David Mccarty [email protected] Ph.D

The Biological Characterization of Galeterone Analogs VNPT-178 and VNLG-74A for the Treatment of Prostate Cancer

Item Type dissertation

Authors McCarty, David

Publication Date 2017

Abstract The expression of truncated androgen receptor spice-variants (AR-Vs) presents a significant clinical challenge for the treatment of advanced stages of prostate cancer and confers complete resistance to abiraterone and enzalutamide. Targeted depletion...

Keywords galeterone; Prostate--Cancer; Receptors, Androgen; Unfolded Protein Response

Download date 05/10/2021 16:14:55

Link to Item http://hdl.handle.net/10713/7073 CURRICULUM VITAE:

David McCarty [email protected]

Ph.D., July 25th, 2017

University of Maryland School of Medicine, August 2011-July 2017 Doctor of Philosophy, July 2017 Molecular Medicine – Cellular and Molecular Biology and Physiology

Saint Cloud State University, August 2003-December 2009 Bachelor of Science, December 2009 Biomedical Sciences Associate of Arts, May 2005 General Education

PENDING PUBLICATION: David J. McCarty, Weiliang Huang, Maureen A. Kane, Puranik Purushottamachar, Lalji K. Gediya. (2017) Novel Galeterone Analogs Act Independently of AR and AR-V7 for the Activation of the Unfolded Protein Response and Induction of Apoptosis in the CWR22Rv1 Prostate Cancer Cell Model. Oncotarget ISSN: 1949-2553

ABSTRACT:

Title of Dissertation: The Biological Characterization of Galeterone Analogs VNPT-178 and VNLG-74A for the Treatment of Prostate Cancer

David J. McCarty, Doctor of Philosophy, 2017

Dissertation Directed by: Dr. Vincent C.O. Njar, Professor of Medicinal Chemistry and

Pharmacology, Department of Pharmacology

The expression of truncated androgen receptor spice-variants (AR-Vs) presents a significant clinical challenge for the treatment of advanced stages of prostate cancer and confers complete resistance to abiraterone and enzalutamide. Targeted depletion of these receptors in addition to antagonism of the androgen signaling axis has been actively evaluated as a superior treatment paradigm but has thus far been unsuccessful in clinical trials. Based on our studies of galeterone and its analogs, we believe these compounds possess a specific activity that is both independent and overlapping of their antagonism of the androgen receptor (AR). In this study, we perform basic biochemical techniques attempting to better delineate the actions of galeterone analogs VNPT-178 and VNLG-

74A and identify a rational, common target for improved study of their utility against all stages of prostate cancer development.

Direct comparisons of VNPT-178 and VNLG-74A to vehicle or equimolar concentrations of abiraterone and enzalutamide reveal both compounds exhibit improved antiproliferative activities in multiple prostate cancer cell models of androgen and AR dependence. VNPT-178 and VNLG-74A directly antagonize the full-length AR in

LNCaP and CWR22Rv1 (22Rv1) cells while promoting the depletion of full-length and truncated receptors with dose- and time-dependency. Activation of the unfolded protein response is rapid and sustained, long preceding appreciable antagonism of the AR in

22Rv1 cells, and followed by CHOP upregulation and PARP cleavage—an effect also seen in PC-3 cells albeit with slower kinetics. Molecular docking of VNPT-178 and

VNLG-74A reveals greater potential for binding the ATPase domains of BiP/Grp78 and

Hsp70-1A compared to the AR’s ligand-binding domain.

Severe or sustained activation of the unfolded protein response can induce apoptosis. Taken together, our data suggest that galeterone analogs VNPT-178 and

VNLG-74A directly modulate the substrate interactions of BiP promoting endoplasmic reticulum stress and apoptosis independent of AR antagonism. Depletion of AR proteins is possibly the result of similar activities against Hsp70. The sensitivity of 22Rv1 cells to

VNPT-178- and VNLG-74A-induced apoptosis may underlie a failure of this cell model to adequately represent the clinical challenge of AR-V-expressing prostate cancers.

iii

The Biological Characterization of Galeterone Analogs VNPT-178 and VNLG-74A for the Treatment of Prostate Cancer

by David J. McCarty

Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, Baltimore in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2017

TABLE OF CONTENTS:

INTRODUCTION……………………………………………………………………………1

PROSTATE CANCER AT A GLANCE……………………………………………………...1

PROSTATE CANCER BIOLOGY AND CLINICAL PRESENTATION………………………….1

HORMONE THERAPY FOR PROSTATE CANCER………………………………………….3

THE ANDROGEN RECEPTOR (AR)……………………………………………………...7

MECHANISMS OF RESISTANCE TO AR-DIRECTED PROSTATE CANCER THERAPIES………………………………………………………………………..11

ANDROGEN RECEPTOR DEPLETING AGENTS (ARDAS) AND GALETERONE…………..12

THE ENDOPLASMIC RETICULUM STRESS/UNFOLDED PROTEIN RESPONSE (ERSR/UPR)……………………………………………………………………..15

PROJECT OVERVIEW AND SPECIFIC AIMS ……………………………………………...19

MATERIALS AND METHODS……………………………………………………………..22

SPECIFIC AIM 1…………………………………………………………………………..31

S1A: COMPARE THE ANTIPROLIFERATIVE ACTIVITY OF NOVEL GALETERONE ANALOGS TO ABIRATERONE AND ENZALUTAMIDE USING THE MTT ASSAY...... 32

S1B: COMPARE THE ABILITY OF GALETERONE ANALOGS FOR THE INHIBITION OF FULL-LENGTH AR TRANSCRIPTIONAL ACTIVATION USING A TRANSIENTLY TRANSFECTED LUCIFERASE REPORTER SYSTEM…………...39

S1C: DETERMINE WHICH OF THE NOVEL ANALOGS CAN INHIBIT THE ENDOGENOUS EXPRESSION OF THE ANDROGEN RECEPTOR IN 22RV1 CELLS………41

SUMMARY AND CONCLUSIONS FROM SPECIFIC AIM 1………………………………...43

SPECIFIC AIM 2…………………………………………………………………………..45

S2A: DETERMINE IF LONGER TREATMENTS WITH LOWER DOSES THAN THE PREDICTED GI50 VALUES COULD IMPROVE RELATIVE ACTIVITY OF GALETERONE ANALOGS TO ESTABLISHED CONTROLS...... 46

S2B: DETERMINE THE MINIMUM CONCENTRATION REQUIRED TO ELICIT AR ANTAGONISM AND DEPLETION WITHIN 24 HOURS IN LNCAP AND 22RV1 CELLS……………………………………………………………………..48

S2C: EVALUATE CELL DEATH ASSOCIATED WITH DOSE AND TIME OF TREATMENT WITH VNPT-178 AND VNLG-74A USING THE TRYPAN BLUE EXCLUSION ASSAY………………………………………………………….52

iii

S2D: EVALUATE CHANGES IN CELL CYCLE PROGRESSION FOLLOWING TREATMENT WITH INCREASING DOSES OF VNPT-178 AND VNLG-7A …………...54

SUMMARY AND CONCLUSIONS FROM SPECIFIC AIM 2 ………………………………..58

SPECIFIC AIM 3…………………………………………………………………………..62

S3A: DETERMINE WHETHER GALETERONE ANALOGS INDUCE NECROSIS OR APOPTOSIS …………………………………………………………………….63

S3B: DETERMINE THE AR-DEPENDENCY OF CELL DEATH INDUCTION IN PROSTATE CANCER CELLS ………………………………………………………...65

S3C: DETERMINE THE ERSR/UPR DEPENDENCY OF CELL DEATH INDUCTION IN PROSTATE CANCER CELLS………………………………………….69

S3D: IDENTIFY NOVEL TARGETS OF GALETERONE ANALOGS VNPT-178 AND VNLG-74A………………………………………………………………….74

SUMMARY AND CONCLUSIONS FROM SPECIFIC AIM 3………………………………...78

SPECIFIC AIM 4…………………………………………………………………………..80

S4A: DETERMINE THE GROWTH INHIBITORY EFFECTS OF GALETERONE ANALOGS AGAINST THE AR-V7-EXPRESSING 22RV1 XENOGRAFT MODEL………..81

S4B: EVALUATE THE SAFETY AND TOLERABILITY OF GALETERONE ANALOGS ADMINISTERED VIA ORAL GAVAGE IN MALE ATHYMIC NUDE MICE……..82

SUMMARY AND CONCLUSIONS FROM SPECIFIC AIM 4………………………………...85

SUMMARY AND FUTURE DIRECTIONS…………………………………………………...86

REFERENCES……………………………………………………………………………..93

LIST OF TABLES:

TABLE 1: SUMMARY OF AR EXPRESSION AND ACTIVITY IN PROSTATE CANCER CELLS…...34

TABLE 2: SUMMARY OF MTT ASSAYS IN LNCAP CELLS...……………………………….35

TABLE 3: SUMMARY OF MTT ASSAYS IN 22RV1 CELLS….………………………………36

TABLE 4: SUMMARY OF MTT ASSAYS IN PC-3 CELLS....…………………………………37

LIST OF FIGURES:

FIGURE 1: THE PROSTATE GLAND………………………………………………………….2

FIGURE 2: THE HYPOTHALAMIC-PITUITARY-GONADAL/ADRENAL (HPG/A) AXIS…………5

FIGURE 3: SYNTHESIS OF STEROID HORMONES FROM CHOLESTEROL………………………6

FIGURE 4: THE STRUCTURE OF THE ANDROGEN RECEPTOR………………………………...8

FIGURE 5: ANDROGEN ACTION IN THE CELL………...……………………………………10

FIGURE 6: THE ANDROGEN RESPONSE ELEMENT (ARE) IS A PALINDROMIC SEQUENCE WITH TWO COMPLEMENTARY HEXAMERS SEPARATED BY THREE BASES ……………………………………………………...………11

FIGURE 7: THE ANDROGEN RECEPTOR (AR) GENE CONTAINS MULTIPLE CODING REGIONS THAT CAN BE ALTERNATIVELY INCORPORATED INTO PROTEIN…………………………………………………………………13

FIGURE 8: THE ENDOPLASMIC RETICULUM STRESS RESPONSE (ERSR)…………………...16

FIGURE 9: CHEMICAL STRUCTURES OF DIHYDROTESTOSTERONE (DHT), GALETERONE, AND GALETERONE ANALOGS…………………………………..33

FIGURE 10: DOSE-RESPONSE CURVES FOR ABIRATERONE, ENZALUTAMIDE, GALETERONE, AND THE TOP FIVE ANTIPROLIFERATIVE ANALOGS …………... 38

FIGURE 11: ANDROGEN RECEPTOR ANTAGONISTS INHIBIT EXPRESSION OF ARR2-LUCIFERASE REPORTER IN LNCAP AND 22RV1 CELLS……………….40

FIGURE 12: GALETERONE ANALOGS CAN REDUCE FULL-LENGTH AND TRUNCATED ANDROGEN RECEPTOR PROTEIN EXPRESSION IN 22RV1 CELLS………………………………………………………………...42

FIGURE 13: GALETERONE ANALOGS VNPT-178 AND VNLG-74A DOSE DEPENDENTLY INHIBIT FORMATION AND GROWTH OF ADHERENT 22RV1 COLONIES………………………………………………………...47-48

FIGURE 14: GALETERONE ANALOGS VNPT-178 AND VNLG-74A DOSE- DEPENDENTLY DECREASE AR AND PSA PROTEIN EXPRESSION IN LNCAP AND 22RV1 CELLS………………………………………………50-51

FIGURE 15: GALETERONE ANALOGS VNPT-178 AND VNLG-74A INHIBIT PROLIFERATION AND INDUCE CELL DEATH IN LNCAP CELLS………………...53

FIGURE 16: CELL CYCLE ANALYSIS OF LNCAP, 22RV1, AND PC-3 CELLS……………55-57

FIGURE 17: CONTINUOUS TREATMENT WITH ARDA ACIDIFIES MEDIA IN LNCAP AND 22RV1 CELLS AFTER 48 HOURS………………………………..59

FIGURE 18: REFRESHING MEDIA AFTER 24 HOURS IMPROVES THE AR- DEPLETIVE EFFECTS OF VNPT-178 AND VNLG-74A IN LNCAP AND 22RV1 CELLS…………………………………………………………...61

FIGURE 19: VNPT-178 AND VNLG-74A INDUCE APOPTOSIS IN PROSTATE CANCER CELLS……………………………………………………………….64

FIGURE 20: 22RV1 CELLS TREATED WITH 20 µM ARDA UNDERGO APOPTOSIS WITH A CONCOMITANT DECREASE OF AR EXPRESSION AND ACTIVITY THROUGH 72 HOURS…………………………….65

FIGURE 21: AR AND AR-V7 DEPLETION IS INSUFFICIENT TO INDUCE APOPTOSIS IN 22RV1 CELLS WITHIN 72 HOURS………………………………67

FIGURE 22: PC-3 CELLS ARE RESISTANT BUT SUSCEPTIBLE TO ARDA MEDIATED APOPTOSIS………………………………………………………..69

FIGURE 23: VNPT-178 AND VNLG-74A INDUCE PARP CLEAVAGE IN LNCAP CELLS……………………………………………………………….70

FIGURE 24: 20 µM ARDA INDUCES ACTIVATION OF THE UPR IN 22RV1 CELLS...……….72

FIGURE 25: PAN-CASPASE INHIBITION DOSE-DEPENDENTLY DECREASES ARDA-INDUCED PARP CLEAVAGE IN 22RV1 CELLS………………………..73

FIGURE 26: CALPEPTIN DOSE-DEPENDENTLY DECREASES ARDA-INDUCED PARP CLEAVAGE IN 22RV1 CELLS…………………………………………..75

FIGURE 27: GALETERONE ANALOGS DEMONSTRATE POTENTIAL TO BIND 70 KDA HEAT SHOCK PROTEINS IN SILICO...………………………………….77

FIGURE 28: GALETERONE ANALOGS MAY REDUCE 22RV1 TUMOR GROWTH IN VIVO………………………………………………………………………82

FIGURE 29: MASS MEASUREMENTS OF EXCISED 22RV1 TUMOR XENOGRAFTS …………...83

FIGURE 30: MICE EXPERIENCED POSITIVE WEIGHT GAIN THROUGHOUT THE XENOGRAFT STUDY ………………………………………………………….84

FIGURE 31: MODEL SUMMARIZING THE POTENTIAL ACTION OF ARDAS VNPT-178 AND VNLG-74A AGAINST THE ER CHAPERONE BIP ……………88

FIGURE 32: CELL-DEPENDENT DIFFERENCES IN RESPONSE TO ARDA OR ANDROGEN TREATMENTS…………………………………………………….92

vii

LIST OF ABBREVIATIONS:

