The Pennsylvania State University. The Graduate School. The Huck Institutes of the Life Sciences ROLES OF PEROXISOME PROLIFERATOR-ACTIVATED RECEPTOR-

Transcription

1 The Pennsylvania State University The Graduate School The Huck Institutes of the Life Sciences ROLES OF PEROXISOME PROLIFERATOR-ACTIVATED RECEPTOR- BETA/DELTA AND B-CELL LYMPHOMA 6 IN PANCREATIC CANCER A Dissertation in Pathobiology by Jeffrey David Coleman 2010 Jeffrey David Coleman Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy December 2010

2 The dissertation of Jeffrey David Coleman was reviewed and approved* by the following: John P. Vanden Heuvel Professor of Veterinary and Biomedical Sciences Dissertation Advisor Chair of Committee Sandeep K. Prabhu Assistant Professor of Veterinary and Biomedical Sciences Adam Glick Associate Professor of Veterinary Science Karam El-Bayoumy Professor of Biochemistry and Molecular Biology Anthony P. Schmitt Assistant Professor of Molecular Immunology Graduate Program Officer for the Department of Pathobiology *Signatures are on file in the Graduate School

3 iii ABSTRACT The Peroxisome proliferator-activated receptors are ligand-activated transcription factors and members of the nuclear receptor superfamily. The PPARs are implicated in the regulation of various cellular processes, including glucose and lipid homeostasis, fatty acid oxidation, inflammation and cell proliferation and invasion. Of the three subtypes PPARα, PPARβ/δ and PPARγ PPARβ/δ is the least well understood in terms of its endogenous ligands and biological role(s). The PPARβ/δ null mouse is more susceptible to hepatotoxicity than the wild-type mouse following challenge with Arsenic or CCl 4, compounds known to induce toxicity via increased reactive oxygen intermediates. Recently, PPARβ/δ has generated much interest in its ability to regulate the inflammatory response in coordination with the transcriptional repressor, B-cell lymphoma-6 (BCL-6). This dissertation will examine the hypothesis that modulation of oxidative stress underlies the protective role of PPARβ/δ in the mouse liver, and that the regulation of inflammation and cell invasion by PPARβ/δ and BCL-6 is a significant role for the receptor in human pancreatic cancer cells. Oxidized very-low density lipoprotein (oxvldl) and its constituents, including 13-S- hydroxyoctadeca-dienoic acid (13-S- HODE) and 4-hydroxynonenal (4-HNE), are endogenous PPARβ/δ activators in PPREdependent reporter assays. A structure-activity relationship was established where 4-HNE and 4-hydroperoxynonenal (4-HpNE) enhanced the activity of the PPARβ/δ subtype. Other oxidative stress mediators, including 4-hyroxy-hexenal (4-HHE), 4-oxo-2-nonenal (4-ONE), and trans-4,5-epoxy-2(e)-decenal, did not activate this receptor in these experiments. Gene expression assays using both wild-type and PPARβ/δ -/- cells

4 iv demonstrated that 4-HNE induced gene expression in a PPARβ/δ-dependent manner. Furthermore, wild-type cells were more resistant to 4-HNE-induced toxicity (EC 50 68µM) compared with PPARβ/δ -/- cells (EC 50 5µM). Consistent with this observation, addition of a synthetic PPARβ/δ activator protected cells from 4-HNE-induced toxicity, while treatment with a PPARβ/δ antagonist reversed this observation. Microarray analyses using both wild-type and PPARβ/δ -/- cells indicated several anti-oxidant genes known to be involved in 4-HNE metabolism that were increased at the transcript level in a PPARβ/δ-dependent manner. Molecular modeling, using the coordinates of PPARβ/δ bound to another endogenous ligand, eicosapentadienoic acid (EPA), indicated that 4- HNE forms a putative hydrogen bonding interaction with His413 in the PPARβ/δ ligandbinding pocket. Thus, 4-HNE binding activates the PPARβ/δ subtype, increasing transcription of anti-oxidant and detoxification genes in a PPARβ/δ-dependent fashion, and this observation may account for the protective role of the receptor in the mouse liver. The identification of several PPARβ/δ-regulated anti-oxidant genes prompted further studies to investigate the role of the receptor in regulating the inflammatory response in human cancer cells. Forced PPARβ/δ over-expression inhibited basal and TNFα-induced NF-κB activity in reporter assays. The synthetic PPARβ/δ agonist, GW induced a physical interaction between the receptor and the p50 subunit of NF-κB in mammalian 2-hybrid assays. Furthermore, GW reduced basal and TNFα-induced NF-κB luciferase activity. Short-hairpin RNAi knock-down of PPARβ/δ attenuated the GW effect on NF-κB luciferase, while knock-down of BCL-6

5 v significantly enhanced TNFα-induced NF-κB activity. The use of GW and shrnai technology identified several anti-inflammatory genes that were increased in a PPARβ/δ-dependent fashion. Moreover, GW inhibited MCP-1 luciferase activity, although this effect was dependent on BCL-6. Real-time PCR experiments identified several pro-inflammatory genes that were regulated in a BCL-6-dependent manner following GW treatment. Finally, GW conditioned pancreatic cancer cell media affected pro-inflammatory gene expression and cell invasion abilities of differentiated THP-1 macrophages. Together, these observations identify that the PPARβ/δ- and BCL-6-dependent anti-inflammatory pathways are active in the human pancreas, and that the PPARβ/δ-specific agonist GW may prove useful in controlling inflammation and immune cell infiltration in pancreatic cancer cells. Several pro-inflammatory genes are consistently over-expressed in pancreatic cancers and are involved in metastasis. The matrix-remodeling matrix metalloproteinase 9 (MMP-9) has been linked to PPARβ/δ and BCL-6 in macrophages, and is one of several genes highly expressed in pancreatic cancers. Indeed, real-time PCR analyses identified both MMP-9 and PPARβ/δ as being significantly elevated in human pancreatic cancer tissue at the mrna level, while BCL-6 was significantly repressed. PPARβ/δ activation by GW was sufficient to reduce the TNFα-induced expression of various genes implicated in metastasis and invasion, including MMP-9 mrna and protein, and reduced the invasion of pancreatic cancer cells through a basement membrane in cell culture models. The use of shrna technologies indicated that the receptor was responsible for the ligand effects observed, while BCL-6 was responsible

6 vi for inhibiting MMP-9 and pro-inflammatory gene expression. These observations indicate that both PPARβ/δ and BCL-6 play important roles in regulating pancreatic cancer cell invasion. Recent studies using other cell lines have implicated PPARβ/δ in the regulation of both the inflammatory response and cell invasion and proliferation. Understanding the mechanisms by which PPARβ/δ agonists inhibit and ameliorate reactive oxygen, the inflammatory response and cell invasion may prove useful in the generation of alternative therapies for diseases in which inflammation and reactive oxygen species play a causative role. Furthermore, the observations presented in this dissertation suggest that PPARβ/δ agonists may be used in concert with chemotherapies to combat and control cancerrelated metastases.

7 vii TABLE OF CONTENTS LIST OF FIGURES...x LIST OF TABLES...xii ACKNOWLEDGEMENTS...xiii Chapter 1: Literature Review Peroxisome Proliferator-Activated Receptors Molecular Biology of PPARβ/δ Structure DNA Binding Endogenous and Specific PPARβ/δ Ligands Regulation of PPARβ/δ Biological Roles of PPARβ/δ Metabolic Syndrome PPARβ/δ in Inflammation Oxidative Stress Cell Proliferation and Cancer B-Cell Lymphoma-6 (BCL-6) Structure Interaction with PPARβ/δ Mechanisms of Transrepression Involvement in the Immune System BCL-6 as a Proto-Oncogene Anti-inflammatory Mechanisms of BCL Oxidative Stress PUFA-Peroxidation Products alkenals hydroxy-2-alkenals Ketoaldehydes hydroxynonenal (4-HNE) Pancreatic Cancer General Remarks Molecular Genetics of Pancreatic Cancer K-ras Oncogenes Tumor Supressor Genes Animal Models Stages...47

8 1.6.4 Descriptive Epidemiology Frequency Age Race Etiology Diabetes Pancreatitis Tobacco and N-Nitroso Compounds Alcohol and Coffee Dietary and Nutritional Factors Treatment Options Conclusions References...57 Chapter 2: The oxidative stress mediator 4-hydroxynonenal is an intracellular agonist of the nuclear receptor peroxisome proliferator-activated receptor-β/δ (PPARβ/δ) Abstract Introduction Materials and Methods Results Discussion References...98 Chapter 3: Peroxisome Proliferator-Activated Receptor β/δ (PPARβ/δ) and B-cell lymphoma 6 (BCL-6) regulate inflammatory signaling in pancreatic cancer cells Abstract Introduction Materials and Methods Results Discussion References Chapter 4: Role of Peroxisome Proliferator-Activated Receptor β/δ (PPARβ/δ) and B-cell lymphoma 6 (BCL-6) in regulation of genes involved in metastasis and migration in pancreatic cancer cells Abstract Introduction Materials and Methods Results Discussion viii

9 ix 4.6 References Chapter 5 Summary and Perspectives Future Directions References...179

10 x LIST OF FIGURES Figure 1-1: Structure of the PPARs....7 Figure 1-2: Chemical structures of selective PPARβ/δ activators...12 Figure 1-3: Domain structure of BCL-6 protein including sites of interaction with various cellular proteins...26 Figure 1-4: PPARβ/δ acts as an inflammatory switch via its association and dissociation with the transcriptional repressor BCL Figure 1-5: Chemical structures of A) reactive aldehydes and B) lipid peroxidation-specific aldehydes Figure 1-6: The non-enzymatic formation of the 4-HNE precursor, 4-HpNE, from 13-HODE via a Hock cleavage mechanism Figure 1-7: Progression of pancreatic cancer from normal pancreatic tissue to malignant neoplasia Figure 2-1: Oxidation of VLDL leads to increased activation of PPARβ/δ Figure 2-2: Structures of 4-HNE and other test molecules examined Figure 2-3: 4-HNE regulates gene expression in a PPARβ/δ-dependent manner Figure 2-4: X-ray structure of PPARβ/δ bound to eicosapentaenoic acid...95 Figure 3-1: Effects of PPARβ/δ, and ligand-activation of PPARβ/δ, on NF-κB activity Figure 3-2: GW treatment contributes to the anti-inflammatory actions of PPARβ/δ by increasing expression of target genes Figure 3-3: The transcriptional repressor BCL-6 contributes to the antiinflammatory actions of GW by suppressing target gene expression Figure 3-4: Effects of PPARβ/δ and BCL-6 knock-down in COX-2 positive human pancreatic cancer cells

11 Figure 3-5: Conditioned media from Miapaca-2 cells influences gene expression in differentiated THP-1 cells Figure 3-6: GW conditioned Miapaca-2 media reduces the percentage of invading THP-1 cells across a basement membrane Figure 4-1: Relative mrna expression in human pancreatic tissues at varying stages of carcinogenesis from chronic pancreatitis to pancreatic cancer Figure 4-2: Effect of PPARβ/δ activation on MMP-9 expression Figure 4-3: Effects of PPARβ/δ and BCL-6 knock-down on Miapaca-2 gene expression Figure 4-4: GW treatment reduces TNFα-stimulated Miapaca-2 and BxPc- 3 cell invasion through a basement membrane Figure 5-1: Proposed model for the regulatory cascade resulting from PPARβ/δ activation by endogenous and synthetic ligands xi

12 xii LIST OF TABLES Table 1-1: List of endogenous and synthetic PPARβ/δ ligands Table 2-1: Microarray analysis of gene expression: Genes statistically regulated by PPARβ/δ in MuSH hepatocytes Table 3-1: List of Real-time PCR primers used in this study Table 4-1: List of Real-time PCR primers used in this study...142

13 xiii ACKNOWLEDGEMENTS I sincerely thank Dr. Jack Vanden Heuvel, my advisor, for graciously accepting me into his laboratory. I would like to thank him for his kind words, his constant willingness to help and his guidance. I would also like to thank my dissertation committee members, Dr. Sandeep Prabhu, Dr. Adam Glick and Dr. Karam El-Bayoumy, for their suggestions and insight. I would also like to thank Dr. Prabhu for his constant assistance and unwavering support since I began my graduate research. I am truly grateful to all the current and past members of the Vanden Heuvel laboratory for their support, friendship and assistance. I would like to specifically mention Dr. Jerry Thompson for training me in standard laboratory protocol, and for always being available to answer questions and offer suggestions for ways in which to improve my research. I would also like to thank Daniel Hannon for training me in RNA techniques and for his consistent and reliable help. I would also like to extend a special note of thanks to my friends and family; particularly to my grandparents, Earl and Charlotte Wilbur, for providing for me; my mother, Karen Wilbur, for her unfailing support and love; my brother, Daniel, for being my best friend; Dr. Ilya and Dr. Tamila Burshteyn, for providing a second home for me in New York; and Inna Burshteyn, for her love.

14 Chapter 1 Literature Review 1.1 Peroxisome Proliferator-Activated Receptors Peroxisome proliferator-activated receptors (PPARs) are ligand-activated transcription factors and members of the nuclear receptor (NR) superfamily. The PPARs consist of three subtypes, PPARα (NR1C1), PPARβ/δ (NR1C2), and PPARγ (NR1C3), that have been identified in in several vertebrate species including Xenopus, mouse, rat, hamster and human (1-11). The three subtypes were so named following their discovery in Xenopus (11), two years after their discovery in the mouse (4). The first subtype, PPARα, was cloned from the mouse and affected proliferation of peroxisomes in response to rodent hepatocarcinogens (4). The two other closely related subtypes, PPARβ/δ (also referred to as NUC1 and FAAR) and PPARγ, were subsequently cloned and found to induce transcription of the acyl coenzyme A oxidase gene via a cognate responsive element (11). Each receptor subtype displays differing affinities for ligand as well as varying tissue distributions. In the human, PPARα is expressed in the heart, kidney, skeletal muscle, and large intestine, PPARγ is found primarily in adipose, colon, liver and heart. PPARβ/δ is ubiquitously expressed and usually at higher levels than the other two subtypes, although it is primarily found in high levels in skin, colon and brain

15 2 (12, 13). In humans, these transcription factors mediate a wide variety of biological processes, and are implicated in the maintenance of lipid and glucose homeostasis, energy metabolism, inflammation, cell proliferation and invasion, and tumorigenesis, to name a few. Of the three subtypes, PPARγ is most understood. The transcription and differential use of three separate promoters, along with splicing variations of three 5 - exons (A1, A2, and B1) drives expression of three mrnas generating two predominant proteins PPARγ1 and PPARγ2 (2). The use of γ1 and γ3 promoters results in the same 477 amino acid protein, while the γ2 promoter generates the 505 amino acid protein known as PPARγ2 (2). Studies conducted involving subjects with varying degrees of obesity determined that PPARγ2, and not PPARγ1, controlled adipocyte function (14). Indeed, while PPARγ1 is expressed several human tissues, PPARγ2 is limited primarily to the adipocyte (15). PPARγ binds and is activated by both ω-3 and ω-6 polyunsaturated fatty acids, including α-linolenic (C18:3), γ-linolenic (C18:3), eicosapentaenoic (C20:5, EPA), docohexaenoic (C22:6, DHA), arachidonic (C20:4), as well as several eicosanoids generated from arachadonic acid, such as 9-HODE (9-hydroxyoctadenoic acid), 13- HODE, and 15-Deoxy-Δ12, 14-PGJ2, arguably its most widely studied endogenous activator. The hypoglycemic agents thiazolidinediones, including Roziglitazone and Troglitazone, are synthetic PPARγ activators. Non-steroidal anti-inflammatory drugs (NSAIDs), including Indomethacin, Fenoprofen and Ibuprofen, further activate PPARγ in saturation binding (SABA) and competition binding assays (COBA) (2). PPARγ plays a major role in adipogenesis by serving as a lipid sensor, and activators of the receptor, particularly the TZDs, are used to control diabetes by improving insulin resistance and

16 3 pancreatic beta-cell function (16, 17). Furthermore, PPARγ is an attractive antiproliferative target, and its activators induce growth arrest in 3T3-L1 cells (adipocytes), primary human liposarcoma cells, and several malignant cancer cell lines, including breast, colon and prostate (15). PPARγ ligands also modulate the inflammatory response, and specific PPARγ activators inhibit inflammatory gene expression and interfere with pro-inflammatory transcription factor signaling pathways in vascular and inflammatory cells (18). PPARα, like PPARβ/δ and PPARγ, also binds fatty acids and eicosanoids, although with much higher affinity (2). Arachidonic acid metabolites, such as 8-(S)- HETE and Leukotriene B4, are strong endogenous PPARα ligands (19). Hypolipidemic agents, particularly Wy-14,643, plasticizers and herbicides are synthetic PPARα activators, and induce peroxisome proliferation and hepatocarcinogenesis in rodents (20). Like PPARγ, PPARα activators are also implicated in regulating the inflammatory and immune response via decreased activity of the master inflammatory regulator NF-κB (18, 21). PPARα is a key mediator of both peroxisomal and mitochondrial β-oxidation pathways, liver fatty acid synthesis, the adaptation to fasting and tumor promotion. PPARβ/δ is probably the least well understood PPAR subtype in terms of its endogenous ligands and biological role. PPARβ/δ binds PUFAs and eicosanoids, although with less affinity than PPARα and PPARγ, and its synthetic ligands include several hypolipidemic agents that induce fatty acid catabolism, increase serum HDL cholesterol, and mediate the inflammatory response (22). Recently, PPARβ/δ has been identified as a key player in mediating the inflammatory response in vivo (23) and in vitro (23, 24). Three major mechanisms for the PPARβ/δ-mediated repression of

17 4 inflammation have been identified: 1) PPARβ/δ-specific ligands induce the dissociation of the transcriptional repressor B cell lymphoma-6 (BCL-6), 2) inhibition of NF-κB activity, and 3) induction of anti-inflammatory target genes (22). The physical association with BCL-6 is a property not shared by either PPARα or PPARγ (24), representing a unique way in which PPARβ/δ influences the inflammatory response. Activation of PPARβ/δ induces terminal differentiation in keratinocytes (25) as well as the induction of an apoptotic-like pathway (26), although the role of PPARβ/δ in cell proliferation remains controversial (26). In keratinocytes, for example, activation of PPARβ/δ by the specific ligand GW induced migration and proliferation (27), implicating the receptor as a master regulator of wound healing. Moreover, PPARβ/δ activation may either potentiate (28) or attenuate cell proliferation and tumorigenesis (29). PPAR activators have attracted considerable attention, particularly because they are master regulators of various cellular processes, and PPAR activation has proven effective in controlling the metabolic syndrome, inflammation, type 2 diabetes, cell proliferation and invasion, and cancer. This dissertation will focus on PPARβ/δ, its association with the transcriptional repressor BCL-6, and the potential therapeutic uses of specific PPARβ/δ activators in the control of inflammation and cell migration in the pancreas.

18 5 1.2 Molecular Biology of PPARβ/δ Structure As a member of the superfamily of steroid/thyroid hormone nuclear receptors, PPARβ/δ displays a typical nuclear receptor domain structure (Figure 1-1) (30). All three PPAR subtypes contain the following domains: an N-terminal A/B domain containing a ligand-independent transactivation function (AF-1); a highly conserved, cysteine-rich C (DNA-binding) domain containing two zinc-finger-like motifs, which is the hallmark of the nuclear receptor superfamily (31); a D domain, called the hinge domain, potentially allowing for flexible interactions between the C and E/F domains; and an E/F domain containing the ligand-dependent transactivation function 2 (AF-2), responsible for ligand binding and effecting ligand-dependent transcriptional activity. The E/F domain also provides an interface for dimerization and protein-protein interaction. PPARα and PPARγ are phosphoproteins, and phosphorylation of the A/B domain by mitogen-activated protein kinases (MAPKs) can modulate the transcriptional activity of these subtypes (32, 33). The DNA-binding domain contains two globular zinc finger moieties, called the proximal (P-box) and the distal (D-box), oriented perpendicular to the P-box (31). The P-box is α-helical, and is responsible for binding the major groove in DNA and recognizing the consensus sequence, AGGTCA, in target gene promoters (34). The D-box is also a helix, and may play a role in spatial orientation of the DNA binding domain. The E/F domain, located at the C-terminus, consists of 13 α helices and a fourstranded β-sheet forming a large Y-shaped hydrophobic ligand cavity (30). This region is

19 6 also responsible for mediating protein-protein interactions with transcriptional coactivators such as steroid receptor coactivator 1 (SRC-1) (36) and CREB-binding protein (CBP) (37) in a ligand-dependent manner. Ligand-binding induces a conformational change effectively closing the binding pocket via helix 12 (termed the mouse trap model (30)) and an interaction with its obligate partner RXR (NR2B), the nuclear receptor for 9-cis retinoic acid (38, 39). The PPARs have a binding cavity of approximately 1300 Å3, which is roughly twice the size of other nuclear receptors, and may contribute to the ability of the PPARs to bind a variety of natural and synthetic ligands. In the case of PPARβ/δ, ligands occupy only about 30-40% of the binding pocket (40). One characteristic of PPAR binding pockets that is absent in other nuclear receptors is the presence of helix 2, which may serve to increase the size of the pocket and participates in an entry channel for ligand (30). The ligand-binding domain (LBD) of PPARβ/δ is unique from PPARα and PPARγ in that it is narrower in the vicinity of the AF-2 helix, a feature that potentially contributes to its particular ligand binding profile. The PPARβ/δ LBD shares 70% and 68% amino acid identity with PPARα and PPARγ, respectively.

20 Figure 1-1: Structure of the PPARs. A) Domain organization of the PPARs and percentage of amino acid identity (compared with PPARα) are depicted. The A/B domain contains the ligand-independent activation function-1 (AF-1). The C domain is the DNA binding domain. The D domain is a hinge region. The E/F domain is the ligand-binding domain, and is responsible for heterodimerization with RXR. The E/F domain also contains the ligand-dependent activation function-2 (AF-2). B) Comparisons of the three-dimensional structures of the LBDs of each PPAR isotype. The solvent-accessible binding pocket is pictured in white. (Source: Hihi, A.K. et al. PPARs: transcriptional effectors of fatty acids and their derivatives. Cell Mol Life Sci 59:790 (2002)) 7

21 DNA Binding The PPARs recognize peroxisome proliferator-response elements (PPREs) in the promoter regions of target genes. The first PPRE to be identified was located in the promoter of the Acyl-CoA Oxidase gene (11, 41). The PPAR consensus sequence is a direct repeat 1 (DR1) motif. Transcription is effected following ligand activation and recruitment and association with RXR. Further analyses using the PPREs located within the promoters of the CYP4A6 and malic enzyme genes identified three other factors that determine PPREs: 1) an extended 5 half-site, 2) an imperfect core DR1 and 3) an adenine as the spacing nucleotide between the two hexamer half-sites (2). A sample PPRE could be the following: 5 - AACTAGGNCA A AGGTCA-3. These factors might infer selectivity to the PPAR:RXR heterodimer, and likewise be more discriminating against homo- and heterodimers of other nuclear receptor complexes that might recognize the consensus DR1 motif, such as androgen receptor-related protein-1 (ARP-1), hepatocyte nuclear factor 4α (HNF-4α, NR2A) and RXRα homodimers (2). The extended 5 half-site facilitate PPAR:RXR binding to DNA, while deletions or mutations within six nucleotides 5' of the DR1 motif dramatically diminished PPAR:RXRα binding (42). The PPAR:RXR heterodimer is oriented on the PPRE with PPAR interacting with the 5 -extended core half-site, while RXR occupies the downstream half-site (43, 44). Interactions of the PPARs with DNA involves a receptor moiety immediately C- terminal to the second zinc finger, called the carboxy-terminal extension (CTE), which is responsible for recognition of the 5 hexamer (2). PPARs cannot bind DNA as monomers, owing to their N-terminal region, but deletion of the A/B domain permitted a

22 9 PPAR:PPAR homodimer to bind DNA (45). Structural analyses using RAR:RXR heterodimers determined that the crucial amino acids involved in PPAR:RXR DNA binding reside in the second zinc finger of the 5 -receptor, outside the D-box, while the second receptor contributes to the dimerization via its CTE (46-48). The differential tissue expression of the three different subtypes infers the possibility of subtype-selective PPREs. Analysis of 16 native PPREs and the relative binding affinities of PPARα, PPARβ/δ and PPARγ enabled the classification of the PPREs into three categories: strong, intermediate and weak elements (44). The number and identities of the 5 -flanking nucleotides, not necessarily the core DR1, which was fairly homologous among the tested PPREs, determined the strength of the PPRE. In all cases, however, PPARγ binds more strongly than the other subtypes, whereas PPARα relies more on conservation of the 5 -flanking nucleotides (44). RXR subtypes also influences PPAR:RXR DNA interactions with binding to strong elements being enhanced by heterodimerization with RXRγ, while its substitution with RXRα confers favorability to binding to weak PPREs (2). Nuclear receptor cross-talk has been demonstrated for most members of this protein family, including the PPARs. The PPAR:RXR unit recognizes the estrogen response element (ERE) in reporter assays. Plasmids containing the ERE-gene upstream of luciferase were transactivated by the PPAR:RXR complex (49). To date, however, no naturally occurring ERE-containing gene is activated by both estrogen receptor (ER) and PPAR:RXR. Conversely, PPAR:RXR transrepresses ER on the vitellogenin gene A2 promoter (49). Furthermore, PPARα competes with thyroid receptor (TR) for heterodimerization with RXR,, hence decreasing TR activity (50).

23 Endogenous and Specific PPARβ/δ Ligands The identification of endogenous ligands for PPARβ/δ has been challenging. Fatty acids and their metabolites were among the first identified PPARβ/δ ligands (2). PUFAs are relatively efficacious PPARβ/δ agonists, while saturated fatty acids are poor PPAR ligands in general (51-53). The preference for PUFAs as PPARβ/δ agonists may be explained by the observation that eicosapentadienoic acid (EPA) forms key hydrogen bonds with the PPARβ/δ AF-2. PUFAs contain long, flexible hydrocarbon tails sufficient for stabilizing interactions yet short enough to fit inside the binding pocket (54) (Table 1-1). As mentioned previously, PPARβ/δ displays a distinct ligand profile, potentially owing to the fact that the binding pocket of PPARβ/δ is narrower than those of its subtype counterparts near the AF-2 helix. Indeed, several hypolipidemic agents, thiazolidinediones, and saturated FAs failed to bind or activate PPARβ/δ in reporter assays (53). Among the PPARβ/δ activators are linoleic acid, arachidonic acid, cpgi, and iloprost, as well as the eicosanoid PGA1 (53). All three subtypes are responsive to 15-Deoxy-Δ12, 14-PGJ2, while cpgi selectively binds PPARα and PPARβ/δ. Thus, it is important to note that while the precise role(s) of PPARβ/δ are still being established, the receptor does recognize and respond to a small set of small molecules, including some that also activate PPARα (53). The difficulty in identifying PPARβ/δ ligands has been exacerbated by speciesspecific ligand binding. In Xenopus, for example, Benzafibrate is a strong PPARβ/δ activator (52), yet displays virtually no activity on the mammalian receptor. Synthetic fibrate derivatives selectively activate human PPARβ/δ and induce interactions with

24 11 nuclear receptor co-activator (CREB-binding protein) in an agonist-dependent manner (55). Studies by Brown et al. identified GW 2433 as a compound that displayed a high, although not selective, affinity for PPARβ/δ (56). Further advances in PPARβ/δ agonist identification have since been made. Minor structural alterations in certain PPARα ligands converts them into potent PPARβ/δ ligands (57). Using GW 2433 as a structural template, Epple et al. synthesized a series of highly potent and selective PPARβ/δ activators with Kd in the nanomolar range and no cross-activation with other subtypes up to micromolar concentrations (58). Selective PPARβ/δ activators also exhibit a Y-shape bearing alkynylallylic moieties (59), or may be achiral, para-alkylthiophenoxyacetic acids (60). The most selective and widely used PPARβ/δ ligands, however, are GW501516, and its analogue GW0742, which were only recently discovered using combinatorial chemistry and structure-based design methods (Figure 1-2). These compounds displayed an EC50 of 1.1 nm for PPARβ/δ, with 1000-fold less selectivity for PPARα or PPARγ (61).

25 Figure 1-2: Chemical structures of selective PPARβ/δ activators. GW501516: X = H, EC 50 = 1.1 nm; GW0742: X = F, EC 50 = 1.0 nm. (Source: Sznaidman, M.L. et al. Novel selective small molecule agonists for peroxisome proliferator-activated receptor delta synthesis and biological activity. Bioorg Med Chem Lett 13:1517 (2003)) 12

26 13 Table 1-1: List of endogenous and synthetic PPARβ/δ ligands. Ligand Class PPARβ/δ Synthetic agonists Fatty acids and derivatives L GW GW0742 GW9578 L L L Docahexanoic acid Arachidonic acid Linoleic acid C6-C8 Eicosanoids 15-deoxy-Δ 12,14 -prostaglandin J 2 PGJ 2 PGB 2 Vitamin Synthetic antagonists Fatty acid inhibitor Other PGA 1/2 All trans retinoic acid GSK0660 Sulindac 2Br-C16 Tetradecylglycidic acid Sphingolipid analogues Perfluorooctanoic acid Perfluorooctane sulfonate

27 Regulation of PPARβ/δ Much effort has been aimed at identifying the biological role(s) of PPARβ/δ; However, the precise regulation of the PPARβ/δ gene has received much less interest, and our current understanding of the processes regulating the expression of this gene is still in its infancy. The PPARβ/δ gene lacks a TATA box, and its transcription is effected via a unique start site located 380 bp upstream of the ATG initiation codon (62). This promoter would later be discovered to be the major region for transcriptional regulation of PPARβ/δ (63). The PPARβ/δ promoter itself is rich in SP-1 binding elements, a hallmark of housekeeping genes (63). PPARβ/δ is also controlled via the binding of AP- 1 and several inflammatory signals (64). While the predominant human PPARβ/δ promoter is established, the possibility that an alternative promoter(s) exist remains, as is the case in its orthologous mouse counterpart (65). Whether or not these variations in alternative splicing contribute to the contradictory findings regarding the role of PPARβ/δ in various diseases remains to be seen (66). While the transcript variants of PPARγ have been known for some time, Lundell et al. (2007) reported a 5 -untranslated exon and multiple 5'-UTRs of PPARβ/δ mrna transcripts that impact translation efficiency (63). Interestingly, their studies also highlight the presence of a 3'-splice variant of human PPARβ/δ (PPARβ/δ2), which may encode a potential dominant negative regulator of gene expression. Differential expression of these PPARβ/δ splice products is seen in various cell lines (HepG2, Huh7, A498, THP-1 and HeLa) and human tissues (skeletal muscle, adult and fetal heart) (63). AP-1 sites in the PPARβ/δ promoter respond to the stress-associated kinase cascade in keratinocytes, inducing PPARβ/δ expression as well as expression of

28 15 endogenous ligands (67). Adding to the complexity is the fact that PPARβ/δ interacts with steroid receptor coactivator-1 (SRC-1) (2), and MAP kinase pathways that phosphorylate SRC-1 (68) may play a role in regulating PPARβ/δ via this interaction. PPARβ/δ is phosphorylated by protein kinase A (PKA) on its DNA-binding domain in in vitro PKA assays, adding another layer of regulation (69). camp-elevating ligands also influence the PPARβ/δ activity by synergistically stimulating PPARβ/δ-mediated transactivation (70).

29 Biological Roles of PPARβ/δ Metabolic Syndrome Metabolic syndrome, also called syndrome X, is an amalgamation of disorders, including obesity, insulin and glucose resistance, hypertension, and dyslipidemia. Activators of PPARα and PPARγ are currently in clinical use as therapeutic agents targeting hypertriglyceridemia and insulin resistance. As our understanding of PPARβ/δ grows, however, it is becoming increasingly clear that the receptor is an attractive molecular target for combating facets of the metabolic syndrome. Dyslipidemia, marked by elevated triglycerides and low HDL cholesterol, is common in metabolic syndrome. The selective PPARβ/δ agonist GW was used in insulin-resistant obese rhesus monkeys to affect a 79% increase in HDL-cholesterol, a 56% decrease in triglycerides, and a 29% decrease in LDL cholesterol (71). In mice, GW increases HDL cholesterol by approximately 50%, regardless of obesity status (72, 73). Increases in HDL cholesterol affected the number, not size, of the HDL particle and increased expression of the reverse cholesterol transporter ABCA1 in human and mouse macrophages, and in human intestinal cells and fibroblasts (73). PPARβ/δ agonists also reduced cholesterol absorption in wild-type mice by significantly reducing expression of the cholesterol absorption protein Niemann-Pick C1-like 1 (Npc1l1) in the intestine (73). Genetic mouse models using constitutively active PPARβ/δ (viral protein 16 PPARδ fusion, VP16-PPARβ/δ) demonstrate PPARβ/δ to be a key regulator in adipose (74) a tissue that plays a key role in the metabolic syndrome (75). Mice expressing

30 17 VP16-PPARβ/δ had increased fat metabolism compared to control littermates, evidenced by decreases in body weight, inguinal fat pad masses, adipocyte triglyceride accumulation, circulating FFAs and triglycerides. Conversely, the PPARβ/δ null mouse is susceptible to weight gain when challenged with a high fat diet (76). PPARβ/δ is very highly expressed in skeletal tissue times higher than PPARα or PPARγ where it regulates muscle fiber reprogramming via increased transcription of genes targeting fatty acid oxidation, mitochondrial respiration and oxidative metabolism (77). Activation of the receptor ameliorated diet-induced obesity and insulin resistance in mice fed a high-fat diet. Furthermore, GW markedly reduced symptoms of diabetes in ob/ob mice by reducing plasma glucose and blood insulin levels (78). The therapeutic effects on insulin sensitivity by PPARβ/δ-specific activation are thought to be mediated by increases in systemic metabolism. Metabolomics studies using ob/ob mice have recently demonstrated that PPARβ/δ activation in the liver dramatically enhanced fatty acid oxidation pathways while simultaneously decreasing gluconeogenesis, with further effects on the TCA cycle, linoleic acid and α-linolenic acid essential fatty acid pathways (79). Furthermore, these changes correlated well with the increased expression of PPAR target genes PPARβ/δ in Inflammation The PPARs are well known regulators of the inflammatory response (80-82). PPARβ/δ regulates the inflammatory response via three unique, and potentially complementary, pathways; 1) the direct regulation of PPARβ/δ responsive genes, 2)

31 18 inhibition of NF-κB signaling, and 3) the release of the transcriptional repressor, B cell lymphoma-6 (BCL-6). Induction of anti-inflammatory genes by PPARβ/δ-specific activators has been reported. In vascular smooth muscle cells (VSMCs) anti-inflammatory TGF-β is a direct PPARβ/δ target gene (83). Several anti-oxidant genes have also been described that contain functional PPREs in their promoters and respond to PPARβ/δ activation, including superoxide dismutase 1 (SOD-1), catalase and thioredoxin (84). The natural inhibitor of IL-1 signaling, IL-1 receptor antagonist (IL-1Ra) is also induced by PPARβ/δ in keratinocytes (85), as well as in VSMCs (86). In cardiomyocytes, TNFα-induced NF-κB luciferase activity was markedly attenuated by co-expression of PPARβ/δ and activation with a specific ligand (81). PPARβ/δ-mediated inhibition of NF-κB activity is believed to be via direct physical interaction between the receptor and the p65 subunit, as demonstrated by coimmunoprecipitation assays (87). This interaction was sufficient to inhibit NF-κB binding to the promoter regions of target genes. Electrophoretic mobility shift assays (EMSA) in H9c2 myotubes further revealed that PPARβ/δ-specific activation significantly decreased LPS-stimulated NF-κB binding activity, with subsequent inhibition of pro-inflammatory gene expression (88). PPARβ/δ inhibition of NF-κB activity has also been described in endothelial cells (89, 90) and adipocytes (91). Although direct protein-protein interaction remains the most widely accepted mechanism of transrepression of NF-κB activity, PPARβ/δ-mediated inhibition of ERK1/2 map kinase phosphorylation with subsequent attenuation of NF-κB was demonstrated in adipocytes (91).

