Curcumin Downregulates Cell Survival Mechanisms in Human Prostate Cancer Cell Lines

Curcumin downregulates cell survival mechanisms in human prostate cancer cell lines

Asok Mukhopadhyay1, Carlos Bueso-Ramos2, Devasis Chatterjee3, Panayotis Pantazis3 and Bharat B Aggarwal*,1

1Cytokine Research Laboratory, Department of Bioimmunotherapy, The University of Texas M. D. Anderson Cancer Center, Houston, Texas, TX 77030, USA; 2Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas, TX 77030, USA; 3Department of Molecular Biology, Cell Biology and Biochemistry, Brown University, Providence, Rhode Island, RI 02912, USA

While the role of nuclear transcription factor activator Introduction protein-1 (AP-1) in cell proliferation, and of nuclear factor-kB (NF-kB) in the suppression of apoptosis are Prostate cancer is the most prevalent malignancy in known, their role in survival of prostate cancer cells is men in the United States, and leads to the highest not well understood. We investigated the role of NF-kB number of cancer-related deaths among them (Parker and AP-1 in the survival of human androgen-independent et al., 1997). Eorts to improve prognosis requires the (DU145) and -dependent (LNCaP) prostate cancer cell understanding of the role of androgens in prostate lines. Our results show that the faster rate of cancer. Although androgens are required for the proliferation of DU145 cells when compared to LNCaP normal growth and development of prostate, they cells correlated with the constitutive expression of may also play a role in prostate carcinogenesis by activated NF-kB and AP-1 in DU-145 cells. The lack acting either as initiators or promoters. Most patients of constitutive expression of NF-kB and AP-1 in LNCaP respond initially to androgen-ablative therapies; the cells also correlated with their sensitivity to the regression of the prostate tumors indicates that antiproliferative eects of tumor necrosis factor (TNF). androgen functions as a survival factor. However, the TNF induced NF-kB activation but not AP-1 activation disease generally recurs within 1 to 3 years of in LNCaP cells. In DU145 cells both c-Fos and c-Jun treatment, and the recurrent tumors no longer require were expressed and treatment with TNF activated c-Jun androgen for growth or survival. Experimental evi- NH2-terminal kinase (JNK), needed for AP-1 activation. dence demonstrates a clonal outgrowth of androgen- In LNCaP cells, however, only low levels of c-Jun was independent prostate cancer cells from androgen- expressed and treatment with TNF minimally activated dependent tumors (Craft et al., 1999a). In addition, JNK. Treatment of cells with curcumin, a chemopreven- these tumors tend to be highly resistant to conventional tive agent, suppressed both constitutive (DU145) and cytotoxic agents such as cisplatin. Presently available inducible (LNCaP) NF-kB activation, and potentiated treatments for advanced hormone-resistant prostate TNF-induced apoptosis. Curcumin alone induced apop- cancer are marginally eective, and so newer agents are tosis in both cell types, which correlated with the needed to selectively kill the cancer cells. downregulation of the expression of Bcl-2 and Bcl-xL How prostate cancer cells develop resistance is not and the activation of procaspase-3 and procaspase-8. fully understood, but Bcl-2 (Rao et al., 1995; Overall, our results suggest that NF-kB and AP-1 may Herrmann et al., 1997a; McConkey et al., 1996), play a role in the survival of prostate cancer cells, and peroxisome proliferator-activated receptor-g (Kubota curcumin abrogates their survival mechanisms. Oncogene et al., 1998), altered expression of nuclear phospho- (2001) 20, 7597 ± 7609 protein pp32 (Kadkol et al., 1999), defective androgen receptor signaling (Craft et al., 1999b), ubiquitously Keywords: Prostate; curcumin; NF-kB; AP-1; Bcl-2; secreted glycoprotein SGP2 (also called clusterin) androgen (Sintich et al., 1999), mutated or nonfunctional p53 (van Brussel et al., 1999; Ewing et al., 1995), nonspeci®c activation of proteosome (Adams et al., 1999; Herrmann et al., 1998), expression of low-anity insulin-like growth factor-binding protein (Degeorges *Correspondence: BB Aggarwal, Cytokine Research Section, et al., 1999), inactivation of tumor suppressor gene Department of Bioimmunotherapy, University of Texas MD PTEN/MMAC (Davies et al., 1999), phosphotidylino- Anderson Cancer Center, 1515 Holcombe Boulevard, Box 143, sitol 3'-kinase (Lin et al., 1999), and high constitutive Houston, Texas 77030, USA; levels of cyclooxygenase (COX)-2 expression (Liu et E-mail: [email protected] Received 27 February 2001; revised 6 September 2001; accepted 18 al., 1998) have been implicated as survival factors for September 2001 prostate cancer. In contrast, mitogen-activated protein Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7598 kinase kinase kinase-1, prostate apoptotic response-4 androgen-dependent cells but not androgen-indepen- (Nalca et al., 1999) and a transcription factor, early dent cells. growth response-1, have been shown to mediate Several cell-survival proteins have been implicated apoptosis in prostate cancer cells (Abreu-Martin et which can induce TNF-resistance (Aggarwal, 2000). It al., 1999; Ahmed et al., 1997). is possible that DU145 cells synthesize the proteins that Within the last few years, several reports have can induce resistance to TNF. Therefore, LNCaP and suggested that nuclear transcription factors NF-kB DU145 were exposed to TNF in the presence of and AP-1 can also act as survival factors and are cycloheximide, an inhibitor of protein synthesis. The required for the proliferation of a variety of dierent results in Figure 1c show that TNF was now cytotoxic tumor cell types (Colotta et al., 1992; Karin et al., to DU145 cells and the level of cytotoxicity was similar 1997; Miyamoto and Verma, 1995; Bargou et al., to that observed with LNCaP cells. Furthermore, when 1997). For instance, the proliferation of Hodgkin's examined for TNF-induced caspase activation by disease is dependent on the constitutive activation of PARP cleavage, both cell types were found to be NF-kB (Bargou et al., 1997). There are some studies equally sensitive. Thus these results suggest that the to suggest NF-kB activation may also play a role in resistance of DU145 cells to TNF is due to synthesis of the proliferation of prostate cancer cells (Herrmann et antiapoptotic proteins. These results also indicate that al., 1997b). Why some prostate cancer cells are TNF signaling machinery is intact in DU145 cells. sensitive to TNF-induced apoptosis while others are resistant is not understood. We hypothesized that this LNCaP and DU145 cells differ for NF-kB dierence is related to the expression of NF-kBand AP-1 in prostate cancer cells. Therefore, in the Because NF-kB has been implicated as a cell-survival present studies, we examined in detail the role of signal (for references see Aggarwal, 2000) and several both constitutive and inducible NF-kB and AP-1 in recent reports suggest that NF-kB activation protects TNF-induced apoptosis in androgen-dependent cells from TNF-induced apoptosis, we investigated the LNCaP and androgen-independent DU145 prostate possibility that DU145 cells resist the antiproliferative cancer cells. Our results indicate that DU145 cells eects of TNF because of their constitutive expression constitutively express the activated form of NF-kB of NF-kB. Therefore, NF-kB activation was examined and AP-1 and were resistant to TNF-induced using EMSA in LNCaP and DU145 cells. As shown in apoptosis, whereas LNCaP cells did not express Figure 2a, LNCaP cells did not express constitutive constitutive NF-kB and AP-1 and were sensitive to NF-kB but DU145 did. Thus, constitutive expression TNF-induced apoptosis. Curcumin, a chemopreventive of NF-kB in DU145 cells may be responsible for the agent, suppressed the constitutive NF-kB and AP-1 in induction of resistance to TNF. DU145 cells and inducible NF-kB in LNCaP cells, TNF is one of the most potent activators of NF-kB. which correlated with augmentation of the TNF- Whether TNF could activate NF-kB in LNCaP cells induced apoptosis. was examined. For this, cells were treated with 0.1 nM of TNF for 30 min and then examined for NF-kB using EMSA. As shown in Figure 2a (right panel), Results TNF induced NF-kB in LNCaP cells. Interestingly, TNF induced NF-kB in the DU145 cells even more. The molecular changes that occur when androgen- The NF-kB induced in prostate cell lines contained dependent prostate tumors become androgen-indepen- p65 (RelA) and p50 subunits, inasmuch as incubation dent is not very well understood. We used androgen- of nuclear extracts with either anti-p65 (Figure 2b) or dependent LNCaP and androgen-independent DU145 anti-p50 antibodies resulted in abrogation in the NF- cell lines to dissect the molecular dierences. kB/DNA complex, whereas preimmune sera had no eect. The speci®city of the TNF-induced NF-kB/ DNA complexes was further con®rmed by demonstrat- TNF inhibits the proliferation of LNCaP cells but not of ing that the binding was abolished in the presence of DU145 cells 100-fold excess unlabeled kB-oligonucleotide, but not We ®rst investigated the rate of proliferation of by mutant oligonucleotide (Figure 2b). prostate cancer cell lines LNCaP and DU145 in the To investigate whether NF-kB in DU145 and LNCaP absence and presence of TNF. To determine this, cells cells was transcriptionally active, NF-kB-dependent were cultured for 2, 4, and 6 days in the absence or reporter activity assay was performed. As shown in presence of 10 nM TNF and then examined for cell Figure 2c (left panel) without any treatment about six- viability by the trypan blue exclusion method. As fold increase in NF-kB-dependent reporter activity over shown in Figure 1, in the absence of TNF the rate of the vector control was observed in DU145 cells, which proliferation of DU145 cells was about two-times increased to 9.5-fold when cells were treated with TNF. faster than LNCaP cells (approximately 30 000 vs On the other hand, LNCaP cells showed a 3.7-fold 17 000 on day 6). The presence of TNF inhibited the increase in NF-kB-dependent reporter activity only after proliferation of LNCaP cells (a) but had little eect on treatment with TNF. These results suggest that both the proliferation of DU145 cells (b). These results constitutive and TNF-inducible NF-kB in prostate cells indicate that TNF inhibits the proliferation of were transcriptionally active.

