Retraction: Fatty acid synthase is a novel therapeutic target in multiple myeloma

Okawa, Y., Hideshima, T., Ikeda, H., Raje, N., Vallet, S., Kiziltepe,T., Yasui, H., Enatsu, S., Pozzi, S., Breitkreutz, I., Cirstea, D., Santo, L.,Richardson, P. & Anderson, K.C. (2008) Fatty acid synthase is a novel therapeutic target in multiple myeloma. British Journal of Haematology, 141, 659–671. The above article, published in the online issue, 141·5 on 7 May 2008 in Blackwell Synergy (http://www.blackwell‐synergy.com), has been retracted by agreement between the authors, the Journal Editor‐in‐Chief and Blackwell Publishing Ltd. Finbarr Cotter Editor‐in‐Chief British Journal of Haematology Summary This study investigated the biological significance of the inhibition of fatty acid synthase (FAS) in multiple myeloma (MM) using the small molecule inhibitor Cerulenin. Cerulenin triggered growth inhibition in both MM cell lines and MM patient cells, and overcame the survival and growth advantages conferred by interleukin‐6, insulin‐like growth factor‐1, and bone marrow stromal cells. It induced apoptosis in MM cell lines with only modest activation of caspase ‐8, ‐9, ‐3 and PARP; moreover, the pan‐caspase inhibitor Z‐VAD‐FMK did not inhibit Cerulenin‐induced apoptosis and cell death. In addition, treatment of MM cells with Cerulenin primarily up‐regulated apoptosis‐inducing factor/endonuclease G, mediators of caspase‐independent apoptosis. Importantly, Cerulenin induced endoplasmic reticulum stress response via up‐regulation of the Grp78/IRE1α/JNK pathway. Although the C‐Jun‐NH2‐terminal kinase (JNK) inhibitor SP600215 blocked Cerulenin‐induced cytotoxicity, it did not inhibit apoptosis and caspase cleavage. Furthermore, Cerulenin showed synergistic cytotoxic effects with various agents including Bortezomib, Melphalan and Doxorubicin. Our results therefore indicate that inhibition of FAS by Cerulenin primarily triggered caspase‐independent apoptosis and JNK‐dependent cytotoxicity in MM cells. This report demonstrated that inhibition of FAS has anti‐tumour activity against MM cells, suggesting that it represents a novel therapeutic target in MM.

