From Celebrex to a Novel Class of Phosphoinositide-Dependent Kinase 1 (Pdk-1) Inhibitors for Androgen-Independent Prostate Caner

Home , Trifluoromethyl group, Vatalanib

FROM CELEBREX TO A NOVEL CLASS OF PHOSPHOINOSITIDE-DEPENDENT KINASE 1 (PDK-1) INHIBITORS FOR ANDROGEN-INDEPENDENT PROSTATE CANER

DISSERTATION

Presented in Partial Fulfillment of the requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Jiuxiang Zhu, M.S.

* * * * *

The Ohio State University 2005

Dissertation Committee: Approved by Professor Ching-Shih Chen, Advisor

Professor Robert Brueggemeier Advisor

Professor Pui-Kai (Tom) Li Graduate Program in Pharmacy

Professor Matthew D. Ringel

ABSTRACT

Celebrex, a nonsteroidal anti-inflammatory drug (NSAID, cyclooxygenase-2

inhibitor), was reported to induce apoptosis in the prostate cancer cell line PC-3 at 50µM.

Early research from our laboratory demonstrated that this apoptotic inducing effect was independent of its COX-2 inhibitory activity. Further investigation showed that PDK-1 was a major protein targeted by celebrex in PC-3 cells to induce apoptosis.

However, celebrex was very weak in inhibiting PDK-1 with IC50 of 48µM. To improve its potency, two series of analogs were designed and synthesized. In the first series of 24 compounds, the 5-position methylphenyl moiety of celebrex was replaced by various aromatic ring systems to explore the optimal hydrophobic group. OSU02067

(IC50=9µM) with phenanthrene at the 5-position was the best inhibitor among this series

and was selected as the lead compound for further modification.

Enzyme kinetics study of PDK-1 inhibition by celebrex indicated that it competed

with ATP for binding. Docking of OSU02067 into the ATP binding domain of PDK-1

showed that the sulfonamide moiety hydrogen bonded to hinge region residue Ala162.

Considering the importance of H-bonding, the sulfonamide moiety was substituted with

various heteroatom-rich functional groups in the 2nd series of 12 compounds.

OSU03012 and OSU03013 stood out as the most potent analogs with IC50s

ii of 5µM and 2µM, respectively. Such improvement was partly attributed to an

additional H-bond formed with hinge region residue Ser160. Exposure of PC-3 cells to

these two agents (5µM or higher) led to significant decreases in Akt and p70s6k kinase activity (both are downstream substrates of PDK-1) and an increase in apoptosis evidenced by nucleosome formation and PARP cleavage.

As OSU03012 was tolerated well by nude mice at a dose of 200mg/kg, it was chosen for further study. In primary chronic lymphocytic leukemia cells and the breast cancer cell line, MDAMB453, the compound induced apoptosis. In addition, Tseng et. al. showed that OSU03012 was equally potent in imatinib-resistant and -sensitive CML cells with IC50 of 6µM. OSU03012 was able to sensitize mutant resistant cells

(Ba/F3p210E255K and Ba/F3p210T315I) to imatinib partly because of the concerted effect on phospho-Akt. Currently, OSU03012 is being studied in lung, thyroid, ovarian and bladder cancers and preclinical studies are underway (including toxicology, pharmacokinetics and pharmacodynamics).

Future structure modification of OSU03012 is expected to be pursued in the following three aspects. Hydrophobic pockets behind the adenine binding domain and at the C-terminal lobe below the ribose binding pocket will be explored. Crystal structures of PDK-1 with this class of inhibitors are needed to provide insights into structure-based design. Adsorption, distribution, metabolism and execretion will be integrated into inhibitor design to achieve clinically applicable drug candidates.

iii

Dedicated to my grandparents, parents and brother

ACKNOWLEDGMENTS

• I would like to acknowledge Dr. Chen for his guidance, constant encouragement,

support and for providing excellent working environment for his students

• I would also like to express my sincere appreciation for the faculty of the division,

especially Drs. Brueggemeier and Li, for all their help and advice

• I want to thank Kathy Brooks and Kelli Ballouz for being able to solve any

problem and make graduate studies going smoothly

• Colleagues and friends: Drs. Kulp, Song, Johnson, Lin and Wang, Ho-Pi, Yvette,

Ping-Hui, Ya-Ting, Joe, Chung-Wai, Yeng-Jeng, Kuen-Feng, Leo, Qiang, Jim,

Chang-Shi, Nicole and Jack for help in all kinds of experimental details,

invaluable discussion about a variety of scientific topics and proofreading thesis

draft.

• Xiaohui and Yan for having lots of funs together and making graduate studies

memorable.

VITA

1993-1997 B.S. Pharmaceutical Science Beijing University, China

1997-2000 M.S. Natural Product Chemistry Peking Union Medical College, China

2000-Present Graduate Teaching and Research Associate College of Pharmacy, Ohio State University

PUBLICATIONS

1. Jiuxiang Zhu, Jui-Wen Huang, Joseph Fowble, Ping-Hui Tseng, Chung-Wai Shiau, Yeng-Jeng Shaw, Samuel K. Kulp, and Ching-Shih Chen From the Cyclooxygenase-2 Inhibitor Celecoxib to a Novel Class of Potent Akt-Targeted Antitumor Agents. Cancer Research, 2004;64(12):4309-18

2. Jiuxiang Zhu, Xueqing Song, Ho-pi Lin, Donn C. Young, Shunqi Yan, Victor E. Marquez, Ching-shih Chen Using Cyclooxygenase-2 Inhibitors as Molecular Platforms to Develop a New Class of Apoptosi s-inducing Agents. Journal of the National Cancer Institute 2002, 94(23), 1745-17

3. Amy J. Johnson, Lisa L. Smith, Jiuxiang Zhu, Nyla A. Heerema, Sara Jefferson, Michael Grever, Ching-shih Chen, John C. Byrd. A Novel Celecoxib Derivative Induces Apoptosis in Primary CLL Cells and Transformed B-cell Lymphoma via a Caspase and Bcl-2 Independent Mechanism. Blood, 2004, Sep 28.

4. Haiming Ding, Chunhua Han, Jiuxiang Zhu, Ching-Shih Chen, Steven M.D’Ambrosio. Celecoxib Derivatives Induce Apoptosis via Disruption of Mitochondrial Membrane Potential and Activation of Caspase 9. Int. J Cancer., 2004 Oct 21.

FIIELDS OF STUDY

Major Field: Pharmacy

TABLE OF CONTENTS Page Abstract……………………………………………………………………………...... ii

Dedication……………………………………………………………………………...iv

Acknowledgments..…………………………………………………………………...... v

Vita……………………………………………………………………………………..vi

List of Tables…………………………………………………………………………...ix

List of Figures………………………………………………………………………...... x

Abbreviations…………………………………………………………………………xiv

Chapter 1 Introduction………………………………………………………………….1 1.1 Hormone-Refractory Prostate Cancer (HRPC) and Therapies……………..1 1.2 PI3K/PDK-1/Akt Signaling Pathway ……………………………………...5 1.3 Protein Kinases (PKs) and Their Inhibitors………………………………...9 1.3.1 Kinases and Cancer…………………………………………………...9 1.3.2 Protein Kinase Inhibitors (PKIs)…………………………………….12 1.4 Cyclooxygenase-2 (COX-2) Inhibitors as Anti-cancer Agents…………....15

Chapter 2 Molecular Target(s) for Celebrex and Project Design……………………...27 2.1 Molecular Target of Celebrex…………………………………………….27 2.1.1 COX-2 Independent Mechanism…………………………………….27 2.1.2 PDK-1, a Major COX-2 Independent Target for Celebrex………….29 2.2 Aims and Project Design………………………………………………….30

Chapter 3 Design, Synthesis and Biological Activities of the1st Series of Celebrex Derivatives…………………………………………..……………………...39 3.1 Design of the 1st Series of Derivatives……………………………..……...39 3.2 Synthesis of the 1st Series of Derivatives…...……………………………..40 3.3 Structures of the Analogs and Their Biological Activities…….…..……...41 3.4 OSU02067………………………………………………………………....42 3.4.1 Effect on Downstream Protein Akt…………………………………...42 3.4.2 Cellular Effect of OSU02067…………………………………………42 3.4.3 In vivo Study of OSU02067………………………………………...... 43

vii

Chapter 4 Structure-Based Design of PDK-1 Inhibitors………………………………49 4.1 Crystal Structure of PDK-1 Catalyic Domain in Complex with ATP…….49 4.2 Docking of OSU02067 to ATP Binding Domain and Design of the 2nd Series Analogs….……………………………………………..……………….50 4.3 Synthesis of the 2nd Series of Compounds……………………………..….52 4.4 Optimal Compounds-OSU03012 and OSU03013………………………...54 4.5 In vitro Effect of OSU03012 and OSU03013 in PC-3……………………55 4.5.1 Effects on Downstream Proteins Akt and p70S6K…………………..55 4.5.2 Cellular Effects of OSU03012 and OSU03013…………………….55

Chapter 5 Application of OSU03012 to Other Cancers……………………………….71 5.1 In Primary Chronic Lymphocytic Leukemia (CLL)………………………71 5.2 In Breast Cancer cells MDAMB453…..………………………..…………71 5.3 In Gleevec Resitant Chronic Myelogenous Leukemia (CML) Cells……...72

Chapter 6 Conclusions and Future Directions………………………………………...73 6.1 Conclusions……………………………………………………………...73 6.2 Future Directions………………………………………………………..73 6.3 Concerns about Anticancer Drug Development………………………...77 6.3.1 Specific PKIs versus Non-Specific (or Broad Spectrum) PKIs…...78 6.3.2 Targeted Cancer Therapy-Lessons from Gefitinib Clinical Trial…79

Chapter 7 Experimental Methods and Material……………………………………….83 7.1 Synthesis of the 1st Series of Compounds……………………………..…..83 7.1.1 Preparation of Starting Material for Synthesis of Compound 1-4….83 7.1.2 Preparation of Starting Material for Synthesis of Compound 10-20.84 7.1.3 General Procedure…………………………………………………..85 7.2 Synthesis of the 2nd Series of Compounds………………………………...87 7.3 Nomenclatures, 1H NMR (proton nuclear magnetic resonance), and HRMS (high resolution mass spectrometry) Characterizations of Compounds 1 – 36………………………..………………..………………….93 7.4 PDK-1 Kinase Assay……………………………………………………...93 7.5 Cell Viability Assay……………………………………………………….94 7.5.1 Cell culture………...………………………………………………..94 7.5.2 MTT Assay………………………………………………………….94 7.6 Immunoblotting……………………………………………………………95 7.7 Immunoprecipitated Akt Kinase Assay…………………………………....96 7.8 Immunoprecipitated p70S6K Kinase Assay………………………………...96 7.9 Cell Death Detection ELISA………………………………………………97 7.10 Molecular Modeling……………………………………………………..98

Bibliography………………………………………………………………………….108

viii LIST OF TABLES

Table 3.1 Structures and IC50s for inhibiting recombinant PDK-1 kinase activity and for inducing apoptotic death in PC-3 cells for celebrex and 24 derivatives………………………………………………………………46

Table 4.1 Structures and potency of derivatives 25–36 for inhibiting recombinant PDK-1 kinase activity and PC-3 cell growth…………………………...65

Table 7.1 Nomenclatures, NMR and HRMS data for compounds 1-36…………101

LIST OF FIGURES

Fig. 1.1 Stem cell model for the organization of prostate epithelium……..……….19

Fig. 1.2 Possible pathways for HRPC development……………………………...... 20

Fig. 1.3 PI3K/PDK-1/Akt signaling pathway…………………………………...... 21

Fig. 1.4 Crystal structure of EGFR with erlotinib (1M17). N-lobe is largely β-sheet and c-lobe consists of many α helices. These two lobes are connected by hinge region amino acids. Erlotinib binds to ATP binding domain……………………………………………………………………..22

Fig. 1.5 ATP binding site of protein kinases. ATP is in red. Dotted line represents H-bonds……………………………………………………………………23

Fig. 1.6 Top. Schematic representation of active and inactive conformation of IRK with emphasis on activation loop and αC helix (33). Bottom. Crystal structure of Abl in complex with imatinib (left) and PD173955 (right) (34). Activation loop is labeled in blue on left (inactive conformation) and in red on right (active conformation)……………………………………………..24

Fig. 1.7 Chemical structures of selected kinase inhibitors…………………………25

Fig. 1.8 Top. The role of cyclooxygenases. Bottom. Structure of celebrex ….....26

Fig. 2.1 A．Western blot analysis showing COX-2 protein levels in parental PC-3 cells and four independent antisense COX-2 clones in the presence (+) or absence (-) of doxycycline (Dox, 2mg/mL) for 10 days. B. Susceptibility of prostate cancer cells to celebrex-induced apoptosis is independent of COX-2 expression levels. Left panel. Effect of 50µM celebrex on the viability of parental PC-3 cells and the COX-2 deficient clone 2F6. Right panel. Effect of 50µM celebrex on the viability of the COX-2 antisense clone 7D9 with (+) or without (-) a DOX pretreatment (2mg/mL)…...……………………………………………………………..34

Fig. 2.2 Structures and characteristics of celebrex and compounds 1–7. The general structure of these molecules is shown at the top. IC50-concentration inhibiting 50% COX-2 activity; T1/2-time required for 50% cell death……………………………………………………….35

Fig. 2.3 A. Dose-dependent inhibition of recombinant PDK-1 kinase activity by celebrex. B. Western blot analysis of Akt and phosphor-Akt in PC-3 cells after treated with celebrex or DMSO for indicated time (63). C. PC-3 cells were susceptible to the treatment of celebrex in a dose and time-dependent manner …...…………………………………………….36

Fig. 2.4 Structures and IC50s for reported PDK-1 inhibitors………………...... 37

Fig. 2.5 Working model outlining the structural features essential for the apoptotic inducing effect of celebrex………………………………...... 38

Fig. 3.1 Synthetic route for aromatic ketone and the 1st series of derivatives.....44

Fig. 3.2 Two isomers formed as diketone is coupled with 4-methylsulfoylphenylhydrazine………………………………………...45

Fig. 3.3 Top: Effect of OSU-02067 versus celebrex on the kinase activity of Akt immunoprecipitated from drug-treated PC-3 cells. Bottom: Phosphorylation status of Akt in PC-3 cells treated with OSU-02067 at different concentrations. Control PC-3 cells received DMSO vehicle…………………………………………………………………...47

Fig. 3.4 A. Time- and dose-dependent effect of OSU-02067 on the viability of PC-3 cells. B. Top: Formation of cytoplasmic nucleosomal DNA in PC-3 cells treated with DMSO vehicle or the indicated concentrations of OSU-02067. DNA fragmentation was quantitatively measured by cell death detection ELISA kit. O.D. absorbance. Bottom: Induction of poly(ADP-ribose) polymerase cleavage (85-kDa fragment) by OSU-02067 in PC-3 cells. C. Antiproliferative effect of 1 and 5µM OSU- 02067 versus 30µM celebrex in PC-3 cells…………………...... 48

Fig. 4.1 Top. Crystal structure of PDK-1 catalytic domain in complex with ATP. Bottom. Enlarged view of van der waals interaction between PDK-1 and adenine moiety in ATP (in orange)……………...... 57

Fig. 4.2 H-bonds formed between ATP and PDK-1……………………………..58

Fig. 4.3 Lineweaver-Burke plots of the competition of celebrex with ATP in PDK-1 kinase activity. Activity of the recombinant PDK-1 toward the peptide substrate was determined using 1–100µM ATP in the presence of 0, 25, and 50µM celebrex………………………………………………..59

Fig. 4.4 A: Global view of OSU02067 docked into ATP binding site. B: H-bond formed between sulfonamide moiety of OSU02067 and Ala162. C: Residues within 6.5Å of OSU02067…………………………………….60

Fig. 4.5 Synthetic route for derivatives 26-30……………………………………61

Fig. 4.6 Synthetic route for derivatives 25 and 31-33……………………………62

Fig. 4.7 Synthetic route for derivatives 34-36……………………………………63 . Fig. 4.8 Two isomers formed as intermediate VII is coupled with 4-nitrophenylhydrazine………………………………………………….64

Fig. 4.9 A: Global view of OSU03012 docked into ATP binding site. B: H-bond formed between NH2 of OSU03012 and CO of Ser160 and NH of Ala162. C: Residues within 6.5Å of OSU03012…………………………………66

Fig. 4.10 A: Global view of OSU03013 docked into ATP binding site. B: H-bond formed between guanidino moiety of OSU03013 and CO of Ser160 and NH of Ala162. C: Residues within 6.5Å of OSU03013………………67

Fig. 4.11 Overlay of OSU02067 (purple), OSU03012 (orange) and OSU03013 (green). Conformations shown here for each compound are extracted from docked structures……………………………………………...... 68

Fig. 4.12 A. Dose-dependent inhibition of recombinant PDK-1 kinase activity by OSU-03012 and OSU-03013. B. Effect of OSU-03012 (top) and OSU-03013 (bottom) on the phosphorylation status of Akt in drug-treated PC-3 cells. C. Effect of OSU-03012 on the kinase activity of immunoprecipitated p70S6K in drug-treated PC-3 cells……………...... 69

Fig. 4.13 A. Time and dose-dependent effects of OSU03012 and OSU03013 on cell viability of PC-3 cells. B. Top: Formation of cytoplasmic nucleosomal DNA in PC-3 cells treated with DMSO vehicle or the indicated concentrations of OSU-03012 (left) or 03013 (right). Bottom: Induction of PARP cleavage by OSU-03012 (left) or 03013 (right) in PC-3 cells as measured by immunoblotting. C. Antiproliferative effect of OSU-03012 (left) and OSU- 03013 (right) in PC-3 cells……………………………..70

xii

Fig. 6.1 (a). Crystal structure of p38 in complex with SB4. Thr106 and fluorophenyl group in inhibitor were highlighted in red rectangle (91). (b). Crystal structure of EGFR in complex with Tarceva. Thr766 and acetyl group of Tarceva were highlighted in red rectangle (90). (c) Crystal structure of CDK2 in complex with NU6102. Two H-bonds formed between Asp86 and Sulfonamide group of NU6102, which are highlighted in red rectangle (39)……………………………………………………..81

Fig. 6.2 Crystal structure of p38α in complex with inhibitors. Yellow represents original conformation in complex with compound. Green represents compound 2 induced conformation with flip at Met109-Gly110 (89)………………………………………………………………………82

Fig. 7.1 Synthesis of methyl aromatic ketone……………………………………83

Fig. 7.2 Synthesis of biphenyl methyl aromatic ketone………………….……....84

Fig. 7.3 General synthetic procedure for compounds 1-24………………………85

Fig. 7.4 Synthesis of OSU02067………………………………..………………..86

Fig. 7.5 Synthesis of the 2nd series of compounds..……………………………..100

xiii

ABBREVIATIONS

1H-NMR 1H-Nuclear Magnetic Resonance ADME Adsorption, Distribution, Metabolism and Execretion Akt Protein Kinase B AR Androgen Receptor ATP Adenosine Triphosphate CDK Cycline Dependent Kinase c-Kit Stem Cell Factor Receptor CLL Chronic Lymphocytic Leukemia CML Chronic Myelogenous Leukemia COX-1/2 Cyclooxygenase-1/2 DHT Dihydrotestestrone EGFR Epidermal Growth Factor Receptor ELISA Enzyme-linked Immunosorbent Assay GIST Gastrointestinal Stromal Tumor GTP Guanine Triphosphate HR-MS High Resolution Mass Spectrum HRPC Hormone Refractory Prostate Cancer IGF-1R Insulin-like Growth Factor-1 Receptor IRK Insulin Receptor Kinase KDR Kinase-insert Domain Receptor LGA Lamarckian Genetic Algorithm LHRH Luteinizing Hormone-Releasing Hormone MAPK Mitogen-Activated Protein Kinase mTOR Mammalian Target of Rapamycin MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide NSAID Nonsteroidal Anti-Inflammatory Drug p-Akt Phospho-Protein Kinase B PARP Poly- (ADP-ribose) polymerase PCa Prostate Cancer PDK-1 3-Phosphoinositide Depdendent Kinase-1 PH Pleckstrin Homology PI3K Phosphoinositide 3 Kinase PIP2 Phosphatidylinositol-4,5-bisphosphate PIP3 Phosphatidylinositol-3,4,5-trisphosphate PK Protein Kinase PKI Protein Kinase Inhibitor PP2A Protein Phosphatase 2A PSTK Protein Serine/Threonine Kinase xiv

PTEN Phosphatase and Tensin Homologue Deleted on Chromosome Ten RTK Receptor Tyrosine Kinase SSKI Spectrum Selective Kinase Inhibitor VEGFR Vascular Endothelial Growth Factor Receptor

xv CHAPTER 1

INTRODUCTION

1.1 Hormone-Refractory Prostate Cancer (HRPC) and Therapies

Prostate cancer (PCa) is the most common type of cancer in American men other than

skin cancer and is the second leading cause of cancer death in men, exceeded only by

lung cancer. While 1 out of 6 men will get prostate cancer during his lifetime, only 1 in

32 men will die of this disease. The American Cancer Society estimates that there will be

about 230,900 new cases of PCa in the United States in 2004 (www.cancer.org). About

29,900 men will die of this disease. Age, nationalities, race, diet and family history are reported to contribute to the risk of PCa (1). It is usually diagnosed in older man with mean age at presentation of about 70 years. High incidence and mortality of PCa are found in the United States and western European countries. African-American men are found to have a higher incidence and lower 5-year survival rate for PCa than Caucasian men. Diet is also a major contributor to PCa in the United States. The typical US diet is rich in animal fat and meats but poor in fruit and vegetables, which contain a variety of antioxidants that have been correlated with reduced risk of PCa. Alternatively, this might explain why Asian men, who have an opposite dietary preference, have a lower incidence.