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7599

Figure 1 TNF inhibits the proliferation of LNCaP cells (a) but not DU145 cells (b). 26103 cells in 0.2 ml in quadruplicate were cultured for 2, 4 and 6 days in a 48-well plate in the presence and absence of 10 nM TNF. After every 2-day interval, the cells were trypsinized, and the viable cell number was determined using the trypan blue dye exclusion method. (c) cytotoxic eect of TNF on LNCaP and DU145 cells in presence of CHX. 106103 cells in 0.1 ml in triplicate were cultured in a 96-well plate in the presence of 10 mg/ml CHX alone or with variable concentrations of TNF. After 24 h, cell viability was determined by the MTT method. (d) TNF-induced degradation of PARP in DU145 and LNCaP cells in presence of CHX. 16106 cells in 1 ml were pretreated with 10 mg/ml CHX for 2 h, and then exposed to none, 0.1, 1, and 10 nM TNF. After 6 h, whole-cell extracts were prepared and analysed for PARP cleavage as mentioned in Materials and methods

constitutive AP-1 activity, which was further enhanced LNCaP and DU145 cells differ for AP-1 by TNF treatment. The DNA-binding assay revealed The AP-1/c-jun gene product is a nuclear protein that the absence of both constitutive and inducible AP-1 in forms a heterodimer with the c-fos gene product to LNCaP cells (Figure 3a). stimulate transcription of genes that contain AP-1/c-jun The lack of either constitutive or inducible AP-1 in response elements (Karin et al., 1997). Expression of LNCaP cells prompted us to examine the expression of these gene products have been implicated for the G0/ two essential components of AP-1, c-Fos and c-Jun. G1 transition of the cell cycle, and they have been For this, nuclear extracts prepared from both cell types reported to protect cells from apoptotic death (Dixit et were probed by Western blot using anti-c-Jun and anti- al., 1989). We therefore analysed nuclear extracts of c-Fos antibodies. DU145 cells expressed a signi®cant DU145 and LNCaP for AP-1, and the results are amount of both c-Jun and c-Fos proteins, but LNCaP shown in Figure 3a. DU145 prostate cells showed cells expressed only small amounts of c-Jun and no c-

