The Role of ROS/Superoxides, NFAT and MEK/ERK Signaling

Home , Proline oxidase

Oncogene (2006) 25, 5640–5647 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE Proline oxidase activates both intrinsic and extrinsic pathways for apoptosis: the role of ROS/superoxides, NFAT and MEK/ERK signaling

Y Liu1, GL Borchert2, A Surazynski1, C-A Hu3 and JM Phang1

1Metabolism and Cancer Susceptibility Section, Laboratory of Comparative Carcinogenesis, Center for Cancer Research, National Cancer Institute, Frederick, MD, USA; 2Basic Research Program, SAIC-Frederick, Inc., Frederick, MD, USA and 3Department of Biochemistry and Molecular Biology, University of New Mexico Health Science Center, Albuquerque, NM, USA

Proline oxidase (POX), often considered a ‘housekeeping Introduction enzyme’ might play an important role in apoptosis. We have shown that POX generated proline-dependent Proline oxidase (POX), localized at the mitochondrial reactive oxygen species (ROS), specifically superoxide inner membrane, catalyzes the conversion of proline to radicals, and induced apoptosis through the mitochondrial pyrroline-5-carboxylate (P5C). Together with P5C (intrinsic) pathway. In our current report, we used DLD-1 reductase, POX mediates the proline cycle to shuttle colorectal cancer cells stably transfected with the POX redox equivalents between mitochondria and cytosol gene under the control of a tetracycline-inducible promoter (Hagedorn and Phang, 1983; Phang, 1985; Phang et al., and found POX-stimulated expression of tumor necrosis 2001). Also important, P5C is the intermediate provid- factor-related apoptosis inducing ligand (TRAIL), DR5 inga direct carbon bridgeconnectingthe tricarboxylic and cleavage of caspase-8. Importantly, apoptosis acid and urea cycles (Phang, 1985; Phang et al., 2001). measured by flow cytometry was partially inhibited by As a residue in proteins, proline contributes special Z-IETD-FMK, a specific inhibitor of caspase-8. These features. In addition to the well-recognized physical findings suggest that the extrinsic (death receptor) path- properties in structural proteins, for example, collagens, way also is activated by POX. Furthermore, the mechan- proline-containingregionsprovide tertiary structure ism of this effect on the extrinsic pathway, specifically, the necessary for protein–protein interactions and are induction of TRAIL by POX, may be mediated by NFAT critical for the function of WW and SH3 domains for transcription factors. Additionally, POX expression also specific recognition and phosphorylation (Macias et al., dramatically decreased phosphorylation of MEK and 2002). Ser/Thr-Pro motifs are the exclusive phosphor- ERK, and the decrease was partially reversed by ylation sites for a large number of Pro-directed protein expression of manganese superoxide dismutase (MnSOD). kinases that play essential roles in signal transduction Overexpression of constitutively active form of MEK, and cell cycle progression (Lu, 2003). Additionally, acMEK, partially blocked POX-induced apoptosis. These hydroxylation at certain conserved proline residues of findings suggest the involvement of MEK/ERK signaling HIF1-a functions as an oxygen sensor (Semenza, 2003). and further confirm the role of ROS/superoxides in POX- Accumulation of proline may cause oxidative stress induced apoptosis. Combined with previously published (Delwing et al., 2003). In pre-clinical cancer research, data, we conclude that POX may induce apoptosis through proline analogs modulate tumor cell growth, perhaps by both intrinsic and extrinsic pathways and is involved in targeting extracellular matrix and/or critical proline nuclear factor of activated T cells (NFAT) signaling and residues of tumor-associated proteins (Hamilton et al., regulation of the MEK/ERK pathway. It is suggested that, 2004). Proline and prolidase, like matrix metalloprotei- as a nutrition factor, POX may modulate apoptosis signals nases, may induce hypoxic and angiogenic effects in induced by p53 or other anti-cancer agents and enhance human cancer cells (our unpublished data). apoptosis in stress situations. Recently, research has focused on proline metabolism Oncogene (2006) 25, 5640–5647. doi:10.1038/sj.onc.1209564; and proline-metabolizingenzymes duringresponses to published online 17 April 2006 stress, especially in programmed cell death. The role of POX in apoptosis has been well documented. First, Keywords: proline oxidase; apoptosis; superoxide radi- usingSerial Analysis of Gene Expression, Polyak et al. cals; NFAT; ERK; metabolism (1997) showed that p53-dependent apoptosis is preceded by the induction of 14 of 7202 genes, including POX. We showed that POX catalyzed the generation of reactive Correspondence: Dr Y Liu or Dr JM Phang, Metabolism and Cancer oxygen species (ROS), particularly superoxide radicals Susceptibility Section, Laboratory of Comparative Carcinogenesis, (Donald et al., 2001; Liu et al., 2005) and overexpression Center for Cancer Research, National Cancer Institute, Frederick, of POX causes apoptotic cell death in a variety of cancer MD 21702, USA. cell types (Donald et al., 2001; Maxwell and Rivera, E-mail: [email protected] or [email protected] Received 2 December 2005; revised 1 March 2006; accepted 1 March 2003; Liu et al., 2005). Although findings from several 2006; published online 17 April 2006 labs suggest that in p53-induced apoptosis, the direct Mechanistic study of POX-induced apoptosis YLiuet al 5641 signaling cascade may be p53 induction of PUMA/ NOXA which interact with BCL-2 or BCL-XL thereby freeingBAX to alter mitochondrial membranes (Naka- no and Vousden, 2001; Yu et al., 2001, 2003), but alternative or modulatingpathways dependingon the metabolic environment may be critical for altering expression of the malignant phenotype. We propose that POX regulates cell survival in situations of nutritional and energetic stress. Therefore, POX may be a good target for matrix/nutrition-dependent mod- ulation of carcinogenesis. Several pathways by which POX induced apoptosis have been studied (Maxwell and Rivera, 2003; Liu et al., 2005; Rivera and Maxwell, 2005). We and others have shown that in POX-induced apoptosis, cytochrome c is released into cytosol from mitochondria and caspase-9 and 3 are activated, suggesting the involvement of the mitochondrial pathway. To better understand POX- induced apoptosis, we expanded our research into several areas and our new data clearly demonstrated that the death receptor pathway also contributed to POX-induced apoptosis. An important mechanism involvingthe induction of tumor necrosis factor-related apoptosis inducingligand (TRAIL) appears to be mediated by nuclear factor of activated T cells (NFAT) stimulation of transcription. Furthermore, the inhibition of MEK/ERK signaling was involved and importantly, our data support the interpretation that the generation of superoxide radicals plays a critical role in many of these processes.

