polarity and motility [29], and to play a negative regulatory role in the immune system and blood cell production [26,30]. In the S49 cell system, the PI3K–Akt pathway determines the apoptosis resistance to DNA damage, in which SHIP-1 thus acts as a negative regulator (or apoptosis sensitizer). On the other hand, this survival pathway does not determine the resistance to ALP and FasL, as noted. Thus there must be functions of SHIP-1 other than the catalytic degradation of PtdIns(3,4,5)P3 that determine the sensitivity to ALP and FasL. One could think of two possible alternative functions: first, the enzymatic product of SHIP-1, PtdIns(3,4)P2, retains the phosphate group on the third position of the inositol ring and thus retains some signalling ability towards PH (pleckstrin homology) domains of distinct proteins such as TAPP-1 (tandem PH-domain-containing protein 1) [25,31]. In this way, SHIP-1 has been considered to be a ‘gatekeeper’ rather than a terminator, acting as a switch to redirect PI3K-dependent signalling towards a set of distinct effectors that are locally and functionally separate from PtdIns(3,4,5)P3-dependent events [15]. Thus in the case of ALP- and Fas-induced apoptosis, SHIP-1 might function to modify phosphoinositide signalling rather than to terminate it. Alternatively, next to the dephosphorylation of PtdIns(3,4,5)P3, SHIP-1 has been shown to serve, via its SH2 domain and other structural motifs, as an adaptor (scaffolding/docking) molecule in the formation of multi-protein complexes at the plasma membrane [14]. Unfortunately, due to poor immunoprecipitation of SHIP-1, we were unable to find a possible physical interaction with SMS1 or any other protein. To discriminate between the catalytic and non-catalytic functions of SHIP-1, mutagenesis in the phosphatase domain would be helpful, but unfortunately, for unknown reasons, our S49AR cells did not survive such manipulations.

The enhanced kinase activity of Akt in S49AR cells is apparent at its major regulatory phosphorylation sites, Ser473 and Thr308, both of which contribute to its full activity. Up-regulation of pSer473 was more pronounced than that of pThr308. Following recruitment of Akt to PtdIns(3,4,5)P3 [and/or PtdIns(3,4)P2] at the plasma membrane, phosphorylation of these sites is differentially regulated by protein kinases and phosphatases [21]. For example, Thr308 is phosphorylated by PDK1 (phosphoinositide-dependent kinase 1), whereas Ser473 can be phosphorylated by mTORC2 (mTOR complex 2; PDK2), ILK (integrin-linked kinase), DNA-PK (DNA-dependent protein kinase) and through autophosphorylation. Depending on these different enzyme activities and the cell type, the effects on Akt and further downstream events, such as cell survival, may vary [21]. Jurkat T-cells, for example, having a similar PTEN- and SHIP1 deficiency as S49AR cells, and therefore high PtdIns(3,4,5)P3 levels, show high pThr308 levels that seem not strictly dependent on PDK1, but rather on low PP2A (protein phosphatase 2A)-like phosphatase activity [32]. We have not investigated these upstream regulators of Akt activity in our S49 cell system.

It is not clear how RNAi-mediated down-regulation of SHIP-1 confers resistance to Fas ligation. We find that Fas expression is reduced, both as total protein as well as on the cell surface. The results differ from a previous paper in which SHIP-1 inhibited Fas-induced apoptosis in T-cells [33]. This inhibition was independent of the SHIP-1 phosphatase activity, but rather due to Fas glycosylation, leading to failure of oligomerization upon stimulation, which is an entirely different concept from ours. Because of the cell-context-specific functions of SHIP-1 [25,26] noted above, apparent controversy exists between models of SHIP-1 function proposed from studies conducted using different cell types [25]. For example, our finding that SHIP-1 expression determines apoptosis sensitivity of S49 cells is in agreement with studies in murine DA-ER haemopoietic cells, where SHIP-1 overexpression promoted apoptosis [34]. However, in other studies, SHIP-1 inhibition triggered apoptosis of blood cancer cell lines [30], whereas restoration of SHIP-1 activity in human leukaemic cells activated an IκB (inhibitory κB) kinase-dependent NF-κB (nuclear factor κB) pathway and promoted survival upon oxidative stress [35]. Furthermore, Jurkat cells expressing SHIP-1 were more resistant to H2O2-induced apoptosis than parental cells [35]. Exposure of S49 cells to ALP not only inhibits lipid metabolism [6,7] (a major cause of apoptosis), but also causes oxidative stress [1,17]. Since SHIP-1 protects T-cells against oxidative stress [35], it might thus be involved in the initial S49 cellular defence against ALP. However, the subsequent down-regulation of SHIP-1 upon prolonged ALP exposure argues for a negative rather than a positive role of this protein in the acquired resistance of S49AR cells.

In conclusion, we have identified SHIP-1 as a common regulator of apoptosis sensitivity towards ALP and related ALPs. Down-regulation of this PtdIns(3,4,5)P3 phosphatase determines resistance to ALP and concomitant cross-resistance to Fas ligation and to DNA damage. Associated with this resistance is the down-regulation of SMS1 and consequent SM deficiency in the plasma membrane. How the combination of SHIP-1 and SMS1 down-regulation makes cells apoptosis resistant remains unknown.

AUTHOR CONTRIBUTION

Maaike Alderliesten measured (inhibition of) apoptosis, phospho-Akt, phospho-ERK and altered gene expression in resistant and genetically manipulated cells. Jeffrey Klarenbeek performed many of the experiments, including sphingolipid analysis and ALP uptake. Arnold van der Luit discovered sphingomyelin and SHIP-1 deficiency in apoptosis-resistant cells. Menno van Lummel performed real-time PCR analyses. Shuraila Zerp performed clonogenic survival assays. David Jones in the laboratory of Nullin Divecha generated the phosphoinositide data. Marcel Verheij was co-leader of the project. Wim van Blitterswijk supervised the project and wrote most of the paper.

FUNDING

This work was supported by the Dutch Cancer Society [grant number NKI 2005-3377]. Funding in the Paterson Insitute for Cancer Research is entirely from Cancer Research UK.

Abbreviations: ALP, alkyl-lysophospholipid; DMEM, Dulbecco's modified Eagle's medium; edelfosine, 1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine; FasL, Fas ligand; ERK, extracellular-signal-regulated kinase; HRP, horseradish peroxidase; MAPK, mitogen-activated protein kinase; MEK, MAPK/ERK kinase; mTOR, mammalian target of rapamycin; PDK, phosphoinositide-dependent kinase; PH, pleckstrin homology; PI3K, phosphoinositide 3-kinase; PTEN, phosphatase and tensin homologue deleted on chromosome 10; RNAi, RNA interference; SH, domain, Src homology domain; SHIP, SH2-domain-containing inositol phosphatase; siRNA, small interfering RNA; SM, sphingomyelin; SMS, sphingomyelin synthase; z-VAD-FMK, benzyloxycarbonyl-Val-Ala-DL-Asp-fluoromethylketone

References

View Abstract