Oncogene (2010) 29, 898–908 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00 www.nature.com/onc ORIGINAL ARTICLE Inhibition of Akt signaling by the lignan matairesinol sensitizes prostate cancer cells to TRAIL-induced apoptosis

E Peuhu1,2, A Rivero-Mu¨ller1,3,7, H Stykki1,2,7, E Torvaldson1,2, T Holmbom4, P Eklund5, M Unkila6, R Sjo¨holm5 and JE Eriksson1,2

1Turku Centre for Biotechnology, University of Turku and A˚bo Akademi University, Biocity, Turku, Finland; 2Department of Biology, A˚bo Akademi University, Tykisto¨katu, Turku, Finland; 3Institute of BioMedicine, University of Turku, Kiinamyllynkatu, Turku, Finland; 4Laboratory of Fiber and Cellulose Technology, A˚bo Akademi University, Porthansgatan 3, Turku/A˚bo, Finland; 5Laboratory of Organic Chemistry, A˚bo Akademi University, Biskopsgatan 8, Turku/A˚bo, Finland and 6Hormos Medical Ltd, Ita¨inen Pitka¨katu 4b, Turku, Finland

Tumor necrosis factor-related apoptosis-inducing ligand et al., 2008). Chemotherapy and androgen withdrawal are (TRAIL) has been shown to be selectively pro-apoptotic in used for slowing down the spread of the disease but cancer cells, with minimal toxicity to normal tissues. currently there are no effective treatments for advanced Although this feature makes TRAIL a promising anti- prostate cancer. Therefore, better understanding of cancer cancer agent, not all cancer cell types are sensitive to cell survival mechanisms and more efficient therapy forms TRAIL-induced apoptosis despite abundant expression of are required for treatment of prostate cancer. In this TRAIL receptors. Thus, combinatorial treatments to regard, death receptor ligands, with capacity to activate sensitize tumor cells to TRAIL-induced apoptosis have apoptosis or programmed cell death specifically in cancer been in the focus of extensive research. Dietary lignans cells, have been suggested as one possible treatment have shown cancer preventive and antitumorigenic activ- approach (reviewed by Fulda and Debatin, 2004). ity, but the mechanisms behind these effects are poorly Activation of the tumor necrosis factor family death known. Here we observed that of the three tested lignan receptors, such as the Fas receptor (CD95/APO-1/ molecules, matairesinol (MAT) was the most effective as DR6) and the tumor necrosis factor-related apoptosis- a death receptor-sensitizing agent. MAT sensitized the inducing ligand (TRAIL) receptors 1 and 2 (DR4 and androgen-dependent LNCaP cells to TRAIL-induced DR5), initiates apoptosis through the extrinsic signal- apoptosis both in the presence and absence of androgens. ing pathway. The cytoplasmic death-inducing signal- Treatment with MAT markedly decreased Akt activity, ing complex (DISC) that is formed to the activated which has been implicated as a key signaling mechanism in death receptors serves as a platform for activation of the TRAIL resistance of LNCaP prostate cancer cells. the initiator cysteine proteases caspase-8 and caspase- The involvement of the pathway in the MAT-mediated 10 (Kischkel et al., 1995; Muzio et al., 1996; Vincenz sensitization was shown in rescue experiments using and Dixit, 1997). The proteolytic activity of the ectopic expression of constitutively active Akt. Owing to downstream effector caspasesleadstoapoptosisof the high activity of phosphatidylinositol 3-kinase/Akt the cell. Various damaging stresses trigger the so-called signaling in cancer, targeting this survival pathway with intrinsic pathway of apoptosis. This signaling pathway MAT could markedly benefit TRAIL-based tumor involves the Bcl-2 family proteins that control mito- therapies, including those aimed at prostate cancer. chondrial integrity, cytochrome c release and activa- Oncogene (2010) 29, 898–908; doi:10.1038/onc.2009.386; tion of caspase-9 (Danial and Korsmeyer, 2004). The published online 23 November 2009 death receptor-mediated extrinsic and the mitochon- drial intrinsic apoptosis pathways are connected Keywords: matairesinol; lignans; apoptosis; TRAIL through Bid, a member of the Bcl-2 family proteins, which is a substrate for caspase-8. When Bid becomes cleaved by activated caspase-8, its truncated form (tBid) translocates to the mitochondria in which it induces the release of cytochrome c and other pro- Introduction apoptotic factors (reviewed in Sprick and Walczak, 2004). The main site of death ligand expression is the Prostate cancer is the most common malignancy and the immune system in which NK and T cells use them for second leading cause of cancer mortality in men (Jemal killing transformed or infected cells (reviewed in Schaefer et al.,2007).TRAILorFas ligand-induced Correspondence: Professor JE Eriksson, Department of Biology, A˚bo apoptosis is thereby an important mechanism in the Akademi University, Tykisto¨katu 6B, Turku, 20520, Finland. immune surveillance against transformed, potentially E-mail: john.eriksson@abo.fi 7These authors contributed equally to this work. carcinogenic cells and, consequently, successful cancer Received 28 April 2009; revised 25 September 2009; accepted 2 October cells usually develop resistance against death receptor 2009; published online 23 November 2009 stimulation. Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 899 The use of Fas ligand as a potential anti-neoplastic inhibition of Akt activity is a key molecular mechanism agent has got into a halt because of induction of severe by which lignans exert their anticancer effects. The liver damage (Ogasawara et al., 1993). However, the presented concept of cancer therapy could be practicable focus has turned to TRAIL because it has been shown to in developing therapies against androgen-dependent induce apoptosis in a range of different tumor cell lines prostate tumors. with minimal pro-apoptotic effects in normal cells. This has lead to extensive research to use TRAIL in treatment of cancer. Several studies have now shown that TRAIL-based therapies show selective antitumor Results activity in mouse models of colon, glioma, lung and prostate cancer, and multiple myeloma (reviewed in Lignans sensitize LNCaP prostate cancer cells to TRAIL-induced apoptosis Koschny et al., 2007). More importantly, recombinant human TRAIL and human monoclonal agonistic Earlier reports show that androgen-dependent LNCaP TRAIL receptor antibodies are currently analysed in prostate cancer cells are TRAIL resistant (Nesterov et al. et al. phase Ib/II clinical trials for cancer therapy. The early , 2001; Thakkar , 2001), although treatment clinical trials are promising and no toxicity has been with the ENL lignan inhibited their proliferation et al. observed in human patients treated with TRAIL or (McCann , 2008) and induced cell death after et al. agonistic antibodies (reviewed by Mahalingam et al., prolonged exposure (Chen , 2007). As it is well 2008). However, targeting of TRAIL receptors alone documented that plant lignans also have potent appears not to be potent enough to kill all cancer cell antitumorigenic activity with minimal side effects et al. et al. types and combination with other sensitizing agents (Thompson , 1996; Bylund , 2005; Saarinen et al. might be needed to induce efficient tumor cell apoptosis. , 2008), we tested whether this activity could Prostate tumors express frequently TRAIL receptors involve sensitization to death receptor signaling. on the cell surface, but these receptors are often not able A number of closely related lignans (Figure 1a) were to trigger TRAIL-induced apoptosis because of in- tested in a non-toxic concentration for their capacity to creased pro-survival signaling (Bucur et al., 2006). sensitize LNCaP prostate cancer cells to TRAIL Research is now focused on discovering combinatorial receptor-mediated apoptosis in the prescence or absence therapies with agents that reduce TRAIL resistance but of androgen. When comparing the effects of hydro- do not harm normal tissues. Lignans are dietary xymatairesinol, ENL and MAT, the results of nuclear phytoestrogens (reviewed by Adlercreutz, 2007) that fragmentation assay showed that both ENL and have shown cancer preventive and antitumorigenic especially MAT were efficient in sensitizing the cells to activity, and thereby may show potential as TRAIL- TRAIL receptor signaling (Figure 1b). Hydroxymatair- esinol had no effect on TRAIL-induced apoptosis. The sensitizing agents. It has been shown that lignans inhibit proliferation and are cytotoxic to prostate cancer cell sensitizing effect of MAT was more pronounced in androgen-deprived cells but observable also in the lines (Lin et al., 2001, Chen et al., 2007). Moreover, androgen-supplied samples that show stronger resis- lignans can prevent tumor growth and induce apoptosis in vivo in prostate cancer xenografts or chemically tance toward TRAIL treatment. At 40 mM, none of the used compounds induced apoptosis in 20 h when induced mammary tumors (Bylund et al., 2005; Saarinen administered alone. We also consistently observed et al., 2008; Thompson et al., 1996) and can also reduce inhibition of cell proliferation by MAT, which corro- experimental metastasis (Li et al., 1999). Plant-derived borates previous work by McCann et al. (2008) (data lignans, such as matairesinol (MAT), are metabolized by not shown). Furthermore, longer pretreatment (24 h) mammalian gut microflora resulting in enterolactone with 40–100 mM MAT was able to some extent enhance (ENL) or enterodiol (Axelson et al., 1982). As a result of TRAIL-induced apoptosis in leukemic Jurkat T cells their promising antitumorigenic effects, we wanted to and HeLa cells, two TRAIL-sensitive cancer cell lines test whether lignans could be used as therapeutic agents, included for comparison (Supplementary Figure 1). The to counteract cancer cell resistance to TRAIL. The purpose of this study was to examine whether prolonged pretreatment with hydroxymatairesinol, ENL or MAT could also sensitize HeLa cells to Fas receptor- lignans could enhance death receptor signaling to enable mediated apoptosis, suggesting that the mechanism TRAIL-based therapies in prostate cancer. To this end, we used the TRAIL-resistant LNCaP cell model for behind the sensitizing effect may be common for different death receptors (Supplementary Figure 2). androgen-sensitive stages of prostate cancer (Nesterov et al., 2001). When screening the lignans for TRAIL sensitization, we observed that MAT was especially MAT sensitizes LNCaP cells effectively to TRAIL efficient in sensitizing cancer cells to TRAIL receptor- but does not enhance doxorubicin-induced apoptosis mediated apoptosis. MAT-sensitized LNCaP cells to As MAT was found to be the most efficient of the tested TRAIL through increased cleavage of Bid, amplified lignans in sensitization of LNCaP cells to TRAIL- mitochondrial depolarization and facilitated caspase induced apoptosis, we examined the concentration activation. The sensitization was dependent on down- dependence of the effect. Androgen-deprived or sup- regulation of the phosphatidylinositol 3-kinase (PI3K)/ plied LNCaP cells were treated with 0–100 mM MAT for Akt survival pathway activity. These results suggest that 1 h before addition of 10 ng/ml izTRAIL (Figure 2a). sensitization to death receptor-mediated apoptosis by After 20 h, the percentage of cells with activated