Results

Proline oxidase expression activated the extrinsic apoptotic pathway It has been reported that POX-induced apoptosis is mediated by the mitochondrial pathway (Maxwell and Rivera, 2003; Liu et al., 2005). We considered that the death receptor pathway might also be involved. To test this, we examined the main components of this pathway. After DOX was removed from culture medium to induce POX, caspase-8, the central component of the death receptor pathway was progressively cleaved and thereby activated (Figure 1a). We also examined the upstream components of this pathway. Although the expression of DR4, Fas, FasL, tBid and FADD (not shown) were not altered, the expression of TRAIL and death receptor 5 (DR5) were induced by POX (Figure 1a). To determine Figure 1 POX activated the extrinsic apoptotic pathway. (a) DLD-1 Tet-Off POX cells were cultured after removal of DOX for whether these changes in protein levels were corrobo- 1, 3, and 5 days. Western blots for different proteins involved in the rated by induction at the transcriptional level, we extrinsic pathway were performed. (b) DLD-1 Tet-Off POX cells performed RNA protection assays and found that were cultured without DOX for 1 or 3 days. The RPA was done DR5 and TRAIL mRNA increased (Figure 1b). These usingAmbion’s RPA III TM and BD PharMingen’s hAPO-3d Multi- w data clearly demonstrated that, in addition to the Probe Set accordingto the manufacturers’ instruction. The ‘ ’ indicates that the expression of these genes has been changed by mitochondrial pathway, components of the death POX. receptor pathway also are activated by POX.

Involvement of NFAT in TRAIL activation by POX 2004), the mechanisms mediatinginduction of TRAIL As TRAIL is widely considered a promisingtargetin by POX are of special interest. In this regard, the NFAT cancer prevention and treatment (Kelley and Ashkenazi, family of transcription factors is an attractive target for