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7600

Figure 2 NF-kB is constitutively active in DU145 cells, whereas it is TNF-inducible in LNCaP cells (a). Cells were cultured in 6- well plates until 80% con¯uence, and were either left untreated or were induced with 0.1 nM TNF for 30 min. We then prepared the nuclear extracts and analysed 8 mg protein for NF-kB by EMSA as described in Materials and methods. (b) Supershift and speci®city analysis of NF-kB. Nuclear extracts were incubated for 15 min at 378C with either medium, anti-p50 antibodies, anti-p65 antibodies, pre-immune sera, unlabeled oligonucleotide or mutant oligonucleotide, and then assayed for NF-kB using EMSA. (c) NF-kB is transcriptionally active in DU145 cells, whereas it is TNF-inducible in LNCaP cells. 1.26106 cells/ml were transfected with pCMVFlag1 (empty vector) alone or with NF-kB-SEAP plasmid for 18 h. Cells were washed with medium and incubated with 1 ml of fresh medium in absence or in presence of 1 nM TNF. After 24 h of incubation, culture supernatants were analysed for SEAP activity as described in Materials and methods

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7601

Figure 3 DU145 cells express constitutive AP-1 but LNCaP cells express neither constitutive nor inducible AP-1. Cells were cultured in a 6-well plate until 80% con¯uence, and then were either left untreated or were treated with 0.1 nM TNF for 30 min. We prepared the nuclear extracts and analysed the 8 ug protein for AP-1 using EMSA as described in Materials and methods. (b), Western blot analysis of c-jun and c-fos. Prostate cancer cell lines were cultured in 100 mm plate until 80% con¯uence, then prepared nuclear extracts, analysed 60 ug protein on 9% SDS ± PAGE, and probed with appropriate antibodies. (c) Kinetics of TNF-induced JNK activation in prostate cancer cell lines. DU-145 and LNCaP cells were cultured in 6-well plate until 80% con¯uence and then induced with 0.1 nM TNF for 0, 10, 20, and 30 min. One hundred micrograms of whole cell lysates were used for kinase assay, as mentioned in Materials and methods. The lower panel shows equal loading

Fos (Figure 3b). This may explain why EMSA results 1991). To determine whether androgen-dependent and showed no constitutive or inducible AP-1 in LNCaP -independent cells dier in their activation of JNK cells. This is the ®rst report to indicate that LNCaP (which if so might explain the dierence in AP-1 cells lack AP-1 and c-Fos expression. activity), we prepared cell extracts from untreated and TNF-induced prostate cell lines and analysed them for JNK activity. As shown in Figure 3c, both cell lines LNCaP and DU145 cells differ for JNK showed some basal level of JNK activity. On treatment The amino-terminal activation domain (Ser-63 and Ser- with TNF, a sixfold increase in kinase activity was 73) of c-jun is phosphorylated by a JNK, thereby observed in DU145 cells, whereas the activation in increasing its transcriptional activity (Pulverer et al., LNCaP cells was marginal (less than twofold).

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7602 results thus suggest that curcumin alone can induce Curcumin inhibits both constitutive and inducible NF-kB apoptosis. and AP-1 NF-kB has been implicated as an antiapoptotic Curcumin potentiates the anti-proliferative effect of TNF transcription factor, which may override the eects of apoptosis pathways in some cells (Van Antwerp et al., Whether or not the suppression of NF-kB and NF-kB- 1996; Wang et al., 1996; Beg and Baltimore, 1996). Our regulated gene products by curcumin leads to the laboratory and several others have shown that potentiation of the antiproliferative eects of TNF was curcumin is a potent inhibitor of NF-kB (Singh and investigated. As shown in Figure 6a, under the same Aggarwal, 1995) and AP-1 (Huang et al., 1991). conditions in which TNF inhibited the proliferation of Whether curcumin can suppress constitutive NF-kB LNCaP cells by only 25% in the absence of curcumin, and AP-1 in DU145 cells and inducible NF-kBin the presence of 10 mM curcumin potentiated the LNCaP cells was investigated. The results in Figure 4a inhibition by almost 75%, suggesting synergism indicate that TNF induced NF-kB in LNCaP cells and between TNF and curcumin. In the case of DU145 curcumin completely suppressed TNF induced NF-kB cells, the potentiation was less dramatic, as the activation in a dose-dependent manner, with maximum cytotoxicity was enhanced from 15% in the absence suppression occurring at around 25 mM. When exam- of curcumin to 35% in the presence of curcumin ined for constitutive NF-kB in DU145 cells, curcumin (Figure 6b). completely suppressed it at around 75 mM (Figure 4b). The cytotoxic eects of TNF are mediated, in part, The dierence in the concentration of curcumin through the activation of caspases, which can be required may depend on the mode of activation of monitored through the cleavage of a PARP substrate. NF-kB, as it is inducible in LNCaP cells and Whether curcumin potentiates TNF-induced caspase constitutive in DU145 cells. activation was examined. As shown in Figure 6c, where Curcumin also completely suppressed AP-1 in both TNF and curcumin by themselves failed to induce DU145 cells at around 50 mM (Figure 4c). While any signi®cant PARP cleavage in LNCaP cells, the suppressing AP-1, curcumin induced another band addition of the two together induced PARP cleavage, below the AP-1 band (Figure 4c). Both the super-shift indicating synergistic eects. In spite of their ability to analysis and the cold-competition revealed that AP-1 suppress cell growth, neither TNF, nor curcumin, nor band consisted of c-Jun and c-Fos (Figure 4d). The the combination of the two had any eect on PARP band below the AP-1 band, however could not be cleavage in DU145 cells (Figure 6d). supershifted by antibodies against c-Jun and c-Fos, indicating that it is a non-speci®c band (NSB). These Curcumin alone induces apoptosis in both types of results indicate that curcumin suppresses both NF-kB prostate cancer cells and AP-1 in prostate cancer cells. Previously we have shown that curcumin inhibits the growth of breast tumor cell lines (Mehta et al., 1997). Curcumin downregulates the expression of antiapoptotic Because curcumin downregulates NF-kB and NF-kB- gene products and activates caspases dependent gene products and activates the caspases in Whether suppression of NF-kB by curcumin leads to prostate cancer cells, we investigated whether curcumin the suppression of NF-kB-regulated gene products was could inhibit the proliferation of these cells. As shown examined. It has been shown that the expression of in Figure 7a, while 10 mM alone had no signi®cant Bcl-2 and Bcl-xL are regulated by NF-kB (Tamatani et eect on the proliferation of either LNCaP or DU145 al., 1999; Hinz et al., 1999). Therefore, the expression cells, 50 mM completely inhibited the proliferation of of these gene products after treating the cells with both cell types. Curcumin also induced caspase- 50 mM curcumin was examined using Western blot dependent cleavage of PARP in a time-dependent analysis. Results indicate that curcumin downregulated manner in both cell types (Figure 7b). The apoptosis the expression of both Bcl-2 and Bcl-xL in LNCaP and of prostate cells by curcumin was further examined by DU145 cells in a time-dependent manner (Figure 5a,b). DNA staining with propidium iodide. The kinetics of The downregulation of the expression of both Bcl-2 apoptosis in LNCaP and DU145 prostate carcinoma and Bcl-xL was not a result of apoptosis of these cells, cells after exposure to curcumin are shown in Figure 8. because the structural protein b-actin was unaltered Curcumin-induced apoptosis in LNCaP cells could be under these conditions. observed as early as 24 h (about 56% PI-positive We also investigated whether curcumin could cells), which reached maximum at 48 h (more than activate the proapoptotic procaspase-3 and procas- 80%). The kinetics of apoptosis by curcumin in pase-8 in prostate cancer cells. As shown in Figure 5, DU145 cells, however, was slower with only 28 and 100 uM curcumin activated both the procaspase-3 (c) 69% cells converted to PI-positive at 48 and 72 h, and procaspase-8 (d) in LNCaP and in DU145 cells. respectively (Figure 8). Microscopic observation re- When examined for the enzymatic activity of caspase-3 vealed plasma membrane blebbing, cytoplasmic and caspase-8, curcumin (50 uM) was found to activate vacuoles, and chromatin condensation in curcumin both the caspases (Figure 5e) in LNCaP cells (left treated LNCaP and DU145 prostate carcinoma cells panel) as well as in DU145 cells (right panel). These (data not shown).

