In vitro and in vivo pharmacology of NXT629, a novel and selective PPARα antagonist

Author’s Accepted Manuscript In vitro and in vivo pharmacology of NXT629, a novel and selective PPARα antagonist Karin J. Stebbins, Alex R. Broadhead, Geraldine Cabrera, Lucia D. Correa, Davorka Messmer, Richard Bundey, Christopher Baccei, Yalda Bravo, Austin Chen, Nicholas S. Stock, Peppi Prasit, Daniel S. Lorrain

PII: DOI: Reference:

www.elsevier.com/locate/ejphar

S0014-2999(17)30322-9 http://dx.doi.org/10.1016/j.ejphar.2017.05.008 EJP71198

To appear in: European Journal of Pharmacology Received date: 12 December 2016 Revised date: 2 May 2017 Accepted date: 4 May 2017 Cite this article as: Karin J. Stebbins, Alex R. Broadhead, Geraldine Cabrera, Lucia D. Correa, Davorka Messmer, Richard Bundey, Christopher Baccei, Yalda Bravo, Austin Chen, Nicholas S. Stock, Peppi Prasit and Daniel S. Lorrain, In vitro and in vivo pharmacology of NXT629, a novel and selective PPARα a n t a g o n i s t , European Journal of Pharmacology, http://dx.doi.org/10.1016/j.ejphar.2017.05.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

In vitro and in vivo pharmacology of NXT629, a novel and selective PPARa antagonist¶

Karin J. Stebbins*, Alex R. Broadhead, Geraldine Cabrera, Lucia D. Correa, Davorka Messmer, Richard Bundey, Christopher Baccei, Yalda Bravo, Austin Chen, Nicholas S. Stock, Peppi Prasit, Daniel S. Lorrain Inception Sciences, 5871 Oberlin Drive, Suite 100, San Diego, CA 92121 *

Correspondence to: Inception Sciences, 5871 Oberlin Drive, suite 100, San Diego, CA 92121. Tel.: +858 224 7714. [email protected]

Abstract Peroxisome-proliferator activated receptors (PPAR) are members of the nuclear hormone receptor superfamily which regulate gene transcription. PPARa is a key regulator of lipid homeostasis and a negative regulator of inflammation. Under conditions of metabolic stress such as fasting or glucose deprivation, PPARa is upregulated in order to control gene expression necessary for processing alternate fuel sources (e.g. fatty acid oxidation) and thereby promote maintenance of cell viability. Clinically, PPARa expression is upregulated in diseased tissues such as melanoma, chronic lymphocytic leukemia, ovarian and prostate cancer. This may allow for cellular proliferation and metastasis. Importantly, genetic knockouts of PPARa have been shown to be protected against tumor growth in a variety of syngeneic tumors models. We hypothesized that a potent and selective PPARa antagonist could represent a novel cancer therapy. Early in our discovery research, we identified NXT629 (Bravo et al., 2014). Herein we describe the pharmacology of NXT629 and demonstrate that it is a potent and selective PPARa antagonist. We identify NXT629 as a valuable tool for use in in vivo assessment of PPARa due

This work did not receive any specific grant from funding agencies in the public, commercial or not-for profit sectors. ¶

to its good systemic exposure following intraperitoneal injection. We explore the in vivo pharmacology of NXT629 and demonstrate that it is efficacious in pharmacodynamic models that are driven by PPARa. Finally, we probe the efficacy of NXT629 in disease models where PPARa knockouts have shown to be protected. We believe that PPARa antagonists will be beneficial in diseases such as ovarian cancer and melanoma where PPARa and fatty acid oxidation may be involved.

Keywords PPARa, PPARa antagonism, fatty acid oxidation, cancer

1. Introduction Peroxisome proliferator activated receptor alpha (PPARa), a member of the nuclear hormone receptor superfamily, exists as a heterodimer with retinoid X receptor (RXR). Upon ligand binding and recruitment of the co-activator protein, the resulting ternary complex binds to the peroxisomal proliferator response element (PPRE) on the promoter region of the target genes and drives gene transcription. Endogenous ligands of PPARa include oleoylethanolamide (OEA) (Fu et al., 2003), which is reported to bind with nanomolar potency (Berger and Moller, 2002), as well as long chain fatty acids and leukotriene B4 (LTB4) which have micromolar potencies (Devchand et al., 1996). Numerous synthetic pharmacological agonists for PPARa have been described including fibrates, WY-14,643 and GW7647 (Willson et al., 2000). PPARa, a key regulator of lipid homeostasis and a negative regulator of inflammation, is widely expressed in normal tissue with predominance in the liver, kidney, brown adipose tissue, skeletal muscle and heart. Expression is upregulated under conditions of metabolic stress such as fasting 1

or glucose deprivation. This promotes maintenance of cell viability by increasing the expression of genes required for fatty acid oxidation in order to facilitate the use of alternate fuel sources (i.e. lipids rather than glucose). From a cancer perspective, expression of PPARa or its target genes is increased in diseased tissues such as melanoma (Eastham et al., 2008), chronic lymphocytic leukemia (Spaner et al., 2013), ovarian (Nieman et al., 2011) and prostate cancer (Collett et al., 2000). In patients with colorectal liver metastases, PPARa staining is associated with worse overall survival (Pang et al., 2015). In PPARa knockout mice, transcription of genes involved in fatty acid oxidation (eg. Cpt1a, Cact) is impaired. Furthermore, genetic knockouts of PPARa have been shown to be protected against tumor growth in a variety of syngeneic tumors models (Kaipainen et al., 2007) . Thus, we hypothesized that a potent and selective PPARa antagonist could represent a novel cancer therapy. To date, a limited number of PPARa antagonists have been disclosed, with GW6471 (Xu et al., 2002) and MK886 (Kehrer et al., 2001) being the most referenced. While GW6471 has been extensively utilized in vitro, it has been used only sparingly in vivo. MK886 is a potent inhibitor of 5-lipoxygenase activating protein (FLAP) and FLAP has previously been implicated in some oncogenic pathways. However in our luciferase-based reporter assay system, MK886 was cytotoxic at the concentrations at which inhibition of PPARa occurred (data not shown), complicating interpretation of results. Eli Lilly has described an acylsulfonamide series of PPARa antagonists (Etgen and Mantlo, 2003) however, these lack stability in murine plasma and are therefore not amenable for use in vivo. Although our ultimate goal was to identify a molecule with oral bioavailablility, we initially identified a compound with good systemic exposure following intraperitoneal injection, as

2

valuable for use in vivo (Bravo et al., 2014). In the studies described herein, NXT629, a potent and selective PPARa antagonist, was profiled in a panel of in vitro and in vivo pharmacological models.

2. Material and methods 2.1. Chemicals NXT629 (Fig. 1A), NXT969 (Fig. 1B) and GW590735 were synthesized at Inception Sciences. WY-16,643 was purchased from Sigma-Aldrich (St. Louis, MO). For all in vitro experiments, a 30 mM stock solution in DMSO was diluted to the appropriate concentration in assay buffer. In vivo dosing solutions were prepared in sterile saline (NXT629), sterile saline containing 2% Tween 20 (NXT969), or in a 1:1 mixture of DMSO and sesame oil (GW590735 and WY16,643). 2.2. In vitro pharmacology 2.2.1. Functional reporter assay Antagonist activity of test compound against human or mouse PPARα receptors was analyzed using commercial kits (Human or Mouse PPARα Reporter Assay System, Indigo Biosciences, State College, PA). PPARα reporter cells are dispensed into 96-well assay plates followed by immediate addition of test compounds in the presence or absence of the PPARα agonist GW7647 (20 nM). Following an overnight incubation, the treatment media is discarded and Luciferase Detection Reagent was added. The luminescence intensity of light emission from the ensuing luciferase reaction is directly proportional to the relative level of PPARα activation in the

3

reporter cells. Luminescence was read using a Molecular Devices FlexStation 3 (Sunnyvale, CA). To determine whether NXT629 was a competitive antagonist of PPARa, the concentration of both NXT629 and GW7647 were varied in the PPARa reporter assay. Inhibitory concentration response curves were generated for NXT629 at 0, 30, 100 and 300 nM against a variable concentration of GW7647 (0.1 nM to 1 µM). We further explored whether potency of NXT629 shifts with increasing concentrations of agonist. This was accomplished by varying the concentration of NXT629 (1 nM to 10 µM) while maintaining a constant agonist concentration of either 100 nM or 1 µM. 2.2.2. Counterscreens To determine selectivity of NXT629 for PPARa, antagonist activity against human PPARδ, PPARg, estrogen receptor b, glucocorticoid receptor and thyroid receptor b was evaluated. NXT629 was also counterscreened against mouse PPARδ and PPARg. Counterscreens were performed using commercially available reporter assay kits (Indigo Biosciences, State College, PA) according to the manufacturer’s instructions.