Oncogene Mechanistic study of POX-induced apoptosis YLiuet al 5642 exploration. Studied primarily in immune cells, stimula- apoptosis by flow cytometry (Figure 3a). These data tion by cytokines increases cellular calcium levels and suggests that the death receptor pathway contributes to activates calcineurin. Calcineurin, a phosphatase, de- the mediation of POX-induced apoptosis in this system. phosphorylates NFAT which undergoes cytoplasmic to We previously showed that the production of ROS, nuclear translocation where it regulates the expression especially superoxides, is critical for POX-induced of target genes as a transcription factor (Crabtree and Olson, 2002). Several lines of evidence prompted us to examine the role of NFAT in the induction of TRAIL by POX. First, Rivera and Maxwell (2005) showed that POX activates NFAT. Secondly, other workers reported that NFAT is found not only in immune cells but also in colon cancer cells (Duque et al., 2005). Thirdly, examiningthe TRAIL promoter, we identified two NFAT bindingsites (Figure 2a and Wang et al., 2000). Therefore, we undertook studies of the signaling mechanisms linkingPOX, NFAT and TRAIL. In our studies usingDLD-1 cells, we showed that the expression of POX increased the expression of NFATc1 and stimulated its relocalization to the nucleus (Figure 2b). NFATc2, on the other hand, was unde- tectable (data not shown), presumably due to its low abundance in DLD-1 cells. Furthermore, the stimula- tory effect of POX on NFAT expression was corroborated by studies usingNFAT promoter-luciferase constructs. POX expression clearly increased NFAT promoter activity (Figure 2c). Having demonstrated that POX increased NFAT expression, we next examined the regulatory effects of NFAT on TRAIL. Surprisingly, the TRAIL promoter activity was induced more than 300-fold by NFATc1 (Figure 2d). To our knowledge, this is the first report that TRAIL promoter activity is activated by NFAT. Finally, the upstream activation by POX was corroborated by the findingthat TRAIL promoter activity was induced by POX (Figure 2e). These data strongly support the hypothesis that POX stimulates TRAIL activity, and this effect, at least in part, is mediated by NFAT.

Manganese superoxide dismutase abrogation of POX- induced apoptosis We first showed that the POX-activation of death receptor components contributed to POX-induced apoptosis. We treated the DLD-1 Tet-Off POX cells with Z-IETD-FMK, a specific inhibitor of caspase-8 and found a significant inhibition of POX-induced

Figure 2 NFAT-mediated induction of TRAIL by POX. (a) Schematic of the 50-upstream region of the human TRAIL gene. (b) The DLD-1 Tet-Off POX cells were cultured for 1, 3, and 5 days without DOX. Cells were harvested and nuclear proteins were prepared. Western blot for NFATc1 was performed usingequal amount proteins. Histone H1 was used as loadingcontrol. ( c) DLD-1 Tet-Off POX cells were transfected with NFAT-Luc construct and then were cultured with or without DOX. After 2 days, cells were harvested and the luciferase assay was performed. (d) DLD-1 cells were transfected with TRAIL-luc alongwith NFATc1 or pcDNA as control. After 36 h, cells were harvested and luciferase assay was performed. (e) DLD-1 Tet-Off POX cells were transfected with TRAIL-Luc construct and then were cultured with or without DOX. After 2 days, cells were harvested and the luciferase assay was performed. * indicates statistically signiﬁcant (Po0.05).

Oncogene Mechanistic study of POX-induced apoptosis YLiuet al 5643 apoptosis (Donald et al., 2001; Liu et al., 2005). This hyperexpress manganese superoxide dismutase earlier work focused on the mitochondria-dependent (MnSOD), an antioxidant enzyme, which decomposes intrinsic pathway. Actually, ROS/superoxides may superoxide radicals to hydrogen peroxide and has been mediate apoptosis through both mitochondrial and reported to inhibit apoptosis induced by a variety of death receptor pathways. As our new data support the agents (Pani et al., 2004). Importantly, we found that involvement of both pathways in POX-induced apop- MnSOD blocked the cleavage/activation not only of tosis, whether ROS/superoxides participate in both caspase-9 and 3, but also caspase-8, the identifying processes became an important question. Our strategy, caspase of the death receptor pathway (Figure 3b). To similar to our previous study (Liu et al., 2005), is to conﬁrm the effects of MnSOD on POX-induced cell behavior, we repeated our previous experiments based on sub-G0/G1 (Liu et al., 2005) but expressed the result based on cell number (Figure 3c) because ROS may stimulate cell proliferation (Loo, 2003) as well as cause apoptosis. These data not only conﬁrm the involvement of superoxide radicals in POX-induced apoptosis, but also show that the generation of ROS/superoxide radicals is involved in activatingboth pathways.