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7603

Figure 4 Curcumin blocks NF-kB and AP-1 activation in prostate cancer cells. A, LNCaP and DU145 cells were cultured in two dierent 6-well plates until 80% con¯uence. LNCaP cells were treated with 0, 1, 5, 10, 25, 50, 75, and 100 mM curcumin for 5 h, and then induced with 0.1 nM TNF for 30 min. Eight cells were treated with 0, 1, 5, 10, 25, 50, 75, and 100 mM curcumin for 5 h. Eight micrograms nuclear protein was analysed for NF-kB by EMSA. (c) DU-145, cells were treated with 0, 10, 25, 50, and 100 mM curcumin for 5 h and then induced with 0.1 nM TNF for 30 min. Eight micrograms of nuclear protein was analysed for AP-1 using EMSA. (d), AP-1 is composed of c-Jun and c-Fos. DU145 cells were treated with 10 mM curcumin for 5 h, nuclear extract was incubated with none, or anti-c-Jun, or anti-c-Fos, or preimmune sera, or cold oligo at 378C for 30 min. Following incubation, samples were analysed by EMSA as mentioned in Materials and methods

Discussion LnCaP cells correlates with a faster proliferation of DU145 cells than LNCaP cells. It also correlated with The mechanism by which prostate cancer cells develop the sensitivity of LNCaP cells and the resistance of resistance to various treatments is not well understood. DU145 cells to the antiproliferative eects of TNF. We investigated the role of NF-kB and AP-1 in the TNF induced NF-kB activation but not AP-1 in induction of resistance to TNF in human androgen- LNCaP cells. While both c-Fos and c-Jun were independent (DU145) and -dependent (LNCaP) pros- expressed in DU145 cells, LNCaP cells expressed only tate cancer cell lines. Our results show that constitu- c-Jun. Also, TNF activated JNK needed for AP-1 tively active NF-kB and AP-1 in DU-145 but not in activation in DU145 cells but not in LnCaP cells.

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7604

Figure 5 Curcumin downregulates the expression of Bcl-2 (a), and Bcl-xL (b), and activates procaspases-3 (c) and procaspase-8 (d). LNCaP and DU145 cells were treated with either 50 mM (a,b) or 100 mM (c,d) curcumin for the indicated times, then prepared the whole-cell extracts and analysed them by Western blot using speci®c antibodies as indicated. (e) curcumin activates caspase-3 and caspase-8 activity in LNCaP cells (left panel) and in DU145 cells (right panel). 36106 cells in 4 ml were treated with 50 mM curcumin for 0, 12, 24, 48, and 72 h. Thereafter, the whole-cell extracts were prepared and 100 mg protein samples were analysed using colorimetric assay as described in Materials and methods

Curcumin, a chemopreventive agent, suppressed both Our results indicate that there are several molecular constitutive (DU145) and inducible (LnCaP) NF-kB dierences between LNCaP and DU145 cells. The ®rst and constitutive (DU145) AP-1 activation, and poten- major dierence is due to the constitutive NF-kBin tiated TNF-induced apoptosis. Curcumin alone in- DU145 cells but not in LNCaP cells. Why NF-kBis duced apoptosis that correlated with the activation of constitutively active in DU145 cells is not clear. It is caspase-3 and caspase-8 and the downregulation of the possible that these cells either have a faster rate of expression of anti-apoptotic proteins, Bcl-2 and Bcl-xL. IkBa (an inhibitor of NF-kB) degradation as seen in B