Fatty acid synthesis occurs in all living organism and provides for essential components of both biological membranes and energy storage. Fatty acid synthase (FAS) is a multifunctional enzyme that catalyzes the terminal steps in the synthesis of the 16-carbon fatty acid palmitate in cells (Wakil, 1989;Asturias et al, 2005). In normal cells, FAS expression levels are relatively low, since fatty acid is generally supplied by dietary fatty acid. In contrast, FAS is expressed at significantly higher levels in a variety of human cancers including breast, colon, ovary, lung, prostate and others (Milgraum et al, 1997;Kuhajda, 2000Kuhajda, , 2006Pizer et al, 2001;Baron et al, 2004;Menendez & Lupu, 2007). Moreover, several reports have shown that FAS expression levels correlate with tumour progression, aggressiveness and metastasis (Kuhajda, 2000(Kuhajda, , 2006Rossi et al, 2003). For example, FAS expression levels are predictive of poor prognosis in breast and prostate cancer (Kuhajda, 2006). As FAS has been strongly linked to tumour cell proliferation (Kuhajda, 2000) and is preferentially expressed in cancer cells, it may represents an attractive target for novel anticancer therapy (Kuhajda, 2006;Menendez & Lupu, 2007). Importantly, Cerulenin, a specific inhibitor of FAS derived from a natural product of Cephalosporium caerulens, inhibits tumour growth and induces apoptosis in a variety of cancer cell lines (Pizer et al, 1996a,b;Heiligtag et al, 2002). It irreversibly inhibits FAS by binding covalently to cysteine at the active site of the b keto acyl synthase moiety, which catalyzes the condensation reaction between the elonging fatty acid chain and each successive acetyl or malonyl residue. Other FAS inhibitors, such as C75 and orlistat, also show antitumour activity in vitro and in vivo Gabrielson et al, 2001;Kridel et al, 2004). The effect of FAS inhibitor on cancer is further verified by RNA interference to downregulate FAS in prostate and breast cancer cells, which induces apoptosis Bandyopadhyay et al, 2006).
This study determined the biological significance of targeting FAS in MM using the small molecule inhibitor Cerulenin. Cerulenin triggers growth inhibition in both MM cell lines and patient MM cells, as well as overcomes the survival and growth advantages conferred by interleukin-6 (IL-6), insulin-like growth factor-1 (IGF-1) and bone marrow stromal cells (BMSC). It induces apoptosis in MM cell lines with only modest activation of caspase -8, -9, -3 and Poly (ADP-ribose) polymerase (PARP); moreover, the pan-caspase inhibitor Z-VAD-FMK does not inhibit Cerulenin-induced apoptosis and cell death. During Cerulenin-induced apoptosis, both apoptosis-inducing factor (AIF) and endonuclease G (Endo G) are released from the mitochondria to cytosol, suggesting that Cerulenin triggers apoptosis, at least in part, via a caspase-independent, AIF/Endo G pathway. Importantly, Cerulenin also induces endoplasmic reticulum (ER) stress response via up-regulation of Grp78/ IRE1a/JNK pathway. Although the C-Jun-NH 2 -terminal kinase (JNK) inhibitor, SP600215, blocks Cerulenin-induced cytotoxicity, it does not inhibit apoptosis and caspase cleavage triggered by Cerulenin. Importantly, Cerulenin triggers synergistic cytotoxicity with various agents including Bortezomib, Melphalan and Doxorubicin, which induce apoptosis via activation of caspases. Our results therefore indicate that inhibition of FAS by Cerulenin triggers primarily caspase-independent apoptosis and JNK-dependent cell death in MM, confirming that FAS represents a novel therapeutic target in MM.
obtained from the Dana-Farber Cancer Institute Institutional Review Board. Informed consent was obtained from all patients in accordance with the Declaration of Helsinki protocol.

Growth inhibition assay
The growth inhibitory effect of Cerulenin in MM cell lines, primary MM cells, PBMNC and normal plasma cells was assessed by measuring 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrasodium bromide (MTT; Sigma) dye absorbance. Cells were plated in 96-well microtitre plates at a density of 2-3 · 10 4 cells per well, and each plate was incubated for 24 and/or 48 h, with MTT added to each well for the last 4 h. The absorbance of each well was measured at 570/630 nm using a spectrophotometer (Molecular Devices, Sunnyvale, CA, USA).
To measure proliferation of MM cells and BMSC, the rate of DNA synthesis was measured, as described previously (Hideshima et al, 2000). MM cells were incubated in 96-well culture plates (Costar, Cambridge, MA, USA) in the presence of Cerulenin and/ or recombinant cytokines IL-6 or IGF-1 for 48 h at 37°C. Cells were pulsed with 18AE5 kBq/well of [ 3 H]-thymidine ([ 3 H]-TdR; Perkin Elmer, Boston, MA, USA) during the last 8 h of culture, harvested onto glass filters with an automatic cell harvester (Cambridge Technology, Cambridge, MA, USA), and counted using the LKB Betaplate scintillation counter (Wallac, Gaithersburg, MD, USA). MTT and 3 H uptake analysis were performed in triplicate or quadruplicate cultures.

Flow cytometric analysis
For detection of apoptotic cells, mitochondrial protein staining was performed with pycoerythrin-labeled anti-Apo2AE7 antibody (Immunotec, Marseille, France). Cells were treated with various concentrations of Cerulenin for 24 h, washed with PBS, and then incubated with Apo2AE7 for 30 min. Apo2AE7 staining was analysed using the RXP cytomics software on an Epics flow cytometer (Coulter Immunology, Hialeah, FL, USA).