2.2.3. Real time-PCR of normal human skin and melanoma In order to examine expression of PPARa, normal healthy or melanoma skin biopsy samples (n=2 donors each) were obtained from ConversantBio (Hunstville, Alabama). Samples were collected into RNALater (Qiagen, Venlo, The Netherlands ). Normal skin and melanoma tumor tissues were weighed then coated and minced while in a solution of Trizol. For every 100 mg of 4

tissue, 1 ml of TRIzol reagent was applied and subjected to homogenization with a Polytron handheld homogenizer. Total RNA was purified according to the manufacturer’s TRIzol extraction protocols (Life Technologies, Carlsbad, CA). The resulting RNA was treated with (1U:25 µl) DNase I (Life Technologies, Carlsbad, CA). Reverse transcription of cDNA was performed using iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA) with 1 µg of RNA. The synthesized cDNA was mixed with primers (Table 3) and SsoFast EvaGreen Supermix With Low ROX (Bio-Rad, Hercules, CA). Real-Time PCR detection on the mix was performed in the CFX96™ Real-Time PCR Detection System (Bio-Rad, Hercules, CA) following manufacturers suggested protocols and using an annealing temperature of 60 °C for 5 seconds.

2.3. In vivo pharmacology 2.3.1. Animals Mice (CD-1, C57BL/6N or Fox1nu females) and jugular vein cannulated Sprague-Dawley rats (male) were obtained from Harlan (Livermore, CA). Animals were given food and water ad libitum and allowed to acclimate for at least 5 days prior to initiation of experiments. All protocols were approved by the Inception Sciences Institutional Animal Care and Use Committee. 2.3.2. Pharmacokinetics Rats were fasted overnight and dosed with NXT629, either per os (p.o.) (10 mg/kg) or intravenously (1 mg/kg). Blood was sampled via the jugular cannula at various time points after

5

drug administration. Plasma was isolated and stored at -80°C until analysis for NXT629 by liquid chromatography-mass spectrometry (LC/MS/MS). CD-1 mice received NXT629 PO (10 mg/kg in 0.5% methyl cellulose) or intraperitoneally (i.p) (10 mg/kg in saline). A pharmacokinetics assessment was performed on the less potent analog NXT969, which served as a negative control. CD-1 mice received NXT969 p.o. (10 mg/kg in 0.5% methyl cellulose) or i.p (30 mg/kg in 2% Tween 20/98% saline). Mice were euthanized at various time points after dosing and blood was collected into EDTA-containing tubes via cardiac puncture. Plasma was isolated and stored at -80°C analysis for NXT629 or NXT969 by LC/MS/MS. 2.3.3. Pharmacodynamics: liver gene expression CD-1 mice received an IP injection of either NXT629 or vehicle control (saline). Two hours later, mice were administered vehicle control (1:1 mixture of DMSO and sesame oil), WY14,643 (3 mg/kg) or GW590735 (1 mg/kg) by oral gavage at 10 ml/kg. These doses are known to induce PPARa target genes in the mouse liver (Hansen et al., 2010; Ye et al., 2001). Six h later, mice were euthanized via isoflurane inhalation, terminal cardiac punctures were performed and livers were dissected. Liver tissue was collected into TRIzol and kept on dry ice until transfer to -80°C for long term storage. Liver tissues were cut to approximately 1 cm3, suspended in TRIzol reagent (100 mg/ml) and subjected to homogenization by Polytron handheld homogenizer. Total RNA was purified according to the manufacturer’s TRIzol extraction protocols (Life Technologies, Carlsbad, CA). Reverse transcription of cDNA and real-time PCR was performed using specific primers (Table 3) as described above for normal and melanoma skin samples. 6

2.3.4. Pharmacodynamics: agonist-induced plasma FGF21 Mice (CD-1) received an IP injection of NXT629 or vehicle control (saline). Two h following antagonist treatment, mice were dosed with oral vehicle control (1:1 mixture of DMSO and sesame oil) or challenged with the agonist GW590735 (1 mg/kg) by oral gavage. Two h after agonist challenge, blood was collected into EDTA tubes via terminal cardiac puncture. Plasma was isolated and analyzed for FGF21 using an ELISA (R&D Systems, Minneapolis, MN) and NXT629 concentration using LC/MS/MS methods.

2.3.5. Pharmacodynamics: fasting-induced plasma FGF21 Mice (C57BL/6N) received an IP injection of vehicle control (saline) or NXT629 daily for 4 consecutive days. For the final 48 h, mice were fasted to induce an elevation of FGF21 in the plasma. Four hours after the final dose, blood was collected into EDTA tubes via cardiac puncture. Plasma was isolated and analyzed for FGF21 using an ELISA (R&D Systems, Minneapolis, MN) and NXT629 concentration using LC/MS/MS methods. 2.3.6. Syngeneic oncology models All studies were performed using C57BL/6N female mice (8-12 weeks old, Harlan, Livermore, CA). 2.3.6.1. Cells B16F10 or Lewis Lung Carcinoma cells were obtained from ATCC (Manassas, VA) and cultures were maintained in-house. Following trypsinization, cells were resuspended in Dulbecco’s phosphate buffered saline (PBS) (+/+) prior to administration in vivo. 7

Using western blot, B16F10 cell expression of PPARa was analyzed. Briefly, cell lysates were prepared in RIPA buffer containing protease inhibitor (Roche, Pleasanton, CA). Lysate was mixed with 6x Laemlli loading buffer and 10 mg of protein was loaded per lane. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transfer to polyvinylidene difluoride (PVDF) membrane were performed using standard methods. The blot was probed with a PPARa antibody (Abnova, Walnut, CA) using b-actin (Licor, Lincoln, NE) as a normalization control. Secondary antibodies were obtained from Licor (Lincoln, NE) and the blot was scanned using a Licor Odyssey NIR scanner. B16F10 and LLC cells were cultured in vitro or tumors were subcutaneously grown in vivo (see below) and RNA was isolated using an RNeasy Mini Kit (Qiagen, Venlo, The Netherlands). cDNA was synthesized, mixed with primers for the genes of interest (Table 3) and real-time PCR was performed as described above for human skin samples. 2.3.6.2. Experimental metastasis On day 0, C57BL/6N mice were injected intravenously with 50,000 B16F10 cells. Beginning on day 1, mice were injected IP once daily with either NXT629 (30 mg/kg) or vehicle control (saline). On day 21, mice were euthanized with isoflurane inhalation and terminal blood samples were collected via cardiac puncture into EDTA-containing tubes for subsequent pharmacokinetic analysis. Lungs were removed and placed into 5 ml Fekete’s solution. After at least 24 h in Fekete’s solution, pulmonary nodules were enumerated. In some experiments, nodules were scored according the size using the following formula: Total score= 1a + 2b + 3c where R= diameter of tumor nodule and a= number of nodules with R < 1mm, b= number of nodules where 1 mm ≤ R < 2 mm and c= number of nodules with R ≥ 2 mm.