Involvement of MEK/ERK signaling in POX-induced apoptosis As a central, important pathway in human cells, mitogen-activated protein kinase (MAPK) signaling is closely linked to apoptotic regulation. However, MAPK signaling is usually oppositely directed from apoptotic signaling. It is noteworthy that like apoptotic signaling, MAPK signaling, itself is regulated by redox status. As the expression of POX initiates perturbations of the redox status, we examined the MAPK pathway and found that POX expression reduced the phosphorylation (inactivation) of all three subtypes of MAPK, for example, ERK, JNK, and p38, as well as the decreased phosphorylation (inactivation) of MEK, the upstream kinase of ERK (Figure 4a). In contrast, the phosphorylation (activation) of AKT was not changed (Figure 4a). Interestingly, the introduction of MnSOD partially reversed the decrease in phosphorylation of ERK and MEK (Figure 4a). To further explore the role of MEK/ERK, we generated DLD-1 cells stably expressinga constitutively active form of MEK, caMEK. The expression and activity were demonstrated by Western blots for phospho-MEK and ERK (Figure 4b). In these cells we examined the apoptotic effects produced by infection with Ad-POX. We found that constitutive expression of MEK markedly reduced POX-induced apoptosis (Figure 4c). These data strongly indicate that the inhibition of MEK/ERK signaling is

Figure 3 MnSOD targeted both intrinsic and extrinsic pathways for inhibitingPOX-induced apoptosis. ( a) The cells were incubated with Z-IETD-FMK or vehicle for 2 h and then switched to the medium without DOX. The cells were harvested on day 3 and subjected to flow cytometry analysis. The percentages of cells in sub-G0/G1 phase were shown. The ‘*’ indicates statistically significant (Po0.05). (b) DLD-1 Tet-Off POX cells were infected with 3000 MOI Ad-MnSOD or Ad-Empty as control. After 1 day cells were switched to the medium without DOX. At day 3, cells were harvested and Western blots were performed. (c) DLD-1 Tet- Off POX cells were seeded in 24-well plates, then were infected with 3000 MOI Ad-MnSOD or Ad-Empty as control. After 1 day, cells were switched to the medium without DOX and cell growth was determined as described in Material and methods. Differences are significant (*Po0.05, **Po0.01).

Oncogene Mechanistic study of POX-induced apoptosis YLiuet al 5644

Figure 4 MEK/ERK pathway was involved in POX-induced apoptosis. (a) DLD-1 Tet-Off POX cells were infected with 1000 or 3000 MOI Ad-MnSOD or Ad-Empty as control. The next day cells were switched to the medium without DOX. At days 1 and 3, cells were harvested and Western blots were performed. (b) The DLD-1 cells were transfected with caMEK. It’s expression and phosphorylation/ activation of downstream ERK were demonstrated by Western blots. (c) The DLD-1 cells were transfected with caMEK or vector. Then the stably transfected cells, DLD-1-caMEK, and DLD-1-Vector were infected with adenoviruses Ad-POX or Ad-GFP. The cells were harvested on day 5 and subjected to ﬂow cytometry analysis. The percentages of cells in sub-G0/G1 phase are shown. The ‘*’ indicates statistically signiﬁcant (Po0.05).

involved in POX-induced apoptosis. Our studies also Wu et al., 2000), which is consistent with POX, itself, suggest the involvement of JNK and p38, but additional beinga PIG (p53-induced gene)(Polyak et al., 1997). In work is needed to elucidate their roles in POX-induced fact, it is possible that the induction of DR5 by p53, at apoptosis. least in some situations, may be indirectly mediated by POX. It is widely recognized that TRAIL-mediated cell killingtargetsonly transformed cells without affecting normal cells (Kelley and Ashkenazi, 2004). This Discussion characteristic makes TRAIL a promisingtargetfor cancer prevention and treatment and prompted us to Apoptosis, also known as programmed cell death, is an investigate the mechanism for its induction by POX. orderly, active metabolic process removingdamagedor Here we showed that one inducer was NFAT. developmentally obsolete cells, and this process plays a The NFAT family of transcription factors was fundamental role in a variety of physiological processes. discovered as a transcriptional regulator of a large Its deregulation contributes to many diseases, including number of inducible genes encoding cytokines and cell tumor development and the acquisition of drugresis- surface receptors essential for producingan immune tance. Two main apoptotic pathways have been response. However, recent evidence established a identiﬁed – the mitochondrial (intrinsic) pathway and regulatory role for NFAT (including apoptosis) for cell the death receptor (extrinsic) pathway. Previous work types other than immune cells. The expression and has demonstrated that POX induces apoptosis through function of NFAT in colon cancer and other types of the mitochondrial pathway, in which ROS, especially cancers have also been demonstrated (Duque et al., superoxide radicals play a critical role (Maxwell and 2005; Rivera and Maxwell, 2005). Rivera’s recent Rivera, 2003; Liu et al., 2005). Given that POX is publication (Rivera and Maxwell, 2005) clearly demon- localized to mitochondrial inner membranes and the strated the activation of Ca2 þ /calcineurin/NFAT signal- mitochondria are the main site of ROS/superoxides ingby p53 and POX in cancer cells. There are two production, it is not surprisingthat the mitochondrial NFAT bindingsites in the TRAIL promoter ( À914/ pathway plays an important role. However, when we À902 and À865/À8530) (Wang et al., 2000) which may extended our studies, we found that the death receptor mediate the induction of TRAIL. Although NFAT pathway also was involved. Thus, POX-induced apop- bindingsites also are present in the promoter regionsfor tosis may be mediated by both mitochondrial and death FasL and TNFa, two other important ligands in the receptor pathways. death receptor pathway (McCaffrey et al., 1994; Latinis Our data showed that with POX-induced activation et al., 1997), we found little change either in mRNA or of the death receptor pathway, two important compo- in protein levels for these ligands with POX expression nents of the pathway were induced, DR5 and TRAIL (Figure 1 or data not shown). Thus FasL and TNF-a (Figure 1). DR5 is a p53-induced gene (Wu et al., 1997; may not be mediators in POX-activated death receptor