178 VNPT-178

22RV1 CWR22RV1 74A VNLG-74A

ABIR ABIRATERONE

ADT ANDROGEN DEPRIVATION THERAPY

AF-1 ACTIVATION FUNCTION 1

AF-2 ACTIVATION FUNCTION 2

AKA ALSO KNOWN AS

AR ANDROGEN RECEPTOR

ARDA ANDROGEN RECEPTOR DEPLETING AGENT

ARE ANDROGEN RESPONSE ELEMENT

AR-V ANDROGEN RECEPTOR SPLICE-VARIANT

ATF-4 ACTIVATING TRANSCRIPTION FACTOR 4

ATF-6 ACTIVATING TRANSCRIPTION FACTOR 6

BIP BINDING IMMUNOGLOBULIN PROTEIN; AKA GRP78

BPH BENIGN PROSTATIC HYPERPLASIA

CDK CYCLIN-DEPENDENT KINASE

CHIP CARBOXY-TERMINAL OF HSP70 INTERACTING PROTEIN

CHOP CCAAT/ENHANCER-BINDING PROTEIN HOMOLOGOUS PROTEIN; AKA DDIT3, GADD153

CP CLIENT PROTEIN; AMBIGUOUS

CRPC CASTRATION-RESISTANT PROSTATE CANCER

CSS CHARCOAL/DEXTRAN-STRIPPED SERUM

CTD CARBOXY-TERMINAL DOMAIN

CYP17A1 CYTOCHROME P450 17A1

DBD DNA-BINDING DOMAIN

DDIT3 DNA DAMAGE INDUCIBLE TRANSCRIPT 3; AKA CHOP, GADD153

viii

DHT DIHYDROTESTOSTERONE

DRE DIGITAL RECTAL EXAM

EGCG EPIGALLOCATECHIN GALLATE

EIF2 EUKARYOTIC INITIATION FACTOR 2 (HETEROTRIMERIC COMPLEX OF Α, Β, Γ SUBUNITS)

ENZA ENZALUTAMIDE

ERSR ENDOPLASMIC RETICULUM STRESS RESPONSE; AKA UPR

ETOP ETOPOSIDE

FAR FULL-LENGTH ANDROGEN RECEPTOR

GADD153 GROWTH-ARREST AND DNA DAMAGE INDUCIBLE PROTEIN 153

GI GROWTH-INHIBITORY

GI50 50% GROWTH-INHIBITORY

GI90 90% GROWTH-INHIBITORY

GRP78 GLUCOSE-REGULATED PROTEIN 78 KDA; AKA BIP

HPG/A HYPOTHALAMIC PITUITARY GONADAL/ADRENAL

HSP HEAT SHOCK PROTEIN

IRE1Α INOSITOL-REQUIRING ENZYME 1 ALPHA

IRES INTERNAL RIBOSOME ENTRY SITE

KDA KILO DALTON

KLK3 KALLIKREIN-LIKE KINASE 3

LBD LIGAND-BINDING DOMAIN

LC3 MICROTUBULE ASSOCIATED PROTEIN 1 LIGHT CHAIN 3

LHRH LUTEINIZING HORMONE RELEASING HORMONE

MCRPC METASTATIC CASTRATION-RESISTANT PROSTATE CANCER

MDM2 MURINE DOUBLE MINUTE 2, HUMAN ORTHOLOG OF

NBD NUCLEOTIDE-BINDING DOMAIN

ND NOT DETERMINED

NLS NUCLEAR LOCALIZATION SEQUENCE

NTD AMINO-TERMINAL DOMAIN

PARP POLY ADENOSINE DIPHOSPHATE (ADP)-RIBOSE POLYMERASE

PERK PROTEIN KINASE R (PKR)-LIKE ENDOPLASMIC RETICULUM KINASE

PSA PROSTATE-SPECIFIC ANTIGEN; AKA KLK3

PTM POST-TRANSLATIONAL MODIFICATION

RIDD REGULATED IRE1-DEPENDENT DECAY

RLU RELATIVE LUCIFERASE UNIT

RPFS RADIOGRAPHIC PROGRESSION-FREE SURVIVAL

S1P SITE-1 PROTEASE

S2P SITE-2 PROTEASE

SBD SUBSTRATE-BINDING DOMAIN

SBMA SPINAL-BULBAR MUSCULAR ATROPHY

TF TRANSCRIPTION FACTOR

UPR UNFOLDED PROTEIN RESPONSE; AKA ERSR

VEH VEHICLE

XBP-1 X-BOX BINDING PROTEIN 1

XBP-1S X-BOX BINDING PROTEIN 1, SPLICED

XBP-1U X-BOX BINDING PROTEIN 1, UNSPLICED

INTRODUCTION:

Prostate Cancer at a Glance:

Prostate cancer continues as one of the most frequently diagnosed and lethal malignancies affecting American men. For 2017, the National Cancer Institute (1) estimates there will be over 160,000 new cases of prostate cancer accounting for nearly

10% of all new cancer diagnoses. Prostate cancer is an age-related, hormone responsive disease that is most frequently diagnosed in men between ages 55-74 with a median age of 66 at time of diagnosis. If detected early, prostate cancer can be curable and increased early screening and treatment identifies more than 90% of new prostate cancers at the local or regional level with five-year relative survival rates approaching 100%. However, not all men are as fortunate and an estimated 5% of new diagnoses will be of metastatic prostate cancer for whom there is an estimated five-year survival rate of less than 30%.

In total, prostate cancer is responsible for nearly 5% of all cancer-related deaths and an estimated 26,730 men will succumb to the disease this year.

Prostate Cancer Biology and Clinical Presentation:

The prostate gland itself has a poorly understood physiological role and is believed to contribute to improved sperm motility in the ejaculate. Situated below the bladder, the prostate gland surrounds the urethra and ejaculatory ducts and is divided into four primary zones (Figure 1) (2). The transitional zone surrounding the urethra is the most frequent site of benign prostatic hyperplasia (BPH) and is the origin of an estimated

20% of new prostate cancer cases. The central zone surrounds the ejaculatory ducts and

1 while fewer than 5% of new prostate cancer cases originate here, they are predicted to be more aggressive and likely to invade the seminal vesicles (3). The peripheral zone is the largest, occupying most of the prostate’s glandular tissue and is situated at the posterior of the gland nearest the rectal wall. Between 70-80% of prostate cancer cases originate here and can be detected as solid lumps during a digital rectal exam (DRE). Since age is the single most common factor for prostate cancer development and DREs can detect nearly 80% of all new cancer cases, prostate examination and cancer screening is typically recommended after age 50.

While early detection and treatment with surgery or radiation therapy have greatly improved overall prostate cancer survival statistics, as indicated, not all cases can be identified by palpation. PSA is a 33 kDa serine protease that is produced and secreted by the prostate epithelia and can be readily detected in blood sera thus making it a valuable biomarker for evaluation of prostate cancer development and progression. While circulating PSA levels are routinely monitored, its production and secretion can be highly variable on an individual basis making it as controversial as it is useful.

The only definitive tool for diagnosing prostate cancer is tissue biopsy. Prostate cancer diagnoses are rated on a four-point scale with stage 4 being the most advanced (4).

Stage 1 and 2 prostate cancers are locally confined, early or advanced with higher

Gleason or PSA scores, respectively. Stage 3 prostate cancers have spread beyond the outer layer of the prostate and may have spread to the seminal vesicles. Stage 4 prostate cancers are metastatic, having spread to nearby tissues and organs including lymph nodes, rectum, or bladder or have spread to distal locations including lung, liver or bone.

Surgery and radiation therapy are largely successful in treating locally confined prostate cancer and contribute greatly to the high relative survival rates. For patients where these treatment paradigms are not recommended, systemic therapies including hormone or chemotherapy are used as first-line options.

Hormone Therapy for Prostate Cancer:

Taxane-based chemotherapies (docetaxel, cabazitaxel) are routinely prescribed for prostate cancer patients but are broadly-acting cytotoxic agents that can cause serious

3 distress to rapidly-dividing normal cells of the gut and other tissues. Alternatively, hormone therapy is often prescribed for the treatment of all stages of prostate cancer (5) having fewer and less-severe side effects than chemo. Hormone therapy includes targeting both the synthesis and availability of androgens (androgen deprivation therapy,

ADT; also known as castration) as well as the specific action of androgens by blocking their interaction with the androgen receptor (AR; referred to as AR antagonism).

In humans, testosterone is the primary male sex hormone and nearly 90% of all testosterone synthesis occurs in the testes with the remaining 10% being converted from corticosteroids in the adrenals. Testosterone synthesis is regulated by the hypothalamic- pituitary-gonadal/adrenal (HPG/A) axis and can in turn regulate its own production via negative feedback inhibition (Figure 2). Traditionally, luteinizing hormone releasing hormone (LHRH) receptor agonists have been used in ADT although more recently,

LHRH receptor antagonists have been employed (6). LHRH receptor agonists promote the production of testosterone, resulting in a ‘testosterone surge’ that then shuts down further production via its negative feedback inhibition on the hypothalamus and pituitary.

LHRH receptor agonists will result in approximately 90% decline in serum testosterone within 2-3 weeks. However, using LHRH receptor antagonists block all downstream signal for testosterone production, eliminating the testosterone surge and reducing serum levels in as little as three days.

While hormonal therapy of this nature can be effective initially, it does not impair de novo synthesis of androgens in tissues outside of the HPA/G axis, namely within tumors themselves. Therefore, specific antagonists targeting the CYP17A1 enzyme, which catalyzes two sequential steps (17α-hydroxylase and 17,20-lyase) converting

4 pregnenolone or progesterone to the androgens dehydroepiandrosterone and androstenedione, respectively (Figure 3), were evaluated clinically. Abiraterone, administered as its prodrug abiraterone acetate, gained FDA approval when combined with prednisone in 2011 for the treatment of metastatic castration-resistant prostate cancer (mCRPC).

Testosterone is the predominant circulating androgen and freely traverses the plasma membrane, entering cells where it can be converted to its more potent metabolite, dihydrotestosterone (DHT) by the 5α-reductase enzyme. In a similar fashion to the

5 aromatase inhibitors used in breast cancer treatments, 5α-reductase inhibitors (finasteride and dutasteride) have been evaluated for use in treatment of BPH and prostate cancer.

However, the FDA warns users of these drugs for treating symptoms of BPH that they have increased risk of developing more aggressive prostate cancers as they can mask signs of its initial development (7). Ultimately, the androgen receptor (AR) is the primary effector of the androgen signaling axis leading to prostate cancer growth and

6 progression. Consequently, specific competitive AR antagonists (bicalutamide and enzalutamide) have been developed and routinely used alone or in combination with

ADT. Unfortunately, most men treated with systemic therapies targeting the androgen signaling axis will eventually progress to its lethal phenotype, mCRPC.

The Androgen Receptor (AR):

The androgen receptor is a member of the steroid hormone family of ligand- dependent nuclear receptors and is necessary for the development of the male phenotype including the prostate gland. The 919-amino acid AR protein is generated from a gene containing 8 exons found within the q12 locus on the X chromosome (Figure 4) (8). The structural organization of the AR protein contains three primary domains: a loosely- ordered amino-terminal domain (NTD) (9, 10) encoded by exon 1; a highly-conserved central DNA-binding domain (DBD) (8) encoded by exons 2 and 3; and a highly-ordered carboxy-terminal domain (CTD) encoded by exons 4-8. Within each primary AR domain are sub-domains and motifs with critical roles for the canonical ligand-mediated activation and translocation of the AR between the cytosol and nucleus (11). The AR-

CTD contains a hydrophobic cleft and alpha helix 12 forming the ligand-binding domain

(LBD), a nuclear localization sequence (NLS), a flexible hinge region, and an activation function (AF-2) domain. The AR-DBD contains two zinc finger motifs that coordinate its interaction with DNA. The AR-NTD contains two sequence motifs FQNLF and

WHTLF and an activation function (AF-1) domain with important roles in coordinating protein-protein interactions. The AF domains are critical recognition points for the AR’s ability to interact with numerous chaperone and transcription cofactor proteins.

The labile conformation of the AR-NTD enables the AR to interact with numerous proteins having diverse biological roles and cofactors able to promote the cell- and tissue-specific activation or repression of AR-regulated genes (12, 13). This feature also makes structure-based drug discovery of novel agents targeting the NTD very difficult and few have had much success in preclinical trials (e.g., EPI-001) (14, 15). The

AR is reported capable of regulating the expression of some 2000 genes due largely to its ability to interact with so many transcription factors and coregulators. More than 200 proteins have been identified to interact directly with the AR and another 100 more are known constituents of AR-containing heteroprotein complexes. Of these, many have biological roles outside of gene transcription including genome maintenance (16) and

DNA repair (17) as well as cell cycle regulation (18). Its broad range of potential actions can make targeting the AR dubious as some cells rely heavily on its actions while its

8 roles in other cells can often be redundant. While this is undoubtedly true on an individual patient basis as prostatic tumors are highly heterogenous, the general consensus remains that targeting the AR is a viable option for specifically inhibiting growth and progression of prostate cancer (19). Though the androgen receptor’s DBD is absolutely required for its transcriptional activity, the sequence homology of this domain is too well-conserved among other members of the nuclear hormone family providing daunting challenges for the development of AR-specific therapeutics targeting this domain (20). Consequently, the rigid structure of the AR-CTD and unique features of its

LBD have enabled researchers to identify multiple agents capable of selectively targeting the AR.

The androgen receptor is maintained in the cytosol in a high-affinity-for-ligand conformation by a heterocomplex of heat shock proteins (HSPs) and co-chaperones (21).

This heterocomplex most notably contains Hsp90; however, recruitment of Hsp90 to the

AR requires an ATP-dependent priming step coordinated by Hsp70. Upon ligand binding, the AR-CTD experiences conformational changes that cause dissociation of the

AR from its chaperones and translocation to the nucleus (22). Translocation is made possible by an N/C folding of the AR involving the FQNLF motif in the NTD and helix

12 of the CTD (23, 24) and the flexibility of the AR hinge region. This N/C interaction causes exposure of the NLS, allowing the AR to interact with the microtubule-associated dynein motors facilitating its translocation to the nucleus (25, 26). Within the nucleus, the AR can bind to exposed androgen response elements (AREs) within the promoter, enhancer, or intronic regions of its regulated genes (Figure 5) (27). The canonical ARE contains two complimentary hexamers of consensus sequence 5’-

AGAACAnnnTGTTCT-3’ (Figure 6) and variability in this sequence can determine the sensitivity of a gene to regulation by the AR under various conditions (28). For instance, it is proposed that deviations in this sequence, either in substitution of nucleotides or presence of only a half-site modulate the affinity of the AR-DBD and the presence of nearby response elements for other transcription factors can, in combination, greatly influence the AR’s transcriptional program (29, 30). Consequently, the regulation of

10 specific ARE-containing genes can be variably sensitive to ADT or AR-antagonism depending largely on the context of the ARE sequence and neighboring response elements; how these factors contribute to different tumor responses amidst therapeutic challenge are continually being investigated.

Mechanisms of Resistance to AR-Directed Prostate Cancer Therapies:

While the goal of antiandrogens is to deplete circulating androgens (i.e., ADT), most AR antagonists compete with androgens for binding the AR-LBD and induce a conformational change that inhibits its activation and nuclear translocation. However, numerous mechanisms of resistance to available interventions have been identified (31).

Depletion of androgens has seen challenge by AR gene amplification resulting in the increased expression of AR proteins sensitizing cancer cells to the very low androgen levels in a castrate environment (32). Numerous mutations in the AR have been

11 identified, conferring ligand promiscuity allowing the AR to be activated by non- androgens or the reversal of antagonists to agonists by modulating the ligand-induced conformational changes in the AR-CTD (33). The androgen receptor is the recipient of multiple post-translational modifications (PTMs) and altering the landscape of protein acetylases, kinases and phosphatases, or ubiquitinases and sumo proteins can promote the activation and nuclear translocation of the AR independently of ligand binding (34-36).