32 19 Perhaps the most interesting manner in which PPARβ/δ negatively regulates the inflammatory response is via its unique association with the transcriptional repressor, BCL-6 (24). PPARβ/δ agonists induce the dissociation of the PPARβ/δ:BCL-6 receptor complex, and the inhibition of inflammation via release of BCL-6 has been reported both in vitro (24) and in vivo (23). In VSMCs, the inhibition of pro-inflammatory adhesion molecules E-selectin, ICAM-1 and VCAM-1 was BCL-6-dependent (84). Direct repression of TNFα-induced VCAM-1 expression was demonstrated by ChIP assays to involve the association of BCL-6 with the VCAM-1 promoter following PPARβ/δ activation. It should be noted, however, that PPARβ/δ-mediated inhibition of these proinflammatory adhesion molecules may also be BCL-6-independent, as GW failed to inhibit the TNFα-induced mrna expression of E-selectin, ICAM-1 and VCAM-1 when a PPARβ/δ, but not control, sirna was used (92). PPARβ/δ activation in the epithelium further reduces the inflammatory burden by inhibiting the major endothelial cell inflammatory responses involved in leukocyte recruitment, as demonstrated in both cell (90) and animal models (92). Moreover, PPARβ/δ activators protect endothelial cells from potentially harmful oxidative stress via direct regulation of α (93). PPARβ/δ is also a master regulator of the inflammatory response in macrophages via its association (pro-inflammatory) and dissociation from (anti-inflammatory) BCL-6 (24, 94). Further information regarding BCL-6 structure and function will be described in detail in Section 1.4. GW treatment reduced the expression of several inflammatory mediators, such as monocyte chemotattractant proteins 1 and 3 (MCP-1, MCP-3), IL-1β, macrophage inflammatory protein 1β (MIP-1β), TNFα and IL-6. To

33 20 date, little is known about the anti-inflammatory role of PPARβ/δ in other inflammatory cells, such as T cells, polymorphonuclear (PMN) leukocytes, or dendritic cells (22). Reports indicate the presence of PPARβ/δ at the mrna level in human CD4 and CD8+ T-cells and CD19+ B cells (95) and in peripheral blood T-cells (96), where PPARβ/δ activation enhances proliferation and inhibits apoptosis Oxidative Stress The PPARs are implicated in the amelioration of oxidative stress (97). PPARβ/δ prevents H2O2-induced apoptosis in H9c2 cardiomyoblasts as demonstrated by cell viability assays and annexin-v staining (98). Catalase, a PPARβ/δ target gene, was responsible for the inhibition of oxidative stress, and the results were PPARβ/δdependent. In a double-blind, randomized study of healthy, moderately overweight subjects, global oxidative stress was measured via urinary concentrations of free 8-iso-prostaglandin-F2α, considered a marker of systemic oxidative stress (99). PPARβ/δ activation by GW501516, and not PPARα activation, reduced by 30% urinary F2- isoprostanes (100). In diabetic male Wistar rats, oxidative stress reduced PPARβ/δ expression (101). Increases in glucose concentrations dose-dependently inhibited the expression of the receptor, indicating that the mechanism of reactive oxygen species (ROS)-mediated cell hypertrophy proceeds via reductions in PPARβ/δ. In another study, rats were fed a liquid diet containing 0% or 37% ethanol, treated with a PPARβ/δ agonist, and then subjected to partial hepatectomy (102). Oxidative stress, measured by immunohistochemical staining and enzyme-linked immunosorbent assay (ELISA) targeting the oxidative stress

34 21 mediators 4-hydroxynonenal (4-HNE) and 8-hydroxy-2'-deoxyguanosine (8-OHdG), was reduced in the livers of GW treated rats compared with control. PPARβ/δ activation also substantially decreased 8-OHdG production during the 24-hour peak regeneration period compared with control and ethanol-treated rats, but failed to completely abolish hepatic DNA damage (102). Thus, PPARβ/δ reduced ethanolmediated hepatic injury, oxidative stress, lipid peroxidation, and insulin resistance while enhancing the regenerative response. In mice, PPARβ/δ activation by specific activators induced heme oxygenase-1 mrna, protein, and enzyme activity (103). Mutations in the heme oxygenase-1 promoter revealed its role as a PPARβ/δ target gene, and constitutively active heme oxygenase-1 further induced PPARβ/δ expression in a positive-feedback loop, while inhibition of PPARβ/δ reduced the protective effect of the receptor in the vascular endothelium. These results are in support of previous studies demonstrating the antiatherogenic properties of PPARβ/δ ligands (23, 94, 104). PPARβ/δ also induces Angiopoietin-like protein 4 (Angptl-4), and this mechanism is responsible for the reduction in fatty-acid-induced oxidative stress in the heart (105) Cell Proliferation and Cancer The role of PPARβ/δ in cell proliferation and carcinogenesis is controversial. For example, there exists evidence for both the potentiation ( ) and attenuation ( ) of colon carcinogenesis. PPARβ/δ is frequently over-expressed in cancers (112), particularly colon cancer, and these findings suggest that the receptor is a downstream target of the APC/β-catenin/Tcf4 pathway. Other studies, however, do not support the

35 22 hypothesis that PPARβ/δ is induced via the APC pathway in colon carcinogenesis, or that the receptor is over-expressed in cancerous tissue. PPARβ/δ mrna is lower in colon carcinomas and adenocarcinomas than in normal tissue in human patients (113). In another study no statistical difference existed between the levels of PPARβ/δ in tumors and normal colon mucosa (114). The influence of PPARβ/δ activation on cancer cell growth is also poorly understood. Honn et al. published some of the first evidence that PPARβ/δ ligands inhibit cancer cell proliferation when the putative PPARβ/δ activator prostacyclin inhibited the growth of the B16 melanoma cell line (115). A report using carbaprostacyclin in conjunction with PPARβ/δ sirna determined that the inhibition of proliferation in a lung cancer cell line was PPARβ/δ-dependent (116). Culturing colonocytes in the presence of the PPARβ/δ specific activator GW inhibited cell proliferation (117). Furthermore, the activation of PPARβ/δ in skin induces target genes that may aid in combating tumor promotion and carcinogenesis via the regulation of ubiquitin C expression (118). The authors conclude, however, that the production of an endogenous PPARβ/δ ligand is necessary for ubiquitin C induction. The reports that PPARβ/δ promotes terminal differentiation and inhibits cell growth are many. The PPARβ/δ null mouse, generated by targeted disruption of the ligand binding domain (119), exhibits enhanced epithelial cell proliferation and susceptibility to skin carcinogenesis, and the use of the PPARβ/δ-specific ligand GW0742 inhibited keratinocyte proliferation and induced the terminal differentiation program (120). Aside from skin cell proliferation, the null mouse model also described roles for the receptor in regulating development, myelination of the corpus callosum, and

36 23 lipid metabolism. In several human cancer cell lines, including the colon cancer lines HT29, HCT116, LS-174T, liver cancer cell lines HepG2, HuH7, and MCF7 (breast) and UACC903 (melanoma), the addition of a PPARβ/δ activator (either GW or GW0742) inhibited cell growth (121, 122). Contradictory evidence also exists for the role of PPARβ/δ in promoting cancer cell survival. Cutler et al. reported that stromal production of prostacyclin confers an anti-apoptotic effect on coloncytes via direct activation of PPARβ/δ (123). As mentioned previously, linoleic acid is an activator of PPARβ/δ, albeit a weak one, and increased intake of linoleic acid increased colon carcinogenesis in experimental animal models (124). GW501516, a much stronger agonist of PPARβ/δ, increased both the size and number of intestinal polyps in APCmin mice (125). In epithelial cells, PPARβ/δ promotes cell proliferation and survival (126), which may account for its increased expression in squamous cell carcinomas (SCC) (127). Tumor angiogenesis is a pivotal step in tumor growth, invasion and metastasis, and, although no role for PPARβ/δ has yet been described, it is noteworthy that microvessel density was significantly higher in tumor cells expressing the receptor (127). Yet another study demonstrated focal areas of accumulation for cyclooxygenase-2 (COX-2), vascular endothelial growth factor (VEGF), and PPARβ/δ in head and neck SCC, suggesting that the receptor is a target for COX-2 metabolites (128). Much of the contradictory evidence surrounding the role of PPARβ/δ in cell proliferation and carcinogenesis might be explained by the variations of genetic mouse models used as well as the type of PPARβ/δ ligand administered (112). It is clear from these reports, however, that our understanding of the complexities of PPARβ/δ signaling

37 24 leaves much to be desired. The mechanism of PPARβ/δ activation in these and other tissues, and their potential therapeutic value in treating and controlling tumor formation and carcinogenesis, is still under critical review.

38 B-Cell Lymphoma-6 (BCL-6) Structure The BCL-6 gene was first discovered by virtue of its frequent mutations (approximately 30-40%) in non-hodgkin lymphoma (129, 130). The BCL-6 gene encodes a 95-kDa nuclear phosphoprotein detectable in most tissues but most strongly in mature B cells (131) during germinal center differentiation (132). The BCL-6 protein structure contains six Kruppel-type C-terminal zinc-finger (ZF) motifs (129), responsible for recognizing consensus DNA sequences in vitro, and an N-terminal POZ domain, which is an autonomous transrepression domain and is highly conserved protein-protein interaction domain (133) (Figure 1-3). The POZ domain of BCL-6 also transrepresses SMRT/mSIN3A/histone deacetylase corepressor complexes (134, 135), and may be involved in hetero- and homodimerization with other POZ-expressing transcriptional regulators (136). Chang et al. demonstrated that BCL-6 could strongly transrepress DNA binding sites regardless of position (3 versus 5 ), orientation and distance from the promoter of target genes. Deletion mutants identified amino acids in the N-terminal domain at two incongruous sites, the POZ domain (amino acids ) and amino acids , as being critical for maximal transrepression by BCL-6. Furthermore, the presence of the six zinc finger motifs suggests that BCL-6 is involved in transcriptional regulation (137). BCL-6 is subject to several forms of post-translational modifications. The middle portion of BCL-6 protein contains several PEST motifs, which are usually found in many rapidly degraded proteins and are proline (P), glutamic acid (E), serine (S), and threonine

39 26 (T) rich (138). In BCL-6, these PEST sequences contain several MAP kinase phosphorylation sites that are required for MAPK-induced, ubiquitin-proteasomemediated phosphorylation and degradation of BCL-6 protein (139, 140). Further negative regulation of BCL-6 protein is effected via p300 (CBP)-mediated acetylation (141). F. 1. BCL6-interacting proteins. The diagram shows the con- Figure 1-3: Domain structure of BCL-6 protein including sites of interaction with various cellular proteins. (Source: Miles, R. et al. Analysis of BCL-6-interacting proteins by Tandem Mass Spectrometry Mol Cell Proteomics 4:1898 (2005))

40 Interaction with PPARβ/δ Only recently has the association of BCL-6 with PPARβ/δ been described (24) (Figure 1-4). Lee et al. observed an inverse correlation in the regulation of the proinflammatory genes MCP-1, MCP-3 and MIP-1β by PPARβ/δ and BCL-6 in macrophages (142). Immunoprecipitation experiments in human embryonic kidney cells (293 cells) demonstrated a protein-protein interaction between epitope-tagged BCL-6 and PPARβ/δ in the absence of ligand. Addition of GW abolished this interaction. Mutant PPARβ/δ lacking AF2 maintained an interaction with BCL-6 regardless of ligand. Glutathione S-transferase (GST) pull-down assays and truncation mutants of BCL-6 indicated that the amino acids (zinc finger domain) were both necessary and sufficient to mediate this interaction, which is unique to the PPARβ/δ subtype (142). The other subtypes, PPARα and PPARγ, were assayed at the same time as PPARβ/δ and showed no affinity for BCL-6.

41 Figure 1-4: PPARβ/δ acts as an inflammatory switch via its association and dissociation with the transcriptional repressor BCL-6. Unliganded PPARβ/δ RXR heterodimers associate with PPREs in the promoters of target genes. Addition of ligand, or genetic deletion of PPARβ/δ, releases BCL-6. BCL-6 then represses inflammatory gene expression. (Source: Barish, G.D. et al. PPAR delta: a dagger in the heart of the metabolic syndrome. J Clin Invest 116:590 (2006)) 28

42 Mechanisms of Transrepression BCL-6 recognizes the following 20-bp DNA sequence in vitro and in vivo: 5 -GAAAATTCCTAGAAAGCATA (143, 144). This consensus motif contains several flanking residues that increase the affinity of BCL-6 for DNA. The N-terminal POZ domain was identified as the transrepression domain, and is position-, orientation- and distance-independent (143). The BCL-6 consensus motif strongly resembles the consensus motif for signal transducers and activators of transcription (STAT) (145), suggesting that BCL-6 mediates the transrepression of these factors. Indeed, BCL-6 inhibits STAT6 in B cells (146). Although the precise mechanisms by which BCL-6 transrepresses target genes remains unclear, STAT6 or BCL-6 consensus motifs were found within the promoter regions of several known BCL-6 target genes (cyclin D2, MIP-1α and MCP-1), suggesting a direct BCL-6:DNA association. ChIP assays have detected the presence of BCL-6 on the VCAM-1 promoter (84). BCL-6 binds to the STAT6 motif with less affinity, and activated STAT6 can replace BCL-6. It is common to find NF-κB consensus motifs located near STAT6 motifs (147), and, indeed, BCL-6 silences NF-κB via direct physical interaction both in vivo and in vitro (148) Involvement in the Immune System Studies in the BCL-6 null mouse have revealed important roles for BCL-6 in the regulation of the immune system. Mice lacking BCL-6 displayed normal B-cell, T-cell and lymphoid-organ development but failed to develop germinal center (GC), potentially owing to a defect in T-cell-dependent antibody responses (149). BCL-6-/- mice also demonstrated an exacerbated systemic inflammation, evidenced by increased eosinophil

43 30 infiltration. At present, the mechanism by which genetic disruption of BCL-6 interferes with GC formation is unclear. In type I TNF receptor-/- or CD19-/- mice, impaired GC formation is a consequence of disturbed lymphoid organ development (150), a response not observed in BCL-6-/- mice. To investigate the mechanism by which BCL-6 deletion disrupts GC formation, RAG1-deficient mice (RAG1-/-) were reconstituted with bone marrow cells from BCL-6-/- mice (the BCL-6-/- mouse displayed growth retardation and mostly died with severe myocardial injury) (151). These reconstitution experiments revealed a deficiency in the B cell development program in BCL-6-/- mice, and suggest a role for BCL-6 in regulating the differentiation of GC B cells. An established role exists for BCL-6 in the regulation of cell survival signaling. Ramos B cells express constitutively active, mutant BCL-6 (BCL6ΔPEST), and are resistant to B cell receptor-induced apoptosis (141). In another study, cytokine-induced cell cycle arrest and apoptosis was inhibited by a mechanism involving BCL-6 repression of STAT-3 signaling (152). Treatment of Burkitt's lymphoma Mutu I cells with Fenretinide (4-(N-hydroxyphenyl) retinamide, 4-HPR), a synthetic retinoid, induced apoptosis and G1/S cell cycle arrest with simultaneous decreases in BCL-6, further suggesting that BCL-6 activity is anti-apoptotic (153). Similar results were obtained in non-immune cell line experiments. When full length BCL-6 cdna was retrovirally transduced into EpH-4 (epithelial) cell line, cells were observed to be resistant to apoptosis as measured by TUNEL assay with prolonged cell cycle (increase of about 3 hours compared to control cells) (154). In addition to a lack of GC formation, BCL-6-/- mice exhibit an increased Th2 immune response, characterized by extensive infiltration of multiple organs by

44 31 eosinophils and immunoglobulin E (IgE)-bearing B cells (149). While the precise mechanism remains unclear, these results suggest a role for BCl-6 in negatively regulating the Th2 immune response pathway. In the immune system, BCL-target genes include those involved in lymphocyte activation and differentiation, Ig subtype switching, inflammation and cell cycle regulation (141) BCL-6 as a Proto-Oncogene Non-Hodgkin lymphoma is a broad set of diseases comprising many different types of cancer of lymphocytes, and may arise from either B- or T-cells. B-cell non- Hodgkin lymphomas include diffuse large B-cell lymphomas (DLBCL). A common chromosomal aberration occurs on the band 3q27 (10-20% of DLBCLs) (155, 156), and molecular cloning studies by several groups identified a bcl-6 gene rearrangement (157). The chromosomal rearrangement affecting the bcl-6 occurs frequently, approximately 30-40%, in DLBCLs (130, 158), and, to a lesser extent, in follicular lymphomas (FL) (approximately 5-10%) (159). Interestingly, bcl-6 remains intact in cancer; the coding region is unchanged, while the 5 non-coding regulatory region is juxtaposed via chromosomal translocation events. The result is a promoter substitution that likely alters the regulatory mechanisms underlying bcl-6 expression (160). Frequent somatic point mutations within the regulatory region of the bcl-6 gene are the most frequent genetic alterations in human B-cell malignancies. Initial in vitro studies using DLBCL-derived BCL-6 alleles conclude that these mutations significantly deregulate bcl-6 expression (141). Despite these in vitro data, however, in vivo results failed to show any association

45 32 between bcl-6 and oncogenesis until several years ago. Using a transgenic mouse model expressing mutant BCL-6 that mimicked the t(3;14)(q27;q32) translocation event observed in DLBCL patients, two independent studies demonstrated a link between BCL- 6 and lymphoma incidence. One study that examined transgenic mice aged months revealed that, depending on the strain, 30 60% of mice developed B-cell lymphomas (161). The second study found that only approximately 13% of transgenic mice developed spontaneous lymphomas. Lymphoma incidence, however, was significantly exacerbated in transgenic mice compared with control mice upon administration of the DNA-alkylating agent and carcinogen N-ethyl-N-nitrosourea (ENU) (162). Although the precise mechanism of BCL-6-mediated lymphomagenesis remains unclear, one recent study identified the tumor suppressor p53 as a BCL-6 target (163). The authors conclude that BCL-6 promotes lymphoma incidence via functional inactivation of p53, conferring resistance to the p53-dependent apoptotic response. The clinical implications of the BCL-6 rearrangement reveal it as an important marker for germinal center-derived Diffuse Large B-Cell lymphoma (GC-derived DLBCL), and two recent studies targeting BCL-6 with specific peptide inhibitors demonstrated success in inducing cell cycle arrest and apoptosis in BCL6-positive lymphoma cells (164, 165). These peptides specifically bind BCL-6 in vivo and inhibit corepressor recruitment, and may prove effective therapeutic agents in combating B-cell lymphomas. Current opinion is divided, however, on the clinical consequences of BCL-6 rearrangement. One hypothesis is that BCL-6 correlates well with favorable clinical prognosis (166), while another study described no relevant clinical differences upon BCL-6 rearrangement (167). The expression pattern of BCL-6 may also be clinically

46 33 relevant. One study that examined BCL-6 expression levels in DLBCL, and not BCL-6 rearrangements, indicated a positive correlation between BCL-6 expression and favorable clinical prognosis (168). In another study that analyzed BCL-6 point mutations in posttransplantation lymphoproliferative disorders, the presence of mutated BCL-6 correlated with poor prognosis (169) Anti-inflammatory Mechanisms of BCL-6 BCL-6 is a master regulator of inflammation, and targeted disruption of BCL-6 results in a highly exacerbated systemic inflammatory response (170). Cytokines are among the various BCL-6-regulated targets. DNA microarray screening experiments in B-cells identified several classes of genes that were repressed by BCL-6 (171), including genes involved in lymphocyte differentiation, cell cycle control and inflammation. The chemokine macrophage inflammatory protein-1α (MIP-1α) was a BCL-6 target, and contains a functional BCL-6 binding element in its promoter. Further analyses in macrophages identified BCL-6 as a key regulator of cytokine expression, and subsequent promoter analyses demonstrated that the pro-inflammatory MCP-1 was repressed by BCL-6 (142). MCP-1 also contains a BCL-6 consensus motif in its promoter, but further assays determined that this site is only responsible for some of the BCL-6-mediated repression of MCP-1 expression (142). This finding is noteworthy in that it may explain how BCL-6 repressed the expression of several genes, notably interferon-inducible protein-10 (IP-10), that do not contain BCL-6 binding sites in their promoter regions. In fact, BCL-6 strongly represses IP-10 expression in transient transfection assays (172).

47 34 BCL-6 regulation of inflammatory signaling is a complicated pathway. MIP-1α, for example, contains a perfect BCL-6 consensus motif in its promoter that binds BCL-6 in transient transfection assays (171). BCL-6-/- macrophages, however, express relatively similar levels of MIP-1α as wild-type macrophages following IL-6 or LPS challenge (142). ChIP assays located BCL-6 on the MIP-1α promoter (172), indicating that the consensus motif is functional. Take together with the previous results, it would seem that BCL-6 repression is reversible under conditions of strong stimulation. Further assays demonstrate that MIP-1α is over-expressed both at the basal and MCP-1- stimulated level in BCL-6-/- macrophages (142). Further complexity is added by the observations regarding BCL-6 regulation of MIP-1α and MCP-1. Both genes contain functional BCL-6 consensus motifs. Although the MIP-1α motif matches the BCL-6 consensus sequence perfectly, it is only repressed by BCL-6 under weakly stimulating conditions. On the other hand, the MCP-1 promoter contains several nucleotide differences when compared with the BCL-6 consensus sequence, yet is strongly repressed under all tested conditions (172). These results clearly indicate that no significant association exists between the promoter binding motif and the amount of repression observed by BCL-6. It is possible to reconcile these observations with the hypothesis that MIP-1α signaling involves the association of various co-factors that override BCL-6 repression, and these co-factors may not interact with the MCP-1 promoter. Another possibility is that BCL-6 negatively regulates MCP-1 via an indirect mechanism. Further regulation of inflammation is effected via direct repression of the master inflammatory regulator, NF-κB. The activated B-cell-like (ABC type) of diffuse large B- cell lymphomas (DLBCLs) are characterized by high NF-κB activity, which is enhanced

48 35 upon BCL-6 silencing (148). Indeed, several well-known NF-κB target genes, including cyclin D1 and TNFα, were significantly increased at the mrna level when BCL-6 was knocked-down via sirna. GST-pulldown assays mapped the BCL-6:NF-κB interaction to the Rel-homology domain (RHD), which interacted with both full-length BCL-6 and the BCL-6 zinc finger truncation (148). These interactions were repeated in vivo, and provide another mechanism for the transcriptional control of pro-inflammatory signaling by BCL-6.

49 Oxidative Stress Oxidative stress results from the production and release of reactive oxygen intermediates, such as superoxide free radicals, hydrogen peroxide, singlet oxygen, nitric oxide (NO), peroxynitrite and 4-hydroxynonenal (4-HNE). The induction of oxidative stress is implicated in the initiation and progression of various diseases, not the least of which being cancer. Reactive oxygen intermediates are constantly present at low levels as a result of normal cellular and physiological conditions, and are usually short-lived due to the scavenging actions of various anti-oxidant defense genes, including superoxide dismutase (SOD), glutathione peroxidase, catalase, and vitamins C and E. Unchecked oxidative stress can result in the destruction of several cell types, including neurons, endothelial cells (ECs), cardiomyocytes, and smooth muscle cells via the induction of apoptosis (173). Oxidative stress is gaining attention as a major initiating event in various disease states, such as diabetes (174), ischemia (175), cognitive loss and Alzheimer s disease (176), pain sensation (177), trauma (178) and cancer (179). The non-enzymatic oxidation of polyunsaturated fatty acids underlies many of the effects of oxidative stress-induced cellular dysfunction. Reactive oxygen species (ROS) initiate the peroxidative degradation of PUFAs via free radical chain reaction, yielding lipid hydroperoxides that ultimately mediate the deleterious effects of oxidative stress (180). Aldehydic end products are one example of oxidative stress mediators. Unlike free radical or reactive oxygen mediators, the aldehydic products of lipid peroxidation are relatively stable and interact with myriad cellular biomolecules, such as DNA, proteins and phospholipids. Moreover, the stability of the aldehydic end products confers an

50 37 ability to diffuse from its cell of origin, allowing them to mediate oxidative stress far from the site of production. Furthermore, oxidative stress mediators affect normal cellular functions even at low levels. The direct modulation of signal transduction, gene expression, cell proliferation, and the oxidative response of the target cell(s) are among the affected pathways PUFA-Peroxidation Products PUFAs, either free or as components of cholesterol esters, phospholipids, and triglycerides, are major targets for reactive oxygen. The free-radical-induced oxidative damage of these molecules results in the production of short-chain aldehydes belonging to one of three families: 2-alkenals, 4-hydroxy-2-alkenals, and ketoaldehydes (181). Examples of these compounds are depicted in Figure alkenals The 2-alkenal oxidative stress mediators are distinguished from the other types on the inclusion of two electrophilic carbon centers in their structures. Cellular macromolecules such as protein and DNA are nucleophilic, and may participate in reactions with 2-alkenals (181). Degradation of PUFAs generates 2-hexenal (180), while cigarette smoke and thermal degradation of industrial chemicals usually produce acrolein and crotonaldehyde (the methyl derivative of Acrolein), two extremely potent electrophiles. Further studies later identified these aldehydes as products of PUFA peroxidation (182, 183).

51 Figure 1-5: Chemical structures of A) reactive aldehydes and B) lipid peroxidationspecific aldehydes. (Source: Uchida K. et al. 4-Hydroxy-2-nonenal: a product and mediator of oxidative stress. Prog Lipid Res 42:318 (2003)) 38

52 hydroxy-2-alkenals The 4-hydroxy-2-alkenals represent the most common PUFA peroxidation products (184). By far, the most well studied 4-hydroxy-2-alkenal is 4-hydroxynonenal (4-HNE), formed from the lipid peroxidation of the ω-6 fatty acids, Linoleic and Arachadonic acid. The same process in ω-3 PUFAs generates 4-hydroxy-2-hexenal. 4- HNE can accumulate at concentrations from 10 µm to 5 mm following the induction of oxidative stress (180), and is implicated in the inhibition of cell growth, changes in glutathione and thiol-containing anti-oxidant molecules, enzyme and protein synthesis inhibition and the exacerbation of neutrophil chemotaxis Ketoaldehydes The final class of lipid peroxidation products is the Ketoaldehydes, of which malondialdehyde (MDA), glyoxal, and 4-oxo-2-nonenal (ONE) are members. MDA is a key marker of oxidative stress, and its detection via 2-thiobarbituric acid is a common assay to determine oxidative burden. ONE is formed via lipid peroxidation reactions on 13-hydroperoxyoctadecadienoic acid (13-HpODE) or via the 4-HNE precursor, 4- hydroperoxy-2e-nonenal (4-HpNE) (185) hydroxynonenal (4-HNE) The precise mechanism of formation of 4-HNE via lipid peroxidation reactions is unclear. The current opinion is that formation of 4-HNE is effected via non-enzymatic Hock cleavage of 13(S)-HpODE between C-9 and C-10 or of 9(S)-HpODE to directly form the racemic precursor, 4-HpNE (186) (Figure 1-6). It is hypothesized that the nine

53 40 carbons comprising the 4-HNE backbone represent the last nine carbons of its ω-6 precursors, and subsequent experimental evidence proposed an alternative enzymatic pathway from PUFA to hydroperoxide to HNE via a key 3Z-alkenal oxygenase-mediated oxygenation (187). Another study described 3Z-nonenal as a molecular substrate for soybean lipoxygenase, which, functioning as an oxygenase, converts 3Z-nonenal to 4- HpNE (188).

54 Figure 1-6: The non-enzymatic formation of the 4-HNE precursor, 4-HpNE, from 13- HODE via a Hock cleavage mechanism. 4-HpNE is further reduced to 4-HNE. 41

55 42 In mammalian cells, 4-HNE is primarily detoxified via glutathione S-transferases (GSTs), aldehyde dehydrogenase, and alcohol dehydrogenase. GSTs effect the Michael conjugation of GSH to 4-HNE; aldehyde dehydrogenase oxidizes 4-HNE to the innocuous 4-hydroxy-2-nonenoic acid; and alcohol dehydrogenase reduces the aldehyde moiety of 4-HNE to the corresponding alcohol (180). Inducible human aldo-keto reductase (AKR1C1) is expressed in response to 4-HNE, and provides a GSHindependent/NADPH-dependent route for cellular detoxification of 4-HNE (189). As previously mentioned, 4-HNE is relatively stable, and is able to interact with a wide range of cellular macromolecules as it passes from one subcellular compartment to another. The deleterious effects of 4-HNE are mediated via covalent Michael addition reactions to specific residues on target proteins, impairing their functions, and DNA, setting the stage for mutation. 4-HNE predominantly covalently modifies cysteine on its sulfhydryl moiety, the imidazole group of histidine (190) and the ε-amino groups of lysine residues (184) on target proteins, which include Na+,K+-ATPase, the neuronal glucose transporter GLUT3, the astrocyte glutamate transporter GLT-1, the GTP-binding protein Gαq/11, tau and the proteosome (180). Polyclonal antibodies directed against HNE-histidine adducts provided the first experimental evidence that 4-HNE covalently modifies proteins in vivo (191). In other animal studies, 4-HNE adducts have been identified in hyperglycemic injury of pancreatic β cells (192), ischemia/reperfusion injury (193), CCL4-induced hepatotoxicity (194) and congestive heart failure (195). Aside from covalent modification of macromolecules, 4-HNE at sublethal concentrations can influence several cell signaling pathways. Cell stress signaling

56 43 pathways are induced by 4-HNE, resulting in the potent expression of c-jun N-terminal kinases (JNK) with subsequent increases in activation of AP-1 (196). Further increases in AP-1 regulated genes, such as adenylate cyclase, epidermal growth factor (EGF) and platelet-derived growth factor (PDGF), are also observed following 4-HNE challenge. 4-HNE also induces the anti-oxidant response in mammalian cells. The exposure of rat liver epithelial cells (RL34) to various aldehydic compounds, including 4-HNE, caused an increase in class π GST (GST-P) transcripts and protein that peaked after 16 hours (197). Several detoxification enzymes contain an anti-oxidant response element (ARE) in their promoter that is activated by 4-HNE (198), inducing the expression of class α GSTs, such as GSTA1 and GSTA4. Subsequent studies identified the transcription factor NF-E2-related factor 2 (Nrf2) are a regulator of anti-oxidant signaling through AREs (199). Nrf2 is sequestered by its regulatory protein Kelch-like ECHassociated protein 1 (Keap1) (200) until electrophilic agents (ie. 4-HNE) interact with Nrf2, inducing its dissociation from Keap1 and activation of the detoxification response. Other pathways affected by 4-HNE have been described: Induction of the inflammatory response via p38 MAP kinase induction of COX-2 expression (201); inhibition of NF-κB-dependent gene expression (202) and proteosome-mediated protein turnover (203); and the apoptosis pathway via the caspase cascade and subsequent cytochrome c release (204).

57 Pancreatic Cancer General Remarks Pancreatic cancer is one of the most lethal of all human cancers, accounting for approximately 213,000 deaths worldwide and is the fourth-leading cause of cancerrelated deaths in the United States (205). Survival rates are poor: the overall median survival rate is a mere six months and the 5-year survival rate compared across all stages of the disease is 5%. The survival rates for patients presenting with advanced cases of pancreatic cancer is <2%. The significant amount of research performed to elucidate the causes of pancreatic cancer has produced a wealth of knowledge about the underlying etiological and molecular causes of the disease, but this has not necessarily translated into better treatment or improved patient survival Molecular Genetics of Pancreatic Cancer K-ras Point mutations that activate the oncogene K-ras are present in 80-90% of all pancreatic cancers (206). These mutations, frequently occurring at codon 12, 13 or 61, abolish the GTPase activity of K-ras, resulting in its constitutive activation. Mutations of K-ras are noteworthy in that they may play a role in the early stages of pancreatic cancer formation (207). In fact, point mutations of K-ras are present in approximately 30% of pancreatic intraepithelial neoplasia (PANIn) lesions (208). Wild-type K-ras is present in pancreatic cancers, although this is very rare, and usually accompanied by a mutation in

58 45 the BRAF oncogene (209). K-ras and BRAF are active in the same Ras/Raf/MAP kinase pathway, and are clearly important in the generation and progression of pancreatic cancer. Several studies have attempted to link mutant K-ras to clinical outcome, albeit with little success. Some studies have reported an association with the detection of mutant K-ras and poor outcome, while others found no significant association (207) Oncogenes Over-expression of oncogenes is also reported in pancreatic cancer. Amplification of the pro-survival C-MYC gene is detected in approximately 50-60% of pancreatic cancers (210). Given the recent advent of small-molecule inhibitors of the MYC pathway, C-MYC may prove a useful therapeutic target in future studies (211). Akt2, a serine-threonine kinase implicated in phosphatidylinositol-3-oh kinase signaling, is rarely over-expressed (5%), but Akt2 signaling is activated in 30-40% of pancreatic cancers (212, 213). Like C-MYC, inhibitors of Akt2 that induce cell cycle arrest or apoptosis may prove useful in the treatment of pancreatic cancer (214). Constitutively active EGFR has also been described (215), and selective EGFR inhibitors have had success in suppression pancreatic cancer cell growth and invasion (216) Tumor Suppressor Genes The tumor suppressor gene p16 (cyclin-dependent kinase CDKN2A/p16) inhibits cell cycle progression at the G1-S checkpoint, and represents the most commonly mutated gene in pancreatic cancer (90%) (217). Loss of p16 is frequently observed in

59 46 pancreatic intraepithelial neoplasias (PANIns) (30% in PanIN-1, 55% of PanIN-2, and 71% of PanIN-3 lesions) (218). P53, a transcription factor that regulates the expression of genes involved in apoptosis, angiogenesis, cell cycle and genome maintenance, is also inactivated in many pancreatic cancers (approximately 50-75%) via mutations on one allele and the loss of the other (207). A key function of p53 is the induction of apoptosis in response to DNA damage. Thus, its inactivation in pancreatic cancer represents a mechanism by which additional genetic damage may accumulate. Deleted in Pancreatic Carcinoma 4 (DPC4) is a candidate tumor suppressor gene that is also inactivated in about 55% of pancreatic cancers (219), but only rarely in other malignancies (220). The consequences of inactive DPC4 are increased proliferation via interferences in cell signaling downstream of cell surface receptors, like those of the TGF-β family Animal Models The study of pancreatic cancer and the search for therapeutic agents has been difficult, owing to the lack of suitable animal models. Mouse xenograft models using athymic nude mice (SCID) are typically used to screen potential drugs since the pancreatic cancer cells display similar genotypic features of the tumor and similar resistances to chemotherapies (221). This model, however, lacks B- or T-cell responses, and therefore makes the examination of immune involvement difficult. Furthermore, as the cells are transplanted subcutaneously, the resulting tumor does not resemble the physiological microenvironment in the pancreas, nor does it metastasize (222).

60 47 The implantation of pancreatic tumor cells or primary pancreatic tissue directly into the mouse pancreas in orthotopic models has seen increasing use. Reflecting the wide range of pathological features of pancreatic cancer, these models are exceptionally useful in the study of drug effects on the tumor microenvironment and metastatic spread (223). Several transgenic mouse models have also been generated that closely mimic human pancreatic cancers. Many of these models use activated K-ras directed to the mouse pancreas via Cre recombinase with subsequent inhibition of either the INK4A/Arf or p53 pathways (224, 225) Stages Stages of pancreatic cancer are determined by the American Joint Commission on Cancer (AJCC) using a tumor-node-mestastasis (TNM) staging system first developed in 1959 and updated every few years. Pancreatic cancer is divided into four stages, each with a unique history and prognosis. The current staging system takes the following into consideration: the size and extent of the primary tumor (T), the presence or absence of regional lymph node metastases (N) and the presence or absence of distance metastases (M) (226). In the most recent AJCC staging guidelines, T1 and T2 type cancers are described as those types being confined to the pancreas, whereas T3 tumors are distinguished from T4 tumors on the basis of resectability. T4 type tumors are not resectable, owing to their advancement to the celiac axis (CA) or superior mesenteric artery (SMA) (226). T3 tumors that do not reach these two points, but may still be extrapancreatic in nature, may still be resected and thus differ from T4 tumors. The

61 48 absence of regional lymph node and/or distant metastases is indicated by N0 and M0, respectively. Conversely, the presence of regional lymph node and/or distant metastases is indicated by N1 and M1. The combinations of these descriptors give rise to the staging system for pancreatic cancer, reflecting the relative differences in survival duration and respectability. Stage I and II pancreatic cancers are resectable; Stage III and Stage IV are not. More precisely, Stage I pancreatic cancers are potentially resectable tumors that are confined to the pancreas itself. Stage II cancers are still potentially resectable, but may involve veinous structures, nearby organs and/or N1. Stage III pancreatic cancers are locally advanced but are unresectable due to CA or SMA involvement. Stage IV pancreatic cancers are unresectable due to M1. Together, patients with Stage III (13%) and Stage IV (55%) account for the majority of newly diagnosed cancers of the pancreas (227) Descriptive Epidemiology Frequency Pancreatic cancer is frequently diagnosed in both the United States and Japan, with approximately 30,000 and 19,700 new cases arising each year, respectively (228). Pancreatic cancer is the fourth leading cause of cancer-related deaths in males, behind lung, prostate and colo-rectum, and is also the fourth leading cause in women, behind lung, breast and colorectum. Males overwhelmingly develop pancreatic cancer in both the United States (40% higher than in females) and in Japan (70% higher), potentially due to lifestyle and occupational hazards (229). The lifetime risk of pancreatic cancer in

62 49 developed countries is approximately 1% (230). In North America, the relative frequency of pancreatic cancer has remained constant, although the number of cases in Europe has risen steadily for several decades (231). Incidence of pancreatic cancer is higher among single individuals compared to married persons (232), as well as among individuals in lower socio-economic class compared to those in higher classes (233) Age In the United States, only 13% of all cases of pancreatic cancer are diagnosed in individuals under the age of 60, usually owing to a positive family history or genetic aberration, indicating that the disease primarily affects the elderly. Indeed, approximately 50% of individuals are at or over the age of 75 at the time of diagnosis (228) Race African Americans are at higher risk for the development of pancreatic cancer than are Caucasians, despite the similarities in survival rates among the two groups (234). PCR amplification studies determined that African Americans have a higher incidence of K-ras mutations to valine (58%) than Caucasians (22%). A large-scale epidemiological study by Longnecker et al. examined over 10,000 individuals diagnosed with pancreatic cancer as reported to USA SEER cancer registry (235), and determined that Asian patients developed less aggressive tumors than either Caucasians or African Americans and had better survival rates.