Immunocytochemistry
Cytospins of cells were prepared on glass slides and fixed with 50% methanol/50% acetone at 4°C. The slides were blocked in PBS with 5% FBS at room temperature (RT), and then incubated with primary antibody overnight at 4°C. Fluorescein isothiocyanate (FITC)-labeled secondary antibody reaction for 30 min was then performed. Coverslips were mounted on the glass slides with VectaShield antifade/4¢,6-diamidino-2-phenylindole (DAPI), which were analysed by Nikon E800 fluorescence microscopy (Nikon Inc., Melville, NY, USA).

Statistical analysis
Statistical significance of differences observed in drug-treated compared with control cultures was determined using the Mann-Whitney test. The minimal level of significance was P < 0AE05. All statistical analyses were determined using GraphPad Prism software (GraphPad Software, Inc. San Diego, CA, USA).

Isobologram analysis
The interaction between Cerulenin and Bortezomib, Melphalan, and Doxorubicin was analysed using CalcuSyn software program (Biosoft, Ferguson, MO, USA) to determine whether the combination was additive or synergistic, as described previously (Chou & Talalay, 1984;Raje et al, 2004). This program is based upon the Chou-Talalay method, which calculates a combination index (CI). Analysis is performed based on the following equation: CI = (D)1/ (Dx)1 + (D)2/(Dx)2 + (D)1(D)2/(Dx)1(Dx)2, where (D)1 and (D)2 are the doses of drug 1 and drug 2 that have x effect when used in combination; whereas (Dx)1 and (Dx)2 are the doses of drug 1 and drug 2 that have the same x effect when used alone. A CI c. 1AE1 represents the conservation isobologram and indicates additive effects, whereas CI c. <0AE9 indicates synergism.

FAS expression in various cells
We first examined baseline expression of FAS in various cells. FAS protein was expressed in all MM cell lines (Fig 1A and B; lane 5), as well as in primary tumour cells from MM patients ( Fig 1B; lane 4, Fig 1C). Importantly, FAS expression was higher in both MM cell lines and primary tumour cells than in normal plasma cells, as assessed by Western blotting (Fig 1B) and confirmed by immunocytochemical analysis (Fig 1D).
including PBMNC and normal plasma cells, using the MTT assay. Cerulenin significantly inhibited the growth of drug-sensitive MM.1S, U266, RPMI8226, OPM1 and OPM2 MM cell lines, with a 50% inhibitory concentration (IC 50 ) at 24 h of 24AE16, 22AE7, 24AE03, 37AE03 and 21AE53 lmol/l, respectively, and IC 50 at 48 h of 12AE59, 11AE12, 17AE08, 11AE45 and 9AE71 lmol/l respectively (Fig 2B and C). Cerulenin also inhibited growth of Dex-resistant MM.1R, Melresistant RPMI-LR5, Dox-resistant RPMI-Dox40 MM cell lines, with IC 50 at 24 h of 22AE59, 86AE21 and 33AE29 lmol/l, and IC 50 at 48 h of 10AE52, 22AE73 and 16AE52 lmol/l respectively (Fig 2B and C). However, Cerulenin did not induce cytotoxicity in PBMNC and normal plasma cells from three healthy volunteers (Fig 2E and F). Importantly, Cerulenin induced dose-dependent cytotoxicity against CD138 positive MM cells (IC 50 at 24 h of 27AE37 lmol/l) isolated from three patients whose disease was refractory to Dexamethasone, Melphalan, Thalidomide, or Bortezomib therapy ( Fig 2D). These results indicate that FAS inhibition by Cerulenin selectively and potently induces cytotoxicity in MM cell lines as well as primary MM cells, even those resistant to conventional and novel therapy.