8

2.3.6.3. Subcutaneous tumor growth On day 0, C57BL/6N mice were lightly anesthetized and injected subcutaneously in the hind flank region with 300,000 B16F10 cells. NXT629 (30 mg/kg) or vehicle control (saline) were administered daily via injection. For therapeutic treatment, mice were randomized into groups when tumor volumes where ≥ 75 mm3 using the formula: tumor volume= (length x width2)/2 at which point drug treatment was initiated. For preventative treatment, dosing was initiated on day 0 following implantation of cells. Typically, dosing continued out to day 21. On day 21, mice were euthanized by isoflurane inhalation and terminal blood was sampled via cardiac puncture for DMPK analysis. Final tumor volumes were determined. In some experiments, tumors were analyzed for test substance concentration or various biomarkers. Day 21 B16F10 subcutaneous tumors were collected, snap frozen on dry ice and stored at -80°C. ATP (Abcam, Cambridge, MA) and lactate (Eton Biosciences, San Diego, CA) and reactive oxygen species (Cell Biolabs, San Diego, CA) were evaluated using a was measured using commercially available kits. Tumor samples were processed according to each kit manufacturer’s instructions.

2.3.6.4. Lewis lung carcinoma (LLC) tumor resection On day 0, LLC cells (1 x 106) were injected subcutaneously into C57BL/6N mice and mice were treated with either vehicle or NXT629. Two weeks later (day 14), tumors were measured and volumes calculated using the formula described above. Mice were then anesthetized with isoflurane and tumors were aseptically surgically excised. Mice were treated with vehicle or

9

NXT629 for another 3 weeks (day 35) at which point terminal blood samples were collected and final tumor volumes were determined as described above. 2.3.7. FGF matrigel plug assay of angiogenesis On day 0, mouse fibroblast growth factor (mFGF) (R&D Systems, Minneapolis, MN) was prepared in matrigel to a final concentration of 500 ng/ml. C57Bl/6N mice were anesthetized with isoflurane, abdomens shaved and 400 ml of the FGF/matrigel injected subcutaneously. Control animals received plugs containing 0.1% BSA. NXT629 or vehicle was dosed daily from day 0 to day 7. On day 8, plugs were collected, photographed and analyzed for luminosity using Adobe Photoshop. 2.3.8. Human Subcutaneous SKOV-3 xenograft model SKOV-3 cells were obtained from ATCC (Manassas, VA) and maintained in culture according to their guidelines. Following trypinsinzation, cells were resuspended in PBS and mixed 1:1 with matrigel. 100 µl of the cell/matrigel mixture was injected subcutaneously into the left flank region of female Fox1 nude mice. Tumors were measured with calipers and tumor volumes calculated as described above. When tumor volumes were ≥ 75 mm3 (day 12 in this study), mice were randomized into groups and treatment with either vehicle (saline i.p.) or NXT629 was initiated. Tumor volumes were determined twice weekly. On day 43, blood was collected via the retro-orbital sinus. On day 44, animals were euthanized by isoflurane inhalation, terminal blood was collected via cardiac puncture and final tumor volume was determined. Tumors were excised, frozen at -80°C and analyzed for NXT629 concentration.

10

2.4. LC/MS/MS analysis of NXT629 and NXT969 in mouse plasma or tissue Known amounts of NXT629 or NXT969 were added to thawed mouse plasma or tissue homogenate to yield a concentration range from 0.8 to 4,000 ng/ml. In general, all standards, test plasma samples (neat) and test tissue samples (homogenized in PBS 1:4, v:v) were precipitated in a 96 well plate using acetonitrile (1:4, v:v) containing the internal standard (ISTD) Buspirone (5ng/ml). Sample plate was shaken for 5 minutes at 900rpm on a plate shaker then centrifuged at 4000 x g for 10 minutes using a swinging platform (Beckman Allegra X-14R). Supernatant was transferred to a new 96 well plate and analyte mixture (10 ml) was injected onto a LC-MS/MS system using a Leap PAL autosampler (Carrboro, NC). The LC/MS/MS system used was a Sciex API-3200 tandem mass spectrometer interfaced to a high-performance liquid chromatography system consisting of two LC10Avp pumps and a staticbed mixer (Shimadzu, Kyoto, Japan). Analysis was performed using an Agilent Zorbax SB-C8 column (2.1 x 50 mm; 5 mm) for chromatographic separation. NXT629, NXT969 and ISTD (Buspirone) were analyzed on a quadrupole mass spectrometer in the positive ion mode (ESI) by multiple reaction monitoring (MRM) using the following transitions (MH+®daughter): 610.1 ® 147.1, 716.8 ® 147.2 and 386.2 ® 122.4, respectively. The LC-MS/MS mobile phases contained 10 mM ammonium acetate in water with 0.05% formic acid (solvent A) and 10mM ammonium acetate in 50% acetonitrile/50% methanol with 0.05% formic acid (solvent B). The flow rate was maintained at 1 ml/min and the total run time was 2.5 min. Analytes were separated using a linear gradient as follows: mobile phase was held at 5%

11

solvent B for 0.5 min., solvent B was increased from 5% to 95% over the next 0.2 min, held constant for 1.3 min. and returned to the initial gradient conditions. The calibration curves were constructed by plotting the peak-area ratio of analyzed peaks against the corresponding nominal concentrations of NXT629 or NXT969 in the standard samples noted above. The data was subjected to linear regression analysis with 1/x2 weighting. The lower limits of quantitation were around 1-4 ng/ml.

2.5. Statistical Analysis Data are presented as mean ± S.E.M.. Statistics were performed using GraphPad Prism (La Jolla, CA). One-way ANOVA, two-way ANOVA or Student’s t-test was performed where appropriate. Statistical significance was P < 0.05.

3. Results 3.1. In vitro NXT629 inhibited agonist-induced activation of human PPARa with an IC50 of 78 nM (Table 1). The selectivity for PPARa over PPARb/d was > 75-fold while that for PPARg was > 200-fold. Furthermore, this molecule was shown to be selective across a panel of other nuclear hormone receptors (Table 1).

12

In the mouse luciferase-based reporter system, NXT629 is less potent on PPARa than in the human reporter system with a calculated IC50 of 2.3 µM. The selectivity of NXT629 in mouse is 4.3-fold over PPARb/d and 3.0-fold selective over PPARg (Table 1). NXT629 is a competitive PPARa antagonist, as determined by Schild analysis in the luciferase reporter assay (Fig. 2A). The IC50 of NXT629 shifts from 190 nM to 2 µM when the concentration of PPARa agonist GW7647 utilized in the assay is increased from 100 nM to 1 µM (Fig 2B). The calculated IC50 for the negative control molecule NXT969 in the human PPARa luciferase reporter assay is 9.6 µM. This molecule is greater than 10-fold selective for PPARa versus PPARb/d and PPARg as well as selective compared to other nuclear hormone receptors (Table 2). In the mouse, NXT969 is at least 20-fold less potent against PPARa than NXT629. 3.2. Pharmacokinetics NXT629 had poor (<10%) oral bioavailability in rats and mice (data not shown). In mice, NXT629 reached a Cmax of 375 nM 1 hour following an i.p. dose of 10 mg/kg (Fig. 3B). Plasma concentrations declined over time to a concentration of 50 nM at 24 hours after dosing. In order to maintain full coverage of mouse PPARa, we selected a dose of 30 mg/kg i.p. once daily for future pharmacological experiments. The less potent analog NXT969 had virtually no exposure after oral dosing and reached a maximum plasma concentration of only 10 nM (data not shown). This negative control molecule however exhibited very good exposure following i.p. dosing, reaching a maximum plasma concentration of 7.2 µM following an i.p. dose of 30 mg/kg (Fig. 3B). We selected a 13