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 900 HMR ENL MAT

O O H O H H MeO HO MeO O O O HO HO HO H H H

OMe OH OMe HO HO

** 30 control HMR ENL MAT 20

*

10 % nuclear fragmentation

0 control TRAIL control TRAIL -androgen +androgen Figure 1 Lignans sensitize cancer cells to Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced apoptosis. (a) Chemical structures of lignans 7-hydroxymatairesinol (HMR), enterolactone (ENL) and matairesinol (MAT). (b) LNCaP cells were cultured in the presence or absence of androgen (1 nM). After treatment with 40 mM HMR, ENL, MAT or solvent control for 1 h izTRAIL (10 ng/ml) was added to cells. Twenty hours later, the cells were disrupted and the nuclei labeled for DNA content with propidium iodide. The samples were analysed by flow cytometry and the fraction of sub-G0/G1 events (nuclear fragmentation) was detected as a measure of apoptotic cell death (mean±s.e.m.; n ¼ 3; Student’s t-test *Po0.05, **Po0.01).

caspase-3 was detected by flow cytometry. Already in the androgen-independent and TRAIL-sensitive PC3 10 mM MAT enhanced TRAIL-induced apoptosis but prostate cancer cells (Supplementary Figure 3). more pronounced effect was observed with 40–100 mM MAT. To reduce the possibility of MAT-induced cytotoxicity with high concentrations, we continued Surface expression of TRAIL receptors or TRAIL–DISC the experiments with 40 mM MAT, as in androgen- formation is not increased in MAT-treated LNCaP cells supplemented cells the almost maximal sensitizing effect To determine the molecular mechanisms underlying the was already observed with this dose. When 40 mM MAT effects of MAT on TRAIL-receptor signaling, we was combined to increasing doses of izTRAIL, both studied different components of the death receptor androgen-deprived and androgen-supplied LNCaP cells signaling pathway. Expression of TRAIL receptors and underwent apoptosis (caspase-3 activation) with lower decoy receptors on the cell surface regulates cellular TRAIL concentrations and higher maximum apoptosis responsiveness to TRAIL (Pan et al., 1997; Sheridan percentage was reached (Figure 2b). Combination of et al., 1997; Jin et al., 2004). To survey the possibility 40 mM MAT to 10 ng/ml izTRAIL also increased nuclear that MAT induces sensitization by increased TRAIL-R fragmentation that was observed by fluorescence micro- or decreased decoy-receptor expression, the amount of scopy of LNCaP cells (Figure 2c). DR4 (TRAIL-R1), DR5 (TRAIL-R2) and TRAIL To test whether the observed effect was specific to decoy receptors (DcR1 and DcR2) expression on the receptor-mediated apoptosis, we combined MAT treat- cell surface was analysed by flow cytometry in LNCaP ment with doxorubicin, a common cancer drug for cells. The results in Figure 3a show no increase of DR4 prostate cancer that induces apoptosis by transcriptional or DR5 expression and no decrease in DcR1 or DcR2 inhibition. Although TRAIL-induced polymerase clea- expression after 1- to 24-h incubation with 40 mM MAT. vage was amplified by MAT, such an effect was not Instead, MAT treatment decreased DR4 expression in observed in doxorubin-induced apoptosis (Figure 2d). both androgen-deprived and androgen-supplemented In contrast to these results with the androgen-responsive cells. Androgen, on the other hand, increased both LNCaP cell line, the same pretreatment with 40 mM MAT DR4 and DR5 expression (Figure 3a), an effect that did could not markedly enhance TRAIL-induced apoptosis not correlate with TRAIL sensitivity.

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 901 *** 50 control TRAIL 40 ** 30 ** *** 20

10 % caspase-3 activation

0 0 10401000 1040100MAT (µM) -androgen +androgen

-androgen +androgen 60 60 control *** ** 40µM MAT 50 ** 50

40 40 ** 30 30

20 20

10 10 % caspase-3 activation % caspase-3 activation

0 0 0 1 10 100 500 0 1 10 100 500 [log] izTRAIL (ng/ml) [log] izTRAIL (ng/ml)

control MAT TRAIL MAT+TRAIL -androgen +androgen

Control 40µM MAT

---10---10TRAIL (ng/ml) - 0.2 1 - - 0.2 1 - Doxorubicin (µM)

MW (kDa) PARP -androgen 80 p85

PARP +androgen 80 p85

Figure 2 Matairesinol (MAT) sensitizes LNCaP prostate cancer cells to tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)- induced apoptosis. (a) LNCaP cells ( þ /À 1nM androgen) were incubated with 0–100 mM MAT or solvent control for 1 h before addition of izTRAIL (10 ng/ml). After 20 h incubation, the percentage of cells with activated caspase-3 was detected by flow cytometry (mean±s.e.m.; n ¼ 3; Student’s t-test **Po0.01, *** Po0.001) (b)LNCaPcells(þ /À 1nM androgen) were incubated with 40 mM MAT or solvent control for 1 h before addition of izTRAIL in different concentrations. After 20 h incubation, the percentage of cells with activated caspase-3 was detected by flow cytometry (mean±s.e.m.; nX2; Student’s t-test **Po0.01, *** Po0.001). (c)LNCaPcells(þ /À 1nM androgen) were incubated with 40 mM MAT or solvent control for 1 h before izTRAIL (10 ng/ml) treatment for 20 h and analysed by fluorescence microscopy for nuclear fragmentation (d)LNCaPcells(þ /À 1nM androgen) were incubated with 40 mM MAT for 1 h before addition of 0–1 mM doxorubicin or 10 ng/ml izTRAIL. After 20 h, whole cell lysates were prepared and polymerase (PARP) cleavage was monitored by western blotting.

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 902 -androgen +androgen 80 80 0h 70 70 1h 60 60 8h 50 50 24h 40 40 30 30 20 20 10 10 Mean fluorescence intensity Mean fluorescence intensity 0 0 Sec Ab DR4 DR5 DcR1 DcR2 Sec Ab DR4 DR5 DcR1 DcR2

Input IP: DR4/DR5 control 40µM MATcontrol 40µM MAT 0 15 60 0 15 60 TRAIL (min)0 15 60 0 15 60 TRAIL (min)

Procaspase-8 Procaspase-8

p43/41 p43/41

p18 FADD -androgen

FADD DR5

Procaspase-8 Procaspase-8

p43/41 p43/41

p18 FADD +androgen

FADD DR5

-androgen +androgen MW (kDa) 0 5 20 0 5 20 MAT (hr)

46 FLIPL

30 25 FLIPS

Procaspase-8 46

25 FADD

46 Actin

Figure 3 Surface expression of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) receptors or TRAIL– death- inducing signaling complex (DISC) formation is not increased in matairesinol (MAT)-treated LNCaP cells. (a) LNCaP cells ( þ /À 1nM androgen) were incubated with 40 mM MAT or solvent control for 1–24 h and labeled on ice by indirect immunostaining for cell surface expression of TRAIL-receptors (DR4 and DR5) or TRAIL-decoy receptors (DcR1 and DcR2). Nonspecific fluorescence was measured using secondary antibody only. Mean fluorescence intensity was measured by flow cytometry (mean±s.e.m.; n ¼ 3). (b) LNCaP cells ( þ /À 1nM androgen) were incubated with 40 mM MAT or solvent control for 1 h before TRAIL–DISC stimulation (izTRAIL 1 mg/ml) for 0–60 min. TRAIL–DISC was immunoprecipitated with anti-DR4 and anti-DR5 antibodies and the amounts of caspase-8, FADD and DR5 in the complex were detected by western blotting. (c) The cells were incubated with 40 mM MAT for 0–20 h after which total cell lysates were prepared and analysed for expression of FLIPL/S (55, 25 kDa), caspase-8 (55/41/43 kDa) and FADD (23 kDa). Actin (45 kDa) expression was measured as loading control.