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7605

Figure 6 Curcumin potentiates the apoptotic eects of TNF in LNCaP and DU145 cells. (a,b) 2.56103 cells in 0.1 ml were plated in triplicate in a 96 well plate, allowed to adhere for 12 ± 14 h, and then exposed to dierent concentrations of TNF alone (open circle) or TNF and 10 mM of curcumin (solid circle). After 72 h, cell viability was determined using the MTT method as described in Materials and methods. (c,d) Induction of PARP cleavage by TNF and curcumin. 16106 cells were treated with either 10 mM curcumin, 3 nM TNF or the indicated combination for 48 h. Then the whole-cell extracts were prepared and analysed by Western blot using anti-PARP antibodies as described in Materials and methods cells (Miyamoto et al., 1994) or these cells produce ratio of AP-1 (positive growth factor) to a protein TNF, a potent activator of NF-kB, as seen in T cells known as QM or Jif (negative growth factor). The (Giri and Aggarwal, 1998). Because constitutive active protein QM interacts with c-Jun and ®nally suppresses NF-kB mediates the proliferation of various tumor AP-1 activity (Colotta et al., 1992; Karin et al., 1997). cells (Bargou et al., 1997; Giri and Aggarwal, 1998; Because AP-1 mediates the proliferation of various Aggarwal, 2000), this might explain why DU145 cells tumor cells (Dixit et al., 1989), this may also explain grow at a faster rate than LNCaP cells. The second why DU145 cells grow at a faster rate than LNCaP dierence between the two cell types, is the constitutive cells. AP-1 activity in DU145 cells but not in LNCaP cells. A third major dierence noted between androgen- Why DU145 cells express constitutive AP-1 is also not dependent LNCaP cells and androgen-independent clear. LNCaP cells, however, do not have constitutive DU145 cells was their ability to undergo apoptosis in AP-1, or it is inducible with TNF. As AP-1 consists response to TNF. LNCaP cells were quite sensitive to primarily of c-Fos and c-Jun, c-Jun was marginally TNF whereas DU145 cells were quite resistant. This expressed in LNCaP cells whereas c-Fos was not at all dierence may also be linked to the constitutively expressed. Why LNCaP cells do not express c-Fos is active NF-kB and AP-1 in DU145 cells and not in not clear. It has been reported that the repression of c- LNCaP cells. That NF-kB activation could down- Fos causes a decline in AP-1 activity by altering the regulate TNF-induced apoptosis has been demon-

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7606

Figure 7 Curcumin inhibits the proliferation of both LNCaP and DU145 cells in a time- and dose-dependent manner. (a,b)26103 cells in 0.2 ml were plated in triplicate in 96-well plates, allowed to adhere for 12 ± 14 h, exposed to 10 mM and 50 mM curcumin for dierent number of days, and then measured the cell viability using the MTT method. (c,d), Induction of PARP cleavage by curcumin. 16106 cells were treated with 50 mM curcumin for indicated times. Then the whole-cell extracts were prepared and analysed by Western blot using anti-PARP antibodies as described in Materials and methods

strated by several groups in a variety of tumors (Van constitutive AP-1 activity in DU145 cells. Alterna- Antwerp et al., 1996; Wang et al., 1996; Beg and tively, curcumin could also act through an antioxidant Baltimore, 1996; Giri and Aggarwal, 1998). The role of mechanism because antioxidants have been shown to NF-kB in preventing TNF-induced apoptosis in block androgen-induced AP-1 and NF-kB activity in prostate cancer cells has also been reported (Sumitomo prostate carcinoma cells (Ripple et al., 1999). We et al., 1999). Similarly, the role of AP-1 in the found that suppression of NF-kB and AP-1 in prostate induction of resistance to apoptosis has been docu- cancer cells enhanced TNF-induced apoptosis. These mented (Karin et al., 1997; Dixit et al., 1989). results are consistent with reports that suggest that We found that curcumin treatment suppressed both NF-kB and AP-1 have an antiapoptotic role. constitutive NF-kB and AP-1 activity in DU145 cells. Our results indicate that the suppression of NF-kB Curcumin also abolished the TNF-inducible NF-kB activation by curcumin correlates with the down- activity in LNCaP cells. These results are in agreement regulation of expression of Bcl-2 and Bcl-xL in both with previous reports on other tumor cell types (Singh prostate cancer cell lines. The expression of Bcl-2 and and Aggarwal, 1995; Huang et al., 1991; Kumar et al., Bcl-xL have been shown to be regulated by NF-kB 1998). How curcumin downregulates NF-kB is not (Tamatani et al., 1999; Hinz et al., 1999). Both Bcl-2 clear, but the suppression of IKK activity, the kinase and Bcl-xl have been reported to inhibit apoptosis by required for IkBa phosphorylation and degradation, suppressing the activation of caspases (Rao et al., has been demonstrated (Jobin et al., 1999). Curcumin 1995; Herrmann et al., 1997a). The downregulation of is also known to inhibit the activation of JNK (Chen Bcl-2 and Bcl-xL by curcumin also coincided with the and Tan, 1998), the kinase required for AP-1 activation of both caspase-8 and caspase-3 acting at activation, which may explain how curcumin blocked upstream and downstream sites, respectively, in the

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7607 regulate the cell survival mechanisms in prostate cancer and other cell types is very appealing. Lack of any known pharmacological toxicity (Li et al., 1993 and references therein) prompts future investigation into the downregulation of tumor cell survival mechanisms.