dose of 30 mg/kg i.p. once daily for a study in which NXT969 would be compared side by side with NXT629 in vivo. 3.3. Pharmacodynamics: Gene expression in the liver At a dose level of 30 mg/kg, NXT629 prevented the induction of PPARa target genes induced in the mouse liver by WY-14,643 (Fig. 4). Cyp4a10 and Ehhadh were highly induced (>50-fold) 6 h following agonist treatment (Fig. 4). NXT629 reduced the induction of Cyp4a10 and Ehhadh, although only Ehhadh reached statistical significance. Aldh3a2 and Acox1 were moderately induced (6- to 10-fold) by WY-14,643 and both were significantly reduced by NXT629 treatment. Of the PPARa target genes measured, Cact and Cpt1a were the least induced (2- to 3-fold) at this 6 hour timepoint. However, NXT629 fully prevented induction of both. In a follow up study, an abbreviated NXT629 dose response curve of GW590735-induced PPARα gene induction was examined. The level of induction of Cact and Cpt1a following GW590735 challenge was similar to that observed with WY-14,643. The ED50 for inhibition of Cact and Cpt1a was approximately 10 mg/kg with a liver IC50 of 1.4 mM (Fig. 5). 3.4. Pharmacodynamics: Inhibition of plasma FGF21 At a dose level of 1 mg/kg, GW590735 elevated FGF21 in the plasma and this was reduced by NXT629 (Fig. 6A). At the time of study termination (4 h post-test compound administration), plasma concentrations of NXT629 were 400 and 30 nM for 30 and 3 mg/kg, respectively. In an expansion of the work that we previously reported (Bravo et al., 2014), NXT629 is a potent inhibitor of fasting-induced FGF21 (Fig. 6B). Plasma concentrations at the doses tested were 40 nM and 500 nM for 3 and 30 mg/kg, respectively, 4 h post administration of NXT629 and were

14

consistent with those in the agonist-induced FGF21 study. NXT629 appears to be a slightly more potent inhibitor of plasma FGF21 than it is of liver gene expression (as shown in Fig. 4)). Since we only tested 2 doses, we cannot calculate a true ED50; however greater than 50% inhibition of plasma FGF21 is achieved with a dose of 3 mg/kg in these models. 3.5. Relevance of PPARa to B16F10 cells and human melanoma Using western blot, we confirmed that B16F10 cells express mouse PPARa (Fig. 7A). Furthermore, PPARa mRNA is highly expressed in the in vivo B16F10 tumor (Fig. 7B). Finally, using LC-MS/MS, we determined that the concentration of oleoylethanolamide (OEA), an endogenous ligand of PPARa, is elevated (400 nM) in B16F10 subcutaneous tumors (sampled on day 21) but not plasma (8 nM) from the same animals. These data suggest that PPARa may be important for the development and growth of this syngeneic tumor cell line. Furthermore, gene expression analysis of human PPARa and some of its target genes in normal versus melanoma skin biopsy samples revealed a differential gene expression profile (Fig. 7C). In human skin, PPARa is more highly expressed than either PPARb/d or PPARg. Furthermore, expression of PPARa is significantly higher in melanoma skin than in non-diseased skin. Finally, we analyzed expression of several PPARα target genes and determined that CPT1a and ALDH3a2 were moderately upregulated while CACT, UCP2 and EHHADH were highly upregulated in melanoma versus non-diseased skin. No difference in FGF21 expression was observed between sample sets.

15

3.6. Experimental metastasis When a prophylactic dosing regimen was used, NXT629 inhibited experimental metastasis of B16F10 melanoma cells to the mouse lung at doses of 3 and 30 mg/kg (Fig. 8A). A dose of 0.3 mg/kg produced a trend toward inhibition which was not statistically significant. Thus, the ED50 in this model is between 3 and 30 mg/kg. At 30 mg/kg, NXT629 increased the concentration of TSP-1 in the plasma (259 ± 8 vs. 354 ± 32 ng/ml, * P <0.05 Student’s t test). Body weight was not significantly altered by 21 day treatment with NXT629 nor did the mice exhibit any adverse effects, demonstrating safety of chronic administration of a PPARa antagonist in this setting over a period of 21 days. When a therapeutic dosing regimen was employed (initiated on day 8), at a dose of 30 mg/kg NXT629 inhibited experimental metastasis slightly greater than 50% (Fig. 8B). A dose of 3 mg/kg did not result in significant inhibition of pulmonary metastasis. In a separate experiment, NXT629 was compared with a structurally similar analog (NXT969) which is 20-fold less potent against mouse PPARa. The degree of inhibition with NXT629 (30 mg/kg) was similar to that observed in the previous 2 experiments. Treatment with the less potent analog NXT969 (30 mg/kg) did not inhibit metastasis of B16F10 cells to the lungs (Fig. 8C). In this experiment, at a sampling time of 3 hours after dosing, plasma concentrations of NXT629 and NXT969 reached concentrations of 500 nM and 5.6 mM, respectively. The 24h trough plasma concentrations were 30 nM for NXT629 and 930 nM for NXT969.

16

3.7. Growth of subcutaneous B16F10 tumors When dosing was initiated at the time of subcutaneous B16F10 tumor cell implantation, NXT629 delayed the growth of subcutaneous tumors (Fig. 9). The concentration of NXT629 in the tumor at study termination was 531 nM. In contrast, when dosing was initiated in mice with established tumors (~80-100 mm3), NXT629 failed to inhibit subcutaneous tumor growth (data not shown) despite reducing ATP concentration in the tumors at study termination (16.8 mM vs. 29 mM, P<0.01).

3.8. Lewis Lung Carcinoma Lewis lung carcinoma tumors express PPARa (Fig. 10A). Similar to the effects observed with preventative treatment of B16F10 tumors, on day 14, LLC tumors were significantly smaller when NXT629 dosing was initiated at the time of tumor implantation versus the vehicle treated group (237 ± 55 mm3 vs. 425 ± 37 mm3, P < 0.05, t-test) (Fig. 10B). Following tumor resection on day 14, we observed regrowth at the resection site over 3 weeks that was significantly reduced by NXT629 treatment. NXT629 was effective at minimizing regrowth of the tumor when dosing was initiated either at the time of resection or 2 weeks prior to resection (Fig. 10C). In this experiment, lung metastases failed to develop in the vehicle treated group by study termination 3 weeks after primary tumor resection, therefore we could not determine effects of NXT629 on lung metastasis in this model.

17

3.9. Subcutaneous SKOV-3 ovarian cancer xenografts In two independent experiments, mice were randomized into groups when tumor volumes were 80-120 mm3 and dosed daily for 6 weeks with either vehicle (saline) or NXT629 (30 mg/kg) IP. For the first 24 days of dosing, NXT629 completely prevented the growth of SKOV-3 tumors (Fig. 11). However, after this point, tumors begin to grow at the same rate as the vehicle treated tumors, approaching vehicle tumor volumes by treatment day 36. NXT629 was nearly equally distributed between plasma and tumor in samples collected 2h after the final dose. Plasma concentration was 410 nM while tumor concentration of NXT629 was 620 nM. No adverse compound-related effects were observed in the mice after 6 weeks of daily dosing (body weight relative to start of treatment: +2.7% (vehicle), +4.5% (NXT629)), supporting earlier evidence that chronic dosing of a PPARa antagonist does not produce harmful side effects. 3.10. In vivo angiogenesis Angiogenesis occurred into matrigel plugs which had been loaded with mFGF (Fig. 12A). This was quantified using Adobe Photoshop (San Jose, CA) to determine luminosity of the plug, where a decrease in luminosity corresponds to an increase in angiogenesis. Daily treatment with NXT629 (30 mg/kg, IP) resulted in approximately 50% inhibition of the angiogenesis/luminosity (Fig. 12A and 12B). This suggests that NXT629 is weakly anti-angiogenic against FGF-induced angiogenesis.