That relatives with PCa increases the possibility for other family members to be

1 diagnosed with this disease indicates a role for family history in PCa incidence.

PCa is usually a slow-growing cancer with a doubling time longer than 2 years (2).

Most PCa is detected in older men and remains latent throughout a person’s life. People

often die of other diseases despite being diagnosed with PCa. So it has been a long-time

debating issue whether patients should be treated in such situation. However, as PCa

becomes clinically very active, it will progress and threaten patient’s life. Treatment

depends very much on the stage of disease. For PCa localized in the prostate, it can be

treated by surgery, radiation therapy and cryotherapy (2). As PCa metastasizes, treatment depends on responsiveness to hormone therapy. In the early 1940s, Huggins et. al. showed dramatic improvement in PCa patients after surgical castration, which subsequently became a widely used first-line treatment for metastatic PCa still relying on hormone for growth (3, 4). Due to the negative psychological and gender-related side effects of surgical castration, medical castration (estrogens and Luteinizing Hormone-

Releasing Hormone analogs) were developed. Estrogens interfere with LHRH release and thus eliminate signals for androgen production by testes. However, estrogens went out of favor when LHRH analogs, which suppress LH production after chronic administration, became available. LHRH analogs have also been used in combination with androgen receptor (AR) antagonists (flutamide, bicalutamide and nilutamide), which is called combined androgen blockade (CAB).

However, for patients whose disease don’t respond to or eventually become resistant

to androgen deprivation treatment (called HRPC), new treatment strategies were

developed (2, 5). Second-line hormonal interventions are based on the thought that

HRPC is so heterogeneous and alternative hormone manipulation might overcome initial

2 resistance to androgen ablation. This second-line hormonal intervention includes

withdrawal of anti-androgens, second anti-androgen and suppression of adrenal

androgens. The rationale for anti-androgen withdrawal is based on the observation that a

flutamide metabolite could stimulate LNCaP prostate cancer cell growth. This was

further confirmed in a number of studies, not only for flutamide, but also for several other

anti-androgens including bicalutamide, megestrol and nilutamide (6-9). However, the

underlying mechanisms are still not clear. The use of a second anti-androgen is based on the idea that each anti-androgen interacts with AR in a unique way. In vitro data demonstrated that flutamide acted as a partial agonist while bicalutamide had an inhibitory activity in LNCaP cells (10). Therefore, alternative anti-androgens might achieve higher response rates. The adrenal cortex, another source of androgen other than testes, is targeted in second-line hormone manipulation to further block androgen production. For example, ketoconazole and megestrol are used to block adrenal androgens. However, these agents have lower response rates, shorter durations of effect and greater toxicity than the first-line therapies.

Chemotherapy hasn’t been recognized as an effective treatment until very recently.

The National Comprehensive Cancer Network (NCCN) currently lists six

chemotherapeutic regimens (ketoconazole/doxorubicin, estramustine/vinblastine,

estramustine/etoposide, estramustine/paclitaxel, docetaxel/estramustine, or

mitoxantrone/prednisone) as options for the management of HRPC (11, 12). Vinblastin,

paclitaxel and docetaxel act on tubulin. Etoposide, mitoxantrone and doxorubicin bind to

DNA and inhibit topoisomerase II. These traditional cytotoxic drugs are associated with

severe side effects. More importantly, these options are solely palliative probably

3 because PCa cells grow slowly and don’t respond very well to drugs targeting on cell

cycle. Therefore development of therapeutic agents with novel mechanisms is necessary.

Recent advances in the understanding of prostate cancer biology and clinical successes

of imatinib mesylate (Bcr-Abl/c-Kit/PDGFR) and trastuzumab (HER2/neu) in patients

have spurred enthusiasm for the discovery of novel agents targeting aberrant signaling

molecules in HRPC (13-16). As indicated in the recent review article about the success

of imatinib, comprehensive understanding of disease progression at the molecular level is

the prerequisite for correct target selection and drug development (17).

Arnold JT and Isaacs JT proposed a working model for androgen-independence

development based on the origin of malignant PCa cells (Fig. 1.1), which determines the

responsiveness to androgen ablation (18). The prostate epithelial compartment is

organized in a hierarchy of expanding stem cell units. The majority of the epithelium

compartment is composed of androgen-dependent glandular cells with lower numbers of androgen-sensitive basal cells and a limited number of androgen-independent basal stem cells. According to this model, PCa can originate from three different cell types. The first type is the androgen-independent stem cells. Since such malignant stem cells can progress down to the androgen-sensitive (but androgen-independent) amplifying cells and androgen-dependent transit cells, PCa of this kind is composed of heterogeneous 3 types of cells and should partially respond to androgen ablation. If PCa originates mono- clonally from androgen-sensitive amplifying cells, differentiation into androgen- dependent transit cells still makes this type sensitive to androgen ablation. Lastly, PCa that is mono-clonally derived from androgen-dependent transit cells should be highly

responsive to androgen ablation as long as heterogeneous development does not take

4 place. So the main issue here is to fully understand the mechanisms underlying the

survival of androgen-independent cells and heterogeneous development of androgen-

dependent cells.

In a recent review article, Feldman BJ and Feldman D summarized four mechanisms

by which HRPC can develop despite successful androgen ablation (Fig. 1.2) (19). In the

hypersensitive pathway, more androgen receptor (AR) is produced (usually by gene

amplification), or AR has enhanced sensitivity to compensate for low levels of androgen,

or more testosterone is converted to dihydrotestosterone (DHT) by 5α-reductase. In the promiscuous pathway, the specificity of AR is broadened so that it can be activated by non-androgenic molecules normally present in the circulation. In the outlaw pathway, receptor tyrosine kinases (RTKs) are activated, and AR is phosphorylated by either Akt or mitogen-activated protein kinase (MAPK), resulting in the ligand-independent activation of AR. In the bypass pathway, signaling, involving proteins such as BAD, obviates the need for AR and its ligand to provide survival signals for cancer cells. All these four mechanisms can possibly account for the heterogeneous development of androgen-dependent prostate cancer cells. Outlaw and bypass pathways are probably responsible for the survival of androgen-independent prostate cancer cells. As Akt is emphasized in both outlaw and bypass mechanisms for HRPC development, PI3K/PDK-

1/Akt signaling is summarized in the following section.

1.2 PI3K/PDK-1/Akt Signaling Pathway

For the past decade, RAS-the first identified oncogene-has been the focus of much effort in cancer research. Extensive studies of signaling proteins upstream or

5 downstream of this GTPase led to the model of mitogenic signaling by RTKs through

RAS and MAPKs (20). Recent approvals of trastuzumab, imatinib mesylate and gefitinib

confirmed the importance and success of targeting tyrosine kinases in cancer. The

PI3K/PDK-1/Akt signaling, downstream of RTKs, has emerged as the second important

pathway for therapeutic targeting in human cancer (Fig. 1.3).

Upon ligand binding to the extracellular domains of RTKs, RTKs become dimerized,

phosphorylated and activated. Phosphorylated tyrosine creates a binding site for class IA

PI3K p85 regulatory unit, which leads to full activation of PI3K. Then the 3-hydroxyl

group of PIP2 is phosphorylated to give PIP3, which in turn serves as a membrane tether for proteins with pleckstrin homology (PH) region, such as Akt and PDK-1. Binding of

Akt PH domain to membrane PIP3 causes the translocation of Akt to plasma membrane and brings Akt into close contact with PDK-1. Phosphorylation of Thr308 by PDK-1 and

Ser473 by DNA-dependent protein kinase results in fully activated Akt. Negative regulation of PI3K/Akt signaling can be achieved by protein phosphatases, including phosphatase and tensin homologue deleted on chromosome ten (PTEN), SH2 domain- containing phosphatase (SHIP1/2) and protein phosphatase 2A (PP2A). PTEN dephosphorylates PIP3 at the 3-position and SHIP converts PI(3,4,5)P3 to PI(3,4)P2 or

PI(4,5)P2 to PI(4)P, thereby counteracting the action of PI3K and inhibiting activation of

Akt. PP2A opposes the activity of PDK-1 by directl y dephosphorylating activated Akt.

Akt conveys its signal to many downstream proteins involved in apoptosis and cell

growth and proliferation (Fig. 1.3) (21-23). Akt phosphorylates apoptotic protein BAD,

which leads to dissociation of BAD from Bcl-2 and Bcl-xL, translocates BAD from

mitochondria to cytoplasm and restores the anti-apoptotic function of Bcl-2 or Bcl-xL.

6 FKHR is a nuclear transcriptional factor that stimulates the transcription of apoptotic

proteins (e.g. Fas ligand, p27). Akt inactivates FKHR by phosphorylating and

sequestering it in cytoplasm as a result of binding to 14-3-3 proteins. NF-κB is usually sequestered in cytoplasm by binding to inhibitory protein IκB. Phosphorylation of

IKKα/β by Akt dissociates IκB from NFκB and allows NFκB to enter the nucleus for regulating anti-apoptotic protein transcription (e.g. FLIP). Akt can also phosphorylate many proteins involved in cell growth and proliferation, such as MDM2, mTOR and

GSK3.

In human tumors, genes coding for proteins involved in these signaling pathways are frequently altered, resulting in abnormal proteins and aberrant signaling (21). Over- expression of upstream RTKs and truncated variant of epidermal growth factor receptor

(EGFR) are found in certain cancers. The gene that encodes p110 catalytic subunit of

PI3K is amplified in some ovarian cancers and amplification of Akt2 is found in breast, ovarian and pancreatic cancers. Truncated catalytic subunit of PI3K is reported in human primary tumor cells. Some human primary colon and ovarian cancers have mutated p85α catalytic domain. More importantly, PTEN mutation or inactivation can occur in a number of human cancers including prostate, lung, ovarian, melanoma and thyroid cancers. All these alterations lead to elevated PI3K/PDK-1/Akt signaling and provide cancer cells a very important survival pathway. This, along with its position as a convergence point for multiple signaling pathways, establishes PI3K/PDK-1/Akt signaling as a relevant and attractive target for therapeutic intervention.

Major proteins within PI3K/PDK-1/Akt signaling pathway have been chosen for drug development purposes. RTKs can be targeted either by antibodies that specifically bind 7 to the extracellular domains to block the natural ligand binding or by small molecules interacting with intracellular catalytic or allosteric domains. Trastuzumab (registered,

HER2/neu) and cetuximab (Phase III, EGFR) are such monoclonal antibodies developed against specific proteins. Gefitinib (EGFR), erlotinib (EGFR) and SU6668 (Phase II,

VEGFR) target the intracellular ATP binding domain and disrupt the activation of downstream proteins. Certainly, many small molecules with same mechanism of inhibition have been developed and are now in different phases of clinical trials.

Compared to RTK inhibitors, agents targeting PI3K/Akt/PDK-1 are relatively few, especially the clinically relevant agents. Wortmannin (irreversible inhibitor) and

LY294002 (reversible inhibitor) represent the early generation of PI3K inhibitors and are broadly used to understand the biological role of PI3K and its effector proteins. Since

PI3K has several isoforms, much effort has been focusing on developing isozyme specific inhibitors. Akt is a serine/threonine protein kinase and most of the Akt inhibitors developed so far are ATP competitors, such as staurosporine (IC50=11nM, most potent

Akt inhibitor known), its derivatives and H-89 (23). In addition, a series of phosphoinositol lipid analogs were designed to disrupt the membrane recruitment of Akt and thus block the activation of Akt. One such inhibitor, perifosine, developed by Keryx

Biopharmaceuticals, showed potential efficacy and tolerability in the treatment of patients with advanced soft tissue sarcoma (phase II trial) (www.keryx.com). Perifosine is so far the only Akt inhibitor in clinical development primarily for the treatment of cancer. Similarly, only a few inhibitors have been developed for PDK-1 (24, 25).

These inhibitors are based on the scaffold of staurosporine (a non-specific protein kinase

inhibitor). Currently, UCN-01(7-hydroxystaurosporine) is in phase II clinical trial.

8 mTOR is one of the most studied downstream substrates of Akt. It has been identified as

“target of rapamycin” by Michael Hall (91). Activation of mTOR by Akt stimulates the

translation of proteins important for cell cycle progression, such as cyclin D1. Inhibition

of mTOR, especially in PTEN inactive cells, induces growth inhibition. Currently, two

mTOR inhibitors (CCI-779 and RAD001) are in clinical trials as cancer therapy.

As the majority of proteins involved in signaling pathways are kinases and much

knowledge has been accumulated in the kinase field during the past 20 years, the next

section will briefly review protein kinases and their inhibitors.

1.3 Protein Kinases (PKs) and Their Inhibitors

1.3.1 Kinases and Cancer

The protein kinase family, one of the largest in the human genome, comprises about

500 genes and is estimated to include more than 2000 kinases (26). All protein kinases

thus far characterized with regard to substrate specificity fall within one of three classes,

serine/threonine (PSTK), tyrosine (PTK) and dual specificity protein kinases. In addition,

a different classification based on catalytic domain sequence alignment divides the

kinases into four families, AGC (PKA, PKG and PKC), CaMK (calcium/calmodulin- regulated kinase), CMGC (CDKs, MAPK, GSK3 and CKII) and OPK (other protein

kinases) (26).

Protein phosphorylation regulates most aspects of cell life, whereas abnormal

phosphorylation is a cause or consequence of disease including cancer. The involvement

of protein kinases in human malignancy may occur by the following mechanisms (28):

9 • Over-expression of kinases leads to deregulated signaling. For example, gene

amplification and transcriptional activation lead to EGFR over-expression,

enhanced dimerization, activation and downstream signaling. Over-expression of

EGFR is found in many cancers, including prostate, breast, lung and ovarian

cancers.

• Genomic rearrangement, such as chromosomal translocations, can result in

oncogenic fusion proteins with constitutive kinase activity. For example, t(9,22)

philadelphia chromosome translocation leads to constitutively active Bcr-Abl

(activation mechanism is not clear) responsible for myeloproliferative disease-

chronic myeloid leukemia.

• Mutation results in constitutively active kinases. Mutations in c-Kit, a receptor

tyrosine kinase, lead to ligand-independent kinase activation and subsequent

stimulation of the downstream signaling including PI3K/Akt (29). Mutations in

the juxtamembrane and kinase regions of c-Kit are associated with gastrointestinal

stromal tumor (GIST) and mast-cell/myeloid leukemia, respectively.

• Loss of tumor suppressor gene elevates the activity of corresponding kinases.

Tumor suppressor retinoblastoma (Rb) negatively regulates cell cycle by binding

to E2F thus blocking its transcriptional factor function (30). Loss of Rb results in

elevated E2F activity, leading to uncontrolled cell proliferation. PTEN, a protein

phosphatase, deactivates Akt signaling by hydrolyzing PIP3 to PIP2 and disrupting

the membrane recruitment of Akt. Mutation of PTEN, giving rise to increased

Akt signaling, is found in prostate, lung, ovarian, melanoma and thyroid cancers.

10 Most protein kinases possess a catalytic domain containing 250-300 amino acid

residues (30kD). There are about 30 kinases with 3-D structures determined so far (26).

For those with unknown 3-D structures, homology models can be used based on sequence

similarity of catalytic domains. The catalytic domain has a bilobal structure (N-terminal

lobe and C-terminal lobe, joined by hinge region residues) (Fig. 1.4). The ATP binding

site is located between these two lobes. This site, together with less conserved

surrounding pockets, has been extensively studied for inhibitor design and is described in

more details in section 1.3.2 (31).

As phosphorylation (or catalysis) process is accompanied by tremendous protein

conformational change (between off/inactive and on/active states), a great deal of attention has been focused on structural study in order to better serve the drug discovery purpose (32, 33). This conformational change usually involves distortions of two to four substructures, namely, activation loop, helix C, glycine-rich loop and ATP binding pocket. The most well known is the activation loop. Shown in Fig. 1.6 are schematic representations of inactive and active forms of the insulin receptor kinase (IRK). Upon phosphorylation the activation loop moves away from the catalytic domain and adopts a conformation allowing substrate binding and catalysis. The same process occurs in Bcr-

Abl (see section 1.3.2 for details). The significance of the conformational change from the inactive to the active form is two fold (34). Firstly, this change places Asp from DFG motif in the position to coordinate with magnesium ion for catalysis. Secondly, the change produces a docking site for the peptide substrate. From a drug discovery perspective, change in the activation loop also significantly affects small molecule binding, which is covered in section 1.3.2. What is also shown in Fig. 1.6 is the big

11 change in the position of helix C. In the active conformation, Glu at the center of helix C is placed in such position that it can form salt bridge with Lys (also binds to ATP phosphate). Such interaction is conserved throughout active kinases and plays an important role in activation. The glycine-rich loop refers to motif GxGxxG in the N-lobe

(Fig. 1.4). Conserved glycine residues allow close contact between the backbone of glycine-rich loop and ATP phosphate. This loop is very flexible and can accommodate various molecules. The ATP binding pocket is delineated by the three substructures mentioned above. So any changes in these three substructures can correspondingly alter the ATP binding pocket.

Detailed structural information not only provides insights into protein structures and functions but also helps researchers design small molecule inhibitors from a more rational approach.

1.3.2 Protein Kinase Inhibitors (PKIs)

Protein kinases are now the second most important drug targets, accounting for 20-

30% of the drug discovery programs of many companies next to G-protein-coupled receptors. A growing interest in developing orally active PKIs culminated in the approvals of imatinib, gefitinib and erlotinib (20).

A number of PKIs with different mechanisms of inhibition have been developed (35).

Monoclonal antibodies, such as trastuzumab (HER2/neu) and bevacizumab (VEGFR), bind to the extracellular domains of receptor protein kinases and disrupt natural ligand binding and subsequent kinase activation. Irreversible inhibitors refer to compounds that form covalent bonds with target proteins. Good examples are EGFR inhibitors, CI-1033 and EKB-569 (Fig. 1.7) (36). They both have an acrylamide moiety on their side chains,

12 which is placed in close proximity to conserved residue Cys 773 and then attacked by the

thiol group of Cys (Michael reaction) to form a covalent bond. Allosteric sites are

regions outside the real ATP binding site, but can regulate kinase activity. The MEK

inhibitor PD-184352 interacts with an allosteric site located in the N-terminal lobe of

MEK and locks the enzyme into an inactive conformation (35). The protein substrate

analog approach is thought to offer higher potency and specificity. Such analogs are

devised based on the assumption that the protein kinase recognizes its protein substrate by the amino acid residues around the phosphorylation site. Some peptide mimics were synthesized and evaluated against kinases but no promising results was seen partly because the assumption may not be valid. In fact, kinases use multiple docking sites other than phosphorylation site for substrate recognition. Bi-substrate analogs mimic the binding of both ATP and protein substrate and thus interrupt the phosphorylation process and shut down the downstream signaling. Theoretically, bi-substrate inhibitors are more specific and potent and a few papers described the academic efforts to develop such inhibitors for insulin-like growth factor-1 receptor (IGF-1R) and insulin receptor kinases

(IRKs) (37, 38). However, in reality no drugs based on this mechanism is seen in clinics.

Targeting ATP binding is so far the most successful approach. The fact that the majority

of small molecule PKIs, including these already on market, are ATP competitors

demonstrates that targeting ATP binding domain is a feasible approach although it has

been doubted for in vivo efficacy and selectivity due to the high intracellular

concentration of ATP (2-10mM) and conserved ATP binding site throughout kinases (35).