Materials and methods

Materials Human prostate cell lines (DU145 and LNCaP) were originally obtained from the American Type Culture Collection (Rockville, MD, USA). The rabbit polyclonal antibodies anti-IkBa, anti-p50, anti-p65, anti-JNK1, anti-c- jun, and anti-procaspases-8 and -3, and the anti-c-Fos and c- Jun antibodies were procured from Santa Cruz Biotechnol- ogy (Santa Cruz, CA, USA). The anti-PARP rabbit polyclonal antibody was purchased from New England Bio Figure 8 Time-dependent curcumin-induced apoptosis of Labs (Beverly, MA, USA). Curcumin and MTT were DU145 and LNCaP cells. 16106 cells in two ml were treated purchased from Sigma Chemicals (St. Louis, MO, USA). with 50 mM curcumin for 0, 12, 24, 48, and 72 h. Thereafter, cells Bacteria-derived recombinant human TNF, puri®ed to were trypsinized, washed, labeled with PI, and percentage PI- homogeneity with a speci®c activity of 56107 units/mg, positive cells were counted in a ¯uorescence microscope was kindly provided by Genentech, Inc. (South San Francisco, CA, USA). RPMI-1640, MEM, and FBS were procured from Life Technologies, Inc. (Grand Island, NY, USA). Protein A/G-sepharose beads were obtained from pathway leading to apoptosis. Curcumin alone at high Pierce (Rockford, IL, USA). concentration caused the suppression of proliferation A20mM of stock solution of curcumin was prepared in in both cell types and the caspase-dependent cleavage dimethyl sulfoxide and then further diluted in cell culture of PARP substrate, suggesting that curcumin can medium. induce apoptosis. Because curcumin is known to inhibit cyclooxygenase-2 (COX-2) expression in colon cancer Cell culture cells (Plummer et al., 1999) and suppression of COX-2 activity in LNCaP cells by NS398, a COX-2-selective DU145 and LNCaP cells were maintained in RPMI-1640 that inhibitor, causes apoptosis (Liu et al., 1998), it is contained 10% FBS and a 16 antibiotic-antimycotic. Each possible that the apoptotic eects of curcumin cell line was split regularly before it attained 70 ± 80% con¯uence. observed in our system are mediated through the downregulation of COX-2. Morphological evidence (not shown), and DNA staining with propidium iodide EMSA further con®rmed that curcumin is a potent inducer of NF-kB activation was analysed using EMSA as described apoptosis in both androgen-dependent as well as previously (Chaturvedi et al., 2000). In brief, 8 mg of nuclear androgen -independent prostate carcinoma cells. It is extracts prepared from the TNF-treated/untreated cells were quite likely that this apoptosis is mediated through the incubated with 32P-end-labeled 45-mer double-stranded NF- downregulation of NF-kB and AP-1, leading to the kB oligonucleotide (16 fmoles DNA) from the HIV-LTR, 5'- suppression of Bcl-2 and Bcl-xL, and resulting in the TTGTTACAAGGGACTTTCCGCTG GGGACTTTCCAG- activation of caspases and apoptosis. Alternatively, it is GGA GGCGT GG-3' (underlined indicates NF-kB binding sites) for 15 min at 378C, and the DNA-protein complex also possible that curcumin-induced apoptosis is not formed was separated from free oligonucleotide on 6.6% mediated through the downregulation of NF-kBand native polyacrylamide gels. A double-stranded mutated AP-1, as LNCaP do not constitutively express these oligonucleotide, 5'-TTGTTACAACTCACTTTCCGCTGCT transcription factors. Because downregulation of Bcl-2 CACTTTCCAGGGAGG CGTGG-3', was used to examine and Bcl-xL in both cell types precede apoptosis, the speci®city of binding of the NF-kB to the DNA. The suggests that curcumin may induce apoptosis through speci®city of binding was also examined by competition with modulation of antiapoptotic proteins. the unlabeled oligonucleotide. The composition and speci®- Curcumin, a major active component of turmeric, city of binding was further determined by supershift of DNA- has been reported to induce growth inhibition or protein complex using speci®c antibodies and preimmune apoptosis in many cancerous cells (Mehta et al., 1997; sera. The dried gels were visualized and radioactive bands quantitated using a PhosphorImager (Molecular Dynamics, Kuo et al., 1996; Jiang et al., 1996; Chen and Huang, Sunnyvale, CA, USA). 1998), by a mechanism that is not fully understood. AP-1 activation was analysed in the same way as that for Our results suggest a mechanism by which curcumin NF-kB except that the nuclear extracts were incubated with may mediate its eects. The chemopreventive activity 32P-end-labeled double-stranded oligonucleotide of AP-1 of curcumin is well documented (Kawamori et al., (Manna et al., 1998). The AP-1/DNA complex was analysed 1999). The fact that curcumin can be used to down- on a 5% native polyacrylamide gel.