63 Etiology A number of risk factors for pancreatic cancer that contribute to its etiology have been described Diabetes There is conflicting evidence pertaining to the contribution of diabetes to pancreatic cancer etiology (236, 237). In a case-control study among French-Canadians in Montreal, the authors determined that long-term diabetes mellitus increases pancreatic cancer risk (236). A recent study, however, claims that diabetes mellitus in patients with pancreatic cancer is a result of the tumor itself, considering the fact that if patients with diabetes of three or more years' duration were considered, the association between pancreatic cancer and diabetes mellitus was not significant (237). In one recent large study following 20,475 men and 15,183 women, the authors conclude that factors associated with abnormal glucose metabolism potentially contribute to pancreatic cancer etiology (238). Diabetes is the manifestation of pancreatic dysfunction, and may play a role in the development of the disease Pancreatitis A history of pancreatitis suggests an increased risk for developing pancreatic cancer, according to several studies. One large case-control study, using multivariate analysis, determined that all types of pancreatitis or chronic pancreatitis significantly elevated the risk for pancreatic cancer (239). A similar study found that the risk of developing cancer from chronic pancreatitis over a 25-year period was around 4% (240).

64 51 Hereditary pancreatitis also plays a role in pancreatic cancer etiology, and is believed to be responsible for approximately 3-6% of all pancreatitis cases. In a study of 21 kindreds, 20% developed pancreatic cancer (241). Another study of 246 individuals with hereditary pancreatitis concluded that the risk of developing cancer by age 70 is 40%. When pancreatitis is inherited via paternal transmission, the risk increases to 75% (242) Tobacco and N-Nitroso Compounds Smoking and tobacco usage is the single established risk factor for pancreatic cancer. Respiratory or gastrointestinal intake of N-Nitroso compounds exposes the pancreatic ductal epithelium to carcinogens via circulation (243). These carcinogens induce DNA strand breaks and adduct formation, overwhelm DNA repair capabilities, and induce tumor development (244). N-Nitroso compounds, such as N-Nitrosodimethylamine (NDMA) and Nicotine-derived nitrosamine ketone (NNK), are ubiquitously present in the environment, but humans gain increased exposure to them through tobacco, dietary and occupational sources. These compounds specifically induce pancreatic cancer in animal models (245), and nearly all studies report that the exposure to tobacco products increases the risk of pancreatic cancer by about 2-fold (246, 247), with increasing risk in relation to increases in cigarettes smoked (248). These studies also highlight an important fact: For the heaviest of smokers, risk levels return to baseline only after a period of 15 years of non-smoking, as compared to 5 years for light smokers.

65 Alcohol and Coffee Chronic alcohol consumption has been a proposed risk factor for pancreatic cancer since the 1960s. One study of chronic alcoholics supported the hypothesis that alcohol plays a role in pancreatic cancer etiology (249), although several subsequent studies have failed to identify a relationship between alcohol consumption and increased pancreatic cancer risk (250). Likewise, early evidence hinted at coffee consumption as a potential pancreatic cancer risk factor (251), but recent studies have also failed to establish a clear link between the two (252, 253). In both frequent alcohol and coffee drinkers, the risk for pancreatic cancer is lower than in non-drinkers (252) Dietary and Nutritional Factors The relationship between diet and pancreatic cancer risk is difficult to establish, particularly because of the rapid onset of the disease and its high mortality rate. Several case-control studies suggest a positive link between increased cholesterol intake and higher pancreatic cancer risk (254). Other studies have linked intake of grilled red meat or butter and saturated fats with increased risk, but subsequent studies have failed to repeat these observations (255). Case studies have also attempted to link increased caloric intake to pancreatic cancer (256). These observations are also inconsistent, as a high calorie diet was shown to be protective against pancreatic cancer in male smokers (257). As for prevention, regular physical exercise, a regular bowel habit and regular consumption of raw vegetables was protective against pancreatic cancer (258). Finally, most studies link obesity and low physical activity to increased risk for developing

66 pancreatic cancer, suggesting that risk increases with increases in body mass index (BMI) (259, 260) Treatment Options Resection is the only treatment option that has the potential to cure the patient, provided that the disease is diagnosed in its early stages. Only 10-20% of patients, however, are candidates for resection (264). A substantial number of patients, approximately 30-40%, present with locally advanced disease, and resection is a potential treatment option if chemotherapies are successful in maintaining or downstaging the disease (265). These so-called neoadjuvant therapies are aimed at converting unresectable pancreatic tumors to resectable, however no Phase III clinical trial data regarding the efficacy of these treatments is available at this time. Unfortunately, resection also has its drawbacks. Approximately 80% of individuals who undergo resection die from the disease anyway, owing to local recurrence (usually due to incomplete resection (266)) and/or distant metastases (267). The current standard for patients receiving resection is to administer post-operative adjuvant chemotherapy (ie. Gemcitabine or 5-FU), and this is supported by data from several Phase III clinical trials (268). Despite these problems, however, resection is still the primary treatment option for patients presenting with Stage I or II pancreatic cancers. Since the majority of patients (80%) present with surgically inoperable (not resectable), locally advanced or metastatic disease, treatment for pancreatic cancer is mostly aimed at improving the quality of life. Chemotherapy in the form of Gemcitabine has been the primary standard since 1997, when Burris et al. demonstrated that

67 54 Gemcitabine increased median survival to 5.65 months compared to 4.41 when 5- fluorouracil (5-FU) was used (261). More importantly, 1-year survival rates increased from 2% to 18%, essentially establishing Gemcitabine as the first line defense against pancreatic cancer. Recent studies have attempted to improve upon the success of Gemcitabine by combining it with various cytotoxic or novel therapeutic agents. In 2007 a randomized Phase III clinical trial showed that the addition of Erlotinib (a human epidermal growth factor receptor type 1-targeting agent) increased median survival as well as 1-year survival compared with Gemcitabine alone in patients with advanced, unresectable pancreatic cancers (262). Another Phase III trial examined the effects of combining Bevacizumab to the Gemcitabine/Erlotinib combination and found no statistically significant increase in overall survival but did find that progression-free survival rates increased (263). Recently, Erlotinib has been approved by the FDA for use with Gemcitibine in the management of advanced pancreatic cancer, and remains the only FDA-approved EGFR-targeting therapy for this disease to date. Chemoradiotherapy and systemic chemotherapies are the current standard for patients with locally advanced or metastatic disease (Stages III and IV), and, although they rarely down-stage the disease, are usually successful in prolonging survival and decreasing pain.

68 Conclusions The PPARs are involved in the regulation of multiple cellular pathways. PPARβ/δ, in particular, is a master regulator of fatty acid oxidation, inflammation, cell growth and cell invasion, and its activation could have profound impacts on the control and management of several disorders including diabetes, inflammatory diseases and cancer. Our current understanding of this PPAR isotype, however, is poor. At present, no truly selective endogenous ligand has been identified. Furthermore, there is considerable controversy regarding the effect(s) of PPARβ/δ activation on the growth and proliferation of cells both in vitro and in vivo. Studies presented in this dissertation first define the lipid peroxidation product 4-hydroxy-2-nonenal (4-HNE) to be both present in oxidized lipids and a PPARβ/δ activator that induces the anti-oxidant response in a PPARβ/δ-dependent manner. These results indicated a role for the receptor in cellular defense against inflammation and oxidative stress-induced cell damage, an observation that is in agreement with a recently discovered role for PPARβ/δ in regulating inflammation. Second, the hypothesis that PPARβ/δ activation is antiinflammatory is further examined in the context of pancreatic cancer cells. Finally, the role of PPARβ/δ in cell migration is further examined in pancreatic cancer cells in an attempt to understand if receptor activation could prove an effective therapeutic target in the control and maintenance of this disorder.

69 Figure 1-7: Progression of pancreatic cancer from normal pancreatic tissue to malignant neoplasia. Our current observations suggest that PPARβ/δ activation limits oxidative stress and the release of inflammatory cytokines, indicating a role for the receptor in preventing pancreatic cancer at multiple stages of the disease. 56

70 References 1. Zhu Y, Alvares K, Huang Q, Rao MS, Reddy JK. Cloning of a new member of the peroxisome proliferator-activated receptor gene family from mouse liver. J Biol Chem 1993;268: Desvergne B, Wahli W. Peroxisome proliferator-activated receptors: nuclear control of metabolism. Endocr Rev 1999;20: Xing G, Zhang L, Heynen T, et al. Rat PPAR delta contains a CGG triplet repeat and is prominently expressed in the thalamic nuclei. Biochem Biophys Res Commun 1995;217: Issemann I, Green S. Activation of a member of the steroid hormone receptor superfamily by peroxisome proliferators. Nature 1990;347: Chen F, Law SW, O'Malley BW. Identification of two mppar related receptors and evidence for the existence of five subfamily members. Biochem Biophys Res Commun 1993;196: Kliewer SA, Forman BM, Blumberg B, et al. Differential expression and activation of a family of murine peroxisome proliferator-activated receptors. Proc Natl Acad Sci U S A 1994;91: Aperlo C, Pognonec P, Saladin R, Auwerx J, Boulukos KE. cdna cloning and characterization of the transcriptional activities of the hamster peroxisome proliferator-activated receptor happar gamma. Gene 1995;162: Schmidt A, Endo N, Rutledge SJ, Vogel R, Shinar D, Rodan GA. Identification of a new member of the steroid hormone receptor superfamily that is activated by a peroxisome proliferator and fatty acids. Mol Endocrinol 1992;6: Sher T, Yi HF, McBride OW, Gonzalez FJ. cdna cloning, chromosomal mapping, and functional characterization of the human peroxisome proliferator activated receptor. Biochemistry 1993;32: Greene ME, Blumberg B, McBride OW, et al. Isolation of the human peroxisome proliferator activated receptor gamma cdna: expression in hematopoietic cells and chromosomal mapping. Gene Expr 1995;4: Dreyer C, Krey G, Keller H, Givel F, Helftenbein G, Wahli W. Control of the peroxisomal beta-oxidation pathway by a novel family of nuclear hormone receptors. Cell 1992;68: Mukherjee R, Jow L, Croston GE, Paterniti JR, Jr. Identification, characterization, and tissue distribution of human peroxisome proliferator-activated receptor (PPAR) isoforms PPARgamma2 versus PPARgamma1 and activation with retinoid X receptor agonists and antagonists. J Biol Chem 1997;272: Auboeuf D, Rieusset J, Fajas L, et al. Tissue distribution and quantification of the expression of mrnas of peroxisome proliferator-activated receptors and liver X receptor-alpha in humans: no alteration in adipose tissue of obese and NIDDM patients. Diabetes 1997;46:

71 14. Rieusset J, Andreelli F, Auboeuf D, et al. Insulin acutely regulates the expression of the peroxisome proliferator-activated receptor-gamma in human adipocytes. Diabetes 1999;48: Kota BP, Huang TH, Roufogalis BD. An overview on biological mechanisms of PPARs. Pharmacol Res 2005;51: Rosenstock J, Einhorn D, Hershon K, Glazer NB, Yu S. Efficacy and safety of pioglitazone in type 2 diabetes: a randomised, placebo-controlled study in patients receiving stable insulin therapy. Int J Clin Pract 2002;56: Ovalle F, Bell DS. Clinical evidence of thiazolidinedione-induced improvement of pancreatic beta-cell function in patients with type 2 diabetes mellitus. Diabetes Obes Metab 2002;4: Moraes LA, Piqueras L, Bishop-Bailey D. Peroxisome proliferator-activated receptors and inflammation. Pharmacol Ther 2006;110: Khan SA, Vanden Heuvel JP. Role of nuclear receptors in the regulation of gene expression by dietary fatty acids (review). J Nutr Biochem 2003;14: Bieri F, Bentley P, Waechter F, Staubli W. Use of primary cultures of adult rat hepatocytes to investigate mechanisms of action of nafenopin, a hepatocarcinogenic peroxisome proliferator. Carcinogenesis 1984;5: Staels B, Koenig W, Habib A, et al. Activation of human aortic smooth-muscle cells is inhibited by PPARalpha but not by PPARgamma activators. Nature 1998;393: Bishop-Bailey D, Bystrom J. Emerging roles of peroxisome proliferator-activated receptor-beta/delta in inflammation. Pharmacol Ther 2009;124: Takata Y, Liu J, Yin F, et al. PPARdelta-mediated antiinflammatory mechanisms inhibit angiotensin II-accelerated atherosclerosis. Proc Natl Acad Sci U S A 2008;105: Lee CH, Chawla A, Urbiztondo N, et al. Transcriptional repression of atherogenic inflammation: modulation by PPARdelta. Science 2003;302: Burdick AD, Bility MT, Girroir EE, et al. Ligand activation of peroxisome proliferator-activated receptor-beta/delta(pparbeta/delta) inhibits cell growth of human N/TERT-1 keratinocytes. Cell Signal 2007;19: Burdick AD, Kim DJ, Peraza MA, Gonzalez FJ, Peters JM. The role of peroxisome proliferator-activated receptor-beta/delta in epithelial cell growth and differentiation. Cell Signal 2006;18: Ham SA, Kim HJ, Kang ES, et al. PPARdelta promotes wound healing by upregulating TGF-beta1-dependent or -independent expression of extracellular matrix proteins. J Cell Mol Med Takayama O, Yamamoto H, Damdinsuren B, et al. Expression of PPARdelta in multistage carcinogenesis of the colorectum: implications of malignant cancer morphology. Br J Cancer 2006;95: Peters JM, Gonzalez FJ. Sorting out the functional role(s) of peroxisome proliferator-activated receptor-beta/delta (PPARbeta/delta) in cell proliferation and cancer. Biochim Biophys Acta 2009;1796: Hihi AK, Michalik L, Wahli W. PPARs: transcriptional effectors of fatty acids and their derivatives. Cell Mol Life Sci 2002;59:

72 31. Schwabe JW, Neuhaus D, Rhodes D. Solution structure of the DNA-binding domain of the oestrogen receptor. Nature 1990;348: Juge-Aubry CE, Hammar E, Siegrist-Kaiser C, et al. Regulation of the transcriptional activity of the peroxisome proliferator-activated receptor alpha by phosphorylation of a ligand-independent trans-activating domain. J Biol Chem 1999;274: Hu E, Kim JB, Sarraf P, Spiegelman BM. Inhibition of adipogenesis through MAP kinase-mediated phosphorylation of PPARgamma. Science 1996;274: Escher P, Wahli W. Peroxisome proliferator-activated receptors: insight into multiple cellular functions. Mutat Res 2000;448: Glass CK. Differential recognition of target genes by nuclear receptor monomers, dimers, and heterodimers. Endocr Rev 1994;15: Onate SA, Tsai SY, Tsai MJ, O'Malley BW. Sequence and characterization of a coactivator for the steroid hormone receptor superfamily. Science 1995;270: Chakravarti D, LaMorte VJ, Nelson MC, et al. Role of CBP/P300 in nuclear receptor signalling. Nature 1996;383: Gearing KL, Gottlicher M, Teboul M, Widmark E, Gustafsson JA. Interaction of the peroxisome-proliferator-activated receptor and retinoid X receptor. Proc Natl Acad Sci U S A 1993;90: Keller H, Dreyer C, Medin J, Mahfoudi A, Ozato K, Wahli W. Fatty acids and retinoids control lipid metabolism through activation of peroxisome proliferatoractivated receptor-retinoid X receptor heterodimers. Proc Natl Acad Sci U S A 1993;90: Wagner RL, Apriletti JW, McGrath ME, West BL, Baxter JD, Fletterick RJ. A structural role for hormone in the thyroid hormone receptor. Nature 1995;378: Tugwood JD, Issemann I, Anderson RG, Bundell KR, McPheat WL, Green S. The mouse peroxisome proliferator activated receptor recognizes a response element in the 5' flanking sequence of the rat acyl CoA oxidase gene. EMBO J 1992;11: Palmer CN, Hsu MH, Griffin HJ, Johnson EF. Novel sequence determinants in peroxisome proliferator signaling. J Biol Chem 1995;270: A IJ, Jeannin E, Wahli W, Desvergne B. Polarity and specific sequence requirements of peroxisome proliferator-activated receptor (PPAR)/retinoid X receptor heterodimer binding to DNA. A functional analysis of the malic enzyme gene PPAR response element. J Biol Chem 1997;272: Juge-Aubry C, Pernin A, Favez T, et al. DNA binding properties of peroxisome proliferator-activated receptor subtypes on various natural peroxisome proliferator response elements. Importance of the 5'-flanking region. J Biol Chem 1997;272: Hsu MH, Palmer CN, Song W, Griffin KJ, Johnson EF. A carboxyl-terminal extension of the zinc finger domain contributes to the specificity and polarity of 59

73 peroxisome proliferator-activated receptor DNA binding. J Biol Chem 1998;273: Kurokawa R, Yu VC, Naar A, et al. Differential orientations of the DNA-binding domain and carboxy-terminal dimerization interface regulate binding site selection by nuclear receptor heterodimers. Genes Dev 1993;7: Mader S, Chen JY, Chen Z, White J, Chambon P, Gronemeyer H. The patterns of binding of RAR, RXR and TR homo- and heterodimers to direct repeats are dictated by the binding specificites of the DNA binding domains. EMBO J 1993;12: Zechel C, Shen XQ, Chambon P, Gronemeyer H. Dimerization interfaces formed between the DNA binding domains determine the cooperative binding of RXR/RAR and RXR/TR heterodimers to DR5 and DR4 elements. EMBO J 1994;13: Keller H, Givel F, Perroud M, Wahli W. Signaling cross-talk between peroxisome proliferator-activated receptor/retinoid X receptor and estrogen receptor through estrogen response elements. Mol Endocrinol 1995;9: Chu R, Madison LD, Lin Y, et al. Thyroid hormone (T3) inhibits ciprofibrateinduced transcription of genes encoding beta-oxidation enzymes: cross talk between peroxisome proliferator and T3 signaling pathways. Proc Natl Acad Sci U S A 1995;92: Wahli W, Devchand PR, A IJ, Desvergne B. Fatty acids, eicosanoids, and hypolipidemic agents regulate gene expression through direct binding to peroxisome proliferator-activated receptors. Adv Exp Med Biol 1999;447: Kliewer SA, Sundseth SS, Jones SA, et al. Fatty acids and eicosanoids regulate gene expression through direct interactions with peroxisome proliferator-activated receptors alpha and gamma. Proc Natl Acad Sci U S A 1997;94: Forman BM, Chen J, Evans RM. Hypolipidemic drugs, polyunsaturated fatty acids, and eicosanoids are ligands for peroxisome proliferator-activated receptors alpha and delta. Proc Natl Acad Sci U S A 1997;94: Xu HE, Lambert MH, Montana VG, et al. Molecular recognition of fatty acids by peroxisome proliferator-activated receptors. Mol Cell 1999;3: Berger J, Leibowitz MD, Doebber TW, et al. Novel peroxisome proliferatoractivated receptor (PPAR) gamma and PPARdelta ligands produce distinct biological effects. J Biol Chem 1999;274: Brown PJ, Smith-Oliver TA, Charifson PS, et al. Identification of peroxisome proliferator-activated receptor ligands from a biased chemical library. Chem Biol 1997;4: Weigand S, Bischoff H, Dittrich-Wengenroth E, et al. Minor structural modifications convert a selective PPARalpha agonist into a potent, highly selective PPARdelta agonist. Bioorg Med Chem Lett 2005;15: Epple R, Azimioara M, Russo R, et al. 1,3,5-Trisubstituted aryls as highly selective PPARdelta agonists. Bioorg Med Chem Lett 2006;16:

74 59. Havranek M, Sauerberg P, Mogensen JP, et al. Novel selective PPARdelta agonists: optimization of activity by modification of alkynylallylic moiety. Bioorg Med Chem Lett 2007;17: Zhang R, Deangelis A, Wang A, et al. Discovery and SAR of paraalkylthiophenoxyacetic acids as potent and selective PPARdelta agonists. Bioorg Med Chem Lett 2009;19: Sznaidman ML, Haffner CD, Maloney PR, et al. Novel selective small molecule agonists for peroxisome proliferator-activated receptor delta (PPARdelta)-- synthesis and biological activity. Bioorg Med Chem Lett 2003;13: Skogsberg J, Kannisto K, Roshani L, et al. Characterization of the human peroxisome proliferator activated receptor delta gene and its expression. Int J Mol Med 2000;6: Lundell K, Thulin P, Hamsten A, Ehrenborg E. Alternative splicing of human peroxisome proliferator-activated receptor delta (PPAR delta): effects on translation efficiency and trans-activation ability. BMC Mol Biol 2007;8: Tan NS, Michalik L, Di-Poi N, et al. Essential role of Smad3 in the inhibition of inflammation-induced PPARbeta/delta expression. EMBO J 2004;23: Larsen LK, Amri EZ, Mandrup S, Pacot C, Kristiansen K. Genomic organization of the mouse peroxisome proliferator-activated receptor beta/delta gene: alternative promoter usage and splicing yield transcripts exhibiting differential translational efficiency. Biochem J 2002;366: Barish GD, Narkar VA, Evans RM. PPAR delta: a dagger in the heart of the metabolic syndrome. J Clin Invest 2006;116: Tan NS, Michalik L, Di-Poi N, Desvergne B, Wahli W. Critical roles of the nuclear receptor PPARbeta (peroxisome-proliferator-activated receptor beta) in skin wound healing. Biochem Soc Trans 2004;32: Rowan BG, Weigel NL, O'Malley BW. Phosphorylation of steroid receptor coactivator-1. Identification of the phosphorylation sites and phosphorylation through the mitogen-activated protein kinase pathway. J Biol Chem 2000;275: Lazennec G, Canaple L, Saugy D, Wahli W. Activation of peroxisome proliferator-activated receptors (PPARs) by their ligands and protein kinase A activators. Mol Endocrinol 2000;14: Hansen JB, Zhang H, Rasmussen TH, Petersen RK, Flindt EN, Kristiansen K. Peroxisome proliferator-activated receptor delta (PPARdelta )-mediated regulation of preadipocyte proliferation and gene expression is dependent on camp signaling. J Biol Chem 2001;276: Oliver WR, Jr., Shenk JL, Snaith MR, et al. A selective peroxisome proliferatoractivated receptor delta agonist promotes reverse cholesterol transport. Proc Natl Acad Sci U S A 2001;98: Leibowitz MD, Fievet C, Hennuyer N, et al. Activation of PPARdelta alters lipid metabolism in db/db mice. FEBS Lett 2000;473: van der Veen JN, Kruit JK, Havinga R, et al. Reduced cholesterol absorption upon PPARdelta activation coincides with decreased intestinal expression of NPC1L1. J Lipid Res 2005;46:

75 74. Kershaw EE, Flier JS. Adipose tissue as an endocrine organ. J Clin Endocrinol Metab 2004;89: Carr DB, Utzschneider KM, Hull RL, et al. Intra-abdominal fat is a major determinant of the National Cholesterol Education Program Adult Treatment Panel III criteria for the metabolic syndrome. Diabetes 2004;53: Wang YX, Lee CH, Tiep S, et al. Peroxisome-proliferator-activated receptor delta activates fat metabolism to prevent obesity. Cell 2003;113: Braissant O, Foufelle F, Scotto C, Dauca M, Wahli W. Differential expression of peroxisome proliferator-activated receptors (PPARs): tissue distribution of PPARalpha, -beta, and -gamma in the adult rat. Endocrinology 1996;137: Tanaka T, Yamamoto J, Iwasaki S, et al. Activation of peroxisome proliferatoractivated receptor delta induces fatty acid beta-oxidation in skeletal muscle and attenuates metabolic syndrome. Proc Natl Acad Sci U S A 2003;100: Roberts LD, Hassall DG, Winegar DA, Haselden JN, Nicholls AW, Griffin JL. Increased hepatic oxidative metabolism distinguishes the action of Peroxisome proliferator-activated receptor delta from Peroxisome proliferator-activated receptor gamma in the ob/ob mouse. Genome Med 2009;1: Bright JJ, Kanakasabai S, Chearwae W, Chakraborty S. PPAR Regulation of Inflammatory Signaling in CNS Diseases. PPAR Res 2008;2008: Smeets PJ, Teunissen BE, Planavila A, et al. Inflammatory pathways are activated during cardiomyocyte hypertrophy and attenuated by peroxisome proliferatoractivated receptors PPARalpha and PPARdelta. J Biol Chem 2008;283: Park SW, Yi JH, Miranpuri G, et al. Thiazolidinedione class of peroxisome proliferator-activated receptor gamma agonists prevents neuronal damage, motor dysfunction, myelin loss, neuropathic pain, and inflammation after spinal cord injury in adult rats. J Pharmacol Exp Ther 2007;320: Kim HJ, Ham SA, Kim SU, et al. Transforming growth factor-beta1 is a molecular target for the peroxisome proliferator-activated receptor delta. Circ Res 2008;102: Fan Y, Wang Y, Tang Z, et al. Suppression of pro-inflammatory adhesion molecules by PPAR-delta in human vascular endothelial cells. Arterioscler Thromb Vasc Biol 2008;28: Chong HC, Tan MJ, Philippe V, et al. Regulation of epithelial-mesenchymal IL-1 signaling by PPARbeta/delta is essential for skin homeostasis and wound healing. J Cell Biol 2009;184: Kim HJ, Kim MY, Hwang JS, et al. PPARdelta inhibits IL-1beta-stimulated proliferation and migration of vascular smooth muscle cells via up-regulation of IL-1Ra. Cell Mol Life Sci. 87. Smeets PJ, Planavila A, van der Vusse GJ, van Bilsen M. Peroxisome proliferator-activated receptors and inflammation: take it to heart. Acta Physiol (Oxf) 2007;191: Planavila A, Rodriguez-Calvo R, Jove M, et al. Peroxisome proliferator-activated receptor beta/delta activation inhibits hypertrophy in neonatal rat cardiomyocytes. Cardiovasc Res 2005;65:

76 89. Liang YJ, Liu YC, Chen CY, et al. Comparison of PPARdelta and PPARgamma in inhibiting the pro-inflammatory effects of C-reactive protein in endothelial cells. Int J Cardiol Rival Y, Beneteau N, Taillandier T, et al. PPARalpha and PPARdelta activators inhibit cytokine-induced nuclear translocation of NF-kappaB and expression of VCAM-1 in EAhy926 endothelial cells. Eur J Pharmacol 2002;435: Rodriguez-Calvo R, Serrano L, Coll T, et al. Activation of peroxisome proliferator-activated receptor beta/delta inhibits lipopolysaccharide-induced cytokine production in adipocytes by lowering nuclear factor-kappab activity via extracellular signal-related kinase 1/2. Diabetes 2008;57: Piqueras L, Sanz MJ, Perretti M, et al. Activation of PPARbeta/delta inhibits leukocyte recruitment, cell adhesion molecule expression, and chemokine release. J Leukoc Biol 2009;86: Liou JY, Lee S, Ghelani D, Matijevic-Aleksic N, Wu KK. Protection of endothelial survival by peroxisome proliferator-activated receptor-delta mediated upregulation. Arterioscler Thromb Vasc Biol 2006;26: Barish GD, Atkins AR, Downes M, et al. PPARdelta regulates multiple proinflammatory pathways to suppress atherosclerosis. Proc Natl Acad Sci U S A 2008;105: Schote AB, Turner JD, Schiltz J, Muller CP. Nuclear receptors in human immune cells: expression and correlations. Mol Immunol 2007;44: al Yacoub N, Romanowska M, Krauss S, Schweiger S, Foerster J. PPARdelta is a type 1 IFN target gene and inhibits apoptosis in T cells. J Invest Dermatol 2008;128: Inoue I, Goto S, Matsunaga T, et al. The ligands/activators for peroxisome proliferator-activated receptor alpha (PPARalpha) and PPARgamma increase Cu2+,Zn2+-superoxide dismutase and decrease p22phox message expressions in primary endothelial cells. Metabolism 2001;50: Pesant M, Sueur S, Dutartre P, et al. Peroxisome proliferator-activated receptor delta (PPARdelta) activation protects H9c2 cardiomyoblasts from oxidative stress-induced apoptosis. Cardiovasc Res 2006;69: Musiek ES, Yin H, Milne GL, Morrow JD. Recent advances in the biochemistry and clinical relevance of the isoprostane pathway. Lipids 2005;40: Riserus U, Sprecher D, Johnson T, et al. Activation of peroxisome proliferatoractivated receptor (PPAR)delta promotes reversal of multiple metabolic abnormalities, reduces oxidative stress, and increases fatty acid oxidation in moderately obese men. Diabetes 2008;57: Yu BC, Chang CK, Ou HY, Cheng KC, Cheng JT. Decrease of peroxisome proliferator-activated receptor delta expression in cardiomyopathy of streptozotocin-induced diabetic rats. Cardiovasc Res 2008;80: Pang M, de la Monte SM, Longato L, et al. PPARdelta agonist attenuates alcoholinduced hepatic insulin resistance and improves liver injury and repair. J Hepatol 2009;50:

77 103. Ali F, Ali NS, Bauer A, et al. PPARdelta and PGC1alpha act cooperatively to induce haem oxygenase-1 and enhance vascular endothelial cell resistance to stress. Cardiovasc Res;85: Graham TL, Mookherjee C, Suckling KE, Palmer CN, Patel L. The PPARdelta agonist GW0742X reduces atherosclerosis in LDLR(-/-) mice. Atherosclerosis 2005;181: Georgiadi A, Lichtenstein L, Degenhardt T, et al. Induction of Cardiac Angptl4 by Dietary Fatty Acids Is Mediated by Peroxisome Proliferator-Activated Receptor {beta}/{delta} and Protects Against Fatty Acid-Induced Oxidative Stress. Circ Res He TC, Chan TA, Vogelstein B, Kinzler KW. PPARdelta is an APC-regulated target of nonsteroidal anti-inflammatory drugs. Cell 1999;99: Gupta RA, Tan J, Krause WF, et al. Prostacyclin-mediated activation of peroxisome proliferator-activated receptor delta in colorectal cancer. Proc Natl Acad Sci U S A 2000;97: Park BH, Vogelstein B, Kinzler KW. Genetic disruption of PPARdelta decreases the tumorigenicity of human colon cancer cells. Proc Natl Acad Sci U S A 2001;98: Harman FS, Nicol CJ, Marin HE, Ward JM, Gonzalez FJ, Peters JM. Peroxisome proliferator-activated receptor-delta attenuates colon carcinogenesis. Nat Med 2004;10: Reed KR, Sansom OJ, Hayes AJ, et al. PPARdelta status and Apc-mediated tumourigenesis in the mouse intestine. Oncogene 2004;23: Yang L, Zhou ZG, Zheng XL, et al. RNA interference against peroxisome proliferator-activated receptor delta gene promotes proliferation of human colorectal cancer cells. Dis Colon Rectum 2008;51:318-26; discussion Peters JM, Hollingshead HE, Gonzalez FJ. Role of peroxisome-proliferatoractivated receptor beta/delta (PPARbeta/delta) in gastrointestinal tract function and disease. Clin Sci (Lond) 2008;115: Notterman DA, Alon U, Sierk AJ, Levine AJ. Transcriptional gene expression profiles of colorectal adenoma, adenocarcinoma, and normal tissue examined by oligonucleotide arrays. Cancer Res 2001;61: Yang L, Zhou ZG, Luo HZ, Zhou B, Xia QJ, Tian C. Quantitative analysis of PPARdelta mrna by real-time RT-PCR in 86 rectal cancer tissues. Eur J Surg Oncol 2006;32: Honn KV, Meyer J. Thromboxanes and prostacyclin: positive and negative modulators of tumor growth. Biochem Biophys Res Commun 1981;102: Fukumoto K, Yano Y, Virgona N, et al. Peroxisome proliferator-activated receptor delta as a molecular target to regulate lung cancer cell growth. FEBS Lett 2005;579: Matthiessen MW, Pedersen G, Albrektsen T, Adamsen S, Fleckner J, Brynskov J. Peroxisome proliferator-activated receptor expression and activation in normal human colonic epithelial cells and tubular adenomas. Scand J Gastroenterol 2005;40:

78 118. Kim DJ, Akiyama TE, Harman FS, et al. Peroxisome proliferator-activated receptor beta (delta)-dependent regulation of ubiquitin C expression contributes to attenuation of skin carcinogenesis. J Biol Chem 2004;279: Peters JM, Lee SS, Li W, et al. Growth, adipose, brain, and skin alterations resulting from targeted disruption of the mouse peroxisome proliferator-activated receptor beta(delta). Mol Cell Biol 2000;20: Kim DJ, Bility MT, Billin AN, Willson TM, Gonzalez FJ, Peters JM. PPARbeta/delta selectively induces differentiation and inhibits cell proliferation. Cell Death Differ 2006;13: Girroir EE, Hollingshead HE, Billin AN, et al. Peroxisome proliferator-activated receptor-beta/delta (PPARbeta/delta) ligands inhibit growth of UACC903 and MCF7 human cancer cell lines. Toxicology 2008;243: Hollingshead HE, Killins RL, Borland MG, et al. Peroxisome proliferatoractivated receptor-beta/delta (PPARbeta/delta) ligands do not potentiate growth of human cancer cell lines. Carcinogenesis 2007;28: Cutler NS, Graves-Deal R, LaFleur BJ, et al. Stromal production of prostacyclin confers an antiapoptotic effect to colonic epithelial cells. Cancer Res 2003;63: Hendrickse CW, Keighley MR, Neoptolemos JP. Dietary omega-3 fats reduce proliferation and tumor yields at colorectal anastomosis in rats. Gastroenterology 1995;109: Gupta RA, Wang D, Katkuri S, Wang H, Dey SK, DuBois RN. Activation of nuclear hormone receptor peroxisome proliferator-activated receptor-delta accelerates intestinal adenoma growth. Nat Med 2004;10: Wahli W. Peroxisome proliferator-activated receptors (PPARs): from metabolic control to epidermal wound healing. Swiss Med Wkly 2002;132: Nijsten T, Geluyckens E, Colpaert C, Lambert J. Peroxisome proliferatoractivated receptors in squamous cell carcinoma and its precursors. J Cutan Pathol 2005;32: Jaeckel EC, Raja S, Tan J, et al. Correlation of expression of cyclooxygenase-2, vascular endothelial growth factor, and peroxisome proliferator-activated receptor delta with head and neck squamous cell carcinoma. Arch Otolaryngol Head Neck Surg 2001;127: Ye BH, Rao PH, Chaganti RS, Dalla-Favera R. Cloning of bcl-6, the locus involved in chromosome translocations affecting band 3q27 in B-cell lymphoma. Cancer Res 1993;53: Ye BH, Lista F, Lo Coco F, et al. Alterations of a zinc finger-encoding gene, BCL-6, in diffuse large-cell lymphoma. Science 1993;262: Cattoretti G, Chang CC, Cechova K, et al. BCL-6 protein is expressed in germinal-center B cells. Blood 1995;86: Allman D, Jain A, Dent A, et al. BCL-6 expression during B-cell activation. Blood 1996;87: Zollman S, Godt D, Prive GG, Couderc JL, Laski FA. The BTB domain, found primarily in zinc finger proteins, defines an evolutionarily conserved family that 65