13 With the availability of 3-D structures of many kinases in complex with ATP,

detailed interaction of ATP with target protein is generalized in Fig. 1.5 and serves as a

guideline for structure-based inhibitor design (28).

-Adenine region. This region contains two important hydrogen bonds formed between

N6H/N1 of adenine and backbone CO and NH of hinge region amino acids. Most of inhibitors use at least one of these two hydrogen bonds to anchor the molecules in ATP binding domain. Adenine also forms van der waals interaction with lipophilic side chains of surrounding amino acids.

-Hydrophobic pocket. This pocket isn’t used by ATP but has been employed to increase the binding affinity and selectivity of many inhibitors (see chapter 5 for more details).

-Hydrophobic channel. This channel opens to solvent and isn’t used by ATP either. It could be exploited to increase the binding affinity although no report in this regard is available yet.

-Sugar region. The 3-OH group of ribose donates a H-bond to Glu in C-terminal lobe

(backbone carbonyl of Glu170 in PKA). Such interaction wasn’t seen so often in

inhibitor design as the hydrophobic interactions mentioned above.

-Phosphate binding region. Triphosphate forms electrostatic interactions with basic side

chains of Lys and Asp (e.g. Lys72, Asp184, Asn171 and Lys168 in PKA). It also

interacts with the backbone amide hydrogen of Gly-rich flap. As mentioned earlier, Gly-

rich flap is very flexible and can accommodate many molecules.

Other than taking advantage of ATP binding site and its surrounding area, different

conformations as a result of substructure distortions (more details in 1.4.1) can be utilized

for drug design. Both STI-571 (imatinib) and PD173955 are inhibitors for Bcr-Abl but

14 PD173955 is 10-fold more potent than STI-571(34). The co-crystal structures of Bcr-Abl

with both inhibitors explained that the potency difference derived from their ability to

bind to active form of Bcr-Abl (Fig. 1.6). Relatively smaller PD173955 can bind to both active and inactive forms of Bcr-Abl. STI-571, however, can only be accommodated in the inactive conformation due to the clash of phenyl and piperazine rings in STI-571 with

the activation loop in the active conformation.

All protein kinases catalyze the same reaction, transfer of γ-phosphate of ATP to the hydroxyl group of serine, threonine or tyrosine. Thus they share similar active conformations. While in inactive state, they have distinct conformations. From a medicinal chemist point of view, targeting inactive conformation for whatever the target protein chosen for drug design might achieve better selectivity. This, however, inevitably compromises the potency as stated in the example of STI-571. So it is a very challenging task to find an optimal balance between selectivity and high potency.

Structurally, PKIs are quite diverse. Fig. 1.7 lists some of chemical structures of

small molecule PKIs. Please refer to the review papers for more details about inhibitors designed for specific protein kinases (31, 36).

1.4 Cyclooxygenase-2 (COX-2) Inhibitors as Anti-cancer Agents

Although this thesis focuses on developing a novel class of PDK-1 inhibitors based

on celebrex skeleton, the whole story began with anticancer effect (or apoptosis inducing

effect) of celebrex. To make the story more complete, background information about its

primary target COX-2 and its role in cancer is given in this section.

15 The cyclooxygenases are responsible for the conversion of arachidonic acid (AA) to

prostaglandins (Fig. 1.8) (40, 41). Phospholipids get hydrolyzed by phospholipase A2 to give AA, an essential polyunsaturated fatty acid. AA is further oxidized in two steps by

COX to give prostaglandin H2 (PGH2). Specific isomerases then convert PGH2 to three groups of cyclic prostanoids, each of which plays a vital role in multiple physiologic and pathologic processes. There are three isoforms of COX, COX-1, COX-2 and COX-3.

COX-1, located on chromosome 9, is constitutively expressed in most tissues and responsible for maintaining physiologic processes such as gastric and renal protection and platelet function. COX-2, located on chromosome 1, is an inducible enzyme that is up-regulated in inflammation, angiogenesis and neoplasia. COX-3 was recently identified as splice variant derived from COX-1(44).

Non-steroidal anti-inflammatory drugs (NSAIDs) are a class of compounds that inhibit COX activities. Selective COX-2 inhibitors have been developed in late 90s to circumvent the gastric and renal toxicity due to the inhibition of COX-1 by NSAIDs (42,

43). As early as 80s, long-term use of aspirin was found to be associated with a decreased incidence of colorectal cancer. At the same time, another NSAID, sulindac, was reported to show regression in colorectal adenomas (45-47). Retrospective epidemiological studies suggest a decreased incidence of oesophagus, stomach, colon and rectum cancers in regular users of NSAIDs (48). However, due to the severe side effects

(e.g. gastrointestinal bleeding) of non-selective COX inhibitors, there was a reluctance to

push them forward. As selective COX-2 inhibitors (e.g., celebrex) were available in late

90s and demonstrated to have relatively mild side effects, much research has focused on

using such inhibitors in cancer treatment and prevention (recent withdrawal of vioxx

16 (COX-2 inhibitor) by FDA due to cardiovascular risk caused much concerns about safety

of other COX-2 inhibitors).

Much evidence has been accumulated to support the role of COX-2 in carcinogenesis

and possible mechanisms have been proposed (41). 1. Xenobiotic metabolism. COX-2

not only catalyzes AA metabolism but also converts procarcinogens to carcinogens. 2.

Inhibit apoptosis. Such phenomenon has been observed in many cell lines but the clear

linkage between COX-2 and apoptosis still is not clear. 3. Enhance angiogenesis. Each

prostaglandin has distinct roles in angiogenesis. For examples, TXA2 can promote

endothelial cell migration. Also, over-expression of COX-2 induces production of

VEGF, PDGF, bFGF and TGF-β, which bind to their corresponding receptors and increase vascular permeability, endothelial cell proliferation and migration. Furthermore,

COX-2 is essential for maintenance of the migration and attachment of endothelial cells through integrin pathways.

Celebrex (Fig. 1.8) was the very first selective COX-2 inhibitor approved by FDA for

the treatment of arthritis, acute pain, osteoarthritis and primary dysmenorrhea in late 1998.

A year later, FDA approved celebrex for the prevention of colon cancer in patients with

familial polyposis coli (FAP) based on Steinbach’s studies (49). In Steinbach’s trial,

FAP patients were randomized in a 2:2:1 fashion to take celebrex either 400mg (BID),

100mg (BID, dose used for osteoarthritis) or placebo for 6 months. Significant regression in tumor number and size was seen in higher dose (400mg) but not in lower dose (100mg). Also no obvious adverse effect was observed. Celebrex is under extensive preclinical and clinical studies either alone or in combination with other

therapies in colorectal, breast, lung, bladder, prostate and other types of cancer (49-53).

17 NCI website currently lists 30 active clinical trials of celebrex either for cancer

prevention or therapy (http://cancer.gov/clinicaltrials). Two of them are for prostate

cancer. One is phase II study of celebrex in patients with prostate cancer in biochemical

relapse after prior definitive radiotherapy or radical prostatectomy. The other is phase I

randomized study of neoadjuvant celebrex followed by prostatectomy in patients with

localized prostate cancer. This randomized, double blinded study evaluates the ability of

celebrex to modulate prostaglandin levels in tissue and determine the effects of this treatment on angiogenic factors.

8). The ndrogen- lium (1 ithe p mposed of a te e is co t prosta f tmen r o bers of androgen-sensitive basal cells tion zed in a hierarchy of expanding stem cell ganiza gani r elium compa the epith l for the or f e artment is o ity o l comp e major Stem cell mod te epithelia Fig. 1.1 prosta units and th dependent glandular cells with lower num

PC development (19). re Publishing Group Copyright) Possible pathways for HR

Fig. 1.2 (under the permission of Natu

Fig. 1.3 PI3K/PDK-1/Akt signaling pathway (21). (under the permission of Nature Publishing Group Copyright)

Helix C

Fig. 1.4 Crystal structure of EGFR with Tarceva (1M17). N-lobe is largely β-sheet and c-lobe consists of many α helices. These two lobes are connected by hinge region amino acids. Small molecule inhibitor (Tarceva) binds to ATP binding domain.

line represents H-bonds (28). P is in red. Dotted nases. AT ATP binding site of protein ki Fig. 1.5

23 αC helix αC helix

Activation loop

HN R HN N+ N O H CO N O N(H C) O N 3 2 3 (H2C)17 O P O H3CO O O N O- H N CI1033 R= Tarceva Perifosine O EGFR Akt Iressa R=OCH3 EGFR O Cl

HN NH N O N NN Cl O O N S N N N O H N N LY294002 PD173955 PI3K Imatinib Bcr-Abl Bcr-Abl

H H Cl H N (CH ) O N OH N 3 3 NN HN O

N O N N N R N O s

R O R N O HN

vatalanib BIRB-796 UCN-01 VEGFR-1/2 p38MAP Kinase chk1/PDK-1

+NH H N O OH HO O COOH Cl HN SU-6668 OH O KDR Rapamycin Flavopiridol mTOR CDK9

Fig.1.7 Chemical structures of selected kinase inhibitors

Phospholipids

PLA2

COOH

"Housekeeping" substances COX-1 IITXA (for platelet clotting) Arachidonic acid 2 (constitutive) IIPGE2 (for kidney function) IIPGI (for stomach protection) Cyclooxygenase 2

O COX-2 COOH "Inflammatory" prostaglandins (inducible) Contribute to O -pain -heat -swelling

PGH2

Prostacycline Prostaglandins Thromboxane A2

F3C

N N

SO2NH2

Fig. 1.8 Top. The role of cyclooxygenases. Bottom. Structure of celebrex

26 CHAPTER 2

MOLECULAR TARGET(S)

FOR CELEBREX AND PROJECT DESIGN

2.1 Molecular Target(s) of Celebrex

2.1.1 COX-2 Independent Mechanism

Although celebrex has been used so broadly in preclinical and clinical studies for many different cancers, the underlying mechanisms remain elusive. It is generally been believed that inhibition of COX-2 is integral to its anti-tumor effect. Many studies demonstrated that prostaglandins and other COX-2 generated downstream mediators promoted tumor cell proliferation, survival and angiogenesis in an autocrine and/or paracrine manner (54-58). It’s been reported that knockout of COX-2 gene can suppress tumorigenesis in mice that have a genetic predisposition to form polyps (59). Moreover, animal studies have indicated that efficient tumor growth requires COX-2 in the host and enhanced COX-2 expression was sufficient to induce mammary gland tumorigenesis (60,

61).

However, there is much evidence to support another argument of COX-2 independent mechanisms. For example, sulindac metabolites without COX-2 inhibitory

27 activity are potent inducers of apoptosis (45). In our laboratory, we focus on COX-2 independent mechanisms in prostate cancer cells. In 2002, Song and Lin successfully demonstrated that celebrex-induced apoptosis was independent of COX-2 by using both biochemical and chemical strategies (62).

PC-3 clones, 2F6, 1F2, 3D9 and 7D9 with various expression levels of COX-2, were generated (Fig. 2.1). COX-2 deficient clone 2F6 was as sensitive to celebrex treatment as parental PC-3 cells. Clone 7D9 expressed much higher level of COX-2 than parental

PC-3 cells in the absence of doxycycline (Dox) but barely detectable COX-2 in the presence of doxycycline. Despite different COX-2 levels with or without Dox, clone

7D9 responded to celebrex treatment in the same patterns. Moreover, chemical strategy was employed to confirm COX-2 independent mechanism from a totally different perspective. Seven celebrex derivatives with various COX-2 inhibitory activities were synthesized and evaluated for their potency in inducing apoptosis in PC-3 cells (Fig. 2.2).

Compound 2, with hydrogen replacing methyl group at para-position of phenyl ring in celebrex, was comparable to celebrex in COX-2 inhibition. However, it wasn’t able to kill 50% PC-3 even at 100 hr while it only took celebrex 2hrs to kill 50% cells at same concentration. Same is true for compound 3. On the other hand, Compound 7 (also called DMC, with 2,5-dimethyl groups on phenyl ring), which was about 2500 fold less potent than celebrex in COX-2 inhibition, killed 50% of the PC-3 cells in 1hr. This phenomenon was also observed in compound 6 with 2, 5-dichloro moiety on phenyl ring.

Therefore, there was no correlation between COX-2 inhibition and apoptosis-inducing potency for these seven analogs.

28 Both biochemical and chemical evidence demonstrated that apoptosis was not

mediated through COX-2 inhibition. Then what would be the possible target(s) for

celebrex?

2.1.2 PDK-1, a Major COX-2 Independent Target for Celebrex

In 2000 J. Biol. Chem. paper, Hsu et.al. reported that celebrex induced apoptosis in

both androgen-dependent LNCaP and androgen-independent PC-3 cells at 50µM (63).

Further mechanistic studies showed that Bcl-2 expression level wasn’t affected and

ectopic Bcl-2 expression didn’t rescue the PC-3 cells from death by celebrex treatment.

But it did decrease phospho-Akt level while total amount of Akt was kept. Transient

transfection of constitutively active Akt rescued 42% of PC-3 from death. However, as

celebrex did not inhibit PI3K, decrease in phospho-Akt wasn’t due to inhibition of PI3K.

In 2002 J. Biol. Chem. paper, Arico et. al. reported that celebrex inhibited PDK-1 in

colon cancer HT-29 cells (64). Similarly, Arico observed the apoptosis-inducing effect

of celebrex in HT-29 cells. Phospho-GSK-3β and phospho-Akt level were decreased

upon celebrex treatment but transfection of constitutively active Akt only had a slight

protection against celebrex-induced apoptosis. So they went upstream of Akt and looked

at the change on PDK-1. Celebrex was shown to inhibit PDK-1 kinase activity and over-

expression of constitutively active PDK-1 reduced 65% of celebrex-induced apoptosis

compared to empty vector.

Therefore, combining these two studies we decided to focus on PDK-1 in androgen-

independent PC-3 cells. In vitro PDK-1 kinase activity assay showed that IC50 for celebrex was about 48µM (Fig. 2.3A), which was parallel with Hsu’s data showing

29 precipitous decrease in phospho-Akt after 1hr celebrex treatment while total amount of

Akt was maintained (Fig. 2.3B). Furthermore, at cellular level 50µM celebrex killed

about 50% PC-3 cells in 2hr and over 90% in 4hr (Fig. 2.3C). As mentioned in chapter 1,

Akt has several downstream proteins (BAD, FKHR) leading to apoptosis. Therefore, the

observed cell killing effect probably resulted from PDK-1 inhibition by celebrex.

In a more recent study by Kulp et. al. to correlate the in vitro activity of celebrex to in

vivo efficacy, PDK-1 was further demonstrated to be a major COX-2 independent target

for celebrex and its analog DMC (65). In vitro, both compounds showed anti-

proliferative effects on PC-3 cells in serum-supplemented medium, which was attributed

to G1 cell cycle arrest at lower concentrations and apoptosis at higher concentrations.

Although in animal study celebrex (200mg/kg) only had marginal effect, DMC at

200mg/kg did inhibit xenograft tumor growth. More importantly such effect was correlated to DMC in vitro potency through inhibiting Akt signaling.

2.2 Aims and Project Design

PDK-1 was discovered in late 90s and much research has been done about PDK-1 (66).

But it is still a relatively new protein from a drug discovery point of view. Only a few

PDK-1 inhibitors have been reported so far, including staurosporine, UCN-01, LY333531 and BIM (bisindolyl maleimides, structures and IC50s are shown in Fig. 2.4). Both

LY333531 (specific for PKCβ) and BIM are nanomolar PKC inhibitors with LY333531

in phase III clinical trial for prevention of diabetes complications. Staurosporine and

UCN-01 are nonspecific PKIs and have been found to inhibit many kinases, e.g. CDK,

PKC. UCN-01 is the only PDK-1 inhibitor in clinical trial for cancer (67). Therefore, 30 the discovery of PDK-1 inhibition by commercial drug celebrex (very simple molecule)

really provided a platform for medicinal chemist to explore.

Due to the minor PDK-1 inhibition by celebrex at therapeutic concentrations (1-5µM), it is far from being used as a PDK-1 inhibitor clinically. So developing more potent

PDK-1 inhibitors based on celebrex skeleton is necessary. Furthermore, as mentioned in

chapter 1, PDK-1/Akt signaling not only represents the convergent point of many RTK

signalings but also is found to be highly active due to mutation of PI3K and PTEN and

over-expression of Akt and upstream RTKs (see details in chapter 1). Therefore it is an

attractive target for the treatment of many different cancers, including prostate, breast,

lung and ovarian cancers. Further structure modification of celebrex should provide a

novel class of PDK-1 inhibitors.

Previous structure-activity relationship (based on cell killing effect) studies provided

us the following working model (Fig. 2.5) (68).

• 5-position methyl phenyl group and 3-position trifluoromethyl group are important

for van der waals interaction. Replacement of methyl group with polar groups,

such as OH, NH2, NO2, OCH3, CN and F, significantly reduced their cyto-toxicities.

Also substitution of methyl phenyl moiety with thiophene, furan and pyridine ring

totally abolished the cell killing effects. But phenyl ring can accommodate alkyl,

CF3 and Cl at different positions without compromising their apoptotic potencies.

In addition, deletion of trifluoromethyl moiety at 3-position gave rise to decreased

potency.

• Pyrazole ring forms electrostatic interaction with target protein. Modeling showed

that two nitrogen atoms in pyrazole ring offered negative electrostatic potential to

31 the bottom of pyrazole ring while relatively positive potential on the opposite site.

Vioxx with totally reversed distribution of electrostatic potential around lactone

ring wasn’t active in PC-3 at all. Structure modification aiming to mimic

electrostatic potential of pyrazole ring in celebrex led to compounds with similar

potency to celebrex.

• Sulfonamide group hydrogen bonds to target protein. Substitution of sulfonamide

(SO2NH2) by methylsulfone (SO2CH3) resulted in a compound totally inactive in

PC-3 cells. For some analogs with carboxamide (CONH2) at this position, similar

potency was observed in PC-3 cells. So as long as H-bonding capability is

preserved at this position, sulfonamide moiety can be modified.

With all the information mentioned above, three goals were set for this project and corresponding strategies were proposed to achieve each of the goals.

Aim 1. Explore the structure modification of celebrex and study the structure features required for PDK-1 inhibition.

Design 1. This model suggested that although the electronegative potential surrounding pyrazole ring is essential for apoptosis-inducing effect, the 5-aryl and sulfonamide (-

SO2NH2) moieties are amenable to modification. So modification can be carried out in

two steps. First is at 5-position. As this position is believed to form van der waals

interaction with target protein, a variety of aromatic rings with alkyl, CF3 and Cl groups will be put at this position (covered in chapter 3). The rest part of structure is kept till the best substituent at 5-position is found. Then the second step of modification can be carried out at sulfonamide moiety (covered in chapter 4). Heteroatom-rich functional groups (able to form H-bond) will replace sulfonamide as 5-position is occupied by

32 optimal aromatic ring. Detailed structure-activity relationship can be generated from

biological activities and structural features.

Aim 2. For the optimized compounds, confirm PDK-1 as a target and correlate PDK-1 inhibition to in vitro potency observed in PC-3 cells.

Design 2. Check PDK-1 downstream protein Akt and p70s6k kinase activity in drug-

treated PC-3 cells. Measure the cyto-toxicity in PC-3 cells and subsequent cellular events

(e.g. apoptosis).

Aim 3. Apply optimized PDK-1 inhibitors to other cancers and explore the potential

clinical translation.

Design 3. Test PDK-1 inhibitors in other major cancers, such as breast, lung, leukemia,

thyroid and oral cancers. Combine such inhibitors with available chemotherapies to see if there is any beneficial effect.

Fig. 2.1 A. Western blot analysis showing COX-2 protein levels in parental PC-3 cells and four independent antisense COX-2 clones in the presence (+) or absence (-) of doxycycline (Dox, 2mg/mL) for 10 days. B. Susceptibility of prostate cancer cells to celebrex-induced apoptosis is independent of COX-2 expression levels. Left panel. Effect of 50µM celebrex on the viability of parental PC-3 cells and the COX-2 deficient clone 2F6. Right panel. Effect of 50µM celebrex on the viability of the COX-2 antisense clone 7D9 with (+) or without (-) a DOX pretreatment (2mg/mL)(62).

ired -time requ 1/2 e general structure of these 50% COX-2 activity; T s of celebrex and compounds 1–7. Th -concentration inhibiting 50 62). ( death cell Structures and characteristic 50% for Fig. 2.2 molecules is shown at the top. IC

B C

Fig. 2.3 A. Dose-dependent inhibition of recombinant PDK-1 kinase activity by celebrex (71). B. Western blot analysis of Akt and phosphor-Akt in PC-3 cells after treated with celebrex or DMSO for indicated time (63). C. PC-3 cells were susceptible to the treatment of celebrex in a dose and time-dependent manner (63).