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7608 using the MTT method as described earlier (Haridas et al., NF-kB-dependent reporter assay 1998). The cells were plated in triplicate in 96-well plates and The NF-kB-SEAP reporter gene expression assay was based allowed to attach for overnight. Thereafter, the medium was on our earlier report (Darnay et al., 1999). In brief, 1.26106 replaced that contained TNF or curcumin or in combination. DU145 or LNCaP cells/1.5 ml were plated in 6-well plates After 24 h or 72 h incubation at 378C, the cell viability was and incubated for 16 h. Cells were then transiently examined by the MTT method. transfected for 18 h with the expression vector (2 mg pCMVFlag1) or pNF-kB-SEAP2 (0.5 mg) by the FuGENE Propidium iodide staining 6 transfection reagent. After transfection, cells were washed and incubated with medium in absence or in presence of LNCaP and DU145 cells were exposed to curcumin for 0 h, 1nM TNF for 24 h. The culture supernatant was removed 12 h, 24 h, 48 h, and 72 h. Thereafter cells were trypsinized, and assayed for SEAP activity. The culture supernatant washed twice with PBS containing 2% BSA and resuspended (25 ml) was mixed with 30 mlof56buer (500 mM Tris Cl, in 500 ml PBS. One-hundred microliter of cell suspension was pH 9, and 0.5% bovine serum albumin) in a total volume of treated with 3 ml of PI (Santa Cruz Biotechnology, Santa 100 ul in a 96-well plate, and the heat-labile endogenous Cruz, CA, USA) for 20 min in dark. Twenty-®ve microliter alkaline phosphatase deactivated by heating the mixture at of cell suspension was applied to a glass slide and covered 658C for 30 min. The plate was chilled on ice for 2 min, 50 ml with a cover-slip. Number of PI-negative cells (in a normal of 1 mM 4 methylumbelliferyl phosphate was added to each light source) and PI-positive cells (in a ¯uorescence source) well, the plates incubated at 378C for 2 h, and ¯uorescence were counted from three representative ®elds and percentage read on a 96-well ¯uorescent plate reader (Fluoroscan II, Lab PI positive cells was determined. Systems, Needham, Heights, MA, USA) with excitation set at 360 nm and emission at 460 nm. The average (+s.e.m.) Assay of caspase-3 and -8 number of relative ¯uorescent light units for each transfection was determined. The caspase assay was essentially based on the manufac- turer's protocol (R&D Systems, Minneapolis, MN, USA). In brief, 100 mg cell lysate was incubated at 378C for 4 h with Western blot analysis 50 ml of reaction buer and 5 ml of colometric substrate for For IkBa,30mg of cytoplasmic protein extracts, prepared as either caspase-3 (DEVD-pNA) or caspase-8 (IETD-pNA). reported earlier (Chaturvedi et al., 2000), was resolved on 9% The optical density of the reaction mixture was read on a sodium dodecyl sulfate-polyacrylamide gel. Resolved samples microplate reader at 405 nm wavelength. were electrotransferred to a nitrocellulose membrane, then blocked with 5% nonfat milk protein for 1 h and incubated PARP degradation by activated caspases with anti-IkBa polyclonal antibodies (1 : 3000) for 1 h. The membrane was washed and treated with HRP-conjugated Apoptosis was examined by measuring the proteolytic secondary antibodies, and ®nally detected by chemilumines- cleavage of PARP (Haridas et al., 1998). LNCaP cells were cence (ECL, Amersham Pharmacia Biotech., Arlington left untreated or were cultured in the presence of 10 mM Heights, IL, USA). To examine c-Jun and c-Fos expression, curcumin, 3 nM TNF, or a mixture of 3 mM TNF and 10 mM cells were induced with 0.1 nM TNF for 30 min, prepared the curcumin for 48 h. Whole-cell extracts were prepared, and nuclear extracts and 60 mg of nuclear protein was resolved on 60 mg of protein was resolved in 7.5% sodium dodecyl 9% sodium dodecyl sulfate-polyacrylamide gel, and probed sulfate-polyacrylamide electrophores gel, which was then with appropriate antibodies. To probe for procaspases, Bcl- electrotransferred to a nitrocellulose membrane, blotted with xl, and Bcl-2 proteins, 70 mg of whole cell proteins were anti-PARP antibodies (1 : 2000), and detected using chemilu- resolved on 10% sodium dodecyl sulfate-polyacrylamide gel minescence. Apoptosis was represented by the cleavage of before being probed with the appropriate antibodies. 116-kDa PARP into an 85-kDa degraded product.

JNK assay The JNK assay was performed using a modi®ed method as described earlier (Kumar and Aggarwal, 1999). In brief, 100 mg of cytoplasmic extracts were incubated with anti- Abbreviations JNK1 antibodies, and the immune-complex that formed was AP-1, activator protein-1; CHX, cycloheximide; EMSA, precipitated with protein A/G-sepharose beads. The kinase electrophoretic mobility shift assay; FBS, fetal bovine assay was performed using washed beads as the source of serum; NF-kB, nuclear transcription factor-kB; IkB, enzyme and glutathione S-transferase-Jun (1 ± 79) as the inhibitory subunit of NF-kB; JNK, c-jun N-terminal substrate (2 mg/sample) in the presence of 10 mCi [32P] ATP kinase; MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl- per sample. The kinase reaction was carried out by incubating tetrazolium bromide]; PARP, poly (ADP-ribose) polymer- the above mixture at 308C in the kinase assay buer for ase; PI, propidium iodide; TNF, tumor necrosis factor, 15 min. The reaction was stopped by adding sodium dodecyl SEAP, secretory alkaline phosphatase sulfate sample buer and boiling the samples. Finally, protein was resolved on a 10% reducing gel. The radioactive bands of the dried gel were visualized and quantitated using a PhosphorImager as mentioned previously.

Cytotoxicity assay Acknowledgments The cytotoxic eects of TNF alone, curcumin alone or the This research was conducted with support from The combination on DU145 and LNCaP cells were determined Clayton Foundation for Research.

Oncogene Role of NF-kB and AP-1 in proliferation of prostate cancer cells A Mukhopadhyay et al 7609 References