As formation of the DISC protein complex at the androgen-free or androgen-supplemented LNCaP intracellular domain of the TRAIL receptors is the first cells that had been preatreated for 1 h with 40 mM intracellular step of the signaling cascade leading to MAT or solvent control. After immunoprecipitation of apoptosis, we examined if MAT increases the amount of the DR4 and DR5 receptor complexes we detected DISC formation and subsequent caspase-8 activation. FADD, caspase-8 and DR5 in the precipitates. MAT TRAIL–R DISC was induced for 15 or 60 min in neither increased nor prolonged FADD or caspase-8

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 903 recruitment to the TRAIL–DISC in androgen treated or androgen TRAIL-induced apoptosis is suppressed up- androgen-deprived LNCaP cells suggesting that MAT stream of Bid cleavage and MAT can enhance the exerts its effect downstream of DISC formation cleavage. In the presence of androgen, apoptosis is (Figure 3b). We noticed that TRAIL–DISC formation blocked downstream of Bid cleavage, most likely by and caspase-8 activation was somewhat more efficient hormone-induced survival signals, and MAT treatment in androgen-supplied LNCaP cells (Figure 3b), which might affect other steps in the apoptotic pathway to supports the reported role of androgen as a mediator increase caspase-3 activation in response to TRAIL. of effective TRAIL–DISC formation in LNCaP cells Separately, we observed that when caspases were (Rokhlin et al., 2002). inhibited, downregulation of full size Bid (22 kDa) was We also examined the protein levels of the major not observed, showing that the downregulation is not TRAIL–DISC proteins, c-FLIPL/S, caspase-8 and because of reduction in the protein level, but rather FADD after 5 and 20 h of MAT treatment. Interest- caused by caspase-mediated cleavage (data not shown). ingly, a clear decrease in protein levels of both c-FLIPL/S The truncated Bid seems to be subjected to active and caspase-8, but not of FADD, was observed after proteasomal degradation, as proteasome inhibition 20-h MAT treatment (Figure 3c). In any case, this through epoximicin (200 nM) results in generation of downregulation was observed later than 5-h MAT the active 15-kDa tBid fragment after TRAIL treatment treatment and thereby is unlikely to affect the composi- (data not shown). We could also observe that Bid–green tion of the TRAIL–DISC that is induced after 1-h MAT fluorescent protein (GFP) transferred from diffuse pretreatment. Taken together, these results suggest that cytoplasmic fluorescence to presumably mitochondrial MAT does not affect the proximal death receptor (granular), localization, when MAT was combined to signaling, but exerts its effect downstream of TRAIL– TRAIL treatment (data not shown), further supporting DISC formation. the western blot results in Figure 4b, showing increased Bid cleavage in response to TRAIL when the cells are also treated with MAT. MAT increases TRAIL-induced Bid cleavage To evaluate the possibility that MAT affects the and mitochondrial depolarization in LNCaP cells mitochondrial signaling pathway by altering protein The involvement of intrinsic cell death signaling in expression, we examined the expression levels of Bid and regulation of death receptor responses of LNCaP cells other Bcl-2 family proteins after MAT treatment. has been shown by sensitization of these cells to TRAIL- Incubation with 40 mM MAT did not affect the expres- induced apoptosis by inhibition of anti-apoptotic sion of Bid, the pro-apoptotic Bcl-2 proteins Bax or Bcl-2 proteins (Kim et al., 2005; Ray et al., 2005). There- BimEL, or the anti-apoptotic BcL-XL (Figure 4). fore, we examined whether MAT could increase the activation of intrinsic mitochondrial apoptotic signal- ing, thereby amplifying death receptor signaling and MAT sensitizes LNCaP prostate cancer cells to apoptosis sensitizing the cells to TRAIL-induced apoptosis. through inhibition of Akt Mitochondrial membrane polarity was evaluated by The LNCaP cells have a high constitutive Akt activity, tetramethylrhodamine labeling 4–24 h after the start of which appears to protect the cells from TRAIL-induced the treatments. The results in Figure 4a show that apoptosis (Chen et al., 2001; Nesterov et al., 2001; mitochondrial depolarization was increased if the cells Rokhlin et al., 2002, 2005). To analyse the possibility were pretreated with MAT for 1 h before TRAIL that MAT would sensitize LNCaP cells to TRAIL by receptor ligation. The increase in TRAIL-induced inhibiting the PI3K/Akt-pathway, we first analysed mitochondrial depolarization was more pronounced in whole cell lysates after 1 h of MAT treatment for the the androgen-deprived cells. This result correlates well levels of phosphorylated Akt (Figure 5a). The activity of with the number of apoptotic cells (Figure 2a). Together Akt, as measured by the fraction of phospho-Akt from these results suggest that the MAT-mediated effect is the total amount of Akt, was quantified from the associated with amplification of the mitochondrial death western blots (Figure 5b). Treatment with MAT rapidly receptor signaling in response to TRAIL. decreased Akt phosphorylation in both androgen- As the caspase-mediated cleavage of Bid serves to link deprived and androgen-stimulated LNCaP cells, indi- the extrinsic and the intrinsic death pathways, we cating a possible mechanism for MAT in the sensitiza- analysed if MAT increases TRAIL-induced Bid clea- tion to TRAIL-induced apoptosis. Inhibition of the vage. This was indeed the case in the absence of PI3K with the small molecule inhibitor LY294002 also androgen, as TRAIL alone was unable to induce results in sensitization LNCaP cells to TRAIL-induced efficient Bid cleavage, but pretreatment with MAT apoptosis (data not shown, Chen et al., 2001). To verify clearly increased the cleavage (Figure 4b). In the the involvement of the Akt pathway, we transfected androgen-supplied LNCaP cells, which have stronger LNCaP cells with mock-GFP or constitutively active caspase-8 activation after TRAIL treatment, Bid clea- Akt-GFP (ca-Akt-GFP), and assayed the cells for vage already occurred with TRAIL treatment alone and MAT-mediated sensitization to TRAIL. In the andro- was not thereby enhanced further by MAT (Figure 4b). gen-deprived cells, expression of ca-Akt efficiently Nevertheless, the androgen-supplemented cells were reversed the MAT-mediated sensitization to TRAIL- found to be more resistant to TRAIL-induced apoptosis induced apoptosis (Figure 5c). The androgen-stimulated (Figure 2b). Our results show that in the absence of cells became greatly sensitized to TRAIL by the

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 904 -androgen +androgen 70 70 Control 60 60 MAT 50 50 TRAIL MAT+TRAIL 40 40 30 30 20 20 10 10 0 0 % mitochondrial depolarization % mitochondrial depolarization % mitochondrial 4h 8h 12h 24h 4h 8h 12h 24h

-androgen +androgen

-- ++-- ++MAT MW (kDa) -+ -+ -+ -+TRAIL 10 ng/ml 25 Bid 17 46 Actin

-androgen +androgen

MW (kDa) 0 5 20 0 5 20 MAT (hr)