18

4. Discussion In the current study, we characterize NXT629, a novel small molecule PPARa antagonist, and highlight its potential as a therapeutic for the treatment of cancer. Specifically, NXT629 is highly potent and selective antagonist for human PPARa over other members of the PPAR and nuclear hormone receptor families. This molecule is significantly less potent against mouse PPARa than it is against human PPARa, which is important to know as all of our efficacy studies are conducted in mouse. While NXT629 does not display suitable pharmacokinetic properties for further clinical development, it serves as useful molecule for in vitro and in vivo exploration of the effects of PPARa antagonism. Intraperitoneal injection of NXT629 produced sustained plasma exposure with detectable concentrations 24 h following a single injection. In order to determine whether we achieve target engagement (i.e. PPARa inhibition) with this route of administration we utilized two separate pharmacodynamics models. In the first model we activated the receptor pharmacologically using agonist stimulation. Administration of two structurally distinct synthetic PPARa agonists (WY14,643, GW590735) increased genes involved in fatty acid oxidation. These were inhibited back to baseline values by a single injection of NXT629, 30 mg/kg, with corresponding plasma concentrations of ~1.1 µM. Secondly, we demonstrated that this same dose of NXT629 could reduce plasma concentrations of fasting-induced elevations of plasma FGF21. In the second model we activated the receptor physiologically by fasting animals. These results are consistent with what has been observed what has been observed in PPARa knockout mice (Lundasen et al., 2007). From these results we chose a dose level of 30 mg/kg for all subsequent efficacy studies. 19

We utilized oncology models that we believe are relevant for exploring PPARa antagonism. As our primary in vivo efficacy screen, we used wildtype mice and the B16F10 cell line which expresses PPARa. Thus, in our experiments, NXT629 could be exerting effects on either the host or on the tumor cells. This is an important point since genetic knockouts of PPARa exhibit diminished tumor growth despite the use of tumor cells such as B16F10 which are themselves PPARa positive (Kaipainen et al., 2007). The degree of inhibition of tumor growth is much greater in PPARa knockout animals than we observed in wildtype mice using our antagonist approach. This discrepancy may be due to changes which occur as a result of ubiquitous absence of PPARa from development onwards in the genetic knockouts. Additionally, we performed our experiments in C57BL/6N mice while PPARa knockouts are on a 129S4/SvJae background which may also account for differences observed in maximal efficacy of NXT629 versus that reported in the literature with genetic ablation of PPARa. In an attempt to explore translation to the clinic, we obtained healthy and diseased skin biopsies. By quantitative PCR methodology, we demonstrated that PPARa and several of its target genes, namely CACT, UCP2, EHHADH and PDK1 were elevated in the melanoma samples. Using NXT629, we recapitulated many of in vivo phenotypes which are observed in PPARa knockout mice. Firstly, we demonstrated that a panel of agonist-induced changes in PPARa driven genes, namely those involved in fatty acid oxidation, in the mouse liver can be inhibited by pre-treatment with NXT629. Likewise, NXT629 inhibited agonist-induced elevations in plasma FGF21. In a parallel publication, we demonstrated that, fasting, acting as an endogenous trigger for PPARa, increases secretion of FGF21 in the plasma and that this can be prevented by pretreatment with NXT629 (Bravo et al., 2014) and we expand upon that data set in this current

20

report. PPARa knockouts display a slight increase in bleeding time (Ali et al., 2009). While beyond the scope of this paper, we have demonstrated that, consistent with the observations in knockout animals, mice treated with NXT629 have approximately a 3-fold increase in bleeding time. Additionally, PPARa knockout mice exhibit an enhanced decrease in fasting-induced changes blood glucose relative to wildtype mice (Kersten et al., 1999). A similar response in blood glucose was seen in mice which had been dosed with NXT629 (data not shown). All of these data strengthen our confidence that NXT629 is exerting its effects in vivo via antagonism of PPARa. We demonstrated that NXT629 inhibited the experimental metastasis of syngeneic B16F10 melanoma cells. This was accompanied by an increase of the anti-angiogenic component TSP-1 in the plasma. These results are consistent with the findings reported by Kaipainen et al. in which genetic knockouts of PPARa are protected in a mouse model of B16F10 experimental metastasis (Kaipainen et al., 2007). The role for PPARa in this model is further strengthened by our finding that NXT969, a structurally similar PPARa-inactive molecule, fails to inhibit experimental lung metastasis of B16F10 cells. When B16F10 cells were implanted subcutaneously, NXT629 inhibited tumor growth only when treatment was initiated at the time of tumor implantation. This suggests that PPARa may be more involved in the adhesion and early proliferation. Once the tumors reach the logarithmic phase of growth, PPARa antagonism alone may be insufficient to prevent tumor growth. Additionally we have demonstrated that NXT629 was effective in preventing growth of the ovarian cancer cell line SKOV-3 in a xenograft model. Recent evidence (Nieman et al., 2011) suggests that ovarian cancer cells preferentially use fatty acids as a fuel source. Fatty acid 21

oxidation has been shown to be an important fuel source for other cancer types, including breast cancer. Silencing RNA to CPT1C has previously been shown to inhibit tumor growth in a model of breast cancer (Zaugg et al., 2011). Since NXT629 demonstrated good inhibition of fatty acid oxidation genes, we therefore hypothesized that PPARa antagonism in turn would lead to tumor cell death by depriving them of one of their fuel sources, namely fatty acids. Indeed, our results suggest that PPARa antagonism may be an effective means of inhibiting ovarian cancer growth in the early stages. However, the inhibition of fatty acid oxidation alone may not be enough. In future experiments, we will explore efficacy of PPARa antagonism in combination with other cancer therapies which target tumor metabolism. Kaipainen et al. (2007) demonstrated that PPARa knockouts are protected against VEGFinduced vascular leak. In the studies described herein, we examine another important angiogenic pathway, namely fibroblast growth factor (FGF). Our results demonstrate that NXT629 displays moderate anti-angiogenic properties in this setting. Thus, one of the mechanisms by which PPARα antagonism inhibits tumor growth and metastasis may be via modulation of angiogenesis. In conclusion, NXT629 is a novel, potent and selective antagonist of PPARa that is efficacious in disease models where PPARa knockouts have been shown to be protected. This molecule proved a valuable means of demonstrating in vivo proof of concept of PPARa antagonism in a panel of pharmacodynamics and efficacy models. To our knowledge, this is the first description of in vivo activity with a potent and selective PPARa antagonist in a solid tumor setting and expands upon data in a chronic lymphocytic leukemia setting which was previously presented elsewhere (Messmer et al., 2015). These data provide a foundation for identifying a drug