Perhaps the most significant and clinically challenging of all resistance mechanisms is the expression of alternatively spliced androgen receptor variants (AR-Vs). AR-Vs are the result of changed splicing activity of AR mRNA (37, 38) leading to the inclusion of cryptic exon sequences or premature stop codons and a representation of many known variants is summarized in Figure 7 (39). The end-product of AR-V expression is often a protein with C-terminal truncation whereby part or all the AR-CTD and consequently the

AR-LBD is absent. While the potential for these proteins to be constitutively active and localized to the nucleus has been known for many years (40), our knowledge of their clinical significance is continually growing (41). Two of the best-known AR-Vs, AR-V7

(42) (also known as AR3 (43)) and ARv567es (44) have been identified in mCRPC patients and the loss of the AR-LBD confers complete resistance to even the most-current antiandrogens (abiraterone) and AR antagonists (enzalutamide) (45).

Androgen Receptor Depleting Agents (ARDAs) and Galeterone

Considering the prolific adaptive responses to AR-directed therapies, particularly with respect to our increased awareness of AR-Vs (46), an approach of targeted and complete AR depletion has been largely adopted as a viable and superior next-generation

12 strategy. Multiple natural and synthetic small molecules have been described as potentiating the post-translational depletion of the androgen receptor and its splice

13 variants. These compounds are colloquially referred to as androgen receptor depleting or down-regulating agents (ARDAs). Many of the naturally occurring compounds are phytochemicals found within specific dietary foods and supplements including long peppers (piperlongumine) (47), green teas (epigallocatechin gallate; EGCG) (48), rosemary extract (carnosic acid) (49, 50), red sage (tanshinone) (51), and ginger varietals (curcumin) (52) while others are anti-bacterial (nigericin) (53) or anti- helminthic (niclosamide) (54) compounds. Niclosamide is currently enrolled in clinical trials in combination with abiraterone (NCT02807805) or enzalutamide (NCT03123978).

Galeterone, a steroidal anti-androgen was first developed similarly to abiraterone as an inhibitor of androgen biosynthesis but was subsequently identified to possess AR antagonistic and depletive activities in prostate cancer models (55). Subsequent efforts have focused on galeterone’s ability to promote the polyubiquitination and proteasome- mediated degradation of the AR and AR-V7 (56-59); however, the specific mechanisms underlying its activity and specific targeting of AR proteins remains to be elucidated.

Since its introduction to clinical trials in 2009 for the treatment of castration-resistant prostate cancer (60), galeterone has garnered much attention for its proposed activity against AR-Vs and was awarded fast track status for Phase 3 evaluation versus enzalutamide for the treatment of patients with AR-V7-expressing mCRPC

(NCT02438007). However, it was identified in 2016 that galeterone would fail to meet its primary endpoint of improved radiographic progression-free survival (RPFS) and was subsequently pulled from further evaluation. The failure of galeterone to perform as anticipated (59) raises question as to its ability to prevent AR-V7 activity and highlights a greater need for expanded understanding of its mechanism (s) of action.

Before its inclusion in clinical trials and the focus of attention on AR-V7, galeterone was evaluated as a potential drug candidate against prostate cancers that do not express the androgen receptor (61). While it was known then that galeterone could directly bind and antagonize the activity of both CYP17A1 and the AR resulting in potent attenuation of the androgen signaling axis, galeterone was actively being investigated for its ability to disrupt protein homeostasis through its activation of the endoplasmic reticulum stress response.

The Endoplasmic Reticulum Stress/Unfolded Protein Response (ERSR/UPR)

The endoplasmic reticulum is a cellular organelle that functions at the center of the folding and processing of proteins destined for secretion or insertion into the plasma membrane. Many cellular stresses can perturb ER function and, as a result, proteins can become misfolded or unfolded. Accumulation of these aberrant proteins within the ER lumen activates an evolutionarily conserved response pathway referred to as the endoplasmic reticulum stress response (ERSR) or the unfolded protein response (UPR)

(Figure 8) (62). At the center of UPR activation is the 78 kDa glucose-regulated protein

(Grp78), also known as binding immunoglobulin protein (BiP). BiP is a member of the

70 kDa heat shock protein family and transcribed by the HSPA5 gene. BiP has multiple functions within the ER lumen including the facilitation of proper protein folding through formation of multimeric chaperone complexes (63), maintaining ER calcium homeostasis

(64), and regulating the activation of the UPR (65). Like other members of the Hsp70 family, BiP contains an amino-terminal nucleotide-binding domain (NBD) with intrinsic

ATPase activity that regulates the conformation and affinity of its carboxy-terminal

15 substrate-binding domain (SBD). Hydrolysis of ATP to ADP results in a conformational change in the SBD increasing BiP’s affinity for its substrates and allows the formation of stable complexes with client proteins (66).

A generally accepted model of UPR activation revolves around the hydrophobic residues of unfolded proteins competing for BiP binding with the lumenal domains of ER transmembrane proteins PERK, IRE1α, and ATF-6 which function as sentinels for the

16 detection of ER stress and initiation of the three arms of the UPR. However, specific studies evaluating the interactions between BiP and PERK, IRE1α or ATF-6 reveal these protein interactions are stable and dissociation is more likely the result of altered ATPase activity, nucleotide exchange, or allosteric mechanisms (67-69). In either case, dissociation of BiP from these sentinel proteins initiates signaling events that attenuate the global translation of proteins, increase the protein-folding capacity of the ER, and promotes the proteasome-mediated or autophagic clearance of ER proteins.

Dissociation of BiP from PERK (protein kinase R (PKR)-like endoplasmic reticulum kinase) results in the homodimerization of PERK and subsequent trans- autophosphorylation of its cytosolic domain leading to activation of its kinase activity.

Activated PERK phosphorylates the alpha subunit of eukaryotic initiation factor 2

(eIF2α) (62), preventing the guanine-nucleotide exchange of eIF2β and subsequent recycling of eIF2 to its active state. Inactive eIF2 prevents the 5’CAP mediated translation of mRNA and halts global protein synthesis. Messenger RNAs containing an internal ribosome entry site (IRES) are unaffected and, consequently, proteins with functions to resolve ER stress including ATF-4 (activating transcription factor 4) are indirectly upregulated. Similarly, dissociation of BiP from IRE1α (inositol-requiring enzyme 1 alpha) results in its homo-oligomerization and trans-autophosphorylation, activating its cytosolic endoribonuclease activity (70). Active IRE1α can degrade non- essential mRNAs through a process known as regulated IRE1-dependent decay (RIDD) or facilitate a unique processing of XBP1 (X-Box binding protein 1) mRNA, resulting in the inclusion of an exon that results in the translation of a functional transcription factor,

XBP-1s. ATF-6 (activating transcription factor 6) dissociation from BiP unmasks a

Golgi-localization signal (71) allowing ATF-6 to translocate to the Golgi where it is proteolytically processed by site-1 and site-2 proteases (S1P and S2P, respectively).

Cleavage of ATF-6 liberates its cytosolic domain—an active transcription factor that translocates to the nucleus.

While the activation of each of the three arms of the UPR generally promotes recovery from ER stress and cell survival, prolonged or severe activation drives the upregulation of genes promoting apoptotic cell death (72-74). In addition to suppressing the UPR, BiP also functions as a gatekeeper against the free diffusion of calcium through the Sec61 translocon complex. The Sec61 translocon is an ER transmembrane pore complex that interacts with active ribosomes in the cytosol to facilitate the insertion of nascent protein chains into the ER lumen. In the absence of the protein chain, BiP associates with the lumenal face of Sec61 closing off the pore (75). Deregulated calcium homeostasis is known to induce apoptosis through a variety of mechanisms (76-80).

Although it was previously determined that galeterone-induced ERSR only promoted quiescence (61), the concentrations of calcium within different microenvironments of the cell can be an influential factor in determining Ca2+ signaling events.

PROJECT OVERVIEW AND SPECIFIC AIMS:

Our primary objective herein is to identify a mechanism by which galeterone analogs can inhibit the proliferation of the AR-V7-expressing CWR22Rv1 cell model.

From a library of novel compounds based on the structure of galeterone, we will identify potential leads that exhibit improved antiproliferative action and disruption of androgen receptor activity through its antagonism and depletion. From here, we will evaluate the potential inhibition of cell cycle progression and/or induction of cell death and attempt to identify a specific modality of inhibited proliferation. Additionally, we will investigate the in vivo safety and efficacy of these leads prior to further preclinical evaluation. Taken together, we anticipate these data to support an investigative bridge between the androgen receptor and a next-generation paradigm of therapeutic intervention of prostate cancer development and progression.

Specific Aim 1: Screen a library of novel galeterone analogs for improved antiproliferative and AR-antagonistic activities in vitro.

Hypothesis: New compounds based on the structure of galeterone that demonstrate improved in vitro antiproliferative kinetics will also exhibit improved anti-AR activities.

S1a: Compare the antiproliferative activity of novel galeterone analogs to abiraterone

and enzalutamide using the MTT assay.

S1b: Compare the potential of galeterone analogs to inhibit full-length androgen

receptor transcriptional activation using a transiently-transfected dual luciferase

reporter system.

S1c: Determine which of the novel analogs can deplete the endogenous expression of

the androgen receptor in 22Rv1 cells.

Specific Aim 2: Evaluate the dose- and time-dependent biological responses to treatment with galeterone analogs to identify a potential therapeutic window.

Hypothesis: Based on prior studies with galeterone, we anticipate a dynamic relationship whereby a concentration range of 1-20 µM and a time range of 0-96 hours is sufficient to induce inhibition of prostate cancer cell growth.

S2a: Determine if longer treatments with lower doses than the predicted GI50 values

could improve relative activity of galeterone analogs to established controls.

S2b: Determine the minimum concentration required to elicit AR antagonism and

depletion within 24 hours in LNCaP and 22Rv1 cells.

S2c: Evaluate cell death associated with dose and time of treatment with VNPT-178

and VNLG-74A using the Trypan Blue exclusion assay.

S2d: Evaluate changes in cell cycle progression following treatment with increasing

doses of VNPT-178 and VNLG-74A.

Specific Aim 3: Identify a potential mechanism of action of VNPT-178 and VNLG-

74A for the induction of death in cellular models of prostate cancer.

Hypothesis: Galeterone analogs induce apoptosis independently of an inhibition of the androgen receptor signaling axis.

S3a: Determine whether galeterone analogs induce necrosis or apoptosis.

S3b: Determine the AR-dependency of cell death induction in prostate cancer cells.

S3c: Determine the involvement of the ERSR/UPR for induction of cell death in

prostate cancer cells.

S3d: Identify novel targets of galeterone analogs VNPT-178 and VNLG-74A.

Specific Aim 4: Evaluate the safety and efficacy of galeterone analogs in vivo using a 22Rv1 tumor xenograft model of castration-resistant prostate cancer.

Hypothesis: The improved in vitro antiproliferative activities of selected galeterone analogs will translate to more efficacious treatments in vivo. The safety profile of galeterone will largely be maintained and these compounds will be well tolerated in vivo.

S4a: Determine the growth-inhibitory effects of galeterone analogs against the AR-

V7-expressing 22Rv1 xenograft model.

S4b: Evaluate the safety and tolerability of galeterone analogs administered via oral

gavage in male athymic nude mice.

MATERIALS AND METHODS:

Cell Lines and Culture

LNCaP (CRL-1740), 22Rv1 (CRL-2505), and PC-3 (CRL-1435) cells were purchased from ATCC (Manassas, VA) and cultured in Phenol red-containing RPMI-

1640 media (Mediatech; Manassas, VA) supplemented with 10% standard fetal bovine serum (GE Healthcare Life Sciences; Logan, UT) and 1% penicillin:streptomycin

(Gemini Bio-Products; West Sacramento, CA). All culture and treatments were performed using tissue culture-treated polystyrene dishes (Corning Inc.; Corning, NY) in humidified incubators at 37°C and 5% CO2. Unless otherwise indicated, media was refreshed at 24-hour intervals during treatments to maintain optimum metabolic conditions and emulate in vivo dynamics. For experiments running longer than 24 hours, but not multiples of 24, a media refresh schedule was customized to allow a final, 24- hour period before sample collection.

Compounds

Galeterone, DHT, VNPT-178, VNLG-74A, and abiraterone were synthesized in- house as previously described (55, 58). Enzalutamide (a.k.a. MDV3100, #SRP016825m) was purchased from Sequoia Research Products (Pangbourne, UK). Methyltrienolone

(a.k.a. R1881, #965M935) was purchased from OChem Incorporation (Des Plaines, IL).

Z-VAD-FMK (FMK001) was purchased from R&D Systems (Minneapolis, MN).

Etoposide (E1383) was purchased from Sigma (St. Louis, MO). Calpeptin (SC-202516) was purchased from Santa Cruz Biotechnology (Dallas, TX). All compounds were

22 solubilized in 95% ethanol (Pharmco-AAPER; Brookfield, CT) or ultrapure DMSO

(AmericanBio; Natick, MA).

MTT Cell Viability

Cells were seeded in 96-well plates at a density of 2000 (PC-3) or 3000 (LNCaP

& 22Rv1) cells per well in 100μL culture media and allowed 24 hours for attachment.

The next day, compound stocks were diluted as a 2x formulation in culture media and

100μL was added topically to each well to a final concentration within the range of 10 nM-100μM (n=6). All plates were returned to culture incubators. After 72 hours, media was aspirated and refreshed (200μL with indicated concentrations of each compound) and plates were returned to incubators for another 72 hours. On the final day, Thiazolyl Blue

Tetrazolium Bromide (M2128; Sigma; St. Louis, MO) was solubilized in culture media as a 5x formulation and added topically to each well to a final concentration of 0.5% w/v and cells were given 2 hours of incubation before the media was aspirated and replaced with 100μL DMSO. To promote cell lysis, plates were agitated on a BT1500 orbital shaker (Benchmark; Edison, NJ) for 10 minutes at 500 rpm. Absorbance was measured at 570 nM wavelength light using an ELx800 plate reader and Gen5 software (BioTek

Instruments; Winooski, VT). Data was exported to and organized in Microsoft Excel;

50% and 90% Growth Inhibitory (GI) values were calculated based on a non-linear regression curve fit using GraphPad Prism4 software (La Jolla, CA). Plots represent the absorbance mean ± SEM of each [compound] versus vehicle-treated controls from at least three independent experiments.

Colony Formation

22Rv1 cells were seeded in 6-well plates at a density of 500 cells per well in 2mL culture media (refreshed 2x weekly) containing vehicle or compound (1 or 5μM) and adherent colony formation was evaluated on day 15 using 0.05% (w/v) Crystal Violet

Staining Solution (10mM PBS containing 1% (v/v) Formaldehyde and 1% (v/v)

Methanol; pH 7.4). Images of the 6-well plates were generated using a digital scanner.

Colonies were manually counted (>30 cells) and total colony area was calculated using

ImageJ software with “Colony Counter” and “ColonyArea” plugins. Representative images and plots of the mean ± standard deviation of treated groups versus vehicle controls are shown. Statistical analysis was performed using Student’s t test and the significance level was determined by the largest p-value within each treatment group.

Luciferase Reporter

Competent E. coli (JM109) were transformed by heat-shock to express either the firefly or Renilla luciferase reporters and grown on ampicillin-containing agar plates.