Bax 17 25 BimEL

30 Bcl-XL 25

Bid 17 46 Actin

Figure 4 Matairesinol (MAT) promotes tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced Bid cleavage and mitochondrial depolarization. LNCaP cells ( þ /À 1nM androgen) were incubated with 40 mM MAT for 1 h before addition of 10 ng/ml izTRAIL. (a) After 4–24 h, the cells were labeled with tetramethylrhodamine methyl ester (TMRM) for mitochondrial membrane potential (MOMP) and analysed by flow cytometry. TMRM negative cells were gated as cells with depolarized mitochondria (Mean±s.e.m.; n ¼ 3) (b) Alternatively, total cell lysates were prepared after 20-h treatment and analysed by western blotting for cleavage of the full size Bid (22 kDa). Actin (43 kDa) expression was measured as loading control. (c) Whole cell lysates were prepared from MAT treated LNCaP cells and analysed by western blotting for protein expression of Bax (20 kDa), Bim EL (23 kDa), BclXL (30 kDa) and Bid (22 kDa). Actin (45 kDa) expression was measured as loading control.

transfection procedure itself and the effect of ca–Akt to TRAIL by MAT involves increased cleavage of Bid expression could not be examined in the presence of and amplification of the mitochondrial death pathway. androgen (data not shown). These results provide MAT downregulates Akt phosphorylation in both evidence that the MAT-mediated inhibition of the androgen-deprived and androgen-stimulated LNCaP PI3K/Akt-pathway indeed is required for the sensitiza- cells and MAT-mediated sensitization to TRAIL can tion to TRAIL, particularly in the absence of androgens. be reversed by ectopically increased Akt activity. Together these results imply that MAT inhibits Akt activity and thereby sensitizes LNCaP cells to TRAIL- Discussion induced apoptosis. In fact, the mechanism underlying LNCaP insensitivity to death receptor stimulation has In this study, we have shown that lignans can sensitize been shown earlier to involve the PI3K/Akt survival androgen-dependent LNCaP prostate cancer cells to pathway (Chen et al., 2001; Nesterov et al., 2001; TRAIL death receptor-mediated apoptosis. Our data Rokhlin et al., 2002, 2005). The constitutive activation show that in LNCaP prostate cancer cells, sensitization of the PI3K/Akt-pathway observed in many cancer

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 905 -androgen +androgen reported to sensitize non-small cell lung cancer cells to

MW (kDa) 0 40 100 0 40 100 MAT (µM) TRAIL by rapid downregulation of c-FLIPL/S expres- 58 sion (Raja et al., 2008). MAT also decreased c-FLIPL/S pAkt(ser473) protein levels but this downregulation occurred after more than 5-h MAT treatment. As our results also show 58 Akt unaltered TRAIL–DISC formation after MAT treat- ment, it is unlikely that the increase in TRAIL-induced

1.0 apoptosis observed with MAT treatment would be caused by increased caspase-8 activation in the DISC. 0.8 ** *** We cannot, however, rule out the possibility that the *** reduced c-FLIPL/S protein levels could influence the 0.6 later stages of TRAIL-induced apoptosis. Instead, our *** 0.4 data show that MAT enhances TRAIL-induced Bid cleavage in androgen-deprived LNCaP cells. According pAkt-ser473/Akt 0.2 to Nesterov et al. (2001) and Chen et al. (2001), Akt

0.0 mediates TRAIL-resistance of LNCaP cells, which 040100 040100 MAT (µM) appears to function through inhibition of Bid cleavage. -androgen +androgen A recent study by Aslan et al. (2009), describes an intriguing link between Akt activity and the pro- * GFP apoptotic actions of Bid in response to TRAIL 45 ** ca-Akt-GFP treatment. They identified PACS-2, a multifunctional 40 35 sorting protein, as an Akt substrate that regulates Bid 30 p-ca-Akt-GFP translocation to the mitochondria after TRAIL-receptor 25 activation. Owing to the importance of Akt activity in 20 TRAIL resistance of cancer cells, identification and 15 10 p-Akt employment of Akt inhibiting compounds, such as 5 MAT, in combinatorial TRAIL treatments may be 0 % apoptotic cells/GFP positive

GFP essential. control MAT TRAIL MAT+TRAIL ca-Akt- MAT-mediated sensitization of LNCaP cells to

mock-GFP TRAIL was less efficient if the cells were cultured with Figure 5 Matairesinol (MAT) sensitizes LNCaP prostate cancer androgen, and the exact mechanism of it remains to be cells to apoptosis through inhibition of Akt. (a) LNCaP cells ( þ /À analysed. When androgen was present, Bid cleavage 1nM androgen) were incubated with 0–100 mM MAT for 1 h after occurred after TRAIL stimulation alone but did not which total cell lysates were prepared. The western blot analysis with anti-Akt and anti-phospho (ser473)-Akt was quantified. lead to mitochondrial depolarization and apoptosis. (b) The histogram presents a relative decrease in pAkt/Akt ratio Therefore, androgen-induced TRAIL-resistance appears compared with control cells (mean±s.e.m.; nX7). (c) LNCaP cells to be conveyed by a mechanism upstream of the were transfected with mock-green fluorescent protein (GFP) or ca– mitochondria but downstream of Bid cleavage. Regula- Akt–GFP. After 24 h of androgen deprivation cells were treated tion of the apoptosis pathway at this level has not been with 40 mM MAT before addition of 10 ng/ml izTRAIL. After 20 h activated caspase-3 was labeled and the percentage of active extensively studied. As the expression of the anti- caspase-3 positive cells in the GFP positive cell population apoptotic Bcl-2 protein is downregulated by androgens was measured (mean±s.e.m.; n ¼ 3; Student’s t-test *Po0.05, (data not shown, Huang et al., 2004), the mechanism **Po0.01). The expression of ca–Akt–GFP recombinant protein that protects androgen-stimulated cells from mitochon- was verified by western blotting with anti-phospho-Akt antibody (inset). drial depolarization is likely to use other factors. Therefore, further studies are required to understand how androgen inhibits the mitochondrial amplification forms can be related, for example, to the loss of PTEN, of death receptor-mediated caspase activation. a lipid phosphatase that negatively regulates the path- Lignans are characterized as structurally similar to way (Vlietstra et al., 1998). Targeting of the PI3K/Akt- endogenous estrogens and belong to the group of pathway is therefore of great importance in treatment of phytoestrogens. Epidemiological and experimental data prostate cancer. suggest that lignans have beneficial effects on human In LNCaP cells, androgen deprivation results in health because of their estrogenic and anti-estrogenic, decreased TRAIL–DISC formation and caspase activa- anti-oxidative, anti-microbial and anti-carcinogenic tion (Rokhlin et al., 2002). We also observed less DR4 properties (reviewed by McCann et al., 2005 and and DR5 surface expression as well as somewhat Adlercreutz, 2007). The molecular mechanism behind decreased TRAIL–DISC formation and caspase-8 anticancer activity of lignans is not yet well character- activation in the absence of androgen. Interestingly, ized. Lignans have been reported to induce down- MAT treatment could overcome this effect by facilitat- regulation of the anti-apoptotic Bcl-2 proteins (Hausott ing TRAIL-induced Bid cleavage and mitochondrial et al., 2003) and to inhibit casein kinase II and protein depolarization. MAT did not increase the levels of DR4 kinase C a (Yokoyama et al., 2003). They can act as or DR5 surface expression, or caspase-8 activation in antioxidants protecting cells from oxidative stress (Kitts the DISC. Honokiol, a phenolic neolignan, was recently et al., 1999), or as inhibitors of steroid metabolizing