H H H N N N O O O O O O

NH N NH N N N

O N N N

LY333531 BIM-1 BIM-2 IC =0.75µM IC =9µM 50 50 IC50=14µM

H H H N N N O O O O O OH

N NH N N N N O H CH3

NH2 NH2 NH

BIM-3 BIM-8 UCN-01 IC =4µM 50 IC50=1µM IC50=0.006µM

Fig. 2.4 Structures and IC50s for reported PDK-1 inhibitors

Van der waals interaction Van der waals interaction F3C

N N

Electronegative area for electrostatic interactions SO2NH2

Hydrogen bonding

Fig. 2.5 Working model outlining the structural features essential for the apoptotic inducing effect of celebrex.

38 CHAPTER 3

DESIGN, SYNTHESIS AND BIOLOGICAL

ACTIVITIES OF 1ST SERIES OF CELEBREX DERIVATIVES

3.1 Design of the 1st Series of Derivatives

In a 2000 J. Med. Chem. paper, Fesik et.al. reported NMR binding data of molecular motifs to 11 proteins (69). It was found that compounds with biphenyl substructure were

5 to 100 times more likely to bind to 5 out of 11 proteins, including enzymes, DNA- binding proteins and proteins interacting with each other. Different substitutions on biphenyl ring offered different specificity to target proteins. The preference for biphenyl structure was explained as follows: biphenyl moiety offers certain degree of flexibility and interaction in binding and interacts favorably with aromatic, hydrophobic and even polar residues in the target protein. So incorporating biphenyl substructure into celebrex molecule became the theme of our first structure modification effort.

The working model of celebrex shown in chapter 2 indicated that 5-position aromatic

ring was believed to form van der waals interactions with target protein PDK-1 (crystal

structure was not published at that time). To further explore the size and shape of this

hydrophobic pocket, biphenyl substructures with various substituents (based on Fesik’s

privileged molecule concept) and other aromatic rings were selected to replace 5-position 39 methylphenyl group in celebrex. Their synthesis and biological activities were illustrated in section 3.2 and 3.3.

3.2 Synthesis of the 1st Series of Derivatives

The 24 compounds were synthesized according to the general procedure outlined in

Fig. 3.1 (see details in Chapter 7). Ar represented the respective aromatic ring structures.

Aromatic ketone (VI) reacted with ethyl trifluoroacetate in the presence of sodium

hydride to give the intermediate (VII) in 90% yield, which was further refluxed with 4-

hydrazinobenzene-1-sulfonamide hydrochloride in ethanol to yield the target product

(80% yield for OSU02067). Both reactions gave very high yields and could easily be

adapted to large scale synthesis for animal study. For the coupling reaction (2nd step), it’s preferred to be carried out in acidic condition (use hydrochloride salt instead of free hydrazine) to minimize the formation of the other isomer. It is been reported that diketone condensed with 4-methylsulfoylphenylhydrazine to give 1:1 mixture of two regioisomeric products, namely, 1, 3-diarylisomer and 1,5-diarylisomer (Fig. 3.2) (70).

When reaction was conducted at PH=1, ratio of isomers changed to 30:1 with 1, 5-diaryl isomer predominating.

For compound 1-4, the starting material (Aromatic ketone, II) was synthesized in

high yield from Friedel-Crafts reaction by reacting substituted benzene (I) with acetyl

chloride. For the biphenyl derivatives 10-20, the aromatic ketone (V) was generated by

Suzuki coupling reaction, in which aryl boronic acid (III) reacted with aryl halide (IV)

under the catalysis of Pd(OAc)2. This is a modified Suzuki coupling reaction as it uses ligandless palladium-Pd(OAc)2 in the presence of tetrabutylammonium bromide in water

40 (99). This condition accelerates the reaction and gives high yield compared to the

traditional Suzuki coupling reaction condition. Also this kind of reaction isn’t affected

by steric hindrance, so biphenyl substructures with various subsitutents can be integrated into analogs.

3.3 Structures of Analogs and Their Biological Activities

Table 3.1 listed structures of 24 analogs and their IC50s in inhibiting PDK-1 and PC-3 cell growth (71).

Compared to celebrex, modification at 5-position with substituted phenyl ring,

biphenyl ring and tricyclic aromatic ring increased the potency to different extents except

for compound 8, which has exceptionally low potency. Generally, tricyclic aromatic

derivatives (21-23) were more potent than biphenyl derivatives (9-19), which in turn

were more potent than phenyl ring derivatives (1-6). This might be due to the increased

van der waals interaction with the target protein and hydrophobic interaction by

desolvation of more apolar surface of target protein (see chapter 4 for details).

Obviously the best leads in this series were compound 13 and 23, which had comparable

potency in PDK-1 (IC50=9µM) and PC-3 growth inhibition (IC50=5µM). This represents about 5 to 6 fold increase over that of celebrex. Compound 24, isomer of compound 23, exhibited a precipitous decrease in PDK-1 inhibitory activity, which indicated the steric hinderance imposed by an unfavorable orientation of tricyclic aromatic ring. There was a correlation between PDK-1 and PC-3 growth inhibition potency.

In parallel with our drug modification, in vivo screening of these two compounds in nude mice PC-3 xenograft model indicated that compound 23 was tolerated well by mice

41 without any observable toxicity (unpublished data). Therefore, it (designated as

OSU02067) was chosen over compound 13 as new lead compound in the design of 2nd series of analogs.

3.4 OSU02067

3.4.1 Effect on Downstream Protein Akt.

The effect on downstream protein Akt was checked to confirm PDK-1 was inhibited by OSU02067 (Fig. 3.3). Top panel compared the ability of OSU02067 to inhibit Akt

kinase activity with that of celebrex. OSU02067 inhibited 50% Akt kinase activity at

5µM, which represented 6-fold increase over that of celebrex (30µM). Similarly, western

blot showed that OSU02067 led to significant decrease in phospho-Akt level at 5µM

while the total amount of Akt was preserved. However, celebrex only exhibited similar

phenomenon at concentration 30µM or higher.

3.4.2 Cellular Effect of OSU02067

Cell growth inhibition was measured in 1%FBS supplemented RPMI1640 medium.

The dose-dependent and time-course effect of OSU02067 on PC-3 cells was shown in

Fig. 3.4A. IC50s for PC-3 cell growth inhibition were 5µM for 24 hr treatment and

2.5µM for 72 hr treatment.

As PDK-1/Akt signaling played an important role in apoptosis, the apoptotic cell death was measured after PC-3 cells were treated with OSU02067 (Fig. 3.4 B). Mono and oligo-nucleosomes, which were quantified by using cell death ELISA kit, formed in a dose-dependent manner. At 5µM or higher, significant numbers of nucleosomes were formed, indicating lots of cells underwent apoptosis. Another evidence of apoptosis is 42 the cleavage of poly (ADP-ribose) polymerase (PARP) from 110kDa to 85kDa fragment.

Consistent with nucleosome formation, cleaved PARP fragment was observed at 5µM or higher.

The growth inhibition effect was also tested in 10%FBS medium (Fig. 3.4 C). At

5µM, OSU02067 completely suppressed the PC-3 cell growth after 6 day treatment.

Even as low as 1µM, It exhibited significant growth inhibition, which was better than the effect of 30µM celebrex.

3.4.3 In vivo study of OSU02067

OSU02067 was further tested in nude mice PC-3 xenograft model for in vivo efficacy.

Mice treated with 200mg/kg daily dose for 28 days had tumors half of size as that in vehicle treated mice (unpublished data) without any obvious toxicity.

OSU02067 represented the optimal derivative from 1st series of analogs not only in inhibiting PDK-1 activity but also in suppressing PC-3 cell growth. As mentioned in chapter 2, further modification would be carried out at sulfonamide group while keeping the phenanthrene moiety at 5-position.

CH COCl 3 O AlCl3/CS2 R R

I II

OH O Pd(OAc) , K CO O R B + Br 2 2 3 R Bu NBr, H O OH 4 2 III IV V H N 2 NH .HCl

F3C Ar N N O OH O CF3COOEt NaH/THF SO2NH2 Ar F3C Ar CH3CH2OH SO NH VI VII 2 2

Fig. 3.1 Synthetic route for aromatic ketone and the 1st series of derivatives

F C NHNH2 3 N N

CH S 3 O O S O O O CH O 3 1,5-diaryl isomer CF 3 F

Diketone F3C N N

O S CH O 3 1,3-diarylisomer

Fig. 3.2 Two isomers formed as diketone is coupled with 4-methylsulfoylphenylhydrazine (70).

IC50 (µM) IC50 (µM) No. Ar PDK- No. Ar PDK- PC-3 PC-3 1 1 Cele- 48 30 13 9 5 Coxib

1 42 18 14 15 8

2 38 17 15 18 8

3 32 17 16 20 11

4 34 18 17 17 9

5 20 9 18 32 15

6 34 18 19 32 15

7 24 11 20 15 8

8 65 31 21 16 9

9 21 11 22 12 7

10 22 9 23 9 5

11 18 10 24 42 23

12 23 10

Table 3.1 Structures and IC50s for inhibiting recombinant PDK-1 kinase activity and for inducing apoptotic death in PC-3 cells for celebrex and 24 derivatives. The general structure of these compounds is shown at the top. Ar represents the respective aromatic ring structures. The reported IC50 values are concentrations at which recombinant PDK-1 kinase activity is inhibited by 50% or at which PC-3 cell death measures 50% relative to DMSO control after 24h exposure in 1% fetal bovine serum-containing RPMI 1640. 46

A B

48 CHAPTER 4

STRUCTURE-BASED DESIGN OF PDK-1 INHIBITORS

4.1 Crystal Structure of PDK-1 Catalytic Domain in Complex with ATP

PDK-1, a monomeric serine/threonine kinase, belongs to AGC kinase superfamily.

It was originally isolated from rabbit skeletal muscle and brain cytosol and ubiquitously expressed in human tissues and cells (72). It contains an amino-terminal kinase domain, followed by a linker region and PH domain at carboxyl terminus. The crystal structure of

PDK-1 catalytic domain in complex with ATP was solved in 2002 and its coordinate was released a year later. Like many other protein kinases, its catalytic core contains an N- terminal lobe mainly β-sheet (yellow), a C-terminal lobe consisting of α-helices (purple) and hinge region which connects C-terminal and N-terminal lobes. (Fig. 4.1 Top). ATP binding site is located between these two lobes. The adenine moiety forms two hydrogen bonds with hinge region residues Ser160 and Ala162 which are believed to anchor ATP into the binding pocket (Fig. 4.2). N(6)H donated a H-bond to backbone

carbonyl (CO) of Ser160 and N1 accepted a H-bond from backbone NH of Ala162.

Adenine moiety also forms van der waals interaction with lipophilic side chain of several

amino acids, namely, Leu88, Val96, Ala109, Val143, Leu159, Ala162 and Leu212 (Fig.

4.1 Bottom). Small molecule inhibitors mimic both H-bonds and van 49 der waals interactions to disrupt ATP binding. But they might not use exactly the same

residues as adenine does for such interactions. The 3-OH of ribose moiety donates an

H-bond to carboxylic acid side chain of Glu166. Triphosphate formed three H-bonds with Ser92, Ser 94 and Lys111 respectively. Ribose and triphosphate binding domains are open to solvent and more hydrophilic so they aren’t used as often as adenine binding domain in small molecule design.

4.2 Docking of OSU02067 to ATP Binding Domain and Design of 2nd Series

Derivatives

Kinetic study of PDK-1 inhibition by celebrex with respect to ATP revealed an inverse relationship between PDK-1 inhibition and ATP concentrations, which indicated that celebrex inhibited PDK-1 by competing with ATP for binding (Fig. 4.3) (71). This

is very common among many protein kinase inhibitors, including imatinib, gefitinib and

tarceva. So as the crystal structure of PDK-1/ATP was available, we were able to use

molecular modeling tool to help us design 2nd series of compounds.

Autodock 3.0 is a docking program developed by A.J. Olson et. al. at Scripps

(http://www.scripps.edu/pub/olson-web/doc/autodock/tools.html) (73). It uses genetic algorithm for global conformation search with energy minimization as a local search method (Larmackian genetic algorithm, LGA). The implementation of LGA allows extensive flexibility (more than 8 rotatable bonds) in ligands to be evaluated while receptors remain rigid throughout. The scoring function shown below is based on

AMBER force field from which protein and ligand parameters are taken.

50 i, j i, j

i, j i, j

It contains five terms (indicated in the subscript of each term): van der waals interaction,

H-bonding, electrostatic interaction, torsional restriction and desolvation. Each of the

five terms is scaled by empirically determined factors (∆G) using linear regression

analysis from 30 protein-ligand complexes with known binding constants.

AutoDock tools (ADT) is a python-based graphical user interface, which helps to set

up input files for the docking in an automatic and visible fashion. Especially during the

process of setting up grid box for energy calculation, the actual visualization of grid box around protein makes parameters adjustment much easier. During the process of docking Sybyl6.9 is used as a generally tool for manipulating small molecules and proteins as it has more functions in this regard than in ADT. Please refer to chapter 7 for

details about setting up input files and analyzing output files.

Docking of OSU02067 to ATP binding domain of PDK-1 showed that binding mode

of this compound was somewhat different from that of ATP (Fig. 4.4). The pyrazole ring

was perpendicular to the ribose ring in ATP although the benzenesulfonamide moiety

occupied the adenine-binding motif. The phenanthrene ring, which was oriented at the

back of ATP triphosphate group, interacted with the anti-parallel β-sheet (containing

glycine rich loop). The potency difference between celebrex and OSU02067 was

explained by desolvation of more aploar moieties in PDK-1 by hydrophobic

phenanthrene ring. Assuming that the measured IC50 values are proportional to the Kds for PDK-1-inhibitor interactions, the relative free energy difference between OSU02067 51 and celebrex is ∆∆Gºcelebrex→OSU02067 = -4 kJ/mol. It’s been reported that upper limit for free

energy associated with desolvation of hydrophobic group was about -136 J/mol•Å2 (74).

32Å2 more apolar surface was covered by PDK-1 in OSU02067 than in celebrex, which gave free energy value of -4.35 kJ/mol. This is quite close to experimentally measured free energy difference. No hydrogen bond was formed between pyrazole ring and PDK-1 despite that ribose of ATP accepted a hydrogen bond from Glu166. Benzene sulfonamide was buried in the same pocket as adenine. Leu88, Val96, Tyr161 and

Leu212 side chains formed van der waals interaction with benzenesulfonamide. The sulfonamide group hydrogen bonded to the backbone NH of Ala162 in the hinge region.

This mimicked one of the two hydrogen bonds formed between adenine and PDK-1 hinge region residues, which were important for anchoring both ATP and small molecule inhibitors at this binding domain. So in the 2nd series of analogs design, heteroatom rich functional groups were chosen to substitute sulfonamide in order to mimic both of the hydrogen bonds. Twelve derivatives were synthesized and evaluated for PDK-1 and PC-

3 growth inhibition (71).

4.3 Synthesis of the 2nd Series of Compounds

Considering the reactivity of cyano and amino functional group, the intermediate (VII) was reacted with 4-hydrazinobenzonitrile, 4-hydrazinophenylacetonitrile and 4- nitrophenylhydrazine hydrochloride to give 26, 31 and VIII respectively, which could further react with various reagents to give heteroatom-rich compounds (Fig. 4.5, 4.6, 4.7).

Compound 26 with cyano group on phenyl ring was reduced to aldehyde by DIBAL-H at

-40ºC, which was easily condensed with hydroxylamine and hydrazine to give products 52 29 and 30. In addition, cyano group could be directly attacked by hydroxylamine and sodium azide to form compounds 27 and 28. Same reactions were employed to synthesize compound 32 and 33. As for intermediate VIII, nitro group was reduced to amine (IX) under the catalysis of palladium. This amino functional group formed an amide with carboxylic group of t-Boc protected glycine in the presence of coupling reagent 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (DEC). The resulted intermediate was de-protected to give the final product 34 with free alkyl amine.

This synthetic route can be applied to most of amino acids. Amino group in IX was also able to serve as nucleophile and react with cyanamide to give compound 35 with guanidino group on phenyl ring. Similarly, attack of sodium isocyanate by amino group in acidic condition resulted in compound 36.

The isomer issue mentioned in chapter 3 still exists here for VIII (Fig. 4.8). 4- nitrophenylhydrazine hydrochloride seems more reactive than its counterparts since the coupling reaction finishes in 2hrs reflux. Isomeric product increases as the reflux continues. This lets us hypothesize that the desired product is kinetically controlled while the side product is thermodynamically controlled as it looks more stable due to the less steric hindrance between substituents on pyrazole ring. As the side product causes much trouble for purification, the reaction needs to be carefully controlled by reaction time and monitored by thin layer chromatography (TLC). To adapt this procedure to large scale synthesis, Dr. Wang modified the procedure by changing the reflux to stirring at room temperature while adding molecular sieve to help the dehydration process of the coupling reaction. This works very well in terms of both avoiding side reaction and getting high yield.

53 4.4 Optimal Compounds-OSU03012 and OSU03013

Twelve compounds were evaluated for their potency in inhibiting PDK-1 and PC-3

cell growth (Table 4.1). To our surprise, IC50 for these 12 compounds varies so much

that some of them (26, 27, 31, 33, 36) are much less potent than parent compound

OSU02067. So increasing the number of heteroatoms doesn’t necessarily improve the

binding affinity as we expected in the analog design. However, we were fortunate

enough to see improved activities in two derivatives-OSU03012 and OSU03013,

showing 1.6 and 4.5 fold increase over OSU02067 in PDK-1 inhibition respectively.

Then could such improvement come from increased H-bonding capability?

Docking of OSU03013 to PDK-1 explained that the increased potency over

OSU02067 might be due to the additional hydrogen bond formed with Ser 160 (Fig. 4.9).

The structure similarity between the guanidino group of OSU03013 and adenine moiety

of ATP supported this model. Moderate increase in the binding affinity was in part due to the non-linear alignment of three atoms (OHN) involved in this hydrogen bond.

Comparing docked structure of OSU03012 and OSU03013 to that of OSU02067 (Fig.

4.10 and 4.11), similar conformations were seen for all three molecules as evidenced by almost identical residues involved in interaction. However, they do have unique conformations. Pyrazole ring in OSU03012 and OSU03013 is projected more outwards and closer to hinge region. Such movement might place glycine moiety in OSU03012 and guanidino moiety in OSU03013 in the right position to form additional H-bond with

Ser160.

As OSU03012 and OSU03013 were the two optimal compounds from this series, they were selected for further biological studies.

54 4.5 In vitro Effect of OSU03012 and OSU03013 in PC-3

4.5.1 Effects on Downstream Proteins Akt and p70S6K

OSU03012 and OSU03013 represented the two most potent inhibitors for PDK-1 with

IC50 of 5 and 2µM respectively (Fig. 4.12A). To further confirm PDK-1 as a target, downstream proteins, Akt and p70S6K, were checked after PC-3 cells were treated with 1,

5 and 10µM of these two compounds for 6hrs in 1%FBS supplemented medium. Both compounds significantly decreased the phospho-Akt level at 10 and 5µM while the total amount of Akt was maintained (Fig. 4.12B). Similarly, OSU03012 showed the

S6K inhibition of p70 kinase activity in a dose-dependent manner (Fig. 4.12C) with IC50 of

5µM, which was consistent with PDK-1 inhibition and phospho-Akt decrease data.

4.5.2 Cellular Effects of OSU03012 and OSU03013

The dose-dependent and time-course effect of OSU03012 and OSU03013 in PC-3 were shown in Fig. 4.13A. IC50s of growth inhibition in 1%FBS supplemented medium

at 24hrs were 5 and 2µM for OSU03012 and OSU03013 respectively. OSU03013 was

more potent at higher concentrations (10 and 5µM) than OSU03012, which reflected its

higher potency in PDK-1 inhibition. But at lower concentrations (2.5 and 1µM) growth

inhibition effect of OSU03013 dropped so dramatically that no effect was observed at

1µM. We proposed that the guanidino group (-NHC(NH)NH2) in OSU03013, which can be positively charged at physiological PH, would lower the amount of drug getting into cells. Such effect would significantly counteract the higher potency in PDK-1 inhibition and result in no effect on growth inhibition. Same pattern of antiproliferative effect was

55 seen in 10%FBS medium after 6 day treatment (Fig. 4.13C). OSU03012 was quite effective even at 1µM but not OSU03013 possibly for the same reason stated above. As

Akt is an important signaling protein in apoptosis, nucleosome formation and PARP cleavage were checked to see whether cells underwent apoptosis or not (Fig. 4.13B).