Oncogene Mechanistic study of POX-induced apoptosis YLiuet al 5645 pathway and POX-induced apoptosis. Nevertheless, to may directly cause mitochondrial membrane permeabi- our knowledge, this is the first demonstration that lization and the loss of cytochrome c (Cai and Jones, NFAT activates TRAIL, and TRAIL joins other pro- 1998; Dussmann et al., 2003), and this may play a apoptotic molecules in the TNF family, that is, TNF-a, critical role for induction of apoptosis in colon cancer FasL, which induce apoptosis by the mediation of cells (Wenzel et al., 2003; Wenzel et al., 2004). The NFAT in certain situations. localization of POX at mitochondrial inner membranes There is increasingevidence for cross talk between and its donation of electrons to electron transport with extrinsic and intrinsic pathways in the process of an interveningFAD may facilitate the production of apoptosis. In the death receptor pathway, once cas- superoxide at a critical site. However, MnSOD also pase-8 is activated it can either directly cleave caspase-3 reversed the decreased phosphorylation of MEK/ERK. or cleave Bid, a Bcl-2 inhibitory BH3-domain-contain- Whether this reflects a similar or distinct mechanism ingprotein. Truncated Bid (tBid) translocates to the from that suppressingactivation of caspase-8 and mitochondrial membrane and induces loss of mitochon- caspase-9 remains to be elucidated. Nevertheless, as drial transmembrane potential and release of apoptotic shown diagrammatically in Figure 5, the pathways factors such as cytochrome c. Our preliminary data involved in POX-induced apoptosis include both the showed no change in tBid, nor did it show other intrinsic and extrinsic pathways, and specific mechan- evidence indicatingthe sequential activation of the two pathways. Thus, the two pathways apparently were activated separately by ROS/superoxides or ROS/superoxides-induced changes of other proteins. All three members of MAPK family have been widely implicated in the apoptotic process. The regulation of apoptosis by MAPKs is complex and often controversial and may depend on the specific element in the MAPK pathway. Generally, p38 and JNK are considered to be double-edged swords, either inducing or inhibiting apoptosis, dependingon the specific inducer, cell type, etc (Wada and Penninger, 2004). On the other hand, activation of the ERK pathway is generally considered to be an anti-apoptotic survival signal induced by growth factors (Wada and Penninger, 2004). It has been reported that ERK activation plays anti-apoptotic roles in the apoptotic signaling of TNF, Fas, TRAIL, chemotherapeutic drugand serum deprivation (Wilson et al., 1999; Cross et al., 2000; Holmstrom et al., 2000; Tran et al., 2001; Wada and Penninger, 2004). More- over, the MEK/ERK pathway blocks apoptosis by inhibitingthe activation of caspase-8 induced by extrinsic stimulation of death receptors (Tran et al., 2001; Soderstrom et al., 2002). Our data is in keeping with these notions and indicate that ERK played an important role in suppressingPOX-induced apoptosis but whether p38 and JNK play similar roles remains unknown and will require additional studies. Multiple lines of evidence indicate that ROS actively Figure 5 Proposed model by which POX induces apoptosis in participate in the process of apoptosis through mito- cancer cells. POX, an inner mitochondrial membrane enzyme that chondrial or death receptor pathway. Whether it is converts proline to P5C, can be upregulated by p53 and other superoxide, hydrogen peroxide or some other derived signals. The role of P5C in apoptosis is still unknown. During the reactive radical that directly modulate target proteins process mediated by POX, ROS/superoxides are generated which play a principal role in POX-induced apoptosis. It activates the remain unknown. Hydrogen peroxide is often impli- intrinsic pathway by causingthe release cytochrome c to cytosol cated because it is longer lived than superoxide. and the activation of caspase-9. It also activates the extrinsic However, the generation of superoxide radicals by pathway by inducingTRAIL and DR5 and then causingthe POX plays a critical step as MnSOD, which converts activation of caspase-8. The two pathways culminate in the superoxide to hydrogen peroxide, inhibited the activa- activation of caspase-3 and the substrates’ cleavage. In this process, MEK/ERK signaling is inhibited; its upregulation may inhibit tion of both caspase-8 and -9 (Figure 3b). Although POX-induced apoptosis. On the other hand, the effect of POX and other antioxidant enzymes, that is, catalase, glutathione ROS/superoxides may be mediated by Ca2 þ /Calcineurin-NFAT peroxidase, may also play a role, it is MnSOD that cascade to activate apoptotic pathways. The detailed mechanism by appears to provide the major defense against oxidative which MEK/ERK signaling inhibits and Ca2 þ /Calcineurin-NFAT pathway mediates POX-induced apoptosis, and how these signal- stress and is widely recognized as an anti-apoptotic ingpathways interact each other, need further elucidation. agent (Pani et al., 2004). It may be that superoxide Abbreviations: PRO, proline; X-PRO, imidodipeptide with C- radicals generated by mitochondrial respiratory chain terminal proline; cyt.c, cytochrome c.