Cerulenin overcomes the growth stimulatory effects of cytokines, and inhibits MM cell growth in culture with BMSCs
It has been demonstrated that IL-6 and IGF-1 induce both growth and inhibition of apoptosis in MM cells (Jourdan et al, 2000;Mitsiades et al, 2002Mitsiades et al, , 2004, therefore, we examined whether FAS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17  inhibition can overcome the effects of exogenous IL-6 and IGF-1. Both IL-6 (10 ng/ml) and IGF-1 (50 ng/ml) triggered increased growth in MM.1S and U266 cells, which was inhibited by Cerulenin (P < 0AE05) (Fig 3A and B). As we previously have shown that the BM microenvironment confered growth and drug resistance in MM cells , we next examined the effect of FAS inhibition on MM cell growth in the BM milieu. MM.1S and U266 cells were cultured for 48 h with or without BMSC, in the presence or absence of Cerulenin. MM cell adherence to BMSC enhanced [ 3 H]-thymidine uptake in MM cells, which was inhibited by Cerulenin (P < 0AE05) (Fig 3C). These results indicate that Cerulenin also blocks the growth stimulatory effect of the BM microenvironment on MM cells.

Cerulenin induces apoptosis via activation of caspase-independent pathway
To analyse the mechanism of cytotoxicity triggered by Cerulenin in MM cells, we next performed Apo2AE7 analysis and immunoblotting in MM.1S and U266 cells. Cerulenin induced apoptosis in a dose-dependent manner, evidenced by Apo2AE7 analysis (Fig 4A, Fig S1C). Cerulenin also triggers caspase -8, -9, -3, and PARP cleavage in MM.1S and U266 (Fig 4B, Fig S1A). Although the pan-caspase inhibitor Z-VAD-FMK blocked Cerulenininduced caspase and PARP cleavage in both MM.1S and U266 cells (Fig 4B, Fig S1A), it did not inhibit Cerulenin-triggered apoptosis (Fig 4D, Fig S1C) and cell death (Fig 4C, Fig S1B). During Cerulenin-induced apoptosis in MM cells, both AIF and Endo G were significantly released from the mitochondria to cytosol in a time-dependent manner ( Fig 4E). These results suggest that Cerulenin triggers apoptosis, at least in part, via a caspase-independent, AIF/Endo G pathway.
Cerulenin induces ER stress response via up-regulation of GRP78, IRE1a, JNK pathway As some reports have demonstrated that FAS inhibitor induced ER stress response in tumour cells (Little et al, 2007), we next examined whether FAS inhibition by Cerulenin triggers activation of GRP78, IRE1a, JNK pathway as an ER stress response in MM cells. The ER stress inducer Tunicamycin significantly upregulated Grp78, IRE1a, ASK1, JNK and XBP-1 in MM.1S cells (Fig 5A). In addition, Cerulenin treatment also induced upregulation of Grp78, IRE1a and p-JNK, as well as its downstream molecules p-ATF-2 and p-c-Jun (Fig 5A). Importantly, JNK inhibitor SP600215 significantly blocks Cerulenininduced cytotoxicity in MM.1S and U266 cells, (P < 0AE05) (Fig 5B, Fig S2A). We also confirmed that SP600215 markedly inhibited JNK signaling pathway in MM.1S cells by Western blotting ( Fig 5E); however, it did not inhibit apoptosis and caspase cleavage triggered by Cerulenin in both MM.1S and U266 cells (Fig 5C and D,Fig S2B and C). In addition, combination pretreatment with JNK inhibitor and ZVAD-FMK did not enhance the inhibitory effect of JNK inhibitor alone, indicating that JNK activation is the major pathway mediating Cerulenin-induced cytotoxicity in MM cells (Fig S3). These results strongly indicate that targeting FAS by Cerulenin induces ER stress response via up-regulation of Grp78/IRE1a/ JNK pathway, associated with JNK-dependent, non-apoptotic cell death in MM cells.
Cerulenin enhanced growth inhibition triggered by Bortezomib, Melphalan and Doxorubicin, as analysed by MTT assay (Fig 6, Fig  S4). Next, we analysed the interaction between Cerulenin and these agents by isobologram analysis using CalcuSyn software program (Raje et al, 2004). Based upon the Chou-Talalay method to calculate the CI, we showed synergistic cytotoxic effects between Cerulenin and Bortezomib, Melphalan, and Doxorubicin, with a maximum CI of 0AE567, 0AE188 and 0AE529, respectively, in MM.1S cells; 0AE63, 0AE404 and 0AE079, respectively, in primary MM cells; and 0AE879, 0AE591, and 0AE301, respectively, in U266 cells. These data demonstrate that these combination therapies enhance Cerulenin-induced cytotoxicity synergistically in MM cells, due to the fact that these agents induce caspase-dependent apoptosis whereas Cerulenin triggers primarily caspase-independent apoptosis.