Both compounds showed significant amount of nucleosome formation and PARP cleavage at concentrations 5µM or higher.

56 N-Lobe

Glycine-rich Hinge Flapp region

ATP

C-Lobe

Fig. 4.1 Top. Crystal structure of PDK-1 catalytic domain in complex with ATP. Bottom. Enlarged view of van der waals interaction between PDK-1 and adenine moiety in ATP (in orange).

O HN O H NH H N H HO O H H (H2C)4 N O N OH O Ser160 Lys111 H Ser94 H N H N O -O N O O O P O- H N N O P O P N H - H O O OH Ser92 O H H OH O Ala 162 H

O- H N O Glu166 O

Fig. 4.2 H-bonds formed between ATP and PDK-1

Fig. 4.4 A: Global view of OSU02067 docked into ATP binding site. B: H-bond formed between sulfonamide moiety of OSU02067 and Ala162. C: Residues within 6.5Å of OSU02067. 60

F3C Ar O OH F3C Ar NN abNN F3C Ar OH 1,1,1-Trifluoro-4-hydroxy-4- phenanthren-2-yl-but-3-en-2-one N 26 CN 27 H2N e d c

F C Ar F C F3C Ar 3 3 Ar

NN NN NN

N N NH NH H N 2 N 30 29 28 N a. 4-cyanophenylhydrazine hydrochloride/EtOH b. hydroxylamine/Na/MeOH/reflux 2hrs c. NH4Cl/NaN3/Toluene/reflux d.1. DIBAL-H/THF(-40°C) 2. NH2OH.HCl/K2CO3/reflux e.1. DIBAL-H/THF(-40°C) 2. NH2NH2.H2O/K2CO3/reflux

Fig. 4.5 Synthetic route for derivatives 26-30

F C Ar F3C Ar O OH 3 NN bcNN F3C Ar

1,1,1-Trifluoro-4-hydroxy-4- NH2 phenanthren-2-yl-but-3-en-2-one CN 32 31 N OH a d

F3C Ar F3C Ar N N NN

HN N 25 CONH 2 33 N N

a. 4-caramoylphenylhydrazine hydrochloride/EtOH/reflux b. 4-hydrazinophenyl acetonitrile hydrochloride/EtOH/reflux c. NH2OH.HCl/EtOH/reflux d. NaN3/Et3N/Toluene

Fig. 4.6 Synthetic route for derivatives 25 and 31-33

F3C Ar F3C Ar O OH a N N b N N F3C Ar 1,1,1-Trifluoro-4-hydroxy-4- VIII IX phenanthren-2-yl-but-3-en-2-one NO2 NH2

c d e

F3C Ar F3C Ar F3C Ar N N NN NN

NH2 NH2 NH2 HN HN HN O 34 O 35 NH 36

a. 4-nitrophenylhydrazine hydrochloride,EtOH, reflux 2h b. PtO2/H2/EtOH c. 1.t-BOC-Glycine/ DEC/THF 2.Conc.HCl/EtoAc d. NH2CN/HCl/EtOH e. NaNCO/HoAc/EtOH/H2O

Fig. 4.7 Synthetic route for derivatives 34-36

Ar F3C

N N

NHNH2.HCl

NO2 Desired product O OH Kinetic product

NO2 F C Ar Ar 3 F3C

VII N N

NO2

Side product Thermodynamic product

Fig. 4.8 Two isomers formed as intermediate VII is coupled with 4-nitrophenylhydrazine.

IC (µM) IC (µM) No. R 50 No. R 50 PDK-1 PC-3 PDK-1 PC-3

25 -CONH2 12 7 31 -CH2CN 42 25 OH N 26 -CN 45 30 32 15 8 NH2 OH H N N H2 N 27 40 25 33 C 45 27 N NH2 N H O N N NH2 28 52 32 34 N 5 5 H N N (OSU03012) NH OH N 29 25 14 35 N NH2 2 3 H H (OSU03013) O C N 30 16 10 36 40 24 H NH2 N NH2 H

Table 4.1 Structures and potency of derivatives 25–36 for inhibiting recombinant PDK-1 kinase activity and PC-3 cell growth. The general structure of these compounds is shown at the top. R represents the respective substitution in place of the sulfonamide moiety. The reported IC50 values are concentrations at which PDK-1 kinase activity is inhibited by 50% or at which PC-3 cell death measures 50% relative to DMSO control after 24 h- exposure in 1% fetal bovine serum-containing RPMI 1640.

65 A.

Fig. 4.9 A: Global view of OSU03012 docked into ATP binding site. B: H-bond formed between NH2 of OSU03012 and CO of Ser160 and NH of Ala162. C: Residues within 6.5Å of OSU03012. 66 A.

Fig. 4.10 A: Global view of OSU03013 docked into ATP binding site. B: H-bond formed between guanidino moiety of OSU03013 and CO of Ser160 and NH of Ala162. C: Residues within 6.5Å of OSU03013.

OSU03013

OSU03012 OSU02067

Fig. 4.11 Overlay of OSU02067, OSU03012 and OSU03013. Conformations shown here for each compound are extracted from docked structures.

Fig. 4.12 A. Dose-dependent inhibition of recombinant PDK-1 kinase activity by OSU- 03012 and OSU-03013. B. Effect of OSU-03012 (top) and OSU-03013 (bottom) on the phosphorylation status of Akt in drug-treated PC-3 cells. C. Effect of OSU-03012 on the kinase activity of immunoprecipitated p70S6K in drug-treated PC-3 cells.

70 CHAPTER 5

APPLICATION OF OSU03012 TO OTHER CANCERS

Although OSU03013 and OSU03012 are the optimal inhibitors from the 2nd series of analogs, OSU03012 was shown to be effective in nude mice PC-3 xenograft model

(unpublished data). Moreover, OSU03012 was tolerated well by nude mice and no obvious toxicity was observed. Therefore it was chosen for further study.

5.1 In Primary Chronic Lymphocytic Leukemia (CLL) (75)

Studies of CLL showed that LC50 (lethal concentration 50%) of OSU03012 was

5.5µM after 72hr treatment. Mechanistic study of apoptosis induced by this compound indicated that it activated not only mitochondrial pathway but also caspase-independent pathway. It’s also noteworthy that such effect was independent of Bcl-2 expression level, which made it distinct from many other therapeutic agents used in leukemia and other cancers.

5.2 In Breast Cancer Cell MDAMB453 (76)

OSU03012 has also been investigated in breast cancer cell MDAMB453. IC50 (50% growth inhibition) was about 5-7.5µM at 24hrs. Decreased phospho-Akt and phospho-

71 GSK3α/β (downstream of Akt) were observed in drug-treated cells. Apoptosis was demonstrated by cleaved 85kDa PARP fragment and nucleosome formation.

5.3 In Gleevec Resitant Chronic Myelogenous Leukemia (CML) Cells (77)

Although recent FDA approved drug imatinib demonstrated success in clinics for the

treatment of CML, resistance has already been observed in patients due to either

amplification of Bcr-Abl or mutation in Abl catalytic domain. Y253F/H, E255K/V,

T315I, and M351T were characterized as the most clinically relevant mutants. These

mutants exhibited significantly reduced sensitivity to imatinib as compared to the wild

type Bcr-Abl, thereby causing cellular resistance to imatinib. In the recent blood paper

by Tseng, OSU03012 was shown to be equally potent in mutant (Ba/F3p210E255K and

T315I Ba/F3p210 ) and sensitive cells with IC50 of 6µM. Moreover, this compound was

able to sensitize these two mutant cell lines to imatinib. This synergistic action was in

part due to the concerted effect on phospho-Akt (a downstream protein of Bcr-Abl).

72 CHAPTER 6

CONCLUSIONS AND FUTURE DIRECTIONS

6.1 Conclusions

Extensive structure modification of celebrex gave optimized compound OSU03012,

which showed not only improved PDK-1 inhibitory activity but also higher cyto-toxicity

in prostate cancer cell PC-3. In vivo efficacy of this compound was demonstrated in PC-

3 xenograft nude mice model. Other than prostate cancer, OSU03012 has been shown to have similar effects in primary CLL, mutant CML and breast cancer cells. In addition, this compound can sensitize the resistant CML with mutations in Abl catalytic domain to imatinib, which made it clinically very important. Currently, OSU03012 is being studied in a broad spectrum of cancers (thyroid, oral, lung and bladder cancers) and preclinical tests (toxicity, pharmacokinetic and pharmacodynamic studies) are underway.

6.2 Future Directions

It has been pointed out in many review articles that ATP competitors have lower

intracellular concentrations (low to high nanomolar range) so they need to be highly

potent in inhibiting kinases in order to overcome high intracellular ATP concentration (2-

10mM) (35). Comparison of OSU03012 with UCN01 and many other protein kinase 73 inhibitors strengthened the need for further modification of OSU03012 and prompted us

to improve the potency in the following aspects:

1. Domains adjacent to ATP binding site are to be explored in order to increase the

binding affinity and selectivity. As indicated in the recent review article by Noble et. al.

in Science (31), one of the principles for structure-based lead optimization is targeting

less conserved pockets, such as the small hydrophobic pocket behind the adenine binding

site (important for the development of selective p38α inhibitors and EGFR inhibitor erlotinib) (78-81) and the C-terminal lobe close to ribose moiety in ATP (exploited by many CDK selective inhibitors) (39). Fig. 6.1a showed the binding of p38α inhibitor

SB4 to ATP binding domain. Fluorophenyl group of SB4 penetrated into hydrophobic pocket at the back of ATP binding domain guarded by Thr106, which couldn’t be seen in other MAP kinases with residues bearing bulkier side chains. Thus this class of inhibitors has higher potency and selectivity for p38α. Similarly, crystal structure of

EGFR in complex with tarceva indicated that acetyl group in erlotinib was projected into the hydrophobic pocket guarded by the small residue Thr766 (Fig. 6.1b). Accordingly, in PDK-1, Phe157, Gly158 and Leu159 are the three residues present before the hinge region residue Ser160 and define the hydrophobic pocket behind adenine binding site.

Although the side chain of Leu159 seems a little bulkier than that of Thr mentioned above in p38α and EGFR, there is still space for the lipophilic moiety to penetrate into that pocket and avoid clash with the side chain of Leu159. So this hydrophobic pocket can be explored in PDK-1 inhibitor design to improve the binding affinity. Moreover,

C-terminal lobe residues close to ribose of ATP molecule was utilized in CDK inhibitors design. Sulfonamide group of NU6102 formed two H-bonds with Asp86 (Fig. 6.1c) (39). 74 - NH2 donated a H-bond to carboxylic acid (COO ) side chain of Asp86 while O of sulfonamide accepted a H-bond from backbone NH of Asp86. Thus a 1000-fold increase in CDK2 inhibition was seen for NU6102 over parent compound. Similarly, in PDK-1, the 3-OH of ribose in ATP formed a H-bond with residue Glu166. Manipulation of substituents on pyrazole ring may lead to potential H-bonding with Glu166 and van der waals interaction with surrounding residues.

2. Experimental structure of PDK-1 in complex with OSU03012 is definitely necessary to

confirm current computer docking model and guide for future inhibitor design. Current version of docking program (Autodock) only considers flexibility in ligand but not in protein, so binding-induced protein conformation change is negelected here. This might be dangerous if we consider the case where binding of inhibitor to p38α caused peptide bond (in hinge region) to flip. The amide bond between Met109-Gly110 flipped as compound 2 bound to ATP binding motif so that carbonyl group in compound 2 formed two hydrogen bonds with backbone NH of Met109 and Gly110 (Fig. 6.2) (79). For

other MAP kinases (ERK and JNK), Glu and Asp occupied the same position as Gly110,

which made peptide flip energetically unfavorable and lower the binding affinity of this

class of compounds. Such information was gotten only through the crystal structure of

the complex. As far as PDK-1 is concerned, Tyr161 and Ala162 are in the positions of

Met109 and Gly110, respectively. The relatively larger side chains of Tyr and Ala make peptide flip not as easy as that seen in p38α but still possible.

In addition, recent research indicated that inhibitors can differentiate between active

and inactive protein conformations. A good example is Bcr-Abl inhibitors. PD173955

is 10-fold more potent than imatinib against Abl, which is attributed to its ability to bind 75 to both active and inactive conformations of Abl while imatinib requires a specific

inactive conformation (34). Comparison of the co-crystal structures indicates that in the

active conformation of Abl the activation loop is placed in such position that it can

accommodate relatively smaller PD173955 very well but collide with phenyl and

piperazine rings of imatinib. While in the inactive conformation, the activation loop

moves away and accommodates both compounds very well. Such a stringent

requirement of imatinib for protein conformation also results in its inability to inhibit

mutated Bcr-Abl (mutations in catalytic domain), which is seen as resistance in clinics.

Such resistance can be overcome by a Src-Abl dual inhibitor, BMS-354825, which binds

to target protein regardless of conformation of the activation loop (82).

As far as PDK-1 is concerned, it’s been reported that different inhibitors induced different conformation changes (83, 84). Intermediate conformation (between active and inactive) was assigned to primary crystal structure of PDK-1 (1H1W) based on “essential dynamics” analysis (compare structure parameters between PDK-1 and PKA). PDK-1 in complex with staurosporine/UCN-01 (IC50=6nM) adopts the most closed conformation

(active form) while most open conformation is seen in PDK-1 complexed with Bim2

(bisindolyl maleimides, IC50=14µM). The preference for different conformations can be roughly correlated to the potency of corresponding small molecules. Generally the more potent the compound, the more closed the conformation of PDK-1. From a physical chemistry point of view, the more closed conformation gains in enthalpy through desolvation of more hydrophobic moieties of protein and small molecule although it is penalized in entropy by reduced translation and vibration at the same time. The net effect is the lower free energy reflected as higher binding affinity. We still do not know how

76 the information above can be integrated into rational PDK-1 inhibitor design. But

definitely, more co-crystal structures and detailed structure analysis will provide insights

in this regard.

3. Adsorption, distribution, metabolism and excretion (ADME) factors need to be

considered throughout the whole drug design process as we aim at clinical translation of such inhibitors. Studies about rational drug design employed by merck advanced clinical candidates indicate that drugs tend to have higher molecular weights and more H-bonding but unchanged lipophilicity (85). As a result, poorer permeability is observed for such molecules. Currently it is difficult to use computational models to predict ADME as these properties are multi-mechanism systems rather than single-mechanism (like receptor inhibitors) and involve many parameters. In practice, filters and rules that sort out the compounds with most undesirable (or worst) ADME parameters should be satisfactory enough. For example, filters predicting the most poorly and highly soluble compounds are practically very useful.

6.3 Concerns about Anticancer Drug Development

Although this thesis focuses specifically on the development of PDK-1 inhibitors

primarily for androgen-independent prostate cancers, I would like to put it in a more

general field-that is protein kinase inhibitors for cancer therapy. Just as much has been

achieved in this area during the past 20 years or so (chapter 1), problems have arisen and

need to be considered by whoever involved in this field. As a medicinal chemist who is

quite concerned about the beneficial effect for cancer patients, I would like to describe a

77 couple of problems related to clinical translation, which will probably affect future

paradigm of protein kinase inhibitor development including PDK-1 inhibitors.

6.3.1 Specific PKIs versus Non-Specific (or Broad Spectrum) PKIs

The reality that majority of PKIs are ATP competitors confirms the feasibility and

success of targeting ATP binding site although many obstacles are still waiting to be

overcome in the future (78, 39). As computational chemistry and structure biology

advance in a fast pace, more drugs from this category should come out. However,

comparison of the efforts invested with the outcome (available drugs) lead us to ponder

on the following questions. Can the available resource be spent in a more efficient way?

Is it worthwhile to put so much effort on improving selectivity? Can approach other than

selective inhibitors achieve the same beneficial effect on patients?

Certainly these questions can’t be answered unanimously as there is much debate

about these issues. But looking closely at the recent clinical trials, I’d like to say there is a trend to change the current focus on selective inhibitors since broad spectrum inhibitors start to get onto the stage.

Exelixis, Inc. has initiated a Phase 1 clinical trial to evaluate the safety, tolerability

and pharmacokinetic profile of XL647, a novel, orally available, proprietary anticancer

compound that targets multiple receptor tyrosine kinases (RTKs) implicated in tumor

proliferation and vascularization. XL647 is the first of several Spectrum Selective

Kinase Inhibitors (SSKIs) that Exelixis intends to advance into clinical testing. Each

SSKI has a different RTK inhibition spectrum, and each has the potential to achieve efficacy through simultaneous inhibition of multiple RTKs. XL647 simultaneously inhibits EGFR, HER2, VEGFR and EphB4 RTKs with high potency and demonstrates

78 excellent activity in target-specific cellular functional assays (www.exelixis.com). SU-

6668 (Sugen, Inc.) elicits potent pro-apoptotic and antiangiogenic effects in vivo by inhibiting FGFR1, KDR and PDGFRβ simultaneously (35). Even for the recent approved drug imatinib, It has multiple targets including Bcr-Abl, c-Kit and PDGF.

Although no paper has reported that its in vivo effect is mediated through inhibition of these three targets simultaneously, I do doubt that single target inhibition can lead to significant in vivo effect.

More importantly, cancer cells are genetically very heterogeneous. Here is the quote

from a paper “over the past 50 years we learned that cancer is not just a disease of proliferation. The hallmarks of cancer are the evasion of apoptosis, deregulated proliferation, failure to senesce, invasion and metastasis, and ability to induce neo- angiogenesis.” (86). All this behavior can be eventually attributed to genetic changes,

which leads to production of abnormal proteins and subsequent aberrant cellular effects.

Nowadays, we see not only much more resistance to single drug treatment but also

increasing numbers of combined therapies targeting on different genetic abnormalities.

In one sense, SSKIs share similar rationale with combination therapy-that is targeting

multiple molecular abnormalities.

Therefore, as long as the clinical efficacy and safety can be demonstrated for SSKIs, they should have their own share on the market. And screening of inhibitors for a panel

of proteins instead of a single one might prevail in both academics and industry.

79 6.3.2 Targeted Cancer Therapy-Lessons from Gefitinib Clinical Trial.

Gefitinib (EGFR) was approved for lung cancer in 2003 after a prolonged debate at

FDA about the clinical trial results. What makes people so confused about is that

gefitinib only showed response in 10% of patients although it’s demonstrated to be quite

effective in animal models of a variety of cancer. In these 10% patients, however,

dramatic improvement or even complete remission was observed. Scientists were so

eager to solve the puzzle in order to comfort the general public. The answer came out

this summer in a N. Eng. J. Med. article (87). It was found that 8 out of 9 patients

responded to gefitinib had mutations in EGFR gene, which corresponded to the mutated

amino acids lining ATP binding domain. This not only enhanced the kinase activity in

response to EGF but also increased the sensitivity to gefitinb.

Many things can be learned from gefitinib clinical trials as they were pointed out in

sawyers review article. The most important scientific lesson is patient selection. For target-directed therapy, patients should be selected on the basis of predisposing molecular

lesions instead of just tumor histology and origin. At present time, this requirement for

clinical trial may represent some technical challenges due to the difficulties in diagnosis

assay and sample acquirement for solid tumor. As this topic is beyond the scope of

thesis, more details can be found in the review article (17).

80 F

N N

NH2 N H

SB4

(b)

N NH

O O

OCH3 OCH3 Tarceva

N O N NH

SO2NH2 NU6102

Fig. 6.1 (a). Crystal structure of p38 in complex with SB4. Thr106 and fluorophenyl group in inhibitor were highlighted in rectangle (91). (b). Crystal structure of EGFR in complex with Tarceva. Thr766 and acetyl group of Tarceva were highlighted in rectangle (90). (c) Crystal structure of CDK2 in complex with NU6102. Two H-bonds formed between Asp86 and Sulfonamide group of NU6102, which are highlighted in rectangle (39). 81

s t n e s e r p e r

w o l l e Y

. s r o t i b i Green represents compound 2 h n i

h t i w

x e l p m o c

n i

α 8 3 flip at Met109-Gly110 (89). p

f o complex with compound 1.

e r u t c u r t s

l a t s y 1 r d C

n 2 . u 6 o

2 p g

d m original conformation in induced conformation with F

n co

82 CHAPTER 7

EXPERIMENTAL METHODS AND MATERIALS

7.1 Synthesis of the 1st Series of Compounds

7.1.1 Preparation of Starting Material for Synthesis of Compound 1-4

For compound 1-4, the starting material, methyl aromatic ketone (not commercially available), was synthesized using Friedel-Crafts reaction (Fig. 7.1).