Oncogene Mechanistic study of POX-induced apoptosis YLiuet al 5646 isms include the ROS/superoxides regulation of Ca2 þ / fected usingLipofectamine 2000. Transfected cells were lysed calcineurin-mediated activation of NFAT and the 36 h after transfection and luciferase activity was measured inhibition of signaling in the MAPK pathway presum- with equal amounts of cell extract usinga TD-20/20 ably by inhibitingkinases. luminometer (Turner Designs, Sunnyvale, CA, USA) and Finally, these findings may provide new insights into normalized with the Renilla activity. the unique bioenergetic and regulatory contributions of proline metabolism previously alluded to by several Cell growth assays The CellTiter 96t aqueous non-radioactive cell proliferation investigators (Adams, 1970; Phang, 1985; Phang et al., assay (MTS assay; Promega, Madison, WI, USA), performed 2001). In the context of our current studies, POX accordingto the manufacturer’s protocol, was used to measure generates two metabolites, P5C and ROS/superoxides. cell growth. Approximately 25 000 cells were seeded in wells of Although the current study emphasizes ROS/super- a 24-well plate and doxycycline or vehicle was added to block oxides, the generation of P5C by POX also may be an or induce the expression of POX by DLD-1 Tet-Off POX cells. important metabolic consequence of POX induction. A solution containingMTS (3-(4,5-dimethylthiazol-2-yl)-5-(3- The exposure of cells to exogenous P5C has marked carboxy-methoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) consequences on glucose metabolism in cultured mam- and PMS (phenazine methosulfate) (20:1 v/v) was added to the malian cells (Phang et al., 1982). In plants and yeast, cells for 2 h at 371C and absorption at 490 nm was determined. P5C has been shown to be cytotoxic and may play a role Each data point was performed in triplicate, and the results were reported as mean absorption7standard error. in cell death. It also has not escaped our notice that collagen is the most abundant protein in the body and Adenovirus infection proline together with hydroxyproline constitutes nearly Recombinant adenoviruses, Ad-GFP, Ad-Empty, Ad- one-third of the residues in collagens. Thus, there is a MnSOD, were purchased from Gene Therapy Vector Core ready reservoir of substrates for mediatingmetabolic Facility of the University of Iowa. Ad-GFP was used to requirements especially duringnutritional stress. Indeed, determine the infection efficiency. The multiplicity of infection identifyingrelated molecules and metabolic interactions (MOI) of different recombinant viruses was used to reach between the proline metabolic enzymes, metabolic 100% infection but did not cause dramatic cell death by virus intermediates and substrate reservoirs on the one hand alone. Ad-POX was supplied by Dr CA Hu of University of (Phang, 1985) and novel signaling mechanisms on the New Mexico. other hand is the focus of our current research. Recent studies demonstratingthe absent or reduced expression Western-blot analysis of POX in kidney tumors compared with their normal The whole-cell lysates were prepared in a buffer containing tissue counterparts (Maxwell and Rivera, 2003) and the 50 mM Tris-HCl (pH 6.8), 2.0% SDS, 10% glycerol and protease inhibitor cocktail tablets (Roche, Mannheim, Ger- anti-tumor effects of proline analogs (Hamilton et al., many). NE-PERt Nuclear and Cytoplasmic Extraction 2004), suggested a cancer suppressor role for POX and Reagents supplied by Pierce Biotechnology (Rockford, IL, thereby identified it as a possible target for cancer USA) was used to extract nuclear protein accordingto the prevention and treatment. manufacturer’s protocol. Equal amounts of extract were electrophoresed on an acrylamide denaturinggeland trans- ferred by electroblottingonto a nitrocellulose membrane. The Materials and methods primary antibodies used were caspase-8, DR4, DR5, TRAIL, FAS, FASL, FADD, NFAT, histone H1(Santa Cruz Co. Cell culture and transfection Santa Cruz, CA), tBID, caspase-3, caspase-9, PARP, ERK, p- The DLD-1 human colon cancer cells are from America Type ERK, JNK, p-JNK, p38, p-p38, MEK, p-MEK, AKT, p-AKT Culture Collection (Manassas, VA, USA). The generation of (Cell Signaling, Beverly, MA, USA), Smac/DIABLO (Calbio- the DLD-1 Tet-Off POX cells has been previously described chem, San Diego, CA, USA), and Actin (Sigma, St Louis, MO, (Donald et al., 2001). The cells were maintained in McCoy’s USA). Blots were developed usingthe chemiluminescence 5A medium (Quality Biological, Gaithersburg, MD, USA) (ECL) procedure (Amersham, Piscataway, NJ, USA). supplemented with 10 % fetal bovine serum (HyClone Laboratories, Logan, UT, USA) in the presence of 0.4 mg/ml RNase protection assay G418, 0.25 mg/ml hygromycin B and 20 ng/ml doxycycline The total RNA of DLD-1 Tet-Off POX cells, which were (DOX). For generation of DLD-1-caMEK cells, DLD-1 cells cultured with or without DOX for 1 or 3 days were isolated were transfected with caMEK expressingconstruct (kindly usingan RNeasy kit (Qiagen,Valencia, CA, USA). The RPA provided by Dr J Pouyssegur, Nice, France) using Lipofecta- was done usingAmbion’s RPA III TM (Austin, TX, USA) and mine 2000 (Invitrogen, Carlsbad, CA, USA) and selected with BD PharMingen’s hAPO-3d Multi-Probe Set (San Diego, CA, geneticin according to manufacture’s protocol. USA) accordingto the manufacturers’ instruction.