79 includes several developmentally regulated genes in Drosophila. Proc Natl Acad Sci U S A 1994;91: Dhordain P, Albagli O, Lin RJ, et al. Corepressor SMRT binds the BTB/POZ repressing domain of the LAZ3/BCL6 oncoprotein. Proc Natl Acad Sci U S A 1997;94: Dhordain P, Lin RJ, Quief S, et al. The LAZ3(BCL-6) oncoprotein recruits a SMRT/mSIN3A/histone deacetylase containing complex to mediate transcriptional repression. Nucleic Acids Res 1998;26: Bardwell VJ, Treisman R. The POZ domain: a conserved protein-protein interaction motif. Genes Dev 1994;8: Kerckaert JP, Deweindt C, Tilly H, Quief S, Lecocq G, Bastard C. LAZ3, a novel zinc-finger encoding gene, is disrupted by recurring chromosome 3q27 translocations in human lymphomas. Nat Genet 1993;5: Rogers S, Wells R, Rechsteiner M. Amino acid sequences common to rapidly degraded proteins: the PEST hypothesis. Science 1986;234: Moriyama M, Yamochi T, Semba K, Akiyama T, Mori S. BCL-6 is phosphorylated at multiple sites in its serine- and proline-clustered region by mitogen-activated protein kinase (MAPK) in vivo. Oncogene 1997;14: Niu H, Ye BH, Dalla-Favera R. Antigen receptor signaling induces MAP kinasemediated phosphorylation and degradation of the BCL-6 transcription factor. Genes Dev 1998;12: Niu H. The proto-oncogene BCL-6 in normal and malignant B cell development. Hematol Oncol 2002;20: Toney LM, Cattoretti G, Graf JA, et al. BCL-6 regulates chemokine gene transcription in macrophages. Nat Immunol 2000;1: Chang CC, Ye BH, Chaganti RS, Dalla-Favera R. BCL-6, a POZ/zinc-finger protein, is a sequence-specific transcriptional repressor. Proc Natl Acad Sci U S A 1996;93: Seyfert VL, Allman D, He Y, Staudt LM. Transcriptional repression by the protooncogene BCL-6. Oncogene 1996;12: Kawamata N, Miki T, Ohashi K, et al. Recognition DNA sequence of a novel putative transcription factor, BCL6. Biochem Biophys Res Commun 1994;204: Harris MB, Chang CC, Berton MT, et al. Transcriptional repression of Stat6- dependent interleukin-4-induced genes by BCL-6: specific regulation of iepsilon transcription and immunoglobulin E switching. Mol Cell Biol 1999;19: Schaffer A, Cerutti A, Shah S, Zan H, Casali P. The evolutionarily conserved sequence upstream of the human Ig heavy chain S gamma 3 region is an inducible promoter: synergistic activation by CD40 ligand and IL-4 via cooperative NFkappa B and STAT-6 binding sites. J Immunol 1999;162: Perez-Rosado A, Artiga M, Vargiu P, Sanchez-Aguilera A, Alvarez-Barrientos A, Piris M. BCL6 represses NFkappaB activity in diffuse large B-cell lymphomas. J Pathol 2008;214: Ye BH, Cattoretti G, Shen Q, et al. The BCL-6 proto-oncogene controls germinalcentre formation and Th2-type inflammation. Nat Genet 1997;16:

80 150. Matsumoto M, Mariathasan S, Nahm MH, Baranyay F, Peschon JJ, Chaplin DD. Role of lymphotoxin and the type I TNF receptor in the formation of germinal centers. Science 1996;271: Fukuda T, Yoshida T, Okada S, et al. Disruption of the Bcl6 gene results in an impaired germinal center formation. J Exp Med 1997;186: Reljic R, Wagner SD, Peakman LJ, Fearon DT. Suppression of signal transducer and activator of transcription 3-dependent B lymphocyte terminal differentiation by BCL-6. J Exp Med 2000;192: Hsieh T, Wu JM. Apoptosis and restriction of G(1)/S cell cycle by fenretinide in Burkitt's lymphoma mutu I cell line accessed with bcl-6 down-regulation. Biochem Biophys Res Commun 2000;276: Alenzi FQ. BCL-6 prevents mammary epithelial apoptosis and promotes cell survival. J Pak Med Assoc 2008;58: Baron BW, Nucifora G, McCabe N, Espinosa R, 3rd, Le Beau MM, McKeithan TW. Identification of the gene associated with the recurring chromosomal translocations t(3;14)(q27;q32) and t(3;22)(q27;q11) in B-cell lymphomas. Proc Natl Acad Sci U S A 1993;90: Deweindt C, Kerckaert JP, Tilly H, Quief S, Nguyen VC, Bastard C. Cloning of a breakpoint cluster region at band 3q27 involved in human non-hodgkin's lymphoma. Genes Chromosomes Cancer 1993;8: Miki T, Kawamata N, Hirosawa S, Aoki N. Gene involved in the 3q27 translocation associated with B-cell lymphoma, BCL5, encodes a Kruppel-like zinc-finger protein. Blood 1994;83: Otsuki T, Yano T, Clark HM, et al. Analysis of LAZ3 (BCL-6) status in B-cell non-hodgkin's lymphomas: results of rearrangement and gene expression studies and a mutational analysis of coding region sequences. Blood 1995;85: Ohno H, Fukuhara S. Significance of rearrangement of the BCL6 gene in B-cell lymphoid neoplasms. Leuk Lymphoma 1997;27: Ye BH, Chaganti S, Chang CC, et al. Chromosomal translocations cause deregulated BCL6 expression by promoter substitution in B cell lymphoma. EMBO J 1995;14: Cattoretti G, Pasqualucci L, Ballon G, et al. Deregulated BCL6 expression recapitulates the pathogenesis of human diffuse large B cell lymphomas in mice. Cancer Cell 2005;7: Baron BW, Anastasi J, Montag A, et al. The human BCL6 transgene promotes the development of lymphomas in the mouse. Proc Natl Acad Sci U S A 2004;101: Phan RT, Dalla-Favera R. The BCL6 proto-oncogene suppresses p53 expression in germinal-centre B cells. Nature 2004;432: Polo JM, Dell'Oso T, Ranuncolo SM, et al. Specific peptide interference reveals BCL6 transcriptional and oncogenic mechanisms in B-cell lymphoma cells. Nat Med 2004;10: Chattopadhyay A, Tate SA, Beswick RW, Wagner SD, Ko Ferrigno P. A peptide aptamer to antagonize BCL-6 function. Oncogene 2006;25:

81 166. Offit K, Louie DC, Parsa NZ, et al. BCL6 gene rearrangement and other cytogenetic abnormalities in diffuse large cell lymphoma. Leuk Lymphoma 1995;20: Pescarmona E, De Sanctis V, Pistilli A, et al. Pathogenetic and clinical implications of Bcl-6 and Bcl-2 gene configuration in nodal diffuse large B-cell lymphomas. J Pathol 1997;183: Takeshita M, Iwashita A, Kurihara K, et al. Histologic and immunohistologic findings and prognosis of 40 cases of gastric large B-cell lymphoma. Am J Surg Pathol 2000;24: Cesarman E, Chadburn A, Liu YF, Migliazza A, Dalla-Favera R, Knowles DM. BCL-6 gene mutations in posttransplantation lymphoproliferative disorders predict response to therapy and clinical outcome. Blood 1998;92: Dent AL, Shaffer AL, Yu X, Allman D, Staudt LM. Control of inflammation, cytokine expression, and germinal center formation by BCL-6. Science 1997;276: Shaffer AL, Yu X, He Y, Boldrick J, Chan EP, Staudt LM. BCL-6 represses genes that function in lymphocyte differentiation, inflammation, and cell cycle control. Immunity 2000;13: Dent AL, Vasanwala FH, Toney LM. Regulation of gene expression by the protooncogene BCL-6. Crit Rev Oncol Hematol 2002;41: Maiese K, Chong ZZ, Hou J, Shang YC. Oxidative stress: Biomarkers and novel therapeutic pathways. Exp Gerontol;45: Wei W, Liu Q, Tan Y, Liu L, Li X, Cai L. Oxidative stress, diabetes, and diabetic complications. Hemoglobin 2009;33: Matsuda S, Umeda M, Uchida H, Kato H, Araki T. Alterations of oxidative stress markers and apoptosis markers in the striatum after transient focal cerebral ischemia in rats. J Neural Transm 2009;116: Sekler A, Jimenez JM, Rojo L, et al. Cognitive impairment and Alzheimer's disease: Links with oxidative stress and cholesterol metabolism. Neuropsychiatr Dis Treat 2008;4: Vaculin S, Franek M, Vejrazka M. Role of oxidative stress in animal model of visceral pain. Neurosci Lett Yazihan N, Uzuner K, Salman B, Vural M, Koken T, Arslantas A. Erythropoietin improves oxidative stress following spinal cord trauma in rats. Injury 2008;39: Wells PG, McCallum GP, Chen CS, et al. Oxidative stress in developmental origins of disease: teratogenesis, neurodevelopmental deficits, and cancer. Toxicol Sci 2009;108: Uchida K. 4-Hydroxy-2-nonenal: a product and mediator of oxidative stress. Prog Lipid Res 2003;42: Uchida K. Role of reactive aldehyde in cardiovascular diseases. Free Radic Biol Med 2000;28: Uchida K, Kanematsu M, Sakai K, et al. Protein-bound acrolein: potential markers for oxidative stress. Proc Natl Acad Sci U S A 1998;95:

82 183. Uchida K, Kanematsu M, Morimitsu Y, Osawa T, Noguchi N, Niki E. Acrolein is a product of lipid peroxidation reaction. Formation of free acrolein and its conjugate with lysine residues in oxidized low density lipoproteins. J Biol Chem 1998;273: Esterbauer H, Schaur RJ, Zollner H. Chemistry and biochemistry of 4- hydroxynonenal, malonaldehyde and related aldehydes. Free Radic Biol Med 1991;11: Lee SH, Blair IA. Characterization of 4-oxo-2-nonenal as a novel product of lipid peroxidation. Chem Res Toxicol 2000;13: Schneider C, Tallman KA, Porter NA, Brash AR. Two distinct pathways of formation of 4-hydroxynonenal. Mechanisms of nonenzymatic transformation of the 9- and 13-hydroperoxides of linoleic acid to 4-hydroxyalkenals. J Biol Chem 2001;276: Gardner HW, Hamberg M. Oxygenation of (3Z)-nonenal to (2E)-4-hydroxy-2- nonenal in the broad bean (Vicia faba L.). J Biol Chem 1993;268: Gardner HW, Grove MJ. Soybean lipoxygenase-1 oxidizes 3Z-nonenal. A route to 4s-hydroperoxy-2e-nonenal and related products. Plant Physiol 1998;116: Burczynski ME, Sridhar GR, Palackal NT, Penning TM. The reactive oxygen species--and Michael acceptor-inducible human aldo-keto reductase AKR1C1 reduces the alpha,beta-unsaturated aldehyde 4-hydroxy-2-nonenal to 1,4- dihydroxy-2-nonene. J Biol Chem 2001;276: Uchida K, Stadtman ER. Modification of histidine residues in proteins by reaction with 4-hydroxynonenal. Proc Natl Acad Sci U S A 1992;89: Toyokuni S, Uchida K, Okamoto K, Hattori-Nakakuki Y, Hiai H, Stadtman ER. Formation of 4-hydroxy-2-nonenal-modified proteins in the renal proximal tubules of rats treated with a renal carcinogen, ferric nitrilotriacetate. Proc Natl Acad Sci U S A 1994;91: Ihara Y, Toyokuni S, Uchida K, et al. Hyperglycemia causes oxidative stress in pancreatic beta-cells of GK rats, a model of type 2 diabetes. Diabetes 1999;48: Urabe T, Yamasaki Y, Hattori N, Yoshikawa M, Uchida K, Mizuno Y. Accumulation of 4-hydroxynonenal-modified proteins in hippocampal CA1 pyramidal neurons precedes delayed neuronal damage in the gerbil brain. Neuroscience 2000;100: Yamamoto H, Yamamoto Y, Yamagami K, et al. Heat-shock preconditioning reduces oxidative protein denaturation and ameliorates liver injury by carbon tetrachloride in rats. Res Exp Med (Berl) 2000;199: Ide T, Tsutsui H, Kinugawa S, et al. Mitochondrial electron transport complex I is a potential source of oxygen free radicals in the failing myocardium. Circ Res 1999;85: Uchida K, Shiraishi M, Naito Y, Torii Y, Nakamura Y, Osawa T. Activation of stress signaling pathways by the end product of lipid peroxidation. 4-hydroxy-2- nonenal is a potential inducer of intracellular peroxide production. J Biol Chem 1999;274:

83 197. Fukuda A, Nakamura Y, Ohigashi H, Osawa T, Uchida K. Cellular response to the redox active lipid peroxidation products: induction of glutathione S- transferase P by 4-hydroxy-2-nonenal. Biochem Biophys Res Commun 1997;236: Tjalkens RB, Luckey SW, Kroll DJ, Petersen DR. Alpha,beta-unsaturated aldehydes increase glutathione S-transferase mrna and protein: correlation with activation of the antioxidant response element. Arch Biochem Biophys 1998;359: Itoh K, Chiba T, Takahashi S, et al. An Nrf2/small Maf heterodimer mediates the induction of phase II detoxifying enzyme genes through antioxidant response elements. Biochem Biophys Res Commun 1997;236: Itoh K, Wakabayashi N, Katoh Y, et al. Keap1 represses nuclear activation of antioxidant responsive elements by Nrf2 through binding to the amino-terminal Neh2 domain. Genes Dev 1999;13: Kumagai T, Nakamura Y, Osawa T, Uchida K. Role of p38 mitogen-activated protein kinase in the 4-hydroxy-2-nonenal-induced cyclooxygenase-2 expression. Arch Biochem Biophys 2002;397: Page S, Fischer C, Baumgartner B, et al. 4-Hydroxynonenal prevents NF-kappaB activation and tumor necrosis factor expression by inhibiting IkappaB phosphorylation and subsequent proteolysis. J Biol Chem 1999;274: Okada K, Wangpoengtrakul C, Osawa T, Toyokuni S, Tanaka K, Uchida K. 4- Hydroxy-2-nonenal-mediated impairment of intracellular proteolysis during oxidative stress. Identification of proteasomes as target molecules. J Biol Chem 1999;274: Ji C, Amarnath V, Pietenpol JA, Marnett LJ. 4-hydroxynonenal induces apoptosis via caspase-3 activation and cytochrome c release. Chem Res Toxicol 2001;14: Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, CA Cancer J Clin 2009;59: Caldas C, Hahn SA, da Costa LT, et al. Frequent somatic mutations and homozygous deletions of the p16 (MTS1) gene in pancreatic adenocarcinoma. Nat Genet 1994;8: Garcea G, Neal CP, Pattenden CJ, Steward WP, Berry DP. Molecular prognostic markers in pancreatic cancer: a systematic review. Eur J Cancer 2005;41: Moskaluk CA, Hruban RH, Kern SE. p16 and K-ras gene mutations in the intraductal precursors of human pancreatic adenocarcinoma. Cancer Res 1997;57: Calhoun ES, Jones JB, Ashfaq R, et al. BRAF and FBXW7 (CDC4, FBW7, AGO, SEL10) mutations in distinct subsets of pancreatic cancer: potential therapeutic targets. Am J Pathol 2003;163: Han H, Bearss DJ, Browne LW, Calaluce R, Nagle RB, Von Hoff DD. Identification of differentially expressed genes in pancreatic cancer cells using cdna microarray. Cancer Res 2002;62:

84 211. Berg T, Cohen SB, Desharnais J, et al. Small-molecule antagonists of Myc/Max dimerization inhibit Myc-induced transformation of chicken embryo fibroblasts. Proc Natl Acad Sci U S A 2002;99: Ruggeri BA, Huang L, Wood M, Cheng JQ, Testa JR. Amplification and overexpression of the AKT2 oncogene in a subset of human pancreatic ductal adenocarcinomas. Mol Carcinog 1998;21: Altomare DA, Tanno S, De Rienzo A, et al. Frequent activation of AKT2 kinase in human pancreatic carcinomas. J Cell Biochem 2002;87: Osaki M, Oshimura M, Ito H. PI3K-Akt pathway: its functions and alterations in human cancer. Apoptosis 2004;9: Holzmann K, Kohlhammer H, Schwaenen C, et al. Genomic DNA-chip hybridization reveals a higher incidence of genomic amplifications in pancreatic cancer than conventional comparative genomic hybridization and leads to the identification of novel candidate genes. Cancer Res 2004;64: Li J, Kleeff J, Giese N, Buchler MW, Korc M, Friess H. Gefitinib ('Iressa', ZD1839), a selective epidermal growth factor receptor tyrosine kinase inhibitor, inhibits pancreatic cancer cell growth, invasion, and colony formation. Int J Oncol 2004;25: Schutte M, Hruban RH, Geradts J, et al. Abrogation of the Rb/p16 tumorsuppressive pathway in virtually all pancreatic carcinomas. Cancer Res 1997;57: Wilentz RE, Geradts J, Maynard R, et al. Inactivation of the p16 (INK4A) tumorsuppressor gene in pancreatic duct lesions: loss of intranuclear expression. Cancer Res 1998;58: Hahn SA, Schutte M, Hoque AT, et al. DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 1996;271: Schutte M, Hruban RH, Hedrick L, et al. DPC4 gene in various tumor types. Cancer Res 1996;56: Rubio-Viqueira B, Jimeno A, Cusatis G, et al. An in vivo platform for translational drug development in pancreatic cancer. Clin Cancer Res 2006;12: Kerbel RS. Human tumor xenografts as predictive preclinical models for anticancer drug activity in humans: better than commonly perceived-but they can be improved. Cancer Biol Ther 2003;2:S Loukopoulos P, Kanetaka K, Takamura M, Shibata T, Sakamoto M, Hirohashi S. Orthotopic transplantation models of pancreatic adenocarcinoma derived from cell lines and primary tumors and displaying varying metastatic activity. Pancreas 2004;29: Hingorani SR, Wang L, Multani AS, et al. Trp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice. Cancer Cell 2005;7: Aguirre AJ, Bardeesy N, Sinha M, et al. Activated Kras and Ink4a/Arf deficiency cooperate to produce metastatic pancreatic ductal adenocarcinoma. Genes Dev 2003;17:

85 226. Katz MH, Hwang R, Fleming JB, Evans DB. Tumor-node-metastasis staging of pancreatic adenocarcinoma. CA Cancer J Clin 2008;58: Bilimoria KY, Bentrem DJ, Ko CY, et al. Validation of the 6th edition AJCC Pancreatic Cancer Staging System: report from the National Cancer Database. Cancer 2007;110: Lowenfels AB, Maisonneuve P. Epidemiology and prevention of pancreatic cancer. Jpn J Clin Oncol 2004;34: Ahlgren JD. Epidemiology and risk factors in pancreatic cancer. Semin Oncol 1996;23: Chappuis PO, Ghadirian P, Foulkes WD. The role of genetic factors in the etiology of pancreatic adenocarcinoma: an update. Cancer Invest 2001;19: Fernandez E, La Vecchia C, Porta M, Negri E, Lucchini F, Levi F. Trends in pancreatic cancer mortality in Europe, Int J Cancer 1994;57: Ernster VL, Sacks ST, Selvin S, Petrakis NL. Cancer incidence by marital status: U.S. Third National Cancer Survey. J Natl Cancer Inst 1979;63: Ghadirian P, Lynch HT, Krewski D. Epidemiology of pancreatic cancer: an overview. Cancer Detect Prev 2003;27: Pernick NL, Sarkar FH, Philip PA, et al. Clinicopathologic analysis of pancreatic adenocarcinoma in African Americans and Caucasians. Pancreas 2003;26: Longnecker DS, Karagas MR, Tosteson TD, Mott LA. Racial differences in pancreatic cancer: comparison of survival and histologic types of pancreatic carcinoma in Asians, blacks, and whites in the United States. Pancreas 2000;21: Ekoe JM, Ghadirian P, Simard A, Baillargeon J, Perret C. [Diabetes mellitus and pancreatic cancer: a case-control study in greater Montreal, Quebec, Canada]. Rev Epidemiol Sante Publique 1992;40: Gullo L. Diabetes and the risk of pancreatic cancer. Ann Oncol 1999;10 Suppl 4: Gapstur SM, Gann PH, Lowe W, Liu K, Colangelo L, Dyer A. Abnormal glucose metabolism and pancreatic cancer mortality. JAMA 2000;283: Bansal P, Sonnenberg A. Pancreatitis is a risk factor for pancreatic cancer. Gastroenterology 1995;109: Lowenfels AB, Maisonneuve P, Cavallini G, et al. Pancreatitis and the risk of pancreatic cancer. International Pancreatitis Study Group. N Engl J Med 1993;328: Kattwinkel J, Lapey A, Di Sant'Agnese PA, Edwards WA. Hereditary pancreatitis: three new kindreds and a critical review of the literature. Pediatrics 1973;51: Lowenfels AB, Maisonneuve P, DiMagno EP, et al. Hereditary pancreatitis and the risk of pancreatic cancer. International Hereditary Pancreatitis Study Group. J Natl Cancer Inst 1997;89: Prokopczyk B, Hoffmann D, Bologna M, et al. Identification of tobacco-derived compounds in human pancreatic juice. Chem Res Toxicol 2002;15: Hecht SS. DNA adduct formation from tobacco-specific N-nitrosamines. Mutat Res 1999;424:

86 245. Risch HA. Etiology of pancreatic cancer, with a hypothesis concerning the role of N-nitroso compounds and excess gastric acidity. J Natl Cancer Inst 2003;95: Silverman DT, Dunn JA, Hoover RN, et al. Cigarette smoking and pancreas cancer: a case-control study based on direct interviews. J Natl Cancer Inst 1994;86: Lin Y, Tamakoshi A, Kawamura T, et al. A prospective cohort study of cigarette smoking and pancreatic cancer in Japan. Cancer Causes Control 2002;13: Durbec JP, Chevillotte G, Bidart JM, Berthezene P, Sarles H. Diet, alcohol, tobacco and risk of cancer of the pancreas: a case-control study. Br J Cancer 1983;47: Burch GE, Ansari A. Chronic alcoholism and carcinoma of the pancreas. A correlative hypothesis. Arch Intern Med 1968;122: Hirayama T. Epidemiology of pancreatic cancer in Japan. Jpn J Clin Oncol 1989;19: MacMahon B, Yen S, Trichopoulos D, Warren K, Nardi G. Coffee and cancer of the pancreas. N Engl J Med 1981;304: Villeneuve PJ, Johnson KC, Hanley AJ, Mao Y. Alcohol, tobacco and coffee consumption and the risk of pancreatic cancer: results from the Canadian Enhanced Surveillance System case-control project. Canadian Cancer Registries Epidemiology Research Group. Eur J Cancer Prev 2000;9: Wynder EL, Hall NE, Polansky M. Epidemiology of coffee and pancreatic cancer. Cancer Res 1983;43: Michaud DS, Giovannucci E, Willett WC, Colditz GA, Fuchs CS. Dietary meat, dairy products, fat, and cholesterol and pancreatic cancer risk in a prospective study. Am J Epidemiol 2003;157: Qiu D, Kurosawa M, Lin Y, et al. Overview of the epidemiology of pancreatic cancer focusing on the JACC Study. J Epidemiol 2005;15 Suppl 2:S Silverman DT, Swanson CA, Gridley G, et al. Dietary and nutritional factors and pancreatic cancer: a case-control study based on direct interviews. J Natl Cancer Inst 1998;90: Stolzenberg-Solomon RZ, Pietinen P, Taylor PR, Virtamo J, Albanes D. Prospective study of diet and pancreatic cancer in male smokers. Am J Epidemiol 2002;155: Inoue M, Tajima K, Takezaki T, et al. Epidemiology of pancreatic cancer in Japan: a nested case-control study from the Hospital-based Epidemiologic Research Program at Aichi Cancer Center (HERPACC). Int J Epidemiol 2003;32: Berrington de Gonzalez A, Sweetland S, Spencer E. A meta-analysis of obesity and the risk of pancreatic cancer. Br J Cancer 2003;89: Michaud DS, Giovannucci E, Willett WC, Colditz GA, Stampfer MJ, Fuchs CS. Physical activity, obesity, height, and the risk of pancreatic cancer. JAMA 2001;286:

87 261. Burris HA, 3rd, Moore MJ, Andersen J, et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997;15: Moore MJ, Goldstein D, Hamm J, et al. Erlotinib plus gemcitabine compared with gemcitabine alone in patients with advanced pancreatic cancer: a phase III trial of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 2007;25: Van Cutsem E, Vervenne WL, Bennouna J, et al. Phase III trial of bevacizumab in combination with gemcitabine and erlotinib in patients with metastatic pancreatic cancer. J Clin Oncol 2009;27: Bilimoria KY, Bentrem DJ, Ko CY, Stewart AK, Winchester DP, Talamonti MS. National failure to operate on early stage pancreatic cancer. Ann Surg 2007;246: Kleeff J, Friess H, Buchler MW. Neoadjuvant therapy for pancreatic cancer. Br J Surg 2007;94: Esposito I, Kleeff J, Bergmann F, et al. Most pancreatic cancer resections are R1 resections. Ann Surg Oncol 2008;15: Kleeff J, Reiser C, Hinz U, et al. Surgery for recurrent pancreatic ductal adenocarcinoma. Ann Surg 2007;245: Klinkenbijl JH, Jeekel J, Sahmoud T, et al. Adjuvant radiotherapy and 5- fluorouracil after curative resection of cancer of the pancreas and periampullary region: phase III trial of the EORTC gastrointestinal tract cancer cooperative group. Ann Surg 1999;230:776-82; discussion

88 Chapter 2 The oxidative stress mediator 4-hydroxynonenal is an intracellular agonist of the nuclear receptor peroxisome proliferator-activated receptor-β/δ (PPARβ/δ) 2.1 Abstract Liver insufficiency and damage is a major cause of death and disease worldwide and may result from exposure to environmental toxicants, specific combinations or dosages of pharmaceuticals and microbial metabolites. The generation of reactive intermediates, in particular 4-hydroxynonenal (4-HNE), is a common event in liver damage caused by a variety of hepatotoxic drugs and solvents. The peroxisome proliferator-activated receptors (PPARs) are nuclear receptors that are involved in the transcriptional regulation of lipid metabolism as well as other biological functions. Importantly, we have observed that the PPARβ/δ-/- mouse is more susceptible to chemically-induced hepatotoxicity than its wildtype counterpart. We hypothesized that PPARβ/δ plays a protective role by responding to toxic lipids and altering gene expression accordingly. In support, oxidized-vldl and its constituents including 13-Shydroxyoctadeca-dienoic acid (13-S-HODE) and 4-HNE were able to activate PPARβ/δ. A structure-activity relationship was established where 4-HNE and 4- hydroperoxynonenal (4-HpNE) enhanced the activity of the PPARβ/δ subtype while 4-

89 76 hyroxy-hexenal (4-HHE), 4-oxo-2-nonenal (4-ONE), and trans-4,5-epoxy-2(e)-decenal did not activate this receptor. Increasing PPARβ/δ activity with a synthetic agonist decreased sensitivity of hepatocytes to 4-HNE and other toxic agents, whereas inhibition of this receptor had the opposite result. Thus, our research establishes 4-HNE as an endogenous modulator of PPARβ/δ activity and raises the possibility that agonists of this nuclear receptor may be utilized to prevent or treat liver disease associated with oxidative damage. 2.2 Introduction Greater than 2.2 million hospitalized Americans suffer adverse drug reactions each year, with liver toxicity presenting as the most common adverse effect, and approximately 100,000 individuals die unintentionally from administration of medications (1, 2). Many additional cases of liver failure occur due to acute, chronic and degenerative disease processes, including those related to acetaminophen overdose, alcohol consumption and solvent exposures. Reactive oxygen intermediates (ROI) elicit oxidative decomposition of polyunsaturated fatty acids (i.e, lipid peroxidation), leading to the formation of a complex mixture of aldehydic end-products, including malondialdehyde (MDA), 4-HNE, and other alkenals (3). These aldehydic molecules have been considered the ultimate mediators of toxic effects elicited by oxidative stress but may also affect cellular function at non-toxic levels via signal transduction, gene expression and cell proliferation. Although the overt toxicity caused by aldehydic endproducts is due primarily to covalent binding to cellular macromolecules, the effects on

90 77 signal transduction are not well-characterized. Since millions of individuals suffer adverse drug reactions each year it is important to understand how the cell responds to intracellular insults such as production of ROI and 4-HNE. The peroxisome proliferator-activated receptors (PPARs) are nuclear receptors that exist as three subtypes (α, β/δ and γ), which exhibit tissue-specific expression, preferential ligand recognition, and distinct biological functions (4-9). Although important as targets of pharmaceutical intervention, there is increasing evidence that the biological niche occupied by the PPARs is that of a receptor for fatty acid and their metabolites. Of the three PPAR genes (α, β/δ, γ), the PPARβ/δ isoform is the least well-studied in terms of its biological functions and endogenous ligands. PPARβ/δ plays an important role in differentiation of epithelial tissues, fatty acid catabolism in skeletal muscle, improvement of insulin sensitivity, attenuated weight gain and elevated HDL levels (10). Emerging evidence suggests the presence of this receptor is important in ameliorating the effects of hepatotoxicants. For example, histological examination of liver and analysis of markers of overt damage to this organ (serum GPT) after treatment with the xenobiotics azoxymethane (AOM), arsenic or carbon tetrachloride demonstrated that the extent of liver toxicity in PPARβ/δ-null mice was more severe than in wild-type mice. While it is remotely possible that the metabolic fate of these hepatotoxicants could be influenced by PPARβ/δ, it is more likely that regulation of oxidative stress underlies the protective role of this receptor in liver. These chemicals share a common mechanism of overt toxicity via production of ROI and oxidized lipid intermediates. For example, CCl 4 affects eicosanoid pathways (11, 12) and increases

91 78 circulating prostaglandin E2 (PGE2) levels (13) and 4-HNE and 4-HNE-protein adducts (3, 14, 15). One possible explanation for the increased susceptibility of PPARβ/δ-/- mice to hepatotoxicity is that oxidative damage increases the production of an endogenous ligand for PPARβ/δ. This putative agonist would in turn stimulate lipid metabolism and degradation of lipid peroxidation intermediates. PPARs are well recognized as transcriptional regulators of lipid metabolism, transport, storage and other activities (16). In the absence of PPARβ/δ the signaling cascade would be disrupted and accumulation of toxic lipids such as 4-HNE would result. If our hypothesis were correct, endogenous ligands of PPARβ/δ should include oxidized lipids, in particular those derived from fatty acids. In support, we discovered that oxidized-vldl and constituents including 13(S)- HODE and 4-HNE are PPARβ/δ agonists. In addition, modulating PPARβ/δ activity affects the sensitivity of hepatocytes to 4-HNE and other toxic agents. This research raises the possibility that PPARβ/δ agonists may be utilized to prevent or treat liver disease associated with the generation of ROIs. 2.3 Materials and Methods Reagents. VLDL (human plasma) was purchased from Calbiochem (La Jolla, CA), LPL was purchased from Sigma (lyophilized powder) and reconstituted in PBS (10 mg/ml), and 13(S)-HODE, 13(S)-HpODE, 4-HHE, 4-ONE, trans-4,5-epoxy-2e-decenal, 4-HpNE, and 4-HNE were purchased from Cayman Chemical (Ann Arbor, MI) and used without

92 79 further purification. The synthesis and purification of linoleic and arachidonic acid oxidation products such as 9-HODE, 12-HpODE, 5-HETE, 9-HETE, 12-HETE, 15- HETE, 5-HpETE, 15-HpETE, 5,15-diHpETE and 5,6-diHETE were performed as described (17). The authenticity of each of the lipid mediators was confirmed using cochromatography as well as gas chromatography-mass spectrometric experiments. UV- Visible spectroscopy was used to determine their respective concentrations, and the compounds were reconstituted in anhydrous ethanol to the desired concentration. The compounds were used within 30 min of reconstitution in ethanol. Plasmids. Plasmids used in reporter assays including pm/mppar-α, -β, and γ, pbk/mppar-α, -β, and γ (murine), ACO Luciferase and pfr-luciferase (Promega) have been described elsewhere (18). Cell culture. 3T3-L1 preadipocyte cells were grown in standard high-glucose Dulbecco s Modified Eagle Medium (DMEM) containing 10% FBS at 37 C and 5% CO2. Murine SV-40-immortalized Hepatocytes (MuSH) were cultured in standard α- MEM containing 10% FBS. The protocol for the isolation and immortalization of hepatocytes are described later in this section. Cells were treated as described in Fig. legends. Transient transfection and treatment. Cells were counted using a hemocytometer following staining with Trypan Blue Solution (0.4%) (Sigma) and plated onto 10-cm dishes at densities between 600, ,000 cells/dish for transfection. Transfections

93 80 were performed at 37 C for 6 h using 12 µg of the appropriate Gal4-PPAR ligand binding domain (LBD) plasmid (pm/mppar-α, -β, or γ) containing a Luciferase reporter, and 24 µg Lipofectamine (Invitrogen). Transfection media was then replaced with fresh high-glucose DMEM after washing the dishes with PBS and the cells were allowed to recover overnight. Cells were re-seeded into 96-well plates (Costar, Corning, NY) 24 h following transfection and allowed to sit undisturbed at 37 C for 2-3 h to adhere to the well bottom. The media was removed and 100 µl of fresh media containing the compound of interest at the desired concentration was added to the cells. Treatments were performed for 12 h at 37 C and 5% CO2. VLDL Experiments. VLDL control experiments used DMSO (the vehicle for positive controls ciprofibrate, tetradecylthioacetic acid (TTA) and rosiglitazone for PPARα, -β/δ and γ, respectively) only in a volume equal to the highest volume used for positive controls, as well as a CuSO4 control at 10 µm. All solutions were prepared in a sterile environment using high-glucose DMEM and were allowed to incubate at 37 C for 1 h prior to treatment. Cells received VLDL and oxvldl at 10 µg/ml. VLDL and oxvldl were treated with 1 µl each of Lipoprotein lipase (LpL, 715 units/µl) and allowed to incubate at 37 C without shaking for 1 h prior to treatment. 13(S)-HODE and 4-HNE experiments. Fatty acid treatment experiments used ethanol (EtOH; the vehicle for 13(S)-HODE and other fatty acids and their metabolites) only in a volume equal to the highest volume used for 13(S)-HODE or 4-HNE. Cells received

94 81 13(S)-HODE and other fatty acids at 50 µm and 4-HNE at 25 µm, except where indicated. Oxidation of VLDL and 13(S)-HODE with Cu2+. Ten-µM copper solutions were prepared fresh for every experiment from a stock solution of aqueous CuSO 4 at 20 mm. VLDL and 13(S)-HODE solutions were treated with an equal volume of aqueous CuSO 4 and allowed to incubate at 37 C in an incubator/shaker for a period of 72 h, unless otherwise indicated. Temperature-sensitive SV40 virus preparation. CV-1 cells were grown to confluence in T-75 flasks at 37 C in αmem-4% FBS. Cells were incubated with stock temperature sensitive SV40 virus (19) at 37 C and agitated every 15 minutes for 2 h. Media was changed to αmem-8% FBS, and the cells were incubated at 34 C for seven days. Viruscontaining media was removed upon gross cell pathology (nuclear inclusion, cytoplasmic vacuolation and lysis) and stored in liquid nitrogen. Quantification of virus titer was not performed. Isolation, maintenance and infection of primary hepatocytes. The isolation and infection of hepatocytes was adapted from a previously described method (20). Purebred wild-type and PPARβ-null mice on a SV/129 background were used and have been previously described (21, 22). Primary hepatocytes were isolated from two male wildtype mice (10 day old) and two male PPARβ-null mice (5 day old). Mice were euthanized by overexposure to carbon dioxide and whole liver was removed and incubated at 37 C

95 82 with 1% collagenase in Hanks buffered saline solution (HBSS) for 5 minutes. Hepatocytes were isolated by centrifugation in 10 % Percoll (Amersham, Piscataway, NJ) in HBSS at 1000 rpm for 20 minutes at 4 C. The cells were washed with 10% fetal bovine serum (FBS) in αmem, centrifuged and resuspended in αmem-10% FBS. Each liver was separated into three populations to be infected separately; individual clones were not selected. Cells were cultured with αmem, 4% FBS, dexamethasone (dex) and 1% penicillin/streptomycin and allowed to grow to approximately 75% confluency at 34 C. The cells were then washed with αmem-10% FBS, overlayed with 1mL of αmem-10% FBS and infected with 200 µl of stock virus per well. Cells were incubated at 34 C for 2-3 hours and gently agitated by hand. Virus-containing media was then aspirated, and the cells were given αmem-4% FBS-0.1 µm dex. Media was changed twice weekly for approximately six weeks. Resultant colonies were isolated as mixed populations or as individual clones and are designated MuSH WT (wild-type) or MuSHβ/δ-/- (PPARβ/δ null). Messenger RNA examination. MuSH WT and MuSH β/δ-/- cells were grown in αmem containing 10% FBS, 100 units/ml penicillin and 100 µg/ml streptomycin. MuSH cells were grown to 75% confluency and treated overnight with the compound of interest. Total RNA was isolated using Tri-Reagent. Quantitative reverse transcriptase polymerase chain reaction (RT-PCR) was performed to measure β-actin, and adipose differentiation related protein (ADRP) mrna levels. The following ADRP gene-specific primers were synthesized and used: 5'-AGTGGAAGAGAAGCATCGGCT-3' (forward)

96 83 and 5'-TCGATGTGCTCAACACAGTGG-3' (reverse); β-actin primers have been described elsewhere (18). Mouse Oligonucleotide Arrays. The Mouse Genome Oligo Set Version 1 was purchased from Operon (Alameda, CA) and contains 6800 optimized 70-mers plus 24 controls, melting temperature normalized to 78 C. Sequences were optimized by the manufacturer using BLAST against all known mouse genes to minimize crosshybridization. Oligonucleotides were printed onto glass slides using GeneMachines Omnigrid (San Carlos, California) with additional controls obtained from Stratagene (SpotReport system, La Jolla, CA) at the Penn State University microarray core facility. Microarray Analysis. Total RNA was isolated by TriReagent (Sigma) and further purified with RNAEasy (Qiagen) according the manufacturers instructions. Labeling and hybridization was performed as discussed previously (18, 20). In the present experiments, co-hybridization was performed with cdna from three separate experiments (all listed as Cy5/Cy3-labeled): MuSHWT / MuSHβ/δ-/-; MuSH WT cells treated with TTA (50µM) compared to DMSO 6 h after treatment (TTA/DMSO), and MuSH WT cells treated with 4-HNE (25µM) compared to DMSO 6 h after treatment (4- HNE/DMSO). Statistical analysis was performed using a Student t-test with a p-value of 0.05 with the additional criteria of being either 2-fold increased or decreased by PPARβ/δ activation. Microarray analysis was performed using GeneSpring version 7.3 software program (Agilent Technologies) and results are given as mean and p-value.