22

candidate with the potential to be used as a therapeutic for diseases such as metastatic melanoma and ovarian cancer in which fatty acid oxidation and PPARa play a role. References Ali, F.Y., Armstrong, P.C.J., Dhanji, A.R.A., Tucker, A.T., Paul-Clark, M.J., Mitchell, J.A., Warner, T.D., 2009. Antiplatelet Actions of Statins and Fibrates Are Mediated by PPARs. Arteriosclerosis, Thrombosis, and Vascular Biology 29, 706-711. Berger, J., Moller, D.E., 2002. The mechanisms of action of PPARs. Annual review of medicine 53, 409435. Bravo, Y., Baccei, C.S., Broadhead, A., Bundey, R., Chen, A., Clark, R., Correa, L., Jacintho, J.D., Lorrain, D.S., Messmer, D., Stebbins, K., Prasit, P., Stock, N., 2014. Identification of the first potent, selective and bioavailable PPARalpha antagonist. Bioorg Med Chem Lett 24, 2267-2272. Collett, G.P., Betts, A.M., Johnson, M.I., Pulimood, A.B., Cook, S., Neal, D.E., Robson, C.N., 2000. Peroxisome proliferator-activated receptor alpha is an androgen-responsive gene in human prostate and is highly expressed in prostatic adenocarcinoma. Clin Cancer Res 6, 3241-3248. Devchand, P.R., Keller, H., Peters, J.M., Vazquez, M., Gonzalez, F.J., Wahli, W., 1996. The PPARalphaleukotriene B4 pathway to inflammation control. Nature 384, 39-43. Eastham, L.L., Mills, C.N., Niles, R.M., 2008. PPARα/γ Expression and Activity in Mouse and Human Melanocytes and Melanoma Cells. Pharmaceutical Research 25, 1327-1333. Etgen, G.J., Mantlo, N., 2003. PPAR ligands for metabolic disorders. Curr Top Med Chem 3, 1649-1661. Fu, J., Gaetani, S., Oveisi, F., Lo Verme, J., Serrano, A., Rodriguez De Fonseca, F., Rosengarth, A., Luecke, H., Di Giacomo, B., Tarzia, G., Piomelli, D., 2003. Oleylethanolamide regulates feeding and body weight through activation of the nuclear receptor PPAR-alpha. Nature 425, 90-93. Hansen, M.K., McVey, M.J., White, R.F., Legos, J.J., Brusq, J.M., Grillot, D.A., Issandou, M., Barone, F.C., 2010. Selective CETP Inhibition and PPAR Agonism Increase HDL Cholesterol and Reduce LDL Cholesterol in Human ApoB100/Human CETP Transgenic Mice. Journal of Cardiovascular Pharmacology and Therapeutics 15, 196-202. Kaipainen, A., Kieran, M.W., Huang, S., Butterfield, C., Bielenberg, D., Mostoslavsky, G., Mulligan, R., Folkman, J., Panigrahy, D., 2007. PPARalpha deficiency in inflammatory cells suppresses tumor growth. PLoS ONE 2, e260. Kehrer, J.P., Biswal, S.S., La, E., Thuillier, P., Datta, K., Fischer, S.M., Vanden Heuvel, J.P., 2001. Inhibition of peroxisome-proliferator-activated receptor (PPAR)alpha by MK886. Biochem J 356, 899-906. Kersten, S., Seydoux, J., Peters, J.M., Gonzalez, F.J., Desvergne, B., Wahli, W., 1999. Peroxisome proliferator-activated receptor alpha mediates the adaptive response to fasting. J Clin Invest 103, 14891498. Lundasen, T., Hunt, M.C., Nilsson, L.M., Sanyal, S., Angelin, B., Alexson, S.E., Rudling, M., 2007. PPARalpha is a key regulator of hepatic FGF21. Biochem Biophys Res Commun 360, 437-440. Messmer, D., Lorrain, K., Stebbins, K., Bravo, Y., Stock, N., Cabrera, G., Correa, L., Chen, A., Jacintho, J., Chiorazzi, N., Yan, X.J., Spaner, D., Prasit, P., Lorrain, D., 2015. A Selective Novel Peroxisome ProliferatorActivated Receptor (PPAR)-alpha Antagonist Induces Apoptosis and Inhibits Proliferation of CLL Cells In Vitro and In Vivo. Molecular medicine 21, 410-419. Nieman, K.M., Kenny, H.A., Penicka, C.V., Ladanyi, A., Buell-Gutbrod, R., Zillhardt, M.R., Romero, I.L., Carey, M.S., Mills, G.B., Hotamisligil, G.S., Yamada, S.D., Peter, M.E., Gwin, K., Lengyel, E., 2011. Adipocytes promote ovarian cancer metastasis and provide energy for rapid tumor growth. Nat Med 17, 1498-1503. 23

Pang, T., Kaufman, A., Choi, J., Gill, A., Drummond, M., Hugh, T., Samra, J., 2015. Peroxisome proliferator-activated receptor-alpha staining is associated with worse outcome in colorectal liver metastases. Mol Clin Oncol 3, 308-316. Spaner, D.E., Lee, E., Shi, Y., Wen, F., Li, Y., Tung, S., McCaw, L., Wong, K., Gary-Gouy, H., Dalloul, A., Ceddia, R., Gorzcynski, R., 2013. PPAR-alpha is a therapeutic target for chronic lymphocytic leukemia. Leukemia 27, 1090-1099. Willson, T.M., Brown, P.J., Sternbach, D.D., Henke, B.R., 2000. The PPARs: from orphan receptors to drug discovery. J Med Chem 43, 527-550. Xu, H.E., Stanley, T.B., Montana, V.G., Lambert, M.H., Shearer, B.G., Cobb, J.E., McKee, D.D., Galardi, C.M., Plunket, K.D., Nolte, R.T., Parks, D.J., Moore, J.T., Kliewer, S.A., Willson, T.M., Stimmel, J.B., 2002. Structural basis for antagonist-mediated recruitment of nuclear co-repressors by PPARalpha. Nature 415, 813-817. Ye, J.M., Doyle, P.J., Iglesias, M.A., Watson, D.G., Cooney, G.J., Kraegen, E.W., 2001. Peroxisome proliferator-activated receptor (PPAR)-alpha activation lowers muscle lipids and improves insulin sensitivity in high fat-fed rats: comparison with PPAR-gamma activation. Diabetes 50, 411-417. Zaugg, K., Yao, Y., Reilly, P.T., Kannan, K., Kiarash, R., Mason, J., Huang, P., Sawyer, S.K., Fuerth, B., Faubert, B., Kalliomaki, T., Elia, A., Luo, X., Nadeem, V., Bungard, D., Yalavarthi, S., Growney, J.D., Wakeham, A., Moolani, Y., Silvester, J., Ten, A.Y., Bakker, W., Tsuchihara, K., Berger, S.L., Hill, R.P., Jones, R.G., Tsao, M., Robinson, M.O., Thompson, C.B., Pan, G., Mak, T.W., 2011. Carnitine palmitoyltransferase 1C promotes cell survival and tumor growth under conditions of metabolic stress. Genes & Development 25, 1041-1051.

Figure Legends Figure 1. Chemical structure of NXT629 (A) and NXT969 (B). Figure 2. NXT629 is a competitive PPARa antagonist by Gaddum-Schild analysis (A). NXT629 potency is influenced by agonist concentration (B). Assays were performed using a commercially available reporter assay system (Indigo Biosciences, State College, PA). Immediately following addition of PPARa reporter cells to the assay plate, a fixed concentration of NXT629 (30, 100 or 300 nM) was added in the presence or absence of variable concentrations (0.1 nM to 1 µM) of the PPARa agonist GW7647(A). Alternatively, a variable concentration of NXT629 (10 nM to 10 µM) was added in the presence or absence of a fixed concentration of the PPARa agonist GW7647 (B). In each case, detection reagent was added twenty four hours later and luminescence was measured.