Luciferase assay Flow cytometry TRAIL and NFAT promoter activity in DLD-1 cells was DLD-1 Tet-Off POX cells were cultured in the presence or measured usingthe Dual-Luciferase t Reporter Assay (Pro- absence of DOX to block or induce the expression of POX. On mega, Madison, WI, USA) according to the manufacturer’s days 0, 2, 4, 6, a total of 1 million cells were harvested and protocol. The cells were co-transfected with the TRAIL-luc washed twice with cold PBS. The pellets were fixed in 70% ice- (kindly provided by Dr BM Evers of UT Galveston Medical cold methanol for a half an hour. Before runningon the flow Branch, TX, USA) or NFAT-luc (Stratagene, La Jolla, CA, cytometer, the cells were washed twice with cold PBS and USA) reporter constructs and pRL-TK, a Renilla construct for incubated with 5 ml RNase (200 units/ml, DNase-free) for normalizingof transfection efficiency. The cells were trans- 15 min. The cells were stained with 10 mg/ml propidium iodide

Oncogene Mechanistic study of POX-induced apoptosis YLiuet al 5647 for at least 1 h in the dark. Stained cells were analysed with an oxygen species; TRAIL, tumor necrosis factor-related apop- EPICS-XL-MCL ﬂow cytometer (Beckman Coulter, Inc., tosis inducingligand. Miami, FL, USA). Each data point was performed in triplicate, and the results Acknowledgements were reported as the mean7standard error. This research is supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Abbreviations Cancer Research. This project also has been funded in part with Federal funds from the National Cancer Institute, Ad, adenovirus; Dox, doxycycline; DR5, death receptor 5; National Institutes of Health, under Contract No. NO1-CO- MnSOD, manganese superoxide dismutase; MOI, multiplicity 12400. The content of this publication does not necessarily of infection; NFAT, nuclear factor of activated T cells; P5C, reﬂect the views or policies of the Department of Health and pyrroline-5-carboxylate; PBS, phosphate-buffered saline; PIG, Human Services, nor does mention of trade names, commer- p53-induced gene; POX, proline oxidase; ROS, reactive cial products, or organizations imply endorsement by the US Government.