97 84 13(S)-HODE Enzyme-linked immunosorbent assay. One hundred fifty-µl of VLDL was oxidized as described previously and incubated with 2 µl LpL at 37 C for 1h. Detection of 13(S)-HODE was achieved by using a 13(S)-HODE Correlate enzyme immunoassay kit (Assay Designs). Expression of mpparβ/δ LBD and 4-HNE-Capture Western immunoblot analysis. Mouse PPARβ/δ ligand binding domain (LBD) was expressed as a His-tagged fusion in bacteria using standard approaches. The mpparβ/δlbd-his protein, thus purified, was determined to be of >60 % purity. To test the nature of interaction of 4-HNE with PPARβ/δ, we incubated the PPARβ/δ (containing domains D, E/F) with 100 µm 4-HNE for 24 h in the dark. The mixture was resolved by PAGE, transferred to a polyvinyl pyrrolidone (PVDF) membrane, and probed with a goat polyclonal antibody raised against 4-HNE mixed protein adducts (Northwest Life Sciences, Vancouver, WA) or PPARβ/δ (Santa Cruz). Appropriate secondary antibodies conjugated with HRP were used followed by autoradiography. Molecular Modeling. Molecular modeling employed Macromodel v. 8.6 and the coordinates of PPARβ/δ bound to eicosapentaenoic acid (PDB ID: 3GWX). The carbonyl of the bound fatty acid was superimposed with the carbonyl of 4-(S)-hydroxynonenal. Ligand and protein residues within 6 Å were minimized to convergence using the OPLS 2003 force field.

98 Results Oxidized VLDL activates PPARβ/δ. We hypothesized that endogenous ligands for PPARβ/δ would include oxidized lipids, such as those derived from VLDL. The VLDL particle is a mixture of free and esterified cholesterols, triglycerides (formed from various fatty acids) and apolipoproteins. As seen in Fig. 2-1 A, the oxidation of VLDL with CuSO 4 (oxvldl) and subsequent incubation with LpL greatly increased PPARβ/δ activity in a dose-dependent manner (Fig. 2-1 B). To determine which potential products released by oxvldl are playing a role in the enhancement of PPARβ/δ activity, we conducted a screen of fatty acids and their CuSO 4 oxidation products. In Fig. 2-1 C we show that, while 13(S)-HODE alone activates PPARβ/δ, its oxidation product increases activity roughly two-fold higher; 15-HETE was also a PPARβ/δ agonist. Both 13(S)- HODE and its hydroperoxy derivative, 13(S)-HpODE, were equally potent in their ability to activate PPARβ/δ (data not shown). Furthermore, we hypothesized that 13(S)-HODE is a component of the oxvldl particle that is hydrolyzed upon incubation with LpL and ultimately induces PPARβ/δ activity. Our results show that the release of 13(S)-HODE increased correspondingly with the amount of oxvldl used (Fig. 2-1 C, Inset).

99 Figure 2-1: A. Oxidation of VLDL leads to increased activation of PPARβ/δ. VLDL (10 µg/ml) was oxidized with CuSO 4 (10 µm) for 72 h and then pretreated with lipoprotein lipase (LpL) for 1 h at 37 C, and the mixture was added to 3T3-L1 cells transfected with GAL4-PPARβ/δ-LBD and the UAS-luciferase constructs. Following treatment the cells were lysed and relative luciferase activity was determined. Results were standardized to the average DMSO-treated samples (n = 8, representative of two independent experiments). B. Dose-dependent activation of PPARβ/δ by LPL-treated oxvldl. 3T3-L1 preadipocytes were transfected with Gal4-LBD-mPPARβ and treated overnight with oxidized VLDL incubated for 1 h with LpL. The cells were lysed and relative luciferase activity was determined. (n =8, representative of at least three independent experiments.) C. Oxidized 13-S-HODE is a strong PPARβ/δ agonist. Chemicals were obtained as described under Materials and methods. UV-visible spectroscopy was used to determine their respective concentrations, and the original solvent was evaporated under a stream of N 2 and the chemical reconstituted in ethanol to 50 µm. Oxidation products were obtained via incubation with CuSO 4 (10 µm) for 72 h. 3T3-L1 preadipocytes were transfected with Gal4-LBD-mPPARβ/δ and UAS-luciferase constructs before being treated with compound overnight. The cells were lysed and relative luciferase activity was determined. Results were normalized to EtOH/H2O. n = 3. *P<0.05 Significantly different than the EtOH control and + is significantly different from theunoxidized counterpart. Inset: Oxidation of VLDL releases 13-S-HODE. VLDL was oxidized using 10 µm CuSO 4 for 72 h. Oxidized VLDL was then incubated with LpL for 1 h. An ELISA specific for 13-S-HODE (Assay Designs) was performed to quantify the amount hydrolyzed from oxvldl. Samples were prepared by diluting 1 and 10 µg of oxvldl with standard diluent to a final volume of 100 µl (n = 2). 86

100 87 The lipid peroxidation product 4-HNE is a PPARβ/δ activator. Based on the increased activity observed following CuSO 4 treatment of lipids, we suspected that 4- HNE or other aldehydic oxidation products formed via a Hock Cleavage mechanism from 13(S)-HODE or 13(S)-HpODE were PPARβ/δ ligands. A structure-activity relationship was established using a representative compound from each family of lipid peroxidation product, 2-alkenals, 4-hydroxy-2-alkenals, and ketoaldehydes (Fig. 2-2 A). Reporter assays clearly showed that 4-HNE enhanced the activity of the PPARβ/δ subtype while 4- HHE, 4-ONE, and trans-4,5-epoxy-2(e)-decenal did not activate this receptor (Fig. 2-2 B). Periods of oxidative stress initiate free-radical mediated oxidative damage of lipids, such as 13(S)-HODE. 4-ONE results in lipid peroxidation and damage, but did not activate this receptor, showing that an indirect oxidative stress response is not responsible for activating PPARβ/δ. The breakdown pathway from lipid to 4-HNE includes a hydroperoxy-derivative, 4-hydroperoxynonenal (4-HpNE), which is rapidly reduced to 4- HNE by cellular peroxidases. In fact, 4-HNE activated PPARβ/δ to a greater extent than its precursor (Fig. 2-2 C) and did so in a dose-dependent manner (Fig. 2-2 D, EC µm). Cellular toxicity was observed at concentrations higher than 30 µm and biological concentration of 4-HNE are known to be within this range following oxidative stress (23).

101 Figure 2-2: A. Structures of 4-HNE and other test molecules examined. B. PPARβ/δ is activated by 4-HNE. 3T3-L1 preadipocytes were transfected with Gal4-LBDmPPARβ/δ and UAS-luciferase constructs followed by treatment with each of the lipid peroxidation products or vehicle control for 12 h. Following the treatment, cells were lysed and luciferase activity was determined and normalized to EtOH. Treatment was with 25 µm 4-HHE, 4-HNE, 4-ONE, and trans-4,5-epoxy-2e-decenal (Decenal; Cayman) overnight. *P<0.05 compared to vehicle control. C. HpNE activates PPARβ/δ. Cells were treated with 25 µm 4-HpNE or 4-HNE for 12 h as described above. *P<0.05 compared to vehicle control. D. Dose response of PPARβ/δ activation by 4-HNE. PPARβ/δ activity is increased with the addition of 25 µm 4-HNE, while cellular toxicity is observed at higher concentrations. n = 8. *P<0.05 compared to vehicle control. 88

102 89 4-HNE induces the expression of PPARβ/δ target genes. In order to analyze if 4-HNE affected gene expression or toxicity in a PPARβ/δ-dependent manner, studies were performed using murine SV40-immortalized hepatocytes prepared from wild-type (MuSH WT) and PPARβ/δ -/- (MuSHβ/δ -/- ) mice (20). The expression of a PPARβ/δregulated gene, adipose differentiation-related protein (ADRP (16)), was up-regulated in a PPARβ/δ-dependent manner in cells treated with 13(S)-HODE and 4-HNE (Fig. 2-3 A). Importantly, the cells lacking PPARβ/δ were more susceptible to 4-HNE cytotoxicity than the wild-type cells (Fig. 2-3 B). The EC 50 value for MuSH WT cells (68 µm) was approximately 10-times higher than that observed for the MuSHβ/δ -/- cells (5 µm). Also, the MuSHβ/δ -/- cells showed overt toxicity at lower concentrations of sodium arsenite than their wildtype counterpart (Fig. 2-3 C) thus confirming the increased hepatotoxicity observed in vivo in the PPARβ/δ -/- mouse model (data not shown). To further examine the role of PPARβ/δ activity on conferring resistance to 4-HNE cytotoxicity, immortalized hepatocytes (MuSH WT) were pretreated with either a PPARβ/δ agonist tetradecylthioacetic acid (TTA) or PPAR antagonist GW9662 (at concentrations known to inhibit PPARβ/δ (24)) prior to 4-HNE administration. Pretreating hepatocytes with TTA resulted in a significant protection from 4-HNE-dependent toxicity (Fig. 2-3 D). Conversely, treating MuSH WT cells with GW9662 increased sensitivity to this toxic lipid (Fig. 2-3 E). This data clearly shows a role for PPARβ/δ in cytoprotection against at least one aldehydic toxic lipid (4-HNE) and a xenobiotic which mediates its toxicity at least in part via oxidative mediators (As). Also, this raises the possibility that one can augment PPARβ/δ activity in order to protect cells from the damaging affects of 4-HNE.

103 90 Gene expression affected by PPARβ/δ activation was examined by gene expression microarray. Of the 6800 unique genes on the arrays, 1001 gave a reliable signal in all three slides. The determination of statistically significant regulated genes was examined as we have previously reported (18, 20). Using this conservative measure of significant (p<0.05 and a two-fold change in gene expression), 41 genes were regulated in a PPARβ/δ manner in MuSH WT cells, in this case all were increased in levels relative to controls (Table 2-1). The biotransformation of 4-HNE to less toxic intermediates can be accomplished by oxidative (aldehyde dehydrogenase; ALDH), reductive (alcohol dehydrogenase; ADH), and conjugative (glutathione S-transferase; GST) pathways (23). Following gene expression microarray experiments (see Table 2-1), we found activation of PPARβ/δ increased mrna for Aldh3a1 (22-fold increase), Gstm3 (2.3-fold increase) and Gsto1 (2.0-fold increase). This suggests that PPARβ/δ activation is capable of increasing the detoxification of 4-HNE, although this will require formal confirmation.

104 Figure 2-3: A. 4-HNE regulates gene expression in a PPARβ/δ-dependent manner. MuSHWT and MuSHβ/δ / cells were treated with 13-S-HODE and 4-HNE overnight and the cells were lysed. Total RNA was extracted using Tri-Reagent and RT-PCR performed for ADRP and corrected for β-actin expression. n = 3. ANOVA with Fishers multiple comparisons test was performed in Minitab (State College, PA). Bars with different letters are significantly different (P<0.05). B. PPARβ/δ is crucial in ameliorating the toxic effects of 4-HNE. MuSHWT and MuSHβ/δ / hepatocytes were treated with 0, 0.1, 0.3, 1, 3, 10, or 25 µm 4-HNE for 24 h. Hepatocytes treated with concentrations of 50 µm or higher 4-HNE do not survive, regardless of genotype. The data are normalized for percentage viable cells, shown on the y-axis. n = 8. Representative of six independent experiments. C. PPARβ/δ / cells are more sensitive to arsenic-induced toxicity. MuSHWT and MuSHβ/δ / hepatocytes were treated to graded concentrations of sodium arsenite for 24 h. The data are normalized for percentage viable cells, shown on the y-axis. n = 5. Representative of two independent experiments. *P<0.05 compared with MuSH PPARβ/δ / hepatocytes. D. Activation of PPARβ/δ leads to protection from oxidative damage. MuSHWT cells were pretreated with vehicle or TTA (25 µm) for 6 h prior to treatment with 4-HNE. n = 5. *P<0.05 compared to vehicle-treated counterpart. E. Inhibition of PPARβ/δ leads to increased susceptibility to oxidative damage. MuSHWT cells were pretreated with vehicle or GW9662 (10 µm) for 6 h prior to treatment with 4-HNE. n = 5. *P<0.05 compared to vehicle-treated counterpart. 91

105 Table 2-1: Microarray analysis of gene expression: Genes statistically regulated by PPARβ/δ in MuSH hepatocytes. 92

106 93 Mapping potential 4-HNE / PPARβ/δ interactions. An interaction with a nuclear receptor and 4-HNE has not been described previously and was examined in more detail using molecular modeling and the coordinates of PPARβ/δ bound to eicosapentaenoic acid (Fig. 2-4 A (25)). PPARβ/δ bound to (S)-4-HNE (CPK model) in a conformation similar to the natural ligand (Fig. 2-4 B). However, a putative hydrogen-bonding interaction between His413 (stick model) and the 4-hydroxyl group of the 4-HNE ligand (CPK model) occurs (Fig. 2-4 C). As described above, PPARβ/δ was activated by 4- HpNE and 4-HNE but not 4-ONE, 4-HHE or decanol. Thus, the association between the hydroxyl group of 4-HNE and His413 is necessary for optimal receptor activation by this class of compound. In data not shown, the fatty acids nonenol and EPA activated PPARβ/δ to a similar extent as 4-HNE. This suggests that the interaction of carboxylic acids with PPARβ/δ proceeds by a different mechanism, one that would not require His413, but is still capable of activating the receptor. Several attempts were made at mutating His413 to either alanine or phenylalanine for a formal testing of the molecular modeling. However, the basal activity or that seen in the presence of PPARβ/δ ligands by these mutants was very low, suggesting that the His residue is critical for proper expression or folding. Although 4-HNE forms covalent adducts with cysteinyl, lysyl, and histidyl of target proteins, we were unable to show covalent interactions between 4-HNE and PPARβ/δ (Fig. 2-4 D and E), strongly suggesting that the association with PPARβ/δ is reversible.

107 Figure 2-4: A. X-ray structure of PPARβ/δ bound to eicosapentaenoic acid. Model is based on coordinates from [25]. B. A molecular model of PPARβ/δ bound to (S)-4- hydroxynonenal (CPK model) in a conformation similar to the natural ligand. C. Visualization of the putative hydrogen-bonding interaction between His413 (stick model) and the 4-hydroxyl group of the ligand (CPK model). D. 4-HNE activates PPARβ/δ via a direct and noncovalent interaction. Mouse His-tagged PPARβ/δ LBD was expressed in bacteria and purified using a nickel column and incubated with 4- HNE. Immunoblot analysis was used to determine 4-HNE-protein adducts. Lane 1, PPARβ/δ-LBD alone; Lane 2, PPARβ/δ-LBD plus EtOH; Lane 3, PPARβ/δ-LBD plus 4- HNE (100 µm); Lane 4, BSA plus 4-HNE (100 µm). E. Identical incubation as performed above with immunoblot analysis of PPARβ/δ. 94

108 Discussion The PPARs are ligand-activated transcription factors that sense a variety of lipophilic molecules and control gene expression to maintain cellular homeostasis. Fatty acids and their metabolites are known endogenous agonists of all PPARs, with PPARβ/δ exhibiting similar structural and geometric preference as PPARα, whereas PPARγ tends to prefer long-chain polyunsaturated fatty acids (26). PPARβ/δ agonists include linoleic acid, oleic acid, and arachidonic acids and EPA has been co-crystallized within the ligand binding domain of this nuclear receptor (25). Prostaglandin A1 (PGA1), PGD2 and PGD1 can activate PPARβ/δ in reporter assays (27). Similar to PPARα and γ, incubation of triglyceride rich lipoproteins with LPL results in the production of PPARβ ligands (28, 29). We have shown herein that several molecules present in oxvldl including 15- HETE, 13(S)-HODE and 4-HNE are agonists of PPARβ/δ. Activation of PPARα and PPARγ by 13(S)-HODE has been observed previously (30), although this molecule has been reported to be an inhibitor of PPARβ/δ in colon epithelial cells (31). Our data indicating 13(S)-HODE being a PPARβ/δ agonist in hepatocytes may suggest cell-type specific phenomena. To our knowledge, this is the first instance of the important toxic lipid 4-HNE being described as a PPAR agonist. Coupled with the fact that PPARβ/δ is expressed virtually ubiquitously and is known to be a regulatory of several important physiological pathways makes this observation intriguing. Examination of the PPARβ/δ null mice has shown a role for this receptor in several biological niches (32). There is mounting evidence that PPARβ/δ plays a role in ameliorating the hepatotoxic effects of a wide range of xenobiotics including arsenic and

109 96 acetaminophen (APAP). Interestingly, a similar hepatoprotective role has been noted for PPARα (33-35) and PPARγ (36) agonists, although the mechanism of this response is not clear. It has been known for some time that part of the liver injury caused by hepatoxicants such as As, CCl 4 or APAP is originated through free radical reactions and subsequent initiation of lipid peroxidation (reviewed in (11, 37)). This ultimately causes oxidative stress and plays a major role in the pathogenesis of both acute and chronic liver damage. For example, the enzymatic oxygenation of arachidonic acid by cyclooxygenase (COX), an enzyme commonly induced by hepatotoxicants, may cause the release of reactive oxygen species (ROS) during the enzymatic reduction of hydroperoxides. The release of ROS may initiate a chain of responses that results in lipid mediators including 4-HNE with subsequent protein and DNA damage. In addition to formation of ROS and lipid peroxides, COX- and lipoxygenase (LOX)-mediated fatty acid metabolism leads to the formation of various prostaglandins and other lipid mediators depending on the various isomerases and reductases present in a tissue-specific manner. These molecules are potent bioactive compounds in vivo (11) and may contribute directly or indirectly to liver damage. The amount of lipid substrate, the activity of phospholipase A2 (PLA2), COX, LOX, CYP as well as various antioxidant defense mechanisms are all dynamically regulated and will ultimately affect the damage caused by a hepatotoxicant. Thus, there are many means by which PPARβ/δ may be protecting the cell from oxidative damage, either by inhibiting the production of ROS and lipid intermediates or by increasing degradation of these molecules. Based on gene expression data performed herein and elsewhere (16), it appears that the most likely candidate would be PPARβ/δdependent metabolism and detoxification of lipid intermediates. The biotransformation of

110 97 4-HNE to less toxic intermediates can be accomplished aldehyde and alcohol dehydrogenases, glutathione S-transferases and perhaps other pathways (23). Following gene expression microarray experiments, we found activation of PPARβ/δ increased mrna for Aldh3a1, Gstm3 and Gsto1. Other microarray experiments have identified GSTA1 and ALDH9A1 as potential PPARβ/δ targets as well (16). The induction of Aldh3a1 and GSTo1 are particularly intriguing due to their known role in protection against lipid peroxidation and toxicity (23). Similar to the PPARβ/δ -/- mouse, mice that lack the transcription factor Nrf2 are more sensitive to the cytotoxic and genotoxic effects of foreign chemicals due to attenuated induction of Gsta1, Gstm1 and Gstm3 (38). Thus, it is possible that these two transcription factors are sharing a common protective mechanism of action. However, in addition to being directly involved in the regulation of expression of these detoxification enzymes, PPARβ/δ also senses, or responds to, the intracellular levels of the toxic intermediate via a feed-back control mechanism. 4-HNE is a marker of lipid peroxidation as well as a mediator of hepatotoxicity and a variety of other pathophysiological conditions including cancer, Alzheimer s disease and atherosclerosis (39). Target proteins that are inhibited by 4-HNE include the glucose transporter GLUT3, tau, the proteasome and IκB kinase (39) and 4-HNE causes a high frequency of G-to-T mutations in codon 249 in the p53 gene (40). Unlike the covalent association of 4-HNE with cysteinyl, lysyl, and histidyl of target proteins, the association with PPARβ/δ appears to be largely reversible, and hence a unique 4-HNEprotein interaction. As 4-HNE is readily conjugated with glutathione in cells, it is noteworthy to mention that a 4-HNE-glutathione adduct may be the actual ligand interacting with PPARβ/δ. In summary, our results show that 4-HNE, previously

111 98 regarded as a mutagen and toxic end-product of lipid peroxidation, is in fact a significant agonist of the nuclear receptor PPARβ/δ. This receptor, PPARβ/δ, when activated provides a feed-back regulation of gene expression that ameliorates toxicity to the liver. The discovery of 4-HNE as an endogenous ligand reveals a crosstalk of PPARβ/δ in cellular events involved in inflammation and energy homeostasis. Given the large number of individuals hospitalized due to chemically-induced liver damage, it is possible that activation of PPARβ/δ may be exploited for more effective means of protecting or treating hepatotoxicity as well as other diseases in which 4-HNE plays a causative role. 2.6 References 1. Lee, W. M.; Senior, J. R. Recognizing drug-induced liver injury: current problems, possible solutions. Toxicol Pathol 33: ; Lee, W. M. Acetaminophen and the U.S. Acute Liver Failure Study Group: l owering the risks of hepatic failure. Hepatology 40:6-9; Comporti, M. Lipid peroxidation and biogenic aldehydes: from the identification of 4-hydroxynonenal to further achievements in biopathology. Free Radic Res 28: ; Vanden Heuvel, J. P. Diet, fatty acids, and regulation of genes important for heart disease. Curr Atheroscler Rep 6: ; Khan, S. A.; Vanden Heuvel, J. P. Role of nuclear receptors in the regulation of gene expression by dietary fatty acids (review). J Nutr Biochem 14: ; Peters, J. M.; Vanden Heuvel, J. P. Peroxisomes, peroxisome proliferators and peroxisome proliferator-activated receptors (PPARs). In: Vanden Heuvel, J. P.; Perdew, G. H.; Mattes, W. B.; Greenlee, W. F., eds. Cellular and Molecular Toxicology. Amsterdam, The Netherlands: Elsevier Science; 2002: Vanden Heuvel, J. P.; Peters, J. M. Biology of the PPAR family of receptors. Comments Toxicol. 7: ; Vanden Heuvel, J. P. Peroxisome proliferator-activated receptors (PPARS) and carcinogenesis. Toxicol Sci 47:1-8; Vanden Heuvel, J. P. Peroxisome proliferator-activated receptors: a critical link among fatty acids, gene expression and carcinogenesis. J Nutr 129:575S-580S; 1999.

112 10. Burdick, A. D.; Kim, D. J.; Peraza, M. A.; Gonzalez, F. J.; Peters, J. M. The role of peroxisome proliferator-activated receptor-beta/delta in epithelial cell growth and differentiation. Cell Signal 18:9-20; Basu, S. Carbon tetrachloride-induced lipid peroxidation: eicosanoid formation and their regulation by antioxidant nutrients. Toxicology 189: ; Kawada, N.; Mizoguchi, Y.; Kobayashi, K.; Yamamoto, S.; Morisawa, S. Arachidonic acid metabolites in carbon tetrachloride-induced liver injury. Gastroenterol Jpn 25: ; Lopez-Parra, M.; Claria, J.; Planaguma, A.; Titos, E.; Masferrer, J. L.; Woerner, B. M.; Koki, A. T.; Jimenez, W.; Altuna, R.; Arroyo, V.; Rivera, F.; Rodes, J. Cyclooxygenase-1 derived prostaglandins are involved in the maintenance of renal function in rats with cirrhosis and ascites. Br J Pharmacol 135: ; Hartley, D. P.; Kolaja, K. L.; Reichard, J.; Petersen, D. R. 4-Hydroxynonenal and malondialdehyde hepatic protein adducts in rats treated with carbon tetrachloride: immunochemical detection and lobular localization. Toxicol Appl Pharmacol 161:23-33; Hartley, D. P.; Kroll, D. J.; Petersen, D. R. Prooxidant-initiated lipid peroxidation in isolated rat hepatocytes: detection of 4-hydroxynonenal- and malondialdehydeprotein adducts. Chem Res Toxicol 10: ; Tachibana, K.; Kobayashi, Y.; Tanaka, T.; Tagami, M.; Sugiyama, A.; Katayama, T.; Ueda, C.; Yamasaki, D.; Ishimoto, K.; Sumitomo, M.; Uchiyama, Y.; Kohro, T.; Sakai, J.; Hamakubo, T.; Kodama, T.; Doi, T. Gene expression profiling of potential peroxisome proliferator-activated receptor (PPAR) target genes in human hepatoblastoma cell lines inducibly expressing different PPAR isoforms. Nucl Recept 3:3; Chang, M.; Rao, M. K.; Reddanna, P.; Li, C. H.; Tu, C. P.; Corey, E. J.; Reddy, C. C. Specificity of the glutathione S-transferases in the conversion of leukotriene A4 to leukotriene C4. Arch Biochem Biophys 259: ; Tien, E. S.; Davis, J. W.; Vanden Heuvel, J. P. Identification of the CREBbinding protein/p300-interacting protein CITED2 as a peroxisome proliferatoractivated receptor alpha coregulator. J Biol Chem 279: ; Chou, J. Y. Establishment of rat fetal liver lines and characterization of their metabolic and hormonal properties: use of temperature-sensitive SV40 virus. Methods Enzymol 109: ; Tien, E. S.; Gray, J. P.; Peters, J. M.; Vanden Heuvel, J. P. Comprehensive gene expression analysis of peroxisome proliferator-treated immortalized hepatocytes: identification of peroxisome proliferator-activated receptor alpha-dependent growth regulatory genes. Cancer Res 63: ; Burdick, A. D.; Kim, D. J.; Peraza, M. A.; Gonzalez, F. J.; Peters, J. M. The role of peroxisome proliferator-activated receptor-beta/delta in epithelial cell growth and differentiation. Cell Signal; Michalik, L.; Desvergne, B.; Tan, N. S.; Basu-Modak, S.; Escher, P.; Rieusset, J.; Peters, J. M.; Kaya, G.; Gonzalez, F. J.; Zakany, J.; Metzger, D.; Chambon, P.; Duboule, D.; Wahli, W. Impaired skin wound healing in peroxisome proliferator- 99

113 activated receptor (PPAR)alpha and PPARbeta mutant mice. J Cell Biol 154: ; Awasthi, Y. C.; Sharma, R.; Cheng, J. Z.; Yang, Y.; Sharma, A.; Singhal, S. S.; Awasthi, S. Role of 4-hydroxynonenal in stress-mediated apoptosis signaling. Mol Aspects Med 24: ; Seimandi, M.; Lemaire, G.; Pillon, A.; Perrin, A.; Carlavan, I.; Voegel, J. J.; Vignon, F.; Nicolas, J. C.; Balaguer, P. Differential responses of PPARalpha, PPARdelta, and PPARgamma reporter cell lines to selective PPAR synthetic ligands. Anal Biochem 344:8-15; Xu, H. E.; Lambert, M. H.; Montana, V. G.; Parks, D. J.; Blanchard, S. G.; Brown, P. J.; Sternbach, D. D.; Lehmann, J. M.; Wisely, G. B.; Willson, T. M.; Kliewer, S. A.; Milburn, M. V. Molecular recognition of fatty acids by peroxisome proliferator- activated receptors. Mol Cell 3: ; Willson, T. M.; Brown, P. J.; Sternbach, D. D.; Henke, B. R. The PPARs: from orphan receptors to drug discovery. J Med Chem 43: ; Yu, K.; Bayona, W.; Kallen, C. B.; Harding, H. P.; Ravera, C. P.; McMahon, G.; Brown, M.; Lazar, M. A. Differential activation of peroxisome proliferatoractivated receptors by eicosanoids. J Biol Chem 270: ; Ziouzenkova, O.; Perrey, S.; Asatryan, L.; Hwang, J.; MacNaul, K. L.; Moller, D. E.; Rader, D. J.; Sevanian, A.; Zechner, R.; Hoefler, G.; Plutzky, J. Lipolysis of triglyceride-rich lipoproteins generates PPAR ligands: evidence for an antiinflammatory role for lipoprotein lipase. Proc Natl Acad Sci U S A 100: ; Chawla, A.; Lee, C. H.; Barak, Y.; He, W.; Rosenfeld, J.; Liao, D.; Han, J.; Kang, H.; Evans, R. M. PPARdelta is a very low-density lipoprotein sensor in macrophages. Proc Natl Acad Sci U S A 100: ; Nagy, L.; Tontonoz, P.; Alvarez, J. G.; Chen, H.; Evans, R. M. Oxidized LDL regulates macrophage gene expression through ligand activation of PPARgamma. Cell 93: ; Shureiqi, I.; Jiang, W.; Zuo, X.; Wu, Y.; Stimmel, J. B.; Leesnitzer, L. M.; Morris, J. S.; Fan, H. Z.; Fischer, S. M.; Lippman, S. M. The 15-lipoxygenase-1 product 13-S-hydroxyoctadecadienoic acid down-regulates PPAR-delta to induce apoptosis in colorectal cancer cells. Proc Natl Acad Sci U S A 100: ; Peters, J. M.; Lee, S. S.; Li, W.; Ward, J. M.; Gavrilova, O.; Everett, C.; Reitman, M. L.; Hudson, L. D.; Gonzalez, F. J. Growth, adipose, brain, and skin alterations resulting from targeted disruption of the mouse peroxisome proliferator-activated receptor beta(delta). Mol Cell Biol 20: ; Mehendale, H. M. PPAR-alpha: a key to the mechanism of hepatoprotection by clofibrate. Toxicol Sci 57: ; Manautou, J. E.; Hoivik, D. J.; Tveit, A.; Hart, S. G.; Khairallah, E. A.; Cohen, S. D. Clofibrate pretreatment diminishes acetaminophen's selective covalent binding and hepatotoxicity. Toxicol Appl Pharmacol 129: ; Nguyen, K. A.; Carbone, J. M.; Silva, V. M.; Chen, C.; Hennig, G. E.; Whiteley, H. E.; Manautou, J. E. The PPAR activator docosahexaenoic acid prevents 100

114 acetaminophen hepatotoxicity in male CD-1 mice. J Toxicol Environ Health A 58: ; Kon, K.; Ikejima, K.; Hirose, M.; Yoshikawa, M.; Enomoto, N.; Kitamura, T.; Takei, Y.; Sato, N. Pioglitazone prevents early-phase hepatic fibrogenesis caused by carbon tetrachloride. Biochem Biophys Res Commun 291:55-61; James, L. P.; Mayeux, P. R.; Hinson, J. A. Acetaminophen-induced hepatotoxicity. Drug Metab Dispos 31: ; Chanas, S. A.; Jiang, Q.; McMahon, M.; McWalter, G. K.; McLellan, L. I.; Elcombe, C. R.; Henderson, C. J.; Wolf, C. R.; Moffat, G. J.; Itoh, K.; Yamamoto, M.; Hayes, J. D. Loss of the Nrf2 transcription factor causes a marked reduction in constitutive and inducible expression of the glutathione S-transferase Gsta1, Gsta2, Gstm1, Gstm2, Gstm3 and Gstm4 genes in the livers of male and female mice. Biochem J 365: ; Uchida, K. 4-Hydroxy-2-nonenal: a product and mediator of oxidative stress. Prog Lipid Res 42: ; Hu, W.; Feng, Z.; Eveleigh, J.; Iyer, G.; Pan, J.; Amin, S.; Chung, F. L.; Tang, M. S. The major lipid peroxidation product, trans-4-hydroxy-2-nonenal, preferentially forms DNA adducts at codon 249 of human p53 gene, a unique mutational hotspot in hepatocellular carcinoma. Carcinogenesis 23: ;

115 Chapter 3 Peroxisome Proliferator-Activated Receptor β/δ (PPARβ/δ) and B-cell lymphoma 6 (BCL-6) regulate inflammatory signaling in human pancreatic cancer cells. 3.1 Abstract Peroxisome proliferator-activated receptor β/δ is a member of the nuclear receptor superfamily and a ligand-activated transcription factor that is implicated in the regulation of the inflammatory response via induction of anti-inflammatory target genes and ligandinduced disassociation from the transcriptional repressor B-cell lymphoma 6 (BCL-6). Chronic inflammation of the pancreas increases the risk for pancreatic cancer by approximately 16-fold. The aim of the present study was to determine if PPARβ/δ antiinflammatory signaling was active in pancreatic cancer cells of ductal origin. Forced PPARβ/δ over-expression in Miapaca-2 cells inhibited basal and TNFα-induced NF-κB activity in reporter assays. The selective PPARβ/δ activator GW induced a physical interaction between pm-pparβ/δ and the p50 sub-unit of NF-κB in mammalian 2-hybrid assays and suppressed TNFα-induced NF-κB luciferase. Short-hairpin RNAi knock-down of PPARβ/δ attenuated the GW effect on NF-κB luciferase, while knock-down of BCL-6 significantly enhanced TNFα-induced NF-κB activity. PPARβ/δ activation induced the expression of several anti-inflammatory genes in a PPARβ/δ-

116 103 dependent manner, and GW inhibited MCP-1 promoter-driven luciferase in a BCL-6-dependent manner. Furthermore, pro-inflammatory gene expression was suppressed in a BCL-6-dependent manner as measured by RT-PCR. Because tumor mass is comprised in large part of macrophages and infiltrating immune cells, we sought to determine if GW treatment affected pancreatic cancer cell-macrophage cross-talk. Indeed, GW treated pancreatic cancer cell media suppressed pro-inflammatory gene expression in differentiated THP-1 macrophages. Finally, GW treated pancreatic cancer cell conditioned media significantly reduced the invasive capabilities of THP-1 cells to invade a basement membrane compared with control. These results demonstrate that PPARβ/δ and BCL-6 regulate anti-inflammatory signaling in human pancreatic cancer cells by inhibiting NF-κB and pro-inflammatory gene expression, and via induction of anti-inflammatory target genes. Moreover, activation of PPARβ/δ may prove useful in controlling pancreatic inflammation and immune infiltration in individuals with pancreatitis. 3.2 Introduction Pancreatic cancer is an extremely aggressive malignancy with poor prognosis, owing mainly to the lack of identifiable treatable early symptoms (1). Early symptoms include jaundice, pain and weight loss, and are so non-specific as to be easily confused for other diseases. Usually, diagnosis of pancreatic cancer is not made until the malignancy has metastasized to other organs, specifically the liver, peritoneum and the lungs, and, in rarer cases, the brain and bone (2). The average survival rate is a mere six