Abreu-Martin MT, Chari A, Palladino AA, Craft NA and Karin M, Liu Z and Zandi E. (1997). Curr. Opin. Cell. Biol., Sawyers CL. (1999). Mol. Cell. Biol., 19, 5143 ± 5154. 9, 240 ± 246. Adams J, Palombella VJ, Sausville EA, Johnson J, Destree Kawamori T, Lubet R, Steele VE, Kello GJ, Kaskey RB, A, Lazarus DD, Maas J, Pien CS, Prakash S and Elliott P. Rao CV and Reddy BS. (1999). Cancer Res., 59, 597 ± 601. (1999). J. Cancer Res., 59, 2615 ± 2622. Kubota T, Koshizuka K, Williamson EA, Asou H, Said JW, Aggarwal BB. (2000). Biochem. Pharmacol., 60, 1033 ± 1039. Holden S, Miyoshi I and Koeer HP. (1998). Cancer Res., Ahmed MM, Sells SF, Venkatasubbarao K, Fruitwala SM, 58, 3344 ± 3352. Muthukkumar S, Harp C, Mohiuddin M and Rangnekar Kumar A, Dhawan S, Hardegen NJ and Aggarwal BB. VM. (1997). J. Biol. Chem., 272, 33056 ± 33061. (1998). Biochem. Pharmacol., 55, 775 ± 783. Bargou RC, Emmerich F, Krappmann D, Bommert K, Kumar A and Aggarwal BB. (1999). Meth. Enzymol., 300, Mapara MY, Arnold W, Royer HD, Grinstein E, Greiner 339 ± 345. A, Scheidereit C and Dorken B. (1997). J. Clin. Invest., Kuo ML, Huang TS and Lin JK. (1996). Biochim. Biophys. 100, 2961 ± 2969. Acta., 1317, 95 ± 100. Beg AA and Baltimore D. (1996). Science, 274, 782 ± 784. Li CJ, Zhang LJ, Dezube BJ, Crumpacker CS and Pardee Chaturvedi MM, Mukhopadhyay A and Aggarwal BB. AB. (1993). Proc. Natl. Acad. Sci. USA, 90, 1839 ± 1842. (2000). Methods Enzymol., 319, 585 ± 602. Lin J, Adam RM, Santiestevan E and Freeman MR. (1999). Chen HW and Huang HC. (1998). Br.J.Pharmacol.,124, Cancer Res., 59, 2891 ± 2897. 1029 ± 1040. LiuXH,YaoS,KirschenbaumAandLevineAC.(1998). Chen YR and Tan TH. (1998). Oncogene, 17, 173 ± 178. Cancer Res., 58, 4245 ± 4249. Colotta F, Polentarutti N, Sironi M and Mantovani A. McConkey DJ, Greene G and Pettaway CA. (1996). Cancer (1992). J. Biol. Chem., 267, 18278 ± 18283. Res., 56, 5594 ± 5599. Craft N, Chhor C, Tran C, Belldegrun A, DeKernion J, Manna SK, Zhang HJ, Yan T, Oberley LW and Aggarwal Witte ON, Said J, Reiter RE and Sawyers CL. (1999a). BB. (1998). J. Biol. Chem., 273, 13245 ± 13254. Cancer Res., 59, 5030 ± 5036. Mehta K, Pantazis P, McQueen T and Aggarwal BB. (1997). Craft N, Shostak Y, Carey M and Sawyers CL. (1999b). Nat. Anticancer Drugs, 8, 470 ± 481. Med., 5, 280 ± 285. Miyamoto S and Verma IM. (1995). Adv. Cancer Res., 66, Darnay B, Ni J, Moore PA and Aggarwal. (1999). J. Biol. 255 ± 292. Chem., 274, 7724 ± 7731. Miyamoto S, Chiao PJ and Verma IM. (1994). Mol. Cell. DaviesMA,KoulD,DhesiH,BermanR,McDonnellTJ, Biol., 14, 3276 ± 3282. McConkey D, Yung WK and Steck PA. (1999). Cancer Nalca A, Qiu SG, El-Guendy N, Krishnan S and Rangnekar Res., 59, 2551 ± 2556. VM. (1999). J. Biol. Chem., 274, 29976 ± 29983. Degeorges A, Wang F, Frierson Jr HF, Seth A, Chung LW Parker SL TT, Bolden S and Wingo PA. (1997). J. Clin., 46, and Sikes RA. (1999). Cancer Res., 59, 2787 ± 2790. 5±27. Dixit VM, Marks RM, Sarma V and Prochownik EV. (1989). Plummer SM, Holloway KA, Manson MM, Munks RJ, J. Biol. Chem., 264, 16905 ± 16909. Kaptein A, Farrow S and Howells L. (1999). Oncogene, Ewing CM, Ru N, Morton RA, Robinson JC, Wheelock MJ, 18, 6013 ± 6020. Johnson KR, Barrett JC and Isaacs WB. (1995). Cancer Pulverer BJ, Kyriakis JM, Avruch J, Nikolakaki E and Res., 55, 4813 ± 4817. Woodgett JR. (1991). Nature, 353, 670 ± 674. Giri DK and Aggarwal BB. (1998). J. Biol. Chem., 273, RaoAJ,PerlmanH,ChenMW,DayML,StreitmanJSand 14008 ± 14014. Buttyan R. (1995). O. Cancer Res., 55, 4438 ± 4445. Haridas V, Darnay BG, Natarajan K, Heller R and Ripple MO, Henry WF, Schwarze SR, Wilding G and Aggarwal BB. (1998). J Immunol., 160, 3152 ± 3162. Weindruch R. (1999). J Natl. Cancer Inst., 91, 1227 ± 1232. Herrmann JL, Menter DG, Beham A, von Eschenbach A and Singh S and Aggarwal BB. (1995). J. Biol. Chem., 270, McDonnell TJ. (1997a). Exp. Cell. Res., 234, 442 ± 451. 24995 ± 25000. Herrmann JL, Beham AW, Sarkiss M, Chiao PJ, Rands MT, Sintich SM, Steinberg J, Kozlowski JM, Lee C, Pruden S, Bruckheimer EM, Brisbay S and McDonnell TJ. (1997b). Sayeed S and Sensibar JA. (1999). Prostate, 39, 87 ± 93. Exp. Cell. Res., 237, 101 ± 109. Sumitomo M, Tachibana M, Nakashima J, Murai M, Herrmann JL, Briones Jr F, Brisbay S, Logothetis CJ and Miyajima A, Kimura F, Hayakawa M and Nakamura McDonnell TJ. (1998). Oncogene, 17, 2889 ± 2899. H. (1999). J. Urol., 161, 674 ± 679. Hinz M, Krappmann D, Eichten A, Heder A, Scheidereit C Tamatani M, Che YH, Matsuzaki H, Ogawa S, Okado H, and Strauss M. (1999). Mol. Cell. Biol., 19, 2690 ± 2698. Miyake S, Mizuno T and Tohyama M. (1999). J. Biol. Huang TS, Lee SC and Lin JK. (1991). Proc. Natl. Acad. Sci. Chem., 274, 8531 ± 8538. USA, 88, 5292 ± 5296. VanAntwerpDJ,MartinSJ,KafriT,GreenDRandVerma Jiang MC, Yang-Yen HF, Yen JJ and Lin JK. (1996). Nutr. IM. (1996). Science, 274, 787 ± 789. Cancer, 26, 111 ± 120. vanBrusselJP,vanSteenbruggeGJ,RomijnJC,Schroder Jobin C, Bradham CA, Russo MP, Juma B, Narula AS, FH and Mickisch GH. (1999). Eur. J. Cancer, 35, 664 ± Brenner DA and Sartor RB. (1999). J. Immunol., 163, 671. 3474 ± 3483. Wang CY, Mayo MW and Baldwin Jr AS. (1996). Science, Kadkol SS, Brody JR, Pevsner J, Bai J and Pasternack GR. 274, 784 ± 787. (1999). Nat. Med., 5, 275 ± 279.

Oncogene