References

Adams E. (1970). Int Rev Connect Tissue Res 5: 1–91. Pani G, Colavitti R, Bedogni B, Fusco S, Ferraro D, Borrello Cai J, Jones DP. (1998). J Biol Chem 273: 11401–11404. S. (2004). Curr Med Chem 11: 1299–1308. Crabtree GR, Olson EN. (2002). Cell 109(Suppl): S67–79. PhangJM. (1985). Curr Top Cell Regul 25: 91–132. Cross TG, Scheel-Toellner D, Henriquez NV, Deacon E, PhangJM, DowningSJ, Yeh GC, Smith RJ, Williams JA, Salmon M, Lord JM. (2000). Exp Cell Res 256: 34–41. Hagedorn CH et al. (1982). J Cell Physiol 110: 255–261. DelwingD, Bavaresco CS, Chiarani F, Wannmacher CM, PhangJM, Hu C-A, Valle D. (2001). The Metabolic and Wajner M, Dutra-Filho CS et al. (2003). Brain Res 991: Molecular Bases of Inherited Disease. Scriver CR, Beaudet 180–186. AL, Sly WS and Valle D (eds). McGraw Hill: New York, pp Donald SP, Sun XY, Hu CA, Yu J, Mei JM, Valle D et al. 1821–1838. (2001). Cancer Res 61: 1810–1815. Polyak K, Xia Y, Zweier JL, Kinzler KW, Vogelstein B. Duque J, Fresno M, Iniguez MA. (2005). J Biol Chem 280: (1997). Nature 389: 300–305. 8686–8693. Rivera A, Maxwell SA. (2005). J Biol Chem 280: 29346–29354. Dussmann H, Kogel D, Rehm M, Prehn JH. (2003). J Biol Semenza GL. (2003). Nat Rev Cancer 3: 721–732. Chem 278: 12645–12649. Soderstrom TS, Poukkula M, Holmstrom TH, Heiskanen Hagedorn CH, Phang JM. (1983). Arch Biochem Biophys 225: KM, Eriksson JE. (2002). J Immunol 169: 2851–2860. 95–101. Tran SE, Holmstrom TH, Ahonen M, Kahari VM, Eriksson Hamilton G, Olszewski U, Ulsperger E, Baumgartner G. JE. (2001). J Biol Chem 276: 16484–16490. (2004). J Clin Oncol 22: 3164. Wada T, Penninger JM. (2004). Oncogene 23: 2838–2849. Holmstrom TH, Schmitz I, Soderstrom TS, Poukkula M, WangQ, Ji Y, WangX, Evers BM. (2000). Biochem Biophys Johnson VL, Chow SC et al. (2000). EMBO J 19: 5418–5428. Res Commun 276: 466–471. Kelley SK, Ashkenazi A. (2004). Curr Opin Pharmacol 4: 333– Wenzel U, Kuntz S, De Sousa UJ, Daniel H. (2003). Int J 339. Cancer 106: 666–675. Latinis KM, Norian LA, Eliason SL, Koretzky GA. (1997). J Wenzel U, Nickel A, Kuntz S, Daniel H. (2004). Carcinogen- Biol Chem 272: 31427–31434. esis 25: 703–712. Liu Y, Borchert GL, Donald SP, Surazynski A, Hu CA, Wilson DJ, Alessandrini A, Budd RC. (1999). Cell Immunol Weydert CJ et al. (2005). Carcinogenesis 26: 1335–1342. 194: 67–77. Loo G. (2003). J Nutr Biochem 14: 64–73. Wu GS, Burns TF, McDonald III ER, JiangW, MengR, Lu KP. (2003). Cancer Cell 4: 175–180. Krantz ID et al. (1997). Nat Genet 17: 141–143. Macias MJ, Wiesner S, Sudol M. (2002). FEBS Lett 513: 30– Wu GS, Kim K, el-Deiry WS. (2000). Adv Exp Med Biol 465: 37. 143–151. Maxwell SA, Rivera A. (2003). J Biol Chem 278: 9784–9789. Yu J, WangZ, Kinzler KW, VogelsteinB, ZhangL. (2003). McCaffrey PG, Goldfeld AE, Rao A. (1994). J Biol Chem 269: Proc Natl Acad Sci USA 100: 1931–1936. 30445–30450. Yu J, ZhangL, HwangPM, Kinzler KW, VogelsteinB. (2001). Nakano K, Vousden KH. (2001). Mol Cell 7: 683–694. Mol Cell 7: 673–682.

Oncogene