117 104 months following diagnosis. Treatment is typically aimed at improving quality of life, as pancreatic tumors are highly resistant to chemotherapeutic regimens, and only 10-20% of pancreatic tumors are resectable (1). Several risk factors for pancreatic cancer have been identified, including age, a high-fat diet, and diabetes mellitus. The most easily controlled risk factor, however, remains tobacco use, which accounts for approximately one quarter of all cases of pancreatic cancer and increases an individual s risk of developing the malignancy by 2- to 3-fold (3). Chronic pancreatitis increases the risk of developing pancreatic cancer by approximately 16-fold, accelerating cancer progression in transgenic models (4). In fact, Carriere et al. found that brief episodes of acute pancreatitis resulted in rapid tumor progression in a mouse model expressing oncogenic K-ras, concluding that pancreatic inflammatory insult may initiate a cascade of events ultimately resulting in pancreatic cancer (4). In each of the risk factors described above, the resulting inflammatory state is a key mediator in driving pancreatic neoplasia in all steps of carcinogenesis (5). A highly inflammatory microenvironment, for example, can promote initiation of carcinogenesis via production of free radicals and oxidative stress mediators that in turn induce genetic abnormalities, ultimately resulting in loss of function of key cancer-related genes (6). Furthermore, the inflammatory response exacerbates tumor growth and spread, likely mediated by the transmission of growth and proliferative signals to both normal and transformed cells (7). Peroxisome proliferator-activated receptor-β/δ (PPARβ/δ) is a member of the nuclear receptor superfamily and a ligand-activated transcription factor that controls several cellular functions. Key among these is the regulation of the inflammatory response (8-11). The PPARα and PPARγ subtypes have generated much interest in their

118 105 abilities to influence the inflammatory response (12). PPARβ/δ, however, differs from the other isoforms in its ubiquitous expression and association with the transcriptional repressor, BCL-6 (11, 13). Unliganded PPARβ/δ resides in the nucleus in a repressive state where it physically associates with BCL-6, inhibiting the latter s anti-inflammatory actions. Upon ligand binding, PPARβ/δ dissociates from BCL-6, dimerizes with retinoid X receptor (RXR) and directly regulates expression of target genes while BCL-6 becomes available to suppress pro-inflammatory gene expression (14). The precise mechanisms by which BCL-6 participates in the inflammatory response may differ by tissue type (15, 16). Further mechanisms by which PPARβ/δ agonists inhibits the inflammatory response include the inhibition of NF-κB signaling as well as the suppression of oxidative stress via direct induction of target genes (14, 15). Although the precise anti-inflammatory mechanisms of PPARβ/δ-ligands are still being elucidated, they have proven potent repressors of inflammatory signaling in both cell and animal models (14). Inflammation of the pancreas, particularly chronic or hereditary pancreatitis, is a major risk factor for developing pancreatic cancer. The PPARβ/δ-BCL-6 antiinflammatory pathway is not active in pancreatic beta cells, however, since those cells lack BCL-6 protein (17). To date, very little evidence exists describing the precise role(s) of PPARβ/δ in the exocrine pancreas. To that end, we sought to determine what role, if any, PPARβ/δ and BCL-6 play in regulating the inflammatory response in pancreatic endothelial duct cells using the PPARβ/δ-specific agonist GW and shrnas to knock-down expression of these inflammatory regulators in the human pancreatic cancer cell lines Miapaca-2 and BxPc-3. Our observations show that GW mediated antiinflammatory signaling is active in two pancreatic cancer cell lines, and may proceed via

119 106 the inhibition of NF-κB activity and TNFα-induced pro-inflammatory gene expression in both PPARβ/δ- and BCL-6-dependent fashions. GW also induced expression of several anti-inflammatory PPARβ/δ target genes. Furthermore, conditioned media experiments using various shrnas implicated PPARβ/δ and BCL-6 as regulators of inflammatory gene expression and macrophage recruitment in differentiated THP-1 cells. 3.3 Materials and Methods Cells and Reagents. Human pancreatic cancer cells, Miapaca-2 (COX-2 negative, CRL- 1420) and BxPc-3 (COX-2 positive, CRL-1687) were purchased from the ATCC (Manassas, VA) and cultured in high-glucose DMEM containing 10% FBS. Human embryonic kidney 293 cells were cultured in DMEM containing 10% FBS. THP-1 cells were cultured in RPMI 1640 media supplemented with 10% FBS. All cell media also contained 100 units each of penicillin and streptomycin, and cells were cultured in a humidified atmosphere at 37 C containing 5% CO2. All media components and fetal bovine serum (FBS) were purchased from Gibco BRL/Life Technologies (Carlsbad, CA). GW (GW), used as the positive control for PPARβ/δ, and Phorbol 12-myristate 13-acetate (PMA), used to differentiate THP-1 cells, were purchased from Sigma Chemical Company. The 2.8 kb mouse MCP-1 (accession # U12470) promoter fragment cloned into the luciferase reporter vector pgl3-basic (Promega) was kindly provided by Dr. Ronald Evans (Salk Institute for Biological Studies, La Jolla, CA) and is described here (11). Transfection control plasmids prl-tk and prlcmv were purchased from

120 107 Promega (Madison, WI). Recombinant human TNFα was purchased from Invitrogen (Carlsbad, CA) and reconstituted in nanopure water. MISSION shrna bacterial glycerol stocks targeted against human PPARβ/δ, Bcl6, IL-1Ra, as well as the nontargeting vector, were purchased directly from Sigma-Aldrich. High Capacity cdna Archive Kit and ABI7300 Real-time PCR System were purchased from Applied Biosystems (Foster City, CA). The ppackh1 packaging plasmids and the pcdna3.1/pparβ/δ-flag plasmid were kindly provided by Dr. Curtis J. Omiecinski (Penn State University). CytoSelectTM 96-well Cell Invasion Assay (Basement Membrane, Fluorometric Format) was purchased from Cell Biolabs, Inc. (San Diego, CA) and used according to the manufacturer s instructions. NF-κB reporter assays. Miapaca-2 cells were seeded at a density of 7.5 x 105 cells in 10-cm tissue culture plates. Cells were then transiently transfected with 9 µg pnf-κbluciferase and 1 µg prlcmv using LipofectAMINE (Invitrogen) reagent for 6 h and allowed to recover overnight. Cells were challenged with the indicated treatments 24 hours post-transfection, and NF-κB luciferase activity was measured using the Dual Luciferase Reporter Assay system (Promega, normalized to control luciferase activity). In experiments where PPARβ/δ was over-expressed, cells were transfected with 5 µg hpparβ/δ-flag, 4 µg pnf-κb-luciferase and 1 µg prlcmv. Mammalian two-hybrid assay. Mammalian two-hybrid studies were performed using the Matchmaker Mammalian-2-Hybrid system (Clontech). Miapaca-2 cells were plated in 24-well plates and allowed to recover overnight. Transfection was carried out with

121 108 LipofectAMINE reagent according to the protocol of the manufacturer. Each well was transfected with 200 ng of pm/pparβ/δ or empty vector, 200 ng of pvp16, pvp16/p50 or pvp16/p65, 100 ng each of pfr-luciferase and prl-tk. Transfected cells were treated with 500 nm GW for 6 h and assayed for luciferase activity. Relative luciferase activity was corrected using the internal transfection control (prl-tk) and the Dual Luciferase Kit (Promega). Isolation of total RNA and Real-time Quantitative RT-PCR. Total RNA was isolated from Miapaca-2 and BxPc-3 cells using Tri-Reagent and the manufacturer s recommended protocol (Sigma). One µg of total RNA was reverse transcribed using the High Capacity cdna Archive Kit (Applied Biosystems, Foster City, CA). PCR primers for quantitative real-time RT-PCR were designed based on published sequences in GenBank and are shown in Table 3-1. The housekeeping gene β-actin was used to normalize all the tested genes. The data shown are representative of three independent experiments with triplicate samples. MCP-1 reporter assay. Miapaca-2 cells transiently expressing non-targeting control, BCL-6 or PPARβ/δ shrna were plated in 10-cm tissue culture plates as described. Cells were transiently transfected with 9 µg mmcp-1-luciferase and 1 µg prlcmv for 6 hours and allowed to recover overnight. Cells were then challenged with the indicated treatments 24 hours post-transfection and MCP-1 promoter-driven luciferase was assayed and corrected using the internal transfection control prlcmv.

122 109 THP-1 cell differentiation with PMA. Differentiation was achieved by resuspending THP-1 cells at a density of 2 x 105 cells/ml in serum-free RPMI 1640 media supplemented with 100 nm PMA for 24 hours. Cells were then allowed to recover in media containing 10% FBS for a further 24 hours before use in experiments. Lentiviral shrna infection. HEK-293 cells were grown to confluency in 10-cm tissue culture plates under the conditions described above. The cells were then transiently transfected with 4.6 µg of either non-targeting shrna, or shrnas targeted against human PPARβ/δ, Bcl6, or IL-1Ra, as well as 2.4 µg each of ppackh1 packaging plasmids, using Lipofectamine Cells were transfected for 6 h and allowed to recover overnight in normal media. Fresh media was added the following morning, and pseudoviral supernatant was generated for 72 h. Supernatant was then harvested and passed through a 0.4 µm filter under sterile conditions. Polybrene (Millipore, Billerica, MA) was then added to a final concentration of 5 µg per ml and the pseudoviral supernatant was then added directly to target cells for 6 h. Infected cells were allowed to recover overnight following the addition of 6 ml complete media and knock-down of target genes was assessed by RT-PCR 48 h post-infection. Conditioned media experiments. Control or knock-down Miapaca-2 cells were plated at a density of 7.5 x 105 cells in 10-cm tissue culture plates. Following overnight recovery the cells were challenged with 1 ng/ml TNFα with or without 500 nm GW for 24 hours. Conditioned media was collected and centrifuged at 200 x g at 20 C for 10 min and any unused media was stored at -80 C. The same volume of normal

123 110 media containing TNFα with or without 500 nm GW was prepared at the start of the experiment and was used as control media. For gene expression assays, undiluted conditioned media from control or knock-down Miapaca-2 cells was added directly to THP-1 cells and the cells were incubated for 24 hours. Conditioned media was removed the following day and total RNA was isolated and reverse-transcribed as described above. Cell migration assay. Cell migration assays were performed using the CytoSelectTM 96-Well Cell Invasion Assay (Basement Membrane, Fluormetric Format) according to the manufacturer s instructions. Briefly, the basement membrane was allowed to reach room temperature for 30 minutes, and rehydrated using warm, serum-free DMEM. THP- 1 cells were then seeded into each well at a density of 2 x 106 cells/ml in serum-free media. Conditioned media, as well as control media, (DMEM containing 10% FBS, along with TNFα with or without GW501516) was added to the feeder tray to act as a chemoattractant and the entire apparatus was placed in an incubator at 37 C containing 5% CO2 for 24 h. CyQuant GR dye/lysis buffer solution was added to the invading cells following completion of the assay and the resulting mixture was incubated at room temperature for 20 minutes. Invading cells were quantified by reading the fluorescence at 480 nm/520 nm. All measurements were performed in triplicate. Statistical Analysis. Quantitative data are presented as mean±sem. ANOVA with p- value<0.05 was used to determine whether differences among variables were significant. Normality was checked using Anderson-Darling test, and the General linear model, followed by the Tukey post hoc test to analyze differences between treatments. All data

124 111 analyses were performed by MiniTAB Ver.14 (MiniTAB, State College, PA) or JMP (SAS Institute, Cary, NC) and data were plotted by Prism 5.01 (GraphPad Software, San Diego, CA). Table 3-1: List of Real-time PCR primers used in this study. Gene Forward Primer Reverse Primer IL-1Ra GGGAACTTTGCACCCAACAT TTGGCAGGTACTCAGCGAATG TGF-β AGGTCCTTGCGGAAGTCAATG CTATTGCTTCAGCTCCACGGA SOD-1 TGCTTCCCCACACCTTCACTGGT ATGGCGACGAAGGCCGTGTG FGF-21 CGCTGGCACAGGAACCTGGA ACCAGAGCCCCGAAAGTCTCCT MCP-1 GGACGCATTTCCCCAGTACA CCGAGAACGAGATGTGGACA MCP-3 ATGAGGTAGAGAAGGGAGGAGCAT CAAACTGGACAAGGAGATCTGTGC TNFα TGGATGTTCGTCCTCCTCACA ATCAATCGGCCCGACTATCTC IL-1β TCCTTAGTCCTCGGCCAAGAC GTGCCATGGTTTCTTGTGACC IL-6 CCGTCGAGGATGTACCGAATT GCCACTCACCTCTTCAGAACG COX-2 CGGTGTTGAGCAGTTTTCTCC AAGTGCGATTGTACCCGGAC

125 Results PPARβ/δ over-expression inhibits basal and TNFα-induced NF-κB activity. Inhibition of NF-κB activity is one of the ways in which PPARβ/δ exerts antiinflammatory actions (14). To evaluate the extent by which PPARβ/δ can interfere with NF-κB activity in human pancreatic cancer cells, Miapaca-2 cells were transfected with a NF-κB-response element-luciferase construct along with either pcdna3.1/pparβ/δ- FLAG or empty vector. In MiaPaca-2 cells, forced over-expression of PPARβ/δ significantly reduced basal (unstimulated) NF-κB reporter activity. Treatment with 1 ng/ml TNFα induced NF-κB reporter activity approximately three-fold in control cells, and this effect was significantly diminished in MiaPaca-2 cells with augmented PPARβ/δ (Fig. 3-1 A). This effect was observed in the absence of a PPARβ/δ ligand, suggesting that under these conditions PPARβ/δ associates with and suppresses NF-κB activity in human pancreatic cancer cells directly and not via downstream signaling. PPARβ/δ physically interacts with the p50 subunit of NF-κB in mammalian 2- hybrid assays. PPARβ/δ physically interacts with the p65 subunit of NF-κB in cardiac cells (18), and similar effects have been shown for PPARα (19). In Miapaca-2 cells mammalian 2-hybrid assays were carried out to map the interaction between PPARβ/δ and NF-κB. Cells were transfected with pm/pparβ/δ or empty vector, along with pvp16/p50, pvp16/p65 or empty vector and luciferase activity was assayed. Treatment with GW significantly increased Gal4-response element reporter activity in cells co-expressing pm/pparβ/δ and pvp16/p50 compared with control, but not in cells

126 expressing the p65 subunit (Fig. 3-1 B). This suggests that PPARβ/δ physically interacts with the NF-κB ciomplex via the p50 subunit. 113 Ligand activation of PPARβ/δ reduces NF-κB reporter activity in Miapaca-2 cells. NF-κB is a master regulator of the inflammatory response in cancer cells (20). To determine if the GW induced interaction of PPARβ/δ with NF-κB could influence NF-κB activity in human pancreatic cancer cells, Miapaca-2 cells expressing shrnas targeted against BCL-6 or PPARβ/δ were transfected with a NF-κB-response element luciferase construct and treated with TNFα in the presence or absence of GW Indeed, in control shrna infected Miapaca-2 cells GW treatment significantly reduced iboth basal and TNFα-stimulated NF-κB-luciferase activity (Fig. 3-1 C). Reporter activity was significantly elevated in BCL-6 knock-down cells following TNFα stimulation, and GW slightly reduced this effect. Interestingly, GW treatment did not significantly alter the basal NF-κB-luciferase activity in BCL-6 knockdown cell line. In PPARβ/δ knock-down Miapaca-2 cells, both basal and TNFα-induced NF-κB reporter activity were unaffected by GW treatment. Taken together, these results indicate that the observed GW reduction in NF-κB activity may require both PPARβ/δ as well as BCL-6, with the former providing ligand-sensitivity and the latter affecting reporter activity.

127 114 A Relative NF$B-luciferase Reporter Activity DMSO TNF# Empty hppar-!/"-flag * * B GW Activation Fold Change a pm plasmid PPAR!/" Empty VP16 Plasmid Empty p50 p65 p50 p65 b a c c C Relative NF$B-luciferase Reporter Activity d,e e c d,e d,e d,e a b d,e e DMSO 500nM GW ng/mL TNF# TNF# + GW c c,d Non-targeting shrna hbcl6 shrna hppar!/" shrna Figure 3-1: Effects of PPARβ/δ, and ligand-activation of PPARβ/δ, on NF-κB activity. A. Over-expression of PPARβ/δ significantly reduces basal and TNFα-induced NF-κB reporter activity. Miapaca-2 cells were transiently transfected with NF-κB-luciferase with or without pcdna3.1/pparβ/δ-flag for 24 hours before challenge with TNFα. Luciferase activity was assayed and corrected for transfection efficiency. *P<0.05 B. PPARβ/δ interacts with the p50 domain of NF-κB in a mammalian 2-hybrid assay. Miapaca-2 cells were transfected with pm-pparβ/δ or empty vector, and either pvp16/p50, pvp16/p65 or pvp16/empty as indicated. Cells were treated with either DMSO or 500 nm GW hours post-transfection and luciferase activity was assayed and corrected for transfection efficiency. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test. C. Effect of PPARβ/δ and BCL-6 knock-down on NF-κB activity in Miapaca-2 cells. Miapaca-2 cells transiently expressing shrnas targeted against PPARβ/δ or BCL-6 were transfected with NF-κBluciferase. Cells were then challenged with the indicated treatments and luciferase activity was assayed and corrected for transfection efficiency. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

128 115 GW induces the expression of several anti-inflammatory genes in a PPARβ/δ-dependent manner. Ligand-dependent activation of PPARβ/δ affects expression of anti-inflammatory mediators in several cell lines (15, 21, 22). To determine if GW treatment induced expression of anti-inflammatory genes, and to determine if the anti-inflammatory effects depended upon the level of expression of PPARβ/δ, Miapaca-2 cells expressing shrnas targeted against BCL-6 and PPARβ/δ were treated with GW or vehicle and gene expression was investigated. GW treatment significantly increased the mrna of IL-1 receptor antagonist (IL- 1Ra), TGF-β, SOD-1 and fibroblast growth factor 21 (FGF-21) in control cells and in cells expressing BCL-6 shrna. In Miapaca-2 cells expressing PPARβ/δ shrna, however, GW failed to induce expression of these target genes, indicating that the anti-inflammatory effects of GW are mediated in part through the direct induction of target genes by PPARβ/δ (Fig. 3-2). The PPARβ/δ-dependent induction of the detoxification genes examined in Chapter 2 (Gstm3, Gsto1 and Aldh3a1) was also examined in Miapaca-2 cells (data not shown), although the mrna levels of these genes were not detectable by RT-PCR.

129 116 IL-1Ra TGF-! SOD-1 FGF-21 Normalized mrna Expression shrnai, Trt b b b a a b Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW b b a a a,b b b b b b b b b Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW a a Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW a a b Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Figure 3-2: GW treatment contributes to the anti-inflammatory actions of PPARβ/δ by increasing expression of target genes. Miapaca-2 cells were transiently infected with lentiviral-mediated shrnas targeted against PPARβ/δ or BCL-6. Cells were then treated with 500 nm GW or vehicle for 24 hours. Gene expression was determined by qrt-pcr and expressed as fold-induction following normalization to β-actin. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

130 117 Inhibition of TNFα-induced pro-inflammatory gene expression by GW is mediated by BCL-6. The anti-inflammatory properties of PPARβ/δ activators are also mediated by BCL-6 in vitro (11, 23) and in vivo (23). Known BCL-6 target genes include MCP-1 (11), MCP-3 and MCP-5 (13) and VCAM-1 (15), to name a few. Further evidence for BCL-6 being the major transcriptional regulator of MCP-1 expression is provided by the fact that no functional PPARβ/δ response element could be found in the 2.8kb 5 -regulatory region of this gene (11). Miapaca-2 cells were infected with shrnas targeting BCL-6 or PPARβ/δ, as well as the 2.8 kb mouse MCP-1 promoter -luciferase reporter vector, and challenged with TNFα to induce the inflammatory response. TNFα increased MCP-1 promoter activity in control cells, while treatment with GW significantly reduced both basal and TNFα-induced reporter activity (Fig. 3-3 A). To further investigate the regulation at the MCP-1 promoter, BCL-6 and PPARβ/δ shrnas were used. Miapaca-2 cells expressing BCL-6 shrna showed significantly higher basal and TNFα-induced MCP-1 promoter activity compared with control cells, which GW treatment failed to attenuate. In Miapaca-2 cells expressing shrna targeted against PPARβ/δ, however, basal MCP-1 promoter activity was significantly reduced compared with control, despite GW treatment. Administration of TNFα significantly induced MCP-1 luciferase activity in the PPARβ/δ shrna infected cells, but not to as great an extent as in control cells, and addition of GW reduced reporter activity (Fig. 3-3 A). To assess whether or not the repressive effects of GW and BCL-6 on the MCP-1 promoter translated into reduced mrna levels, real-time PCR was performed

131 118 using total RNA isolated from control Miapaca-2 cells as well as cells expressing the indicated shrnas. Indeed, GW activation significantly reduced the mrna expression of MCP-1, MCP-3, TNFα and IL-1β in control cells, and this effect was not observed in Miapaca-2 cells expressing shrna directed against BCL-6 (Fig. 3-3 B). Furthermore, the mrna expression of these genes was increased in BCL-6 knock-down cells compared with control, with or without GW treatment. Consistent with previous reports, cells expressing PPARβ/δ shrna exhibited lower mrna levels of these pro-inflammatory markers despite TNFα treatment (11). Taken together, these results further confirm that the anti-inflammatory properties of PPARβ/δ-specific ligands are mediated in part via BCL-6 repression of target gene expression.

132 119 A Relative Mcp-1-luciferase Reporter Activity c,d d b d c c a a d d DMSO 500nM GW ng/mL TNF! TNF! + GW c,d d Non-targeting shrna hbcl6 shrna hppar"/# shrna

133 120 B MCP-1 MCP-3 TNF# IL-1! Normalized mrna Expression shrnai, Trt a b a b b a b Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW b a c c a c Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW b a b,c c a c Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW a a a b b b Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Figure 3-3: The transcriptional repressor BCL-6 contributes to the anti-inflammatory actions of GW by suppressing target gene expression. A. Effects of PPARβ/δ and BCL-6 knock-down on TNFα-induced MCP-1 promoter-driven luciferase. Miapaca-2 cells transiently expressing the indicated shrnas were transfected with pgl3/mmcp-1 promoter luciferase and treated with vehicle, 500 nm GW or 1 ng/ml TNFα with or without GW for 24 hours before luciferase activity was assayed. B. Repression of TNFα-induced pro-inflammatory target genes by BCL-6. Miapaca-2 cells expressing the indicated shrnas were treated with TNFα with or without GW for 24 hours. Gene expression was determined by qrt-pcr and expressed as fold-induction following normalization to β-actin. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

134 121 GW exerts anti-inflammatory effects in human pancreatic cancer cells expressing COX-2. The human pancreatic cancer cell line Miapaca-2 does not express COX-2 or IL-6, and since IL-6 is a PPARβ/δ-regulated target in adipocytes (24) as well as in mouse macrophages (23), we next sought to determine the effects of PPARβ/δ activation in a COX-2-positive human pancreatic cancer cell line. BxPc-3 cells were transiently infected with control, BCL-6- or PPARβ/δ-targeting shrna and stimulated with TNFα with or without GW As in Miapaca-2 control cells, TNFα induced expression of IL-6 and COX-2 mrna, but GW only reduced mrna levels of IL- 6 with a negligible effect on COX-2 expression (Fig. 3-4). In both cases, knock-down of BCL-6 increased the inflammatory response that was not significantly affected by PPARβ/δ activation. Knock-down of PPARβ/δ resulted in significantly lower expression of IL-6 despite PPARβ/δ activation, indicating that the GW regulation is effected via BCL-6, while COX-2 expression was only slightly lower.

135 122 IL-6 COX-2 Normalized mrna Expression shrnai, Trt a a,b b,c c c c Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW a a,b b b b b Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Figure 3-4: Effects of PPARβ/δ and BCL-6 knock-down in COX-2 positive human pancreatic cancer cells. BxPc-3 cells transiently expressing the indicated shrnas were challenged with TNFα with or without GW for 24 hours. Gene expression was determined by qrt-pcr and expressed as fold-induction following normalization to β-actin. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

136 123 Human pancreatic cancer cell conditioned media affects gene expression in differentiated THP-1 macrophages. Tumor-associated macrophages secrete gene products that contribute to an immunosuppressive tumor microenvironment that allows the malignancy to escape immune surveillance (25), and are implicated in the progression of cancer by promoting tumor angiogenesis, growth and metastasis (26). To determine if conditioned media from human pancreatic cancer cells could influence gene expression in macrophages, Miapaca-2 cells expressing shrnas targeted against either BCL-6, PPARβ/δ, IL-1Ra or control were stimulated with TNFα in the presence of absence of GW for 24 hours, and conditioned media was added directly to PMAdifferentiated THP-1 cells for 24 hours and gene expression was examined. Unconditioned media (media placed in incubator in absense of cells) and media conditioned by untransfected Miapaca-2 cells were used as a control. Our results show that significant repression of the pro-inflammatory mediators MCP-1, MCP-3, TNFα, IL- 1β and IL-6, but not COX-2, in THP-1 macrophages was effected using both control media and conditioned media from control shrna-expressing Miapaca-2 cells (Fig. 3-5). When BCL-6 was knocked-down in Miapaca-2 cells, conditioned media from TNFαstimulated cells resulted in strong increases in pro-inflammatory mediators in THP-1 cells, particularly TNFα and IL-6. GW treatment of the MiaPaca cells was able to significantly reduce this effect in THP-1 cells. Conversely, conditioned media from PPARβ/δ knock-down Miapaca-2 cells did not stimulate the production of TNFα or IL-6 to as great an extent as control (Fig. 3-5, C and E). Surprisingly, knock-down of IL-1Ra in Miapaca-2 cells did not significantly alter TNFα-induced gene expression in THP-1

137 124 cells compared with control media or control shrna. MCP-1 and MCP-3, however, were repressed in THP-1 macrophages following the addition of conditioned media from GW treated IL-1Ra knock-down Miapaca-2 cells, suggesting that IL-1Ra does not play a significant role in the cross-talk between human pancreatic cancer cells and macrophages.

138 125 A MCP-1 mrna Expression shrnai a,b a,b c,d d Ctrl media Ctrl a Bcl6 d b,c IL-1Ra d 0.1 ng/ml TNF# TNF# + GW b,c,d PPAR!/" d B MCP-3 mrna Expression shrnai a,b Ctrl media c a,b Ctrl c a Bcl6 c a,b IL-1Ra c b,c PPAR!/" c C TNF# mrna Expression shrnai b,c Ctrl media e b Ctrl d,e a Bcl6 b b,c,d c,d,e d,e e IL-1Ra PPAR!/" D E F IL-1! mrna Expression shrnai a,b Ctrl media c a,b Ctrl c a Bcl6 c b,c IL-1Ra c b,c PPAR!/" c IL6 mrna Expression shrnai c,d Ctrl media e,f a b c d,e,f Ctrl Bcl6 c,d,e d,e,f IL-1Ra f PPAR!/" f COX-2 mrna Expression shrnai Ctrl media Ctrl Bcl6 IL-1Ra PPAR!/" Figure 3-5: Conditioned media from Miapaca-2 cells influences gene expression in differentiated THP-1 cells. Miapaca-2 cells were transiently infected with the indicated shrnas for 48 hours before treatment with TNFα with or without GW for 24 hours. Equal aliquots of conditioned media or control media were then added to THP-1 cells that had been previously differentiated. Gene expression of pro-inflammatory genes MCP-1 (A), MCP-3 (B), TNFα (C), IL-1β (D), IL-6 (E) and COX-2 (F) was determined by qrt-pcr and expressed as fold-induction following normalization to β-actin. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

139 126 Conditioned media from GW treated Miapaca-2 cells influences THP-1 cell invasion. Since PPARβ/δ activation suppresses the production of genes that positively influence macrophage chemotaxis, particularly MCP-1 and MCP-3, as well as various other pro-inflammatory mediators, we next sought to determine if Miapaca-2 conditioned media could influence the migration of THP-1 cells across a basement membrane. THP- 1 cells were seeded into the upper chamber of an invasion plate in serum-free media, and conditioned media from human pancreatic cancer cells expressing the indicated shrnas and treated with TNFα with or without GW was used as a chemoattractant. Consistent with the observation that PPARβ/δ activation reduces MCP-1 expression, GW conditioned media was able to significantly reduce the amount of invading THP-1 cells except in media conditioned by PPARβ/δ knock-down Miapaca-2 cells, where GW addition had no significant effect (Fig. 3-6). Of particular note, however, is the observation that conditioned media from BCL-6 knock-down Miapaca-2 cells stimulated with TNFα alone resulted in an approximately 70% increase in invading THP-1 cells compared with controls, while media conditioned by PPARβ/δ knock-down Miapaca-2 cells resulted in an approximate 30% reduction in invading THP-1 cells regardless of TNFα stimulation or GW treatment.

140 ng/ml TNF# 200 a TNF# + GW Percent Invading Cells b d b d b,c,d b,c d c,d d 0 Ctrl media Ctrl Bcl6 IL-1Ra PPAR!/" shrnai Figure 3-6: GW conditioned Miapaca-2 media reduces the percentage of invading THP-1 cells across a basement membrane. Conditioned media was isolated from Miapaca-2 cells expressing the indicated shrnas and treated with TNFα with or without GW for 24 hours. THP-1 monocytes were allowed to migrate across the membrane for 24 hours and relative invasion was quantified using the CytoSelect 96-well cell invasion assay with fluorometric readings at 480 nm / 520 nm. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

141 Discussion Similar to other nuclear receptors, PPARβ/δ is a regulator of the inflammatory response (14) and a potential target of therapeutic intervention. Agonists of this receptor exert their anti-inflammatory effects via several potential mechanisms, such as induction of anti-inflammatory target genes, inhibition of NF-κB activity, and the release of the transcriptional repressor BCL-6. There has been much interest recently in the role of the PPARs in regulating inflammation (12), but to date little is known about the role of PPARβ/δ in regulating inflammation in the pancreas. This study provides evidence that the PPARβ/δ-specific ligand GW reduces TNFα-induced NF-κB activity and inflammatory gene expression via induction of target genes as well as BCL-6 suppression of pro-inflammatory mediators in human pancreatic cancer cells. Moreover, pancreatic cancer cell-conditioned media was able to effect gene expression and invasive potential of THP-1 cells, indicating that PPARβ/δ activators may influence the cross-talk between pancreatic cancer cells and macrophages, providing a link between inflammation and cancer (27). PPARβ/δ over-expression significantly reduced both basal and TNFα-induced NF-κB reporter activity in a process likely mediated, in part, by direct physical interaction between the two transcription factors, although co-immunoprecipitation studies would be required to confirm these observations. Furthermore, GW reduced NF-κB activity in control, but not BCL-6 or PPARβ/δ knock-down cells, indicating that the potential repressive mechanism is dependent on adequate levels of both BCL-6 and PPARβ/δ. The p65 subunit of NF-κB is constitutively activated in

142 129 human pancreatic cancer cells (28), and activation of this transcription factor is implicated in the progression of pancreatic tumorigenesis (29). NF-κB activity is inhibited in endothelial cells and cardiomyocytes following PPARβ/δ activation (30, 31), although whether or not GW mediated inhibition of NF-κB activity in pancreatic cancer cells proceeds solely via direct interaction or subsequent inhibition of ERK1/2 map kinase phosphorylation as in adipocytes (24) remains to be seen. Consistent with reports where BCL-6 was silenced (32), TNFα-stimulated NF-κB activity was significantly elevated in BCL-6 knock-down Miapaca-2 cells, indicating that the transcriptional repressor does participate in regulating NF-κB activity. In diffuse large B- cell lymphomas, for example, BCL-6 interacts physically with NF-κB both in vivo and in vitro (32). Although unexplored in this study, it is possible that this repressive mechanism is active in human pancreatic cancer cells. Previous studies indicate promise for the role of NF-κB inhibitors in the control of the inflammatory response and cell growth (29). Thus, compounds such as GW may stimulate the association of NFκB with two inhibitor proteins (PPARβ/δ and BCL-6.) and may provide an effective approach to controlling NF-κB activity and hence cell survival and inflammation in the pancreas. Another mechanism by which PPARβ/δ activators exert their anti-inflammatory effects is through induction of target genes. GW treatment increases the mrna levels of IL-1Ra, TGF-β, SOD-1 and FGF-21 in human pancreatic cancer cells. IL-1Ra, a natural antagonist of IL-1β (33), and TGF-β are PPARβ/δ target genes in VSMCs (21, 34) where they contribute to the inhibition of IL-1β-induced migration and proliferation. PPARβ/δ also induces the expression of various anti-oxidant genes in vascular cells, such

143 130 as SOD-1, which reduces the oxidative stress burden by attenuating superoxide anion, and contains a functional PPRE in its 5 -flanking region (15). Ligands for PPARα and PPARγ also induce expression of SOD-1 in endothelial cells (35). FGF-21 is regulated by PPAR-dependent pathways in mice, and is responsible for protecting pancreatic acini from pancreatitis-induced injury (36). We demonstrate here that PPARβ/δ activation in human pancreatic cancer cells reduces the inflammatory response in part through direct induction of these anti-inflammatory PPARβ/δ-target genes, and that the antiinflammatory properties of GW in the pancreas are mediated at least in part by PPARβ/δ. Unique to PPARβ/δ is its association with the transcriptional repressor, BCL-6. PPARα and PPARγ exhibit no affinity for BCL-6 (11), which contributes in large part to the anti-inflammatory properties of PPARβ/δ ligands in various cellular and animal models. In the mouse macrophage, for example, PPARβ/δ activation and subsequent release of BCL-6 is responsible for suppressing MCP-1 reporter and gene expression, as well as MCP-3, IL-1β and MIP1β (11). These effects also translate well in animal models, as Angiotensin II treated LDLR-/- mice dosed with the PPARβ/δ-specific agonist GW0742 showed increased levels of total and free BCL-6, and attenuated AngIIaccelerated atherosclerosis. This effect was associated with decreased levels of proinflammatory gene expression (MCP-1, PAI-1, TNFα, IL-6 and macrophage infiltration leptin, for example) in the artery (23). Our observations in Miapaca-2 and BxPc-3 pancreatic cancer cells demonstrate that the BCL-6 anti-inflammatory pathway is active and responds to PPARβ/δ activation, reducing the TNFα-induced promoter activity of MCP-1 and pro-inflammatory gene expression in a BCL-6-depedent fashion.

144 131 Pancreatic tumor mass is comprised in large part of macrophages and the inflammatory response therein promotes tumor growth and spread (37). Increased macrophage infiltration into tumors results in a poor prognosis (38), and deletion of macrophages in breast (39) and Ewing s sarcoma (40) animal models have resulted in decreased tumor incidence. The control of infiltrating macrophages, therefore, presents an attractive therapeutic method of reducing tumor burden. Indeed, in an orthotopic mouse model of pancreatic cancer, the inhibition of VEGF reduced macrophage infiltration and subsequent tumor growth (37). We show that conditioned media from TNFα-stimulated Miapaca-2 cells effected gene expression in differentiated THP-1 macrophages, inducing the expression of various pro-inflammatory markers and chemotactic agents, while conditioned media from pancreatic cancer cells treated with a PPARβ/δ-specific agonist simultaneously with TNFα resulted in significantly lower proinflammatory markers in THP-1 cells. Using lentiviral-mediated knock-down of various genes, we also demonstrate that both PPARβ/δ and BCL-6 are key players in the regulation of cross-talk between pancreatic cancer cells and macrophages. More importantly, the GW mediated suppression of these pro-inflammatory and chemotactic mediators affected THP-1 cell migration toward pancreatic cancer cellconditioned media, resulting in an approximate 30% reduction in invasive potential. Clearly, the anti-inflammatory actions of BCL-6 and PPARβ/δ following ligandactivation show potential for limiting the recruitment of tumor-associated macrophages via reduced pro-inflammatory gene expression. Further experimentation is required to fully elucidate the role(s) of both PPARβ/δ and BCl-6 in the exocrine pancreas, but our data support the use of PPARβ/δ activators in both reducing the inflammatory burden and

145 132 affecting potential cross-talk between pancreatic carcinomas and infiltrating macrophages. 3.6 References 1. Flanders TY, Foulkes WD. Pancreatic adenocarcinoma: epidemiology and genetics. J Med Genet 1996;33: Borad MJ, Saadati H, Lakshmipathy A, et al. Skeletal metastases in pancreatic cancer: a retrospective study and review of the literature. Yale J Biol Med 2009;82: Wolpin BM, Stampfer MJ. Defining determinants of pancreatic cancer risk: are we making progress? J Natl Cancer Inst 2009;101: Carriere C, Young AL, Gunn JR, Longnecker DS, Korc M. Acute pancreatitis markedly accelerates pancreatic cancer progression in mice expressing oncogenic Kras. Biochem Biophys Res Commun 2009;382: Greer JB, Whitcomb DC. Inflammation and pancreatic cancer: an evidence-based review. Curr Opin Pharmacol 2009;9: Coussens LM, Werb Z. Inflammation and cancer. Nature 2002;420: Greer JB, Lynch HT, Brand RE. Hereditary pancreatic cancer: a clinical perspective. Best Pract Res Clin Gastroenterol 2009;23: Lim HJ, Park JH, Lee S, Choi HE, Lee KS, Park HY. PPARdelta ligand L ameliorates Western diet-induced hepatic lipid accumulation and inflammation in LDLR-/- mice. Eur J Pharmacol 2009;622: Kalinin S, Richardson JC, Feinstein DL. A PPARdelta agonist reduces amyloid burden and brain inflammation in a transgenic mouse model of Alzheimer's disease. Curr Alzheimer Res 2009;6: Ding G, Cheng L, Qin Q, Frontin S, Yang Q. PPARdelta modulates lipopolysaccharide-induced TNFalpha inflammation signaling in cultured cardiomyocytes. J Mol Cell Cardiol 2006;40: Lee CH, Chawla A, Urbiztondo N, et al. Transcriptional repression of atherogenic inflammation: modulation by PPARdelta. Science 2003;302: Moraes LA, Piqueras L, Bishop-Bailey D. Peroxisome proliferator-activated receptors and inflammation. Pharmacol Ther 2006;110: Barish GD, Atkins AR, Downes M, et al. PPARdelta regulates multiple proinflammatory pathways to suppress atherosclerosis. Proc Natl Acad Sci U S A 2008;105: Bishop-Bailey D, Bystrom J. Emerging roles of peroxisome proliferator-activated receptor-beta/delta in inflammation. Pharmacol Ther 2009;124:

146 15. Fan Y, Wang Y, Tang Z, et al. Suppression of pro-inflammatory adhesion molecules by PPAR-delta in human vascular endothelial cells. Arterioscler Thromb Vasc Biol 2008;28: Piqueras L, Sanz MJ, Perretti M, et al. Activation of PPARbeta/delta inhibits leukocyte recruitment, cell adhesion molecule expression, and chemokine release. J Leukoc Biol 2009;86: Kharroubi I, Lee CH, Hekerman P, et al. BCL-6: a possible missing link for antiinflammatory PPAR-delta signalling in pancreatic beta cells. Diabetologia 2006;49: Planavila A, Laguna JC, Vazquez-Carrera M. Nuclear factor-kappab activation leads to down-regulation of fatty acid oxidation during cardiac hypertrophy. J Biol Chem 2005;280: Delerive P, De Bosscher K, Besnard S, et al. Peroxisome proliferator-activated receptor alpha negatively regulates the vascular inflammatory gene response by negative cross-talk with transcription factors NF-kappaB and AP-1. J Biol Chem 1999;274: Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature 2008;454: Kim HJ, Ham SA, Kim SU, et al. Transforming growth factor-beta1 is a molecular target for the peroxisome proliferator-activated receptor delta. Circ Res 2008;102: Chong HC, Tan MJ, Philippe V, et al. Regulation of epithelial-mesenchymal IL-1 signaling by PPARbeta/delta is essential for skin homeostasis and wound healing. J Cell Biol 2009;184: Takata Y, Liu J, Yin F, et al. PPARdelta-mediated antiinflammatory mechanisms inhibit angiotensin II-accelerated atherosclerosis. Proc Natl Acad Sci U S A 2008;105: Rodriguez-Calvo R, Serrano L, Coll T, et al. Activation of peroxisome proliferator-activated receptor beta/delta inhibits lipopolysaccharide-induced cytokine production in adipocytes by lowering nuclear factor-kappab activity via extracellular signal-related kinase 1/2. Diabetes 2008;57: Morse MA, Hall JR, Plate JM. Countering tumor-induced immunosuppression during immunotherapy for pancreatic cancer. Expert Opin Biol Ther 2009;9: Mancino A, Lawrence T. Nuclear factor-kappab and tumor-associated macrophages. Clin Cancer Res;16: Sica A, Allavena P, Mantovani A. Cancer related inflammation: the macrophage connection. Cancer Lett 2008;267: Wang W, Abbruzzese JL, Evans DB, Larry L, Cleary KR, Chiao PJ. The nuclear factor-kappa B RelA transcription factor is constitutively activated in human pancreatic adenocarcinoma cells. Clin Cancer Res 1999;5: Yip-Schneider MT, Nakshatri H, Sweeney CJ, Marshall MS, Wiebke EA, Schmidt CM. Parthenolide and sulindac cooperate to mediate growth suppression and inhibit the nuclear factor-kappa B pathway in pancreatic carcinoma cells. Mol Cancer Ther 2005;4:

147 30. Rival Y, Beneteau N, Taillandier T, et al. PPARalpha and PPARdelta activators inhibit cytokine-induced nuclear translocation of NF-kappaB and expression of VCAM-1 in EAhy926 endothelial cells. Eur J Pharmacol 2002;435: Smeets PJ, Teunissen BE, Planavila A, et al. Inflammatory pathways are activated during cardiomyocyte hypertrophy and attenuated by peroxisome proliferatoractivated receptors PPARalpha and PPARdelta. J Biol Chem 2008;283: Perez-Rosado A, Artiga M, Vargiu P, Sanchez-Aguilera A, Alvarez-Barrientos A, Piris M. BCL6 represses NFkappaB activity in diffuse large B-cell lymphomas. J Pathol 2008;214: Bresnihan B, Cunnane G. Interleukin-1 receptor antagonist. Rheum Dis Clin North Am 1998;24: Kim HJ, Kim MY, Hwang JS, et al. PPARdelta inhibits IL-1beta-stimulated proliferation and migration of vascular smooth muscle cells via up-regulation of IL-1Ra. Cell Mol Life Sci. 35. Inoue I, Goto S, Matsunaga T, et al. The ligands/activators for peroxisome proliferator-activated receptor alpha (PPARalpha) and PPARgamma increase Cu2+,Zn2+-superoxide dismutase and decrease p22phox message expressions in primary endothelial cells. Metabolism 2001;50: Johnson CL, Weston JY, Chadi SA, et al. Fibroblast growth factor 21 reduces the severity of cerulein-induced pancreatitis in mice. Gastroenterology 2009;137: Dineen SP, Lynn KD, Holloway SE, et al. Vascular endothelial growth factor receptor 2 mediates macrophage infiltration into orthotopic pancreatic tumors in mice. Cancer Res 2008;68: Talmadge JE, Singh RK, Fidler IJ, Raz A. Murine models to evaluate novel and conventional therapeutic strategies for cancer. Am J Pathol 2007;170: Lin EY, Pollard JW. Tumor-associated macrophages press the angiogenic switch in breast cancer. Cancer Res 2007;67: Zeisberger SM, Odermatt B, Marty C, Zehnder-Fjallman AH, Ballmer-Hofer K, Schwendener RA. Clodronate-liposome-mediated depletion of tumour-associated macrophages: a new and highly effective antiangiogenic therapy approach. Br J Cancer 2006;9 134

148 Chapter 4 Role of Peroxisome Proliferator-Activated Receptor β/δ (PPARβ/δ) and B-cell lymphoma 6 (BCL-6) in regulation of genes involved in metastasis and migration in pancreatic cancer cells 4.1 Abstract Pancreatic cancer is malignant neoplasia with poor prognosis due in part to the lack of established biomarkers, preventative strategies and therapeutic intervention. Peroxisome proliferator-activated receptor-beta/delta (PPARβ/δ) is a ligand-activated transcription factor that regulates various cellular functions via induction of target genes either directly or in concert with the associated transcriptional repressor, B-cell lymphoma-6 (BCL-6) protein. Matrix remodeling proteinases, particularly MMP-9, are frequently over-expressed in pancreatic cancer and are involved with metastasis. The present study tested the hypothesis that PPARβ/δ is expressed in human pancreatic cancer cells and that its activation could regulate genes such as MMP-9 and decrease cancer cells ability to transverse the basement membrane. In human pancreatic cancer tissue there was significantly higher expression of MMP-9 and PPARβ/δ, and significantly lower levels of BCL-6 mrna. Human pancreatic cancer cells (MiaPaca and BxPc-3) expressed functional PPARβ/δ as evidenced by GW induction of the target gene ADRP. PPARβ/δ activation was sufficient to reduce the TNFα-induced expression of

149 136 various genes implicated in metastasis and invasion, including MMP-9 mrna and protein, and reduced the invasion of pancreatic cancer cells through a basement membrane in cell culture models. Through the use of short hairpin RNA inhibitors of PPARβ/δ, BCL-6 as well as MMP-9, it was evident that PPARβ/δ was responsible for the ligand-dependent effects whereas BCL-6 release from the complex upon GW treatment was ultimately responsible for decreasing MMP-9 expression and hence invasion activity. These results suggest that PPARβ/δ plays a role in regulating pancreatic cancer cell invasion through regulation of genes via ligand-dependent release of BCL-6, and that activation of the receptor may provide an alternative therapeutic method for controlling pancreatic cancer cell migration and metastasis. 4.2 Introduction Pancreatic cancer is the fourth leading cause of cancer-related deaths of both men and women in the United States. The American Cancer Society estimates for 2009 predicted approximately 42,470 new cases of pancreatic cancer and that 35,240 of those cases would result in death. Lack of suitable identifiable symptoms or biomarkers combined with a 4% five-year survival rate makes pancreatic cancer one of the deadliest of malignancies (1). Although the cancer itself is difficult to detect in its early stages, several known risk factors exist, with smoking being the most well-documented etiologic agent and believed to be responsible for 20-25% of all known cases of pancreatic cancer

150 137 (2). Several other risk factors include age, diets high in fat (3), excessive alcohol consumption (4), diabetes mellitus (5) and chronic pancreatitis (6). Common chemotherapeutic treatments have had little success in improving survival rates or restraining the highly metastatic malignancies (7) with the median survival rate of less than six months and surgical resection as the only effective treatment (8). Prevention strategies and alternative treatments for pancreatic cancer are sorely needed. Peroxisome proliferator-activated receptor-β/δ (PPARβ/δ) is a member of the nuclear receptor superfamily of ligand-activated transcription factors. The PPARs consist of three isoforms; PPARα (NR1C1), PPARβ/δ (NR1C2; NUC1; FAAR fatty acidactivated receptor) and PPARγ (NR1C3). The PPARs effect gene transcription in response to various stimuli, such as fatty acids and their metabolites, xenobiotics and isoform-specific drugs, through a heterodimerization with retinoid X recptors (RXRs) and subsequent recognition and binding to peroxisome proliferator-responsive elements (PPREs) within the promoter regions of target genes (9, 10). PPARβ/δ, unlike PPARα or PPARγ which have distinct tissue expression patterns and synthetic ligands, is ubiquitously expressed, often at higher levels than the other isoforms. This receptor regulates fatty acid oxidation and lipid homeostasis (11), cell proliferation and differentiation (12), cell survival (13) and the inflammatory response (14). The latter response may be via its association with the transcriptional repressor BCL-6, which is released upon activation of PPARβ/δ (15). In the pancreas, PPARβ/δ is expressed in islet cells to a greater extent than either PPARα or PPARγ, and in beta cells where it regulates the inflammatory response (16). Expression profiling analyses in the mouse demonstrated high PPARβ/δ expression in the cytoplasm of delta cells of the islet of

151 138 Langerhans, suggesting a potential role for the receptor in the regulation of glucose metabolism (17). Pancreatic ductal adenocarcinomas are by far the most common of pancreatic malignancies (18), and the role(s) of PPARβ/δ in pancreatic ductal cells is poorly understood. The matrix metalloproteinases are a family of zinc-dependent proteolytic enzymes that degrade extracellular matrix (ECM) proteins and are well-known regulators of pancreatic cancer cell metastasis and invasion (19, 20). Matrix metalloproteinase-9 (MMP-9, also known as Gelatinase B) in particular is highly expressed in both clinical and experimental models of pancreatic cancer (21). Furthermore, pancreatic cancer cells display extremely high basal MMP-9 expression, which is further inducible by phorbol 12-myristate 13-acetate (PMA) (22). Recently, several studies have linked PPARβ/δ to MMP-9; in PPARβ/δ null macrophages, basal MMP-9 expression is reduced (15), and in vascular smooth muscle cells (VSMCs) PPARβ/δ activation suppressed the expression of both MMP-2 and MMP-9, with further inhibition on VSMC migration and proliferation (23). The role(s) of PPARs, particularly PPARβ/δ, in tumorigenesis and cancer cell invasion remains controversial. For example, inhibition of PPARγ suppressed pancreatic cancer cell motility in Capan-1 and Panc-1 cells (24), while its activation in AsPC-1 cells by the specific ligand Rosiglitazone increased levels of the tumor suppressor PTEN and decreased levels of phosphorylated Akt (25) and induced caspase-mediated apoptosis in Miapaca-2 cells (26). PPARβ/δ is an APC-regulated target of non-streroidal antiinflammatory drugs (NSAIDs), suggesting that NSAIDs inhibit tumorigenesis via PPARβ/δ inhibition (27), and genetic disruption of PPARβ/δ contributes to the growth-

152 139 inhibitory effects of APC (28). Opposing evidence exists suggesting that PPARβ/δ activation increases (29-31) and decreases cell proliferation (32, 33) in various cell types. Previous evidence, however, establishes a clear link between PPARβ/δ, BCL-6 and MMP-9, and we sought to elucidate the role(s) of PPARβ/δ activation on potential target genes involved in pancreatic cancer invasion and metastasis. We used the PPARβ/δspecific activator GW and shrnas to decrease expression of PPARβ/δ, BCL-6 and MMP-9 in two human pancreatic cancer cell lines, Miapaca-2 (COX-2 negative) and BxPc-3 (COX-2 positive). The experiments show that ligand-dependent activation of PPARβ/δ causes a BCL6-dependent repression of MMP-9 and other genes involved in cancer metastasis and decreases indices of cell migration. This suggests that PPARβ/δ agonists may be a beneficial tool in the prevention and treatment of pancreatic cancer. 4.3 Materials and Methods Cells and Reagents. Human pancreatic cancer cells, Miapaca-2 (COX-2 negative, CRL- 1420) and BxPc-3 (COX-2 positive, CRL-1687) were purchased from the ATCC (Manassas, VA) and cultured in high-glucose DMEM containing 10% FBS and 1% penicillin/streptomycin in a humidified atmosphere at 37 C containing 5% CO 2. Human embryonic kidney 293 cells were cultured in DMEM containing 10% FBS and 1% penicillin/streptomycin. All media components and fetal bovine serum (FBS) were purchased from Gibco BRL/Life Technologies (Carlsbad, CA). Ciprofibrate (Cipro),

153 140 purchased from Sigma Chemical Co. (St Louis, MO), was used as the positive control for PPARα. GW (GW), purchased from Sigma Chemical Co., was used as the positive control for PPARβ/δ. Rosiglitazone (Rosi), purchased from Cayman Chemicals (Ann Arbor, MI), was used as the positive control for PPARγ. Recombinant human TNFα and human MMP-9 ELISAs were purchased from Invitrogen (Carlsbad, CA) and used according to the manufacturer s instructions. Human pancreatic cancer, chronic pancreatitis and pancreas tissue samples were obtained from Dr. Gerhard Leder, (Abt. Allgemein- und Viszeralchirurgie, St. Josef Hospital - Klinikum der Ruhr, University of Bochum, Germany). MISSION shrna bacterial glycerol stocks targeted against human PPARβ/δ, Bcl6, MMP-9, as well as the non-targeting vector, were purchased directly from Sigma-Aldrich. High Capacity cdna Archive Kit and ABI7300 Real-time PCR System were purchased from Applied Biosystems (Foster City, CA). The ppackh1 packaging plasmids were kindly provided by Dr. Curtis J. Omiecinski (Penn State University). CytoSelect TM 96-well Cell Invasion Assay (Basement Membrane, Fluorometric Format) was purchased from Cell Biolabs, Inc. (San Diego, CA) and used according to the manufacturer s instructions. Isolation of total RNA and Real-time Quantitative RT-PCR. Total RNA was isolated from Miapaca-2 and BxPc-3 cells using Tri-Reagent and the manufacturer s recommended protocol (Sigma). Human pancreatic tissue samples were briefly homogenized in 1mL Tri-Reagent and total RNA was isolated. One µg of total RNA was reverse transcribed using the High Capacity cdna Archive Kit (Applied Biosystems,

154 141 Foster City, CA). PCR primers for quantitative real-time RT-PCR were designed based on published sequences in GenBank and are shown in Table 4-1. The housekeeping gene β-actin was used to normalize all the tested genes. The data shown are representative of three independent experiments with triplicate samples. Quantification of MMP-9 protein by ELISA. MMP-9 protein levels were quantified using the human MMP-9 ELISA according to the manufacturer s instructions (Invitrogen). Briefly, control Miapaca-2 cells or shrna knock-down cells were plated in 6-well tissue culture plates and treated 1 ng/ml TNFα with or without 500 nm GW for 24 h. At the end of the incubation time, the media was removed and diluted 1:40 in standard diluent buffer. Diluted media samples and MMP-9 standards were added to a 96-well microtiter plate containing human MMP-9 antibody-coated wells and allowed to incubate at room temperature for 2 h. Following the incubation, the media was aspirated and each well washed 5 times with wash buffer. One hundred µl Biotinylated anti-mmp-9 (Biotin Conjugate) solution was added to each well and the plate was incubated for 1 h at room temperature. Each well was washed a second time with wash buffer, and 100 µl of Streptavidin-HRP working solution was added and the plate was allowed to incubate at room temperature for 30 minutes. After a third wash, 100 µl of stabilized chromogen was added to each well and the plate was incubated at room temperature for 30 minutes in the dark, after which time 100 µl of stop solution was added and the absorbance read at 450 nm.

155 142 Cell migration assay. Either control or knock-down human pancreatic cancer cells were grown to confluence in 10-cm tissue culture plates and then pre-treated with TNFα with or without GW for 24 h, as above. Cell migration assays were performed using the CytoSelect TM 96-Well Cell Invasion Assay (Basement Membrane, Fluormetric Format) according to the manufacturer s instructions. Briefly, the basement membrane was allowed to reach room temperature for 30 minutes, and rehydrated using warm, serum-free DMEM. Human pancreatic cancer cells were then seeded into each well at a density of 2 x 10 6 cells/ml in serum-free media. Normal cell media (DMEM containing 10% FBS, along with TNFα with or without GW501516) was added to the feeder tray and the entire apparatus was placed in an incubator at 37 C containing 5% CO 2 for 24 h. CyQuant GR dye/lysis buffer solution was added to the invading cells following completion of the assay and the resulting mixture was incubated at room temperature for 20 minutes. Invading cells were quantified by reading the fluorescence at 480 nm/520 nm. All measurements were performed in triplicate. Lentiviral shrna infection. HEK-293 cells were grown to confluency in 10-cm tissue culture plates under the conditions described above. The cells were then transiently transfected with 4.6 µg of either non-targeting shrna, or shrnas targeted against human PPARβ/δ, Bcl6, or MMP-9, as well as 2.4 µg each of ppackh1 packaging plasmids, using Lipofectamine Cells were transfected for 6 h and allowed to recover overnight in normal media. Fresh media was added the following morning, and pseudoviral supernatant was generated for 72 h. Supernatant was then harvested and

156 143 passed through a 0.4 µm filter under sterile conditions. Polybrene (Millipore, Billerica, MA) was then added to a final concentration of 5 µg per ml and the pseudoviral supernatant was then added directly to target cells for 6 h. Infected cells were allowed to recover overnight following the addition of 6 ml complete media and knock-down of target genes was assessed by RT-PCR 48 h post-infection. Statistical Analysis. Quantitative data are presented as mean±sem. ANOVA with p- value<0.05 was used to determine whether differences among variables were significant. Normality was checked using Anderson-Darling test, and the General linear model, followed by the Tukey post hoc test to analyze differences between treatments. All data analyses were performed by MiniTAB Ver.14 (MiniTAB, State College, PA) or JMP (SAS Institute, Cary, NC) and data were plotted by Prism 5.01 (GraphPad Software, San Diego, CA).

157 144 Table 4-1: List of Real-time PCR primers used in this study. Gene Forward Primer Reverse Primer E-selectin TCCTATTCCAGCCTGCAATGT AACCCATTGGCTGGATTTGTC ICAM-1 ACTCAGCGGTCATGTCTGGAC GGCATAGCTTGGGCATATTCC VCAM-1 AGTGGTGGCCTCGTGAATG CACGCTAGGAACCTTGCAGC IL-1β TCCTTAGTCCTCGGCCAAGAC GTGCCATGGTTTCTTGTGACC MCP-1 GGACGCATTTCCCCAGTACA CCGAGAACGAGATGTGGACA MMP-9 AGCGGTCCTGGCAGAAATAG ACGCACGACGTCTTCCAGTAC BCL-6 GCTCACGGCTCACAACAATG TCCGGAGTCGAGACATCTTGA PPARβ/δ AGGCCATTCACCAACTGCTT ATTGTGGCAGGCAGAGAAGG

158 Results Tissue samples from human pancreatic ductal carcinomas show significantly increased levels of MMP-9 mrna. It is well known that the matrix metalloproteinases are key regulators of cell proliferation and migration in human pancreatic cancer cells (34) and that MMP-9 protein is increased in the pancreatic juice from patients diagnosed with pancreatic ductal adenocarcinomas (35). Recently, MMP-9 has been linked to PPARβ/δ and the transcriptional repressor BCL-6; in PPARβ/δ-/- macrophages, for example, there was lower MMP-9 expression compared with wild-type cells (15). Tissue samples from patients diagnosed with chronic pancreatitis or pancreatic cancer were obtained, and we set out to assess the differences in expression of several genes involved in inflammation and metastasis. Indeed, there was a 10-fold increase in MMP-9 gene expression in ductal carcinomas compared with samples from patients diagnosed with chronic pancreatitis (Fig. 4-1). Interestingly, PPARβ/δ expression was also elevated while mrna expression of the transcriptional repressor BCL-6 was almost 3-fold lower in tumor samples compared with those from chronic pancreatitis patients. Despite the low expression of BCL-6 in ductal carcinomas, however, the relative expression of two BCL-6 target genes, VCAM-1 (36) and MCP-1, were not significantly elevated in tumor samples. A PPARβ/δ target gene, ADRP, was also not different between tumor, pancreatitis and other pancreatic tissue samples (data not shown).

159 146 E-selectin ICAM-1 VCAM-1 IL-1! MCP-1 MMP-9* Bcl-6* PPAR!/"* Relative Expression Pancreatitis Other Ductal Carcino. Pancreatitis Other Ductal Carcino. Pancreatitis Other Ductal Carcino. Pancreatitis Other Ductal Carcino. Pancreatitis Other Ductal Carcino. a a b Pancreatitis Other Ductal Carcino. a a,b b Pancreatitis Other Ductal Carcino. a a b Pancreatitis Other Ductal Carcino. Figure 4-1: Relative mrna expression in human pancreatic tissues at varying stages of carcinogenesis from chronic pancreatitis to pancreatic cancer. Total mrna was isolated using standard Tri-reagent protocol and reverse-transcribed. Gene expression was determined using qrt-pcr and expressed as fold induction after normalization to β-actin. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

160 147 Regulation of MMP-9 expression by PPARβ/δ and BCL6 in pancreatic cancer cells. PPARβ/δ activation negatively influences MMP-9 gene expression in IL-1β-stimulated vascular smooth muscle cells (23). To study the mechanism of this response and to determine its applicability to another cell type, Miapaca-2 cells were transiently infected with non-targeting control, hbcl-6, hpparβ/δ or hmmp9 lentiviral shrnas (Figure 4-2 A). Cells transiently infected with non-targeting control shrna showed no alterations in BCL-6, PPARβ/δ or MMP9 mrna expression. MiaPaca cells infected with an shrna targeted against BCL-6 showed approximately 50% reduction in BCL-6 mrna levels, and cells infected with an shrna targeting PPARβ/δ or MMP9 showed approximately 70% reduction in corresponding mrna levels (Fig. 4-2 A). To determine if gene expression is altered after lentiviral-mediated BCL-6 or PPARβ/δ repression, the PPAR target gene ADRP was examined upon treatment with three isoform-specific PPAR agonists (ciprofibrate, PPARα; GW501516, PPARβ/δ; rosiglitazone, PPARγ). Control cells and those transiently expressing the indicated shrnas were treated with 20 µm Ciprofibrate, 500 nm GW or 10 µm Rosiglitazone. Miapaca-2 cells contain functional PPARs as indicated by the ligand-induced expression of ADRP, with each treatment resulting in a three-fold increase in transcript levels (Fig. 4-2 B). Cells expressing PPARβ/δ-specific shrna did not induce expression of ADRP in response to GW at a concentration that activates only PPARβ/δ (37), while cells expressing BCL-6-targeting shrna retained inducible expression of ADRP by all three isoformspecific ligands. Following BCL-6, PPARβ/δ or MMP-9 knock-down, cells were treated with 1 ng/ml TNFα with or without 500 nm GW for 24 hours, and MMP-9 protein levels were assessed by ELISA. MMP-9 protein levels were significantly

161 148 elevated in BCL-6 knock-down Miapaca-2 cells following TNFα challenge compared with control cells, while PPARβ/δ knock-down Miapaca-2 cells showed a significant reduction in TNFα-induced MMP-9 protein levels (Fig. 4-2 C), consistent with previous reports in PPARβ/δ -/- macrophages. While GW co-treatment significantly suppressed TNFα-induced MMP-9 protein levels in control (non-targeting) Miapaca-2 cells, this effect was not observed in either of the BCL-6 or PPARβ/δ knock-down cells. Not unexpectedly, lentiviral shrna targeted against MMP-9 significantly reduced both mrna and protein expression in Miapaca-2 cells, and GW activation of PPARβ/δ did not further reduce MMP-9 protein levels in these cells.

162 149 A B Normalized mrna Expression Non-targeting shrna hbcl6 shrna hppar!/" shrna hmmp-9 shrna * hbcl-6 mrna hppar!/" mrna hmmp-9 mrna * * Normalized ADRP Expression shrnai, Trt Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl Cipro Bcl6 Cipro PPAR!/" Cipro * * * Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl Rosi Bcl6 Rosi PPAR!/" Rosi C D Human MMP-9 [pg/ml] 250 TNF# a TNF# + GW a,b b Non-targeting shrna hbcl6 shrna hppar!/" shrna hmmp9 shrna c,* c c,d c,d d Percent Invading Cells 125 Non-targeting shrna hmmp9 shrna * 25 0 Figure 4-2: Effect of PPARβ/δ activation on MMP-9 expression. A. Miapaca-2 cells were transiently infected with non-targeting control, hbcl-6, hpparβ/δ or hmmp9 lentiviral shrnas for 48 hours. Total mrna was isolated and gene-specific knockdown was assessed using qrt-pcr. *P<0.05 B. Miapaca-2 cells contain functional PPARs. Miapaca-2 cells were transiently infected with the indicated shrnas and treated with the indicated PPAR isoform-specific agonists. Induction of the PPAR-target gene ADRP was determined using qrt-pcr. *P<0.05 C. Miapaca-2 cells were stimulated with human TNFα with or without GW for 24 hours following transient infection with the indicated shrnas. Human MMP-9 protein expression was quantified using MMP-9-specific ELISA (Invitrogen). Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test. D. Miapaca-2 cells with reduced MMP-9 expression are less invasive than control Miapaca-2 cells. Following infection with human MMP-9-targeting shrna, Miapaca-2 cells were seeded in 96-well invasion plates (Cell Biolabs, Inc.) and allowed to invade the basement membrane overnight. Relative cell invasion was quantified using the CytoSelect 96-well cell invasion assay with fluorometric readings at 480 nm / 520 nm. *P<0.05

163 150 MMP-9 knock-down reduces Miapaca-2 cell invasion. Because MMP-9 is a key regulator of human pancreatic cancer cell invasion and metastasis, we further examined the effect of lentiviral shrna-mediated MMP-9 knock-down on the basal ability of Miapaca-2 cells to invade a basement membrane. Miapaca-2 cells treated with nontargeting control or MMP-9-targeting lentiviral shrnas were seeded into a 96-well invasion plate and allowed to migrate across a membrane for 24 hours. Using the migration assay described, we found that MMP-9 knock-down significantly reduced the basal migration of Miapaca-2 cells (Fig. 4-2 D). PPARβ/δ activation decreases TNFα-induced expression of pro-inflammatory and cell adhesion genes in human pancreatic cancer cells. PPARβ/δ associates with the transcriptional repressor BCL-6 that, upon PPARβ/δ activation, is released and decreases expression of target genes. To determine if the PPARβ/δ / BCL-6 pathway is active in human pancreatic cancer cells, we used shrna knock-down of PPARβ/δ and BCL-6 in conjunction with PPARβ/δ-specific activation by GW to analyze the gene expression changes. Miapaca-2 cells transiently expressing non-targeting control shrna, or shrnas targeted against PPARβ/δ or BCL-6, were stimulated with 1 ng / ml TNFα with or without GW for 24 hours. In control Miapaca-2 cells, TNFα stimulation induced the robust expression of the cell adhesion molecules E-selectin, ICAM-1 and VCAM-1, the pro-inflammatory genes IL-1β and MCP-1 and the promigratory gene MMP-9, while co-treatment with 500 nm GW significantly suppressed their expression at the mrna level (Fig. 4-3). Treatment of Miapaca-2 cells with a BCL-6-targeting shrna attenuated the GW inhibitory effect on the genes

164 151 tested, indicating a role for BCL-6 in GW mediated repression. Consistent with the findings of Lee et al. in PPARβ/δ -/- RAW264.7 macrophage cells, PPARβ/δ knockdown Miapaca-2 cells displayed significantly lower levels of these genes when challenged with TNFα alone, and PPARβ/δ activation with GW had no further significant repressive effect. Of note is the fact that although BCL-6 repression resulted in increased MMP-9 protein levels in media, it did not concordantly increase the mrna expression of this gene.

165 152 E-Selectin ICAM-1 VCAM-1 IL-1! MCP-1 MMP Normalized mrna Expression a,b a b,c b,c a c b,c a c c a,b c b,c a c c a,b c a,b a c c b c a,b a b,c c a c b a c c a,b c shrnai, Trt 0.0 Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Ctrl Ctrl Bcl6 Ctrl PPAR!/" Ctrl Ctrl GW Bcl6 GW PPAR!/" GW Figure 4-3: Effects of PPARβ/δ and BCL-6 knock-down on Miapaca-2 gene expression. Miapaca-2 cells were transiently infected with non-targeting control, hbcl-6 or hpparβ/δ-specific shrnas and then stimulated with human TNFα with or without GW for 24 hours. Total mrna was isolated and gene expression was determined using qrt-pcr. Data is normalized to β-actin and indicated as fold change. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

166 153 PPARβ/δ activation inhibits human pancreatic cancer cell migration. To examine if PPARβ/δ activation by GW and subsequent repression of pro-inflammatory and pro-migratory genes via BCL-6 influenced their ability to invade a basement membrane, Miapaca-2 (COX-2 negative, Fig. 4-4 A) and BxPc-3 (COX-2 positive, Fig. 4-4 B) were treated with non-, PPARβ/δ- or BCL-6-targeting shrna, and the effects of GW on cell migration were examined. In control cells, GW treatment negatively influenced the ability of either Miapaca-2 or BxPc-3 cells to migrate across a membrane (50% reduction). Lentiviral-mediated knock-down of BCL-6, however, increased cell migration in both cell lines compared with control cells. GW treatment of BCL-6 repressed cells did not have an effect on MiaPaca (Fig. 4-4 A) but did have an effect on comparable BxPc-3 cells (Fig. 4-4 B). Interestingly, Miapaca-2 and BxPc-3 cells transiently expressing an shrna targeted against PPARβ/δ showed significantly reduced cell migration compared with control cells with or without GW treatment. These results suggested that the transcriptional repressor BCL-6 mediates the anti-migratory actions of GW in human pancreatic cancer cells, but does so in a PPARβ/δdependent manner.

167 154 A Miapaca-2 B BxPc-3 Percent Invading Cells b c a a,b c TNF# TNF# + GW c Percent Invading Cells b c a b c c 0 Non-Targeting BCL-6 PPAR!/" 0 Non-Targeting BCL-6 PPAR!/" shrna shrna Figure 4-4: GW treatment reduces TNFα-stimulated Miapaca-2 and BxPc-3 cell invasion through a basement membrane. Human pancreatic cancer cells were transiently infected with the indicated shrnas and stimulated with human TNFα with or without GW Cells were allowed to invade a basement membrane overnight, and relative invasion was quantified using the CytoSelect 96-well cell invasion assay with fluorometric readings at 480 nm / 520 nm. GW treatment inhibits the invasion of the COX-2 negative pancreatic cancer cell line Miapaca-2 A and the COX-2 positive pancreatic cancer cell line BxPc-3 B through a basement membrane. Different letters indicate a statistical difference at P<0.05 using Tukey s multicomparison test.

168 Discussion The PPAR nuclear receptors are key regulators of inflammation and cell proliferation in human pancreatic cells (16, 38, 39). Although several studies have implicated PPARγ activation to inhibition of pancreatic cancer cell growth, little is known about the role of PPARβ/δ, save for its role in suppressing inflammation via release of the transcriptional repressor, BCL-6 (16). In general, the role of PPARβ/δ in cancer cell growth and tumorigenesis remains controversial. In colorectal cancer cells, for example, non-steroidal anti-inflammatory drugs (NSAIDs) inhibit tumorigenesis through inhibition of PPARβ/δ (27), and PPARβ/δ promotes intestinal carcinogenesis (40). Studies in the PPARβ/δ null mouse, however, show that ligand activation of PPARβ/δ induces terminal differentiation (41), and, indeed, PPARβ/δ-specific ligands inhibit the growth of keratinocytes in vivo (42, 43) and in vitro (32). Furthermore, PPARβ/δ activation is linked to inhibition of IL-1β-stimulated proliferation and migration of vascular smooth muscle cells (VSMC) (23) via regulation of IL-1Ra and TGF-β and negative regulation of MMP-2 and MMP-9. Our results show that PPARβ/δ activation by the specific ligand GW suppresses the TNFα-induced expression of MMP-9 in human pancreatic cancer cells via BCL-6, with further inhibition on the ability of two cell lines, Miapaca-2 and BxPc-3, to invade a basement membrane. Consistent with previous work (35), analysis of the relative expression levels of several genes in both ductal carcinomas and chronic pancreatitis showed significantly elevated levels of the matrix-remodeling gene MMP-9. Several studies have linked increased MMP-9 levels to increased invasiveness and metastasis (44, 45). Indeed,

169 156 MMP-9, along with MMP-2, is a critical player in the early stages of tumor invasion by degrading basement membrane Type IV collagen (46), which is considered to be a crucial step in tumor cell invasion (47). MMP-9 also participates in the degradation of the various components of the ECM (48). Inhibition of MMP activity by orally bioavailable matrix metalloproteinase inhibitors has shown promise in decreasing tumor metastasis in several clinical trials (46). In the human pancreatic cancer cell lines BxPc-3 and Miapaca-2, treatment with the neurotransmitter Norepinephrine increased cell invasiveness in a dose-dependent manner via augmented MMP-2, MMP-9, and VEGF (49), while treatment with the β-blocker propranolol inhibited these effects. Clearly the regulation of MMP activity is important in controlling, and possibly treating, pancreatic cancer. The association between MMP-9, PPARβ/δ and BCL-6 was recently established by Lee et al. (15), who showed that MMP-9 expression in PPARβ/δ -/- macrophages is significantly repressed compared to wild type. Ligand-dependent activation of the receptor significantly decreased the expression of several pro-inflammatory markers, suggesting that BCL-6 released from the PPARβ/δ complex may play a role in the negative regulation of MMP-9. A similar result was obtained in VSMCs, although the authors concluded that GW activation reduced MMP-9 expression through a PPARβ/δ- and TGF-β-dependent mechanism (23). In the present studies, PPARβ/δ mrna was increased in ductal carcinomas, while BCL-6 expression was decreased. In colorectal cancer tissue samples, PPARβ/δ expression increased during multistage carcinogenesis and was tightly associated with a highly malignant morphology (50). It is possible that PPARβ/δ plays a pivotal role in human pancreatic cells, but whether PPARβ/δ contributes to pancreatic cancer cell

170 157 metastasis or if its over-expression is the result of some altered signaling pathway remains unclear. Since the regulatory region of the PPARβ/δ contains several AP-1 response elements and is controlled by a variety of inflammatory-signals (51), the increased expression of this nuclear receptor may be indicative of stress-response and not causally related to the tumor phenotype. However increased expression of PPARβ/δ in the unactivated state may sequester BCL-6 in an inactive complex and hence increase the expression of genes normally controlled by this transcriptional repressor. Of particular note is the observation that the relative expression of BCL-6, a proto-oncogene known to suppress genes involved in cell cycle progression, particularly cyclin D1 (52), and inflammation (53), was significantly lower in ductal carcinomas compared with pancreatitis. BCL-6 is frequently mutated by genetic aberrations in several disorders (54-56) and has been implicated in cell line immortalization and oncogenic transformation by overriding cellular senescence downstream of p53 (57). Interestingly, DNA-chip hybridization assays identified both BCL-6 and BCL-10 as novel candidate genes in pancreatic cancer that were over-expressed in pancreatic cancer cell lines and primary tumor samples (58). Contrary to ductal cells, BCL-6 is absent in pancreatic beta cells, potentially explaining the lack of anti-inflammatory PPARβ/δ signaling in this cell type (16). Our results suggest that the BCL-6 pathway may be disrupted as inflamed pancreatic tissue transforms into a tumor, with the possibility that a loss of BCL-6 expression or activity eventually leads to increased cancer invasion via increased MMP-9 expression. Activation of PPARβ/δ decreased the TNFα-induced expression of MMP-9 at the protein level as measured by ELISA, while knock-down of BCL-6 using shrna

171 158 techniques significantly increased TNFα-induced MMP-9 protein production. Activation of PPARβ/δ in BCL-6 knock-down cells showed no significant effect on reducing MMP- 9 protein levels, suggesting a key role for BCL-6 in the regulation of MMP-9. Furthermore, knocking down PPARβ/δ in human pancreatic cancer cells significantly reduced MMP-9 protein levels to those comparable to GW treated control cells despite PPARβ/δ activation. It is our hypothesis that MMP-9 is an indirect PPARβ/δ target gene, and that PPARβ/δ activation allows for release of BCL-6 which may then relocate to the MMP-9 promoter. Of course, further studies, such as chromatin immunoprecipitation assays, for example, would be needed to substantiate this point. Our results, however, are in agreement with previous work indicating that low levels of PPARβ/δ result in decreased MMP-9 expression (15), and we believe that, in the absence of PPARβ/δ, BCL-6 is then more available to repress target genes, either through direct repression on target gene promoters as in the case of VCAM-1 (36) or through interactions with other cell signaling mediators, such as NF-κB (59). The relationship between MMP-9 and metastasis in pancreatic cancer has been well documented. Indeed, treatment of Miapaca-2 cells with shrna targeted against MMP-9 reduced protein levels by approximately 50% regardless of GW treatment, with corresponding inhibition on the ability of the cell line to invade a basement membrane, where the percentage of invading cells was effectively reduced by 60% compared with control cells. The cell line Miapaca-2 is considered a highly metastatic cell line (60), and our results strongly support the idea that regulating MMP-9 expression may be a key to effectively controlling human pancreatic cancer cell invasion and metastasis.

172 159 PPARβ/δ activation has been associated with reduced inflammatory and adhesion cell markers in several studies, and indeed our results show that in the Miapaca-2 cell line the PPARβ/δ:BCL-6 anti-inflammatory pathway is active and represses the target genes E-selectin, ICAM-1, VCAM-1, IL-1β, MCP-1 and MMP-9 following GW treatment. Repression of the genes VCAM-1 (36) and MCP-1 (15) in particular are dependent upon dissociation of BCL-6 from its complex with PPARβ/δ and subsequent relocation to the corresponding promoters. Our results demonstrate that the pro-adhesion molecules E-selectin, ICAM-1 and VCAM-1, the pro-inflammatory IL-1β and MCP-1 and the pro-metastasis gene MMP-9 are BCL-6 regulated genes in human pancreatic cancer cells. Treatment with shrna against BCL-6 attenuated the GW mediated inhibition of TNFα-induced expression of these molecules, while treatment with a PPARβ/δ shrna significantly reduced their expression at the mrna level regardless of GW treatment. The expression of the cell adhesion molecules E-selectin, ICAM-1 and VCAM-1 is of importance in pancreatic cancer, where they may participate in the detachment of cells from the primary tumor and contribute to cancer spread (61), and their over-expression in pancreatic adenocarcinomas is associated with a stimulation in tumor growth, increased metastatic ability and potentially shorter post-operative survival following tumor resection (62). IL-1β induces MMP-9 expression in other cell lines (63-65), and enhances the invasiveness of human pancreatic cancer cells (66). Monocyte chemotactic protein-1 (MCP-1) is produced by pancreatic cancer cells in response to TNFα challenge and may contribute to the accumulation of tumor-associated macrophages (67), which have also been shown to influence key events in the tumor invasion process (68). Taken together, these results suggest that PPARβ/δ activation and

173 160 subsequent suppression of pro-adhesion and pro-migratory genes via BCL-6 might prove useful in the control of pancreatic cancer cell invasion and metastasis. GW treatment also inhibited the TNFα-promoted invasion of a basement membrane by the pancreatic ductal cell lines Miapaca-2 and BxPc-3. Pancreatic cancer cells transiently expressing BCL-6 shrna were significantly more invasive, while GW treatment attenuated their invasive potential close to control levels. Conversely, cells transiently expressing PPARβ/δ shrna were significantly less invasive than control cells, and GW treatment showed no significant effect in further reducing invasion. We hypothesize that cells expressing lower levels of the transcriptional repressor BCL-6 are more invasive owing to a lack of control over promigratory gene regulation and increased protein levels of MMP-9, while knocking down PPARβ/δ via shrna methods allows for a greater population of free, un-associated BCL-6 which is available to repress pro-migratory genes resulting in significantly lower invasion. Although we present here a fairly simplified mechanism, it is possible that more complex signaling cascade is taking place resulting in the inhibition of cell invasion. In VSMCs, repression of MMP-9 activity was effected in a TGF-β-dependent manner following PPARβ/δ activation (23), and indeed, TGF-β suppresses MMP-9 expression in monocytes through a prostaglandin E2- and camp-dependent mechanism (69). TGF-β is a PPARβ/δ target gene in VSMCs (70), and it is possible that the PPARβ/δ - TGF-β signaling pathway could in fact be active in human pancreatic cancer cells. Our results do suggest, however, that BCL-6 and PPARβ/δ play critical roles in suppressing both pro-migratory gene expression at the mrna level and in ultimately controlling human pancreatic cancer cell invasion.

174 161 Taken together, our observations show that ligand-activation of PPARβ/δ by a specific agonist reduces the TNFα-induced mrna levels of several genes known to be involved in the regulation of human pancreatic cancer cell invasion and metastasis, and this negative regulation is also manifested at the protein level, as demonstrated by MMP- 9 ELISA. Furthermore, we show that PPARβ/δ activation further reduces Miapaca-2 and BxPc-3 invasion through a basement membrane, and that the transcriptional repressor BCL-6 plays a critical role in the pathway(s) regulating human pancreatic cancer cell invasion. This is not the first time PPAR activators have been shown to negatively influence pancreatic cancer cell invasion, and perhaps further in vivo studies using these mouse-transplantable cell lines could provide more useful insight into the potential therapeutic uses of PPARβ/δ activators in the control and regulation of pancreatic cancer. 4.6 References 1. Maitra A, Hruban RH. Pancreatic cancer. Annu Rev Pathol 2008;3: Lowenfels AB, Maisonneuve P. Epidemiology and risk factors for pancreatic cancer. Best Pract Res Clin Gastroenterol 2006;20: Michaud DS, Liu S, Giovannucci E, Willett WC, Colditz GA, Fuchs CS. Dietary sugar, glycemic load, and pancreatic cancer risk in a prospective study. J Natl Cancer Inst 2002;94: Michaud DS, Vrieling A, Jiao L, et al. Alcohol intake and pancreatic cancer: a pooled analysis from the pancreatic cancer cohort consortium (PanScan). Cancer Causes Control. 5. Everhart J, Wright D. Diabetes mellitus as a risk factor for pancreatic cancer. A meta-analysis. JAMA 1995;273: Raimondi S, Maisonneuve P, Lowenfels AB. Epidemiology of pancreatic cancer: an overview. Nat Rev Gastroenterol Hepatol 2009;6: Adrian TE. Inhibition of pancreatic cancer cell growth. Cell Mol Life Sci 2007;64: Goldstein D, Carroll S, Apte M, Keogh G. Modern management of pancreatic carcinoma. Intern Med J 2004;34:

175 9. Chawla A, Repa JJ, Evans RM, Mangelsdorf DJ. Nuclear receptors and lipid physiology: opening the X-files. Science 2001;294: Lemberger T, Desvergne B, Wahli W. Peroxisome proliferator-activated receptors: a nuclear receptor signaling pathway in lipid physiology. Annu Rev Cell Dev Biol 1996;12: Fredenrich A, Grimaldi PA. PPAR delta: an uncompletely known nuclear receptor. Diabetes Metab 2005;31: Cimini A, Ceru MP. Emerging roles of peroxisome proliferator-activated receptors (PPARs) in the regulation of neural stem cells proliferation and differentiation. Stem Cell Rev 2008;4: Michalik L, Desvergne B, Wahli W. Peroxisome proliferator-activated receptors beta/delta: emerging roles for a previously neglected third family member. Curr Opin Lipidol 2003;14: Bishop-Bailey D, Bystrom J. Emerging roles of peroxisome proliferator-activated receptor-beta/delta in inflammation. Pharmacol Ther 2009;124: Lee CH, Chawla A, Urbiztondo N, et al. Transcriptional repression of atherogenic inflammation: modulation by PPARdelta. Science 2003;302: Kharroubi I, Lee CH, Hekerman P, et al. BCL-6: a possible missing link for antiinflammatory PPAR-delta signalling in pancreatic beta cells. Diabetologia 2006;49: Higashiyama H, Billin AN, Okamoto Y, Kinoshita M, Asano S. Expression profiling of peroxisome proliferator-activated receptor-delta (PPAR-delta) in mouse tissues using tissue microarray. Histochem Cell Biol 2007;127: Pandol S, Edderkaoui M, Gukovsky I, Lugea A, Gukovskaya A. Desmoplasia of pancreatic ductal adenocarcinoma. Clin Gastroenterol Hepatol 2009;7:S Halbersztadt A, Halon A, Pajak J, Robaczynski J, Rabczynski J, St Gabrys M. [The role of matrix metalloproteinases in tumor invasion and metastasis]. Ginekol Pol 2006;77: Deryugina EI, Quigley JP. Matrix metalloproteinases and tumor metastasis. Cancer Metastasis Rev 2006;25: Bloomston M, Zervos EE, Rosemurgy AS, 2nd. Matrix metalloproteinases and their role in pancreatic cancer: a review of preclinical studies and clinical trials. Ann Surg Oncol 2002;9: Roomi MW, Monterrey JC, Kalinovsky T, Rath M, Niedzwiecki A. Patterns of MMP-2 and MMP-9 expression in human cancer cell lines. Oncol Rep 2009;21: Kim HJ, Kim MY, Hwang JS, et al. PPARdelta inhibits IL-1beta-stimulated proliferation and migration of vascular smooth muscle cells via up-regulation of IL-1Ra. Cell Mol Life Sci. 24. Nakajima A, Tomimoto A, Fujita K, et al. Inhibition of peroxisome proliferatoractivated receptor gamma activity suppresses pancreatic cancer cell motility. Cancer Sci 2008;99: Farrow B, Evers BM. Activation of PPARgamma increases PTEN expression in pancreatic cancer cells. Biochem Biophys Res Commun 2003;301:

176 26. Hashimoto K, Farrow BJ, Evers BM. Activation and role of MAP kinases in 15d- PGJ2-induced apoptosis in the human pancreatic cancer cell line MIA PaCa-2. Pancreas 2004;28: He TC, Chan TA, Vogelstein B, Kinzler KW. PPARdelta is an APC-regulated target of nonsteroidal anti-inflammatory drugs. Cell 1999;99: Park BH, Vogelstein B, Kinzler KW. Genetic disruption of PPARdelta decreases the tumorigenicity of human colon cancer cells. Proc Natl Acad Sci U S A 2001;98: Bastie C. [PPARdelta and PPARgamma: roles in fatty acids signalling, implication in tumorigenesis]. Bull Cancer 2002;89: Glinghammar B, Skogsberg J, Hamsten A, Ehrenborg E. PPARdelta activation induces COX-2 gene expression and cell proliferation in human hepatocellular carcinoma cells. Biochem Biophys Res Commun 2003;308: Romanowska M, al Yacoub N, Seidel H, et al. PPARdelta enhances keratinocyte proliferation in psoriasis and induces heparin-binding EGF-like growth factor. J Invest Dermatol 2008;128: Burdick AD, Bility MT, Girroir EE, et al. Ligand activation of peroxisome proliferator-activated receptor-beta/delta(pparbeta/delta) inhibits cell growth of human N/TERT-1 keratinocytes. Cell Signal 2007;19: Hollingshead HE, Killins RL, Borland MG, et al. Peroxisome proliferatoractivated receptor-beta/delta (PPARbeta/delta) ligands do not potentiate growth of human cancer cell lines. Carcinogenesis 2007;28: Kilian M, Gregor JI, Heukamp I, et al. Matrix metalloproteinase inhibitor RO decreases liver metastasis by reduction of MMP-2 and MMP-9 concentration in BOP-induced ductal pancreatic cancer in Syrian Hamsters: inhibition of matrix metalloproteinases in pancreatic cancer. Prostaglandins Leukot Essent Fatty Acids 2006;75: Tian M, Cui YZ, Song GH, et al. Proteomic analysis identifies MMP-9, DJ-1 and A1BG as overexpressed proteins in pancreatic juice from pancreatic ductal adenocarcinoma patients. BMC Cancer 2008;8: Fan Y, Wang Y, Tang Z, et al. Suppression of pro-inflammatory adhesion molecules by PPAR-delta in human vascular endothelial cells. Arterioscler Thromb Vasc Biol 2008;28: Stephen RL, Gustafsson MC, Jarvis M, et al. Activation of peroxisome proliferator-activated receptor delta stimulates the proliferation of human breast and prostate cancer cell lines. Cancer Res 2004;64: Kawa S, Nikaido T, Unno H, Usuda N, Nakayama K, Kiyosawa K. Growth inhibition and differentiation of pancreatic cancer cell lines by PPAR gamma ligand troglitazone. Pancreas 2002;24: Eibl G. The Role of PPAR-gamma and Its Interaction with COX-2 in Pancreatic Cancer. PPAR Res 2008;2008: Mackenzie GG, Rasheed S, Wertheim W, Rigas B. NO-Donating NSAIDs, PPARdelta, and Cancer: Does PPARdelta Contribute to Colon Carcinogenesis? PPAR Res 2008;2008:

177 41. Marin HE, Peraza MA, Billin AN, et al. Ligand activation of peroxisome proliferator-activated receptor beta inhibits colon carcinogenesis. Cancer Res 2006;66: Kim DJ, Bility MT, Billin AN, Willson TM, Gonzalez FJ, Peters JM. PPARbeta/delta selectively induces differentiation and inhibits cell proliferation. Cell Death Differ 2006;13: Kim DJ, Prabhu KS, Gonzalez FJ, Peters JM. Inhibition of chemically induced skin carcinogenesis by sulindac is independent of peroxisome proliferatoractivated receptor-beta/delta (PPARbeta/delta). Carcinogenesis 2006;27: Yang X, Staren ED, Howard JM, Iwamura T, Bartsch JE, Appert HE. Invasiveness and MMP expression in pancreatic carcinoma. J Surg Res 2001;98: Maatta M, Soini Y, Liakka A, Autio-Harmainen H. Differential expression of matrix metalloproteinase (MMP)-2, MMP-9, and membrane type 1-MMP in hepatocellular and pancreatic adenocarcinoma: implications for tumor progression and clinical prognosis. Clin Cancer Res 2000;6: Evans JD, Ghaneh P, Kawesha A, Neoptolemos JP. Role of matrix metalloproteinases and their inhibitors in pancreatic cancer. Digestion 1997;58: Lee CS, Montebello J, Georgiou T, Rode J. Distribution of type IV collagen in pancreatic adenocarcinoma and chronic pancreatitis. Int J Exp Pathol 1994;75: Senior RM, Griffin GL, Fliszar CJ, Shapiro SD, Goldberg GI, Welgus HG. Human 92- and 72-kilodalton type IV collagenases are elastases. J Biol Chem 1991;266: Guo K, Ma Q, Wang L, et al. Norepinephrine-induced invasion by pancreatic cancer cells is inhibited by propranolol. Oncol Rep 2009;22: Takayama O, Yamamoto H, Damdinsuren B, et al. Expression of PPARdelta in multistage carcinogenesis of the colorectum: implications of malignant cancer morphology. Br J Cancer 2006;95: Tan NS, Michalik L, Di-Poi N, et al. Essential role of Smad3 in the inhibition of inflammation-induced PPARbeta/delta expression. EMBO J 2004;23: Glauser DA, Schlegel W. The FoxO/Bcl-6/cyclin D2 pathway mediates metabolic and growth factor stimulation of proliferation in Min6 pancreatic beta-cells. J Recept Signal Transduct Res 2009;29: Toney LM, Cattoretti G, Graf JA, et al. BCL-6 regulates chemokine gene transcription in macrophages. Nat Immunol 2000;1: Wlodarska I, Nooyen P, Maes B, et al. Frequent occurrence of BCL6 rearrangements in nodular lymphocyte predominance Hodgkin lymphoma but not in classical Hodgkin lymphoma. Blood 2003;101: Tibiletti MG, Martin V, Bernasconi B, et al. BCL2, BCL6, MYC, MALT 1, and BCL10 rearrangements in nodal diffuse large B-cell lymphomas: a multicenter evaluation of a new set of fluorescent in situ hybridization probes and correlation with clinical outcome. Hum Pathol 2009;40:

178 56. Tzankov A, Schneider A, Hoeller S, Dirnhofer S. Prognostic importance of BCL6 rearrangements in diffuse large B-cell lymphoma with respect to Bcl6 protein levels and primary lymphoma site. Hum Pathol 2009;40:1055-6; author reply Shvarts A, Brummelkamp TR, Scheeren F, et al. A senescence rescue screen identifies BCL6 as an inhibitor of anti-proliferative p19(arf)-p53 signaling. Genes Dev 2002;16: Holzmann K, Kohlhammer H, Schwaenen C, et al. Genomic DNA-chip hybridization reveals a higher incidence of genomic amplifications in pancreatic cancer than conventional comparative genomic hybridization and leads to the identification of novel candidate genes. Cancer Res 2004;64: Perez-Rosado A, Artiga M, Vargiu P, Sanchez-Aguilera A, Alvarez-Barrientos A, Piris M. BCL6 represses NFkappaB activity in diffuse large B-cell lymphomas. J Pathol 2008;214: Takada M, Hirata K, Ajiki T, Suzuki Y, Kuroda Y. Expression of receptor for advanced glycation end products (RAGE) and MMP-9 in human pancreatic cancer cells. Hepatogastroenterology 2004;51: Tempia-Caliera AA, Horvath LZ, Zimmermann A, et al. Adhesion molecules in human pancreatic cancer. J Surg Oncol 2002;79: Friess H, Berberat P, Schilling M, Kunz J, Korc M, Buchler MW. Pancreatic cancer: the potential clinical relevance of alterations in growth factors and their receptors. J Mol Med 1996;74: Wu CY, Hsieh HL, Sun CC, Yang CM. IL-1beta induces MMP-9 expression via a Ca2+-dependent CaMKII/JNK/c-JUN cascade in rat brain astrocytes. Glia 2009;57: Gordon GM, Ledee DR, Feuer WJ, Fini ME. Cytokines and signaling pathways regulating matrix metalloproteinase-9 (MMP-9) expression in corneal epithelial cells. J Cell Physiol 2009;221: Li L, Xing FQ, Chen SL. [Role of interleukin-1beta in regulating human cultured endometrial cell MMP-9 and TIMP-3 expressions in the mid-secretory phase]. Nan Fang Yi Ke Da Xue Xue Bao 2006;26: Greco E, Basso D, Fogar P, et al. Pancreatic cancer cells invasiveness is mainly affected by interleukin-1beta not by transforming growth factor-beta1. Int J Biol Markers 2005;20: Takaya H, Andoh A, Shimada M, Hata K, Fujiyama Y, Bamba T. The expression of chemokine genes correlates with nuclear factor-kappab activation in human pancreatic cancer cell lines. Pancreas 2000;21: Solinas G, Marchesi F, Garlanda C, Mantovani A, Allavena P. Inflammationmediated promotion of invasion and metastasis. Cancer Metastasis Rev. 69. Vaday GG, Schor H, Rahat MA, Lahat N, Lider O. Transforming growth factorbeta suppresses tumor necrosis factor alpha-induced matrix metalloproteinase-9 expression in monocytes. J Leukoc Biol 2001;69: Kim HJ, Ham SA, Kim SU, et al. Transforming growth factor-beta1 is a molecular target for the peroxisome proliferator-activated receptor delta. Circ Res 2008;102:

179 Chapter 5 Summary and Perspectives The PPARs are master regulators of several cellular pathways, including cell growth and proliferation, invasion, energy homeostasis, and inflammation and specific small molecule activators of the three PPAR isotypes are used in the treatment of various diseases. The fibrate class of hypolipidemic drugs, activators of PPARα, are currently in use for treating dyslipidemia. PPARγ activators, including the thiazolidinediones, are used clinically in the management of insulin sensitivity and diabetes mellitus. Selective PPARβ/δ ligands have only recently become available. GW and L are two PPARβ/δ-specific activators that have been used clinically to increase HDL levels and decrease triglyceride and insulin levels, suggesting that specific activators of this subtype are of therapeutic use in the management of the metabolic syndrome. Furthermore, the association of PPARβ/δ with BCL-6, a physical protein-protein interaction unique to this member of the NR1C family, has been identified, providing a key insight into the anti-inflammatory properties of PPARβ/δ-specific activators. Unlike PPARα or PPARγ, however, the role(s) of PPARβ/δ is not conclusively established. Activators of PPARβ/δ are effective in controlling inflammation, cell proliferation and cell invasion both in vitro and in vivo, and further clinical research efforts that examine

180 167 the potential therapeutic uses of PPARβ/δ-specific ligands will benefit from an increased understanding of the mechanism(s) that govern their action. In general, the PPARs must be activated by ligand to effect the transcription of target genes. Several endogenous ligands of PPARβ/δ have been identified (free fatty acids, triglycerides, prostacyclin, and retinoic acid (1)). The PPARβ/δ isotype is established as a very low-density lipoprotein (VLDL) sensor in macrophages (2). Triglycerides, cleaved from the VLDL particle by the actions of lipoprotein lipase (LpL) are the transcriptionally active components, inducing the expression of PPARβ/δ target genes (such as ADRP) in a PPARβ/δ-dependent manner. In the absence of the receptor, the transcriptional activity induced by VLDL is abolished. The oxidation of these lipoproteins is implicated in the progression of various inflammation-related disorders, including atherosclerosis (3), and previous studies have demonstrated that PPARγ is a sensor of oxidized lipids (4). Exposure to oxidized lipoproteins induces the expression of pro-inflammatory genes (such as TNFα, IL-1α, IL-1β, and IL-6 (4)), and enhances macrophage chemotaxis towards the site(s) of oxidative insult. Although low-density lipoproteins and their oxidized counterparts were traditionally regarded as the only risk factors for atherogenesis, recent human studies suggested similar roles for VLDL (5). Given that the lipids in LDL, oxldl and VLDL activate the PPARs, and are relevant in inflammation-related disorders, examining the relationships between PPARβ/δ, oxidized lipids and inflammation is prudent. Furthermore, the PPARβ/δ null mouse is susceptible to various injuries following oxidative insult (i.e. increased hepatotoxicity) compared

181 168 with wild-type mice, implying that the receptor plays a critical role in the detection and amelioration of toxic lipids and inflammation (6). Studies in this dissertation have characterized the oxidative stress mediator 4- hydroxy-2-nonenal (4-HNE) as an endogenous PPARβ/δ activator. The oxidation of native VLDL (oxvldl) and subsequent treatment with LpL enhanced PPARβ/δ reporter activity in a dose-dependent manner. Screening various fatty acids and their oxidized counterparts identified 13-(S)-HODE as a potent PPARβ/δ activator, while oxidized 13- (S)-HODE significantly increased PPARβ/δ reporter activity compared to the native fatty acid. We hypothesized that 13-(S)-HODE was a component of oxvldl, and, indeed, the concentration of 13-(S)-HODE increased correspondingly with increases in oxvldl as measured by ELISA. The oxidation of these lipids usually proceeds via non-enzymatic free radical chain reaction, producing toxic aldehydic end products. One proposed mechanism, a Hock-cleavage reaction, produces the toxic oxidative stress mediator 4- HNE from both 13-(S)-HODE and its hydroperoxy derivative, 13-(S)-HpODE (7). Structure-activity relationship experiments using representative compounds from each family of oxidized lipids (2-alkenals, 4-hydroxy-2-alkenals, and ketoaldehydes) confirmed that 4-HNE dose-dependently activated PPARβ/δ at sub-lethal levels, and did so to a greater extent than its hydroperoxy precursor, 4-HpNE. Furthermore, the activation of PPARβ/δ by 4-HNE induced the expression of various PPARβ/δ target genes, confirmed by using both wild-type and PPARβ/δ-/- mouse hepatocytes. Microarray analyses revealed that 4-HNE induced increases in the expression of three genes involved in detoxification (Aldh3a1, 22-fold increase, Gstm3, 2.3-fold increase and

182 169 Gsto1, 2.0-fold increase) in a PPARβ/δ-dependent manner. These results indicated that PPARβ/δ is an oxidized lipid sensor, and plays a key role in ameliorating oxidative insults via direct induction of anti-oxidant target genes. Consistent with this hypothesis, wild-type mouse hepatocytes were significantly less sensitive to 4-HNE insult (EC µm) compared to PPARβ/δ-/- cells (EC 50 5 µm) in cell viability assays. Since interactions between 4-HNE and a nuclear receptor had not previously been reported, we sought to map potential 4-HNE / PPARβ/δ interactions. Western blot analyses were unable to demonstrate a covalent modification of PPARβ/δ, but molecular modeling did reveal a putative hydrogen bonding interaction between the hydroxy group of 4-HNE and His413 in the PPARβ/δ ligand binding pocket. These results suggested that 4-HNE activates PPARβ/δ via a reversible interaction, and this modulation of receptor activity may play an important role in mediating the detoxification of oxidized lipids and subsequent protection of target cells from oxidative stress. The induction of detoxification genes by a PPARβ/δ ligand suggested a role for the receptor in mediating the inflammatory process. Indeed, several recent studies have highlighted the regulation of the inflammatory response as a major role for the receptor in several cell types and animal models (8-11). The anti-inflammatory properties of PPARβ/δ-specific activators proceeds via three general mechanisms: 1) the direct induction of anti-inflammatory PPARβ/δ target genes, 2) the inhibition of proinflammatory signaling pathways (ie. NF-κB) and 3) the ligand-induced dissociation of the transcriptional repressor, BCL-6 (8), a function unique to the PPARβ/δ subtype (9). As mentioned previously, PPARβ/δ is ubiquitously expressed, but its activity in the

183 170 pancreas remains largely unexamined. Activation of PPARβ/δ is anti-inflammatory in macrophages (9), colon epithelium (12) and vascular endothelial cells (13). In pancreatic beta cells, however, PPARβ/δ-specific ligands do not induce the anti-inflammatory response, owing to a lack of BCL-6 expression (14). To date, no studies have examined the role(s) of PPARβ/δ and BCL-6 in regulating inflammation in pancreatic ductal cells, the predominant site for cancer production in humans. Unchecked inflammation and chronic pancreatitis in these cells increases the risk of developing pancreatic cancer by 10- to 20-fold (15). Studies presented herein demonstrate that PPARβ/δ is expressed and functional in pancreatic cancer cells of ductal origin (Miapaca-2 and BxPc-3) and that this receptor has anti-inflammatory effects. Forced over-expression of PPARβ/δ inhibited both basal and TNFα-induced NF-κB reporter activity. Inhibition of NF-κB by PPARβ/δ is reported to occur via several mechanisms including inhibition of ERK1/2 map kinase phosphorylation (16). In pancreatic cancer cells, PPARβ/δ inhibition of NF-κB activity proceeds via direct physical interaction between the receptor and the p50 subunit of NFκB. Mammalian 2-hybrid assays demonstrated that this interaction is enhanced upon GW treatment. In NF-κB reporter assays, PPARβ/δ activation was sufficient to inhibit basal and TNFα-induced NF-κB activity in control cells. Interactions between BCL-6 and NF-κB have been described both in vitro and in vivo (17). Knock-down of PPARβ/δ and BCL-6 using shrnas indicated that the observed repression of NF-κB by GW may require both the ligand-sensitivity of PPARβ/δ and the presence of BCL-6 to negatively affect reporter activity.

184 171 The anti-inflammatory properties of the PPARβ/δ-specific ligand GW were also examined in the context of gene expression. Treatment of pancreatic cancer cells with GW induced the expression of several anti-inflammatory genes (IL-1 receptor antagonist, TGF-β, SOD-1 and FGF-21) in a PPARβ/δ-dependent manner. Proinflammatory MCP-1 reporter activity and gene expression was also significantly reduced in a BCL-6-dependent manner following PPARβ/δ activation. In pancreatic duct cells expressing IL-6 and COX-2 (BxPc-3), PPARβ/δ activation was not sufficient to reduce the expression of these genes. Knock-down of BCL-6 increased the inflammatory response, while knock-down of PPARβ/δ resulted in lower mrna expression of IL-6, suggesting that the GW regulation is effected via BCL-6. Taken together, these results indicate that the PPARβ/δ / BCL-6 anti-inflammatory pathway is active in pancreatic duct cells and responds to PPARβ/δ activation. The ability of PPARβ/δspecific activators to reduce the inflammatory response in the pancreas in vivo is supported by these observations, although this will require further examination. Conditioned media experiments presented in this dissertation further examined the potential cross-talk between human pancreatic cancer cells and macrophages. Tumorassociated macrophages are implicated in enhancing tumor growth and spread, and also allowing it to evade immune surveillance (18). Treatment of control, PPARβ/δ, BCL-6 and IL-1Ra knock-down human pancreatic cancer cells with GW significantly reduced pro-inflammatory gene expression in macrophages, supporting the notion that PPARβ/δ ligands will be effective in both reducing the inflammatory burden via crosstalk between pancreatic carcinomas and infiltrating macrophages. Furthermore,

185 172 GW conditioned media from pancreatic cancer cells attracted significantly fewer THP-1 invading macrophages compared with TNFα-stimulated pancreatic cancer cells in cell invasion assays using a basement membrane. These results suggest that PPARβ/δ activators might play a therapeutic role in controlling local pancreatic inflammation, such as pancreatitis, but also in reducing the overall inflammatory burden by reducing the attraction of circulating monocytes and macrophages. The final studies in this dissertation examine a potential role for PPARβ/δ and BCL-6 in regulating pancreatic cancer cell invasion. Human pancreatic cancer, chronic pancreatitis and pancreatic tissue samples were screened for the expression of several genes involved in inflammation and pancreatic cancer metastasis using real-time PCR techniques. Ductal carcinomas expressed approximately 10-fold higher mrna levels of the pro-invasion MMP-9 gene compared with chronic pancreatitis. This is consistent with previous reports indicating that MMP-9 is typically over-expressed in pancreatic cancers (19). MMP-9 is a key player in the early stages of tumor development and invasion by degrading basement membrane Type IV collagen (20). Studies have linked increased MMP-9 expression to increased tumor invasion and metastasis (21). Interestingly, BCL-6 was significantly repressed in pancreatic ductal carinomas compared to pancreatitis, while PPARβ/δ was over-expressed. Although there is conflicting evidence regarding the role of PPARβ/δ in tumorigenesis, its expression is potentially controlled via AP-1 transcription factors. The regulatory region of PPARβ/δ contains several AP-1 response elements (22), and its over-expression in several cancers (23) and in ductal carcinomas may be the result of an exacerbated stress response and not

186 173 causally related to tumor formation. BCL-6, on the other hand, is a proto-oncogene that is typically mutated in various disorders (24). BCL-6 controls genes that are involved in cell cycle progression (25), and these results suggest a disruption of BCL-6 signaling as inflamed pancreatic tissue converts to tumor. Of particular note is the fact that BCL-6, PPARβ/δ and MMP-9 have been associated in macrophages (9). PPARβ/δ null macrophages display significantly decreased MMP-9 levels, potentially due to a lack of BCL-6 sequestration in the absence of the receptor. These results suggest that pancreatic ductal carcinomas have less ability for BCL-6 to repress MMP-9 due to increased PPARβ/δ (which sequesters BCL-6 and is considered pro-inflammatory) and mutations in the BCL-6 gene itself. Mutational analyses of these tissues could provide insight into whether or not the proto-oncogene BCL-6 is altered by genetic aberrations in pancreatic carcinomas. The role of MMP-9 in cell invasion and metastasis is clearly established (26-28). Studies in this dissertation examined the roles of PPARβ/δ and BCL-6 in the regulation of MMP-9 mrna and protein expression. In control human pancreatic cancer cells, GW treatment resulted in a significant decrease in TNFα-induced MMP-9 protein as measured by ELISA. Knock-down of BCL-6 resulted in a significant increase in TNFα-induced MMP-9 protein despite GW treatment, suggesting that the transcriptional repressor plays a role in the regulation of MMP-9. Conversely, knockdown of PPARβ/δ resulted in a significant decrease in MMP-9 protein to levels comparable to GW treated control cells. These results indicated that ligandactivated PPARβ/δ releases BCL-6, leading to subsequent repression of MMP-9

187 174 expression. In the absence of the receptor (PPARβ/δ knock-down or PPARβ/δ -/- ), BCL-6 is free to repress MMP-9 without PPARβ/δ activation. The use of shrnas targeted against MMP-9 resulted in a significant decrease (approximately 50%) in invading pancreatic cancer cells in cell invasion assays, suggesting that the regulation of MMP-9 levels in human pancreatic cancer cells is important in controlling invasion and, potentially, metastasis. Studies in this dissertation also examined the potential for PPARβ/δ-specific activators to inhibit pro-migratory gene expression in human pancreatic cancer cells. Indeed, GW significantly reduced the mrna expression of E-selectin, ICAM-1, VCAM-1, IL-1β, MCP-1 and MMP-9 in control cells. Cells infected with BCL-6 shrna expressed much higher mrna levels of these genes despite GW treatment. Conversely, pancreatic cancer cells expressing PPARβ/δ shrna expressed significantly lower mrna levels of these genes regardless of TNFα or GW treatment, again suggesting that the regulation of theses genes is effected via BCL-6. Finally, the effects of PPARβ/δ activation on the invasive potential of two human pancreatic cancer cell lines was examined. Miapaca-2 (COX-2 negative) and BxPc-3 (COX-2 positive), and their corresponding BCL-6- and PPARβ/δ-knock-down counterparts were treated with GW and cell invasion was measured. In both control Miapaca-2 and BxPc-3 cells, GW reduced the percent invading cells by approximately 50%. BCL-6 knock-down cells were significantly more invasive than control cells, and GW was only effective in BxPc-3 cells in reducing invasive potential. PPARβ/δ knock-down cells were significantly less invasive than control cells

188 175 despite GW treatment. To date, no in vivo studies have been performed that examined the role of PPARβ/δ and/or BCL-6 in the pancreas. These results strongly support the notion that PPARβ/δ and BCL-6 are anti-inflammatory in the human pancreas, and that specific activation of the receptor presents a potentially effective therapeutic option for controlling pancreatic cancer cell invasion and metastasis. In summary, this dissertation characterizes the effects of PPARβ/δ activation by endogenous and synthetic ligands on the anti-oxidant and inflammatory response, and supports the notion of PPARβ/δ as a novel molecular target for the treatment and control of inflammatory disorders and pancreatic cancer. Figure 5-1 highlights the significant aspects of this pathway.

189 Figure 5-1: Proposed model for the regulatory cascade resulting from PPARβ/δ activation by endogenous and synthetic ligands. Nuclear PPARβ/δ sequesters the transcriptional repressor BCL-6. Endogenous ligands (intracellular fatty acids and their metabolites) and synthetic agonists (GW501516) induce the dissociation of the PPARβ/δ / BCL-6 protein complex and subsequent heterodimerization with RXR. Activated PPARβ/δ recruits co-activator complexes and induces target gene transcription, and/or physically interacts with and inhibits the activity of NF-κB. Free BCL-6 physically associates with the regulatory region of target genes and NF-κB. Together, PPARβ/δ and BCL-6 mediate the anti-cancer and anti-inflammatory properties of PPARβ/δ activators. 176