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 906 enzymes (Adlercreutz et al., 1993). Chen et al. (2007) the culture mediums were replaced by medium with or without recently showed that ENL induces spontaneous apop- androgen and the cells were cultured for further 24 h. tosis of LNCaP cells. The apoptotic mechanism was suggested to involve caspase activation through mito- Assessment of apoptosis chondrial depolarization. Concomitantly, downregula- LNCaP cells were treated with 0–100 mM lignans or solvent tion of Akt phosphorylation was observed, but no direct control for 1 h after which recombinant human isoleucine relationship or a possible mediator between these events zipper-TRAIL (izTRAIL; a kind gift from H Walczak) was was provided (Chen et al., 2007). We did not observe added to cells at indicated concentrations. Twenty hours later, the floating attached cells were collected and analysed for spontaneous apoptosis during the 24-h exposure by any apoptosis. The cells were disrupted and the nuclei labeled for of the lignans used in this study. The dose of MAT used DNA content with propidium iodide (Sigma-Aldrich). The in the LNCaP cells did not either induce apoptosis samples were analysed by flow cytometry (FL-2 channel, whereas combination with low concentrations of FACSCalibur; BD Pharmingen, San Diego, CA, USA) and the TRAIL was lethal to cells. More recent report shows fraction of sub-G0/G1 events (nuclear fragmentation) was that ENL inhibits insulin-like growth factor-1 receptor detected as a measure of apoptotic cell death. For further signaling, Akt activity and cell migration in PC3 examination of nuclear integrity, the cells were washed once in prostate cancer cells (Chen et al., 2009). These results phosphate-buffered saline and fixed for 15 min in 3% support the view that the anticancer activity of lignans is paraformaldehyde. Cytospin (Thermo Scientific, Waltham, largely because of their ability to inhibit the Akt MA, USA) preparations were made and mounted in DAPI Vectashield (Vector Laboratories, Burlingame, CA, USA). signaling. Whether this action is direct or indirect, like The nuclei were viewed with Leica DMRE fluorescence through the inhibition of the insulin-like growth factor-1 microscope (40 Â objective) and the images were collected receptor signaling, remains to be elucidated. with Wasabi (1.4) software. Activated caspase-3 in cells was To our knowledge, this is the first report showing that labeled with phycerythrin-conjugated antibody according to lignans can influence death receptor signaling by inhibit- manufacturer’s protocol (BD Pharmingen) and analysed by ing downstream survival mechanisms such as Akt activity, FACSCalibur flow cytometer (FL-2, FSC). For flow cyto- thus, providing a new mechanism for anticancer activity metric analysis of transfected LNCaP cells GFP-positive cells of lignans. Therapeutic strategies that target the PI3K/ were gated (FL-1) and measured for caspase-3 activation (FL- Akt pathway may be especially useful in treatment of 2). For doxorubicin-induced apoptosis LNCaP cells were prostate cancer. The combination of lignans and TRAIL treated with 0.2 or 1 mM doxorubicin (Sigma-Aldrich) for 20 h. has therapeutic potential against hormone-dependent prostate cancer and deserves further pre-clinical testing. TRAIL–R DISC immunoprecipitation This possibility is currently under investigation. To stimulate TRAIL receptors LNCaP cells were detached and pelleted (1000 revolutions per minute. for 10 min) and resus- pended in 1 ml of the collected medium. Later, 1 mg izTRAIL was added to the cell suspension. The cells were incubated at 37 1Cfor Materials and methods 15–60 min after which the reaction was stopped by adding 10 ml of ice-cold phosphate-buffered saline. Control cells were incubated without TRAIL at 37 1C for 60 min. After washing, the cells were Cell culture and reagents lysedin1mloflysisbuffer(20mM Tris–HCl, pH 7.4, 150 mM The LNCaP prostate cancer cells (clone FGC; EACC) were NaCl, 10% glycerol, 0.2% Nonidet P40, and complete protease cultivated in RPMI-1640 medium (Sigma-Aldrich, St Louis, MO, inhibitor mixture (Roche, Basel, Switzerland)) for 30 min on ice. USA). Medium was supplemented with 10% fetal calf serum The cell debris was removed by centrifugation at 15 000 g for (BioClear, Wiltshire, UK), 2 mM L-glutamin, 100 U/ml penicillin, 15 min at 4 1C. Equal amount of protein from each sample was 100 mg/ml streptomycin and 1 mM sodium pyruvate (Sigma- precleared with Sepharose CL-4B for 1,5 h at 4 1C. A total of Aldrich). Two days after LNCaP seeding the medium was 2.5 mg of monoclonal anti-TRAIL–R1 (Clone HS101; Alexis, San changed to phenol red free RPMI-1640 (Invitrogen, Carlsbad, Diego, CA, USA) and 2.5 mg of monoclonal anti-TRAIL–R2 CA, USA) with or without 1 nM synthetic androgen ([7a,17b]-17- (Clone HS201, Alexis) were added to samples and immunoprecipi- hydroxy-7,17-dimethylestr-4-en-3-one). This medium was supple- tated with 15 ml of protein G beads (Amersham Biosciences, mented with 10% charcoal/dextran-treated fetal bovine serum Buckinghamshire, UK) for 2.5 h at 4 1C. The beads were washed (Hyclone, Logan, UT, USA), L-glutamin, penicillin and strepto- six times with lysis buffer, resuspended in 3 Â Laemmli sample mycin. The cells were incubated with MAT, ENL, hydroxyma- buffer, and finally boiled for 5 min. The immunoprecipitated tairesinol or solvent control before apoptosis induction. Lignans samples and corresponding cell lysates (input) were analysed by were stored at þ 4 1C in 100 mM stock solutions in ethanol. 10% Criterion sodium dodecyl sulfate–polyacrylamide gel elec- trophoresis (BioRad, Hercules, CA, USA). Western blot analysis Plasmid construction and DNA transfection was performed using anti-FADD (BD Transduction Labora- The coding sequence for a constitutively active gag-Akt (kindly tories, Lexington, KY, USA), caspase-8 (clone C15; Alexis) and provided by Julian Downward) was amplified from the pSG5- anti-DR5 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). PKBGAG vector by PCR and cloned in-frame into pEGFP-N1 The results are representative of three independent experiments. vector (Clontech, Mountain View, CA, USA) in EcoRI and KpnI sites. LNCaP cells were seeded on 24-well plates 1 day Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and before lipofection and grown to 30–50% confluency. The immunoblotting JetPEI transfection complexes (PolyPlus-transfection SA, Ill- Whole cell lysates were prepared by lysing floating and attached kirch, France) were prepared and transfected according to the cells in Laemmli sample buffer and boiling the samples for manufacturer’s protocol. Empty pEGFP-N1 vector was 10 min after which proteins were separated with sodium dodecyl transfected to control cells. Twenty hours post-transfection, sulfate–polyacrylamide gel electrophoresis. Western blotting was