Discussion
This report demonstrated that FAS inhibition by Cerulenin induced cytotoxicity in drug-sensitive and drug-resistant MM cell 1997; Kuhajda, 2000Kuhajda, , 2006Pizer et al, 2001;Baron et al, 2004;Menendez & Lupu, 2007) and further suggesting that it represents a novel therapeutic target in MM. The BM microenvironment confers drug resistance in MM cells via at least two different mechanisms: (i) cell adhesion mediated drug resistance (CAM-DR) through adhesion of MM cells to fibronectin and BMSC; and (ii) activation of phosphatidylinositol 3-kinase (PI3-K)/Akt and/or Janus kinase 2 (JAK2)/signal transducers and activators of transcription 3 (STAT3) signaling due to high levels of cytokines in the BM milieu (such as IL-6 and IGF-1) (Damiano et al, 1999;Hazlehurst et al, 2000;Hideshima et al, 2001Hideshima et al, , 2004. Importantly, treatments targeting the BM microenvironment as well as the MM cell can overcome drug resistance in both preclinical and clinical studies Hideshima et al, 2007). Our data indicate that Cerulenin overcomes the survival and growth advantages conferred by IL-6, IGF-1 and BMSC, strongly suggesting its utility to overcome conventional drug resistance in MM.
Although caspases are the key executors of apoptosis, recent studies suggest that caspase activation is not the sole pathway for inducing apoptosis or necrosis (Jaattela & Tschopp, 2003;Abraham & Shaham, 2004;Lockshin & Zakeri, 2004). For example, caspase-independent cell death in MM cells is triggered by AS 2 O 3 and inosine monophosphate dehydrogenase inhibitor VX-944 treatment (McCafferty-Grad et al, 2003;Ishitsuka et al, 2005). The AIF/Endo G pathway is a major caspase-independent apoptotic cascade (Daugas et al, 2000;Ahn et al, 2004;Cande et al, 2004;Cregan et al, 2004). In this pathway, enhanced Bcl-2 family expression induces mitochondrial membrane permeability, thereby releasing AIF and Endo G from the mitochondria to the cytosol and nucleus, with subsequent chromatin condensation and cell death. The present study showed that MM cell apoptosis triggered by Cerulenin was primarily mediated via a caspaseindependent pathway. Although cleavage of caspase -8,-9,-3 and PARP was induced by Cerulenin treatment, Z-VAD-FMK did not block completely Cerulenin-induced apoptosis and cell death. Moreover, both AIF and Endo G were released from the mitochondria to cytosol during Cerulenin-induced apoptosis. These results suggest that Cerulenin triggers apoptosis, at least in part, via a caspase-independent AIF/Endo G pathway, with only modest activation of caspase -8,-9,-3 and PARP. As most conventional and novel drugs induce apoptosis in MM cells via activation of caspases, we next examined the effect of combining Cerulenin and other drugs. Cerulenin combined with Bortezomib, Melphalan or Doxorubicin triggered synergistic MM cell cytotoxicity, indicating that it may be useful to sensitize or overcome drug resistance. The ER stress response is a choreographed series of cellular events activated by specific insults that result in altered ER function (Kaufman, 2004;Xu et al, 2005). The combined effect of this response is the activation of genes associated with an adaptation response. Upon prolonged stress, the adaptation mechanism of the ER stress response is saturated, resulting in cell death. Several studies have demonstrated important associations between lipid synthesis pathways and the ER stress response (Cox et al, 1997;van der Sanden et al, 2004;Sriburi et al, 2004;Tessitore et al, 2004). Inhibition of phospholipid synthesis, especially that of phosphatidylcholine, induces ER stress-related pathways (van der Sanden et al, 2004). Fatty acid synthesis in general, and FAS activity in