24

Figure 3. Time concentration plasma profiles for NXT629 in mouse (A). Pharmacokinetics of NXT969, a structurally related but PPARa inactive molecule, in mice (B). Mice were dosed intraperitoneally with NXT629 (10 mg/kg) or NXT969 (30 mg/kg) and blood was sampled at 1, 2, 4, 6 (NXT629 group only) and 24 h. Replicate plasma concentrations for each time point are graphed. n=2 individual mice per timepoint Figure 4. NXT629 inhibition of genes induced in the mouse liver by WY-14,643. C57Bl/6N mice (n=4 per group) received an intraperitoneal dose of NXT629 (30 mg/kg) which was followed 2 hours later by oral challenge with D/oil (1:1 DMSO:sesame oil) or WY-14,643 (3 mg/kg). Six h later, mice were euthanized, livers harvested into TRIzol, RNA isolated and PCR performed on cDNA for a panel of PPARa target genes. Cyp4a10: cytochrome P450, family 4, subfamily 1, polypeptide 10; Cact: carnitine-acylcarnitine translocase; Cpt1a: carnitine palmitoyltransferase 1a, liver; Ehhadh: enoyl-Coenzyme A, hydratase/e-hyddroxyacyl Coenzyme A dehydrogenase; Aldh3a2: aldehyde dehydrogenase family 3, subfamily A2; Acox1: acyl-Coenzyme A oxidase 1, palmitoyl. # P < 0.05, ## P < 0.01, ### P < 0.001, #### P < 0.0001 veh/WY vs. veh/no agonist. * P < 0.05, ** P < 0.01 NXT629/WY vs veh/WY. Two-tailed t test. Figure 5. Dose response curves for NXT629 inhibition of genes induced in the mouse liver by GW590735. C57Bl/6N mice (n=4 per group) received an intraperitoneal dose of NXT629 (3, 10 or 30 mg/kg) which was followed 2 h later by oral challenge with D/oil (1:1 DMSO:sesame oil) or GW590735 (1 mg/kg). Six h later, mice were euthanized, livers harvested into TRIzol, RNA isolated and PCR performed on cDNA for a panel of PPARa target genes. Cact: carnitineacylcarnitine translocase; Cpt1a: carnitine palmitoyltransferase 1a, liver. # P < 0.05, ## P < 0.01 veh/GW vs veh/no agonist, two-tailed t-test. * P <0.05 NXT629/GW vs veh/GW, one-way ANOVA with Dunnett’s multiple comparison post hoc. 25

Figure 6. Inhibition of GW590735-induced (A) or fasting-induced (B) plasma FGF21 by NXT629. (A) CD-1 mice (n=3-4 per group) received an intraperitoneal injection of NXT629 (3, 30 mg/kg) and two h later received an oral challenge with D/oil (1:1 DMSO:sesame oil) or GW590735 (1 mg/kg). Two h later, blood was collected and plasma concentration of FGF21 was determined by ELISA. # P < 0.05 veh/GW vs control, * P < 0.05 NXT629/GW vs veh/GW, one-way ANOVA with Dunnett’s multiple comparison post hoc. (B) C57Bl/6N mice (n=4 per group) received a total of 4 daily doses of vehicle or NXT629 via intraperitoneal injection. For the final 48 h, mice were fasted to induce FGF21. Four h after the last dose, blood was collected and plasma FGF21 concentration was determined by ELISA. ## P < 0.01 vs fed, * P < 0.05 vs vehicle/fasted, one-way ANOVA with Dunnett’s multiple comparison post hoc. Figure 7. PPARa protein (A) and message (B) are expressed in B16F10 cells. PPARa and its target genes are upregulated in melanoma versus normal human skin (C). B16F10 cells or subcutaneous tumor were analyzed for PPARa expression using standard western blotting (representative blot) or RT-PCR techniques (n=2 in vitro cultures, n=5 in vivo tumors). Normal (n=5) and melanoma (n=4) human skin biopsy samples were obtained from Conversant Bio and expression of PPARa and its target genes were determined by RT-PCR. Figure 8. NXT629 inhibits experimental metastasis of B16F10 cells to the mouse lung following preventative (A) and therapeutic (B) dosing regimens. The structurally related but PPARa inactive molecule NXT969 does not inhibit metastasis of B16F10 cells to the mouse lung. C57Bl/6 mice received an intravenous injection of B16F10 cells on day 0. Intraperitoneal dosing of NXT629 (30 mg/kg), NXT969 (30 mg/kg) or saline vehicle was initiated on day 1 or 26

day 8 (NXT629 only). On day 21, animals were euthanized, lungs dissected into Fekete’s solution and tumor nodules enumerated. n=8 per group for all data sets. * P < 0.05, ** P < 0.01, *** P < 0.001 NXT629 vs vehicle, one-way ANOVA with Dunnett’s multiple comparison post hoc. Figure 9. NXT629 inhibits growth of subcutaneous B16F10 tumors when administered preventatively. On day 0, B16F10 cells were subcutaneously implanted into C57Bl/6 mice. NXT629 (30 mg/kg) or saline vehicle was administered by intraperitoneal injection beginning on day 0 (n=5 mice per group). After tumors became palpable, tumor size was measured with calipers 1-2 times per week until study termination on day 22. ** P < 0.01 day17 NXT629 vs day 17 vehicle, **** P < 0.0001 day 22 NXT629 vs day 22 vehicle, two-way ANOVA with Bonferroni’s post hoc. Figure 10. Lewis lung carcinoma cells express PPARa protein (A) (n=2 in vitro cultures, n=6 in vivo tumors). Preventative treatment with NXT629 decreases growth of subcutaneous LLC tumors (n=6, * P < 0.05, t-test) (B) and prevents tumor regrowth following resection of the primary tumor (n=3 to 6, P < 0.05 vs. vehicle, one-way ANOVA with Dunnett’s multiple comparison test post hoc) (C). Therapeutic administration of NXT629 initiated at the time of tumor resection prevents subsequent tumor regrowth (C). RNA was isolated, cDNA synthesized and RT-PCR performed on in vitro LLC cells and subcutaneously grown in vivo tumors. For assessment of effects on in vivo tumor growth C57Bl/6 mice were subcutaneously implanted with LLC cells on day 0. Mice were treated with either vehicle or NXT629 (preventative group) for 14 days, at which time subcutaneous tumors were measured with calipers and resected. At this time, the vehicle group was divided into two groups, one of which continued to receive

27

vehicle, the other which received NXT629 (therapeutic group). Dosing continued for another 3 weeks. Figure 11. NXT629 delays growth of subcutaneous SKOV-3 tumors in nude mice. n=6-8 per group per time point. Graph is composed of data from 2 independent studies with tumor measurements performed on different treatment days in each study. SKOV-3 tumor cells were mixed 1:1 with matrigel and subcutaneously implanted on day 0. When tumor volumes were ≥75 mm3, intraperiotoneal administration of NXT629 or vehicle was initiated. Tumors were measured with calipers two times per week and tumor volumes calculated. **** P < 0.0001 overall treatment effect vs vehicle, two-way ANOVA with Bonferroni multiple comparison post hoc. Figure 12. NXT629 partially inhibits FGF-induced angiogenesis in a matrigel plug assay. A, representative images of matrigel plugs at study termination. B, Quantified angiogenesis using the Adobe PhotoShop luminosity feature, where luminosity is inversely proportional to the degree of angiogenesis. On day 0, mouse FGF was mixed with matrigel and 200 ng of FGF was implanted subcutaneously into C57Bl/6 mice (n=6-9 per group). Mice received an intraperitoneal dose of either saline vehicle or NXT629 (30 mg/kg) from day 0 to day 7. On day 0, control animals (n=3) received matrigel plugs containing 0.1% BSA in matrigel. On day 8, matrigel plugs were removed and photographed. Photographs were analyzed for luminosity which inversely correlates with the degree of angiogenesis. ## P < 0.01 vs veh/BSA, * P < 0.05 vs veh/FGF, one-way ANOVA with Dunnett’s multiple comparison test post hoc.

28

Table 1. Antagonist potency and selectivity profile of NXT629 Species

Human

Mouse

Assay

IC50 (nM)

PPARa

78

PPARb/d

6,007

PPARg

15,483

Estrogen Receptor b

12,403

Glucocorticoid Receptor

26,403

Thyroid Receptor b

45,100

PPARa

2,327

PPARb/d

35,061

PPARg

6,913

Table 2. Antagonist potency and selectivity profile of NXT969 Species

Human

Mouse

Assay

IC50 (nM)

PPARa

9,647

PPARb/d

>100 mM

PPARg

>100 mM

Estrogen Receptor b

34,810

Glucocorticoid Receptor

>100 mM

Thyroid Receptor b

72,973

PPARa

>100 mM

PPARb/d

>100 mM

PPARg

>100 mM

Table 3. Forward and reverse primers used for real-time PCR reactions Gene

Forward

Reverse 29

Ppara (m)

AGAGCCCCATCTGTCCTCTC

ACTGGTAGTCTGCAAAACCAAA

Cyp4a10 (m)

TTCCCTGATGGACGCTCTTTA

GCAAACCTGGAAGGGTCAAAC

Cact (m)

GACGAGCCGAAACCCATCAG

AGTCGGACCTTGACCGTGT

Cpt1a (m)

CTCCGCCTGAGCCATGAAG

CACCAGTGATGATGCCATTCT

Ehhadh (m)

ATGGCTGAGTATCTGAGGCTG

GGTCCAAACTAGCTTTCTGGAG

Aldh3a2 (m)

CCTGAGCAAAAGTGAACTCAATG

TCTTAGCCGGTCTCGCAGAA

Acox1 (m)

CCGCCACCTTCAATCCAGAG

CAAGTTCTCGATTTCTCGACGG

Cyclophilin (m)

ATTTCTTTTGACTTGCGGGC

AGCTAGACTTGAAGGGGAATG

B-actin (m)

GGCTGTATTCCCCTCCATCG

CCAGTTGGTAACAATGCCATGT

PPARa (h)

GGCAAGACAAGCTCAGAAC

TTATCTATGAAGCAGGAAGCAC

PPARb/d (h)

CAGGGCTGACTGCAAACGA

GCCACCTGTGGGTTGTACTG

PPARg (h)

GGGATCAGCTCCGTGGATCT

TGCACTTTGGTACTCTTGAAGTT

Cyclophilin (h)

CTCGAATAAGTTTGACTTGTGTTT

CTAGGCATGGGAGGGAACA

CPT1a (h)

TCCAGTTGGCTTATCGTGGTG-

CTAACGAGGGGTCGATCTTGG

ALD3H3a2 (h)

CTGAAGCAGCGATTTGACCAC

CCCTCCCAGTTCAAGAGTCAC

CACT (h)

GACCAGCCAAAACCCATCAG

AGAGGGTGACCGACGAACA

UCP2 (h)

CCCCGAAGCCTCTACAATGG

CTGAGCTTGGAATCGGACCTT

EHHADH (h)

CGGAGCATCGTGGAAAACAGCA

CCGAGTCTACAGCAATCACAGG

(m): mouse, (h): human

30

Figure 1

Figure 1

B

A

0 -11

20000

40000

60000

80000

100000

Figure 2

PPARa activity (RLU)

A

Figure 2

-10

-8

-7

[GW7647] (LogM)

-9

Kb=24.4nM Schild Slope=1.22±0.1

-6

Gaddum/Schild Analysis

-5

no antagonist 30nM 100nM 300nM

B

% Inhibition 0

50

100

IC50 1.9e-007 2.0e-006

-9

-7

-6

[NXT629] (LogM)

-8

shift: 10.2 fold

GW7647 amount 100nM 1µM

-5

-4

Agonist concentration effects antagonist potency

A

Plasma [NXT629] (mM)

Figure 3

Figure 3

0.0

0.1

0.2

0.3

0.4

0.5

0

4

8

16

time (h)

12

NXT629 (10 mg/kg)

20

24

28

B Plasma [NXT969] (mM) 0

2

4

6

8

10

0

4

8

16

time (h)

12

NXT969 (30 mg/kg)

20

24

0

20

40

60

80

0

50

100

150

200

Figure 4

Figure 4 revised

D/oil

D/oil

NXT629

NXT629

**

WY-14,643 (3 mg/kg)

Saline

####

Ehhadh

WY-14,643 (3 mg/kg)

Saline

#

Cyp4a10 relative mRNA abundance (fold of control) relative mRNA abundance (fold of control)

relative mRNA abundance (fold of control)

relative mRNA abundance (fold of control)

0

5

10

15

0

1

2

3

4

D/oil

D/oil

NXT629

*

NXT629

**

WY-14,643 (3 mg/kg)

Saline

###

Aldh3a2

WY-14,643 (3 mg/kg)

Saline

#

Cact relative mRNA abundance (fold of control) relative mRNA abundance (fold of control)

0

2

4

6

8

0.0

0.5

1.0

1.5

2.0

2.5

D/oil

D/oil

NXT629

*

NXT629

*

WY-14,643 (3 mg/kg)

Saline

##

Acox1

WY-14,643 (3 mg/kg)

Saline

#

Cpt1a

A

Figure 5

relative mRNA abundance (fold of control)

Figure 5 revised

0

1

2

3

4

0 D/oil Saline (IP)

3

10

30

*

GW590735 (1 mg/kg PO)

30

#

GW590735 (1 mg/kg IP)

10

1

2

3

Cpt1a

NXT629 (mg/kg IP)

3

*

B

NXT629 (mg/kg IP)

D/oil Saline (IP)

##

Cact relative mRNA abundance (fold of control)

plasma FGF21 (pg/ml)

A

Saline (IP)

0 30

*

500 nM

NXT629 (mg/kg IP)

GW590735 (1 mg/kg PO)

3

*

2000

1000

30 nM

D/oil

#

3000

4000

5000

6000

Figure 6

Figure 6 revised

0

1000

2000

3000

4000

B plasma FGF21 (pg/ml)

NXT629 (mg/kg, IP)

Saline (10 m/kg, IP)

30

*

500 nM

3

*

40 nM

Fed Fasted

##

B

A

relative mRNA abundance normalized to cyclophilin

Figure 7

Figure 7

0

1

2

3

4

5

in vitro

PPARa

in vivo

b-actin

PPARa

C Fold expression over cyclophilin 0

20

40

60

80

0

1

2

3

0

10

20

30

40

Fold expression over cyclophilin Fold expression over cyclophilin

CACT

CPT1

PPARa

PPARg

UCP2

EHHADH

ALDH3a2

PPARd

normal human skin human melanoma skin biopsy

nodules/lung

A

0

10

20

30

h ve

Figure 8

3

30

***

NXT629 (mg/kg, IP)

3 0.

*

Preventative

B

nodules/lung

Figure 8

0

5

10

15

20

h ve

30

NXT629 (mg/kg, IP) (dosed d8-d18)

3

***

Therapeutic

Nodule score 0

10

20

30

40

50

C

Veh

NXT629 NXT969

**

NS

Active vs. Negative

0

2000

4000

6000

8000

10000

Figure 9

Figure 9 revised

Tumor volume (mm3)

0

5

15

day

10

start dosing NXT629 (30 mg/kg IP)

** 25

****

20

Preventative vehicle NXT629

relative mRNA abundance normalized to cyclophilin

0

2

4

6

8

10

A

in vivo

PPARa

in vitro

Figure 10

Figure 10

0

100

200

300

400

500

B

Veh

NXT629

*

Lewis Lung Carcinoma day 14, prior to resection Tumor volume (mm3)

Tumor volume (mm3)

0

2000

4000

6000

C

Veh

*

NXT629 (30 mg/kg, qd)

Preventative Therapeutic

*

Lewis Lung Carcinoma day 35

Figure 11

Figure 11 style rev

relative tumor volume vs. start of treatment 0

1

2

3

4

5

6

7

0

14 21 28 35 42

Treatment day

7

SKOV-3

veh NXT629

**** P < 0.0001

Figure 12

Figure 12

B

A

NXT629

Vehicle

0

50

100

150

200

FGF

FGF

0.1% BSA

luminosity

200 ng FGF

BSA

NXT629

veh

veh

##

*