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 907 performed using antibodies against poly (ADP-ribose) polymer- Mitochondrial depolarization ase (clone C-2-10; Sigma-Aldrich), Bid, Bax, Bim Akt and Tetramethylrhodamine, methyl ester ( Invitrogen) was stored as phospho(Ser473)-Akt (Cell Signalling Technology, Davers, MA, 20 mM dimethylsulfoxide stock and diluted in medium before USA), BclXL (Santa Cruz Biotechnology), Actin (clone AC-40; use. Floating and trypsinized cells were incubated in 20 nM Sigma-Aldrich) and c-FLIP (clone NF6; Alexis). Horseradish tetramethylrhodamine for 10 min in 37 1C waterbath. The cells peroxidase-conjugated secondary antibodies were from Southern were placed on ice and analysed immediately by FACSCalibur Biotechnology Associates (Birmingham, AL, USA), Promega flow cytometer (BD Pharmingen). (Madison, WI, USA) and Amersham Biosciences. The results were visualized using the ECL method (Amersham Biosciences) on X-ray film. The presented western blot results are represen- Conflict of interest tative of at least three independent experiments. For densito- metric analysis of western blots, the X-ray films were scanned and the analysis was carried out with the MCID M5 þ software. The authors declare no conflict of interest. The values were normalized to the untreated control sample, which was given the value 1. The histogram presents a relative decrease in pAkt/Akt ratio compared with control cells. Acknowledgements

Tumor necrosis factor-related apoptosis-inducing ligand receptor We thank Henning Walczak for providing the recombinant cell surface expression izTRAIL, Julian Downward for the ca-Akt construct and Surface expression of TRAIL receptors was evaluated by Perttu Terho for technical assistance in flow cytometry. indirect immunostaining using the anti-TRAIL–R1 (clone We also thank Cecilia Sahlgren and all the members of our HS101), anti-TRAIL–R2 (clone HS201), anti-TRAIL–R3 (clone laboratory for critical comments on the paper and technical HS301) and anti-TRAIL–R4 (clone HS402) primary antibodies help during the course of this study. This work was supported followed by Alexa 488 conjugated anti-mouse secondary anti- by the Academy of Finland, the Sigrid Juse´lius Foundation, body (Alexis). Nonspecific fluorescence was assessed using the the Research Institute of the A˚bo Akademi University, the secondary antibody only. Flow cytometric analyses were Foundation of the A˚bo Akademi University, K Albin performed using a FACSCalibur and the mean fluorescence Johansson Foundation, Liv och Ha¨lsa Foundation and intensity of the FL-1 channel was measured. Magnus Ehrnrooth Foundation.

References

Adlercreutz H. (2007). Lignans and human health. Crit Rev Clin Lab Hausott B, Greger H, Marian B. (2003). Naturally occurring lignans Sci 44: 483–525. efficiently induce apoptosis in colorectal tumor cells. J Cancer Res Adlercreutz H, Bannwart C, Wahala K, Makela T, Brunow G, Hase T Clin Oncol 129: 569–576. et al. (1993). Inhibition of human aromatase by mammalian lignans and Huang H, Zegarra-Moro OL, Benson D, Tindall DJ. (2004). isoflavonoid phytoestrogens. J Steroid Biochem Mol Biol 44: 147–153. Androgens repress bcl-2 expression via activation of the Aslan JE, You H, Williamson DM, Endig J, Youker RT, Thomas L retinoblastoma (RB) protein in prostate cancer cells. Oncogene 23: et al. (2009). Akt and 14-3-3 control a PACS-2 homeostatic switch 2161–2176. that integrates membrane traffic with TRAIL-induced apoptosis. Jemal A, Thun MJ, Ries LA, Howe HL, Weir HK, Center MM et al. Mol Cell 34: 497–509. (2008). Annual report to the nation on the status of cancer, 1975– Axelson M, Sjovall J, Gustafsson BE, Setchell KD. (1982). Origin of 2005, featuring trends in lung cancer, tobacco use, and tobacco lignans in mammals and identification of a precursor from plants. control. J Natl Cancer Inst 100: 1672–1694. Nature 298: 659–660. Jin Z, McDonald ER, Dicker DT, El-Deiry WS. (2004). Deficient Bucur O, Ray S, Bucur MC, Almasan A. (2006). APO2 ligand/tumor tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) necrosis factor-related apoptosis-inducing ligand in prostate cancer death receptor transport to the cell surface in human colon cancer therapy. Front Biosci 11: 1549–1568. cells selected for resistance to TRAIL-induced apoptosis. J Biol Bylund A, Saarinen N, Zhang JX, Bergh A, Widmark A, Johansson A Chem 279: 35829–35839. et al. (2005). Anticancer effects of a plant lignan 7-hydroxymatair- Kim KM, Song JJ, An JY, Kwon YT, Lee YJ. (2005). Pretreatment of esinol on a prostate cancer model in vivo. Exp Biol Med (Maywood) acetylsalicylic acid promotes tumor necrosis factor-related apopto- 230: 217–223. sis-inducing ligand-induced apoptosis by down-regulating BCL-2 Chen LH, Fang J, Li H, Demark-Wahnefried W, Lin X. (2007). gene expression. J Biol Chem 280: 41047–41056. Enterolactone induces apoptosis in human prostate carcinoma Kischkel FC, Hellbardt S, Behrmann I, Germer M, Pawlita M, LNCaP cells via a mitochondrial-mediated, caspase-dependent Krammer PH et al. (1995). Cytotoxicity-dependent APO-1 (Fas/ pathway. MolCancerTher 6: 2581–2590. CD95)-associated proteins form a death-inducing signaling complex Chen LH, Fang J, Sun Z, Li H, Wu Y, Demark-Wahnefried W et al. (DISC) with the receptor. EMBO J 14: 5579–5588. (2009). Enterolactone inhibits insulin-like growth factor-1 Kitts DD, Yuan YV, Wijewickreme AN, Thompson LU. (1999). receptor signaling in human prostatic carcinoma PC-3 cells. J Nutr Antioxidant activity of the flaxseed lignan secoisolariciresinol 139: 653–659. diglycoside and its mammalian lignan metabolites enterodiol and Chen X, Thakkar H, Tyan F, Gim S, Robinson H, Lee C et al. (2001). enterolactone. Mol Cell Biochem 202: 91–100. Constitutively active akt is an important regulator of TRAIL Koschny R, Walczak H, Ganten TM. (2007). The promise of sensitivity in prostate cancer. Oncogene 20: 6073–6083. TRAIL—potential and risks of a novel anticancer therapy. J Mol Danial NN, Korsmeyer SJ. (2004). Cell death: critical control points. Med 85: 923–935. Cell 116: 205–219. Li D, Yee JA, Thompson LU, Yan L. (1999). Dietary supplementation Fulda S, Debatin KM. (2004). Signaling through death receptors in with secoisolariciresinol diglycoside (SDG) reduces experimental cancer therapy. Curr Opin Pharmacol 4: 327–332. metastasis of melanoma cells in mice. Cancer Lett 142: 91–96.