particular, occurs in the ER . Importantly, because of the direct connection between FAS activity and phospholipid synthesis, previous reports demonstrated that pharmacologic blockade of FAS activity induced ER stress in tumour cells via activation of PERK/eIF2a and IRE1a/XBP-1 pathways Little et al, 2007). In addition, Nakamura et al (2006) reported that activation of the ER stress pathway was associated with survival of MM cells, as they contained a well-developed ER due to M-protein production. Our data and other reports indicated that increased FAS expression in tumour cells is important for ER function; and conversely, suggest a role for ER stress in the antitumour effects of FAS inhibitors. The present study showed that Cerulenin treatment induced upregulation of Grp78, IRE1a and p-JNK, as well as its downstream molecules p-ATF-2 and pc-Jun. IRE1a can activate the ASK1/JNK mitogen-activated protein kinase pathway (Urano et al, 2000;Xu et al, 2005;Li et al, 2006), and activation of JNK is a common response to many forms of stress known to influence the cell-death machinery (Davis, 2000). Although studies show that multiple pathways can contribute to ER-stress-induced apoptosis (Li et al, 2006), the mechanisms by which unresolved ER stress causes cell death are not fully characterized (Szegezdi et al, 2006;Zhao & Ackerman, 2006). Bandyopadhyay et al (2006) proposed a model for apoptotic signaling in which inhibition of FAS leads to accumulation of malonyl-CoA, which in turn inhibits carnitine palmitoyltransferase-1 (CPT-1) and results in up-regulation of ceramide, followed by induction of the proapoptotic genes BNIP3, tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TNFSF10), and death-associated protein kinase 2 (DAPK2) . BNIP3 not only induces apoptotic cell death, but also is implicated in necrosis and autophagy (Vande Velde et al, 2000;Daido et al, 2004). TNFSF10 has been shown to induce apoptosis in a wide range of transformed cells (Kelley & Ashkenazi, 2004). In addition, a human single-chain fragment variable (HW1) antibody against TRAIL receptor 2 induced autophagic cell death predominantly via the JNK pathway in a caspase-independent manner (Park et al, 2007). DAPK2 is associated with ceramide-mediated apoptosis (Pelled et al, 2002;Yamamoto et al, 2002), but also induces another cell death pathway mediating autophagic or type II cell death (Inbal et al, 2002;Bialik & Kimchi, 2004) without caspase activation (Inbal et al, 2002). Our data and others suggest that inhibition of FAS by Cerulenin activates multiple death signal pathways in MM cells, including caspase-independent apoptosis and JNK dependent non-apoptotic cell death. JNK and its downstream molecules ATF-2 and c-Jun induce apoptosis in various tumours (Bossy-Wetzel et al, 1997;Podar et al, 2007). Moreover, caspase-independent JNK activation promotes autophagic cell death induced by TNF-a (Jia et al, 2006) and other chemicals (Borsello et al, 2003;Yu et al, 2004;Park et al, 2007). JNK can activate the mitochondrial pathway including AIF, Endo G, cytochrome C and Smac (Weston & Davis, 2007). However, the exact role of c-Jun in mediating cell differentiation, growth, survival and apoptosis is not fully understood (Jochum et al, 2001;Vogt, 2001).
In summary, our data showed that inhibition of FAS by Cerulenin induced primarily caspase-independent apoptosis via upregulation of AIF/Endo G pathway, and JNK-dependent cell death in MM cells. This report demonstrated that inhibition of FAS had anti-tumour activity against MM cells, identifying FAS as a novel therapeutic target. Taken together, our data provide the preclinical rationale for clinical trials targeting FAS to improve patient outcome in MM.