CH COCl Br 3 Br O AlCl3/CS2

I II

Fig. 7.1 Synthesisofmethyl aromatic ketone

1-[4-(2-bromoethyl)phenyl]-ethanone (II).

A suspension of aluminum chloride (0.792g) in a mixture of acetyl chloride (0.384mL) and carbon disulfide (15mL), under N2, is cooled in ice bath and treated dropwise with the mixture of 2-bormoethylbenzene (I, 1g) and acetyl chloride. At the end of addition, a

83 brown solution was obtained. The cooling bath is removed and the stirring is continued

at rt for 2hrs. The reaction mixture is poured into a mixture of ice and conc. HCl. The

resulted mixture is extracted into ether and dried over MgSO4. Solvent is removed under reduced pressure and about 1.1 g of intermediate II was yielded (90%).

As the methyl aromatic ketone was synthesized, the following procedures were the same as that shown in section 7.1.3. Compound 3 and 4 were synthesized from corresponding compounds 1 and 2 by the following procedure:

Add 1 eq of compound 1 or 2 to 1.5 eq of 0.5M NaN3 (DMSO) and stir at rt for 2 days. Evaporate the DMSO under reduced pressure and the product was extracted with

CH2Cl2. Re-crystallization gave compound 3 and 4 in high yield (85%).

7.1.2 Preparation of Starting Material for Synthesis of Compounds 10-20

The starting material biphenyl ketone (V) for the synthesis of compound 10-20 was

derived from the Suzuki coupling reaction (an example was shown in Fig. 7.2) (88).

Once the methyl aromatic ketones were gotten, the rest of synthesis was same as

described in section 7.1.3.

OH O O Pd(OAc) , K CO B + Br 2 2 3 OH Bu4NBr, H2O III IV V

Fig. 7.2 Synthesis of biphenyl methyl ketone

84 1-[(1,1'-biphenyl)-4-yl]-ethanone (V)

To a 25 mL flask were added a stir-bar, 1g of 4-bromoacetophenone (IV), 0.636g of

phenylboronic acid (III), 2.2mg of Pd(OAc)2, 1.74g of powdered K2CO3, and 1.62g of

Bu4NBr. The flask was flushed with argon and equipped with a rubber septum. Water (7

mL) was added with a syringe and the resulting suspension was energetically stirred. The

mixture was stirred and heated for 1 h at 70 °C under argon. It was then cooled to room

temperature, diluted with water, and extracted with EtOAc. The solution was dried with

Na2SO4 and concentrated to yield a white solid. Chromatography on silica gel afforded

0.88g of title compound V (90%).

7.1.3 General Procedure

Compounds 1-24 were synthesized according to a two-step general procedure

described in Fig. 7.3, in which Ar represents the respective aromatic ring structures.

H N 2 NH .HCl

F3C Ar N N O OH O CF3COOEt NaH/THF SO2NH2 Ar F3C Ar CH3CH2OH SO NH VI VII 2 2

Fig. 7.3 General synthetic procedure for compound 1-24

85 Synthesis of compounds 5-9 and 21-23 followed the exact same procedures via

commercially available starting material (methyl aromatic ketone, VI) and respective

intermediates with different aromatic ring structures. Compound 23 (OSU-02067) is

used here as an example to illustrate the synthesis of this series of compounds (Fig. 7.4).

H N 2 NH .HCl

F3C O OH O CF3COOEt NN NaH/THF SO2NH2 F3C CH3CH2OH Step 1 Step 2 SO NH 1,1,1-Trifluoro-4-hydroxy-4- 2 2 2-acetylphenanthrene phenanthren-2-yl-but-3-en-2-one OSU02067

Fig. 7.4 Synthesis of OSU02067

1,1,1-Trifluoro-4-hydroxy-4-phenanthren-2-yl-but-3-en-2-one (step 1). To a suspension of sodium hydride (NaH; 0.13 g, 5.4 mmol) in 5 ml of anhydrous tetrahydrofuran (THF) was added ethyl trifluoroacetate (CF3COOEt; 0.64 g, 4.5 mmol) under argon. After stirring at 25 oC for 10 min, 2-acetylphenanthrene (1 g, 4.5 mmol) in

5 ml of THF was added dropwise to the solution. The mixture became clear and orange

hued within 30 min, and after stirring for an additional 2 h, was concentrated under

vacuum. The residue was suspended in water, and extracted with ethyl acetate (15 ml)

twice. The organic phase was separated, dried over sodium sulfate, and concentrated to

dryness under vacuum to give the product (yellow solid; 1.29 g, 90% yield). The product

was used directly without purification.

86 OSU-02067 (step 2). 4-Hydrazinobenzene-1-sulfonamide hydrochloride (1.1 g, 4.9

mmol) was added to a stirred solution of 1,1,1-trifluoro-4-hydroxy-4-phenanthren-2-yl-

but-3-en-2-one (1.29 g, 4.1 mmol) in 40 ml of ethanol. The mixture was refluxed for 12

h, cooled to room temperature, and concentrated to dryness under vacuum. The residue

was dissolved in ethyl acetate, and washed with water. The organic layer was dried over

sodium sulfate, and concentrated under vacuum. The crude product was purified by

silica gel flash chromatography to yield OSU-02067 (1.52 g, 80% yield).

7.2 Synthesis of the 2nd Series of Compounds

Compounds 25–36 were synthesized using 1,1,1-trifluoro-4-hydroxy-4-phenanthren-

2-yl-but-3-en-2-one, product of the aforementioned step 1, as a common precursor (Fig.

7.5).

4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenecarboxamide (25) (step 3). (4-Carbamoylphenyl)-hydrazine hydrochloride

(0.92g, 4.9mmol) was added to a stirred solution of 1,1,1-trifluoro-4-hydroxy-4- phenanthren-2-yl-but-3-en-2-one (1.29 g, 4.1 mmol) in 40 ml of ethanol at 25 oC. The mixture was refluxed for 12 h, cooled to room temperature and concentrated to dryness under vacuum. The residue was dissolved in ethyl acetate, and washed with water. The organic layer was dried over sodium sulfate, and concentrated under vacuum. The crude product was purified by silica gel flash chromatography (ethyl acetate-hexane, 1:1), yielding 25 (1 g, 60% yield).

87 4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-benzonitrile (26)

(step 4). To a stirred solution of 1,1-trifluoro-4-hydroxy-4-phenanthren-2-yl-but-3-en-2- one (2.45 g, 7.7 mmol) in 60 ml of ethanol was added 4-cyanophenylhydrazine hydrochloride (2.53g, 15 mmol) at 25 oC. The mixture was stirred under reflux for 12 h, cooled to room temperature and concentrated to dryness under vacuum. The residue was dissolved in methylene chloride, and washed with water. The organic layer was dried over sodium sulfate, and concentrated under vacuum. The crude product was purified by silica gel flash chromatography (ethyl acetate-hexane, 1:4) to afford 26 (2.7g, 85% yield).

4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-N-hydroxy- benzamidine (27) (step 5). Hydroxylamine hydrochloride (25 mg, 0.36 mmol) was added to a suspension of Na metal (8.3 mg, 0.36 mmol) in methanol (3 ml). The mixture was stirred at room temperature for 10 min, and compound 26 (124 mg, 0.3 mmol) was added. The mixture was refluxed for 2 hr, then stirred at 25 oC for additional 16 h, and concentrated under vacuum. The residue was purified by silica gel flash chromatography

(ethyl acetate-hexane, 1:4 to 1:1) to give 27 (120 mg, 76% yield).

5-(2-Phenanthrenyl)-3-(trifluoromethyl)-4-(1H-tetrazol-5-ylphenyl)-1H- pyrazole (28) (step 6). A mixture containing compound 26 (125 mg, 0.3 mmol), NH4Cl

(123.7 mg), and NaN3 (58.5 mg, 0.9 mmol) in 5 ml of toluene was stirred under reflux for

12 h, cooled to room temperature, 5 ml of 10% HCl was added, and extracted with 20 ml

of methylene chloride twice. The organic phase was dried over sodium sulfate, and

88 concentrated to dryness under vacuum. The crude product was purified by silica gel flash

chromatography (ethyl acetate-hexane, 1:4) to give 28 (96 mg, 70% yield).

4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-benzaldehyde oxime (29) (step 7). DIBAL-H (3.1 ml, 3.1 mmol, 1.0 M in hexane) was added dropwise to a solution of compound 26 (0.417 g, 1.1 mmol) in 5 ml of THF at –40 oC. The mixture was stirred for 8 h, poured into 5 ml of 10% acetic acid, and stirred for 30 min. The organic layer was dried over sodium sulfate, and concentrated to dryness under vacuum.

The crude product was purified by silica gel flash chromatography (ethyl acetate-hexane,

1:4) to give an aldehyde intermediate (141 mg, 0.34 mmol) that was immediately added

to a solution containing hydroxylamine hydrochloride (211 mg) and K2CO3 in 5 ml of

ethanol. The mixture was stirred under reflux for 16 h. After removal of solvent, the

residue was extracted with CH2Cl2 and washed with water. The organic layer was dried

over magnesium sulfate, and purified by silica gel chromatography eluting with ethyl

acetate/hexane (1/2) to give 29 (116mg, 83%).

4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-benzaldehyde hydrazone (30) (step 8). Compound 30 (124 mg, 85 % yield) was synthesized in the

same manner as 29 except that hydrazine monohydrate (153 mg, 3.1 mmol) was used

instead of hydroxylamine hydrochloride.

{4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl}-

acetonitrile (31) (step 9). (a) Preparation of (4-Hydrazinophenyl)acetonitrile

89 hydrochloride. A solution of sodium nitrite (3.15 g, 45.7 mmol) in water (20 ml) was

added dropwise to a cooled (-15 oC), stirred suspension of 4-aminobenzonitrile (5 g, 42.3 mmol) in a concentrated hydrogen chloride solution (55 ml) at such a rate as to maintain a temperature below –10 0C. After addition was finished, the reaction mixture was quickly filtered to remove solids. The filtrate was added in portions to a cooled (–20 0C) and stirred solution of SnCl2•2H20 (47.7 g, 0.21 mol) in a concentrated hydrogen chloride

solution (37 ml) at such a rate as to keep the temperature below -10 0C. After stirring the

solution for additional 15 min, the solid was collected, washed with diethyl ether (4 x 25

ml), and dried to give (4-hydrazinophenyl)acetonitrile hydrochloride (5.6 g, 78%). (b)

Compound 31. A mixture of (4-hydrazinophenyl)acetonitrile hydrochloride (0.32 g, 1

mmol) and 1,1,1-trifluoro-4-hydroxy-4-phenanthren-2-yl-but-3-en-2-one (0.18g, 1.1

mmol) in ethanol (20 ml) was stirred under reflux for 24 h, cooled to room temperature,

concentrated to dryness under vacuum, and dissolved in ethyl acetate. The organic layer

was dried over magnesium sulfate, and concentrated to dryness under vacuum. The

crude product was purified by silica gel column chromatography (hexane-ethyl acetate,

2:1) to give compound 31 (0.35g, 81% yield).

2-{4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl}-N- hydroxy-acetamidine (32) (step 10). A solution of compound 31 (0.43 g, 1 mmol) and hydroxylamine hydrochloride (0.075 g, 1.1 mmol) in ethanol (10 ml) was stirred under reflux for 8 h, concentrated to dryness under vacuum. The residue was dissolved in water, brought to pH 8-9 by addition of saturated NaHCO3 solution, and extracted with ethyl acetate. The organic layer was dried over magnesium sulfate, and concentrated to dryness

90 under vacuum. The crude product was recrystallized in diethyl ether-hexane to give

compound 32 (0.32 g, 71% yield).

5-(2-Phenanthrenyl)-3-(trifluoromethyl)-4-(1H-tetrazol-5-ylmethylphenyl)-1H- pyrazole (33) (step 11). A mixture containing compound 31 (0.43 g, 1 mmol), sodium azide (0.08 g, 1.2 mmol), and triethylamine hydrochloride (0.12 g, 1.2 mmol) in toluene

(5 ml) was stirred at 100 °C for 5 h, cooled to room temperature, and extracted with water

(10 ml). To the aqueous phase was added dropwise a 36% hydrogen chloride solution to salt out the resulting tetrazole 33. After filtration, the solid was dried under vacuum,

yielding compound 33 (0.39 g. 84% yield).

1-(4-Nitrophenyl)-5-phenyl-3-(trifluoromethyl)-1H-pyrazole (VIII) (step 12).

To a solution of 1,1,1-trifluoro-4-hydroxy-4-phenanthren-2-yl-but-3-en-2-one (1.29 g, 4.1 mmol) in 40ml of ethanol was added 4-nitrophenylhydrazine hydrochloride (0.93 g, 4.9 mmol) under stirring, refluxed for 1h, cooled to room temperature, and concentrated to dryness under vacuum. The residue was dissolved in ethyl acetate, and washed with water. The organic phase was dried over magnesium sulfate, and concentrated to dryness under vacuum. The crude product was purified by silica gel column chromatography to afford compound VIII (0.88 g, 50% yield).

4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]phenylamine (IX)

(step 13). To a solution of compound VIII (0.88 g, 2 mmol) in 20 ml ethanol was added platinum oxide(27 mg, 0.12 mmol), stirred under H2 for 12 h, filtered to remove the

91 catalyst, and concentrated to dryness under vacuum. The crude product was purified by

silica gel column chromatography to yield compound IX (0.57 g, 70% yield).

2-Amino-N-{4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-

phenyl}-acetamide (34) (steps 14 and 15). To a solution of t-butyloxycarbonyl (tBOC)-

glycine (0.25 g, 1.4 mmol) and compound IX (0.57 g, 1.4 mmol) in 10 ml of

tetrahydrofuran was added 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide

hydrochloride (0.41 g, 2.1 mmol), stirred at 25 oC for 12 h, and concentrated to dryness

under vacuum in a rotary evaporator. The residue was suspended in water, and the product was extracted with ethyl acetate. The organic phase was dried over magnesium sulfate, and concentrated to dryness under vacuum to give t-Boc protected compound 34

(0.67 g, 85 % yield). This intermediate (0.67 g, 1.2 mmol) was dissolved in 8 ml of ethyl

acetate containing 0.7 ml of concentrated HCl solution, stirred at room temperature for 2

h, and concentrated to dryness under vacuum. The crude product was purified by silica

gel column chromatography to yield compound 34 as a white powder (0.49 g, 90 %)

4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl-

guanidine (35) (step 16). To a solution of compound IX (0.57g, 1.4mmol) in 7 ml of

ethanol was added cyanamide (89 mg, 2.1 mmol) and 1.5 ml of 1N HCl. The mixture

was refluxed for 24 h, and concentrated to dryness under vacuum. The product was

purified by silica gel column chromatography to give compound 35 as white solid (0.25 g,

40% yield)

92 4-[5-(2-Phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl-urea (36)

(step 17). Into a 250 ml round bottom flask containing acetic acid (50 ml), water (12 ml),

and ethanol (20 ml) was added compound IX (2.25 g, 5.6 mmol), followed by sodium

isocyanate (0.74 g, 11.2 mmol). The reaction was stirred for 1.5 h, and then neutralized

with the addition of 1 N sodium hydroxide solution followed by sodium hydroxide pellets

until the pH had changed to 7.0. The product was extracted with 200 ml of ethyl acetate.

The organic phase was separated and then washed with 100 ml of water, dried with

magnesium sulfate, and then solvent was removed to obtain the crude product.

Purification was performed by silica gel chromatography with (hexane-ethyl acetate, 3:2

to hexane-acetone, 1:3) to afford compound 36.

7.3 Nomenclatures, 1H NMR (proton nuclear magnetic resonance), and HRMS (high

resolution mass spectrometry) characterizations of compounds 1 – 36 (Table 7.1).

7.4 PDK-1 Kinase Assay

This in vitro assay was performed using a PDK-1 kinase assay kit (#17-280, Upstate,

Lake Placid, NY) according to the vendor’s instructions. This cell-free assay is based on

the ability of recombinant PDK-1, in the presence of DMSO vehicle or the test agent, to

activate its downstream kinase serum- and glucocorticoid-regulated kinase (SGK) which,

in turn, phosphorylates the Akt/SGK-specific peptide substrate RPRAATF with [γ-32P]-

ATP. The [32P]-phosphorylated peptide substrate was then separated from the residual

[γ-32P]-ATP using P81 phosphocellulose paper and quantitated by a scintillation counter

after three washes with 0.75% phosphoric acid. The reported values represent the means 93 of three independent determinations with duplicates for each concentration in each

determination. The IC50 values (concentration to give 50% enzyme inhibition) were

determined by median-effect plot using software calcusyn (www.biosoft.com) (90).

7.5 Cell Viability Assay

7.5.1 Cell Culture

PC-3 cells were obtained from American Type Culture Collection (ATCC, Manassas,

VA). PC-3 cells were cultured at 37ºC in 10% fetal bovine serum supplemented

RPMI1640 medium.

7.5.2 MTT Assay

PC-3 cells were seeded into 96 well plates at 5000 cells/well in 10% FBS-containing

RPMI1640. After 24 hours, cells were treated with compounds in 1%FBS-containing

RPMI1640 medium at different concentrations for different time period. The drug solution was replaced by 200µL/well MTT solution (0.5mg/mL, 3-(4,5-dimethylthiazol-

2-yl)-2,5-diphenyl-2H-tetrazolium bromide in 10%FBS RPMI medium) and cells were incubated at 37ºC for 2 hours. MTT solution was aspirated, 200µL DMSO was used to dissolve the reduced dye 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide and plates were read at 570nM. The cell viability was calculated as follows

CV=(ODTreatment-ODBackground)/(ODDMSO-ODBackground)×100%. For each concentration, six replicates were used. The IC50 values (concentration to give 50% growth inhibition) were determined by median-effect plot using software calcusyn (www.biosoft.com) (90).

94 7.6 Immunoblotting

The general procedure for the western blot analysis of Akt, phospho-Akt and PARP is

described as follows. Cells (floating and attached) were collected by scraping and

centrifuged at 1500 rpm for 10 min. Then cells were washed with 1mL PBS and

centrifuged again at 13200rpm for 2min. 100µL of lysis buffer was added to cell pellets

and it was left on ice for 30min before it was centrifuged at 13200 rpm for 20min. 2µL

of supernatant was used to measure the protein concentration using bio-rad reagent while

the rest was denatured in equal volume of 2×sample buffer. Then it was boiled for 5 min

and were resolved in 10% SDS-polyacrylamide gels on a minigel apparatus, and

transferred to a nitrocellulose membrane using a semi-dry transfer cell. The transblotted

membrane was washed 15min three times with TBS containing 0.1% Tween 20 (TBST).

After blocking with TBST containing 5% nonfat milk for 60 min, the membrane was

incubated with the primary antibody at 1:1,000 dilution in TBST-5% low fat milk at 4 oC for 12 h, and was then washed 15 min three times with TBST (0.1%Tween 20). The membrane was probed with goat anti-rabbit IgG-HRP conjugates (1:1,000) for 1 h at room temperature, and was washed with TBST three times. The immunoblots were visualized by enhanced chemiluminescence. Here is the detailed information for harvesting samples. For PARP, 4×105 were seeded into T-25 flask and after 24 hrs cells were treated with 10, 5, 1µM of OSU compounds in serum free or 1%FBS supplemented

FBS medium for 8hrs (for 10µM, treatment was 4 hrs. Since lots of cells died at 10µM, 3

T-25 flasks of cells were used in order to get enough protein for assay.). As for Akt and

p-Akt, same condition was used except that the treatment was 6hr for all concentration.

Experiments were carried out twice on different occasions. 95 7.7 Immunoprecipitated Akt Kinase Assay

Akt immunoprecipitation was carried out according to a modified, published

procedure (JBC paper). PC-3 cells were treated with DMSO vehicle or the test agents at

the indicated concentrations for 2 h, and then lysed at 4 oC for 1 h in buffer A containing

50 mM Tris-HCl, pH 7.5, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 50 mM sodium fluoride, 10 mM sodium β-glycerophosphate, 0.1% 2-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, and 1 µg/ml each of aprotinin, pepstatin, and leupeptin.

Cell lysates were centrifuged at 10,000g for 5 min, and the supernatant was treated with

Akt antibody at 4 oC for 60 min, followed by protein G-agarose beads for additional 60 min. The immunoprecipitate was used to analyze Akt kinase activity by using the

Akt/SGK-specific peptide substrate RPRAATF as described in PDK-1 kinase assay.

Values represent the means of three independent determinations with duplicates for each concentration in each determination.

7.8 Immunoprecipitated p70S6K Kinase Assay

Immunoprecipitation of p70 S6 kinase (p70S6K) was carried out according to a

modification of a published procedure (89). In brief, 1.5 x 106 PC-3 cells were cultured in T-75 flasks and treated with OSU-03012 at the indicated concentrations in 1% FBS- containing RPMI 1640 medium for 6 h. Both floating and adherent cells were collected and lysed in 1 ml lysis buffer [50 mM Tris (pH 7.4), 150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 1 mM EGTA, and 10% protease inhibitor mixture (Calbiochem,

San Diego, CA)] for 30 min on ice. Lysates were centrifuged at 10,000 x g at 4 oC for 20 min. The supernatants were immunoprecipitated with anti-p70S6K antibody (sc-8418; 96 Santa Cruz Biotechnology, Santa Cruz, CA) on ice for 1 h, which was followed by

incubation with Protein A Sepharose beads for 2 h. The immunocomplex was washed

twice with lysis buffer and once with assay buffer [20 mM 4-morpholinepropanesulfonic acid (pH 7.2), 25 mM β-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, and 1 mM DTT], and then resuspended in assay buffer. The assay was carried out using a p70S6K assay kit (17-136, Upstate, Lake Placid, NY) according to the manufacturer’s instruction. Duplicates were used for each concentration and experiments were carried

out in three different occasions.

7.9 Cell Death Detection ELISA

In brief, 4 x 105 PC-3 cells were cultured in 10%FBS supplemented RPMI1640 in T-

25 flask for 24 h before treatment. Cells were treated with DMSO vehicle or test agents

at the indicated concentrations for 6–24h. Both floating and attached cells (trypsinization) were collected and centrifuged. Cell pellets were re-suspended in 1mL medium and numerated (12-25µM). 105 cells were used for each sample and lysed in 0.5mL incubation buffer. Cell lysates equivalent to 2 x 103 PC-3 cells were used for each well in

ELISA assay, which was carried out by using cell death detection ELISA kit from Roche

Diagnostics (1-544-675, Mannheim, Germany). The assay was carried out by exactly following manufacturer’s procedure except for the last step-incubation. The final data came from 50 min incubation with the substrate instead of 20-30 min suggested by the supplier’s protocol. The reason is because even after such long time incubation, the O.D. for the DMSO control still kept very low. So in one sense the relative O.D. is more

97 meaningful than the absolute values. Experiments were carried out in two different

occasions.

7.10 Molecular Modeling

AutoDock 3.05 predicts the bound conformations of a small, flexible ligand to a

nonflexible macromolecular target of known structure (73). This software is an

automated docking package that combines a rapid grid-based method for energy

evaluation with a Lamarckian genetic algorithm (LGA) method of conformation search.

The crystal structure of PDK-1 in complex with ATP was obtained from the

Brookhaven Protein Data Bank (entry code 1H1W). In this crystal structure, residues

233-242 were missing. Although this substructure was a kind of far away from ATP

binding domain, I added them all to PDK-1 structure in SYBYL 6.9 (Tripos Associate; St.

Louis, MO, 2002). The resulted structure was optimized by energy minimization in

subset 233-242. The final structure was then read into AutoDock Tools (ADT, available

at http://www.scripps.edu/pub/olson-web/doc/autodock/tools.html) for docking setup.

Firstly the protein was subject to the deletion of heteroatoms (H2O, ATP, glycerol and

sulfate) and addition of polar hydrogens. Then as it was selected for grid setup, it was

shown automatically to have non-integral charges probably due to missing atoms in the

input structure, which could be corrected by using “repair missing atoms” function in

ADT. After that, it was added with kollman charge and atomic solvation parameters.

Grid box was set to be centered at (42.3456, 19.1569, 5.0) with 60 grid points on each

axis.

98 The 3D structures of small molecules were generated using SYBYL 6.9. Gasteiger charges were computed and energy minimization was carried out with default parameters.

As small molecules were read into ADT, system automatically detected and merged non- polar hydrogens and checked for aromatic carbons. Then root atom and rotable bonds could be assigned automatically or manually. At the end, pdbq file was generated for each small molecule.

All the gpf and dpf input files were set by using default parameters. Autodock 3.05

(http://www.scripps.edu/pub/olson-web/people/gmm/autodock/obtaining.html) was used

to generate glg and dlg output files. As sybyl6.9 is a very good graph interface, dlg file

was converted to dlg.pdb file in order to be read into sybyl6.9. All molecular modeling

calculations and manipulations were performed on Silicon Graphics O2 (Silicon Graphics

Inc.; Mountain View, CA).

series of compounds nd Synthesis of 2 Fig. 7.5

100 Compd Description 4-[5-(4-(2-bromoethyl)phenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenesulfonamide 1H-NMR δ 3.16 (t, J = 6.4, 2.0Hz, 2H), 3.60 (t, J = 6.4, 2.0 Hz, 2H), 4.90 (s, 1 2H), 6.75 (s, 1H), 7.13 (d, J = 8.0Hz, 2H), 7.20 (d, J = 8.0 Hz, 2H), 7.47 (d, J = 8.5Hz, 2H), 7.91 (d, J = 8.5Hz, 2H) + C18H15BrF3N3O2S; HRMS (M + Na ): theoretical mass, 495.9913; actual mass, 495.9943 4-[5-(4-(3-bromopropyl)phenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenesulfonamide 1H-NMR δ 2.16 (m, 2H), 2.81 (t, J = 7.1 Hz, 2H), 3.41 (t, J = 6.4 Hz, 2H), 2 5.08 (s, 2H), 6.76 (s, 1H), 7.15 (d, J = 8.2 Hz, 2H), 7.25 (d, J = 8.2 Hz, 2H), 7.47 (d, J = 8.5 Hz, 2H), 7.90 (d, J = 8.5 Hz, 2H) + C19H17BrF3N3O2S; HRMS (M + Na ): theoretical mass, 510.0069; actual mass, 510.0042 4-[5-(4-(2-azidoethyl)phenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenesulfonamide 1H-NMR δ 2.90 (t, J = 6.8 Hz, 2H), 3.51 (t, J = 6.8 Hz, 2H), 5.49 (s, 2H), 6.76 3 (s, 1H), 7.17 (d, J = 8.3 Hz, 2H), 7.24 (d, J = 8.3 Hz, 2H), 7.42 (d, J = 8.7 Hz, 2H), 7.85 (dd, J = 8.7, 2.0 Hz, 2H) + C18H15F3N6O2S; HRMS (M + Na ): theoretical mass, 459.0821; actual mass, 459.0817 4-[5-(4-(3-azidopropyl)phenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenesulfonamide 1H-NMR δ 1.83 (m, 2H), 2.64 (t, J = 7.5 Hz, 2H), 3.20 (t, J = 7.5 Hz, 2H), 4 5.31 (br s, 2H), 6.67 (s, 1H), 7.07 (m, 4H), 7.35 (dd, J = 7.5, 2.0 Hz, 2H), 7.79 (dd, J = 7.5, 2.0 Hz, 2H) + C19H17F3N6O2S; HRMS (M + Na ): theoretical mass, 473.0978; actual mass, 473.0946 4-[5-(4-butylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenesulfonamide 1H-NMR δ 0.93 (t, J = 7.2 Hz, 3H), 1.36 (m, 2H), 1.64 (m, 2H), 2.63 (t, J = 5 7.6 Hz, 2H), 5.54 (s, 2H), 6.76 (s, 1H), 7.15(d, J = 8.3 Hz, 2H), 7.20 (d, J = 8.3 Hz, 2H), 7.45 (dt, J =8.8, 2.0 Hz, 2H), 7.88 (dt, J = 8.8, 2.0 Hz, 2H) + C20H20F3N3O2S; HRMS (M + Na ): theoretical mass, 446.1120; actual mass, 446.1149

Continued

Table 7.1 Nomenclatures, NMR and HRMS data for compounds 1-36

101 Table 7.1 continued

4-[5-(4-t-butylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenesulfonamide 1H-NMR δ 1.33 (s, 9H), 4.90 (s, 2H), 6.53 (s, 1H), 7.32 (dd, J = 9.7 Hz, 4H), 6 7.42 (d, J = 8.8 Hz, 2H), 8.02 (d, J = 8.8 Hz, 2H) + C20H20F3N3O2S; HRMS (M + Na ): theoretical mass, 446.1120; actual mass, 446.1118 4-[5-(2-naphthalenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenesulfonamide 1H-NMR δ 5.47 (s, 2H), 6.89 (s, 1H), 7.18 (dd, J = 8.6, 1.6 Hz, 1H), 7.42 (bd, J 7 = 8.6 Hz, 2H), 7.51-7.55 (m, 2H), 7.78-7.83 (m, 6H) + C20H14F3N3O2S; HRMS (M + Na ): theoretical mass, 440.0651; actual mass, 440.0657 4-[5-(3-indolyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide 1 H-NMR δ(acetone-d6) 6.69 (br s, 1H), 7.03-7.08 (m, 2H), 7.19 (t, J = 7.2 Hz, 1H), 7.40 (d, J = 7.8 Hz, 1H), 7.50 (d, J = 7.8 Hz, 1H), 7.67 (d, J = 8.7 Hz, 2H), 8 7.92 (d, J = 8.7 Hz, 2H) + C18H13F3N4O2S; HRMS (M + Na ): theoretical mass, 429.0603; actual mass, 429.0606 4-[5-(4-biphenylyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide 1H-NMR δ 4.81(s, 2H), 6.75 (s, 1H), 7.23 (d, J = 8.5 Hz, 2H), 7.34-7.56 (m, 9 5H), 7.56 (m, 4H), 7.86 (d, J = 8.5 Hz, 2H) + C22H16F3N3O2S; HRMS (M + Na ): theoretical mass, 466.0807; actual mass, 466.0811 4-[5-(4’-chloro[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 5.42 (s, 2H), 6.83 (s, 1H), 7.30 (d, J = 8.2 Hz, 2H), 7.40-7.59 (m, 10 8H), 7.92 (d, J = 8.2 Hz, 2H) + C22H15ClF3N3O2S; HRMS (M + Na ): theoretical mass, 500.0418; actual mass, 500.0432 4-[5-(3’,5’-dichloro[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 4.85 (s, 2H), 6.82 (s, 1H), 7.30 (d, J = 8.8 Hz, 2H), 7.36 (s, 1H), 11 7.37-7.57 (m, 6H), 7.93 (d, J = 8.8 Hz, 2H) + C22H14Cl2F3N3O2S; HRMS (M + Na ): theoretical mass, 534.0028; actual mass, 534.0016

Continued

102 Table 7.1 continued

4-[5-(2’,3’-dichloro[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 4.85 (s, 2H), 6.76 (s, 1H), 7.18-7.25 (m, 3H), 7.35-7.49 (m, 6H), 12 7.88 (d, J = 8.6 Hz, 2H) + C22H14Cl2F3N3O2S; HRMS (M + Na ): theoretical mass, 534.0028; actual mass, 533.9999 4-[5-(2’,4’,5’-trichloro[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 4.86 (s, 2H), 6.77 (s, 1H), 7.25 (dt, J = 8.6, 2.0 Hz, 2H), 7.37 (dt, J = 13 8.6, 2.0 Hz, 2H), 7.39 (s, 1H), 7.46 (dt, J = 8.8, 2.0 Hz, 2H), 7.54 (s, 1H), 7.88 (dt, J = 8.9, 1.2 Hz, 2H) + C22H13Cl3F3N3O2S; HRMS (M + Na ): theoretical mass, 567.9638; actual mass, 567.9679 4-[5-(4’-methyl[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 2.32 (s, 3H), 4.75 (s, 2H), 6.72 (s, 1H), 7.18-7.21 (m, 4H), 7.39-7.52 14 (m, 6H), 7.84 (d, J = 8.9 Hz, 2H) + C23H18F3N3O2S; HRMS (M + Na ): theoretical mass, 480.0964; actual mass, 480.0961 4-[5-(4’-trifluoromethyl[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 5.19 (s, 2H), 6.86 (s, 1H), 7.36 (d, J = 8.0 Hz, 2H), 7.53 (d, J = 8.5 15 Hz, 2H), 7.65 (m, 6H), 7.92 (d, J = 8.5 Hz, 2H) + C23H15F6N3O2S; HRMS (M + Na ): theoretical mass, 534.0681; actual mass, 534.0677 4-[5-(4’-bromomethyl[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 3.92 (s, 2H), 4.93 (s, 2H), 6.66 (s, 1H), 7.03-7.26 (m, 8H), 7.38 (d, J 16 = 8.6 Hz, 2H), 7.82 (d, J = 8.6 Hz, 2H) + C23H17BrF3N3O2S; HRMS (M + Na ): theoretical mass, 558.0069; actual mass, 558.0112 4-[5-(3’,5’-dimethyl[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 2.40 (s, 6H), 5.38 (br s, 2H), 6.83 (s, 1H), 7.05 (s, 1H), 7.25 (m, 17 4H), 7.50 (dd, J = 6.7, 1.7 Hz, 2H), 7.59 (dd, J = 6.7, 1.7 Hz, 2H), 7.92 (dd, J = 6.7, 1.7 Hz, 2H) + C24H20F3N3O2S; HRMS (M + Na ): theoretical mass, 494.1120; actual mass, 494.1119

Continued

103 Table 7.1 continued

4-[5-(4’-butyl[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 0.96 (t, J = 7.5 Hz, 3H), 1.41 (m, 2H), 1.66 (m, 2H), 2.68 (t, J = 7.5 18 Hz, 2H), 5.20 (br s, 2H), 6.84 (s, 1H), 7.29 (dd, J = 8.2, 2.0 Hz, 4H), 7.53 (dt, J = 8.2, 2.0 Hz, 4H), 7.62 (d, J = 8.5 Hz, 2H), 7.93 (d, J = 8.5 Hz, 2H) + C26H24F3N3O2S; HRMS (M + Na ): theoretical mass, 522.1433; actual mass, 522.1466 4-[5-(4’-tert-butyl[1,1’-biphenyl]-4-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 1.35 (s, 9H), 4.87 (s, 2H), 6.59 (s, 1H), 7.44-7.57 (m, 6H), 7.58 (d, J 19 = 7.5 Hz, 2H), 7.92 (d, J = 8.7 Hz, 2H), 8.12 (d, J = 7.5 Hz, 2H) + C26H24F3N3O2S; HRMS (M + Na ): theoretical mass, 522.1433; actual mass, 522.1401 4-[5-(4-(phenylmethyl)phenyl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 3.71 (s, 2H), 4.74 (s, 2H), 6.52 (s, 1H), 6.91-7.11 (m, 9H), 7.27 (d, J 20 = 8.9 Hz, 2H), 7.69 (d, J = 8.9 Hz, 2H) + C23H18F3N3O2S; HRMS (M + Na ): theoretical mass, 480.0964; actual mass, 480.0938 4-[5-(9H-fluoren-2-yl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 3.88 (s, 2H), 4.64 (s, 2H), 6.68 (s, 1H), 7.26-7.38 (m, 4H), 7.56 (d, J 21 = 8.7 Hz, 2H), 7.74-7.81 (m, 3H), 7.90 (d, J = 8.7 Hz, 2H) + C23H16F3N3O2S; HRMS (M + Na ): theoretical mass, 478.0807; actual mass, 478.0771 4-[5-(9-anthracenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide 1H-NMR δ 4.63 (s, 2H), 6.93 (s, 1H), 7.33 (d, J = 6.8 Hz, 2H), 7.45-7.55 (m, 22 8H), 8.04 (d, J = 6.8 Hz, 2H), 8.60 (s, 1H) + C24H16F3N3O2S; HRMS (M + Na ): theoretical mass, 490.0807; actual mass, 490.0769 4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ (600 MHz) 4.89 (s, 2H), 6.92 (s, 1H), 7.37 (dd, J = 8.5, 1.4 Hz, 1H), 23 7.51 (d, J = 8.6 Hz, 2H), 7.65-7.69 (m, 3H), 7.80 (d, J = 8.8 Hz, 1H), 7.86-7.92 (m, 4H), 8.64 (d, J = 8.4 Hz, 2H) + C24H16F3N3O2S; HRMS (M + Na ): theoretical mass, 490.0807; actual mass, 490.0805

Continued

104 Table 7.1 continued

4-[5-(9-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1- yl]benzenesulfonamide 1H-NMR δ 4.76 (s, 2H), 6.90 (s, 1H), 7.43-7.84 (m, 11H), 8.72 (t, J = 7.8 Hz, 24 2H) + C24H16F3N3O2S; HRMS (M + Na ): theoretical mass, 490.0807; actual mass, 490.0833 4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl] benzenecarboxamide 1H-NMR δ 5.75-6.05 (br d, 2H), 7.0 (s, 1H), 7.50 (dd, J = 8.5, 1.4 Hz, 1H), 7.55 25 (d, J = 8.5 Hz, 2H), 7.77 (m, 3H), 7.88 (m, 3H), 7.90 (m, 2H), 8.72 (m, 2H) + C25H16F3N3O; HRMS (M + Na ): theoretical mass, 454.0038; actual mass, 454.1142 4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-benzonitrile 1H-NMR δ 6.91(s, 1H), 7.46 (s, 1H), 7.50 (d, J = 2.0 Hz, 2H), 7.63-7.79 (m, 26 5H), 7.83 (d, J = 2.0 Hz, 2H), 7.92 (m, 1H), 8.64 (d, J = 8.4 Hz, 2H) + C25H14F3N3; HRMS (M + Na ): theoretical mass, 436.1032; actual mass, 436.1032 4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-N-hydroxy- benzamidine 1H-NMR δ 7.10 (s, 1H), 7.34 (dd, J = 4.0, 0.9 Hz, 1H), 7.36 (d, J = 0.9 Hz, 1H), 27 7.37 (d, J = 0.9 Hz, 1H), 7.42-7.45 (m, 3H), 7.46 (d, J = 0.8 Hz, 1H), 7.51-7.52 (m, 2H), 7.53 (d, J = 0.9 Hz, 1H), 7.57 (s, 1H), 7.89 (s, 1H), 7.91 (s, 1H) + C25H17F3N3O; HRMS (M + Na ): theoretical mass, 469.1220; actual mass, 469.1247 5-(2-phenanthrenyl)-3-(trifluoromethyl)-4-(1H-tetrazol-5-ylphenyl)-1H- pyrazole 1H-NMR δ 6.82 (s, 1H), 7.28 (d, J = 1.8 Hz, 1H), 7.38 (d, J = 8.7 Hz, 2H), 7.48- 28 7.74 (m, 5H), 7.74 (d, J = 2.5 Hz, 2H), 7.95 (d, J = 8.7 Hz, 2H), 8.47 (d, J = 8.7 Hz, 2H) + C25H15F3N6; HRMS (M + Na ): theoretical mass, 479.1202; actual mass, 479.1225 4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-benzaldehyde oxime 1H-NMR δ 6.81 (s, 1H), 7.27-7.30 (m, 3H), 7.47 (d, J = 8.7 Hz, 2H), 7.52-7.57 29 (m, 4H), 7.68 (d, J = 8.8 Hz, 2H), 7.75-7.79 (m, 2H), 8.48-8.53 (m, 2H) + C25H16F3N3O; HRMS (M + Na ): theoretical mass, 454.1137; actual mass, 454.1106

Continued

105 Table 7.1 continued

4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-benzaldehyde hydrazone 1H-NMR δ 6.81 (s, 1H), 7.27-7.30 (m, 2H), 7.33 (d, J = 1.8 Hz, 1H), 7.42 (d, J 30 = 8.6 Hz, 1H), 7.53-7,55 (m, 2H), 7.57-7.60 (m, 2H), 7.68 (d, J = 8.9 Hz, 2H), 7.75 (d, J = 1.7 Hz, 1H), 7.80 (s, 1H), 8.48-8.55 (m, 2H) + C25H17F3N4; HRMS (M + Na ): theoretical mass, 453.1297; actual mass, 453.1302 {4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl}- acetonitrile 1H-NMR δ 3.77 (s, 2H), 6.93 (s, 1H), 7.29-7.43 (m, 4H), 7.66-7.86 (m, 6H), 31 8.65 (t, J = 7.0 Hz, 3H) + C26H16F3N3; HRMS (M + Na ): theoretical mass, 450.1151; actual mass, 450.1188 2-{4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl}-N- hydroxy-acetamidine 1H-NMR δ 3.30 (s, 1H), 3.38 (s, 1H), 6.83 (s, 1H), 7.20-7.41 (m, 4H), 7.59-7.89 32 (m, 6H), 8.55-8.60 (m, 3H) + C26H19F3N4O; HRMS (M + Na ): theoretical mass, 461.1580; actual mass, 461.1584 5-(2-phenanthrenyl)-3-(trifluoromethyl)-4-(1H-tetrazol-5-ylmethylphenyl)-1H- pyrazole 1H-NMR δ 4.45 (s, 2H). 7.15 (s, 1H), 7.42 (s, 4H), 7.53 (d, J = 6.9 Hz, 1H), 33 7.66-7.76 (m, 3H), 7.89 (d, J = 7.2 Hz, 1H), 8.01 (m, 2H), 8.78 (t, J = 6.9 Hz, 2H) + C26H17F3N6; HRMS (M + Na ): theoretical mass, 493.1335; actual mass, 493.1359 2-amino-N-{4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]- phenyl} acetamide 1H-NMR δ 3.48 (s, 2H), 6.92 (s, 1H), 7.35 (d, J = 8.8 Hz, 2H), 7.42 (dd, J = 8.6, 34 1.7 Hz, 1H), 7.62-7.72 (m, 5H), 7.79 (d, J = 8.8 Hz, 1H), 7.85-7.94 (m, 2H), 8.62 (t, J = 8.5 Hz, 2H), 9.56 (br s, 1H) + C26H19F3N4O; HRMS (M + Na ): theoretical mass, 483.1403; actual mass, 483.1389 4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl-guanidine 1H-NMR δ 6.90 (s, 1H), 7.19 (d, J = 8.7 Hz, 2H), 7.34 (dd, J = 8.7, 2.0 Hz, 1H), 7.39 (d, J = 8.7 Hz, 2H), 7.61-7.67 (m, 3H), 7.79 (d, J = 9.0 Hz, 1H), 7.84-7.91 35 (m, 3H), 8.62 (d, J = 8.3 Hz, 2H), 9.95(s, 1H) + C25H18F3N5; HRMS (M + H ): theoretical mass, 446.1587(M+H); actual mass, 446.1596(M+H)

Continued

106 Table 7.1 continued

4-[5-(2-phenanthrenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]-phenyl-urea 1H-NMR δ 6.98 (s, 1H), 7.19 (dt, J = 8.9, 2.1 Hz, 2H), 7.34-7.42 (m, 3H), 7.51- 36 7.62 (m, 4H), 7.70 (d, J = 9.0 Hz, 1H), 7.81-7.85 (m, 2H), 8.59-8.64 (m, 2H) + C25H17F3N4O; HRMS (M + Na ): theoretical mass, 469.1252; actual mass, 469.1199

107 BIBLIOGRAPHY

1. Nelson WG, Marzo AM, Isaacs WB. Prostate cancer. N. Engl. J. Med. 2003: 349(4): 366-381.

2. Carol Balmer. Chapter 92 prostate cancer In Herfindal ET, Gourley DR. (Clinical pharmacotherapy) Textbook of therapeutics : drug and disease management. Lippincott Williams & Wilkins, c2000

3. Huggins C, Hodges CV. Studies on prostatic cancer: I the effect of castration of estrogens and androgen injection on serum phosphatases in metastatic carcinoma of the prostate. Cancer Res. 1941;1: 293-7

4. Huggins C, Stevens R, Hodges CV. Studies on prostatic carcinoma: II. The effect of castration on advanced carcinoma of the prostatic gland. Arch. Surg. 1941; 43: 209- 33.

5. Harris KA, Reese DM. Treatment options in hormone-refractory prostate cancer: current and future approaches. Drugs. 2001;61(15):2177-92

6. Kelly WK, Scher HI. Prostate specific antigen decline after antiandrogen withdrawal: the flutamide withdrawal syndrome. J Urol 1993; 149 (3): 607-9

7. Small EJ, Carroll PR. Prostate-specific antigen decline after casodex withdrawal: evidence for an antiandrogen withdrawal syndrome. Urology 1994; 43 (3): 408-10

8. Dawson NA, McLeod DG. Dramatic prostate specific antigen decrease in response to discontinuation of megestrol acetate in advanced prostate cancer: expansion of the antiandrogen withdrawal syndrome. J Urol 1995; 153 (6): 1946-7

9. Bissade NK, Kaczmarek AT. Complete remission of hormone refractory adenocarcinoma of the prostate in response towithdrawal of diethylstilbesterol. J Urol 1995; 153 (6): 1944-5

10. Veldscholte J, Berrevoets CA, Mulder E. Studies on the human prostatic cancer cell line LNCaP. J Steroid Biochem Mol Biol. 1994; 49(4-6):341-6.

108 11. Kulp SK, Chen KF and Chen CS. Chemotherapy for prostate cancer. In Chang, c. and Messing, E. (eds). Hormone Therapy of Prostate Cancer. World Scientific Publishing: Hackensack NJ. In press, 2004.

12. Scherr D, Swindle PW, Scardino PT. National Comprehensive Cancer Network guidelines for the management of prostate cancer. Urology 2003; 61:14-24

13. Druker BJ, Sawyers CL, Kantarjian H, et al. Activity of a specific inhibitor of the BCR-ABL tyrosine kinase in the blast crisis of chronic myeloid leukemia and acute lymphoblastic leukemia with the philadelphia chromosome. N Engl J Med. 2001; 344:1038-1042

14. Joensuu H, Roberts PJ, Sarlomo-Rikala M, et al. Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor. N Engl J Med. 2001; 344:1052-1056

15. Lynch T. J., Bell D. W., Sordella R. et. al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non–small-cell lung cancer to gefitinib. N Engl J Med. 2004; 350:2129-2139

16. Paez JG, Janne PA, Lee JC. et. al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science. 2004;304(5676):1497-500

17. Sawyers CL. Opportunities and challenges in the development of kinase inhibitor therapy for cancer. Genes Dev. 2003;17(24):2998-3010

18. Arnold JT, Isaacs JT. Mechanisms involved in the progression of androgen- independent prostate cancers: it is not only the cancer cell's fault. Endocr Relat Cancer. 2002;9(1):61-73

19. Feldman BJ, Feldman D. The development of androgen-independent prostate cancer. Nat Rev Cancer. 2001;1(1):34-45

20. Cohen P. Protein kinases-the major drug targets of the twenty-first century? Nature reviews-drug discovery 2002; 1:309-315

21. Vivanco I., Sawyers CL. The Phosphatidylinositol 3-Kinase-Akt Pathway in Human Cancer. Nat Rev Cancer. 2002; 2: 489

22. Dancey J, Sausville EA. Issues and progress with protein kinase inhibitors for cancer treatment. Nat Rev Drug Discov. 2003; 2(4):296-313.

23. Li Q., Zhu GD. Targeting serine/threonine protein kinase B/Akt and cell-cycle checkpoint kinases for treating cancer. Current topics in medicinal chemistry 2002; 2: 939-971.

109 24. Komander D, Kular GS, Schuttelkopf AW, Deak M, Prakash KR, Bain J, Elliott M, Garrido-Franco M, Kozikowski AP, Alessi DR, van Aalten DM. Interactions of LY333531 and other bisindolyl maleimide inhibitors with PDK1. Structure. 2004; 12(2):215-26.

25. Komander D, Kular GS, Bain J, Elliott M, Alessi DR, Van Aalten DM. Structural basis for UCN-01 (7-hydroxystaurosporine) specificity and PDK1 (3- phosphoinositide-dependent protein kinase-1) inhibition. Biochem J. 2003; 375:255- 62.

26. Fischer PM. The Design of drug candidate molecules as selective inhibitors of therapeutically relevant protein kinases. Curr. Med. Chem. 2004; 11(12): 1563

27. Hanks S.K., Quinn A.M., Hunter T. The protein kinase family: conserved features and deduced phylogeny of the catalytic domains. Science, 2002; 241: 42

28. Fabbro D. et al. Protein kinases as targets for anticancer agents: from inhibitors to useful drugs. Pharmacology&therapeutics 2002;93: 79-98

29. Bluen-Jensen P, Hunter T. Oncogenic kinase signaling. Nature. 2001; 411: 355-365.

30. Sielecki TM, Boylan JF, Benfield PA, Trainor GL. Cyclin-dependent kinase inhibitors: useful targets in cell cycle regulation. J Med Chem. 2000;43(1):1-18

31. Noble MEM., Endicott JA., Johnson LN.. Protein kinase inhibitors: insights into drug design from structure. Science 2004; 303: 1800-1805

32. Engh RA, Bossemeyer D. Structural aspects of protein kinase control-role of conformational flexibility. Pharmacol Ther. 2002;93(2-3):99-111

33. Huse M, Kuriyan J. The conformational plasticity of protein kinase cell. 2002; 109: 275-282

34. Nagar B, Bornmann WG, Pellicena P, Schindler T, Veach DR, Miller WT, Clarkson B, Kuriyan J. Crystal structures of the kinase domain of c-Abl in complex with the small molecule inhibitors PD173955 and imatinib (STI-571). Cancer Res. 2002;62(15):4236-43.

35. Parang K, Sun G. Design strategies for protein kinase inhibitors. Current opinion in drug discovery & development 2004; 7(5): 617-629

36. Bridges AJ. Chemical inhibitors of protein kinases. Chem. Rev. 2001; 101: 2541- 2571.

110 37. Blum, G. et al. Substrate competitive inhibitors of IGF-1 receptor kinase. Biochemistry 2000; 39: 15705–15712

38. Parang, K. et al. Mechanism-based design of a protein kinase inhibitor. Nat. Struct. Biol. 2001; 8: 37–41

39. Davies TG, Bentley J, Arris CE, et.al. Structure-based design of a potent purine-based cyclin-dependent kinase inhibitor. Nat Struct Biol. 2002;9(10):745-9.

40. Karamouzis MV, Papavassiliou AG. COX-2 inhibition in cancer therapeutics: a field of controversy or a magic bullet? Expert Opin Investig Drugs. 2004;13(4):359-72.

41. Zha S, Yegnasubramanian V, Nelson WG, Isaacs WB, De Marzo AM. Cyclooxygenases in cancer: progress and perspective. Cancer Lett. 2004;215(1):1-20.

42. Grossman HB Selective COX-2 inhibitors as chemopreventive and therapeutic agents. Drugs Today. 2003;39(3):203-12

43. Toki A.T. and Masferrer J.L. Celecoxib: a specific COX-2 inhibitor with anticancer properties. Cancer Control. 2002;9(2 Suppl):28-35.

44. Chandrasekharan NV, Dai H, Roos KL, Evanson NK, Tomsik J, Elton TS, Simmons DL COX-3, a cyclooxygenase-1 variant inhibited by acetaminophen and other analgesic/antipyretic drugs: cloning, structure, and expression..Proc Natl Acad Sci U S A. 2002;99(21):13926-31.

45. Lim JT, Piazza GA, Han EK, et.al. Sulindac derivatives inhibit growth and induce apoptosis in human prostate cancer cell lines. Biochem Pharmacol. 1999;58(7):1097-107.

46. Piazza GA, Rahm AL, Krutzsch M, et al. Antineoplastic drugs sulindac sulfide and sulfone inhibit cell growth by inducing apoptosis. Cancer Res. 1995;55(14):3110-6

47. Waddell WR, Loughry RW. Sulindac for polyposis of the colon. J Surg Oncol. 1983; 24(1):83-7.

48. Sharma R.A. Translational medicine:targeting cyclooxygenase isozymes to prevent cancer. QJM. 2002;95(5):267-73

49. Steinbach G, Lynch PM, Phillips RK, et.al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med. 2000;342(26):1946-52

50. Koehne CH, Dubois RN. COX-2 inhibition and colorectal cancer. Semin Oncol. 2004;31(2 Suppl 7):12-21.

111

51. Arun B, Goss P. The role of COX-2 inhibition in breast cancer treatment and prevention. Semin Oncol. 2004;31(2 Suppl 7):22-9.

52. Sandler AB, Dubinett SM. COX-2 inhibition and lung cancer. Semin Oncol. 2004;31(2 Suppl 7):45-52.

53. Sabichi AL, Lippman SM COX-2 inhibitors and other nonsteroidal anti inflammatory drugs in genitourinary cancer. Semin Oncol. 2004;31(2 Suppl 7):36-44.

54. Hla T, Ristimaki A, Appleby S, Barriocanal JG. Cyclooxygenase gene expression in inflammation and angiogenesis. Ann. NY. Acad Sci 1993;696:197–204.

55. Dubois RN, Abramson SB, Crofford L, Gupta RA, Simon LS, Van De Putte LB, et al. Cyclooxygenase in biology and disease. FASEB J 1998;12:1063–73.

56. Taketo MM. Cyclooxygenase-2 inhibitors in tumorigenesis (part I). J Natl. Cancer Inst 1998;90:1529–36.

57. Hla T, Bishop-Bailey D, Liu CH, Schaefers HJ, Trifan OC. Cyclooxygenase-1 and -2 isoenzymes. Int J Biochem Cell Biol 1999;31:551–7.

58. Prescott SM, Fitzpatrick FA. Cyclooxygenase-2 and carcinogenesis. Biochim Biophys Acta 2000;1470:M69–78.

59. Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E,et al. Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 1996;87:803–9.

60. Liu CH, Chang SH, Narko K, Trifan OC, Wu MT, Smith E, et al. Overexpression of cyclooxygenase-2 is sufficient to induce tumorigenesis in transgenic mice. J Biol Chem 2001;276:18563–9.

61. Williams CS, Tsujii M, Reese J, Dey SK, DuBois RN. Host cyclooxygenase-2 modulates carcinoma growth. J Clin Invest 2000;105:1589–94.

62. Song X, Lin HP, Johnson AJ, Tseng PH, Yang YT, Kulp SK, Chen CS. Cyclooxygenase-2, player or spectator in cyclooxygenase-2 inhibitor-induced apoptosis in prostate cancer cells. J Natl Cancer Inst 2002;94:585–91.

63. Hsu AL, Ching TT, Wang DS, Song X, Rangnekar VM, Chen CS. The cyclooxygenase-2 inhibitor celecoxib induces apoptosis by blocking Akt activation in human prostate cancer cells independently of Bcl-2. J Biol Chem 2000;275:11397– 403.

112 64. Arico S, Pattingre S, Bauvy C, et al. Celecoxib induces apoptosis by inhibiting 3- phosphoinositide-dependent protein kinase-1 activity in the human colon cancer HT- 29 cell line. J Biol Chem 2002;277:27613–21.

65. Kulp SK, Yang YT, Hung CC, et al. PDK-1/Akt signaling represents a major COX-2- independent target for celecoxib in prostate cancer cells. Cancer Res 2004; 64: 1444- 51

66. Alessi DR, James SR, Downes CP, Holmes AB, Gaffney PR, Reese CB, Cohen P. Characterization of a 3-phosphoinositide-dependent protein kinase which phosphorylates and activates protein kinase Balpha. Curr Biol. 1997;7(4):261-9

67. Force T, Kuida K, Namchuk M, Parang K, Kyriakis JM. Inhibitors of protein kinase signaling pathways: emerging therapies for cardiovascular disease. Circulation 2004; 1196

68. Zhu J, Song X, Lin HP, Young DC, Yan S, Marquez VE, Chen CS. Using cyclooxygenase-2 inhibitors as molecular platforms to develop a new class of apoptosis inducing agents. J Natl Cancer Inst 2002;94:1745–57.

69. Hajduk PJ., Bures M., Praestgaard J., Fesik SW. Privileged Molecules for Protein Binding Identified from NMR-Based Screening. J. Med. Chem. 2000; 43: 3443-3447

70. Talley JJ, Bertenshaw SR, Carter JS, et. al. Discovery of selective COX-2 inhibitors. Actulitaes de chimie thaerapeutique. 1999;25: 125-134.

71. Zhu J, Huang JW, Tseng PH, Yang YT, Fowble J, Shiau CW, Shaw YJ, Kulp SK, Chen CS. From the cyclooxygenase-2 inhibitor celecoxib to a novel class of 3- phosphoinositide-dependent protein kinase-1 inhibitors. Cancer Res. 2004;64(12):4309-18

72. Storz P, Toker A. 3'-phosphoinositide-dependent kinase-1 (PDK-1) in PI 3-kinase signaling. Front Biosci. 2002; 7:d886-902.

73. Morris GM, Goodsell DS, Halliday RS, et al. Automated docking using a Lamarckian genetic algorithm and empirical binding free energy function. J Comput Chem 1998;19:1639–62.

74. Juffer AH, Eisenhaber F, Hubbard SJ, Walther D, Argos P. Protein sc. 1995; 4: 2499-2509.

75. Johnson AJ, Smith LL, Zhu J, Heerema NA, Jefferson S, Mone A, Grever M, Chen CS, Byrd JC. A novel celecoxib derivative, OSU03012, induces cytotoxicity in primary CLL cells and transformed B-cell lymphoma via a caspase and Bcl-2 independent mechanism. Blood. 2004 Sep 28

113 76. Kucab JE, Chen CS, Zhu J, Gilks CB, Cheang M, Huntsman D, Yorida E, Emerman J, Pollak M, Dunn SE. Celcoxib analogs distup Akt signaling which is commonly activated in primary breast tumors. (submitted)

77. Tseng PH, Lin HP, Zhu J, Chen KF, Byrd JC, Grever M, Druker BJ, and Chen CS. Synergistic Interactions between Imatinib and the Novel Phosphoinositide-Dependent Kinase-1 Inhibitor OSU-03012 in Overcoming Imatinib Resistance. Blood. 2005; Jan 21

78. Tong L, Pav S, White DM, Rogers S, Crane KM, Cywin CL, Brown ML, Pargellis CA. A highly specific inhibitor of human p38 MAP kinase binds in the ATP pocket. Nature Struct. Biol. 1997; 4: 311.

79. Fitzgerald CE, Patel SB, Becker JW, Cameron PM, Zaller D, Pikounis VB, O'Keefe SJ, Scapin G. Structural basis for p38alpha MAP kinase quinazolinone and pyridol- pyrimidine inhibitor specificity. Nature Struct. Biol. 2003; 10: 764.

80. Stamos J, Sliwkowski MX, Eigenbrot C. Structure of the epidermal growth factor receptor kinase domain alone and in complex with a 4-anilinoquinazoline inhibitor. J Biol Chem. 2002; 277(48):46265-72.

81. Wang, Zhulun; Canagarajah, Bertram J; Boehm, Jeffrey C; Kassisà, Skouki; Cobb, Melanie H; et. al. Structural basis of inhibitor selectivity in MAP kinases. Structure 1998; 6(9):1117-1128

82. Shah NP., Tran C., Lee FY., Chen P., Norris D., Sawyers CL. Overriding Imatinib Resistance with a Novel ABL Kinase Inhibitor. Science. 2004; 305:399-401

83. Biondi RM, Komander D, Thomas CC, Lizcano JM, Deak M, Alessi DR, van Aalten DM.High resolution crystal structure of the human PDK1 catalytic domain defines the regulatory phosphopeptide docking site. EMBO J. 2002;21(16):4219-28.

84. Komander D, Kular GS, Schuttelkopf AW, Deak M, Prakash KR, Bain J, Elliott M, Garrido-Franco M, Kozikowski AP, Alessi DR, van Aalten DM. Interactions of LY333531 and other bisindolyl maleimide inhibitors with PDK1. Structure (Camb). 2004;12(2):215-26

85. Lipinski CA. Drug-like properties and the causes of poor solubility and poor permeability. J. Pharma. Tox. Meth. 2000; 44: 235-249.

86. Hickman, JA., Lazo, JS.. Changes and challenges-the world post-Gleevec. Current opinion in pharmacology 2002; 2:357-360

87. Rich JN, Rasheed BK, Yan H. EGFR mutations and sensitivity to gefitinib. N Engl J Med. 2004;351(12):1260-1

114

88. Badone, D.; Baroni, M.; Cardamone, R.; Ielmini, A.; Guzzi, U. Highly Efficient Palladium-Catalyzed Boronic Acid Coupling Reactions in Water: Scope and Limitations. J.Org.Chem. 1997; 62(21): 7170-7173

89. Flynn P, Wongdagger M, Zavar M, Dean NM, Stokoe D Inhibition of PDK-1 activity causes a reduction in cell proliferation and survival. Curr Biol. 2000;10(22):1439-42.

90. Chou, T.-C., and Talalay, P. Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Adv. Enz. Regul. 1984; 22: 27-55

91. Heitman J, Movva NR, Hall MN. Targets for cell cycle arrest by the immunosuppressant rapamycin in yeast. Science. 1991;253(5022):905-9

115