Oncogene Lignans sensitize prostate cancer cells to TRAIL E Peuhu et al 908 Lin X, Switzer BR, Demark-Wahnefried W. (2001). Effect of by TNFR family ligands via multiple signaling pathways in LNCaP. mammalian lignans on the growth of prostate cancer cell lines. Oncogene 24: 6773–6784. Anticancer Res 21: 3995–3999. Rokhlin OW, Taghiyev AF, Guseva NV, Glover RA, Syrbu SI, Mahalingam D, Szegezdi E, Keane M, Jong SD, Samali A. (2008). Cohen MB. (2002). TRAIL-DISC formation is androgen-dependent TRAIL receptor signalling and modulation: are we on the right in the human prostatic carcinoma cell line LNCaP. Cancer Biol Ther 1: TRAIL? Cancer Treat Rev 35: 280–288. 631–637. McCann MJ, Gill CI, Linton T, Berrar D, McGlynn H, Rowland IR. Saarinen NM, Warri A, Dings RP, Airio M, Smeds AI, Makela S.. (2008). Enterolactone restricts the proliferation of the LNCaP (2008). Dietary lariciresinol attenuates mammary tumor growth and human prostate cancer cell line in vitro. Mol Nutr Food Res 52: reduces blood vessel density in human MCF-7 breast cancer 567–580. xenografts and carcinogen-induced mammary tumors in rats. Int J McCann MJ, Gill CI, McGlynn H, Rowland IR. (2005). Role of Cancer 123: 1196–1204. mammalian lignans in the prevention and treatment of prostate Schaefer U, Voloshanenko O, Willen D, Walczak H. (2007). TRAIL: cancer. Nutr Cancer 52: 1–14. A multifunctional cytokine. Front Biosci 12: 3813–3824. Muzio M, Chinnaiyan AM, Kischkel FC, O’Rourke K, Shevchenko Sheridan JP, Marsters SA, Pitti RM, Gurney A, Skubatch M, Baldwin A, Ni J et al. (1996). FLICE, a novel FADD-homologous ICE/ D et al. (1997). Control of TRAIL-induced apoptosis by a family of CED-3-like protease, is recruited to the CD95 (Fas/APO-1) death– signaling and decoy receptors. Science 277: 818–821. inducing signaling complex. Cell 85: 817–827. Sprick MR, Walczak H. (2004). The interplay between the bcl-2 family Nesterov A, Lu X, Johnson M, Miller GJ, Ivashchenko Y, Kraft AS. and death receptor-mediated apoptosis. Biochim Biophys Acta 1644: (2001). Elevated AKT activity protects the prostate cancer cell 125–132. line LNCaP from TRAIL-induced apoptosis. J Biol Chem 276: Thakkar H, Chen X, Tyan F, Gim S, Robinson H, Lee C et al. (2001). 10767–10774. Pro-survival function of Akt/protein kinase B in prostate Ogasawara J, Watanabe-Fukunaga R, Adachi M, Matsuzawa A, cancer cells. relationship with TRAIL resistance. J Biol Chem 276: Kasugai T, Kitamura Y et al. (1993). Lethal effect of the anti-fas 38361–38369. antibody in mice. Nature 364: 806–809. Thompson LU, Seidl MM, Rickard SE, Orcheson LJ, Fong HH. Pan G, Ni J, Wei YF, Yu G, Gentz R, Dixit VM. (1997). An (1996). Antitumorigenic effect of a mammalian lignan precursor antagonist decoy receptor and a death domain-containing receptor from flaxseed. Nutr Cancer 26: 159–165. for TRAIL. Science 277: 815–818. Vincenz C, Dixit VM. (1997). Fas-associated death domain protein Raja SM, Chen S, Yue P, Acker TM, Lefkove B, Arbiser JL et al. interleukin-1beta-converting enzyme 2 (FLICE2), an ICE/Ced-3 (2008). The natural product honokiol preferentially inhibits cellular homologue, is proximally involved in CD95- and p55-mediated FLICE-inhibitory protein and augments death receptor-induced death signaling. J Biol Chem 272: 6578–6583. apoptosis. Mol Cancer Ther 7: 2212–2223. Vlietstra RJ, van Alewijk DC, Hermans KG, van Steenbrugge GJ, Ray S, Bucur O, Almasan A. (2005). Sensitization of prostate Trapman J. (1998). Frequent inactivation of PTEN in prostate carcinoma cells to Apo2 L/TRAIL by a bcl-2 family protein cancer cell lines and xenografts. Cancer Res 58: 2720–2723. inhibitor. Apoptosis 10: 1411–1418. Yokoyama T, Okano M, Noshita T, Funayama S, Ohtsuki K. (2003). Rokhlin OW, Taghiyev AF, Guseva NV, Glover RA, Chumakov PM, Characterization of (À)-matairesinol as a potent inhibitor of casein Kravchenko JE et al. (2005). Androgen regulates apoptosis induced kinase I in vitro. Biol Pharm Bull 26: 371–374.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene