PHARMACOLOGY OF SELECTIVE ANDROGEN RECEPTOR MODULATORS (SARMS) DISSERTATION. Presented in Partial Fulfillment of the Requirements for

Transcription

1 PHARMACOLOGY OF SELECTIVE ANDROGEN RECEPTOR MODULATORS (SARMS) DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University By Wenqing Gao, B.S. The Ohio State University 2004 Dissertation Committee: Dr. James T. Dalton, Advisor Approved by Dr. Robert W. Brueggemeier Dr. M. Guillaume Wientjes Dr. Thomas D. Schmittgen Advisor Pharmacy Graduate Program

2

3 ABSTRACT The tissue selectivity of a new generation of selective androgen receptor modulators (SARMs) was characterized in castrated and intact male rats. Studies focused on two SARMs, S-1 and S-4. In castrated animals, SARMs showed strong agonist activity in the anabolic tissues by maintaining and/or restoring castration induced loss in levator ani muscle mass, soleus muscle strength, and total body bone mineral density; but weak agonist activity in maintaining and stimulating prostate growth. Although SARMs still worked as full agonists in the muscle of intact animals, S-1 and S-4 behaved as antagonists in the prostate with the presence of endogenous androgens. Furthermore, SARM also showed agonist activity in the pituitary, which could help maintain the feedback regulation of plasma LH and FSH levels. The tissue-selective agonist activity of SARMs in the anabolic tissues and the pituitary suggests that this novel class of nonsteroidal AR ligands might serve as better alternatives for male hormone replacement therapy and treatment of benign prostate hyperplasia (BPH). Another significant advantage of SARMs, compared to steroidal ligands, is the oral bioavailability of this class of compounds. Pharmacokinetic studies showed that SARMs were orally available. In the castrated animal model, SARM also showed strong anabolic activity after oral administration, even more potent than oxandrolone, an orally available anabolic steroid, suggesting that SARM could also be used in the treatment of disease-related muscle wasting. The mechanism underlying the tissue selectivity of SARMs was studied in detail. In vitro experiments using transiently expresvhgkxpdq -reductase showed that SARMs were not substrates for -UHGXFWDVH7KXV -reductase inhibition by SARMs could not contribute to the antagonist activity of SARM in the prostate in intact animals. The tissue selectivity of SARM was more related to the fact that testosterone activity in the prostate is amplified by conversion to dihydrotestosterone (DHT), a more potent androgen receptor agonist, while SARM activity was not amplified. ii

4 The other possible mechanism of action was the formation of active metabolites by liver metabolizing enzymes. In vitro and in vivo metabolism studies were conducted for S-4. The major metabolite identified was the B-ring deacetylation product of S-4. Although species differences were observed in S-4 metabolism due to the species difference in N-acetyltransferase expression, the metabolite was not active and could not contribute to the pharmacologic activity of S-4. Gene expression profiling using a prostate cancer cell line, LNCaP, revealed the ligand-specific regulation of gene expression by S-4 as compared to DHT, suggesting that the tissue selectivity might not be simply due to the differences in the potency of these two ligands. In conclusion, these SARMs, S-1 and S-4 demonstrated strong agonist activity in the muscle and pituitary, but weak agonist activity in the prostate with or without the presence of endogenous androgens. In vitro and in vivo studies showed that the tissue selective pharmacologic activity of SARMs was markedly different from steroidal androgens and could be related to the tissue speclilfh[suhvvlrqri - reductase and ligand-specific regulation of gene expression in the prostate. iii

5 Dedicated to my parents Huaiqi Gao and Chaorong Huang iv

6 ACKNOWLEDGMENTS I would like to express my sincere gratitude to my advisor, Dr. James T. Dalton, for his friendship, guidance, encouragement and unconditional support. I would also like to thank the members of my dissertation committee, Dr. Robert W. Brueggemeier, Dr. M. Guillaume Wientjes, and Dr. Thomas D. Schmittgen, for their comments and discussions. I wish to thank Dr. Duane D. Miller and his research group, Dr. Kiwon Chung, Dr. Vipin A. Nair for synthesizing the compounds studied in this dissertation; Dr. David W. Russell (University of 7H[DV'DOODV7;IRUSURYLGLQJWKHKXPDQ -reductase expression plasmids; Dr. Charles R. Yates and Mr. Leslie B. Stuart for performing the real-time PCR analysis; and Dr. Peter J. Reiser for technical support and valuable scientific discussions. This research was supported in part by grants from NIH (R01 DK59800) and GTx Inc. Memphis, TN to Dr. James T. Dalton. v

7 VITA June 20, Born Chengdu, China 1997 B.S. Pharmacology, China Pharmaceutical University Research Assistant, New Drug Screening and Development Center China Pharmaceutical University Graduate Research Assistant, University of Tennessee, Memphis, TN 2000-present Graduate Research Associate, The Ohio State University PUBLICATIONS Research Publications 1. Fu Y, Matta SG, Gao W and Sharp BM. Local Alpha-bungarotoxin-sensitive Nicotine Receptors in The Nucleus Accumbens Modulate Nicotinine-Stimulated Dopamine Secretion In vivo. Neuroscience, 101(2): , Fu Y, Matta SG, Gao W, Brower VG and Sharp BM. Systemic Nicotine Stimulates Dopamine Release in NucleusAccumbens: Re-evaluation of the Role of N-Methyl-D-aspartate Receptors in the Ventral Tegmental Area. Journal of Pharmacology and Experimental Therapeutics, 294(2): , Yin D, Gao W, Kearbey JD, Xu H, Chung K, Miller DD, and Dalton JT. Pharmacodynamics of Selective Androgen Receptor Modulators. Journal of Pharmacology and Experimental Therapeutics, 304(3): , Wu Z, Gao W, Miller DD, and Dalton JT. The Favorable Effects of Weak Acids on Negative- Ion Electrospray Mass Spectrometry. Analytical Chemistry, 76(3):839-47, vi

8 5. Marhefka CA, Gao W, Chung K, Kim J, He Y, Yin D, Bohl CE, Dalton JT, and Miller DD. Design and Synthesis of Novel Selective Androgen Receptor Modulators: Toward in vivo Metabolic Stability. Journal of Medicinal Chemistry, 47(4):993-8, Kearbey JD, Wu D, Gao W, Chung K, Miller DD, and Dalton JT. Pharmacokinetics of S-3-(4- acetylamino-phenoxy)-2-hydroxy-2-methyl-n-(4-nitro-3-trifluoromethyl-phenyl)-propionamide in rats, a nonsteroidal selective androgen receptor modulator. Xenobiotica, 34(3): , FIELD OF STUDY Major Field: Pharmacy vii

9 TABLE OF CONTENTS Abstract... iii Dedication... v Acknowledgments... vi Vita... vii List of Tables... xiii List of Figures... xvii Chapters: 1. Introduction Selective Androgen Receptor Modulator Androgen Action Tissue-VSHFLILF([SUHVVLRQRI -reductase Androgen Receptor Mediated Signaling Pathways Structural Basis for Ligand-specific Regulation of Gene Expression Scope and Objectives of Dissertation Tissue Selectivity of SARMs in Intact, Hemi-orchidectomized and Orchidectomized Rats Introduction Materials and Methods Materials Animals Pharmacologic Effects of S-4 and S-1 in Male Rats of Different Hormonal Status Comparison of the Pharmacologic Effects of S-4 and S-1 to Oxandrolone in Orchidectomized Rats Results Pharmacological Effects of S-4 and S-1 in Male Rats of Different Hormonal Status Endocrine Properties of S-4 and S-1 in Male Rats of Different Hormonal Status Comparison of the Tissue Selectivity of S-4, S-1, and Oxandrolone in Orchidectomized Rats Discussion viii

10 3. Comparison of the Pharmacological Effects of S-1, Finasteride and Hydroxyflutamide in Intact Male Rat: New Approach for Benign Prostate Hyperplasia (BPH) Introduction Materials and Methods Materials Animals Experimental Design In Vitro -reductase Assays Results Pharmacological Effects of S-1, Finasteride and Hydroxyflutamide in Intact Male Rats Endocrine Properties of S-1, Finasteride and Hydroxyflutamide in Intact Male Rats S-1, S-DQG -Reductase Discussion S-4 Treatment Improves Muscle Strength and Body Composition, and Prevents Bone Loss in Orchidectomized Rats Introduction Materials and Methods Materials Animals Experimental Design Soleus Muscle Strength Measurement Body Composition and Bone Mineral Density (BMD) Measurement Electrophoretic Separation of Skeletal and Cardiac Myosin Heavy Chain (MHC) Isoforms Results Anabolic Effects of S-4 on Soleus Muscle Strength in Orchidectomized Rats Tissue-selective Restoration of the Androgen-dependent Tissues by S-4 in Orchidectomized Rats Effects of S-4 on Plasma Levels of IGF-1 and Osteocalcin Effects of S-4 on the Body Composition and BMD in Orchidectomized Rats Effects of S-4 on MHC Isoform Expression in Skeletal and Cardiac Muscles Discussion Species Difference in the Metabolism of S Introduction Materials and Methods Materials Animals In Vitro Metabolism Reaction Using Different Liver Enzyme Preparations Identification of the Phase I Metabolite of S Covalent Binding of S-4 Metabolites to Human Liver Microsomal Protein Pharmacokinetic Studies of S-4 and Its Primary Metabolite Compartmental Analysis of the Concentration versus Time Profiles of S-4 and M ix

11 5.3 Results Identification of the Phase I Metabolites and Metabolic Pathways of S-4 Using Human Liver Microsomes (HLM) and Recombinant Human CYP Enzymes Characterization of the Kinetics of Phase I Metabolism of S Covalent Binding of S-4 Metabolites to Human Liver Microsomal Protein Conversion of the Amine-derivative (M1) Back to S-4 by N-Acetyltransferase (NAT) In Vitro Characterization of the Species Difference in NAT Expression In Vivo Characterization of the Species Difference in S-4 Metabolism Discussion Effects of S-1 and S-4 on Cytochrome P450 Enzyme Expression in Primary Culture of Human Hepatocytes Introduction Materials and Methods Materials Cytotoxicity Measurement in HepG2 Cells Primary Culture of Human Hepatocytes Treatment of Human Hepatocyte Culture CYP Enzyme Function Assays Western Immunoblot Analysis Real-time PCR Analysis Results Cytotoxicity of S-1 and S-4 in HepG2 Cells Effects of S-1 and S-4 on CYP Enzyme Function, Protein Expression, and mrna Levels Discussion Profiling of SARM (S-4) - Regulated Gene Expression in LNCaP cells Introduction Materials and Methods Materials LNCaP Cell Growth Curve With Treatment of S-4 and DHT Study Design Microarray Data Analysis Results LNCaP Cell Growth Curve Comparison of S-4 and DHT-Regulated Gene Expression in LNCaP Cells Discussion Summary and Conclusions BIBLIOGRAPHY APPENDICES APPENDIX A Data Relevant to Chapter APPENDIX B Data Relevant to Chapter x

12 APPENDIX C Data Relevant to Chapter APPENDIX D Data Relevant to Chapter APPENDIX E Data Relevant to Chapter APPENDIX F Data Relevant to Chapter xi

13 LIST OF TABLES Table Page 2.1 Animal Groups and Experimental Design for Plasma concentrations of testosterone (ng/ml), LH (ng/ml), FSH (ng/ml) and PRL (ng/ml) in different treatment groups (n=5) ,QKLELWLRQRIKXPDQ -reductase isozymes expressed in transfected COS cells Body weight, soleus muscle weight, optimal length (L 0 ), cross sectional area (CSA), and soleus muscle weight to body weight ratio (n=7-8) in different treatment groups Contractile properties of the soleus muscle (n=7-8) in different treatment groups Specific substrates for recombinant human CYP enzymes, NAT1, and NAT2, and the internal standard used for HPLC analysis Pharmacokinetic parameters of S-4 and M1 in rats (n=5) Pharmacokinetic parameters of S-4 and M1 in dogs (n=3) Oligonucleotide sequences for real-time PCR analysis Genes (metabolism/proliferation) showed more than 2 fold change in expression after treatment with either DHT (1 nm) or S-4 (1 nm) Genes (structure/signal transduction) showed more than 2 fold change in expression after treatment with either DHT (1 nm) or S-4 (1 nm) Genes (protease/protein synthesis/transcription regulation) showed more than 2 fold change in expression after treatment with either DHT (1 nm) or S-4 (1 nm) Genes showed significant differences in expression between DHT (1 nm) and S-4 (1 nm) treated LNCaP cells xii

14 A.1 Normalized prostate weight (normalized by body weight, and presented as % of intact group) of male rats with different hormonal status a after 14 days treatment A.2 Normalized seminal vesicle weight (normalized by body weight, and presented as % of intact group) of male rats with different hormonal status a after 14 days treatment A.3 Normalized levator ani muscle weight (normalized by body weight, and presented as % of intact group) of male rats with different hormonal status a after 14 days treatment A.4 Plasma levels of testosterone and LH in male rats with different hormonal status a after 14 days treatment with S-4 (0.5 mg/day), S-1 (0.5 mg/day), and TP (0.5 mg/day) A.5 Plasma levels of FSH and PRL in male rats with different hormonal status a after 14 days treatment A.6 Normalized prostate weight (normalized by body weight, and presented as % of intact group) of castrated male rats treated with vehicle, S-1 or S-4 for 14 days via daily oral gavage A.7 Normalized seminal vesicle weight (normalized by body weight, and presented as % of intact group) of castrated male rats treated with vehicle, S-1 or S-4 for 14 days via daily oral gavage A.8 Normalized levator ani muscle weight (normalized by body weight, and presented as % of intact group) of castrated male rats treated with vehicle, S-1 or S-4 for 14 days via daily oral gavage A.9 Normalized prostate, seminal vesicle, and levator ani muscle weights (normalized by body weight, and presented as % of intact group) of castrated male rats treated with oxandrolone for 14 days via daily oral gavage B.1 (Q]\PHNLQHWLFVRIWUDQVLHQWO\H[SUHVVHGKXPDQW\SHDQGW\SH -reductase using testosterone (T) as a substrate (n=3) B.2,QKLELWLRQRIWUDQVLHQWO\H[SUHVVHGKXPDQW\SHDQGW\SH -reductase by S-1, S-4, and finasteride (n=3) B.3 Normalized ventral prostate weight (normalized by body weight, and presented as % of intact group) of male rats after 3 days treatment with hydroxyflutamide, S-1 and finasteride xiii

15 B.4 Normalized ventral prostate weight (normalized by body weight, and presented as % of intact group) of male rats after 6 days treatment with hydroxyflutamide, S-1 and finasteride B.5 Normalized ventral prostate weight (normalized by body weight, and presented as % of intact group) of male rats after 9 days treatment with hydroxyflutamide, S-1 and finasteride B.6 Normalized seminal vesicle weight (normalized by body weight, and presented as % of intact group) of male rats after 3 days treatment with hydroxyflutamide, S-1 and finasteride B.7 Normalized seminal vesicle weight (normalized by body weight, and presented as % of intact group) of male rats after 6 days treatment with hydroxyflutamide, S-1 and finasteride B.8 Normalized seminal vesicle weight (normalized by body weight, and presented as % of intact group) of male rats after 9 days treatment with hydroxyflutamide, S-1 and finasteride B.9 Normalized levator ani muscle weight (normalized by body weight, and presented as % of intact group) of male rats after 3 days treatment with hydroxyflutamide, S-1 and finasteride B.10 Normalized levator ani muscle weight (normalized by body weight, and presented as % of intact group) of male rats after 6 days treatment with hydroxyflutamide, S-1 and finasteride B.11 Normalized levator ani muscle weight (normalized by body weight, and presented as % of intact group) of male rats after 9 days treatment with hydroxyflutamide, S-1 and finasteride B.12 Plasma levels of testosterone, LH, and FSH in intact, castrated, and finasteride (5 mg/kg) treated male rats after 3, 6, and 9 days of treatment B.13 Plasma levels of testosterone, LH, and FSH in hydroxyflutamide treated male rats after 3, 6, and 9 days of treatment B.14 Plasma levels of testosterone, LH, and FSH in S-1 treated male rats after 3, 6, and 9 days of treatment C.1 Body weight and tissue weights of the ventral prostate, seminal vesicle, levator ani xiv

16 muscle, and soleus muscle in orchidectomized (ORX) male rats after 8 weeks treatment with S-4 or DHT C.2 Plasma levels of IGF-1, osteocalcin in orchidectomized (ORX) male rats after 8 weeks treatment with S-4 or DHT C.3 Soleus muscle optimal length (L 0 ) and cross sectional area (CSA) in orchidectomized (ORX) male rats after 8 weeks treatment with S-4 or DHT C.4 Total body bone mineral content (BMC) (measured by DEXA scans) in orchidectomized (ORX) male rats during the 8 weeks treatment with S-4 or DHT C.5 Total body bone mineral density (BMD) (measured by DEXA scans) in the orchidectomized (ORX) male rats during the 8 weeks treatment with S-4 or DHT C.6 Tissue mass, lean mass and fat mass (measured by DEXA scans) in the orchidectomized male rats before the 8 weeks treatment with S-4 or DHT C.7 Tissue mass, lean mass and fat mass (measured by DEXA scans) in the orchidectomized male rats after the 8 weeks treatment with S-4 or DHT D.1 S-4 deacetylation rate by different CYP enzymes as measured by the formation rate of product (n=2) D.2 In vitro metabolism kinetics of S-4 by CYP3A4 and HLM (n=3) D.3 In vitro metabolism kinetics of M1 acetylation by recombinant human NAT1 and NAT2 (n=3) E.1 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -naphthoflavone %1) 0RQ&<3V$&&'DQG$DFWLYLWLHVQ E.2 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -naphthoflavone (BNF) (50 0RQ&<3V$&&'DQG$P51$OHYHOVQ -4) F.1 LNCaP cell growth rates in the presence of DHT or S-4 (n=4-5) xv

17 LIST OF FIGURES Figure Page 1.1 Chemical structures of AR ligands Structural organization of human androgen receptor Androgenic and anabolic effects of S-4 (0.5 mg/day), S-1 (0.5mg/day) and TP (0.5 mg/day) in intact, hemi-orchidectomized and orchidectomized rats (n=5) Comparison of the tissue selectivity of S-1, S-4 and oxandrolone in orchidectomized rats (n=4-5) Comparison of the tissue-selectivity of hydroxyflutamide, S-1 and finasteride in the ventral prostate, seminal vesicle and levator ani muscle in intact male rats after different treatment periods (n=5) Effects of hydroxyflutamide, S-1 and finasteride on plasma concentrations of testosterone, LH and FSH in intact male rats after different treatment periods (n=5) Comparison of the anabolic effects of hydroxyflutamide, S-1 and finasteride at doses that showed equal potency in suppressing the prostate after 9 days treatment (n=5) Twitch and tetanus measurements in rat soleus muscle Peak tetanic tension (P 0 ) and normalized P 0 /CSA in different treatment groups (n=7-8) Normalized prostate, seminal vesicle and levator ani muscle weights (n=7-8) in different treatment groups Plasma IGF-1 and osteocalcin levels (n=7-8) in different treatment groups Total body BMD and BMC (n=7-8) in different treatment groups Body composition and body weight change (n=7-8) in different treatment groups during 8 weeks treatment xvi

18 4.7 Electrophoretic separation of the MHC isoforms expressed in soleus muscle samples from (n=7-8) different treatment groups Electrophoretic separation of the MHC- DQG0+&- LVRIRUPVH[SUHVVHGLQ the ventricular muscle samples from (n=7-8) different treatment groups Flutamide (A) and bicalutamide (B) metabolism Compartment models proposed for simultaneous fitting of S-4 and M1 data HPLC separation of 14 C-S-4 and its metabolites after incubation with human liver microsome (HLM) Phase I metabolism pathways of 14 C-S-4 (uniformly labeled B-ring) as determined in HLM Identification of the major CYP enzymes that are responsible for phase I metabolism of S-4 using Supersome prepared from insect cells that express different recombinant human CYP enzymes (n=2) Characterization of the enzyme kinetics of S-4 metabolism by CYP3A4 and S-4 deacetylation by HLM (n=2) Covalent binding of 14 C-S-4 metabolites to human liver microsomal proteins (n=2) Species differences in S-4 metabolism are related to the species differences in NAT expression. Deacetylation product of S-4 was observed after incubation with rat, human and dog liver microsomal preparations In vitrofrqyhuvlrqri0 0WR6-4 in the presence of acetyl-coa (A) and NADPH (B) in different enzyme preparations (n=3) In vitro conversion of M1 to S-4 by recombinant human N-acetyl transferase 1 and 2 (NAT1 and NAT2) (n=3) Pharmacokinetic profiles of S-4 and its metabolite M1 in rat and dog Representative fitting results of S-4 and M1 concentration-time profiles in rat (A) and dog (B) Cytotoxicity of S-1 and S-4 in HepG2 cells measured by SRB assay (n=3) after 72 hour treatment xvii

19 6.2 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -naphthoflavone %1) 0RQ&<3$DFWLYLW\DQGH[SUHVVLRQ Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -naphthoflavone %1) 0RQ&<3&DFWLYLW\DQGH[SUHVVLRQ Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -naphthoflavone %1) 0RQ&<3&DFWLYLW\DQGH[SUHVVLRQ Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -naphthoflavone %1) 0RQ&<3'DFWLYLW\DQGHxpression Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -naphthoflavone %1) 0RQ&<3$DFWLYLW\DQGH[SUHVVLRQ LNCaP cell growth curves in the presence of DHT or S-4 (n=4-5) Overlapping but distinct gene expression profiles of DHT and S-4 in androgen-deprived LNCaP cells Correlation of gene expression levels for LNCaP cells treated with 1 nm S-4 and 1 nm DHT for 24 hours D.1 MS and MS/MS spectra of deacetylation product M1 with m/z xviii

20 CHAPTER 1 INTRODUCTION 1.1 Selective Androgen Receptor Modulator In the past several years, the successful marketing of selective estrogen receptor modulators (SERMs) has raised the possibility of developing selective ligands for other members of the nuclear receptor superfamily. The concept of selective androgen receptor modulators (SARMs) [1, 2] also emerged: a compound that is an antagonist or weak agonist in the prostate, but agonist in the pituitary and muscle; and orally available with low hepatic toxicity. Tissue-selective activation of the androgen receptor (AR) by SARMs could not only greatly improve the side effects profile of currently available antiandrogens, like flutamide, in the treatment of BPH; but might also be used to treat muscle-wasting conditions and age-related frailty with less concern for the stimulation of potential prostate diseases. The application of endogenous (Figure 1.1 A) and synthetic steroidal AR ligands [3] is not only limited by the lack of tissue selectivity, but also poor oral bioavailability and potential hepatotoxicity. Therefore, another important property for SARM is improved oral bioavailability and low hepatotoxicity. Since the development of a new series of androgen receptor antagonists: flutamide, nilutamide and bicalutamide, more progress has been made in identifying nonsteroidal SARMs. Non-steroidal SARMs of several different structural classes have been reported, as recently reviewed by Allan and Sui [4]. Our laboratory first discovered a series of flutamide and bicalutamide analogs as a novel class of SARMs in 1998 [5]. Further structural modification improved the binding affinity, intrinsic activity, and metabolic profile of the compounds [6-9]. Two leading compounds, S-1 and S-4 (Figure 1.1 B), were identified in our 1

21 previous study [10] using castrated male rats as in vivo model to evaluate the pharmacologic activity of AR ligands. In male rats, castration causes rapid shrinkage of both the androgenic tissues like prostate and seminal vesicle, and anabolic tissue like levator ani muscle. The levator ani muscle in rat shows androgendependence similar to that of the prostate and seminal vesicle, and is widely used as a bioassay marker for the anabolic effects of androgens [49]. Testosterone propionate (TP) maintains the tissue weights of both the androgenic and anabolic tissues without any selectivity; while SARMs, especially S-4, fully maintains the levator ani muscle weight without stimulating prostate growth. The relative efficacy of S-4 in the prostate was less than 30%, compared to that of TP. Compared to the first generation of SARMs [9], S-1 and S-4 had very similar binding affinity to AR, about 10% of that of DHT, but showed significantly improved in vivo activity, suggesting improved pharmacokinetic properties. However, the study was conducted using subcutaneously implanted osmotic pumps. The oral bioavailability and the mechanism of the tissue selectivity of these compounds were yet to be determined. 1.2 Androgen Action The pharmacologic actions of androgen are mainly classified into two categories: androgenic (reproductive) and anabolic (muscular and bone). The endogenous AR ligand, testosterone (T), is considered as a non-tissue-selective androgen since it works as a full agonist in both the androgenic and anabolic tissues. Androgen therapy (still mainly T preparations) has been used to treat many male disorders, including hypogonadism. However, the androgenic effects (proliferation of prostate) of T have limited its application in hormone replacement therapy (HRT). The circulating male hormone gradually decreases in aging men, which causes osteopenia, loss of lean body mass, and reduced sexual function. At the same time, men also experience an increased risk of benign prostatic hyperplasia (BPH) and prostate cancer, which are androgen-dependent. The ideal reagent for HRT needs to have agonist activity in most target tissues (muscle and bone), but antagonist activity in the prostate. 2

22 The activities of T in different tissues are mediated through different mechanisms. In anabolic tissues like muscle and bone, and the central nervous system including the pituitary and hypothalamus, T action is directly mediated by the AR. However, in the prostate, the androgenic activity of T is amplified by FRQYHUVLRQWRGLK\GURWHVWRVWHURQH'+7DPRUHSRWHQWDQGURJHQE\ -reductase. Therefore, the prostate is generally considered a DHT-dependent tissue, while the muscle is considered a DHT-independent tissue. Also, T is converted to estrogen by aromatase, and acts through the estrogen receptor to promote bone growth. 1.3 Tissue-VSHFLILF([SUHVVLRQRI -reductase 7ZR -reductase isozymes have been identified in human and rats [11-13]W\SHDQGW\SH - reductase. The prostate expresses predominantly the type 2 isozyme, while the liver and skin express primarily the type 1 isozyme. Both isozymes are expressed at a much lower level in other peripheral tissues, including skeletal muscle and the pituitary [12, 13]. Due to the tissue-vshflilfh[suhvvlrqri - reductase, the prostatic DHT concentration is much higher than the prostatic testosterone concentration [14, 15], and DHT is believed to be the major endogenous androgen that is responsible for prostate growth. -reductase inhibition was used as a strategy to achieve tissue-selective blockage of androgen action in the SURVWDWHDVREVHUYHGZLWK -reductase inhibitors in the treatment of BPH. Possible LQWHUDFWLRQVEHWZHHQ6$50DQG -reductase might also contribute to the tissue selectivity of SARMs. 1.4 Androgen Receptor Mediated Signaling Pathways Full-length human AR contains 919 amino acids [16]. Similar to most nuclear receptors, AR contains four main functional domains: N-terminal transactivation domain (NTD), DNA binding domain (DBD), hinge region, and C-terminal ligand-binding domain (LBD) (Figure 1.2). Two transactivational functions have been identified. The N-terminal activation function (AF-1) is not conserved in sequence and is ligand-independent (constitutively active), whereas the C-terminal activation function (AF-2) is 3

23 conserved in sequence and functions in a ligand-dependent manner [16]. A nuclear localization signal spans the region located between the DBD and the hinge region. Similar to all the other steroid receptors, unbound AR is associated with a complex of heat shock proteins (HSPs) through interactions with LBD, and is mainly located in the cytoplasm [17]. Upon agonist binding, the HSPs dissociate from AR LBD, and the transformed AR undergoes N-terminal and C-terminal interaction (N/C interaction), dimerization, phosphorylation, and translocation to the nucleus. Recruitment of certain transcription coregulators (co-activators or co-repressors) and transcriptional machinery happens at the same time to ensure the transactivation of AR-regulated gene expression. This is the wellcharacterized genomic pathway of AR action. Besides the genomic pathway, the non-genomic pathway of AR has also been reported in oocytes [18], skeletal muscle cells [19], osteoblasts [20], and prostate cancer cells [21]. Different from the genomic pathway, non-genomic action of steroid receptor action is characterized by the rapidity of the action, which varies from seconds to an hour or so; and the interaction with the plasma membrane-associated signaling pathway [22]. Separation of the genomic and non-genomic functions of steroid receptors using specific ligands was also proposed as a new strategy to achieve tissue selectivity [20, 23]. 1.5 Structural Basis for Ligand-Specific Regulation of Gene Expression Co-regulator recruitment is required for the genomic function of AR, and is directly regulated by the ligand-induced conformational changes in LBD. The agonist-induced conformational change generates a cognate surface (AF2 region as shown in Figure 1.2) for co-activator interaction, and the so-called coactivator nuclear receptor box LxxLL motif [7] from the nuclear-receptor-interacting domain (NID) of coactivator can specifically bind to this surface [24]. The agonist-erxqg/%'vrirwkhu15olnh(5 [25] and 75 ZHUHFR-crystallized with a short peptide from the NID that contains the LxxLL motif. Structural analysis revealed a surface hydrophobic groove formed by the C-terminal of helix 3, loop 3-4, helix 4, and helix 12, a region that covers the highly conserved nuclear receptor LBD signature motif [26] and the AF2 core, which indicates that a similar structure is present in other NR as well. Hydrophobic interactions between the leucine residues in LxxLL motif and the hydrophobic groove hold the peptide in place; and 4

24 the hydrogen bonds between a main-chain peptide bond and two conserved residues, a lysine at the C- terminal of helix 3 and a glutamate in helix 12, form a charge clamp to further stabilize the interaction (Figure 1.2). Ligand-specific recruitment of co-activators induced by SERMs was observed with both ER DQG (5 [27]. The tissue selectivity of SERMs was found to be related to ligand-specific regulation of gene expression [28, 29], which could be affected by both tissue-specific expression of the coregulators and ligand-specific recruitment of the coregulators. Peptide binding studies [30] using the AR LBD showed that peptides containing the LxxLL motif have different binding affinities for the AF2 region. The interaction was affected by the specific ligand binding and the flanking sequence of the LxxLL motif, which suggested that ligand-induced specific conformational changes in the AF2 region could affect the recruitment of transcription coregulators and downstream regulation of gene expression. 1.6 Scope and Objectives of Dissertation A previous study in our laboratory [10] identified two members of a new generation of nonsteroidal SARMs, S-1 and S-4, showing improved pharmacologic activity and tissue selectivity in castrated male rats. However, little is known about the mechanism of the tissue selectivity of these compounds and their pharmacologic activities in the presence of endogenous androgen. The long-term goal of this research project is to define the mechanism of the tissue selectivity of SARMs. The specific objectives of the current studies were to: 1. Further characterize the in vivo pharmacologic effects of SARMs in the presence of endogenous androgen; and evaluate the anabolic effects of SARM in different anabolic tissues; 2. Explore the mechanism of tissue selectivity of SARMs. The following hypotheses were tested: - SARMsFRXOGEHLQDFWLYDWHGE\ -reductase and/or aromatase in the prostate, or SARM FRXOGLQKLELWWKH -reductase activity in vivo; - The pharmacologic activity of SARMs is mediated by the formation of active metabolites 5

25 through hepatic metabolism, and the tissue specific distribution of the metabolite contributes to the tissue selectivity of SARM; - The tissue selectivity of SARM is due to ligand-specific regulation of AR target genes in the prostate. 6

26 A. OH testosterone O 5 alpha-reductase aromatase OH OH O H dihydrotestosterone (DHT) HO estrogen B. O 2 N F O S-1 F 3 C N H O OH O 2 N NHCOCH 3 O S-4 F 3 C N H O OH Figure 1.1 Chemical structures of AR ligands. A. Primary endogenous steroidal androgens; B. Nonsteroidal SARMs, S-1 and S-4. S-1 = S-3-(4-fluorophenoxy)-2-hydroxy-2-methyl-N-(4-nitro- 3-trifluoromethylphenyl)propionamide; S-4 = S-3-(4-acetylamino-phenoxy)-2-hydroxy-2-methyl-N-(4-nitro-3- trifluoromethyl-phenyl)-propionamide. 7

27 23 FxxLF WxxLF AF2 core AF AF NLS 634 NH 2 - AF2 -COOH N-C Interaction Co-regulator Recruitment LxxLL Motif NTD DBD LBD Hinge V 730 K 720 H4 I 737 M 734 V 716 Q 738 H5 W 741 L 712 E 897 H3 H12 AF2 H10 H9 H1 H8 H4 H7 H5 H3 H11 H12 H6 Figure 1.2 Structural organization of human androgen receptor (AR). The crystal structure of the LBD bound with DHT is presented in the right lower panel. A detailed view of the AF2 region is shown on the left. The LxxLL binding motif from co-regulators and the FxxLF motif from the N-terminal of the receptor both interact with LBD through the AF2 region. 8

28 CHAPTER 2 TISSUE SELECTIVITY OF SARMS IN INTACT, HEMI-ORCHIDECTOMIZED AND ORCHIDECTOMIZED RATS 2.1 Introduction Our previous study characterized the tissue selectivity of SARMs in orchidectomized male rats [10], and identified lead compounds S-1 and S-4 as partial agonist in the prostate but full agonist in the levator ani muscle. This study was designed to further characterize the tissue selectivity and endocrine properties of S-1 and S-4 in intact and hemi-orchidectomized male rats. Considering their partial agonist activity in the prostate, SARM might work as an antagonist in the prostate in the presence of endogenous androgens: testosterone and DHT, while no significant changes were expected in the levator ani muscle, due to their full agonist activity in the anabolic tissue. Two important clinical applications for SARMs are benign prostate hyperplasia (BPH) and disease-related muscle wasting. BPH treatment requires suppressive effects in the prostate, and muscle wasting treatment requires strong anabolic activity in the muscle. Although reduced circulating testosterone levels are normally observed in older men [1], endogenous androgens are not depleted, so orchidectomized animal is not the best model to mimic the hormonal status in patients. In this study, intact and hemiorchidectomized male rats were included to evaluate the tissue selectivity of SARMs in the presence of normal or decreased endogenous testosterone levels. A dose rate of 0.5 mg/day, which is similar to the ED 50 values of S-1 and S-4 in the prostate as identified in our previous study [10], was used in this experiment. 9

29 Besides the tissue selectivity, another important property of non-steroidal SARM is the improved oral bioavailability. Pharmacokinetic study of S-1 (data to be published) and S-4 [31] showed that both SARMs were orally available. In separate experiment, we also examined the androgenic and anabolic activity of S-1 and S-4 in orchidectomized rats after oral administration in comparison to oxandrolone, a steroidal AR ligand, and the only orally available anabolic reagent used to treat muscle wasting in cancer patients [32, 33]. 2.2 Materials and Methods Materials Compounds S-1 and S-4 were synthesized by Dr. Duane Miller s research group at the University of Tennessee [6]. The purities of these compounds were greater than 99%, as determined by HPLC. Testosterone propionate (TP), polyethylene glycol 300 (PEG 300, reagent grade), polyethylene glycol 400 (PEG 400, reagent grade), and dimethylsulfoxide (DMSO, reagent grade) were purchased from Sigma Chemical Company (St Louis, MO). Oxandrolone was purchased from Steraloids Inc. (Wilton, NH). Ethyl alcohol USP was purchased from AAPER Alcohol and Chemical Company (Shelbyville, KY). Alzet osmotic pumps (model 2002) were purchased from Alza Corp. (Palo Alto, CA) Animals Male Sprague-Dawley rats were purchased from Harlan Biosciences (Indianapolis, IN). The animals were maintained on a 12-hour light-dark cycle with food and water available ad libitum. The animal protocol was reviewed and approved by the Institutional Laboratory Animal Care and Use Committee of The Ohio State University. 10

30 2.2.3 Pharmacologic Effects of S-4 and S-1 in Male Rats of Different Hormonal Status Male Sprague-Dawley rats ( g) were randomly distributed into 12 groups (n=5/group). Treatments for all the groups are described in Table 2.1. Compounds S-1 and S-4 were dissolved in EtOH:PEG300 (5:95, v/v). TP was dissolved in EtOH:DMSO:PEG300 (10:35:55, v/v/v). All drugs were delivered via Alzet osmotic pumps (Model 2002): one pump per animal for S-4 and S-1 treated groups; two pumps per animal for TP treated groups due to the poor solubility of TP. On day 0, unilateral or bilateral orchidectomy was performed; at the same time, the pre-filled Alzet osmotic pumps were implanted subcutaneously. After 14 days of treatment, rats were weighed, anesthetized, and sacrificed. Blood samples were collected by venipuncture of the abdominal aorta. Plasma samples were analyzed for testosterone, FSH, LH and prolactin (PRL). Testosterone concentrations were measured by AniLytics Inc. (Gaithersburg, MD). FSH, LH and PRL levels were measured by the National Hormone and Peptide Program (Dr. A. F. Parlow, UCLA, CA) using rat LH, FSH and PRL RIA kits. The ventral prostates, seminal vesicles, and levator ani muscle were removed and weighed. All the organ weights were normalized to body weight, and analyzed for statistically significant differences between groups using single-factor ANOVA with p< Comparison of the Pharmacologic Effects of S-4 and S-1 to Oxandrolone in Orchidectomized Rats Male Sprague-Dawley rats ( g) were randomly distributed into 10 groups (n=4-5). All animals were orchidectomized the day before the initiation of drug treatment. Increasing doses of compounds S-1, S-4 and oxandrolone (0.1, 0.3, 0.5, 0.75 and 1 mg/day) were administered by oral gavage on daily basis. All drugs were prepared in EtOH:PEG400 (3:97, v:v). After 14 days of drug treatment, animals were weighed, anesthetized, and sacrificed. The ventral prostates, seminal vesicles, and levator ani muscle were removed and weighed. All organ weights were normalized to body weight. Statistical analysis was performed as described in

31 2.3 Results Pharmacological Effects of SARMs in Male Rats of Different Hormonal Status Similar to the results from our previous study [10], in orchidectomized (ORX) rats (Figure 2.1 white bars), 14-day treatment of compounds S-4 (0.5mg/day) and S-1 (0.5mg/day) maintained the prostate weight to 32.5% and 17.4%, respectively, of that observed in vehicle treated intact animals, and maintained the seminal vesicle weight to 16.5% and 10.3%, respectively, of that observed in vehicle treated intact animals (Figure 2.1). TP (0.5mg/day) increased the prostate and seminal vesicle weights to 76.8% and 63.5%, respectively, of that observed in intact animals. On the other hand, compounds S-4, S-1 and TP increased the levator ani muscle weight to 100.7%, 69.9% and 87.3%, respectively, of that observed in intact animals. At the dose rate of 0.5 mg/day, both compounds S-4 and S-1 showed 4 to 5 fold higher efficacies in the androgenic tissues than in the anabolic tissue. However, TP showed similar efficacy in both androgenic and anabolic tissues. When compounds S-4 and S-1 of the same dose were given to the intact animals (Fig. 2.1 black bars), which had normal plasma testosterone levels, both compounds significantly decreased the prostate weight to 79.4% and 64.1%, respectively, of that observed in vehicle treated intact rats, while no significant change was seen in TP treated group. Similar changes were also observed in seminal vesicle weights (90.5% for S-4, 75.9% for S-1), although not statistically significant. None of the three treatments caused any significant changes in levator ani muscle weight in intact animals. Compared to the ORX animals, unilateral orchidectomy (Figure 2.1 gray bars) didn t cause any significant changes in any of the organ weights. In the hemi-orchidectomized animals, compounds S-4 and S-1 also decreased the prostate weight to 74.7% and 62.9%, respectively, of that observed in vehicle treated intact animals, although the changes in S-4 treated animals were not significant compared to the vehicle treated hemi-orchidectomized animals. Also, compounds S-4 and S-1 significantly decreased the seminal vesicle weights to 78.6% and 79.7%, respectively, of that observed in intact animals. No significant changes in prostate or seminal vesicle weights were seen in TP treated animals. Similar to the results 12

32 observed in intact animals, none of the three treatments significantly changed the levator ani muscle weights in the hemi-orchidectomized animals Endocrine Properties of S-4 and S-1 in Male Rats of Different Hormonal Status Besides the organ weights, the plasma levels of testosterone, FSH, LH, and PRL were also measured (Table 2.2). Similar to others [34] reported before, 14 days after hemi-orchidectomy, there were no significant changes in testosterone and LH levels, although the increase in FSH level was significant. The results suggested that hemi-orchidectomy is not a good model to mimic the decreased circulating levels of testosterone in older men. ORX caused dramatic increase in FSH and LH levels, and both were partially reversed by TP (0.5 mg/day) treatment. However, at the same dose rate, TP did not significantly change plasma testosterone, FSH or LH level in either intact or hemi-orchidectomized animals. Compound S-4, at 0.5 mg/day, behaved very similar to TP in intact or hemi-orchidectomized animals, it didn t cause any significant changes in any of these parameters. However, in ORX animals, S-4 was not able to reverse the changes in testosterone, FSH and LH levels caused by orchidectomy. Compared to compound S-4, compound S-1 showed stronger agonist activity in the pituitary: 1) it significantly decreased plasma testosterone level in intact animals at the same dose rate (0.5 mg/day); 2) it also significantly decreased the LH level in castrated animals without affecting the FSH level at the same time. It s been well documented [34-38] that testosterone can modify the release of PRL and vice versa. Hemi-orchidectomy does not change PRL level in rats, while orchidectomy can cause decrease in PRL level. On the other hand, when hemi-orchidectomized or ORX rats were treated with testosterone, the exogenous testosterone can increase PRL levels in these animals. Similar changes of PRL levels in vehicle or TP treated groups (Table 2.2) were observed in our experiment as well. Both SARMs were very similar to TP in regulating PRL secretion in intact, hemi-orchidectomized or ORX animals. 13

33 2.3.3 Comparison of the Tissue Selectivity of S-4, S-1, and Oxandrolone in Orchidectomized Rats This experiment was designed to compare the tissue selective anabolic effects and the oral bioavailability of oxandrolone and compounds S-4 and S-1 using ORX animals. The same dose rates (0.1, 0.3, 0.5, 0.75, 1 mg/day) were used as in our previous study [10]. Instead of using subcutaneously implanted osmotic pumps, the drugs were given by oral administration. Although different administration route was used, very similar results were observed in S-4 and S-1 treated animals (Figure. 2.2), which confirmed that S-4 and S-1 were orally available. Orchidectomy decreased the prostate, seminal vesicle, and levator ani muscle weight to 6%, 8% and 41%, respectively, of that observed in intact animals. Both S- 4 and S-1 were able to reverse these changes after oral administration, and showed much higher efficacy in the levator ani muscle (anabolic tissue) than in the prostates and seminal vesicles (androgenic tissues). However, oxandrolone was not able to show its pharmacologic effects in any of these tissues at the same dose rates. 2.4 Discussion The androgenic effects of testosterone in the prostate and seminal vesicle are mainly mediated by '+7DPRUHSRWHQWPHWDEROLWHFRQYHUWHGE\ -UHGXFWDVH'XHWRWKHWLVVXHVSHFLILFH[SUHVVLRQRI - reductase, levator ani muscle is generally viewed as a DHT-independent tissue (lozh[suhvvlrqri - reductase), while the prostate and seminal vesicles are considered DHT-dependent tissues (high H[SUHVVLRQRI -reductase) [11]. Previous studies [39, 40] have shown that testosterone, by itself, is also an AR partial agonist. When ORX rats were treated with testosterone and finasteride -reductase inhibitor), the levator ani muscle size was fully maintained, but the prostate weight was only partially restored. Similar to testosterone, S-1 and S-4 showed partial agonist activity in the prostate in ORX animals (Figure 2.2). Furthermore, S-1 and S-4 also demonstrated tissue selectivity in the intact animals by decreasing the prostate weight, but without affecting the levator ani muscle (Figure 2.1). The suppressive effects of SARM in prostate of intact animals showed that they could be used for the treatment of benign 14

34 prostate hyperplasia (BPH), and the suppressive effects were further evaluated in more detailed studies (see Chapter 3). Gonadotropins, especially LH, contribute to the regulation of production and secretion of endogenous testosterone. Testosterone, in turn, affects the release of LH and FSH through negative feedback regulation at both pituitary and hypothalamus. At the hypothalamic level, testosterone indirectly regulates LH and FSH secretion via its ability to influence GnRH release. At the pituitary level, testosterone directly inhibits LH release. The steady state plasma concentration of TP achieved in these studies was sufficient to maintain the size of androgenic and anabolic tissues in castrated animals, but below the concentration needed to suppress pituitary gonadotropin (i.e., LH and FSH) production. S-1 decreased plasma concentrations of testosterone and LH, but had no effect on FSH concentrations (Table 2.2), suggesting that the observed decreases in LH were mediated through negative feedback effects of S-1 in the pituitary. No such effects were observed with TP and S-4, indicating that maximal pharmacologic effects in the muscle could be achieved at concentrations below that needed to regulate pituitary LH production with these compounds. The greater hydrophobicity of S-1 (log P = 5.67 ± 0.66) as compared to testosterone (log P = 3.48 ± 0.28) and S-4 (log P = 4.35 ± 0.61) suggests that this is an important physicochemical property governing negative feedback of androgens in the CNS. Additional studies with SARMs of varying hydrophobicity are needed to confirm this hypothesis. The regulatory effects of testosterone in pituitary and hypothalamus can also be mediated by the active metabolites of testosterone, estrdglroe\durpdwl]dwlrqdqg'+7e\ -reduction. However, species differences are known to exist. Bhatnagar et al [41] showed that aromatase inhibitors increase plasma LH in men but not in male rats, indicating that the pituitary regulation of LH secretion by testosterone is related to its conversion to estradiol in men, but not in male rats. In vitro transcriptional activation studies with estrogen receptor (ER) (data not shown) showed no interaction between ER and S-1 and S-4. Thus, the ability of S-1 to regulate pituitary LH secretion was most likely due to its interactions with the AR. PRL is also involved in the regulation of testicular testosterone secretion. In adult men, PRL increases plasma testosterone concentration [42], possibly, by increasing the number of LH receptors and LH-stimulated testosterone production in Leydig cells after binding to its specific receptor [43, 44]. PRL production is not under direct feedback control by plasma testosterone. Sharma et al. showed that 15

35 testosterone administration to castrated rats prevented the accumulation of thyrotropin-releasing hormone (TRH) in the posterior pituitary (PP) without affecting the expression level of TRH in the hypothalamus [37]. TRH is implicated as a regulator of PRL release [45]. In our study, S-1 increased PRL levels to a similar extent as TP, suggesting that SARMs might regulate PRL release via their actions in the pituitary and that PRL may be a useful plasma marker for androgenic effects in this animal model. In addition to characterizing the pharmacological effects of SARMs in intact male rats, we also compared the potency and efficacy of S-1 and S-4 to that of oxandrolone, a known anabolic reagent. Oxandrolone has been used to treat disease-related muscle wasting and is orally available. However, oxandrolone failed to show measurable pharmacological effects at the dose rates we used in this study. For the first time, we demonstrated the pharmacologic activity of SARMs after oral administration: both S-1 and S-4 were more potent and efficacious than oxandrolone as anabolic reagent. In summary, our previous [10] and current studies demonstrated that this novel series of nonsteroidal SARMs: (1) were active after oral administration, (2) act as partial agonists in the prostate and full agonists in levator ani muscle in both intact and ORX animals, and (3) have varying effects on pituitary gonadotropin production. As summarized by Negro-Vilar [2], SARMs are a rationale first step towards greatly expanding the clinical applications of androgen therapy. SARMs, such as S-1 and S-4, may be used directly or further optimized to achieve drugs with clinical utility in the treatment of muscle wasting and BPH. 16

36 Group # Surgical Status Drug Dose (mg/day) Number of animals 1 Intact Vehicle NA 5 2 Intact S Intact S Intact TP Hemi-orchidectomized Vehicle NA 5 6 Hemi-orchidectomized S Hemi-orchidectomized S Hemi-orchidectomized TP Orchidectomized Vehicle NA 5 10 Orchidectomized S Orchidectomized S Orchidectomized TP Table 2.1 Animal Groups and Experimental Design for

37 Intact Hemi-orchidectomized Orchidectomized Testosterone (ng/ml) Vehicle 2.67 ± ± 1.05 ND S-4 (0.5 mg/day) 1.83 ± ± 0.81 ND S-1 (0.5 mg/day) 0.86 ± ± 0.63 ND TP (0.5 mg/day) 1.48 ± ± ± 0.10 LH (ng/ml) Vehicle 0.16 ± ± ± 1.71 S-4 (0.5 mg/day) 0.03 ± ± ± 2.80 S-1 (0.5 mg/day) 0.03 ± ± ± 2.67 TP (0.5 mg/day) 0.17 ± ± ± 1.51 FSH (ng/ml) Vehicle 13 ± 1 18 ± 2 69 ± 6 S-4 (0.5 mg/day) 14 ± 2 15 ± 2 70 ± 12 S-1 (0.5 mg/day) 12 ± 3 17 ± 2 67 ± 10 TP (0.5 mg/day) 11 ± 2 17 ± 3 58 ± 7 PRL (ng/ml) Vehicle 41 ± ± 9 29 ± 6 S-4 (0.5 mg/day) 48 ± ± ± 16 S-1 (0.5 mg/day) 41 ± ± ± 20 TP (0.5 mg/day) 53 ± 5 55 ± 9 55 ± 22 Table 2.2 Plasma concentrations of testosterone (ng/ml), LH (ng/ml), FSH (ng/ml) and PRL (ng/ml) in different treatment groups (n=5). p<0.05 compared to vehicle treated control group of the same surgical status. ND: Not detectable. 18

38 % of Intact Control Prostate Intact Hemi-Orchidectomized Castrated Seminal Vesicle % of Intact Control Levator Ani Muscle % of Intact Control Vehicle S-4 S-1 TP Figure 2.1 Androgenic and anabolic effects of S-4 (0.5 mg/day), S-1 (0.5mg/day) and TP (0.5 mg/day) in intact, hemi-orchidectomized and orchidectomized rats (n=5). Prostate, seminal vesicle and levator ani muscle weights were measured after 14 days treatment via subcutaneously implanted osmotic pumps. All organ weights were normalized by body weight, and were shown as the percentage of the weights in the intact control group. p<0.05, compared to the vehicle-treated control group of the same surgical status. 19

39 S-1 S-4 Oxandrolone Levator Ani Muscle Prostate Seminal Vesicle % of Intact Control (mg/day) (mg/day) (mg/day) Figure 2.2 Comparison of the tissue selectivity of S-1, S-4 and oxandrolone in orchidectomized rats (n=4-5). Prostate, seminal vesicle and levator ani muscle weights were measured after 14 days treatment via oral administration. All organ weights were normalized by body weight, and were shown as the percentage of the weights in the intact control group. 20

40 CHAPTER 3 COMPARISON OF THE PHARMACOLOGICAL EFFECTS OF S-1, FINASTERIDE AND HYDROXYFLUTAMIDE IN INTACT MALE RAT: NEW APPROACH FOR BENIGN PROSTATE HYPERPLASIA (BPH) 3.1 Introduction Previous in vivo studies have successfully demonstrated the tissue-selective AR agonist activities of compounds S-1 and S-4 in both orchidectomized [10] and intact male rats (Chapter 2). In ORX animals, the tissue selectivity of S-1 and S-4 was apparently related to their partial agonist activity in the prostate. However, in intact animals, the suppressive effects of S-1 and S-4 in the prostate could also be explained by other mechanisms, such as: 1). Interactions with 5 -reductase and/or aromatase, including inactivation of the compound or the inhibition of the enzyme, 2). Suppression of endogenous LH and testosterone release, and 3). Formation of other active metabolites by hepatic enzymes (see Chapter 5). Previous study has shown that SARM does not interact with aromatase (data not shown). In FXUUHQWVWXG\WKHLQWHUDFWLRQEHWZHHQ6$50DQG -reductase was investigated with in vitro enzyme DVVD\VXVLQJWUDQVLHQWO\H[SUHVVHGKXPDQW\SHDQGW\SH -reductase to further explore the mechanism for the tissue selectivity of SARM. On the other hand, the suppression of prostate growth by SARM in intact animals suggested that SARM could be used as a novel approach for BPH treatment. To further characterize the endocrine properties and pharmacological activities of SARM in intact animals, and explore the possible mechanism of action in vivo, more detailed in vivo experiment with SARM in intact male rats were performed in comparison to two other prostate-suppressive reagents: WKH -reductase inhibitor finasteride and antiandrogen hydroxyflutamide, which are clinically used for BPH treatment. 21

41 Benign prostatic hyperplasia (BPH) is one of the most common diseases in men older than 50. Urinary obstruction is the main symptom of BPH, and it appears to be caused by both obstructing adenoma (static ouphfkdqlffrpsrqhqwdqgrufrqwudfwlrqvrivprrwkpxvfohvxqghu -receptor-mediated sympathetic stimulation (dynamic component) [46]. Currently, both surgical and medical options are available for treatment. The medical theraplhvlqfoxghdqgurjhqvxssuhvvlrq -blockade, aromatase inhibitors and phytotherapy [46-48]. Alpha-blockers improve the symptoms by reducing the muscular tone, while androgen suppression and aromatase inhibition counteract the static component or mechanic enlargement. Androgen suppression primarily causes the regression of the epithelial elements of the prostate while aromatase inhibitors are believed to suppress the size of the stromal component and the stromal-epithelial interactions in the prostate. The most commonly used reagents in androgen suppression include anti-androgens K\GUR[\IOXWDPLGHDQG -reductase inhibitors (finasteride). Anti-androgens directly block the androgen DFWLRQDWWKHUHFHSWRUZKLOHWKH -reductase inhibitors suppress the androgen action by inhibiting the conversion of testosterone to dihydrotestosterone (DHT) [46, 49]. Due to the complete blockage of androgen action in both the prostate and the pituitary, anti-androgens can cause significant increases in plasma testosterone and LH levels [50] -reductase inhibitors reduce prostatic and plasma DHT concentrations [15, 51] with fewer side effects on the plasma hormone levels, although increases in prostatic [15] and plasma testosterone [51] levels have been observed. The increased prostatic and/or plasma testosterone levels might contribute to the increase in prostatic concentration of estrogen, since more substrate is available to aromatase. The increased prostatic estrogen concentration might also promote the proliferation of prostate tissue, especially the stromal components. On the other hand, increased prostatic testosterone can still activate AR, although with lower affinity and intrinsic activity compared to DHT. Different from either hydroxyflutamide or finasteride, with the weak agonist activity in the prostate, SARMs could provide a completely new approach for androgen suppression in BPH treatment, with improved side effects profile due to their strong agonist activity in the anabolic tissue (Figures 2.1) and the pituitary (Table 2.2). As mentioned above, one aim of this in vivo study was to compare the potency and efficacy of S-1 in suppressing the prostate growth in intact male rats, in comparison with other androgen suppression 22

42 treatments: the antidqgurjhqk\gur[\ioxwdplghdfwlyhirupriioxwdplghdqgwkh -reductase inhibitor finasteride, and to evaluate the side effects profile as well. S-1 was used in this experiment, due to its stronger inhibitory effect in the prostate of intact animals (Figure 2.1). Since both SARM and hydroxyflutamide share the same target, AR, multiple doses of SARM and hydroxyflutamide were used to characterize the dose-response relationship; and multiple time points were included to compare the onset of the effects at different doses. Finasteride was only given at 5 mg/kg/day, the lowest dose that produces the maximum inhibitory effects in the prostate [52], due to the limited supply of the compound. The other aim of this in vivo study was to find out if the suppressive effects of SARM in the prostate are mediated by the suppression of endogenous LH and testosterone release. Instead of using osmotic pumps, drugs were administered via daily subcutaneous injections to better mimic the pulsatile release of the hormones under physiological conditions. 3.2 Methods and Materials Materials Compounds S-1 and S-4 [6], hydroxyflutamide and finasteride were synthesized by Dr. Duane Miller s research group at the University of Tennessee. The purities of these compounds were greater than 99%, as determined by HPLC. Polyethylene glycol 300 (PEG 300, reagent grade), and dimethylsulfoxide (DMSO, reagent grade) were purchased from Sigma Chemical Company (St Louis, MO). Testosterone EIA kit was purchased from Diagnostic Systems Laboratories Inc. (Webster, TX). Lipofectamine Reagent was purchased from Invitrogen Corp. (Carlsbad, California) Animals Male Sprague-Dawley rats were purchased from Harlan Biosciences (Indianapolis, IN). The animals were maintained on a 12-hour light-dark cycle with food and water available ad libitum. The 23

43 animal protocol was reviewed and approved by the Institutional Laboratory Animal Care and Use Committee of The Ohio State University Experimental Design To compare the tissue selectivity of S-1, hydroxyflutamide and finasteride in intact male rats, male Sprague-Dawley rats ( g) were randomly distributed into groups of 5 animals. The intact male rats were treated with hydroxyflutamide (0.5, 1, 5, 10 or 25 mg/kg), finasteride (5 mg/kg), S-1 (0.5, 1, 5, 10, 25mg/kg) or vehicle for 3, 6, or 9 days. The drugs were dissolved in DMSO:PEG300 (20:80, v:v) and administered via daily subcutaneous injections. A group of castrated rats (n=5) was also included as control for each time point. By the end of each treatment period, the animals were weighed, anesthetized, and sacrificed within 8 hours after the last dose. The androgenic and anabolic tissues (ventral prostate, seminal vesicle and levator ani muscle) were removed and weighed, and blood samples were collected and used for the measurement of serum markers, including FSH, LH and testosterone. Testosterone concentrations were measured by a commercially available testosterone EIA kit. FSH and LH levels were measured by the National Hormone and Peptide Program (Dr. A F Parlow, UCLA, CA) using rat LH and FSH RIA kits. The organ weights were normalized with the body weights. Percentage changes were determined by comparison to intact animals. Statistical analyses of all the parameters were performed by single-factor ANOVA with the alpha value set a priori at p< In Vitro -reductase Assays In vitro -reductase assays using cell lysate were performed as described by Thigpen et al [53]. Briefly, COS1 cells were transiently transfected with pcmv plasmid for expression of type 1 and type 2 KXPDQ -reductase (obtained from Dr. David W. Russell, Southwestern Medical Center, Dallas, TX) using Lipofectamine Reagent (Invitrogen). Forty-eight hours after transfection, cells were harvested in PBS, and cell pellets were lysed in 10 mm potassium phsphate, ph 7.0, 150 mm KCl, 1 mm EDTA using 24

44 Pro200 homogenizer (Pro Scientific). Total protein concentration in the lysate was measured by the Bradford methogxvlqjerylqhvhuxpdoexplqdvvwdqgdug -Reductase assays were conducted in 0.1 M Tris-citrate buffers at the indicated ph (Table 3.1). Cell lysate (1- JSURWHLQZDVLQFXEDWHGDWƒ&ZLWK testosterone and/or S-1, S-4 for minutes. The reaction was initiated by the addition of NADPH (final concentration 1 mm) and stopped by the addition of ice-cold acetonitrile (1:1, v:v) containing internal standard for HPLC (Agilent 1100 HPLC system) analysis. Precipitated protein was pelleted by centrifugation, and the supernatant was subjected to HPLC analysis. Testosterone concentration in the incubate was determined by HPLC using a reversed-phase FROXPQ %RQGD3DN C 18, mm, Waters Corporation, Milford, MA) and a mobile phase of acetonitrile and de-ionized water (50:50, v:v) at a flow rate of 1 ml/min, with UV absorbance at 254nm. The intra-day coefficient of variation of the HPLC analysis was 10%. The reaction velocity was calculated based on the decrease in testosterone concentration. Apparent K m, V max and IC 50 values were determined by non-linear regression analysis using WinNonlin (version 4.0, Pharsight Corporation, Mountain View, CA); apparent K i values were derived from the IC 50 values of the inhibition curves. 3.3 Results Pharmacologic Effects of S-1, Hydroxyflutamide and Finasteride in Intact Male Rats Castration caused rapid decreases in both the androgenic and anabolic tissues. After 3, 6 and 9 days treatment, prostate weight decreased to 45%, 22% and 15%, respectively; seminal vesicle weight decreased to 30%, 24% and 14%, respectively, of that in intact animals; and levator ani muscle weight decreased to 71%, 65% and 62%, respectively, of intact level (Figure 3.1). Antiandrogens block the binding of endogenous androgens, including testosterone and DHT, to AR. Hydroxyflutamide treatment caused very similar effects in both androgenic (DHT-dependent) and anabolic (DHT-independent) tissues as castration. After only three days treatment, at all the doses tested, hydroxyflutamide significantly decreased the tissue weights of prostate, seminal vesicle and levator ani muscle to near castration level measured at the same time point. However, fluctuations in the changes were 25

45 observed over time, and the dose-response relationship was only observed after 9 days treatment. At 25mg/kg dose, hydroxyflutamide significantly decreased prostate, seminal vesicle and levator ani muscle weights to 45%, 35% and 69%, respectively, of intact level. Although hydroxyflutamide (5mg/kg) decreased prostate weight to a similar extend (65%) compared to finasteride (55%) at the same dose after 9 days treatment, it also significantly decreased the levator ani muscle weight to 84% of intact level, showing its lack of tissue selectivity. Due to the tissue specific expressionri -reductase, finasteride caused significant decreases in DHT-dependent tissues (prostate and seminal vesicle) without affecting DHT-independent tissue (levator ani muscle). At 5mg/kg dose, finasteride caused significant decreases in the prostate and seminal vesicle weights in as short as three days, and the changes in tissue weights were similar after 3, 6 and 9days treatment. The prostate and seminal vesicle weights decreased to 55% and 29%, respectively, of the intact level after 9 days treatment, while no significant changes in the levator ani muscle weight was observed. Intact male rats were also treated with S-1 at the same doses as hydroxyflutamide for 3, 6 and 9 days. After 9 days treatment, S-1 selectively decreased the prostate and seminal vesicle weights without affecting the levator ani muscle in a dose-dependent manor. The maximum inhibitory effects were observed at the higher doses: 5, 10 and 25mg/kg, the prostate and seminal vesicle weights were decreased to a similar extend, 50% and 45%, respectively, of the intact level; which is comparable to the efficacy of finasteride (55% and 29%, respectively) (Figure 3.3). Although S-1 showed similar tissue selectivity as finasteride, fluctuations in the changes were observed in during 3 days and 6 days treatment. Compare three treatments, finasteride and S-1 showed better tissue selectivity than hydroxyflutamide for DHT-dependent tissues (prostate and seminal vesicle), which would cause less adverse effects in other peripheral tissues during the treatment of BPH Endocrine Properties of S-1, Finasteride and Hydroxyflutamide in Intact Male Rats As mentioned earlier, the tissue weights in hydroxyflutamide and S-1 treated animals fluctuated over time. Stable dose-response relationship was only observed after 9 days treatment. The fluctuations in tissue responses to these treatments could be due to the fluctuations in plasma concentrations of 26

46 testosterone. Therefore, we also measured plasma concentrations of testosterone, LH and FSH to compare the endocrine properties of these compounds. Hydroxyflutamide treatment significantly increased plasma testosterone, LH and FSH concentrations in intact animals in a dose-dependent manner (Figure 3.2) by blocking the negative feedback regulation of testosterone at pituitary and hypothalamus levels. The initial blockage of the feedback loop caused a rapid increase in plasma testosterone, FSH and LH levels, which caused another transient increase in tissue weights at the 6 days time point. However, once a new balance between the androgen blockage and feedback increase in plasma hormone concentrations was established after 9 days treatment, more apparent dose-response relationship was observed. Similar to hydroxyflutamide, S-1 also caused transient increases in plasma testosterone (Figure 3.2 A) after three days treatment, which could be due to its competitive binding to AR. The plasma testosterone level returned to normal level after 6 days and 9 days treatment. Different from hydroxyflutamide, S-1 is an AR partial agonist, it does not block the negative feedback loop, and instead, it helps maintain the regulation by working as an agonist in the pituitary. Therefore, the plasma LH and FSH levels (Figure 3.2 A and 3.2 C) were not increased by S-1 treatment. Plasma FSH level was slightly increased after 3 and 6 days treatment, while it returned to normal after 9 days treatment. No significant changes in plasma LH level was observed after 3 and 6 days treatment, while after 9 days treatment, higher doses of S-1, 5, 10, and 25 mg/kg, significantly suppressed the LH secretion to undetectable level, further confirmed its strong agonist activity in the pituitary. Also, significant suppression of the LH release was only observed after 9 days treatment, while significant suppression of the prostate weight was observed as early as the three day time point (Figure 3.1). The early onset of the suppressive effects of S-1 in the prostate does not correlate with the increased plasma testosterone level and unchanged plasma LH levels at the same time point (Figure 3.1 and 3.2). Therefore, the suppressive effects of S-1 in the prostate are not mediated by the suppression of endogenous LH and testosterone release. Finasteride, on the other hand, did not cause significant increases in plasma FSH concentration in LQWDFWPDOHUDWV'XHWRWKHLQKLELWLRQRI -reductase at tissue level, finasteride significantly increased plasma testosterone levels, while decreased plasma LH level in intact animals after 9 days treatment. 27

47 As demonstrated in Figure 3.1, all three different treatments were able to suppress prostate proliferation in intact animals. Hydroxyflutamide (5, 10, 20 mg/kg), S-1 (5, 10, 20 mg/kg) and finasteride (5 mg/kg) were similarly potent in decreasing prostate weight in intact animals (Figure 3.3). However, compared to hydroxyflutamide and finasteride at equal-potent doses, S-1 showed better pharmacological profile in other tissues: 1). S-1 did not cause significant decrease in levator ani muscle, and 25 mg/kg dose even significantly increase the muscle weight by 10%, showing strong agonist activity in muscle; 2). S-1 did not increase plasma testosterone, LH and FSH levels in intact animals, showing its agonist activity in the pituitary; while hydroxyflutamide dramatically increased testosterone, LH and FSH levels. In summery, S-1 showed suppressive effects in the prostate of intact male rats, with similar potency compared to hydroxyflutamide and finasteride, two clinically approved reagents for the treatment of BPH, suggesting that SARM might represent a new approach to treat BPH S-1, S-DQG -reductase In vitrohq]\phdvvd\zlwkwudqvlhqwo\h[suhvvhgkxpdqw\shru -reductase showed that S-1 and S-DUHQRWVXEVWUDWHVIRUHLWKHUW\SHRI -reductase. Experiments were performed at the optimum ph as determined by Thigpen et al. [53]: type 1 isozyme at ph 7.0 and type 2 isozyme at ph 5.0. Inhibition study did show that S-1 could inhlelw -reductase activity at high micro molar concentrations (Table 3.1), with K i YDOXHVRIDQG 0IRUW\SHDQG -reductase, respectively, while finasteride showed PXFKKLJKHUDFWLYLW\IRUERWKW\SHDQG -reductase, with K i values of 484 nm and 23 nm, respectively. Pharmacokinetic and tissue distribution studies of S-1 and/or S-4 have shown that neither plasma nor tissue concentrations of S-1 at similar dose were high enough for effective inhibition of the prostatic - reductase. Therefore, the inhibitory effects observed in vitro will not actually contribute to the in vivo activity of S-1. In summary, S-1 (SARM) and finasteride share similar tissue selective effects in vivo, although their mechanisms of action are completely different, both are closely related to the tissue-specific H[SUHVVLRQRI -reductase [12, 13]+LJKH[SUHVVLRQRI -reductase in the prostate (type 2) makes the tissue more DHT-GHSHQGHQW HLWKHU -reductase inhibition or AR competitive binding by partial agonist 28

48 (i.e. SARM) can effectively suppress the androgenic effects of DHT in the prostate; 2). Low expression of -reductase in the skeletal muscle and the pituitary makes these tissues DHT-LQGHSHQGHQW QHLWKHU - reductase inhibition nor partial activation of the androgen receptor suppresses the anabolic effects of testosterone, or attenuates the feedback regulation at the pituitary level. In comparison, antiandrogen blocks androgen action without any selectivity, causing more adverse effects in DHT-independent tissues and endocrine physiology. Therefore, SARM represents a better approach with high tissue-selectivity for BPH treatment Discussion Endogenous testosterone can be converted by 5 -reductase to a more potent form: GLK\GURWHVWRVWHURQH'+77ZR -reductase isozymes have been identified in human and rats [11-13]: type 1 and type 2 5 -reductase, with the prostate expressing predominantly the type 2 isozyme, and the liver and skin expressing primarily the type 1 isozyme. Both isozymes are expressed at a much lower level in other peripheral tissues, including skeletal muscle and pituitary [12, 13]. Due to the tissue-specific H[SUHVVLRQRI -reductase, the prostatic DHT concentration is much higher than the prostatic testosterone concentration [14, 15], and DHT is believed to be the major endogenous androgen that is responsible for the prostate growth. Therefore, prostate is considered as DHT-dependent tissue, while the pituitary and skeletal muscle are considered DHT-independent compared to the prostate. For BPH treatment, $5DQG -reductase are the major targets for androgen suppression. SARM DQGDQWLDQGURJHQVDUH$5OLJDQGVWKH\VXSSUHVV'+7DFWLRQE\FRPSHWLWLYHELQGLQJWR$5ZKLOH - reductase inhibitors suppress DHT action by decreasing DHT formation in the prostate. Antiandrogens (flutamide and bicalutamide) are AR antagonists, they block AR binding and suppress the tissue uptake of circulating testosterone [54], causing tissue androgen ablation similar to castration. Although hydroxyflutamide efficiently decreases prostate volume in male rats (Figure 3.1) [55, 56] and in BPH patients [57, 58], the action is not tissue-selective, and significant androgen suppression in other tissues causes major side effects: 1) Symptoms of androgen depletion: hot flashes, impotence [58]; suppression of the anabolic effects of androgen in the muscle (Figure 3.1); 2) Breast tenderness (42% to 29

49 52%) and gynecomastia (12% to 17%) [58] caused by increased tissue estrogen concentration, which is probably caused by the increased circulating testosterone levels after the blockage of the negative feedback loop at the pituitary and hypothalamus levels. Therefore, an AR ligand of better tissue selectivity: antagonist or weak agonist in the prostate, but agonist in the pituitary, is needed to reduce the adverse effects of antiandrogen; and the discovery of the non-steroidal SARM (bicalutamide derivative) finally accomplished this goal. As AR partial agonist, S-1 showed very low androgenic activity in the prostate, less than 15% efficacy of testosterone, which makes it a strong antagonist in the DHT-dependent tissues. In intact male rats, S-1 decreased the prostate size by half, showing similar potency to hydroxyflutamide (Figure 3.3). However, in the DHT-independent tissues, S-1 worked as AR full agonist, demonstrating greatly improved tissue-selectivity: 1). Its agonist activity in the pituitary maintained the feedback regulation of endogenous testosterone and LH secretion, so S-1 didn t cause any increases in plasma testosterone and LH levels (Figure 3.2); which should prevent the occurrence of breast tenderness and gynecomastia; 2). Its agonist activity in other peripheral tissues will help induce androgen depletion syndromes; 3). Its strong agonist activity in the skeletal muscle (Figures 3.1, 3.2 and 3.4) could also be used to treat muscle wasting and age-related frailty. Different from SARM and antiandrogen, finasteride decreases the circulating [51] and prostatic [15] DHT concentrations by LQKLELWLQJW\SH -reductase. Compared to antiandrogen, finasteride showed greatly improved tissue selectivity and endocrine properties (Figure 3.1 and 3.2) due to the tissue specific H[SUHVVLRQRIW\SH -reductase. No significant increases in circulating LH and FSH levels were observed in either animal study (Figure 3.2) or in BPH patients [51]. However, even without changing feedback UHJXODWLRQRIHQGRJHQRXVWHVWRVWHURQHHIIHFWLYHLQKLELWLRQRIWKH -reductase in tissues still causes the elevation of both circulating (Figure 3.2) and prostatic testosterone levels in rats [14, 52, 59, 60] and in human [51, 61]. Side effects like breast tenderness and gynecomastia were still observed in some BPH patients (<1%) [62-64], but significantly improved, compared to flutamide (15% to 52%) [58]. As it decreases the DHT formation in the prostate, finasteride also increases the intra-prostatic testosterone level in a reciprocal fashion [15]. Testosterone, by itself, can also stimulate the prostate epithelial elements, although with lower potency compared to DHT [39, 40]. Since more testosterone is now available for the 30

50 aromatase, intra-prostatic estrogen to androgen ratio was significantly increased by finasteride [65] as well, which could attenuate its overall suppressive effects on the prostate by further stimulating the stromal FRPSRQHQW$OWKRXJKFRPELQDWLRQWKHUDS\XVLQJ -reductase inhibitor and aromatase inhibitor has been proposed as a more potent method to suppress both the epithelial and stromal components, the effects have not been confirmed by large-scale placebo-controlled clinical trial. As shown in Table 1, S-KDVYHU\ORZDIILQLW\IRU -UHGXFWDVHWKHHIIHFWLYHLQKLELWLRQRIWKH - reductase couldn t be achieved in vivo. Therefore, although finasteride and S-1 shared similar potency and tissue selectivity in suppressing the prostate growth in intact male rats (Figure 3.1 and 3.3), and both are somehow related to the DHT-dependency of the prostate, they work on different targets and have completely different mechanism of action. Compared to finasteride, S-1 still showed potential advantage for BPH treatment by effectively maintaining the balance of the feedback regulation of endogenous testosterone, as reflected by the unchanged circulating testosterone level after 9 days treatment (Figure 3.2). The plasma LH concentration was even decreased after 9 days treatment with S-1 (Figure 3.2), suggesting that further decrease in circulating testosterone level could happen after longer treatment. Also, the lack of interactions between SARM and ER, SARM and aromatase (data not shown) ruled out the possible estrogenic effects by S-1 itself. Furthermore, S-ZLOOQRWLQFUHDVHWKHSURVWDWLFHVWURJHQE\LQKLELWLQJ - reductase, so stimulation of the stromal components will not be concern in SARM treatment. Other side effects for finasteride include decreased libido, decreased ejaculation and impotence [62-64], and only occur in less 10% of the patients. The effects of SARMs on libido, sexual behavior, spermatogenesis is under investigation in our laboratory. In summary, SARM (S-1) suppressed the prostate growth in intact animals with similar potency and efficacy compared to hydroxyflutamide and finasteride, but greatly improved tissue selectivity, endocrine properties and oral bioavailability (data to be published). The tissue selectivity of SARM in LQWDFWUDWVLVQRWGXHWRWKHLQWHUDFWLRQZLWKWKH -reductase or mediated by the suppression of endogenous LH and testosterone release, but its partial agonist activity in the prostate. The full agonist activity of SARM in other tissues, like muscle and the pituitary, could greatly improve the safety of the treatment. Therefore, SARM represents a novel approach with high tissue-selectivity for BPH treatment. 31

51 Type 1 (ph 7.0) Type 2 (ph 5.0) Substrates K m 0 V max (nmol/(min PJ K m 0 V max (nmol/(min PJ Testosterone 4.7 ± ± ± ± 0.5 S S Inhibitors Ki Ki Finasteride 484 ± 266 nm 23 ± 4 nm S S DEOH,QKLELWLRQRIKXPDQ -reductase isozymes expressed in transfected COS cells. 32

52 % of Vehicle-treated Intact Control Hydroxyflutamide # # Prostate S Finasteride 0.5 mg/kg 1 mg/kg 5 mg/kg 10 mg/kg 25 mg/kg Castrated % of Vehicle-treated Intact Control Hydroxyflutamide # # # Seminal Vesicle S Finasteride # Levator Ani Muscle % of Vehicle-treated Intact Control Hydroxyflutamide # # # # # # # S-1 # Finasteride Days Days Days Figure 3.1 Comparison of the tissue-selectivity of hydroxyflutamide, S-1 and finasteride in the ventral prostate, seminal vesicle and levator ani muscle in intact male rats after different treatment periods (n=5). All organ weights were normalized by body weight, and were shown as the percentage of the weights in vehicle-treated intact control group. Data are presented as means ± SD. p>0.05, compared to the vehicletreated intact control group of the same time point. # p>0.05, compared to the vehicle-treated castrated control group of the same time point. 33

53 A. Testosterone Plasma Testosterone Concentration (ng/ml) Day Day Day B. LH Plasma LH Concentration (ng/ml) Day Day Day C. FSH 3 Day 6 Day 9 Day Plasma FSH Concentration (ng/ml) Intact Castrated Hydroxyflutamide S-1 Dose (mg/kg) Finasteride 0 Intact Castrated Hydroxyflutamide S-1 Dose (mg/kg) Finasteride 0 Intact Castrated Hydroxyflutamide S-1 Dose (mg/kg) Finasteride Figure 3.2 Effects of hydroxyflutamide, S-1 and finasteride on plasma concentrations of testosterone, LH and FSH in intact male rats after different treatment periods (n=5). Intact male rats were treated with hydroxyflutamide (0.5, 1, 5, 10, 25 mg/kg), S-1 (0.5, 1, 5, 10, 25 mg/kg) and finasteride (5 mg/kg) for 3, 6, and 9 days. Vehicle-treated intact and castrated groups were also included as control. Plasma concentrations of testosterone, LH and FSH were measured at the end of each treatment period. Data are presented as means ± SD. p<0.05, compared to the intact control group of the same time point. 34

54 % of Vehicle-treated Intact Control Intact Hydroxyflutamide 120 S-1 Finasteride Prostate Levator Ani Muscle Dose (mg/kg) Figure 3.3 Comparison of the anabolic effects of hydroxyflutamide, S-1 and finasteride at doses that showed equal potency in suppressing the prostate after 9 days treatment (n=5). After 9 days treatment, hydroxyflutamide (5, 10, 25 mg/kg), S-1 (5, 10, 25 mg/kg) and finasteride (5 mg/kg) decreased the prostate weight in the intact rats to a similar extend, the anabolic effects of these treatments in the levator ani muscle were also compared. Data are presented as means ± SD. p>0.05, compared to the vehicle-treated intact control group. 35

55 CHAPTER 4 S-4 TREATMENT IMPROVES MUSCLE STRENGTH AND BODY COMPOSITION, AND PREVENTS BONE LOSS IN ORCHIDECTOMIZED RATS 4.1. Introduction Testosterone therapy results in major improvements in muscle function [66], body composition and bone mineral density [66, 67], but the major concern for testosterone therapy in elderly men is the increased risk of prostate cancer. As discussed in Chapters 2 and3, two SARMs showed strong anabolic effects in levator ani muscle without stimulating prostate growth in ORX animals, suggesting that they may serve as better alternatives for hormone replacement therapy (HRT) in men. In this study, the anabolic effects of SARM on muscle, bone, and body composition were further characterized in ORX animals. The ability of SARM to stimulate prostate growth after ORX-induced full shrinkage of the organ was also evaluated. The strong anabolic activity of SARM in muscle can be used to treat disease-related muscle wasting or improve muscle performance in hypogonadal men. Although S-4 showed stronger agonist activity in maintaining levator ani muscle weight in ORX animals[10] (Figure 2.1 and 2.2), the maintenance of the levator ani muscle weight does not provide direct evidence for the improvement in muscle performance. Therefore, the effects of SARM treatment on skeletal muscle strength in ORX animals were measured directly using isolated soleus muscle. Soleus muscle is a slow twitch muscle in the rear leg of the rat that contains mainly slow muscle fibers that are rich in myosin heavy chain I (MHC-I). The soleus muscle is generally considered as an antigravity muscle. It is also one of the most commonly used models for skeletal muscle function. 36

56 Orchidectomy significantly decreases the soleus muscle weight and strength in male rats [68], and the effect is reversed by DHT treatment [69]. In this study, the effects of S-4 (stronger anabolic reagent compared to S-1, as shown in Chapter 2) and DHT in restoring the soleus muscle mass and strength were compared in ORX animals. Muscle contractile properties, including the kinetic properties and contractile force, were measured. Peak tetanic tension (P 0 ) was used as the major parameter for muscle strength comparison. Myosin is a motor protein that interacts with actin to generate the force for muscle contraction. It is a hexameric protein consisting of two myosin heavy chain (MHC) subunits (200 kda) and two pairs of non-identical light chain subunits (17-23 kda) [70]. MHC is the most abundant myofibrillar protein expressed in muscle. The major MHC isoforms expressed in skeletal muscle include the fast fibers (i.e., MHC-IIa, IIb, and IId), and the slow fiber (i.e., MHC-I, the main form expressed in the soleus muscle) [70]. Cardiac muscle expresses mainly two slow forms: MHC- DQG MHC provides both the motor and filament-forming functions of the intact myosin molecule. Changes in whole muscle contractile force are very likely to be related to the MHC isoform expression since the contractile properties of the muscle, including shortening velocity and maximal force, are correlated with the MHC composition [71, 72]. The expression of certain MHC isoforms appears to be directly regulated by androgen, as demonstrated in cardiac MHC- H[SUHVVLRQLQVSRQWDQHRXVO\ hypertensive rats (SHR) [73, 74]. MHC isoform expression in the soleus muscle samples from different treatment groups was also compared to explore the possible mechanism of action of SARMs. Besides anabolic effects in muscle, androgen treatment can also improve body composition (i.e., increase lean mass and decrease fat mass) and prevent bone loss in hypogonadal men [75-79]. The effects of testosterone in bone may be mediated directly by AR or indirectly via aromatization of testosterone into estrogen and subsequent stimulation of the ER, as impaired skeletal development and growth was observed in aromatase inhibitor-treated male rats [80], aromatase knockout (ArKO) [81]DQG(5 NQRFNRXW (5.2PLFH[82]. Furthermore, testosterone treatment, but not estradiol treatment, prevented bone loss LQRUFKLGHFWRPL]HG(5.2PDOHPLFH[82]. The anti-resorptive effects of AR have been confirmed in AR knockout mice (ARKO) [86]. The direct action of testosterone in bone via the AR-mediated pathway is essential for its anabolic effects in bone. However, conversion of testosterone WR'+7E\ -reductase is 37

57 QRWUHTXLUHGIRUWKHSURFHVVVLQFHILQDVWHULGHD -reductase inhibitor) treatment did not affect the bone mineral density (BMD) in rats [83] or humans [84]. In this study, a nonaromatizable androgen, DHT, was used as a positive control to avoid the indirect actions of androgens through conversion to estrogen. S-4 does not interact with the ER, and is not aromatized. Thus, the effects of S-4 on bone should only be mediated by direct action on the AR, providing a valid and direct comparison to DHT. Considering the fact that both muscle and bone are DHT-independent tissues, we hypothesized that S-4 would have very similar anabolic activity to DHT in these tissues, and both treatments would improve muscle strength and body composition and restore ORX-caused bone loss. Since the decline in both muscle strength and bone remodeling are relatively slow processes [68, 85], androgen treatment was not initiated until 12 weeks after ORX to allow significant decreases in muscle strength and BMD to occur, which is different from the immediate treatment design (i.e., treatment started right after ORX) used in previous studies (Chapter 2 and 3). The effects of S-4 on body composition and bone in ORX rats were measured by dual X-ray absorptiometry (DEXA) during the study. Changes in serum markers for bone formation, such as IGF-1 and osteocalcin, were also measured after the treatment. Additionally, studies employing immediate treatment only tested the ability of SARM to maintain the androgen-dependent tissues in ORX animals, while the delayed treatment design used in this study allowed us to test the ability of SARM to restore tissue growth in ORX animals Materials and Methods Materials Compound S-4 [5] was synthesized by Dr. Duane Miller s research group at the University of Tennessee. The purities of these compounds were greater than 99%, as determined by HPLC. Polyethylene glycol 300 (PEG 300, reagent grade), dimethylsulfoxide (DMSO, reagent grade), dihydrotestosterone (DHT) and urea were purchased from Sigma Chemical Company (St Louis, MO). Insulin-like growth factor I (IGF-I) concentrations were determined using a commercially available enzyme immunoassay (EIA) kit purchased from Diagnostic Systems Laboratories Inc. (Webster, TX). Rat osteocalcin 38

58 concentrations were determined using a commercially available EIA kit purchased from Biomedical Technologies Inc. (Stoughton, MA). Acrylamide, bis-acrylamide, ammonium persulfate, sodium dodecyl sulfate, tetramethylethylenediamine (TEMED), dithiothreitol (DTT), were purchased from Bio-Rad Laboratories (Hercules, CA) Animals Male Sprague-Dawley rats were purchased from Harlan Biosciences (Indianapolis, IN). The animals were maintained on a 12-hour light-dark cycle with food and water available ad libitum. The animal protocol was reviewed and approved by the Institutional Laboratory Animal Care and Use Committee of The Ohio State University Experimental Design Male Sprague-Dawley rats (12 weeks old) were orchidectomized at the beginning of the study. A group of sham-operated male rats was also included as intact control. The ORX animals ( g) were randomly distributed into groups of 7 or 8 animals. Animals were maintained for 12 weeks after orchidectomy to allow for the maximum decrease in soleus muscle mass and strength [68, 86], and were then treated with S-4 (3 or 10 mg/kg), DHT (3 mg/kg), or vehicle for 8 weeks. The intact animals were also treated with vehicle during the treatment period. The dosage was adjusted weekly based on animals body weights. The drugs were dissolved in DMSO:PEG300 (20:80, v:v) and administered via daily subcutaneous injections. At the end of treatment, animals were weighed, anesthetized, and sacrificed within 8 hours after the last dose. The soleus muscle from the left rear leg was dissected immediately and was used for muscle strength measurements. After strength measurements, the soleus muscle was frozen in liquid nitrogen and preserved at 80 C for electrophoretic analysis of MHC isoform expression. Besides the soleus muscle, the heart sample was also frozen and preserved to examine the MHC isoform expression in the left ventricles. 39

59 The androgenic (ventral prostate and seminal vesicle) and anabolic tissues (levator ani muscle), and the soleus muscle from the right rear leg were removed and weighed. Blood samples were collected and used for the measurement of serum markers, including IGF-I and osteocalcin. (Stoughton, MA). Statistical analyses of all the parameters were performed by single-factor ANOVA with the alpha value set a priori at p< Soleus Muscle Strength Measurement The soleus muscle was isolated with care so as not to damage the muscle and its tendons, and then mounted in the experimental chamber. The muscle was perfused in oxygenated (95% O 2, 5% CO 2 ) Krebs- Ringer solution (ph at room temperature, NaCl 137 mm, KCl 5 mm, NaHCO 3 13 mm, KH 2 PO mm, CaCl 2 2 mm, MgSO 4 1 mm, and glucose 11 mm) at room temperature (20-25 C). The proximal tendon was attached to a rigid post, and the distal tendon was attached to a Kulite BG1000 transducer (Kulite Semiconductor Products, Inc., Leonia, NJ) with 4.0 silk. The muscle was stimulated using a Grass S48 Stimulator (Quincy, MA) through two plantium field electrodes attached to the chamber walls. The output from the transducer was recorded using ASI Dynamic Muscle Control and Analysis Software from Aurora Scientific Inc. (Aurora, Canada). Twitch kinetics and amplitude (P t ) were measured before the tetanus amplitude (P 0 ) was measured. A force tension curve was obtained by stimulating the muscle at supermaximal voltage (10 Volts) for 2 ms while stretching the muscle across 1 mm increment. Once the optimal length of the muscle (L 0, muscle length at which maximal twitch tension was achieved) was identified, maximal tetanic tension (P 0 ), time to peak twitch tension (tp t ), time to one half relaxation (t 1/2 R) were measured at L 0. These parameters are illustrated in Figure 4.1. Peak twitch tension was elicited with supermaximal voltage (10 Volts) as square-wave pulses of 2 ms duration delivered at 0.1 Hz frequency. Sixteen continuous twitches were recorded, and one out of every three twitches was analyzed. The average of five measurements for each parameter was reported for each muscle sample. Tetanus was evoked with 3.0-s trains of stimuli at 2 ms duration and 40 Hz frequency. Three tetani were obtained for each muscle, the average of three measurements of P 0 was reported. 40

60 After the measurements, the soleus muscle was weighed, and the cross sectional area (CSA) of the muscle was estimated using following equation [87]: CSA (mm 2 ) = muscle mass (mg) / [muscle optimal length (mm) muscle density (mg/mm 3 )] Muscle density was assumed to be 1 mg/mm 3 as previously determined in rat skeletal muscle [88]. Contractile force measurements were normalized to the CSA of the muscle prior to statistical comparison Body Composition and Bone Mineral Density (BMD) Measurement During the study, all the animals were analyzed monthly by dual X-ray absorptiometry (DEXA) using (GE, Lunar Prodigy ) using the small animal software (Lunar encore, version ). The animals were anesthetized with ketamine:xylazine (87:13 mg/kg) for the scanning. Total body bone mineral content (BMC), density (BMD), and body composition (e.g., lean mass and fat mass) were measured. DEXA analyses were completed the same day for all the animals to avoid potential errors associated with inter-day variability Electrophoretic Separation of Skeletal and Cardiac Myosin Heavy Chain (MHC) Isoforms Samples of the soleus and left ventricular muscle were homogenized (Pro200 homogenizer, Pro Scientific, Monroe, CT) for 5 to 10 seconds in sample buffer [89, 90] OSHUPJWLVVXH6DPSOHEXIIHU contained 6 M urea, 2 M thiourea, M DTT, 0.05 M Tris base, 3% SDS, and ph was adjusted to 6.8. Dissolved samples were further diluted (1:10/v:v) with sample buffer before loading on the gel. An aliquot ORIeach sample was loaded on each lane of the gel. For soleus muscle sample analysis, the stacking and separating gels (0.75 mm thick) consisted of 4 and 7 % acrylamide (wt/vol), respectively, with acrylamide:bis-acrylamide ratio of 50:1[90]. The stacking gel included 5% glycerol, and the separating gel included 30% glycerol. 2-Mercaptomethanol was added to the upper electrode buffer at a final concentration of 10 mm. Gels were run in a Hoefer SE600 unit (Hoefer Scientific, San Francisco, CA) at 8 C, with a constant voltage of 330 V for 23 hours. The ventricular sample analyses were conducted in a similar manner, except that the separating gel consisted of 6 % 41

61 acrylamide and 5 % glycerol and the gels were run at 200 V for 20 hours [89]. The gels were then fixed and silver-stained. The stained gels were analyzed using a GS 300 scanning densitometer (Hoefer Scientific) Results Anabolic Effects of S-4 on Soleus Muscle Strength in Orchidectomized Rats The body weight and soleus muscle weight of the ORX animals were significantly lower than those of the intact control animals twenty weeks after orchidectomy (Table 4.1.). Although S-4 and DHT treatment slightly increased the body weight and the soleus muscle weight in ORX animals, the changes were not significant compared to the ORX control group. When the soleus muscle weight was normalized by the body weight, no significant change was observed in any of the treatment groups. Besides the decreases observed in muscle size, the optimal length (L 0 ) of the soleus muscle (the length at which the maximal twitch tension could be achieved) also significantly decreased in ORX animals. Both S-4 and DHT treatment returned the optimal length back to that observed in intact animals. No significant difference in CSA was observed between treatment groups. Therefore, the decreases in soleus muscle weight and optimal length (L 0 ) observed in ORX animals were more likely due to the decrease in animal body weight. Neither S-4 nor DHT treatment in the ORX animals caused significant changes in the normalized soleus muscle size, corresponding to prior reports that the soleus muscle weight wass less androgen-sensitive [68] than the levator ani muscle. The peak tetanic tension (P 0 ) is often used as a measure of the contractile force of the soleus muscle [68]. Although S-4 and DHT didn t cause significant changes in muscle size, both treatments significantly increased the P 0 of the soleus muscle in ORX animals (Figure 4.2 A). The P 0 of the soleus muscle decreased from 0.85 N in intact animals to 0.57 N in ORX animals, while S-4 (3 mg/kg) and DHT (3 mg/kg) increased the P 0 of the soleus muscle to 0.86 N and 0.95 N, respectively. S-4 (10 mg/kg) increased P 0 further to 1.02 N, which was significantly higher than the P 0 observed in intact animals. Since the CSA of the soleus muscle was not changed by either ORX or any of the treatments (Table 4.1), 42

62 identical results were observed (Figure 4.2 B) when P 0 was normalized by the CSA and compared between treatment groups. As summarized in Table 4.2, other tetanus and twitch parameters including twitch tp t, tetanus and twitch t 1/2 R were not significantly between ORX, S-4 and DHT treatment groups. Different from our observations in peak tetanic tension (Figure 4.2), the peak twitch tension was not significantly changed by ORX. Only the higher dose of S-4 (10 mg/kg) significantly increased the peak twitch tension in ORX animals. In summary, ORX decreased rat body weight, soleus muscle weight, and peak tetanic tension (P 0 ). S-4 and DHT treatments significantly increased soleus muscle strength (P 0 ) of ORX animals to that of the intact animals without affecting the contraction kinetics (tp t and t 1/2 R). S-4 and DHT also increased soleus muscle mass in ORX animals, but the change was not significant due to the large variability of the data Tissue-selective Restoration of the Androgen-dependent Tissues by S-4 in Orchidectomized Rats Drug treatment was initiated immediately after ORX in our prior studies (Chapter 2). In the current study, we examined the ability of S-4 and DHT to restore androgen-dependent tissues after androgen deprivation. As such, drug treatment was initiated 12 weeks after ORX. Prolonged androgen deprivation (i.e., 20 weeks) caused significant decreases in the size of the prostate, seminal vesicle, and levator ani muscle, with these organs shrinking to 3.6%, 6.7% and 41.4%, respectively, of those observed in intact animals. Treatment with DHT (3 mg/kg) for weeks significantly increased the prostate and seminal vesicle weights by more than two fold compared to the intact animals, and increased the levator ani muscle to 131% of that observed in intact controls. S-4 (3 mg/kg) for weeks 13 through 20, selectively restored the levator ani muscle weight to that observed in intact animals, but only partially restored the prostate and seminal vesicle weights to less than 20% of that observed in intact animals, and less than 10% of that observed in DHT (3mg/kg) treated animals. S-4 (10 mg/kg) showed very similar effects in the levator ani muscle compared to the lower dose group, but stronger androgenic effects in the prostate and seminal vesicle. 43

63 These studies demonstrate that 1). S-4 can restore androgen-dependent tissue growth after prolonged androgen depletion; 2). S-4 (3 mg/kg) and DHT (3 mg/kg) demonstrate similar anabolic effects in increasing soleus muscle strength (Figure 4.2) and levator ani muscle weight (Figure 4.3), 3). S-4 demonstrates weak androgenic activity in the prostate at doses capable of maximally promoting anabolic activity Effects of S-4 on Plasma Levels of IGF-1 and Osteocalcin Besides their effects in skeletal muscle, androgens may also be anabolic in the skeleton. Plasma levels of IGF-1 and osteocalcin [80, 85, 91] levels are commonly used markers of anabolic activity and bone turnover rate. Twenty weeks after ORX, circulating IGF-1 levels in ORX animals (Figure 4.4 A) were similar to those observed in intact animals, S-4 (3 or 10 mg/kg) treatment did not affect IGF-1 levels. However, DHT (3 mg/kg) significantly decreased plasma IGF-1 concentrations to 271 ng/ml, approximately 70% of the level observed in intact animals. Plasma osteocalcin levels were not significantly different between intact and ORX animals when measured 20 weeks after ORX (Figure 4.4 B). However, eight weeks treatment with S-4 (3 or 10 mg/kg) or DHT (3 mg/kg) in ORX animals significantly decreased the plasma osteocalcin levels to about 70% and 50%, respectively, of the level in intact animals. Deceases in plasma osteocalcin concentration suggest that S-4 and DHT decreased bone turnover rate in ORX animals Effects of S-4 on the Body Composition and BMD in Orchidectomized Rats The direct effects of S-4 and DHT on the skeleton were also assessed by monthly DEXA scans. Total body bone mineral density (BMD, g/cm 2 ) and content (BMC, g) in ORX animals were significantly lower than that observed in intact animals, g/cm 2 and g, respectively (Figure 4.5 A) within 12 weeks after ORX. Animals were then treated with vehicle, S-4 (3 or 10 mg/kg) or DHT (3 mg/kg) for another 8 weeks. By the end of the treatment, the total body BMD and BMC in vehicle-treated intact animals increased by g/cm 2 and 1.45 g (Figure 4.5 B), reaching g/cm 2 and g, 44

64 respectively; while the total body BMD and BMC in the ORX animals only increased by g/cm 2 and 0.65 g (Figure 4.5 B), reaching g/cm 2 and g, respectively. S-4 (3 and 10 mg/kg) treated ORX animals showed significantly greater increases in total body BMD compared to vehicle treated ORX animals (Figure 4.5 B). The change in the BMD in S-4 treated ORX animals was similar to that observed in intact animals, with the total body BMD of both dose groups increasing to g/cm 2, and was significantly higher than that observed in vehicle-treated ORX animals. S-4 (10 mg/kg) treated ORX animals also showed significantly higher increases in total body BMC compared to the vehicle treated ORX animals, with the total body BMC increasing to g (Figure 4.5 A) and the change in BMC being similar to that observed in intact animals (Figure 4.5 A). However, changes in BMD and BMC for DHTtreated animals were smaller than that observed in intact animal and S-4-treated ORX animals, and were not significantly from the changes in ORX animals that received vehicle. Body composition of the animals (i.e., total tissue mass, fat mass, and lean mass) was also measured using DEXA. The body weights of all ORX animals were significantly lower than that observed in the intact animals (Figure 4.6 B) when measured 12 weeks after ORX and prior to drug treatment. Although there was no significant difference in total body weight between the S-4 treated and vehicletreated ORX animals by the end of the treatment (20 weeks post-orx) (Figure 4.6 B), S-4 treated ORX animals did gain more weight than the vehicle treated ORX animals during the 8-week treatment (Figure 4.6 A, Grey bars). DHT-treated ORX animals, however, showed similar body weight change during the treatment period compared to the vehicle-treated control group. The intact animals gained a similar amount of fat mass and lean mass (approximately 15 g for each) during the last 8 weeks of the study. However the vehicle-treated ORX animals lost about 6 g of lean mass while gaining the same amount of fat mass as intact controls. DHT treatment significantly increased the lean mass in ORX animals by more than 20 g in 8 weeks, and decreased the fat mass by more than 10 g. S-4 (3 or 10 mg/kg) treatment also prevented the loss of lean mass in ORX animals, and significantly increased the lean mass by 15 g in 8 weeks. However, different from DHT treatment, S-4 treatment did not decrease the fat mass in ORX animals. 45

65 Effects of S-4 on MHC Isoform Expression in Skeletal and Cardiac Muscles As described in 4.3.1, S-4 and DHT fully restored soleus muscle strength (measured as P 0, Figure 4.1) in ORX animals, only partially restored the muscle mass, suggesting that the increase in muscle strength was not simply due to an increase in muscle size. Changes in MHC expression provide another possible mechanism for the observed changes in muscle contractile force. MHC isoform expression is different in different types of skeletal muscle. Fast muscle like the extensor digitorum longus (EDL) expresses mainly MHC-IIb (fast fiber, Figure 4.7); while slow muscles like the soleus muscle expresses more MHC-I (slow fiber, Figure 4.7). The expression of MHC isoforms in both the soleus muscle sample and cardiac ventricular sample were analyzed using modified SDS-PAGE analysis [89, 90]. Similar to literature reports [71], in male Sprague-Dawley rats, two isoforms were detected in the soleus muscle samples: MHC-I and MHC-IIa. In most samples analyzed, MHC-I expression counted for more than 85% of the total MHC expressed (Figure 4.7). In intact animals, only two out of the seven samples expressed MHC-IIa, while 6 or 7 out of the 8 samples in the ORX sample expressed MHC-IIa (Figure 4.7, labeled lanes). Likewise, since no difference was observed between the S-4 and DHT-treated animals and vehicle-treated animals, the slow-to-fast shift observed in ORX animals did not seem to account for the increase in soleus muscle strength reported in section MHC- DQG0+&- DUHWKHWZRPDMRU0+&LVRIRUPVH[SUHVVHGLQFDUGLDFPXVFOH)LJXUH$ In intact animals, the expression of MHC- DFFRXQWHGIRURIWKHWRWDO0+&H[SUHVVHG)LJXUH% and androgen depletion (ORX) significantly decreased MHC- H[pression to 44%. S-4 (3 mg/kg) treatment in ORX animals increased MHC- H[SUHVVLRQWRRIWKHWRWDO0+&H[SUHVVHGZKLFKZDVVLJQLILFDQWO\ higher than that observed in the vehicle-treated ORX animals, but still significantly lower than that observed in intact animals. Both S-4 (10 mg/kg) and DHT (3 mg/kg) increased MHC- H[SUHVVLRQWRD similar level compared to that observed in intact animals. The androgen-uhjxodwhgµ WR VKLIW LQFDUGLDFPXVFOHPLJKWEHUHODWHGWRWKHIXQFWLRQRI androgen in the heart [74]. However, the slow to fast shift observed in the soleus muscle does not seem to be related to the androgen-induced increase in muscle strength in ORX animals. 46

66 4.4. Discussion Androgen treatment improves skeletal muscle performance in both animal models [69, 92] and in men [93, 94]. In our study, S-4 and DHT treatment fully restored soleus muscle strength in ORX animals (Figure 4.2) without changing the kinetics of muscle contraction (Table 4.2). Both S-4 (3 mg/kg) and DHT (3 mg/kg) were similarly potent in restoring the P 0 of the soleus muscle in ORX rats. Some clinical studies have indicated that androgen-induced increases in muscle performance in healthy men are related to muscle fiber hypertrophy [95]. However, the S-4 and DHT-induced increase in soleus muscle strength cannot be fully explained by an increase in muscle mass. Although the soleus muscle size was increased in S-4 and DHT-treated ORX animals as compared to that in vehicle-treated ORX animals, it was still smaller than that observed in intact animals (Table 4.1). Furthermore, the cross sectional area of the whole muscle (Table 4.1), estimated based on muscle mass and optimal length, was not different among any of the treatment groups. Muscle strength in S-4-treated ORX animals was similar to (S-4, 3 mg/kg) or higher (10 mg/kg) than that observed in intact animals (Figure 4.2). Thus, the change in muscle size did not correlate with the change in muscle strength, indicating that the improvement in muscle strength after androgen treatment can not be fully explained by the increase in muscle size [69]. Another possible mechanism for the increase in muscle strength that we observed in our study could have been changes in MHC expression in the soleus muscle. The functional importance of MHC in muscle contraction and direct regulation of certain MHC isoforms expression by androgen is well known [73, 74]. Castration significantly decreased MHC- H[SUHVVLRQLQWKHYHQWULFOHRI6+5[73], and the change was reversed by testosterone treatment. Similar changes were observed when rat cardiomyocytes were treated with testosterone [74]. Furthermore, computational analysis of the promoter region recognized potential ARE binding sites in both human and rat MHC- JHQHVXJJHVWLQJWKDW0+&LVRIRUPH[SUHVVLRQ could be directly regulated by androgen. Regulation of MHC- H[SUHVVLRQE\DQGURJHQZDVDOVRREVHUYHGLQRXUVWXG\)LJXUH6-4 and DHT treatment increased MHC- H[SUHVVLRQLQWKHKHDUWVRI25;DQLPDOs. However, no significant change in MHC isoform expression was observed in the soleus muscle samples (Figure 4.7). Although more animals expressed MHC-IIa in the ORX group, S-4 and DHT treatments did not reverse the change, 47

67 which suggests that the increase in MHC-IIa expression may not be related to androgen regulation. As a slow antigravity muscle, the soleus muscle is more sensitive to changes in gravity. Spaceflight causes a significant slow-to-fast shift in MHC isoform expression in rat soleus muscle (i.e., higher percentage of MHC-IIa and lower percentage of MHC-I) [96-98]. Thus, shift in MHC expression observed in our experiment could have been due to the significantly lower body weight (30-50 g less compared to the intact control group, Table 4.1) in ORX animals. As androgens did not reduce MHC-IIa expression in the soleus muscle, the change in whole muscle strength that we observed was not related to MHC isoform expression. Previous studies by Gentile et al [69] showed similar changes in soleus muscle strength in ORX animals treated with DHT. These earlier studies revealed that DHT treatment stimulated remodeling of the neuromuscular junction by regulating gene expression. Likewise, animal studies [99] also observed that testosterone treatment in male rats decreased diaphragm neuromuscular transmission failure as well. Therefore, the anabolic effects of S-4 treatment in improving muscle strength could also be related to changes in neuromuscular junction remodeling, which needs to be further investigated in future studies. Besides their strong anabolic effects in skeletal muscle, S-4 and DHT also improved body composition (Figure 4.6) in the ORX animals. Interestingly enough, these two processes might not be completely unrelated [100, 101]. Many clinical studies have shown that testosterone treatment causes reciprocal changes in muscle and fat mass [76, 78, 93, ]. Similar results were observed with the DHT-treated ORX animals (Figure 4.6 A) in our study. However, S-4 treatment only increased the muscle mass without changing the fat mass in the ORX animals (Figure 4.6). Recent studies [100, 101] using pluripotent, mesenchymal C3H 10T1/2 cells that are capable of differentiating into muscle, fat, cartilage, and bone cells, a model widely used to study the regulation of myogenic and adipogenic lineage determination, showed that testosterone and DHT promoted the differentiation of these cells to myogenic lineage and inhibited their differentiation into the adipogenic lineage by up-regulating MyoD and MHC (markers for myogenic differentiation) expression and down-regulating PPAR- -DQG&(%3 PDUNHUV for adipogenic differentiation) expression, respectively. Since both testosterone and DHT showed similar effects in both processes, the regulation is believed to be mediated by AR. Gene expression profiling performed in our laboratory (Chapter 7) revealed differential regulation of gene expression by S-4 and DHT in a prostate cancer cell line. Therefore, the lack of reciprocal changes in muscle and fast mass in S-4-48

68 treated ORX animals could be related to the different effects of S-4 and DHT in regulating the expression of the gene markers that are responsible for the differentiation of the mesenchymal cells. More detailed gene expression experiments need to be conducted to confirm this hypothesis. Another important target organ of androgen is bone. In the mature adult, bone undergoes a continuous remodeling process; a process consisting of new bone formation by osteoblasts and bone resorption by osteoclasts [84]. The remodeling process is also regulated by mechanical factors, systemic hormones (sex steroids, parathyroid hormone, growth hormone, etc.), and locally produced factors (cytokines, growth factors). Orchidectomy or ovariectomy increases bone turnover rate in animals. Both estrogen and androgen treatment in ORX animals has been shown to have anti-resorptive effects in bone by decreasing the bone remodeling turnover rate [20, 23, 84]. Although both S-4 and DHT treatments prevented bone loss (Figure 4.5) in the ORX animals, S-4 and DHT treatments showed significant differences in some parameters, indicating possible differences in their mechanism of action. DHT was very effective in improving body composition and muscle strength in ORX animals. However, it was not as effective in restoring ORX-induced bone loss in these animals (Figure 4.5). S-4 was more potent than DHT in preventing bone loss in the ORX animals. Recent studies have shown that ER and AR can act through both genomic (i.e., gene expression regulation) and nongenomic pathways (i.e., cross-talk with other signaling pathway through direct protein-protein interaction without directly regulating gene expression) [23] in regulating bone remodeling. The non-genotropic effects appear to be very important to the anti-apoptotic activities of estrogen and androgen in osteoblast [20, 23]. More importantly, different ER ligands showed varying ability to stimulate genomic or nongenomic pathways [20]. The fact that S-4 is more potent in restoring ORX-induced bone loss could be related to a difference in potency of S-4 and DHT in stimulating the AR non-genomic pathway in bone. In mature rats, ORX-induced bone loss is associated with increased bone turnover in the first few months after ORX, followed by a lower turnover state [85]. Osteocalcin is a noncollageous protein associated with the mineralized matrix, and is accepted as a highly specific osteoblastic marker for bone formation. Plasma osteocalcin levels are thought to reflect changes in bone turnover [85, 105]. In our study, the plasma osteocalcin level in ORX animals was similar to that observed in intact animals, suggesting that animals had transitioned to a lower bone turnover rate within five months after ORX (Figure 4.4 B). Both 49

69 S-4 and DHT treatments further decreased the plasma osteocalcin levels in ORX animals, reflecting an even lower turnover rate in these animals. Changes observed in DHT-treated ORX animals were similar to the observations in previous studies [85, 91]. These results suggest that S-4 might have anti-resorptive activity as well. IGF-1 is a growth factor that increases bone turnover rate by stimulating osteoblast proliferation and osteoclast differentiation, with a net increase in bone accumulation [106]. The effects of IGF-1 in bone are more related to the local concentration of IGF-1, which is related to both circulating IGF-1 and tissuespecific expression of IGF-1 and IGFBP proteins [106]. Although circulating IGF-1 is mainly released from liver [107], which may not reflect the tissue concentration of IGF-1 in the bone, changes in plasma IGF-1 could still reflect the effects of S-4 and DHT on IGF-1 and IGFBP expression. S-4 treatment tends to increase IGF-1 expression in ORX animals (Figure 4.4 A), while DHT significantly decreased IGF-1 expression in these animals, providing another example for possible differential regulation of gene expression by S-4 and DHT. Furthermore, gene expression profiling in LNCaP cells also revealed possible differential regulation of another important player of the IGF-1 signaling pathway, IRS-1, by S-4 and DHT treatment. The possible differences in S-4 and DHT regulated gene expression could contribute to the tissue-specific pharmacological activities of these AR ligands. In summary, S-4 treatment greatly improved the muscle strength and body composition, and restored lost bone in ORX rats. The anabolic effects of S-4 in muscle, bone, and body composition were very similar to those observed in DHT-treated ORX animals. However, at an equipotent dose that induced similar changes in bone and muscle (3 mg/kg dose), S-4 only restored the prostate growth to less than 10% of the level observed in DHT treated animals, showing minimum stimulation of the prostate compared to DHT treatment. Significant differences between S-4 and DHT treatments in regulating fat mass change and bone turnover rates in ORX animals were also observed, suggesting that possible differences in the mechanism of action of S-4 and DHT could exist even though they both work through AR-mediated pathways. 50

70 A. Tetanus P 0 B. Twitch P t TPT t 1/2 R Figure 4.1 Twitch and tetanus measurements in rat soleus muscle. A. Maximal tetanic tension (P 0 ) was measured at the plateau of the tetanus. B. Maximal twitch tension (P t ), time to peak twitch tension (TPT), and time to one half relaxation (t 1/2 R) were measured at the optimal length (L 0 ) of the muscle. 51

71 Intact + Veh ORX + Veh ORX + S-4 ORX + S-4 ORX + DHT (3 mg/kg) (10 mg/kg) (3 mg/kg) Body Weight (g) 437 ± ± ± ± ± 14 Soleus Muscle (mg) 161 ± ± ± ± ± 23 Soleus Muscle Weight / Body Weight (mg/g) 0.37 ± ± ± ± ± 0.05 Soleus Muscle L 0 (mm) 28 ± 1 25 ± 2 27 ± 2 # 28 ± 1 # 27 ± 1 # Soleus Muscle CSA (mm 2 ) 7.0 ± ± ± ± ± 0.6 Table 4.1 Body weight, soleus muscle weight, optimal length (L 0 ), cross sectional area (CSA), and soleus muscle weight to body weight ratio (n=7-8) in different treatment groups. Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), and DHT (3 mg/kg) for 8 weeks. Data are presented as mean ± SD. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle treated ORX group. 52

72 A. Peak Tetanic Tension (P 0 ) # # # 1.0 P 0 (N) B. P 0 /CSA # # # 1.4 P 0 /CSA (kn/m 2 ) Intact Veh S-4 ( 3mg/kg) S-4 (10 mg/kg) DHT (3 mg/kg) ORX Figure 4.2 Peak tetanic tension (P 0 ) and normalized P 0 /CSA in different treatment groups (n=7-8). Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for 8 weeks. Data are presented as mean ± SD. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle-treated ORX group. 53

73 Intact + Veh ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) Tetanus P 0 (N) 0.85 ± ± ± 0.25 # 1.02 ± 0.17 # 0.95 ± 0.21 # P 0 /CSA (kn/m 2 ) 1.24 ± ± ± 0.37 # 1.56 ± 0.22 # 1.39 ± 0.31 # t 1/2 R (ms) 256 ± ± ± ± ± 63 Twitch P t (N) 0.21 ± ± ± ± 0.06 # 0.22 ± 0.05 P t /CSA (kn/m 2 ) 0.30 ± ± ± ± 0.10 # 0.33 ± 0.08 TPT (ms) 160 ± ± ± ± ± 17 t 1/2 R (ms) 222 ± ± ± ± ± 49 Table 4.2 Contractile properties of the soleus muscle (n=7-8) in different treatment groups. Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for 8 weeks. The contractile properties of the soleus muscle were determined after treatment. Data are presented as mean ± SD. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle-treated ORX group. 54

74 % of Vehicle-treated Intact Control # # # Intact ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) # # Intact ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) # # # Intact ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) # Prostate Seminal Vesicle Levator Ani Muscle Figure 4.3 Normalized prostate, seminal vesicle and levator ani muscle weights (n=7-8) in different treatment groups. Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for 8 weeks. All organ weights were normalized by body weight, and are shown as the percentage of the weights in vehicle-treated intact control group. Data are presented as mean ± SD. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle treated ORX group. 55

75 A. IGF Plasma IGF-I (ng/ml) # B. Osteocalcin Plasma Osteocalcin (ng/ml) # # # 10 0 Intact Veh S-4 ( 3mg/kg) S-4 (10 mg/kg) DHT (3 mg/kg) ORX Figure 4.4 Plasma IGF-1 and osteocalcin levels (n=7-8) in different treatment groups. Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for 8 weeks. Plasma IGF- 1 and osteocalcin levels were measured using EIA kits. Data are presented as mean ± SD. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle treated ORX group. 56

76 A. Total Body BMD B. Change in BMD Total Body BMD (g/cm 2 ) # # Change in Total Body BMD (g/cm 2 ) # # Total Body BMC 1.8 Change in BMC Total Body BMC (g) # Change in Total Body BMC (g) # Intact ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) 12 weeks post-orx (Pre-treatment) Intact ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) 20 weeks post-orx (Post-treatment) Intact ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) Figure 4.5 Total body BMD and BMC (n=7-8) in different treatment groups. A). Total body BMD and BMC measured at 12 weeks (before S-4 and DHT treatment) and 20 weeks (after S-4 and DHT treatment). B). Changes in total body BMD and BMC during 8 weeks treatment with S-4 and DHT (between 12 and 20 weeks post-orx). Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for 8 weeks. Total body BMD and BMC were measured by DEXA before and after the treatment. Data are presented as mean ± SEM. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle treated ORX group. 57

77 A. Body Composition Change Changes in Body Composition (g) # # # # # Fat Mass Lean Mass Total Body Weight -20 Intact Veh S-4 (3 mg/kg) S-4 (10 mg/kg) # DHT (3 mg/kg) ORX B. Body Weight Change Total Body Weight (g) Intact ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) weeks 16 weeks 20 weeks Figure 4.6 Body composition and body weight change (n=7-8) in different treatment groups during 8 weeks treatment. A). Body composition change between 12 weeks (before S-4 and DHT treatment) and 20 weeks (after S-4 and DHT treatment) post-orx. B). Animal body weight measured during 8 weeks treatment with S-4 and DHT (between 12 and 20 weeks post-orx). Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for another 8 weeks. Body composition was measured by DEXA before and after the treatment. Data are presented as mean ± SEM. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle treated ORX group. 58

78 MHC Isoform Expression in Soleus Muscle MHC-IIa MHC-IId MHC-IIb MHC-I Ctrl Intact ORX + Veh ORX + S-4 (3 mg/kg) MHC-IIa MHC-I ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) Figure 4.7 Electrophoretic separation of the MHC isoforms expressed in soleus muscle samples from (n=7-8) different treatment groups. Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for another 8 weeks. Soleus muscle strength was measured at the end of the treatment, and the soleus muscle samples were homogenized after the measurement and subjected to electrophoretic separation (7% acrylamide gel, with 30% glycerol) of the MHC isoforms. The gels were fixed and silver-stained. A mixed control sample (Ctrl) containing both EDL and diaphragm was included to show the separation of the MHC isoforms. Indicates the presence of MHC-IIa isoform. 59

79 A. Electrophoretic Separation of MHC Isoforms in Ventricular Samples Intact ORX + Veh ORX + Veh ORX + S-4 (3 mg/kg) ORX + S-4 (10 mg/kg) MHC- MHC- MHC- MHC- MHC- MHC- ORX + S-4 (10 mg/kg) ORX + DHT (3 mg/kg) B. Cardiac MHC-alpha Expression 100 Cardiac MHC-alpha Expression (% of total MHC) Intact Veh # S-4 (3 mg/kg) # S-4 (10 mg/kg) # DHT (3 mg/kg) ORX Figure 4.8 Electrophoretic separation of the MHC- DQG0HC- LVRIRUPVH[SUHVVHGLQWKHYHQWULFXODU muscle samples from (n=7-8) different treatment groups. Twelve weeks after ORX, animals were treated with vehicle, S-4 (3 or 10 mg/kg), or DHT (3 mg/kg) for another 8 weeks. The ventricular muscle samples were homogenized and subjected to electrophoretic separation (6% acrylamide gel, with 5% glycerol) of MHC isoforms. The gels were fixed and silver-stained, and the relative expression levels (% of total MHC expressed) of the MHC- DQG0+&- LVRIRUPVZHUHPHDVXUHd by densitometer. Data were presented as mean ± SD. p< 0.05, compared to the intact control group. # p< 0.05, compared to the vehicle-treated ORX group. 60

80 CHAPTER 5 SPECIES DIFFERENCE IN THE METABOLISM OF S Introduction Our previous studies (Chapter 3) have shown that S-4 could not be metabolized in AR target WLVVXHVE\HLWKHU -reductase or aromatase. However, pharmacokinetic study of S-4 [31] showed that it had a short half life of 3-4 hours in rat, suggesting that S-4 still underwent extensively hepatic metabolism. We previously improved the metabolic stability of a thio-ether linked SARM, increasing the in vivo half life from less than one hour [9] to 3-4 hours in rat [31], by structural modification to an ether linkage (S-4). However, the ideal half life required for once a day administration, as proposed by Negro-Vilar [2], is yet to be achieved. Better understanding of the hepatic metabolism profile of S-4 will be helpful to guide additional structural modification. Also, the formation of active metabolites in vivo represents another possible mechanism for the observed tissue selectivity of SARMs. As a matter of fact, the pharmacological activity of non-steroidal AR antagonist, flutamide, is mediated by its active metabolite, hydroxyflutamide [108] (Figure 5.1 A). Although S-LVQRWDVXEVWUDWHIRUHLWKHU -reductase or aromatase, it is unclear whether any active hepatic metabolite(s) contribute to the pharmacologic activity of S-4. The studies described herein were designed to address these issues. Both flutamide and bicalutamide are structural analogs of S-4. Hydrolysis of the amide bond and oxidation of the aromatic rings are the major phase I metabolism pathways for both flutamide and bicalutamide (Figure 5.1). Species differences were observed in bicalutamide metabolism [109, 110] (Figure 5.1 B). Hydrolysis of the amide bond of bicalutamide was only observed in rats, but not in humans. 61

81 The metabolic profile and possible species differences in the metabolism of S-4 were compared using recombinant human cytochrome P450 (CYP) enzymes and pooled liver enzyme preparations from rats, dogs and humans. The most commonly used liver enzyme preparations include liver microsomes, cytosolic fractions, and S9 fractions [111]. Although fresh human liver hepatocytes can also be purchased from commercial sources, they are very expensive and not always available. We chose to use the subcellular systems mentioned above for our studies, because they are much easier to store and more convenient to use. The hepatic metabolizing enzymes include phase I (e.g., hydrolysis, reduction and oxidation) and phase II (e.g., conjugations including glucuronidation, sulfation, and acetylation) enzymes [112]. Different subcellular fractions contain different enzymes. Microsomal preparations contain most of the CYP enzymes, cytosolic fractions contain many phase II enzymes like N-actyltransferase (NAT) but not the CYPs. Liver S9 fractions contain both microsomal and cytosolic fractions. Another advantage of these liver enzyme preparations is that they allow us easily compare drug metabolism between different species, and help choose the most appropriate animal model for further in vivo studies. In many cases, the pooled liver samples are used to obtain general information regarding drug metabolism. However, individual human samples can be used as well to study the possible influence of genetic polymorphisms. On one hand, recombinant human CYP enzymes prepared from stably transfected mammalian cells or baculovirus-infected insect cells are commercially available and provide a valuable approach to identify metabolites and CYP pathways for metabolism, without concerns related to heterogeneity or enzyme interactions. However, the recombinant CYP enzymes cannot be used in studies related to multiple enzymes due to the lack of correlation with the in vivo system. On the other hand, the pooled liver enzyme preparations are mixtures of many different enzymes, are easily prepared, and provide better correlation with the in vivo system. In our studies, recombinant CYP enzymes were used to identify metabolite pathways, while the pooled liver enzyme preparations were used to study possible interactions between the enzymes. 62

82 5.2 Materials and Methods Materials 14 C-S-4 was synthesized by Dr. Michael Darby. Compounds S-4 and M1 were synthesized by Dr. Duane Miller s research group at the University of Tennessee. Recombinant human CYP enzyme (Supersome ), recombinant human N-acetyl transferase (NAT) 1 and 2, pooled human, rat, and dog liver microsome, cytosol, and S9 preparations were purchased from BD Gentest (Woburn, MA). 4 -Hydroxydiclofenac, 4 -hydroxy-mephenytoin, mephenytoin, 1 -hydroxy-bufuralol, and bufuralol were also purchased from BD Gentest. EcoLite (+) scintillation cocktail was purchased from ICN Research Products Division (Costa Mesa, CA). All other chemicals and reagents were purchased from Sigma Chemical Company (St Louis, MO). All analytical columns were purchased from Waters Corporation (Milford, MA) Animals Male Sprague-Dawley rats (about 250 g) were purchased from Harlan Biosciences (Indianapolis, IN). Beagle dogs (about 9 kg) were purchased from an approved vendor through University Laboraty Animal Resources at The Ohio State Univeristy. The animals were maintained on a 12-hour light-dark cycle with food and water available ad libitum. The animal protocol was reviewed and approved by the Institutional Laboratory Animal Care and Use Committee of The Ohio State University In Vitro Metabolism Reaction Using Different Liver Enzyme Preparations In vitro enzyme reactions were conducted according to the instructions provided by BD Gentest. All phase I reactions using Supersome, liver microsomes, cytosol, or liver S9 preparations, were conducted at 37 C with the presence of 1 mm NADPH (Nicotinamide adenine dinucleotide phosphate, reduced form) in 100 mm phosphate buffer (ph 7.4) for various times. The reaction time, substrate and 63

83 enzyme concentrations are noted in figure legends. In general, incubation time was 1 to 2 hours for metabolite identification, while incubation time for the determination of enzyme kinetics parameters was much shorter, ranging from 5 to 30 minutes. The kinetic parameters, Km and Vmax, were determined by nonlinear regression analysis using WinNonlin (version 4.0, Pharsight Corporation, Mountain View, CA) and the sigmoidal Emax model. Reactions were stopped by adding ice-cold acetonitrile (v:v/1:1) containing internal standard (Table 5.1) for HPLC analysis. Protein present in the reaction mixture was precipitated by centrifugation (> 16,000 g, 30 min at 4 C), and the supernatant was either diluted with appropriate mobile phase or directly used for HPLC analysis. Enzyme activity in different preparations was also verified using the specific substrates as listed in Table 5.1. HPLC methods for the CYP enzyme specific substrates and their metabolites are listed in section Recombinant human NAT1 and NAT2 probe substrates and their metabolites were separated on a reversed-phase column (NovaPak C18, mm) with a mobile phase of 10% acetonitrile, 1% acetic acid, and 3.6 mm triethylamine in deionized water, at a flow rate of 1.5 ml/min, and were detected by their UV absorbance at 254 nm. Different from the CYP enzyme reaction, NAT1 and NAT2 reaction buffer contained 50 mm triethanolomine (ph 7.5), 1 mm EDTA, 1 mm DTT, 0.1 mm acetyl-coa, and an acetyl- CoA re-generating system (4.6 mm acetyl-d.l-carnitine and 0.06 units of carnitine acetyl transferase). For some mixed reactions that contained both microsomal enzyme and NAT, 100 mm phosphate buffer (ph 7.4) was used. NADPH and/or acetyl-coa (including the re-generating system) was added according to experimental design. The disappearance of S-4 or the appearance of the metabolites was determined using HPLC analysis. S-4 and M1 were separated on a reversed-phase column (Symmetry C8, mm) with a mobile phase of 45% acetonitrile and 50 mm phosphate buffer (ph 4.8) in deionized water, at a flow rate of 1.0 ml/min, and were detected by their UV absorbance at 230 nm. A structural analog of S-4, CM-II-87, was used as internal standard for the HPLC analysis of S-4. 64

84 5.2.4 Identification of the Phase I Metabolite of S-4 14 C-S- 0ZDVLQFXEDWHGZLWKKXPDQOLYHUPLFURVRPH+/0PJPODWƒ&IRUKRXUV After precipitation of the microsomal protein with acetonitrile (v:v/1:1), the supernatant of the reaction mixture was used for HPLC analysis. HPLC conditions similar to that described above were used, except that 50 mm phosphate buffer was not included in the mobile phase. The eluted fractions (30 seconds per fraction) from the HPLC were collected, and the total radioactivity (DPM) in each fraction and the resuspended microsomal protein was counted in EcoLite (+) scintillation cocktail. A similar experiment was conducted using non-radiolabeled S-4. In this case, the eluted factions containing possible metabolites were analyzed using negative-ion electrospray ionization (ESI - ) mass spectrometry (ThermoFinnigan LCQ DECA ion trap mass spectrometer, San Jose, CA) as previously described [113]. The samples were injected directly through the syringe pump. For the MS system, the heated capillary temperature was set at 280 C, spray voltage was 3.5 kv, and the sheath gas and auxiliary gas flow rate were 96 and 56 ml/min, respectively. All other parameters were set to the optimized conditions for ionization and detection of S-4. Data acquisition was controlled by Xcalibur software (Revision B, ThermoQuest Corp., San Jose, CA). The data was collected in full-scan negative ion mode at a range of m/z Covalent Binding of 14 C-S-4 Metabolites to Human Liver Microsomal Protein In vitro metabolism reactions using recombinant human CYP enzymes were performed with 14 C- S-4. The microsomal proteins were precipitated by the addition of acetonitrile (v:v/1:1). The precipitated protein pellets were washed three times with 50% acetonitrile in deionized water and resuspended in EcoLite (+) scintillation cocktail. The total radioactivity present in the protein pellets was counted. For the measurement of non-specific binding, microsomal protein was denatured before being added to the reaction. Since the recombinant human CYP enzymes were expressed in insect cells (Sf9 cells), the microsomal protein prepared from control insect cells (IC) was also included as a negative control. 65

85 5.2.6 Pharmacokinetic Studies of S-4 and Its Primary Metabolite The pharmacokinetic profile of S-4 in rats was determined in our previous study [31]. In vitro and in vivo studies demonstrated that deacetylation to an amine metabolite (M1) represented the primary metabolic conversion of S-4. After the identification of the species difference in S-4 metabolism, the pharmacokinetic profiles of S-4 and the deacetylated metabolite M1 in rat (n=5) and dog (n=3) were determined as well. In rats, the right jugular vein was catheterized 18 hours before dosing. In dogs, the saphenous vein catheter was implanted immediately prior to dosing. Intravenous (i.v.) doses (10 mg/kg) were administered via the catheters. Blood samples were withdrawn from the catheters before and after each dose. For the 10 PJNJGRVHRI0LQUDWVEORRGVDPSOHV OHDFKZHUHFROOHFWHGDW and 1440 min after the i.v. dose. For the 10 mg/kg dose of S-4 or M1 in dogs, blood samples (3 ml each) were collected at 2, 5, 10, 15, 30, 60, 120, 240, 480, 720, 1440 min after the i.v. dose. All dogs were treated with S-4 (10 mg/kg) during the first week of the study. One week later, M1 (10 mg/kg) was administered to the same group of animals. Plasma samples were separated by centrifugation (3000 rpm, 20 min at 4 C). Plasma proteins were precipitated by the addition of acetonitrile (v:v/1:1, containing internal standard, CM-II-87). Protein pellets were separated by centrifugation (>16,000 g, 30 min at 4 C), and the supernatant was used for HPLC analysis. The HPLC method for the analysis of S-4 and M1 was the same as that described in section The interday variability was less than 8% at the lower limit of quantitation (0 JPO Compartmental Analysis of the Concentration versus Time Profiles of S-4 and M1 A variety of compartmental models were employed in order to identify the model that bestdescribe the pharmacokinetic data. The goodness-of-fit between models was compared using the AIC, observed variability in the estimated PK parameters, and visual inspection of plots of residual differences between actual and computer-estimated concentration-time profiles. The models presented below represented those which best represented the data for each study. 66

86 Plasma concentration-time profile of S-4 in the rat after an i.v. dose of S-4 were computer-fitted individually using a standard one-compartment model in WinNonlin. The volume of distribution (V d1 ) and the elimination rate constant (k 10 ) of S-4 were estimated as model parameters using a weighting scheme of 1/Y. The S-4 and M1 data from each rat after an i.v. dose of M1 were computer-fitted using a userdefined model (Figure 5.2 A) and WinNonlin software. The plasma concentrations of S-4 and M1 were best described by the following equations: Cp = Dose V d 2 e ( k a + k m ) t (Equation 5.1) for M1; Dose ka k t ( k k ) t ( 10 e e a + m Cp = ) (Equation 5.2) for S-4. V ( k + k k ) d1 a m 10 The mean values of k 10 ( min -1 ) and V d1 (112 ml) determined from the S-4 data after an i.v. dose of S-4 were used as fixed constants in Equation 5.2. This was necessary because the studies were conducted in separate animal groups. Both M1 and S-4 data were fitted simultaneously using Equations 5.1 and 5.2 to determine constants A, C, and k, which were defined as: Dose A = (Equation 5.3) V d 2 Dose k C = a (Equation 5.4) V ( k + k k10 ) d1 a m k = k a + k m (Equation 5.5) After the three constants were determined by simultaneous fitting of the S-4 and M1 data, the secondary parameters ka, km, and Vd2 were determined using Equations 5.3 through 5.5. Other secondary parameters like Cl and AUC were determined by: Cl m = km Vd 2 (Equation 5.6) A AUC = for M1, or k AUC m C C = for S-4. (Equation 5.7) k k 10 The fraction of M1 that was converted to S-4 (f m ) was determined by: 67

87 f m AUC = m (Equation 5.8) AUC + AUCm Since the S-4 and M1 pharmacokinetics studies were completed in a cross-over design with the same group of dogs, the three sets of pharmacokinetic data were fitted simultaneously using the twocompartment model proposed in Figure 5.2 B. The plasma concentrations of S-4 after an i.v. dose of S-4 were best described by: Cp = α t A e β t + B e (Equation 5.9) ZLWK$% DQG GHILQHGDV Dose ( k + α) = 21 k A h and V ( β α) α + β = k α β = k 12 d k k h h + k + k k + k k 10 k h Dose ( β k ) = 21 k B h (Equation 5.10) V ( β α) d1 (Equation 5.11) A B AUC + α β = (Equation 5.12) The plasma concentrations of M1 after an i.v. dose of M1 were best described by: Cp = α t A e β t + B e (Equation 5.13) ZLWK$ % DQG GHILQHGDV Dose ( k α ) A = 21 and V ( β α ) α β = k d1 α + β = k k k m 12 + k m Dose ( β k ) B = 21 (Equation 5.14) V ( β α ) d1 (Equation 5.15) Separate fitting of the S-4 data after an i.v. dose of S-4 and the M1 data after an i.v. dose of M1 VHSDUDWHO\UHYHDOHGWKDW DQG YDOXHVZHUHYHU\FORVH7KXVWKHSODVPDFRQFHQWUDWLRQVRI0DIWHUDQ i.v. dose of S-4 were described by a simplified equation: β t Cp = C ( e β t α t α t e ) + D ( e e ) (Equation 5.16) with C and D defined as: 68

88 Dose k = 12 k21 k C h ; V ( β β ) ( α β ) ( α β ) d1 Dose k = 12 k21 k D h (Equation 5.17) V ( α α ) ( α β ) ( α β ) d1 AUC m = C ( ) + D ( ) (Equation 5.18) β β α α Equations 5.10, 5.14, and 5.16 were used to fit the three sets of data from each dog simultaneously. The fraction of S-4 that was deacetylated to M1 (fm) was determined using Equation Results Identification of the Phase I Metabolites and Metabolic Pathways of S-4 Using Human Liver Microsomes (HLM) and Recombinant Human CYP Enzymes After incubating 14 C-S- 0ZLWKKXPDQOLYHUPLFURVRPHV+/0DWƒ&IRUWZRKRXUVWKH metabolites and S-4 were separated on HPLC (Figure 5.3 A). About 21% of 14 C-S-4 (unchanged parent drug) remained in the reaction after two hours (Figure 5.3 B), two other fractions (M1 and M2) that accounted for about 40% of the total radioactivity were also identified during HPLC. About 30% of the total radioactivity was found in the protein pellets. The same experiment was repeated with non-radiolabeled S-4. The fraction corresponding to the M1 fraction (Figure 5.3 A) was collected and subjected to MS analysis. Only one metabolite, identified as the B-ring deacetylated product (m/z 398) was identified in this fraction (Figure 5.4). The identity of M1 was confirmed by comparison of the MS spectra, to the MS profile of the standard compound synthesized by Dr. Miller s group. The other fraction, M2 was highly hydrophilic. Under most of the separation conditions we tested, this faction co-eluted with the void peak. Due to the high salt content of the fractions eluting within the first 3 min of the chromatogram, LC-MS analysis of this sample became very difficult. Considering the high hydrophilicity of the metabolite, and the known hydrolysis that occur during flutamide and bicalutamide metabolism, we speculate that the M2 fraction could be the hydrolysis product resulting from cleavage of the amide bond of S-4 (Figure 5.4). However, difficulties that were encountered during organic synthesis of 69

89 these theoretical metabolites precluded development of an LC/MS analytical method or confirmation of their existence in our incubates. After the identification of the deacetylation product, the major CYP enzymes responsible for S-4 metabolism were identified using recombinant human CYP enzymes (Figure 5.5 A). By measuring the disappearance of S-4, CYP3A4 was identified as the major CYP enzyme that was responsible for S-4 PHWDEROLVPDW 0)LJXUH$OHIWSDQHO+RZHYHUDW 06-4 was deacetylated to M1 by multiple CYP enzymes, including CYP 3A4, 1A2, 2C19 and 2D6 (Figure 5.5 A, right panel). Among the five enzymes tested, only CYP2C9 showed no activity towards S-4. The ability of different CYP enzymes to catalyze S-4 deacetylation was estimated by measuring M1 formation rate (Figure 5.5 B). The activities RI&<3V$&DQG'ZHUHRQO\GHWHFWHGDWFRQFHQWUDWLRQVKLJKHUWKDQ 0$WWKHORZHU concentrations tested, only CYP3A4 showed detectable activity for S-4 deacetylation. At clinical relevant doses, the plasma concentration of S-LQKXPDQLVDURXQG 0GDWDQRWVKRZQ7KHUHIRUH&<3$ is most likely to contribute to the in vivo metabolism of S Characterization of the Kinetics of Phase I Metabolism of S-4 The disappearance of S-4 was determined as an initial measure to estimate the enzyme kinetics parameters of CYP3A4 (Figure 5.6 A). S-4 showed similar affinity to CYP3A4 ( as testosterone (13 0EXWDORZHU9PD[Smole/(pmolemin)) compared to testosterone (7.6 pmole/(pmolemin), data not shown). After authentic M1 was synthesized, the formation rate of M1 was used to estimate the kinetic parameters of CYP3A4-mediated deacetylation. As shown in Figure 5.5 B, the CYP3A4 catalyzed GHDFHW\ODWLRQUHDFWLRQSUHGRPLQDWHGDWORZHUFRQFHQWUDWLRQVZLWKD.PRI 0DQG9PD[RI pmole/(pmolemin). However, the kinetic parameters of CYP3A4-mediated deacetylation of S-4 in HLM preparation (Figure 5.6 B) varied significantly from those observed in recombinant enzyme preparations, ZLWK.PRI 0DQG9PD[RISPROHPJSURWHLQPLQ7KHGLIIHUHQFHLQNLQHWLFSDUDPHWHUV (i.e., Km and Vmax) was most likely, due to the presence of multiple CYP enzymes in the HLM, since other CYP enzymes could also deacetylate S-4 at high concentrations (Figure 5.5 B). 70

90 5.3.3 Covalent Binding of S-4 Metabolites to Human Liver Microsomal Protein As shown in Figure 5.3, 14 C-S-4 and its metabolites showed covalent binding to human microsomal proteins after 2 hours incubation. S-4 was uniformly labeled with 14 C on the B-ring (Figure 5.4). Thus, both 14 C-S-4 and some metabolites could potentially form a covalent bond with microsomal proteins during the incubation. 14 C-S- 0ZDVLQFXEDWHGZLWKGLIIHUHQW&<3HQ]\PHVZLWKRUZLWKRXW NADPH (Figure 5.7), a co-factor required for phase I reaction, to determine which CYPs were involved in covalent bond formation. When 14 C-S-4 was incubated with denatured microsomal protein (from liver enzyme preparation, or from insect control sample (IC)), non-specific binding was less than 0.5% of the total activity used. On the other hand, when 14 C-S-4 was incubated with IC, no radioactivity was recovered from the protein pellets no matter whether NADPH was included or not. This suggested that 14 C-S-4,by itself, could not form covalent bond non-specifically with the microsomal protein present in the Supersome preparation. Similarly, no covalent binding was observed when 14 C-S-4 was incubated with individual CYP enzymes without NADPH (no metabolism), suggesting that 14 C-S-4 did not covalently bind to CYP enzymes either. However, when 14 C-S- 0ZDVLQFXEDWHGZLWK&<3HQ]\PHVLQWKHSUHVHQFHRI1$'3+ to 3% of the total radioactivity added was found in the protein pellets of samples containing CYP3A4 and CYP2C19. Since CYP3A4 catalyzed deacetylation of S-4 at this concentration (Figure 5.5), the results suggested that M1 might covalently bind to CYP3A4. When other CYP enzymes (CYP 1A2, 2C9, 2D6), which showed low activity in S-4 deacetylation at similar concentrations (Figure 5.5), were co-incubated with CYP3A4 in the presence of NADPH, the protein bound radioactivity increased dramatically to 8-19% of the total radioactivity added; further suggesting that the protein binding was related to the formation of the metabolites and their interactions with other CYP enzymes Conversion of the Amine-derivative Back to S-4 by N-Acetyltransferase (NAT) 71

91 The experiments discussed above were specifically focused on phase I metabolic reactions, since only microsomal preparations (HLM or Supersome ) were used. We postulated that phase II metabolic enzymes might also play an important role in S-4 metabolism and disposition due to the discovery of the deacetylated metabolite M1 and the potential for conversion of M1 to S-4 by N-acetyltransferase (NAT) (Figure 5.8). More importantly, species differences in NAT expression could introduce species differences to the metabolism and diposition of S-4. The conversion of M1 to S-4 by NAT was first confirmed using in vitro enzyme preparations (Figure 5.9). As NAT is located in the cytosol [112], HLM preparations would not contain NAT. Also, since NAT is not expressed in dog [112], dog liver S9 do not contain NAT activity HLWKHU:KHQ0 0ZDVLQFXEDWHGZLWKGLIIHUHQWOLYHUHQ]\PHSUHSDUDWLRQVLQWKHSUHVHQFHRIDFHW\O- CoA (Figure 5.9 A), M1 to S-4 conversion was only observed human liver cytosol, S9, and rat liver S9, but not in HLM or dog liver S9. The rate of M1 acetylation was highest in rat liver S9, with the percent of M1 converted to S-4 being about 3-fold and 20-fold higher than that observed in human and dog liver S9, respectively, over the 2 hour incubation period In Vitro Characterization of the Species Difference in NAT Expression The opposing reactions of S-4 deacetylation and M1 acetylation suggested a reversible metabolic transformation might exist in some species. To test the reversibility of the reaction in different species, S-4 was incubated with different liver enzyme preparations in the presence of NADPH, with or without acetyl- CoA, a necessary cofactor for NAT (Figure 5.9 B). Without acetyl-coa, S-4 was deacetylazed to M1 by HLM, human liver cytosol, human liver S9, rat liver S9, and dog liver S9. M1 was formed in rat, dog, and human S9 liver preparations, suggesting that species differences in S-4 deacetylation are small. However, the rate of S-4 deacetylation in rats appeared to be significantly slower than that observed in dogs and humans. Since a similar amount of total protein was included in these reactions, the higher deacetylation activity observed in human and dog liver preparations suggested higher CYP enzyme activities in these species. However, other cytosolic enzymes might also catalyze the deacetylation of S-4 since M1 formation was also observed in reactions with human liver cytosol preparations. 72

92 On the other hand, when the same enzyme preparations were incubated with S-4 in the presence of NADPH and acetyl-coa, no change in the extent of S-4 deacetylation was observed with HLM and dog liver S9, due to the absence of NAT activity. In contrast, S-4 deacetylation was completely abolished in reactions containing rat liver S9, and tended to decrease in reactions containing human liver cytosol and S9, but the change was not significant. Since the reaction buffer used in the mixed reaction was not optimized for NAT reaction, the NAT activity in human liver cytosol and S9 might be under estimated. The differences observed in this in vitro experiment suggested that species differences in S-4 metabolism in vivo might exist. Comparing the results in Figures 5.9 A and B, it is apparent that both rat liver S9 and HL cytosol showed greater NAT activity (Figure 5.9 A), but lower deacetylation activity (Figure 5.9 B, without acetyl- CoA) compared to human liver S9. In the presence of both NADPH and acetyl-coa, the rate of S-4 deacetylation reaction was decreased to a greater extent with these preparations compared to that observed with human liver S9, which suggested that the relative content of CYP enzyme and NAT might change the metabolism profile of S-4 in vivo. Furthermore, the strong NAT activity observed in rat liver S9 preparation suggested that M1 might not be observed in vivo. Genetic polymorphism in NAT in humans is well documented, especially human NAT2 [114, 115]. The possible interaction between M1 and human NAT was characterized by using recombinant human NAT1 and NAT2 enzymes (Figure 5.10). M1 could be acetylated by both NAT1 and NAT2, and NAT2 showed higher affinity for M1 than NAT1. However, due to the limited solubility of M1 in the reaction buffer used for NAT enzyme reaction, the Vmax was not reached in these experiments In Vivo Characterization of the Species Differences in S-4 Metabolism We examined the pharmacokinetics of S-4 and M1 after intravenous doses of S-4 or M1 to rats and dogs. In rats, after an i.v. bolus dose (10 mg/kg) of S-4, only S-4, but no M1, was detected in the plasma (Figure 5.11 A upper panel). After an i.v. dose (10 mg/kg) of M1, both M1 and S-4 were observed in the plasma (Figure 5.11 A lower panel). Opposite results were observed in dog. After an i.v. dose of S-4 (10 73

93 mg/kg), both S-4 and M1 were detected in the plasma (Figure 5.11 B, upper panel). However, after an i.v. dose of M1 (10 mg/kg), no S-4 was detected in the plasma of dogs (Figure 5.11 B, lower panel). Since the pharmacokinetics of S-4 and M1in rats were studied in separate groups of animals, the data were computer-fitted separately. Plasma concentration-time profiles of S-4 (Figure 5.11 A) after i.v. dose of S-4 were computer-fitted with one-compartment model. Pharmacokinetic parameter estimates are summarized in Table 5.2. The volume of distribution for S-4 in rats was 112 ml, and elimination rate constant k 10 was min -1 (Table 5.2). Plasma concnetration0time profile of M1 and S-4 after an i.v. dose of M1 were computer-fitted using the model shown in Figure 5.2. Since no M1 was detected in rats after i.v. dose of S-4, we assume that the acetylation rate in vivo was much higher than the deacetylation rate. Therefore, only the acetylation of M1 (k a ) was considered in the proposed model (Figure 5.2 A). The M1 and S-4 PK data obtained after i.v. dose of M1 were fitted simultaneously using the proposed model. S- 4 and m1 data from each animal were fitted individually, and the mean and standard deviation of each parameter are listed in Table 5.2. The elimination rate constant of M1 (k m ) and the acetylation rate constant (k a ) were min -1 and min -1, respectively. After the i.v. dose of M1, 94 % of the M1 converted to S-4. The acetylation occurred so rapidly that M1 disappeared from the systemic circulation within 1 hour after dosing. The plasma concentration-time profiles of both S-4 and M1 after an i.v. dose of each in dog showed two exponential phases (Figure 5.11). When S-4 and M1 data were fitted separately using twocompartment model, the results suggested that the volume of the central compartment was about 400 to 600 ml (similar to the results shown in Table 5.3), which is close to the blood volume of dog [116]. Since the deacetylation occurs in the hepatocytes, we consider the liver as part of the peripheral compartment, and deacetylation (k h ) only happens in the peripheral compartment (Figure 5.2 B). Plasma concentration-time profiles of S-4 and M1 after i.v. dose of either S-4 or M1 were computer-fitted simultaneously using the model shown in Figure 5.2 B. The fitting results showed that after i.v. dose of S-4 in dogs, 29% of S-4 was metabolized through the deacetylation pathway. Representative fitting results are shown in Figure Although the clearance of M1 was more than two times higher than that of S-4, both S-4 and M1 had very similar terminal half-life for about two hours. The elimination rate 74

94 constant of M1 (k m ) was two times of the deacetylation rate constant (k h ), showing that deacetylation of S-4 is the rate-limiting process. 5.4 Discussion The major in vitro metabolism pathway of S-4 in rats, dogs, and humans is the deacetylation of the B-ring acetamide group. Deacetylation of the B-ring acetamide group was also observed with another thioether linked analog in our previous study [8]. Among the CYP enzymes tested, CYP3A4 is the major CYP enzyme that was responsible for the deacetylation reaction observed in vitro. Kinetics studies showed that S-4 has similar affinity (Km) to CYP3A4 as testosterone. At clinically relevant concentrations, it is likely that CYP3A4 would be responsible for the deacetylation of S-4. However, the deacetylation reaction observed with the phase I enzymes was reversed by NAT in vitro and in vivo. Species differences in NAT expression also introduced species differences into S-4 metabolism and pharmacokinetics in rats, dogs, and humans. The deacetylated product, M1, was converted back to S-4 in vitro by rat and human liver enzyme preparations that contained NAT activity. However, such conversion was not observed in dog liver enzyme preparation, due to the absence of NAT expression. Similar results were observed during in vivo experiments. In rats, due to the strong activity of NAT, M1 was not detected after an i.v. dose of S-4. Further, 94% of M1 was converted back to S-4 after an i.v. dose of M1. On the other hand, due to the lack of NAT activity in dog, about 30% of S-4 was metabolized by deacetylation, but S-4 was not observed after an i.v. dose of M1. Therefore, the pharmacokinetics of S-4 and M1 in dogs are qualitatively and quantitatively different in dogs than in rats and humans. Furthermore, although both rats and humans express NAT, the relative expression levels of CYP enzymes and NAT could also affect the in vivo metabolism of S-4 and result in inter-species differences. As observed during in vitro experiments, the strong NAT activity in rat liver S9 completely abolished the deacetylation reaction of S-4, which might explain the fact that no M1 was detected in rat plasma samples after i.v. doses of S-4. In human liver enzyme preparations, the NAT activity was not strong enough to completely reverse the deacetylation process. Thus, M1 is very likely to be observed in humans after administration of S-4. As a matter of fact, M1 was detected in human plasma sample after oral dose of S-4 75

95 during phase I clinical trial (data not shown.). Detailed analysis of clinical data may determine if the NAT polymorphisms [114, 115] will have any effects on S-4 phamacokinetics and metabolism in human. Another very important question is: does M1 have any pharmacologic activity? In competitive ELQGLQJDVVD\VXVLQJUDWSURVWDWHF\WRVRO0GLGQRWVKRZVLJQLILFDQWELQGLQJWR$5DW 07KHUHIRUH it s very unlikely that M1 will have any pharmacologic activity through AR-mediated pathway. The extremely short half-life of M1 in rat, 8 min (Figure 5.11 A, Table 5.2), also excluded the possibility that M1 might contribute to the pharmacologic effects of S-4 observed in male rats (Chapter 2 and 3). Even though M1 does not contribute to the pharmacologic activity of S-4, exposure to M1 might be related to the possible toxicity of S-4. Preclinical safety and toxicology studies conducted by a pharmaceutical sponsor showed that rats and monkeys tolerated higher doses of S-4 than dogs (data not shown). Our pharmacokinetics studies showed that M1 exposure in dogs after an i.v. dose of S-4 was significantly higher than that in rats (M1 not detected). M1 might also covalently bind to microsomal protein in vivo as indicated by the in vitro experiment using 14 C-S-4, in which 14 C-S-4 did not covalently bind any of the human CYP enzyme when incubated alone. This might also explain the effects of S-4 treatment on CYP enzyme activity in fresh human hepatocytes (Chapter 6). However, it is important to note that the toxicity observed during in vitro experiments could be exaggerated. First of all, to facilitate the identification and analysis of the metabolites, the concentration used in the in vitro system was much higher than clinical relevant concentrations. Secondly, the in vitro system is relatively static. Drug toxicity could be overestimated due to the accumulation of metabolites over time. The potential hepatotoxicity needs to be further evaluated using appropriate in vivo models. Nonsteroidal antiandrogens, flutamide and nilutamide, also showed covalent binding to hepatic enzymes [ ] during incubation with cultured hepatocytes. Free radical formation, redox cycling, and the formation of electrophilic metabolites were proposed to explain covalent binding observed, but it is still unclear what metabolite actually binds to the protein. Although in vitro enzyme covalent binding of these AR ligands are well documented, no severe hepatotoxicity was observed when normal pharmacological doses of flutamide and nilutamide are used. Also, to completely block androgen action in vivo, the pharmacological doses of these antiandrogens used in human are much higher than that of S-4. Therefore, it is unlikely that S-4 would cause any hepatotoxicity in humans at the clinically relevant doses. 76

96 Inter-species scaling is often used to estimate the appropriate dosage for humans based on the pharmacokinetic profile of the drug in animals. However, the species differences in drug metabolism and disposition of S-4 could greatly affect the accuracy of this estimation. The relatively higher CYP enzyme activity in human liver preparation, absence of NAT in dog, and known genetic polymorphisms in human NAT (Figure 5.9) suggest that inter-species scaling of S-4 might be confounded by numerous issues. Therefore, in vitro-in vivo correlation smight be a better approach to predict the in vivo pharmacokinetic and metabolism of S-4 in humans. In summary, both in vitro and in vivo data showed that S-4 could be metabolized by the deacetylation of the B-ring acetamide group, and that species differences exist for the metabolism of S-4. However, the deacetylated metabolite, M1, does not seem to contribute to the in vivo pharmacologic activity of S-4, which excluded the possibility that the tissue selectivity of S-4 is mediated by active metabolites. 77

97 A. Flutamide Metabolism O 2 N O F 3 C N H CH 3 CH 3 Hydrolysis Oxidation O 2 N F 3 C NH 2 O 2 N F 3 C N H O OH CH 3 Oxidation CH 3 O 2 N OH F 3 C NH 2 B. Bicalutamide Metabolism NC O O F F 3 C N H OH S O Hydrolysis Rat Rat Dog Human Oxidation NC NC O O F F 3 C NH 2 + O O F F 3 C N H OH S O OH HO OH S O Figure 5.1 Flutamide (A) [108] and bicalutamide (B) [109, 110] metabolism. 78

98 Enzyme Substrates Metabolites Internal Standard for HPLC Analysis CYP1A2 Phenacetin Acetamidophenol 2-Acetamidophenol CYP2C9 Diclofenac 4 -Hydroxy-diclofenac Isoxicam CYP2C19 Mephenytoin 4 -Hydroxy-mephenytoin Phenobarbital CYP2D6 Bufuralol 1 -Hydroxy-bufuralol CM-II-87 CYP3A4 Testosterone -Hydroxy-testosterone Dexamethasone NAT1 p-aminosalicylic Acid 4-Acetamidosalicylic Acid NAT2 Sulfamethazine N-acetyl Sulfamethazine Table 5.1 Specific substrates for recombinant human CYP enzymes, NAT1, and NAT2, and the internal standard used for HPLC analysis. No internal standard was used for NAT1 and NAT2 specific substrate reactions. Structural analog of S-4. 79

99 A. Compartment model for M1 pharmacokinetics in rat M1 k a S-4 k m k 10 B. Compartment model for S-4 pharmacokinetics in dog M1 (Peripheral) k h S-4 (Peripheral) k 21 k 12 k 21 k 12 M1 (Central) S-4 (Central) k m k 10 Figure 5.2 Compartment models proposed for simultaneous fitting of S-4 and M1 data. A. M1 pharmacokinetics in rat; k a is the rate constant for M1 acetlyation in rat. B. S-4 pharmacokinetics in dog. k h is the rate constant for S-4 deacetylation in dog. 80

100 A. Total radioactivity in HPLC fractions S-4 Rxn supernatant 14 C-S-4 standard 200 DPM 150 M2 100 M Time (min) B. Relative amount of the total radioactivity in different fractions of the reaction mixture S-4 standard Rxn Supernatant M % S % S % Others 7% M2 17.6% Others 11.65% Protein Pellet 28.54% Figure 5.3 HPLC separation of 14 C-S-4 and its metabolites after incubation with human liver microsome (HLM). S- 0ZDVLQFXEDWHGZLWK+/0PJPODWƒ&IRUKRXUV0LFURVRPDOSURWHLQZDV precipitated after the reaction. S-4 and the metabolites in the supernatant were separated by HPLC, eluted fractions were collected. The total radioactivity in each HPLC fraction and the precipitated protein pellet was counted. A. Total radioactivity (DPM) in different fractions eluted from the HPLC. B. Relative amount of the total radioactivity in different fractions of the reaction mixture. 81

101 O 2 N A O B NHCOCH 3 S-4 F 3 C N H OH O 14 C Hydrolysis (Proposed Pathway) Deacetylation O 2 N O 2 N O NH 2 HO F 3 C NH 2 + O O 14 C OH NHCOCH 3 M2 F 3 C N H O OH M1 m/z C O or NH 2 HO OH O 14 C Figure 5.4 Phase I metabolism pathways of 14 C-S-4 (uniformly labeled B-ring) as determined in HLM. Deacetylation product of the B ring acetamide group, M1, was confirmed by LC-MS analysis. Hydrolysis of the amide bond is a proposed pathway yet to be confirmed. 82

102 A % of S-4 Remaining CYP1A2 CYP2C9 CYP2C19 CYP2D6 CYP3A4 M1 Formation (pmole/(pmole P450 min)) S-4 2 um 0.18 S um CYP1A2 CYP2C9 CYP2C19 CYP2D6 CYP3A4 B M1 Formation Rate (pmole / (pmole CYPsmin)) CYP2C CYP2D6 CYP1A CYP3A CYP3A4 K m = 1.65 um V max = pmole/(pmolemin) S-4 (um) Figure 5.5 Identification of the major CYP enzymes that are responsible for phase I metabolism of S-4 using Supersome prepared from insect cells that express different recombinant human CYP enzymes (n=2). A. S- 0RU 0ZDVLQFXEDWHGZLWKGLIIHUHQWUHFRPELQDQWKXPDQ&<3HQ]\PHV10 to 40 pmole /reaction) at 37 C for 2 hours, with the presence of NADPH. The disappearance of S-4 or appearance of M1 was measured by HPLC analysis. B. Formation of M1 after 20 minutes incubation of various concentrations of S-4 with different recombinant human CYP enzymes (10 pmole/reaction) in the presence of NADPH at 37 C. 83

103 A. In vitro metabolism of S-4 by recombinant human CYP3A4 Disappearance of S-4 (pmole/pmole/min) K = 16.1 um 0.6 m V = 1.6 pmole/(pmole min) max S-4 (um) B. S-4 Deacetylation by HLM M1 Formation Rate (pmole/(min mg protein)) K = 58.4 um m V = pmole/(mg protein min) max S-4 (um) Figure 5.6 Characterization of the enzyme kinetics of S-4 metabolism by CYP3A4 and S-4 deacetylation by HLM (n=2). A. S- 0RU 0ZDVLQFXEDWHGZLWKGLIIHUHQWUHFRPELQDQWKXPDQ&<3HQ]\PHV (10 to 40 pmole /reaction) at 37 C for 2 hours, with the presence of NADPH. The disappearance of S-4 or appearance of M1 was measured by HPLC analysis. B. Formation of M1 after 20 minutes incubation of various concentrations of S-4 with different recombinant human CYP enzymes (10 pmole/reaction) in the presence of NADPH at 37 C. The kinetic parameters Km and Vmax were determined by nonlinear regression analysis using WinNonlin and the sigmoidal Emax model. 84

104 % Total DPM NSB IC NSB IC CYP3A4 CYP1A2 CYP2C9 CYP2C19 CYP2D6 IC/NADPH CYP3A4 only CYPs 1A2+3A4 CYPs 2C9+3A4 CYPs 2C19+3A4 CYPs 2D6+3A4 CYP 1A2 CYP 2C9 CYP 2C19 CYP 2D6 w/o NADPH w/ NADPH. Figure 5.7 Covalent binding of 14 C-S-4 metabolites to human liver microsomal proteins (n=2). S- 0 was incubated with Supersome prepared from insect cells that express different recombinant human CYP enzymes (1 nmole/ml, with total protein content about 0.5 mg/ml) at 37 C for 2 hours, with or without the presence of NADPH. Microsomal protein prepared from blank insect cells (IC) was used as control. Microsomal protein was precipitated by the addition of acetonitrile (v:v/1:1) after the reaction. The total radioactivity in the precipitated protein pellet was counted, and presented as the percentage of the total radioactivity added to the reaction. Non-specific binding (NSB) was determined by incubation with predenatured microsomal protein. 85

105 O 2 N O NHCOCH 3 S-4 F 3 C N H OH O Rat Human Dog Deacetylation CYPs Acetylation N-Acetyl Transferase (NAT) Rat Human Dog O 2 N F 3 C N H O OH O NH 2 M1 Figure 5.8 Species differences in S-4 metabolism are related to the species differences in NAT expression. Deacetylation product of S-4 was observed after incubation with rat, human and dog liver microsomal preparations. In human and rat, M1 can be converted back to S-4 by N-acetyl transferase (NAT), but similar conversion does not happen in dog since NAT is not expressed in dog. 86

106 A 100 NAT + Acetyl-CoA M1 S-4 % of M1 Converted to S Control HLM HL Cytosol Human S9 Rat S9 Dog S9 B NAT + Acetyl-CoA M1 S-4 35 CYPs + NADPH % of S-4 Converted to M HLM HL Cytosol Human S9 Rat S9 Dog S9 HLM HL Cytosol Human S9 Rat S9 Dog S9 with acetyl-coa Figure 5.9 In vitrofrqyhuvlrqri0 0WR6-4 in the presence of acetyl-coa (A) and NADPH (B) in different enzyme preparations (n=3). Appearance of S-4 was measured after 2 hour incubation, and the result is presented as mean ± S.D. A. Conversion of M1 to S-4 in the presence of acetyl-coa was only observed in human liver (HL) cytosol, S9 and rat liver S9, which contain NAT. Similar conversion was not observed in human liver microsomes (HLM) or dog S9, in which NAT is either not present or not expressed. B. Metabolism of S-4 to M1 in the presence of NADPH could be reversed by co-incubation with acetyl-coa with preparations containing NAT, such as human liver cytosol, S9, and rat liver S9. p<0.05, compared to corresponding reaction performed without acetyl-coa. 87

107 70 S-4 Formation Rate (nmole/(minmg protein)) NAT2 NAT M1 (um) Figure 5.10 In vitro conversion of M1 to S-4 by recombinant human N-acetyl transferase 1 and 2 (NAT1 and NAT2) (n=3). Data is presented as mean ± S.D. Various concentration of M1 were incubated with recombinant human NAT1 and NAT2 (0.25 mg/ml total protein) at 37 C for 5 min. The appearance of S-4 was measured by HPLC analysis. 88

108 A. S-4 and M1 PK in Rat B. S-4 and M1 PK in Dog 100 S-4 (iv bolus, 10 mg/kg) PK in Rats (n=5) 1000 S-4 (iv bolus, 10 mg/kg) PK in Dogs (n=3) Plasma Concentration (ug/ml) 10 S-4 Plasma Concentration (ug/ml) S-4 M Time (min) Time (min) 100 M1 (iv bolus, 10 mg/kg) PK in Rats (n=5) 100 M1 (iv bolus, 10 mg/kg) PK in Dogs (n=3) Plasma Concentration (ug/ml) 10 1 M1 S-4 Plasma Concentration (ug/ml) 10 1 M Time (min) Time (min) Figure 5.11 Pharmacokinetic profiles of S-4 and its metabolite M1 in rat and dog. Data are presented as mean ± S.D. A. S-4 and M1 pharmacokinetic profile in rats. M1 was not observed in rats after an i.v. dose of S-4, but S-4 was observed in rats after an i.v. dose of M1. B. S-4 and M1 pharmacokinetic profile in dogs. M1 was observed in dogs after an i.v. dose of S-4, but S-4 was not observed in dogs after an i.v. dose of M1. 89

109 Parameter Unit Mean ± S.D. S-4 iv V d1 ml 112 ± 6 k 10 min ± t 1/2 min 160 ± 30 Cl ml/min 0.49 ± 0.07 AUC JPLQPO 5144 ± 738 M1 iv V d2 ml 115 ± 20 k a min ± k m min ± t 1/2 (M1) min 8 ± 1 Cl m ml/min 1.88 ± 0.39 AUC (M1) JPLQPO 248 ± 38 AUC m JPLQPO 4099 ± 275 f m 0.94 ± 0.01 Table 5.2 Pharmacokinetic parameters of S-4 and M1 in rats (n=5). 90

110 Parameter Unit Mean ± S.D. V d1 ml 423 ± 49 k 10 min ± t 1/2 min 123 ± 23 Cl ml/min 19.7 ± 4.7 AUC JPLQPO 2217 ± 379 V d2 ml 624 ± 242 k h min ± k m min ± t 1/2 (M1) min 121 ± 31 Cl m ml/min 53.3 ± 7.1 AUC m JPLQPO 905 ± 212 f m 0.29 ± 0.05 Table 5.3 Pharmacokinetic parameters of S-4 and M1 in dogs (n=3). 91

111 A. Plasma concentration of S-4 and M1 in rats 100 Concentration (ug/ml) 10 1 S-4 M Time (min) B. Plasma concentration of S-4 and M1 in dogs M1 after iv dose of S-4 S-4 after iv dose of S-4 M1 after iv dose of M1 Concentration (ug/ml) Time (min) Figure 5.12 Representative fitting results of S-4 and M1 concentration-time profiles in rat (A) and dog (B). 92

112 CHAPTER 6 EFFECTS OF S-1 AND S-4 ON CYTOCHROME P450 ENZYME EXPRESSION IN PRIMARY CULTURE OF HUMAN HEPATOCYTES 6.1 Introduction The direct regulation of hepatic metabolizing enzyme expression by nuclear receptors (NR) is well documented [ ]. However, the effects of AR on cytochrome P450 enzyme expression are not as well characterized as the regulatory effects of the orphan receptors [120, 121, 125] and other steroid receptors. Up to date, there is no direct evidence to show that any of the CYP genes are directly regulated by AR [125], although the presence of various nuclear receptor elements in many CYP gene promoter regions [120] suggests that AR may play a direct role in regulating CYP enzyme expression. Even so, AR ligands could still indirectly regulate hepatic metabolizing enzyme expression through cross interaction with other NR, secondary effects of direct regulation of other gene targets, or cross-talk between AR and other NR signaling pathway [122]. Although S-4 shows much better specificity for the AR than other steroidal ligands, it is unclear if it interacts with any of the orphan receptors that are more important in direct regulation of CYP enzyme expression, or whether activates cross-talk between AR and other NR. Enzyme induction or suppression by drug treatment is one of the major causes for drug-drug interactions and hepatotoxicity [126]. Since both S-1 and S-4 are substrates for CYP3A4, the in vivo pharmacologic effects of S-1 and S-4 could also be affected by changes in hepatic enzyme expression. Therefore, the possible effects of S-1 and S-4 on the expression of the major CYP enzymes (CYPs 1A2, 2C9, 2C19, 2D6, and 3A4) that are responsible for the metabolism of a majority of drugs [127] were tested in primary cultures of human hepatocytes. 93

113 Both animal models and in vitro systems, including isolated liver tissue or primary culture, are used to assess the effects of drugs on hepatic enzyme expression. In vitro systems are more widely used due to their low cost and convenience [127]. Primary culture of human hepatocytes has been used as the gold standard in vitro model to assess the inductive and/or suppressive effects of drug treatment on CYP enzyme expression [127] due to its inherent ability to avoid concerns about possible species differences. Although the expression of CYP enzymes decreases in cultured hepatocytes over time, the hepatocytes remain specifically and quantitatively responsive to model inducers [127, 128]. Enzyme induction and suppression can be assessed by measuring: the enzyme activity using specific substrates, enzyme protein level using immunoblotting, and mrna level using quantitative PCR. However, measuring only one of these markers is not advisable, due to differences in mechanism, and the potential that a good correlation between changes in enzyme function and changes in protein and mrna expression might not exist. A comparison of the changes in enzyme transcription, translation and function are most often employed to better understand the mechanism of regulation [127]. Two well characterized model inducers, rifampicin (RIF, CY3$LQGXFHUDQG -naphthoflavone (BNF, CYP1A2 inducer), were included as positive controls in our study to test the viability of the hepatocyte culture. A solvent control was also included since the presence of an organic solvent like DMSO might also affect the enzyme expression [129, 130]. Since enzyme activity in isolated hepatocytes is more stable after 4 days in culture, and maximum changes in enzyme expression were observed after 72 hours treatment [131], we began drug treatment 4 days after isolation of the hepatocytes and continued it for 72 hours to allow for maximal effects. The cytotoxicity of S-1 and S-4 and plasma concentrations of S-4 at clinically relevant doses were both considered in choosing the appropriate concentration for treatment. 6.2 Materials and Methods Materials Recombinant human CYP enzymes (Supersome ), liver microsome preparation, fresh human hepatocytes, Hepato-STIM medium and supplements, and primary antibodies for human CYPs 1A2, 2C9, 94

114 2C19, 2D6, 3A4 were purchased from BD Gentest (Woburn, MA). 4 -Hydroxy-diclofenac, 4 -hydroxymephenytoin, mephenytoin, 1 -hydroxy-bufuralol, and bufuralol were also purchased from BD Gentest. Rabbit anti-actin IgG, goat anti-mouse IgG, and rabbit anti-goat IgG were purchased from Santa Cruz %LRWHFKQRORJLHV6DQWD&UX]&$ -Hydroxy-testosterone was purchased from Steraloids Inc. (Newport, RI). Enhanced chemiluminescence kit was purchased from Amersham Biosciences (Buckinghamshire, UK). Trizol reagent and Superscript First-strand Synthesis System were purchased from Invitrogen Corp. (Carlsbad, California). PCR primers were synthesized by IDT, Inc. (Coralville, IA). SYBR green nucleic acid gel stain was purchased from Molecular Probes (Eugene, OR). Smart Cycler additive was purchased from Cepheid (Sunnyvale, CA). All other chemicals and reagents were purchased from Sigma Chemical Company (St Louis, MO). All analytical columns were purchased from Waters Corporation (Milford, MA) Cytotoxicity Measurement in HepG2 cells HepG2 cells were plated in 24 well plates, and were treated with solvent (0.1% DMSO) or various FRQFHQWUDWLRQVWR 0RI6-1 or S-4 for 72 hours. Three wells were included for each concentration. The medium was changed every 24 hours. At the end of treatment, cell number was measured using the colorimetric sulforhodamine B (SRB) assay, and was reported as a percentage of that observed in control samples Primary Culture of Human Hepatocytes Primary cultures of human hepatocytes isolated from one donor (BD Gentest, Lot # 54, Donor # HH129, female Caucasian, 49 year old, died of stroke) were plated into 24 or 48 well plates, and were shipped 24 hours after isolation. The cultures were maintained in Hepato-STIM medium at 37 C. The culture medium did not include phenol red, but was supplemented with epidermal growth factor (EGF, 1 PJPODQGGH[DPHWKDVRQH 0 95

115 6.2.4 Treatment of Human Hepatocyte Culture The hepatocytes were maintained in the Hepato-STIM medium for two days after arrival to allow for recovery from shipment, and were then incubated with S- 06-0ULIDPSLFLQ5,) 0 -QDSKWKRIODYRQH%1) 0RUVROYHQW'062IRUKRXUV)LIWHHQZHOOVZHUHLQFOXGHG for each treatment, and three wells were used for each activity assay. Cells without any treatment were also included as a control. Drug-containig solutions were prepared freshly everyday in DMSO, and then diluted to the desired concentration in culture medium. Culture medium with drugs was changed every 24 hours CYP Enzyme Function Assays After three days treatment in 48 well plates, the intact hepatocytes were washed three times with blank medium, and then incubated with the CYP enzyme-specific substrdwhvlqfoxglqjskhqdfhwlq 0 &<3$GLFORIHQDF 0&<3&PHSKHQ\WRLQ 0&<3&EXIXUDORO 0 &<3'DQGWHVWRVWHURQH 0&<3$IRURQHKRXUDWƒ&WRWHVWWKHHQ]\PHDFWLYLWLHV7KUHH wells were used for each reaction, and the appearance of the metabolites in the medium was assessed by HPLC analysis. Cells were lysed after the functional assay, and the total protein content in the lysate was determined using the Bradford method (Bio-Rad Protein Assay). All enzyme activity data was normalized by total protein content of each well. Acetamidophenol (CYP1A2 metabolite) was separated on a reversed-skdvhfroxpq %RQGD3DN C18, mm) with a mobile phase of 15% acetonitrile in deionized water at a flow rate of 1.5 ml/min, and was detected by its UV absorbance at 244 nm. 4 -Hydroxy-diclofenac (CYP2C9 metabolite) was separated on a reversed-phase column (NovaPak C18, mm) with a mobile phase of 40% acetonitrile and 0.5% formic acid (ph 2.65) in deionized water at a flow rate of 1 ml/min, and was detected by its UV absorbance at 280 nm. 4 -Hydroxy-mephenytoin (CYP2C19 metabolite) was separated on a reversed-phase column (NovaPak C18, mm) with a mobile phase of 25% acetonitrile and 25 mm potassium phosphate (ph 7.4) in deionized water at a flow rate of 1 ml/min, and was detected by its UV absorbance at 214 nm. 1 -Hydroxy-bufuralol (CYP2D6 metabolite) was separated on a reversed-phase 96

116 column (NovaPak C18, mm) with a mobile phase of 50% acetonitrile and 2 mm perchloric acid in deionized water at a flow rate of 1 ml/min, and was detected using a fluorescence detector with H[FLWDWLRQZDYHOHQJWKRIQPDQGHPLVVLRQZDYHOHQJWKRIQP -Hydroxy-testosterone (CYP3A4 metabolite) was separated on a reversed-phase column (NovaPak C18, mm) with a mobile phase of 40% acetonitrile in deionized water at a flow rate of 1 ml/min, and was detected by its UV absorbance at 254 nm Western Immunoblot Analysis Cell lysate prepared after the functional assay were used for immunoblotting analysis of CYPs. eta-actin was also analyzed as a loading control. Signal was detected using an enhanced chemiluminescence kit from Amersham Biosciences (Buckinghamshire, UK). The standard curve for each CYP isoform was constructed using the recombinant human CYP Supersome with known enzyme content. The band density was analyzed using ImageJ software ( Real-Time PCR Analysis A separate 24 well plate of hepatocytes was treated similarly as described in Four wells were included for each treatment. After 72 hours treatment, total RNA was extracted using Trizol reagent. cdna samples were prepared from the isolated total RNA sample using Superscript First-strand Synthesis System, and then was used for real-time PCR analysis. Gene specific primers were designed for CYP1A1, 2C9, 2C19, 2D6, 3A4 and GAPDH (Table 6.1) using the Primer 3 program ( Amplification was carried out using the Smart Cycler (Cepheid, Sunnyvale, CA) as follows: 300s at 95 C, 35 cycles of 20s at 95 C, 30s at 58 C, and 30s at 72 C. An extension cycle of 300s at 72 C was followed by melt analysis starting at 60 C and increasing to 95 C at a rate of 0.2 C/s. Negative first derivative peaks, which are characteristic of the PCR product melt temperature, were used to identify VSHFLILF3&5SURGXFWV5HDFWLRQFRQGLWLRQV O-QJF'1$VROXWLRQ O'(3&ZDWHU3&5 97

117 buffer (10 mm Tris-+&OS+P0.&ODQGP00J&O 0G1730 nm of both sense and antisense primers, 1:12500 dilution of SYBR green nucleic acid gel stain 10,000X in DMSO, 1.0 unit RI7DT'1$SRO\PHUDVHDQG O6PDUW&\FOHU DGGLWLYHIRUDWRWDOYROXPHRI OSHUUHDFWLRQ(DFK cdna sample was subjected to a reaction consisting of duplicate runs for each CYP isoform and for GAPDH. The comparative Ct method [132] was used for mrna quantification. This method compares the relative expression of the gene of interest to a reference gene such as GAPDH. The number of cycles UHTXLUHGWRUHDFKDQDUELWUDU\IOXRUHVFHQFHWKUHVKROGYDOXH&WZDVXVHGWRFDOFXODWH'HOWD&W &WE\ VXEWUDFWLQJWKH&WRIWKHUHIHUHQFHJHQHIURPWKH&WRIWKHWDUJHWJHQH &Wwas calculated for the control LQFXEDWLRQLQPHGLDDQGH[SHULPHQWDOO\WUHDWHGFHOOV6XEWUDFWLQJWKH &WRIWKHH[SHULPHQWDOJURXSIURP WKH &WRIWKHFRQWUROJURXS\LHOGHG &W7KHIROG-change relative to the control was determined using the formula 2 - &W. Statistical analyses of all the parameters were performed by single-factor ANOVA with the alpha value set a priori at p< Results Cytotoxicity of S-1 and S-4 in HepG2 cells The cytotoxicity of S-1 and S-4 was assessed in HepG2 cells, a hepatocellular carcinoma cell line, to determine the concentration that would be used to treat human hepatocytes. S-1 and S-4 showed some WR[LFLW\LQ+HSFHOOVDWFRQFHQWUDWLRQVJUHDWHUWKDQ 0)LJXUH+RZHYHUQRW[RFLFLW\ZDV observed at lower concentrations. The concentration required for toxicity was much higher than the highest SODVPDFRQFHQWUDWLRQVDERXW 0RI6-4 observed during phase I clinical trials. To fully assess the effects of S-1 and S-4 on CYP expression without causing cywrwr[lflw\dfrqfhqwudwlrqri 0ZDV chosen for the human hepatocyte studies Effects of S-1 and S-4 on CYP Enzyme Function, Protein Expression, and mrna levels 98

118 Results for individual CYP enzyme are summarized in Figures 6.2 through 6.6, including enzyme activity measured as conversion of the probe substrate to metabolite, CYP enzyme protein expression level determined by immunoblot, and the relative mrna level quantified by real-time PCR. CYP1A2 activity (Figure 6.2) in untreated control cells was around 50 pmole/(mgmin). Solvent (0.1% DMSO), S- 0DQG6-0WUHDWPHQWGLGQRWFDXVHDVLJQLILFDQWFKDQJHLQ&<3$ activity, protein expression level, or mrna levels. mrna signal in solvent treated samples was not detected due to the limithgdprxqwriwrwdo51$dydlodeoh%1) 0DNQRZQ&<3$LQGXFHU significantly increased CYP1A2 activity by 10 fold, with a concomitant increase in CYP1A2 protein expression. Correspondingly, the mrna level of the enzyme was also increased by 7.56 fold after BNF treatment, showing that the increase in CYP1A2 activity was due to the induction of enzyme expression by BNF. Both CYP2C9 (Figure 6.3) and CYP2C19 (Figure 6.4) enzyme activity and expression showed very similar changes in response to the different treatments. Solvent and BNF treatment showed little effects on the enzyme actiylwlhvdqgh[suhvvlrqohyhovri&<3v&dqg&5,) 0WUHDWPHQW increased CYP2C9 and CYP2C19 activity by 2 fold and 6 fold, respectively. Increases in protein expression were also observed. S-1 and S-4 did not cause significant changes in CYPs 2C9 and 2C19 protein expression or mrna levels. However, both treatments decreased enzyme activity towards their probe substrate. S-4 decreased the CYPs 2C9 and 2C19 activities by 57% and 73%, respectively; while S-1 decreased the CYPs 2C9 and 2C19 activities by 16% and 47%, respectively. Since no significant change in enzyme expression was observed, and considering that S-4 had some affinity for CYP&DW 0 concentrations (see Chapter 5), the decrease in enzyme activity in S-1 and S-4 treated samples could be related to the direct interaction between S-4 or its metabolites and the enzyme. The expression of CYP2D6 (Figure 6.5) was increased by RIF and BNF to twice of the protein levels observed in control samples, CYP2D6 activity was increased by more than 2 fold as well. No significant changes in CYP2D6 enzyme, mrna, or activity were observed after treatment with S-1 or S-4. Although S-4 is a substrate for CYP3A4, it did not affect the enzyme expression or activity in hepatocytes (Figure 6.6). Similar results were observed in solvent, S-1, and BNF-treated samples as well. 99

119 RIF is a stronger inducer of CYP3A4, significantly increasing the enzyme activity (7 fold), and the enzyme expression at both mrna (3.69 fold) and protein levels (more than 10 fold). 6.4 Discussion,QVXPPDU\5,) 0VLJQLILFDQWO\LQFUHDVHGWKHHQ]\PHDFWLYLW\RI&<3V&&DQG 3A4 by 2, 6, and 7 fold, respectively. BNF 0VLJQLILFDQWO\LQFUHDVHGWKHHQ]\PHDFWLYLW\RI&<3$ by 10 fold. Similar changes were also observed in the protein expression levels of these enzymes as well. In hepatocytes treated with S-4, enzyme activity of CYP2C9 and 2C19 was decreased by 57% and 73% respectively. In hepatocytes treated with S-1, the enzyme activity of CYP2C9 and 2C19 was decreased by 16% and 47%, respectively. However, no significant changes in protein or mrna levels of any of these enzymes were observed in S-1 and S-4 treated cells. As mentioned in the introduction, NR play a very important role in regulating CYP enzyme expression. Both model inducers used in this study, RIF and BNF, are NR ligands. BNF induces CYP1A2 expression by activating aryl hydrocarbon receptor (AhR) [127, 133], while RIF induces CYPs 2C9 and 3A4 expression by activating human PXR (Pregnane X Receptor) [120]. Recent studies with CYP2C19 promoter also identified binding sites for CAR (constitutive androstane receptor) and GR [134]. Gel-shift assay showed that human PXR binds to the CAR response element as well [134], which suggested that CYP2C19 expression could also be directly regulated by PXR ligand (i.e., rifampicin). CYP2C gene induction study using primary human hepatocyte also showed that rifampicin induced the expression of CYPs 2C9 and 2C19 at both protein and mrna levels [134]. The results observed in this study are consistent with those findings. +XPDQ&<3'H[SUHVVLRQLVUHJXODWHGE\DQRWKHURUSKDQUHFHSWRUKHSDWRF\WHQXFOHDUIDFWRU +1) [123]. There is no evidence to show that RIF or %1)DUH+1) OLJDQGV$OWKRXJK+1) directly regulate the gene expression of PXR and CAR [135], it is not clear if PXR and CAR regulate +1) H[SUHVVLRQ7KHLQGXFWLRQRI&<3'H[SUHVVLRQE\5IF and BNF could be related to the cross- WDON EHWZHHQ3;5&$5DQG+1) 100

120 SARM treatment did not cause any significant changes in the major CYP enzyme expression, suggesting that AR might not be involved in the direct regulation of the expression of these CYP enzymes, and that is unlikely that any interactions between SARM and the orphan receptors that directly regulate CYP enzyme expression occur. Unexpectedly, the mrna level of CYP2C9 (Figure 6.3) was increased in solvent treated samples, but no changes in enzyme activity or protein level were observed in these samples. This could be related to the real-time PCR analysis, since only one housekeeping gene, GAPDH, was used as the internal control. A prior literature report [136] showed that normalization using a single gene leads to relatively larger errors in measurement. Normalization using multiple housekeeping genes [136]or 18S ribosomal RNA [137] could be used in future studies to reduce the error in measurement. Although S-1 and S-4 did not show any regulatory effects on CYP enzyme expression at either transcription or protein expression level, S-1 and S-4 treatment decreased the enzyme activity of CYPs 2C9 and 2C19, which could be the results of direct inhibition of the enzyme by the residual amount of drugs or metabolites left in the culture, a common problem observed in enzyme induction studies using hepatocytes [127]. Therefore, S-1 and S-4 do not induce or suppress the expression of the major CYP enzymes in primary human hepatocytes, although these drugs could directly inhibit the enzyme activity of CYPs 2C9 and 2C19. HoZHYHUWKHGUXJFRQFHQWUDWLRQXVHGLQWKLVVWXG\ 0ZDVPRUHWKDQIROGKLJKHUWKDQ the plasma concentration that could be achieved at clinical relevant doses. Thus, inhibitory effects of S-1 and S-4 on CYPs 2C9 and 2C19 might not be observed in vivo. 101

121 120 Cell Numbers (% of Control) S-1 S-4 0 Ctrl 0.1% DMSO 1 um 5 um. 10 um 20 um 50 um 100 um Figure 6.1 Cytotoxicity of S-1 and S-4 in HepG2 cells measured by SRB assay (n=3) after 72 hour treatment. Data was presented as mean ± SD. 102

122 Primer Sequence (5 to 3 ) 7 T m CYP1A2 1 Forward CAGAATGCCCTCAACACCTT 89 C Reverse CTGACACCACCACCTGATTG CYP2C9 2 Forward AAGAACCTTGACACCACTCCA 89 C Reverse TAATGCCCCAGAGGAAAGAG CYP2C19 3 Forward TGGGACAGAGACAACAAGCA 88 C Reverse TGGGGATGAGGTCGATGTAT CYP2D6 4 Forward AGGGAACGACACTCATCACC 92 C Reverse CAGGAAAGCAAAGACACCAT CYP3A4 5 Forward AATAAGGCACCACCCACCTA 86 C Reverse CTTGGAATCATCACCACCAC GAPDH 6 Forward GTCAGTGGTGGACCTGACCT 91 C Reverse TGAGCTTGACAAAGTGGTCG Table 6.1 Oligonucleotide sequences for real-time PCR analysis. 1 Based upon published CYP1A2 sequence (NM_ ) 2 Based upon published CYP2C9 sequence (NM_000771) 3 Based upon published CYP2C19 sequence (NM_000769) 4 Based upon published CYP2D6 sequence (NM_000106) 5 Based upon published CYP3A4 sequence (NM_017460) 6 Based upon published GAPDH sequence (NM_002046) 7 Amplicon melting temperature (T m ) obtained from melt curve analysis 103

123 Phenacetin O-deethylase Activity pmole/(mgmin) CYP1A2 Ctrl Solv S-4 S-1 RIF BNF HLM CYP1A2 (fmole) -actin mrna level 1.00 ND Figure 6.2 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -QDSKWKRIODYRQH%1) 0 on CYP1A2 activity and expression. CYP enzyme activity was measured in triplicate, and the result is presented as mean ± S.D. Enzyme content was estimated by comparing the band density to the standard curve constructed with Supersome SUHSDUDWLRQVDQGQRUPDOL]HGE\ -actin expression level. Human liver microsome (HLM) sample was included as positive control for the immunoblot. The fold change in mrna level was normalized to the control samples. 104

124 Diclofenac 4 -hydroxylase Activity pmole/(mgmin) CYP2C9 Ctrl Solv S-4 S-1 RIF BNF CYP2C9 (fmole) -actin HLM mrna level Figure 6.3 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -QDSKWKRIODYRQH%1) 0 on CYP2C9 activity and expression. CYP enzyme activity was measured in triplicate, and the result is presented as mean ± S.D. Enzyme content was estimated by comparing the band density to the standard curve constructed with Supersome SUHSDUDWLRQVDQGQRUPDOL]HGE\ -actin expression level. Human liver microsome (HLM) sample was included as positive control for the immunoblot. The fold change in mrna level was normalized to the control samples. 105

125 Diclofenac 4 -hydroxylase Activity pmole/(mgmin) CYP2C19 Ctrl Solv S-4 S-1 RIF BNF CYP2C19 (fmole) -actin HLM mrna level Figure 6.4 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -QDSKWKRIODYRQH%1) 0 on CYP2C19 activity and expression. CYP enzyme activity was measured in triplicate, and the result is presented as mean ± S.D. Enzyme content was estimated by comparing the band density to the standard curve constructed with Supersome SUHSDUDWLRQVDQGQRUPDOL]HGE\ -actin expression level. Human liver microsome (HLM) sample was included as positive control for the immunoblot. The fold change in mrna level was normalized to the control samples. 106

126 Bufuralol 1 -hydroxylase Activity pmole/(mgmin) CYP2D6 Ctrl Solv S-4 S-1 RIF BNF CYP2D6 (fmole) -actin HLM mrna level Figure 6.5 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -QDSKWKRIODYRQH%1) 0 on CYP2D6 activity and expression. CYP enzyme activity was measured in triplicate, and the result is presented as mean ± S.D. Enzyme content was estimated by comparing the band density to the standard curve constructed with Supersome SUHSDUDWLRQVDQGQRUPDOL]HGE\ -actin expression level. Human liver microsome (HLM) sample was included as positive control for the immunoblot. The fold change in mrna level was normalized to the control samples. 107

127 Testosterone 6beta-hydroxylase Activity pmole/(mgmin) CYP3A4 Ctrl Solv S-4 S-1 RIF BNF HLM CYP3A4 (fmole) -actin mrna level Figure 6.6 Effects of S- 06-0ULIDPSLFLQ5,) 0DQG -QDSKWKRIODYRQH%1) 0 on CYP3A4 activity and expression. CYP enzyme activity was measured in triplicate, and the result is presented as mean ± S.D. Enzyme content was estimated by comparing the band density to the standard curve constructed with Supersome SUHSDUDWLRQVDQGQRUPDOL]HGE\ -actin expression level. Human liver microsome (HLM) sample was included as positive control for the immunoblot. The fold change in mrna level was normalized to the control samples. 108