The firefly luciferase gene is fused to an ARR2-modified rat Probasin promoter (ARR2-

Luc; kindly provided by Dr. Yun Qiu) while the Renilla luciferase reporter is pRL-null

(Promega, #E2271). A single bacterial colony containing each reporter was selected and grown in 40mL LB broth overnight (37°C, 220 rpm constant agitation). Plasmid DNA was purified using an EndoFree Plasmid Maxi Kit (Qiagen, #12362) and quantified using a BioPhotometer Plus (Eppendorf, #6132). LNCaP and 22Rv1 cells were seeded in 12- well tissue culture-treated plates and grown overnight in phenol red-free RPMI-1640 supplemented with 10% charcoal/dextran-treated FBS and 1% penicillin/streptomycin

(C/D Media) to reduce androgen availability. The cells were co-transfected with the

24 firefly and Renilla luciferase reporters using Effectene (#301425) transfection reagents in accordance with the manufacturer’s suggested protocol (Qiagen; Hilden Germany) for six hours in C/D media. After transfection, media was aspirated and replaced with C/D media containing vehicle or 10μM of each compound for a four-hour pre-treatment before the addition of DHT (C/D media with vehicle or 10μM compound and 20 nM

DHT to be added topically 1:1 to each well to a final concentration of 10 nM). Cells were incubated for an additional 20 hours at 37°C and 5% CO2. Media was aspirated from each well and 200μL of 2.5X Passive Lysis Buffer (Promega) was added. Plates were agitated on automated plate shakers for four hours at 4°C to complete cell lysis.

Full-length AR transcriptional activation was analyzed using a Dual-Luciferase Reporter

Assay System (E1980) in accordance with the manufacturer’s recommendations

(Promega; Madison, WI). Luciferase activity was measured in white 96-well flat bottom plates (Corning) using a Tecan M1000 plate reader. For three independent experiments

(n=4), mean relative luciferase units (RLU) and standard errors were calculated using

Excel and presented as fold-change versus vehicle-treated positive control.

AR/AR-V7 Knockdown

Short hairpin RNAs (shRNA) specific for the human androgen receptor (AR) and

AR-V7 (aka AR3) were packaged into the pLKO.1-puro vector for the transient transfection of 22Rv1 cells. The oligo sequences used for plasmid construction are as follows: fAR 5’-TGC ACT GCT ACT CTT CAG CAT TCA AGA GAT GCT GAA

GAG TAG CAG TGC TTT TTT C-3’ and 5’-TC AGA AAA AAG CAC TGC TAC TCT

TCA GCA TCT CTT GAA TGC TGA AGA GTA GCA GTG CA-3’; AR-V7 5’-TGT

AAT AGT GGT TAC CAC TCT TCA AGA GAG AGT GGT AAC CAC TAT TAC

TTT TTT TTC-3’ and 5’-TCG AGA AAA AAA AGT AAT AGT GGT TAC CAC TCT

CTC TTG AAG AGT GGT AAC CAC TAT TAC A-3’.

22Rv1 cells were seeded in 6 well plates at a density of 150,000 cells per well and incubated overnight. The following day, the media was aspirated and refreshed containing an optimal dosage of mature lentivirus (empty vector, fAR shRNA, or AR-V7 shRNA). 22Rv1 cells were cultured in the presence of the lentivirus for 48 hours before the media was aspirated and refreshed with media containing vehicle or VNPT-178 for another 24 hours.

Microscopy

Representative LNCaP, 22Rv1, and PC-3 cell images were taken at 200x total magnification using a Fisher Scientific Micromaster inverted digital microscope and the

Micron software. Images shown are a 300x300 pixel crop (600x600 for vehicle controls) of the original 3MP (2048x1536) image for clarity purposes.

Western Blotting

Antibodies (primary 1:500) purchased from Santa Cruz Biotechnology (Santa

Cruz, CA) include: AR N20 (SC-816, Lot# I1014), p21 (SC-397, Lot# K300).

Antibodies (primary 1:1000; secondary 1:3000) purchased from Cell Signaling

Technology (Boston, MA) include: PSA/KLK3 (#5877, Lot 1), PARP (#9542, Lot 14),

Caspase-3 (#9665, Lot 7), β-Tubulin (#2128, Lot 3), GAPDH (#2118, Lot 8), CHOP

(#2895, Lot 3), LC3-A/B (#12741, Lot 3), Beclin-1 (#3495, Lot 2), PERK (#5683, Lot

5), (p)eIF2α (#3398, Lot 2), eIF2α (#5324, Lot 3), ATF-4 (#11815, Lot 4), ATF-6

(#65880, Lot 1), IRE1α (#3294, Lot 9), XBP-1s (#12872, Lot 3), Calnexin (#2679, Lot

4), BiP/Grp78 (#3177, Lot 8), ERO-1Lα (#3264, Lot 4), PDI (#3501, Lot 3), anti-

Rabbit IgG (#7074, Lot# 26), anti-Mouse IgG (#7076, Lot# 31). Antibodies (primary

1:1000) purchased from Abcam (Cambridge, MA) include: (p)PERK (ab192591, Lot

GR185975-32). Whole cell lysates were collected for each condition in RIPA buffer

(#R0278; Sigma) supplemented with 0.02% EDTA (#E8008; Sigma), Complete protease inhibitor cocktail (#11697498001; Roche), and phosphatase inhibitor cocktails 1 and 2

(#P2850, #P5726; Sigma). Samples containing 15-30μg total protein were resolved by

SDS-PAGE using 8, 10, or 12% polyacrylamide gels. Proteins were captured on PVDF membrane (#162-0177) purchased from Bio-Rad Laboratories (Hercules, CA) and HRP- conjugated secondary antibodies were detected using SuperSignal West Pico chemiluminescent substrate (#34080) purchased from Thermo Scientific (Rockford, IL) and X-Ray film.

Molecular Docking

The ligands VNPT-178 and VNLG-74A were docked to the AR, BiP/Grp78 and

Hsp70-1A binding sites by sampling their rotational and translational degrees of freedom and the generated poses were subsequently scored by their estimated free energy of binding (ΔG). Docking computations were performed on the AR-LBD, BiP/Grp78, and

HSP70-1A crystal structures (Protein Data Bank: 2PNU, 3IUC and 5AR0, respectively).

The ligand SMILES strings were first converted to 2D MOL format by Indigo (EPAM) then to 3D MOL format by OpenBabel (81). Subsequently, the 3D MOLs were converted to PDBQT format by using a ligand preparation script of ADT (82). The Vina docking algorithm developed by Oleg Trott et al. was used to perform the docking computations

(83). The initial binding center coordinates were derived from the sc-PDB database or

27 original ligands co-crystallized in the structures (84). Essential hydrogen atoms and

Gasteiger charges were added. Non-polar hydrogens, lone-pairs, water molecules and non-standard residues were removed. Position, orientation, and torsions of the ligand molecule were initialized by a random seed. The binding search space was set to 22Å per dimension, and the exhaustiveness of conformational sampling was set to eight to balance search thoroughness and simulation time.

Cell Cycle Analysis

Cells were seeded onto 10cm tissue culture-treated petri dishes at a density of

5 6x10 cells per dish and incubated overnight at 37°C and 5% CO2. Cells were treated with ARDA or controls at indicated concentrations for 96 hours before a gently being washed and detached by trypsinization. Cells were gently pelleted by centrifugation, washed in DPBS to remove residual media and trypsin solution, and resuspended in ice- cold ethanol (95%) during gentle vortexing to prevent aggregate formation. Cells were fixed for at least 24 hours at -20°C. Ethanol was diluted out through two wash steps in

DPBS and cells were suspended at 1x106 cells per mL in DPBS containing 10 µg/mL

RNase A for at least 30 minutes at room temperature. Cell number was estimated based on pellet size. Propidium Iodide solution (#421301, BioLegend; San Diego, CA) was added per the manufacturer’s recommendations and cells were incubated on ice for at least 15 minutes before data acquisition and analysis. Cell cycle phase population data was acquired for 10,000 cells per condition using a BD FACSCanto II flow cytometer

(Becton Dickinson; Franklin Lakes, NJ) and analyzed using FlowJo software (FlowJo;

Ashland, OR). Cell cycle phase populations are plotted as the mean ± S.D. of total viable cells (G1+ S + G2/M) from three independent experiments.

22Rv1 Tumor Xenografts

Male athymic nude mice age 5 weeks were acquired from the National Cancer

Institute-Frederick Cancer Research Center (Frederick, MD), tagged, and provided a week of acclimation time within UMB’s animal facility before initiation of the study.

Mice were housed in a pathogen-free environment under light- and humidity-controlled conditions and provided food and water ad libitum. At 6 weeks of age, mice flanks were injected subcutaneously with 2x106 22Rv1 cells suspended in 1:1 mix of RPMI-1640 media and Matrigel basement membrane matrix (#354234, Corning; Corning, NY). Mice were monitored until tumors developed to an average of at least 90mm3 and stratified into five groups of six mice using a random number generator (RNG; www.random.org/).

Mice were ordered based on estimated initial tumor volume and assigned a treatment group using an RNG-based system to ensure that each treatment group had a mouse from each of 6 strata of tumor volumes.

The average mouse body weight was used to determine the formulation of compound mixtures in β-cyclodextrin. 100 µL of vehicle or galeterone, VNPT-178 and

VNLG-74A at a concentration of 148mmol/L or enzalutamide at a concentration of

25mg/kg were administered twice daily by oral gavage. Enzalutamide was intended on a schedule of 25mg/kg once daily and corrected after day 7 to include 1 administration of drug and 1 administration of vehicle for the remainder of the experiment. On day 8 it was discovered that one mouse from the VNPT-178 treatment group had died from an apparent esophageal or gastric hemorrhage. Mice were observed daily for signs of overt toxicity. Tumor volumes were estimated using digital caliper measurements and the formula V = (L x W x W)/2 where L represents the longest dimension and W represents

29 the shortest dimension. Tumor volume estimates and body mass measurements were recorded twice weekly. After a single group reached an estimated mean tumor volume of

1750mm3 all mice were euthanized and tumors extracted for mass measurements. Plots represent the mean ± standard error for each condition. All mice were handled in accordance to UMB policies and an IACUC approved methods (protocol #0512013).

Statistical Analysis

Where applicable, significance of change between treatment groups and vehicle controls were evaluated using a two tailed, two-samples of equal variance Student’s T test in Excel. Significance is determined at the 95, 99, and 99.9% confidence intervals.

SPECIFIC AIM 1:

A library of novel galeterone analogs has been synthesized featuring modifications to the carbon-carbon bond organization within the steroid backbone, the C3 or C16 positions, or the C17 benzimidazole moiety of galeterone. A partial account of these analogs is depicted in Figure 9. Within this specific aim, our goal is simply to identify which compounds demonstrate both an antiproliferative effect and attenuation of androgen receptor activity. Antiproliferation will be judged based on the estimated dose- dependent decline in cell number of analog-treated cells versus vehicle treatment and compared against galeterone and the clinically-available controls abiraterone and enzalutamide. Antagonism of the full-length androgen receptor will be determined using a dual luciferase reporter assay in a direct competition model. Depletion of androgen receptor at the protein level will be performed using Western blotting of whole cell lysates following dose- and time-dependent treatments in the 22Rv1 cell model.

Maintaining alignment with the currently accepted paradigm of androgen receptor- dependent prostate cancer development and progression, identification of new lead compounds that effectively impair both androgen receptor action and cell proliferation will establish a foundation for successful, further preclinical evaluation of lead drug candidates.

Specific Aim 1a: Compare the antiproliferative activity of novel galeterone analogs to abiraterone and enzalutamide using the MTT assay.

Our first objective is to establish a list of potential leads following an assessment of the net antiproliferative activity of these compounds versus galeterone and the clinical controls, abiraterone and enzalutamide. Using the MTT assay, LNCaP, 22Rv1, and PC-3 cells were seeded in 96-well plates and treated with each analog or control within a concentration range of 10 nM-100 µM for six days. These three cell lines were selected based on their diverse dependence on androgens and the androgen receptor and together represent a comprehensive model of prostate cancer progression from androgen dependence to both androgen and androgen receptor independence (Table 1). Following treatment, relative cell numbers were determined by the metabolism of MTT reagent (3-

(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) and analysis of absorbance at 570 nM wavelength light. Non-linear regression curves of the mean absorbance for each concentration were generated using GraphPad Prism software and a summary of the estimated concentrations whereby 50% and 90% growth inhibition (GI50 and GI90) occurred within each cell line, as well as the inverse fold-change to each of the three controls, is provided in Tables 2-4.

Arbitrarily, we assigned conditional formatting to the table to identify compounds that were at least twice as potent as abiraterone and enzalutamide and improved upon galeterone by at least 15% (Green background). Instances in which compounds were less than half as effective as abiraterone or enzalutamide or 85% of galeterone are highlighted in red. White boxes reflect instances in which only marginal improvements were observed and dashes indicate conditions in which the GI50 or GI90 concentration could

32 not be determined (ND) or where an estimated baseline value exceeded 100 µM and therefore irrelevant for comparison here. As our primary objective revolves around

33 identifying compounds that may be more effective in prostate cancers expressing AR-V7, the greatest emphasis for selection was placed on performance within the 22Rv1 cell model. We therefore selected five leads for continued evaluation and a depiction of their regression curve-fit plots are provided (Figure 10). An understated observation when performing these assays is the shape of the plots that can be generated. If a curve is generated with too steep a slope, the therapeutic window may be very narrow and can provide difficult to work within depending on the stability and action of the compounds.

Conversely, if it is too shallow, the compound may require a concentration that is difficult to achieve in vivo due to poor solubility or bioavailability.

Specific Aim 1b: Compare the potential of galeterone analogs to inhibit full-length androgen receptor transcriptional activation using a transiently-transfected dual luciferase reporter system.

As the androgen receptor (AR) remains a primary target for systemic therapies against prostate cancer, our next objective was to determine the ability for the five selected galeterone analogs to impair the transcriptional activation of the AR. Using an established method of selectively identifying androgen-sensitive gene expression, LNCaP and 22Rv1 cells were transfected with dual luciferase reporters. The first reporter is the firefly luciferase gene fused to an ARR2-modified rat Probasin promoter. The ARR2 denotation identifies two sequential androgen response elements (AREs). The second reporter is a promoter-less Renilla luciferase gene construct to be used as a negative control for non-specific activation. LNCaP and 22Rv1 cells seeded into 12-well culture plates were incubated for 24 hours in media supplemented with charcoal/dextran-stripped serum (CSS) to deplete residual androgens before transfection. Following transfection, the cells were pre-incubated with 10 µM of each compound before addition of 10 nM dihydrotestosterone (DHT) to stimulate AR transcriptional activity. After addition of luciferin substrate, the relative luciferase units (RLUs) of firefly and Renilla luciferase activity was determined using a light spectrophotometer and the mean fold-change RLUs of three independent experiments are plotted (Figure 11). As negative controls, cells that were exposed to both constructs in the absence of Effectene reagent or were transfected normally but not exposed to DHT gave a virtually-zero reading of light production.

Addition of 10 nM DHT serves as the primary positive control of AR-mediated luciferase expression and activity (85). Abiraterone, enzalutamide, and galeterone all serve as

39 secondary positive controls for the inhibition of DHT-mediated activation of the AR.

While all five analogs performed better than both abiraterone and galeterone, only

VNPT-178 and VNLG-74A exhibited strong AR antagonism nearly equaling that of enzalutamide. As mentioned, this assay depicts the androgen-sensitive activation of the firefly luciferase construct, i.e. the transcriptional activity of the full-length AR, and successfully excludes any potential AR-V7 activity (compare—vehicle only).

Specific Aim 1c: Determine which of the novel analogs can deplete the endogenous expression of the androgen receptor in 22Rv1 cells.

As has been proposed by many groups, depletion of the AR protein may serve to improve upon direct AR antagonism by further reducing the pool of activatable receptors.

However, a potential caveat also exists as it has been identified that androgen-bound AR can bind elements in its own gene enhancer regions and regulate its further expression.

This is believed to function as a negative feedback loop to maintain appropriate AR protein levels in the normal prostate. As androgen deprivation therapy (ADT) may contribute to relief of this negative feedback inhibition and thus promote the upregulation of the AR, we sought to determine whether galeterone analogs could successfully promote AR depletion in the context of ADT. 22Rv1 cells were seeded in 10cm petri dishes and cultured overnight in CSS-supplemented media to deplete residual androgens.

Galeterone or its analogs were administered for 3-24 hours in a hybrid time- and dose- response assay and whole cell lysates were collected for evaluation using SDS-PAGE and

Western blotting. The relative abundance of full-length and truncated AR isoforms was compared against beta-tubulin and plotted (Figure 12). Representative blots are provided below the plots for each compound tested. From these data, we can see that administration of galeterone and its analogs promote an initial rise in AR protein expression that can be attenuated at later time points or with higher concentrations.

Summary and Conclusions from Specific Aim 1:

The MTT assays allowed us to sort compounds based on their relative potencies in comparison with galeterone, abiraterone, and enzalutamide. As the 22Rv1 cell model represents a challenging, clinically relevant model of aberrant AR expression and activity, it served as our priority for selection. The galeterone analogs VNLG-74A,

VNPP-334, VNPP-397, VNPT-178, and VNPT-191 demonstrated improvement in at least half of the direct, fold-change comparison tests against abiraterone, enzalutamide, and galeterone in 22Rv1 cells with similarly positive results in the LNCaP and PC-3 models. In the presence of 10 nM DHT, VNPT-178 and VNLG-74A performed nearly equally as well as the potent AR antagonist enzalutamide at inhibiting the AR-mediated transcriptional activation of the ARR2-luciferase construct. These data are particularly important as not only do they narrow the selection to two leads, they also indicate a potential for specific action versus broad cytotoxicity. The Western blots within 22Rv1 cells reveal an additional potential activity extending beyond the full-length androgen receptor with the depletion of the truncated AR-V isoform as well.

Taken together, these data helped us identify five galeterone analogs with improved antiproliferative activities in three diverse cellular models of prostate cancer.

Further consideration of full-length androgen receptor antagonism and potential depletion of total androgen receptor protein expression helped us identify VNPT-178 and VNLG-

74A as our top two leads for characterization. Interestingly, though, the predicted GI50 and GI90 values for these two analogs do not appear to be dependent on AR status and deviate only slightly between the three cell lines examined. While it therefore seems these compounds can interact with at least two targets in prostate cancer cells, it is

43 nonetheless important to establish an origin of anti-AR activity. From here we can more reliably expand our investigation into mechanisms of action while potentially maintaining at least partial selectivity for malignant prostatic versus normal tissues.

Of final note is the marked structural differences between VNPT-178 and VNLG-

74A about the C3 position. The bulky C3 substituent within VNPT-178 may be tolerated at lower concentrations (around the GI50) but offer an advantage at higher concentrations

(around GI90). We therefore maintain an open mind that the AR may be a secondary target that can modulate a cell’s response to disruption of a primary target or that its antagonism within the time points evaluated here is inconsequential.

SPECIFIC AIM 2:

Our second objective is to determine a potential experimental window within which a specific dose and duration of treatment can reliably inhibit the proliferation and

AR signaling of prostate cancer cells in vitro. For the preclinical evaluation of new drug candidates, it is important to know an approximate baseline or minimum effective concentration and its temporal dependence. This is especially important when targeting transcription factors (TF) as the anticipated outcome is potentially dependent upon the loss or gain of multiple TF-regulated gene products over time. We will therefore expand upon the MTT assays to determine if lower concentrations over longer durations improve the antiproliferative outcome compared to additional controls. Next, we will evaluate the minimum concentration needed to elicit marked depletion of endogenous AR and PSA expression in LNCaP and 22Rv1 cells within 24 hours. Finally, we will determine whether changes in proliferative kinetics are dependent upon increased rates of cell death or attenuated cycle progression. Together, these data will help us identify a modality in which a desired combination of impaired AR signaling and cell growth can be reliably tested. Based on prior galeterone studies, we hypothesize that ARDA treatments within a concentration range of 1-20 µM for a duration of 0-96 hours will effectively disrupt AR signaling and inhibit prostate cancer cell growth.

Specific Aim 2a: Determine if longer treatments with lower doses than the predicted GI50 values could improve relative activity of galeterone analogs to established controls.

With the identification of two potential lead galeterone analogs, our next objective was to determine whether lower doses treated over longer time points could improve upon the relative antiproliferative activity of the compounds compared to established controls. Using an adherent colony formation assay, we evaluated the relative antiproliferative activities of VNPT-178 and VNLG-74A compared to multiple controls.

The MTT results in the 22Rv1 model predict cell growth could be inhibited by approximately 50% with the addition of 5 µM of either analog. As galeterone studies have identified decreased cell cycle kinetics as a primary means of activity in PC-3 cells, we anticipate this effect may be more pronounced extending treatments from six to 15 days. 22Rv1 cells seeded at very low density in 6-well plates were treated with various anticancer agents and the number and size of colonies formed were determined (Figures

13A and 13B).

Our MTT data predicted that approximately 10 or 20 µM of abiraterone or enzalutamide, respectively, would inhibit 22Rv1 proliferation by 50%. Galeterone was slightly more potent than abiraterone with an estimated GI50 of 7.5 µM. Here, 5 µM concentrations of abiraterone and galeterone demonstrated slightly improved kinetics in

22Rv1 cells with an approximate reduction of colony number and size by 50%.

Enzalutamide, on the other hand, failed to attenuate the growth of 22Rv1 colonies. Three additional controls were included here for comparison. Palbociclib is a potent CDK4/6 inhibitor and is being investigated for use in triple-negative breast cancers. In this assay,

46 inhibition of cell cycle progression with 0.5 µM Palbociclib reduced colony formation by approximately 80% compared to vehicle-treated controls. CGP-57380 is an experimental

Mnk1 kinase inhibitor. Disruption of global protein translation with 5 µM CGP-57380 reduced 22Rv1 colony formation nearly as well as both abiraterone and galeterone.

VNPT-55, a galeterone analog, has been reported to target multiple cellular pathways

47 including the AR axis, cell cycle progression, and the 5’CAP-mediated translational complexes. VNPT-178 and VNLG-74A appear to parallel the activities of VNPT-55 as 5

µM concentrations significantly reduced 22Rv1 colony formation. An important observation here, though, is the relative lack of activity by any galeterone analog at 1

µM.

Specific Aim 2b: Determine the minimum concentration required to elicit AR antagonism and depletion within 24 hours in LNCaP and 22Rv1 cells.

Dr. Bruno, previous member of our lab, estimated that while galeterone could effectively antagonize the CYP17A1 enzyme in the nanomolar range, more than 1 µM

48 would be necessary for effective antagonism of the androgen receptor and more than 10

µM is required for significant depletion of AR proteins. Additionally, it was determined that upwards of 20 µM concentrations of galeterone could be achieved in vivo. To create a parallel to these studies, we evaluated increasing concentrations of VNPT-178 and

VNLG-74A for the ability to antagonize and deplete endogenous AR proteins. LNCaP and 22Rv1 cells seeded in 10cm petri dishes were treated with 1, 2.5, 5, 10, or 10 µM

VNPT-178 or VNLG-74A for 24 hours and AR and PSA expression in whole cell lysates were compared to vehicle-treated controls (Figures 14A and 14B). Predictably, increasing concentrations of both analogs resulted in a nearly linear depletion of PSA expression in LNCaP cells. AR depletion experienced a lag-phase reminiscent of treatments in CSS conditions and, in accordance with Dr. Bruno’s observations, was markedly enhanced with 10 and 20 µM concentrations of both compounds. 22Rv1 cells do not exhibit the same kinetics of AR antagonism/depletion as do LNCaP cells with at least 20 µM or 24 hours being required for significant depletion of AR and PSA expression in these conditions.

Specific Aim 2c: Evaluate cell death associated with dose and time of treatment with VNPT-178 and VNLG-74A using the Trypan Blue exclusion assay.

While the androgen receptor’s role in promoting growth and differentiation of the prostate is well-supported, evidence underlying assumptions that its antagonism or removal directly activates cell death is much less abundant. The apparent presumption seems to be that, in an environment where both pro- and anti-apoptotic signals exist, the attenuation of one signal vs the other will drive the cell towards death or survival. The consequence of this model as it relates to prostate cancer therapy requires, then, that a sufficient pro-apoptotic signal be present upon the removal of AR’s anti-apoptotic signal.

In the clinic, men undergoing ADT or other AR-directed therapies inevitably develop resistance and progress to a lethal phenotype of metastatic castration-resistant prostate cancer (mCRPC). Therefore, we may surmise that while depletion of AR activity may attenuate anti-apoptotic signals in the cell, without concomitant pro-apoptotic therapies, the cancerous tissues persist in a dormant state until successful acquisition of new malignancy drivers.

It is our goal here to identify whether treatment of LNCaP cells could result in observable cell death or if our compounds merely promote cytostasis. We performed trypan blue assays to determine if reduction of cell numbers relative to control might be, at least partially, the result of activating a cellular death pathway. In Figure 15, we can see LNCaP cells grown in the presence of vehicle for 96 hours experience minimal turnover. Conversely, treating these cells with 10 or 20 µM galeterone, VNPT-178 or

VNLG-74A results in increased dye uptake concomitant with decreased overall cell

52 numbers suggesting there is, in these conditions, a sufficient signal for activation of cell death.

Specific Aim 2d: Evaluate changes in cell cycle progression following treatment with increasing doses of VNPT-178 and VNLG-74A.

Previous studies with galeterone in AR-negative cell lines attributed its ability to inhibit proliferation to its induction of an endoplasmic reticulum stress response and subsequent decrease in the translation of cyclin-D1 mRNA to protein. We have seen throughout our work that galeterone analogs can proficiently inhibit cell proliferation and this may be largely attributable to the accumulation of cells in the G1 phase of the cell cycle. While the study evaluated the cell cycle progression of serum starved cells trapped in G1 following co-administration of galeterone and fully-supplemented media for 24 hours, we instead believe a more useful model would be the capture of cell populations spontaneously trapped in a specific cycle phase upon treatment with our compounds.

LNCaP, 22Rv1, and PC-3 cells seeded in 10cm petri dishes were treated with anticancer agents for 96 hours before collection, fixation, propidium iodide (PI) staining, and analysis by flow cytometry (Figures 16A-16C). As all three cell lines have doubling times less than 48 hours, we anticipate an observable shift in cell cycle phase populations upon treatments with increasing concentrations of our compounds. The CDK4/6 inhibitor Palbociclib was used as the primary positive control for the accumulation of cells into G1 as these kinases are strictly required for progression to S phase. CGP-57380 seemed to have variable effect in each of the three cell lines, however increased concentrations may be necessary to identify more significance. Docetaxel promoted S phase accumulation and cell death (

54 of VNPT-178 seemed to promote G1 arrest in LNCaP and 22Rv1 cells and neither analog had an appreciable effect on PC-3 cell cycle progression. Of important note, however, is the increase in

µM of both galeterone analogs not seen here in PC-3 cells. The sub-G1 phase as determined here is often characterized as ‘apoptotic’ as DNA is actively being broken down in an otherwise fully-constituted cell. S phase are cells actively growing and replicating DNA in preparation for division and, although statistical significance could not be determined for G1 phases, significant decreases in S phase numbers could support prior claims of G1 arrest.

Summary and Conclusions from Specific Aim 2:

Our colony formation assay provided valuable information by identifying that, while increasing the duration of treatment can improve the relative responses of 22Rv1 cells to concentrations of ARDAs lower than the predicted GI50 values, 5 µM remains a realistic absolute minimum concentration to achieve significant reduction in tumor growth. Comparing the responses of LNCaP and 22Rv1 cells to increasing concentrations of ARDAs over 24 hours supports the potential for AR-Vs to regulate both AR and PSA expression during antagonism of the full-length AR. Additionally, though 5 µM treatments may be effective inhibitors of 22Rv1 cell proliferation over longer durations, within 24 hours the AR signaling axis remains largely intact.

Furthermore, caspase-3 cleavage does not strictly correlate negatively with AR or PSA expression indicating cell death observed by 24 hours is likely due to an extraneous mechanism (Figures 14A and 14B). By examining the cells directly using a light microscope we can see clearly that galeterone and its analogs VNPT-178 and VNLG-74A reduce LNCaP cell numbers after 96 hours compared to vehicle-treated controls.

Collecting all cells and performing a Trypan blue exclusion assay confirms that VNPT-

178 and VNLG-74A are capable of inducing cell death at 10 and 20 µM concentrations.

Performing a parallel assay using propidium iodide staining of fixed cell DNA and analysis by flow cytometry, we see that in 22Rv1 and LNCaP cells, 10 µM concentrations are sufficient to induce cell death but neither compound largely impairs cell cycle progression in PC-3 cells within these conditions.

As an aside, we frequently observed culture media acidification upon treatments with higher concentrations of ARDA as determined by the colorimetric pH scale of

Phenol red-containing RPMI-1640 (Figure 17A). These observations were most prominent within LNCaP cells treated with VNLG-74A (Figure 17B) and appear to be

59 dose-dependent. We found that the compounds themselves do not alter media pH (data not shown) and the magnitude of the pH change was largely dependent on the total number of viable cells upon treatment initiation but not a strictly a function of cell death induction as etoposide elicited no change in media pH. Concerned how these changes could impact our results we subsequently analyzed LNCaP and 22Rv1 cells for AR and

PSA expression after three different treatment conditions (Figure 18). Cells were treated with vehicle or 20 µM VNPT-178, VNLG-74A, etoposide, abiraterone, or enzalutamide as either a 24-hour treatment, a continuous 48-hour treatment, or two serial 24-hour treatments totaling 48 hours and whole cell lysates were examined via Western blotting.

Surprisingly, although all four AR-directed compounds maintain durable antagonism of the AR across all six conditions, AR expression fluctuates markedly in the LNCaP cell line. After a single, continuous 48-hour treatment with VNPT-178 and VNLG-74A, AR expression is fully recovered despite maintained depletion of PSA. Refreshing the media after 24 hours and examining the cells after 48 hours reveals our anticipated results whereby AR depletion is further enhanced compared to a single 24-hour treatment.

Expression of the AR is under control of numerous transcription factors that act during multiple cellular stresses including metabolic (PPARγ), cell proliferation (FOXO1), DNA damage (p53), or inflammation (NFκB) and, consequently, allowing continuous treatments may introduce too many confounding variables to clearly identify ARDA- specific mechanisms of action.

Taken together, our data from Specific Aim 2 have allowed us to identify a concentration of 20 µM VNPT-178 and VNLG-74A as reliably effective for experimental evaluation of ARDA activity in prostate cancer cells under short-term treatments and

60 daily media refreshment is likely obligatory for evaluation of time points longer than 24 hours. Additionally, further evidence has been obtained indicating the induction of cell death following ARDA administration may be independent of the AR signaling axis.

SPECIFIC AIM 3:

Our third objective is to identify a potential, specific mechanism of action by which galeterone analogs induce cell death in prostate cancer cells. The androgen receptor is largely considered a driver for cellular differentiation and survival. Using a model in which the relative strength of pro-survival and pro-death signals determines the fate of a cell, the simple attenuation of a pro-survival signal is insufficient to induce cell death if it remains the dominant signal. Based on our observations in AR-expressing and

-null prostate cancer cells, we hypothesize that VNPT-178 and VNLG-74A actively promote cell death induction via up-regulation of pro-death signals and this action is independent yet potentially overlapping any disruption of AR signaling.

To test our hypothesis, we will determine whether galeterone analogs induce apoptosis and if this event is related to depletion of the androgen receptor. We will examine again the role of the endoplasmic reticulum stress response and attempt to identify a novel target or class of targets for ARDA interaction. Together, our aim is to better delineate the consequences of ARDA exposure and reconcile how potentially diverse actions within prostate cancer cells can, together, be effectively employed as a next generation anticancer approach.

Specific Aim 3a: Determine whether galeterone analogs induce necrosis or apoptosis.

Our first objective is to determine whether the galeterone analogs VNPT-178 and

VNLG-74A promote apoptosis or necrosis. Apoptotic cell death is an important consideration during development of new cancer therapies as necrosis can promote a pro- tumorigenic inflammatory response. As we have seen with our Western blot (Figures

14A and 14B) and trypan blue exclusion (Figure 15) assays already, there is a good chance that both VNPT-178 and VNLG-74A activate either the intrinsic or extrinsic apoptotic pathways. Indeed, studies with galeterone and VNPT-55 indicate that mitochondrial dysfunction, cytochrome c release and intrinsic apoptosis may cause the cell death observed here but considering the potential for media acidification to contribute to these consequences, we sought to evaluate LNCaP, 22Rv1, and PC-3 cells at multiple time points over 72 hours to identify the kinetics of cell death induction.

Depicted in Figure 19, digital micrographs were collected following treatments of 20 µM

VNPT-178 (top panels) or VNLG-74A (bottom panels) and assembled for easy visual comparison to initial cell appearance or cells treated with vehicle for 72 hours (right land left large panels, respectively). In LNCaP cells, morphological signs of apoptosis including membrane blebbing, cell shrinkage, and cell detachment are present as early as

24 hours. 22Rv1 cells appear slightly more sensitive to treatments exhibiting morphological signs of apoptosis as early as 18 hours while PC-3 cells appear markedly less sensitive to cell death with barely detectable membrane blebbing seen through 72 hours. This is not unlike the incidental cell death observed during the flow cytometric analysis of PI-stained cells after 96 hours. As we have identified a peculiar relationship

63 between AR expression/activity and caspase-3 cleavage in our earlier analysis of 22Rv1 cells, we performed western blots using whole cell lysates for each time point. AR protein expression declines steadily over time in 22Rv1 cells treated with VNPT-178 correlating with loss of PSA and increases in the cleavage of both caspase-3 and PARP

(Figure 20).

Specific Aim 3b: Determine the AR-dependency of cell death induction in prostate cancer cells.

While we can routinely see a correlation between a decline in AR activity, AR expression, and induction of cell death, we are not certain as to the AR-dependency of this phenomenon. Galeterone was promoted as being better than castration or AR antagonism alone and in combination as it could also deplete the expression of the androgen receptor. However, this argument is difficult to test as AR depletion has never been absolute and thus a minimum threshold of AR expression necessary to confer its survival benefits is unknown. Therefore, to evaluate whether attenuated AR expression

65 itself could promote the effects observed following treatment of 22Rv1 cells with VNPT-

178, we transiently transfected these cells with shRNA capable of selectively targeting the full-length AR and AR-V7. After 72 hours of specific or combined knockdown of

AR mRNA, we compared the effects of vehicle treatment to that of 20 µM VNPT-178 treatment for 24 hours via western blotting (Figure 21).

The plots represent the fold-change in mean full-length AR and AR-V7 expression normalized to β-tubulin controls for each condition compared to vehicle- treated cells not treated with either empty virus or virus loaded with the shRNA against the AR. From these data, we can determine that the shRNA constructs are specific for each indicated AR isoform and that their combination does not additively or synergistically enhance depletion in comparison to each agent alone. Additionally, the loss of either the full-length or truncated AR individually or in combination failed to appreciably deplete PSA expression. Furthermore, shRNA against the AR resulting in approximately 30% reduction in fAR and 50% reduction in AR-V7 did not result in the cleavage of caspase-3 or PARP. Altogether these results are not surprising given our estimation that the androgen receptor is abundantly expressed in these cell lines and, in general, prostatic tumors and incomplete reduction in its expression is unlikely a significant contributor to the overall effect observed in these models. Contrarily, a 24-

67 hour treatment of VNPT-178 was sufficient to antagonize the AR and induce robust caspase-3 and PARP cleavage across all conditions of vehicle or shRNA treatments.

Comparing VNPT-178 to vehicle in the presence or absence of empty virus we see an approximate 20% decline in AR protein with a substantial reduction in PSA.

Combining VNPT-178 with shRNA against fAR or AR-V7 further reduces the loss of

AR proteins from 30% and 50%, respectively, to 65% and 70%. However, this observation is not paralleled by further decreases in PSA expression. This does not exclude the argument that VNPT-178 induces apoptosis due to a more pronounced reduction in AR expression concomitant with antagonism of the AR compared to shRNA treatments alone, but it does indicate a greater depletion of AR protein is necessary.

Returning to Dr. Bruno’s work with galeterone in PC-3 cells, it was identified that induction of the endoplasmic reticulum stress response (ERSR) effected the inhibition of cell proliferation. When we look outside the AR axis here, we can see that VNPT-178 very clearly induces the expression of CHOP and promotes the lipidation of LC-3 which are not observed in response to shRNA or vehicle treatments alone. Dr. Bruno concluded that galeterone did not induce cell death but rather promoted quiescence.

Since we also observe induction of the ERSR, we were curious if extending our treatments of PC-3 cells to 96 hours could elicit similar responses. Indeed, PC-3 cells treated with greater concentrations of VNLG-74A for 96 hours appear to undergo apoptotic cell death concomitant with CHOP expression and LC-3 lipidation (Figure 22).

Additionally, re-evaluating LNCaP cells treated with 20 µM concentrations of ARDAs,

Etoposide, abiraterone, or enzalutamide for 48 hours (with a media refresh at 24 hours) reveals that what was potent antagonism of the AR by abiraterone and enzalutamide

(Figure 18) as marked by decreased AR and PSA expression was not coincidental with

PARP cleavage unlike treatments with VNPT-178 and VNLG-74A (Figure 23).

Therefore, we conclude that any observed activation of apoptosis within 72 hours not attributable to potent disruption of AR expression and/or activity.

Specific Aim 3c: Determine the involvement of the ERSR/UPR for induction of cell death in prostate cancer cells.

Our objective here is to evaluate the activation of the endoplasmic reticulum stress/unfolded protein responses and its sufficiency to induce apoptosis following administration of galeterone analogs. As stated, previous studies examining the ERSR following galeterone treatments failed to identify a specific mechanism of action or the

69 culmination of apoptosis. Herein we aim to better understand the kinetics of the

ERSR/UPR to determine a potential target for VNPT-178 and VNLG-74A. After treating 22Rv1 cells with 20 µM ARDA for up to 72 hours we evaluated changes in the expression of initiator and effector proteins of the UPR via western blot (Figure 24 upper panels).

Within three hours, phosphorylation of PERK and eIF2α is observed concomitant with upregulation of ATF-4. CHOP is abundantly expressed by six hours and XBP1 splicing steadily increases expression of this potent transcription factor through 72 hours.

These data clearly indicate the rapid induction of the UPR ahead of appreciable down- regulation or antagonism of the androgen receptor. Additionally, the UPR is wholly engaged before detectable caspase-3 and PARP cleavage and not just present during these events. Interestingly, a steady up-regulation of the UPR sentinels ATF-6 and IRE1α is observed in response to VNPT-178. As the dissociation of BiP/Grp78 from these proteins (and PERK) is argued to cause the UPR in the first place, an increase in these

70 proteins would likely expand the pool of unbound UPR activators. We therefore examined the abundance of ER chaperones expecting to see a concomitant increase in these proteins to manage the accumulation of misfolded proteins within the ER lumen.

To our surprise, the chaperone proteins calnexin, BiP/Grp78, and PDI remain stably abundant during VNPT-178 treatments (Figure 24 lower panels). While this seems counter-intuitive, other studies have proposed IRE1α plays a critical role in the ER quality control and autophagic clearance of unfolded or aggregated proteins (86). The concomitant increase in LC-3 lipidation without change in Beclin-1 expression may support these claims and suggest the severity of UPR induction following ARDA treatments.

We next sought to identify whether the UPR was required for the activation of

PARP cleavage in these cells. As we still do not know how the ARDAs initiate the UPR, we decided to work backwards and first determine whether cleavage of PARP is in fact a consequence of the caspase-3 cleavage we have repeatedly observed. We treated 22Rv1 cells with vehicle, VNPT-178, or VNLG-74A in the presence or absence of increasing concentrations of the pan caspase inhibitor Z-VAD-FMK for a total of 26 hours and performed western blots using the whole cell lysates (Figure 25). This inhibitor is a consensus peptide sequence mimetic that specifically inhibits the catalytic domains of both initiator and effector caspases. Neither vehicle nor increasing doses of Z-VAD-

FMK caused appreciable cleavage of caspase-3 or PARP. Treatments with VNPT-178 and VNLG-74A alone both caused marked caspase-3 and PARP cleavage and increasing the presence of Z-VAD-FMK resulted in the dose-dependent decrease of PARP cleavage.

However, with respect to caspase-3, even concentrations of 100 µM Z-VAD-FMK were insufficient to completely block the appearance of the 19/20 kDa fragment.

As caspase-3 is natively expressed as a 32 kDa zymogen that is cleaved by initiator caspases to complementary 19/20 and 12 kDa fragments and subsequently undergoes rapid autoproteolytic processing of the 19/20 kDa fragment resulting in a 17 kDa fragment, we expected to see a complete attenuation of caspase-3 cleavage products during Z-VAD-FMK inhibition (87). However, lingering abundance of a 19/20 kDa fragment may indicate that a non-caspase protease is responsible, at least in part, for the activation of caspase-3. Since it was identified that calcium efflux from the ER to the

73 cytosol was a primary consequence of galeterone treatments in PC-3 cells and calpains have known roles in cell death induction (88), we replicated these treatment conditions using the calpain 1 and 2 inhibitor calpeptin. Calpains are calcium-dependent, neutral pH cysteine proteases with very broad target specificity thought to be determined primarily by tertiary conformation compared to the strict sequence-specific cleavage by caspases.

Treatments of 22Rv1 cells with 100 µM calpeptin and vehicle exhibited a slight induction of PARP cleavage without observable caspase-3 contribution (Figure 26). However, calpeptin dose-dependently decreased PARP cleavage following the administration of

VNPT-178 and VNLG-74A for 24 hours. Calpeptin also reduced, but did not completely abolish caspase-3 cleavage compared to ARDA treatments alone. Relief of caspase inhibition by Z-VAD-FMK resulted in normal caspase cleavage product abundance with prominent bands appearing at 17 and 12 kDa exclusively.

Specific Aim 3d: Identify novel targets of galeterone analogs VNPT-178 and

VNLG-74A.

Our primary objective here is to evaluate whether the functional assays performed up to now are sufficient to identify novel targets for galeterone analogs VNPT-178 and

VNLG-74A. With these data, we conclude that VNPT-178 and VNLG-74A rapidly activate the UPR like galeterone treatments and do so independently of their action against the androgen receptor signaling axis. The activation of the UPR is either in response or concomitant to deregulated calcium signaling resulting in the activation of calpains and subsequent induction of apoptosis within 24 hours. As both VNPT-178 and

VNLG-74A are based on a steroidal scaffold, their hydrophobic properties enable them to

74 easily traverse cellular membranes. Dr. Bruno’s work left with the conclusion that galeterone, due to its hydrophobicity, could interact with any number of membrane- localized proteins. However, the specific and intense activation of the UPR suggests instead there may be a single, alternate protein target or a domain/motif shared by multiple proteins that have essential roles for maintaining protein homeostasis. As the androgen receptor contains a large, hydrophobic cleft for ligand binding, we used this and the requirement for UPR activation to focus our attention on the Hsp70 chaperone

BiP/Grp78.

BiP is the key ER chaperone protein ensuring proper protein folding within the

ER lumen and specifically binds the luminal domains of the UPR sentinels PERK,

IRE1α, and ATF-6 (65). Bound by BiP, these UPR sentinels are maintained in their inactive, monomeric conformations. Additionally, BiP binds the lumenal face of the

Sec61 transclocon complex acting as a gatekeeper against the efflux of calcium from the

ER to the cytosol (89). Since Hsp70 family proteins contain a hydrophobic nucleotide binding domain (i.e. ATPase domain) that allosterically regulate the affinity of their substrate binding domains (66), it seems probable that VNPT-178 and VNLG-74A could interact with BiP.

Using an in silico molecular docking approach, the crystallographically resolved structures of BiP’s ATPase domain (90) were evaluated for potential interaction with

ARDAs. We have determined previously that VNPT-178 can bind the AR-LBD using surface plasmon resonance but did not pursue determining Kd values as its interaction was notably weaker than that of galeterone (data not shown). Therefore, to provide a baseline for predicted interactions, we first determined the predicted binding energy of

VNPT-178 and VNLG-74A with the AR-LBD (Figures 27A and 27B). Both compounds exhibit a potential for spontaneous binding with the AR-LBD and the free energy of this association (ΔG°) is predicted to be -7.2 and -6.6 kcal/mol for VNPT-178 and VNLG-

74A, respectively. Docking each compound into the ATPase domain of BiP revealed a greater potential for interaction with free energies of binding predicted to be -8.0 and -8.3 kcal/mol for VNPT-178 and VNLG-74A, respectively (Figures 27C and 27D). Inspired by these results, we also evaluated whether the ARDAs could interact with another member of the Hsp70 chaperone family, Hsp70-1A. Using the same methods, we estimate Gibb’s free energy of binding for VNPT-178 and VNLG-74A to be -7.8 and -7.1 kcal/mol, respectively (Figures 27E and 27F). Taken together, these molecular docking

76 studies predict that both galeterone analogs VNPT-178 and VNLG-74A have a greater potential to bind the ATPase domains of Hsp70 chaperones BiP/Grp78 and Hsp70-1A than they do the ligand-binding domain of the androgen receptor.

Summary and Conclusions from Specific Aim 3:

Direct observation of cells in response to treatment is substantially more telling than relying on biochemical reporters and our side-by-side micrograph comparison of

ARDA treatments in LNCaP, 22Rv1 and PC-3 cells clearly indicates that 22Rv1 cells are very sensitive to 20 µM treatments of VNPT-178 and VNLG-74A. Not only do they exhibit morphological signs of apoptosis earlier than LNCaP and PC-3 cells but a greater proportion of the cells appear to be actively dying. Performing western blots of 22Rv1 cell lysates following treatment with VNPT-178 at multiple time points supports our visual observations of cell death induction. However, based on these blots alone one could still make the argument that cell death induction is a graded consequence of AR antagonism and depletion. We therefore evaluated whether AR depletion via shRNA knockdown or AR antagonism alone were sufficient to induce PARP cleavage. Selective or combined AR knockdown as well as direct AR antagonism are insufficient to induce cell death. ARDA treatments are, contrarily, necessary and sufficient to promote PARP cleavage in LNCaP, 22Rv1, and PC-3 cells. Additionally, despite the apparent negative correlation of AR expression and activity to cell death induction, induction of the UPR is rapid and sustained through 72 hours of treatment. This is an important observation as it more realistically accounts for time when considering the stability of mRNA and proteins transcriptionally regulated by the AR—just because the AR is lost doesn’t mean cells are instantly and linearly sensitive to apoptosis.

The molecular docking studies proved paramount for the reconciliation of our observations here and elsewhere in the literature in which compounds targeting the AR and promoting its depletion simultaneously promote UPR and cell death. As mentioned,

78 we believe AR antagonism/depletion is insufficient to induce the apoptotic effects observed, especially within the time frame of 24-72 hours. By identifying an enhanced potential for binding BiP/Grp78 and Hsp70-1A we have given plausibility to an argument that the AR is not the primary effector of ARDA action.

CHOP is a putative pro-apoptotic transcription factor (72, 73) and its expression is an early response to ARDAs and is maintained at high levels at all time points where apoptosis is observed. Additionally, based on the inhibitor studies using Z-VAD-FMK and calpeptin, our data expand on observations by Dr. Bruno that calcium efflux from the

ER to the cytosol may sensitize prostate cancer cells to apoptosis. While the roles of calcium and calpains are not yet fully elucidated, we conclude that VNPT-178 and

VNLG-74A actively promote apoptosis by inhibiting BiP and their overlapping ability to antagonize the AR axis my tune the sensitivity of prostate cancer cells to cell death induction.

SPECIFIC AIM 4:

Herein we will evaluate the safety and efficacy of VNPT-178 and VNLG-74A in vivo using a 22Rv1 tumor xenograft model. Based on previous studies using Galeterone and VNPT-55, we anticipate oral gavage administration of these analogs twice daily will both be relatively safe and inhibit the outgrowth of 22Rv1 tumors. We will monitor tumor growth and, upon termination of the study, will evaluate the efficacy of galeterone and its analogs compared to vehicle and enzalutamide based on change in tumor mass.

Mouse body weights and behavior will be the primary determinant for safety and tolerability of the vehicle and each compound.

Specific Aim 4a: Determine the growth-inhibitory effects of galeterone analogs against the AR-V7-expressing 22Rv1 xenograft model.

Six-week old male athymic nude mice injected subcutaneously with 22Rv1 cells suspended in Matrigel were randomly stratified into six groups of six upon development of tumors measuring an estimated 100 cubic millimeters. Treatment groups included β- cyclodextrin vehicle control, enzalutamide (25mg/kg), and galeterone, VNPT-178, and

VNLG-74A (148mmol/kg twice daily). Vehicle or compound was delivered by oral gavage every day until tumors grew to an estimated 1750 cubic millimeters. Mice were evaluated daily for overt signs of toxicity and both body weights and tumor dimensions were measured twice weekly. Without intervention, the 22Rv1 xenografts grew rapidly and exceeded the primary endpoint within two and a half weeks (Figure 28). Caliper measurements and estimated tumor volumes are plotted as a function of the mean volume

± the standard error of the mean. Estimated tumor volumes indicated both galeterone and

VNPT-178 may reduce tumor growth up to 75% while enzalutamide and VNLG-74A were less effective, reducing tumor growth by an estimated 50%. However, the mass measurements of excised tumors indicate only that galeterone treatments differed with statistical significance versus vehicle-treated controls (Figure 29). No statistically significant differences were observed upon comparing tumor volume or mass of galeterone or its analog-treated groups to the enzalutamide-treated positive controls.

Specific Aim 4b: Evaluate the safety and tolerability of galeterone analogs administered via oral gavage in male athymic nude mice.

Throughout the xenograft study all mice in each group exhibited normal behavior, were relatively equal in appearance and activity, and fought with fervor against oral gavage needle. Measurement of body weights throughout the xenograft study resulted positive weight gain within all treatment groups, and this weight gain was more than the measured tumor masses indicating a low level of toxicity for all treatment conditions

(Figure 30). One mouse in the VNPT-178 group died during the study and we expect this

83 was the result of it having bitten and damaged the gavage needle upon administration causing trauma to the esophagus. The plastic needles were damaged and frequently replaced during the study.

Summary and Conclusions from Specific Aim 4:

These data indicate that although not overtly toxic and potentially safe at doses more than 148mmol/kg, additional inquiry is necessary to evaluate the in vivo bioavailability and activity of these novel galeterone analogs. As the dosing schedules used here were determined from previous studies comparing galeterone and VNPT-55, we will need to do a thorough examination of dosing quantity and route of administration for continued study of these analogs. Additionally, the rapid growth of the 22Rv1 xenografts may have contributed greatly to the variability in tumor development within all mouse groups. Injections of fewer cells, 5x105 instead of 2x106 for instance, may improve the consistency of tumor development, reducing deviations between mice, and enabling collection of more-statistically significant data.

SUMMARY AND FUTURE DIRECTIONS:

The androgen receptor (AR) remains a primary target for the treatment of all stages of prostate cancer development and progression. Despite many advances in the development of new and more potent agents targeting the AR signaling axis, mechanisms of therapy resistance continually emerge (91). This observation alone is significant when viewed through the lens of Darwinian evolution as it strictly implies these aberrations and subsequent reactivation of the AR axis are advantageous to and sufficient for tumor proliferation. Aptly, it has been proposed by many that depletion of the AR would be a superior treatment paradigm compared to attenuation of its activity through androgen deprivation or direct antagonism, alone or in combination (92). Indeed, multiple compounds have been described as depleting AR availability through upregulation of post-translational degradation and these studies also often depict increases in cell death induction, namely apoptosis. However, there is much we do not know about the precise mechanisms by which the AR is selectively targeted by the ubiquitin-proteasome pathway (93, 94) and how its loss can directly promote apoptosis versus quiescence.

Our objectives herein were straightforward: identify novel galeterone analogs with the potential for improved antiproliferative and anti-AR activity; determine a modality in which those analogs could be reliably evaluated in vitro; identify a mechanism of apoptosis induction and its dependency on AR status; and finally, evaluate their activity in vivo using a xenograft model of castration-resistant prostate cancer. In general, our hypotheses in which galeterone analogs exhibiting improved antiproliferative activity would also be more potent AR antagonists and depletive agents, and that these compounds would be effective at concentrations at or less than 20 µM and within 96

86 hours are supported by our data. Further, a critical finding that both VNPT-178 and

VNLG-74A exhibit a greater potential for binding the ATPase domains of Hsp70 chaperones BiP/Grp78 and Hsp70-1A than the AR-LBD provides plausible explanation for how ARDAs induce apoptosis extraneously to effects on the androgen signaling axis and a summary model is depicted in Figure 31. However, our tumor xenograft experiment failed to support our hypothesis of greater in vivo efficacy of ARDAs VNPT-

178 and VNLG-74A compared to galeterone, enzalutamide, or vehicle-treated controls despite supporting their relative tolerability.

The experiments performed within Specific Aim 3 significantly advance our understanding of how galeterone analogs can affect malignant cell proliferation.

Although prostate cancer therapies may continue to benefit from restricting the activation and expression of the AR over many months, we can almost universally reject any notion that these analogs inhibit malignant cell proliferation within 96 hours via targeting the fAR or AR-V7. Firstly, only VNPT-178 and VNLG-74A, and not abiraterone or enzalutamide, induced PARP cleavage in LNCaP cells (Figure 23). Therefore, antagonism of the AR is not sufficient to induce apoptosis within 48 hours. Secondly, only VNPT-178, and not shRNA knockdown, induced apoptosis in 22Rv1 cells (Figure

21). Therefore, incomplete loss of full-length and/or truncated AR proteins is insufficient to induce apoptosis within 72hrs. Thirdly, VNLG-74A induced caspase-3 and PARP cleavage in PC-3 cells (Figure 22). Therefore, galeterone analogs do not require the expression, activation, or losses thereof of the androgen receptor to induce apoptosis within 96 hours. Additionally, the differences in AR and PSA expression kinetics between LNCaP and 22Rv1 cells (Figures 14A and 14B) as well as the AR-V7 shRNA

88 knockdown in 22Rv1 cells reveal that neither galeterone analog sufficiently antagonizes

AR-V7 activity. However, we do concede that these experiments were designed around a global ARDA-mediated induction of apoptosis and do not strictly exclude the potential for a partial or specific inhibition of AR-V7 function unrelated to this endpoint.

Additionally, as these compounds simultaneously target the AR and Hsp70 proteins, we cannot be certain that the overlapping of effects is not significant at longer time points.

Future studies expanding on our identification of novel ARDA targets will greatly improve our understanding of the context-specific activities and limitations of galeterone analogs and their use as anticancer agents. Based on our rationalization of functional assays, we identified two heat shock proteins containing nucleotide-binding domains with which VNPT-178 and VNLG-74A can directly interact. However, adenosine phosphates are the primary energy currency within cells and, consequently, numerous proteins contain ATP/ADP-binding domains and may thus be inhibited or promoted by galeterone analogs. Hsp70-1A and BiP/Grp78 share only 67% sequence homology within the NBD; however, their 3D structures as determined by crystallography are highly similar (90).

Therefore, it seems likely that we have only partially described the totality of potential

ARDA-interaction partners and there may be examples with even greater affinity than

BiP. Likewise, evaluation of galeterone analogs in comparison to epigallocatechin gallate (EGCG), which also binds BiP, Hsp70, and the AR, may provide additional clarity versus enzalutamide comparisons (48, 95, 96).

A functional characterization of ARDA-bound BiP would help us improve our model of UPR and apoptosis induction. We presume, based on our studies, that VNPT-

178 and VNLG-74A likely emulate an ATP-bound BiP conformation which favors

89 transient interaction with its substrates. As the exact mechanisms for BiP dissociation from PERK, IRE1α, and ATF-6 are still being investigated (68-71), it would be valuable to know whether galeterone analogs compete with adenosine phosphates or modulate the kinetics of ATP hydrolysis or ADP-ATP exchange. Additionally, the effects on Hsp70 and, subsequently the AR and AR-V7, have yet to be fully resolved. Using current models of Hsp70 action for the recruitment of Hsp90 to its client proteins (CPs), which include nuclear hormone receptors, it is plausible that accumulation of Hsp70 in its ATP- bound state and inhibiting subsequent hydrolysis to ADP could elicit the proteasome- mediated degradation of Hsp90 clients (i.e. fAR). However, studies on Hsp70-mediated clearance of polyQ AR in neurodegenerative disorders demonstrate that efficient clearance of these androgen receptors requires ADP-bound Hsp70 to recruit the E3 ligase

CHIP (97, 98). Additionally, AR-Vs lacking the AR-CTD do not require Hsp90 chaperone activity for stabilization and activity (99-101) but can still be targeted by CHIP

(102). Therefore, it remains to be determined whether Hsp70 is necessary to recruit

CHIP to AR-Vs and, if so, whether Hsp70 can efficiently promote this interaction locked in an ATP-bound conformation. As the kinetics of AR-V protein depletion are slow, taking nearly 72 hours, this may not be an efficient process and thus independent of a stable Hsp70-AR interaction.

Additionally, the involvement of calcium and calpains for the induction of cell death is not fully resolved. Outside of cell-specific differences in calpain expression or localization, additional factors, namely p53, have been implicated in calcium-mediated apoptosis (80, 103, 104). While performing parallel studies comparing action of galeterone analogs to R1881 in LNCaP, 22Rv1 and PC-3 cells we noticed that p21

90 induction was markedly upregulated in LNCaP and 22Rv1 but not PC-3 cells following treatments with 20 µM VNPT-178, VNLG-74A, and R1881 (Figure 32). P21 is a potent cyclin-dependent kinase inhibitor (CKI) capable of universally inhibiting CDK complexes and stalling cell cycle progression in all phases (105, 106). While we did not fully pursue this avenue of investigation as the role of p53 in cancer development is broad, it may indeed be a critical factor in regulating cell responses to ARDA treatment.

An obvious, related observation is the relative insensitivity of PC-3 cells to both VNPT-

178- and VNLG-74A-mediated cell death despite rapid induction of apoptosis in LNCaP and 22Rv1 cells. P53 activation could sensitize cells to early death induction while sustained ERSR and CHOP upregulation inevitably promote apoptotic death over longer treatment durations.

Altogether, continued evaluation of VNPT-178 and VNLG-74A, in addition to revised re-screening of galeterone analogs, could enable the identification of highly effective anticancer agents for all stages of prostate cancer development and progression.

While 22Rv1 cells have been extensively used as a model of AR-V7-mediated castration- resistant prostate cancer, it is apparent from these studies that galeterone analogs VNPT-

178 and VNLG-74A are potent inhibitors of 22Rv1 cell proliferation through AR- and

AR-V-independent mechanisms. Therefore, evaluation in additional AR-V-expressing models of CRPC, including LuCaP and VCaP, may aid our determination of the clinical utility of galeterone analogs in a broader context of AR-V expression and tumor resistance to currently available AR-directed therapies.

REFERENCES:

1. Howlader N NA, Krapcho M, Miller D, Bishop K, Kosary CL, Yu M, Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA (2017) SEER Cancer Statistics Review, 1975-2014, National Cancer Institute. 2. De Marzo AM, et al. (2007) Inflammation in prostate carcinogenesis. Nature reviews. Cancer 7 (4):256-269. 3. Lee JJ, et al. (2015) Biologic differences between peripheral and transition zone prostate cancer. The Prostate 75 (2):183-190. 4. Board PDQATE (2002) Prostate Cancer Treatment (PDQ (R)): Patient Version. PDQ Cancer Information Summaries, (National Cancer Institute (US), Bethesda (MD)). 5. Saad F & Fizazi K (2015) Androgen Deprivation Therapy and Secondary Hormone Therapy in the Management of Hormone-sensitive and Castration-resistant Prostate Cancer. Urology 86 (5):852-861. 6. Crawford ED & Hou AH (2009) The role of LHRH antagonists in the treatment of prostate cancer. Oncology (Williston Park, N.Y.) 23 (7):626-630. 7. Andriole GL, et al. (2010) Effect of dutasteride on the risk of prostate cancer. The New England journal of medicine 362 (13):1192-1202. 8. Tan ME, Li J, Xu HE, Melcher K, & Yong EL (2014) Androgen receptor: structure, role in prostate cancer and drug discovery. Acta pharmacologica Sinica. 9. Lavery DN & McEwan IJ (2008) Structural characterization of the native NH2-terminal transactivation domain of the human androgen receptor: a collapsed disordered conformation underlies structural plasticity and protein-induced folding. Biochemistry 47 (11):3360-3369. 10. Lavery DN & McEwan IJ (2008) Functional characterization of the native NH2-terminal transactivation domain of the human androgen receptor: binding kinetics for interactions with TFIIF and SRC-1a. Biochemistry 47 (11):3352-3359. 11. Cutress ML, Whitaker HC, Mills IG, Stewart M, & Neal DE (2008) Structural basis for the nuclear import of the human androgen receptor. Journal of cell science 121 (Pt 7):957- 968. 12. Heemers HV & Tindall DJ (2007) Androgen receptor (AR) coregulators: a diversity of functions converging on and regulating the AR transcriptional complex. Endocrine reviews 28 (7):778-808. 13. Bolton EC, et al. (2007) Cell- and gene-specific regulation of primary target genes by the androgen receptor. Genes & development 21 (16):2005-2017. 14. Myung JK, et al. (2013) An androgen receptor N-terminal domain antagonist for treating prostate cancer. The Journal of clinical investigation 123 (7):2948-2960.

15. Andersen RJ, et al. (2010) Regression of castrate-recurrent prostate cancer by a small- molecule inhibitor of the amino-terminus domain of the androgen receptor. Cancer cell 17 (6):535-546. 16. Kim SH, et al. (2010) Androgen receptor interacts with telomeric proteins in prostate cancer cells. The Journal of biological chemistry, United States), Vol 285, pp 10472- 10476. 17. Schiewer MJ, et al. (2012) Dual roles of PARP-1 promote cancer growth and progression. Cancer discovery, (2012 Aacr., United States), Vol 2, pp 1134-1149. 18. Litvinov IV, et al. (2006) Androgen receptor as a licensing factor for DNA replication in androgen-sensitive prostate cancer cells. Proceedings of the National Academy of Sciences of the United States of America, United States), Vol 103, pp 15085-15090. 19. Zegarra-Moro OL, Schmidt LJ, Huang H, & Tindall DJ (2002) Disruption of androgen receptor function inhibits proliferation of androgen-refractory prostate cancer cells. Cancer research 62 (4):1008-1013. 20. Dalal K, et al. (2014) Selectively targeting the DNA-binding domain of the androgen receptor as a prospective therapy for prostate cancer. The Journal of biological chemistry 289 (38):26417-26429. 21. Pratt WB & Toft DO (2003) Regulation of signaling protein function and trafficking by the hsp90/hsp70-based chaperone machinery. Experimental biology and medicine (Maywood, N.J.) 228 (2):111-133. 22. Jasuja R, et al. (2009) Kinetic and thermodynamic characterization of dihydrotestosterone-induced conformational perturbations in androgen receptor ligand-binding domain. Molecular endocrinology (Baltimore, Md.) 23 (8):1231-1241. 23. Thompson J, Saatcioglu F, Janne OA, & Palvimo JJ (2001) Disrupted amino- and carboxyl-terminal interactions of the androgen receptor are linked to androgen insensitivity. Molecular endocrinology (Baltimore, Md.) 15 (6):923-935. 24. Elhaji YA, et al. (2006) Impaired helix 12 dynamics due to proline 892 substitutions in the androgen receptor are associated with complete androgen insensitivity. Human molecular genetics 15 (6):921-931. 25. Darshan MS, et al. (2011) Taxane-induced blockade to nuclear accumulation of the androgen receptor predicts clinical responses in metastatic prostate cancer. Cancer research 71 (18):6019-6029. 26. Thadani-Mulero M, Nanus DM, & Giannakakou P (2012) Androgen receptor on the move: boarding the microtubule expressway to the nucleus. Cancer research 72 (18):4611-4615. 27. Meehan KL & Sadar MD (2003) Androgens and androgen receptor in prostate and ovarian malignancies. Frontiers in bioscience : a journal and virtual library 8:d780-800. 28. Wilson S, Qi J, & Filipp FV (2016) Refinement of the androgen response element based on ChIP-Seq in androgen-insensitive and androgen-responsive prostate cancer cell lines. Scientific reports 6:32611.

29. Massie CE, et al. (2007) New androgen receptor genomic targets show an interaction with the ETS1 transcription factor. EMBO reports 8 (9):871-878. 30. Denayer S, Helsen C, Thorrez L, Haelens A, & Claessens F (2010) The rules of DNA recognition by the androgen receptor. Molecular endocrinology (Baltimore, Md.) 24 (5):898-913. 31. Eisermann K, Wang D, Jing Y, Pascal LE, & Wang Z (2013) Androgen receptor gene mutation, rearrangement, polymorphism. Translational andrology and urology 2 (3):137-147. 32. Chen CD, et al. (2004) Molecular determinants of resistance to antiandrogen therapy. Nat Med 10 (1):33-39. 33. Bohl CE, Gao W, Miller DD, Bell CE, & Dalton JT (2005) Structural basis for antagonism and resistance of bicalutamide in prostate cancer. Proceedings of the National Academy of Sciences of the United States of America, United States), Vol 102, pp 6201-6206. 34. van der Steen T, Tindall DJ, & Huang H (2013) Posttranslational modification of the androgen receptor in prostate cancer. International journal of molecular sciences 14 (7):14833-14859. 35. Guo Z, et al. (2006) Regulation of androgen receptor activity by tyrosine phosphorylation. Cancer cell, United States), Vol 10, pp 309-319. 36. Dehm SM, Regan KM, Schmidt LJ, & Tindall DJ (2007) Selective role of an NH2-terminal WxxLF motif for aberrant androgen receptor activation in androgen depletion independent prostate cancer cells. Cancer research 67 (20):10067-10077. 37. Nakata D, et al. (2017) The RNA helicase DDX39B and its paralog DDX39A regulate androgen receptor splice variant AR-V7 generation. Biochemical and biophysical research communications 483 (1):271-276. 38. Liu LL, et al. (2014) Mechanisms of the androgen receptor splicing in prostate cancer cells. Oncogene 33 (24):3140-3150. 39. Cao S, Zhan Y, & Dong Y (2016) Emerging data on androgen receptor splice variants in prostate cancer. Endocrine-related cancer 23 (12):T199-t210. 40. Rundlett SE, Wu XP, & Miesfeld RL (1990) Functional characterizations of the androgen receptor confirm that the molecular basis of androgen action is transcriptional regulation. Molecular endocrinology (Baltimore, Md.) 4 (5):708-714. 41. Chan SC, Li Y, & Dehm SM (2012) Androgen receptor splice variants activate androgen receptor target genes and support aberrant prostate cancer cell growth independent of canonical androgen receptor nuclear localization signal. The Journal of biological chemistry 287 (23):19736-19749. 42. Hu R, et al. (2009) Ligand-independent androgen receptor variants derived from splicing of cryptic exons signify hormone-refractory prostate cancer. Cancer research 69 (1):16- 22.

43. Guo Z, et al. (2009) A novel androgen receptor splice variant is up-regulated during prostate cancer progression and promotes androgen depletion-resistant growth. Cancer research 69 (6):2305-2313. 44. Sun S, et al. (2010) Castration resistance in human prostate cancer is conferred by a frequently occurring androgen receptor splice variant. The Journal of clinical investigation 120 (8):2715-2730. 45. Antonarakis ES, et al. (2014) AR-V7 and resistance to enzalutamide and abiraterone in prostate cancer. The New England journal of medicine 371 (11):1028-1038. 46. Yu Z, et al. (2014) Rapid induction of androgen receptor splice variants by androgen deprivation in prostate cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 20 (6):1590-1600. 47. Golovine KV, et al. (2013) Piperlongumine induces rapid depletion of the androgen receptor in human prostate cancer cells. The Prostate 73 (1):23-30. 48. Ren F, Zhang S, Mitchell SH, Butler R, & Young CY (2000) Tea polyphenols down- regulate the expression of the androgen receptor in LNCaP prostate cancer cells. Oncogene 19 (15):1924-1932. 49. Petiwala SM, et al. (2014) Rosemary (Rosmarinus officinalis) extract modulates CHOP/GADD153 to promote androgen receptor degradation and decreases xenograft tumor growth. PloS one 9 (3):e89772. 50. Petiwala SM, et al. (2016) Carnosic acid promotes degradation of the androgen receptor and is regulated by the unfolded protein response pathway in vitro and in vivo. Carcinogenesis 37 (8):827-838. 51. Zhang Y, et al. (2012) Tanshinones from Chinese medicinal herb Danshen (Salvia miltiorrhiza Bunge) suppress prostate cancer growth and androgen receptor signaling. Pharmaceutical research 29 (6):1595-1608. 52. Yamashita S, et al. (2012) ASC-J9 suppresses castration-resistant prostate cancer growth through degradation of full-length and splice variant androgen receptors. Neoplasia (New York, N.Y.) 14 (1):74-83. 53. Mashima T, Okabe S, & Seimiya H (2010) Pharmacological targeting of constitutively active truncated androgen receptor by nigericin and suppression of hormone-refractory prostate cancer cell growth. Molecular pharmacology 78 (5):846-854. 54. Liu C, et al. (2014) Niclosamide inhibits androgen receptor variants expression and overcomes enzalutamide resistance in castration-resistant prostate cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 20 (12):3198-3210. 55. Handratta VD, et al. (2005) Novel C-17-heteroaryl steroidal CYP17 inhibitors/antiandrogens: synthesis, in vitro biological activity, pharmacokinetics, and antitumor activity in the LAPC4 human prostate cancer xenograft model. Journal of medicinal chemistry 48 (8):2972-2984.

56. Purushottamachar P, et al. (2007) First pharmacophore-based identification of androgen receptor down-regulating agents: discovery of potent anti-prostate cancer agents. Bioorg Med Chem, England), Vol 15, pp 3413-3421. 57. Purushottamachar P, et al. (2008) Potent anti-prostate cancer agents derived from a novel androgen receptor down-regulating agent. Bioorg Med Chem, England), Vol 16, pp 3519-3529. 58. Purushottamachar P, et al. (2013) Systematic structure modifications of multitarget prostate cancer drug candidate galeterone to produce novel androgen receptor down- regulating agents as an approach to treatment of advanced prostate cancer. Journal of medicinal chemistry 56 (12):4880-4898. 59. Kwegyir-Afful AK, Ramalingam S, Purushottamachar P, Ramamurthy VP, & Njar VC (2015) Galeterone and VNPT55 induce proteasomal degradation of AR/AR-V7, induce significant apoptosis via cytochrome c release and suppress growth of castration resistant prostate cancer xenografts in vivo. Oncotarget 6 (29):27440-27460. 60. Montgomery B, et al. (2016) Androgen Receptor Modulation Optimized for Response (ARMOR) Phase I and II Studies: Galeterone for the Treatment of Castration-Resistant Prostate Cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 22 (6):1356-1363. 61. Bruno RD, Gover TD, Burger AM, Brodie AM, & Njar VC (2008) 17alpha- Hydroxylase/17,20 lyase inhibitor VN/124-1 inhibits growth of androgen-independent prostate cancer cells via induction of the endoplasmic reticulum stress response. Molecular cancer therapeutics 7 (9):2828-2836. 62. Kepp O, et al. (2015) eIF2alpha phosphorylation as a biomarker of immunogenic cell death. Seminars in cancer biology 33:86-92. 63. Lee YK, Brewer JW, Hellman R, & Hendershot LM (1999) BiP and immunoglobulin light chain cooperate to control the folding of heavy chain and ensure the fidelity of immunoglobulin assembly. Molecular biology of the cell 10 (7):2209-2219. 64. Schauble N, et al. (2012) BiP-mediated closing of the Sec61 channel limits Ca2+ leakage from the ER. The EMBO journal 31 (15):3282-3296. 65. Bertolotti A, Zhang Y, Hendershot LM, Harding HP, & Ron D (2000) Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response. Nature cell biology 2 (6):326-332. 66. Zhu X, et al. (1996) Structural analysis of substrate binding by the molecular chaperone DnaK. Science (New York, N.Y.) 272 (5268):1606-1614. 67. Stan RC, Silva RL, & de Camargo MM (2017) Human GRP78 affinity towards its signaling partners Ire1alpha and PERK is differently modulated by an unfolded protein client. Biochemical and biophysical research communications. 68. Shen J, Snapp EL, Lippincott-Schwartz J, & Prywes R (2005) Stable binding of ATF6 to BiP in the endoplasmic reticulum stress response. Molecular and cellular biology 25 (3):921- 932.

69. Oikawa D, Kimata Y, Kohno K, & Iwawaki T (2009) Activation of mammalian IRE1alpha upon ER stress depends on dissociation of BiP rather than on direct interaction with unfolded proteins. Experimental cell research 315 (15):2496-2504. 70. Pincus D, et al. (2010) BiP binding to the ER-stress sensor Ire1 tunes the homeostatic behavior of the unfolded protein response. PLoS biology 8 (7):e1000415. 71. Shen J, Chen X, Hendershot L, & Prywes R (2002) ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of Golgi localization signals. Developmental cell 3 (1):99-111. 72. Matsumoto M, Minami M, Takeda K, Sakao Y, & Akira S (1996) Ectopic expression of CHOP (GADD153) induces apoptosis in M1 myeloblastic leukemia cells. FEBS letters 395 (2-3):143-147. 73. Teske BF, et al. (2013) CHOP induces activating transcription factor 5 (ATF5) to trigger apoptosis in response to perturbations in protein homeostasis. Molecular biology of the cell 24 (15):2477-2490. 74. Iurlaro R & Munoz-Pinedo C (2016) Cell death induced by endoplasmic reticulum stress. The FEBS journal 283 (14):2640-2652. 75. Alder NN, Shen Y, Brodsky JL, Hendershot LM, & Johnson AE (2005) The molecular mechanisms underlying BiP-mediated gating of the Sec61 translocon of the endoplasmic reticulum. The Journal of cell biology 168 (3):389-399. 76. Sharma AK & Rohrer B (2004) Calcium-induced calpain mediates apoptosis via caspase- 3 in a mouse photoreceptor cell line. The Journal of biological chemistry 279 (34):35564- 35572. 77. Das A, et al. (2006) Calpeptin provides functional neuroprotection to rat retinal ganglion cells following Ca2+ influx. Brain research 1084 (1):146-157. 78. Park KS, Poburko D, Wollheim CB, & Demaurex N (2009) Amiloride derivatives induce apoptosis by depleting ER Ca (2+) stores in vascular endothelial cells. British journal of pharmacology 156 (8):1296-1304. 79. Hammadi M, et al. (2013) Modulation of ER stress and apoptosis by endoplasmic reticulum calcium leak via translocon during unfolded protein response: involvement of GRP78. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 27 (4):1600-1609. 80. Giorgi C, et al. (2015) p53 at the endoplasmic reticulum regulates apoptosis in a Ca2+- dependent manner. Proceedings of the National Academy of Sciences of the United States of America 112 (6):1779-1784. 81. O'Boyle NM, et al. (2011) Open Babel: An open chemical toolbox. Journal of cheminformatics 3:33. 82. Morris GM, Goodsell, D. S., Halliday, R. S., Huey, R., Hart, W. E., Belew, R. K., Olson, A. J. (1998) Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. Journal of computational chemistry 19 (14):1639-1662.

83. Trott O & Olson AJ (2010) AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. Journal of computational chemistry 31 (2):455-461. 84. Meslamani J, Rognan D, & Kellenberger E (2011) sc-PDB: a database for identifying variations and multiplicity of 'druggable' binding sites in proteins. Bioinformatics (Oxford, England) 27 (9):1324-1326. 85. Zhang J, Thomas TZ, Kasper S, & Matusik RJ (2000) A small composite probasin promoter confers high levels of prostate-specific gene expression through regulation by androgens and glucocorticoids in vitro and in vivo. Endocrinology 141 (12):4698-4710. 86. Ogata M, et al. (2006) Autophagy is activated for cell survival after endoplasmic reticulum stress. Molecular and cellular biology 26 (24):9220-9231. 87. Lopez-Hernandez FJ, Ortiz MA, Bayon Y, & Piedrafita FJ (2003) Z-FA-fmk inhibits effector caspases but not initiator caspases 8 and 10, and demonstrates that novel anticancer retinoid-related molecules induce apoptosis via the intrinsic pathway. Molecular cancer therapeutics 2 (3):255-263. 88. Gao G & Dou QP (2000) N-terminal cleavage of bax by calpain generates a potent proapoptotic 18-kDa fragment that promotes bcl-2-independent cytochrome C release and apoptotic cell death. Journal of cellular biochemistry 80 (1):53-72. 89. Hamman BD, Hendershot LM, & Johnson AE (1998) BiP maintains the permeability barrier of the ER membrane by sealing the lumenal end of the translocon pore before and early in translocation. Cell 92 (6):747-758. 90. Wisniewska M, et al. (2010) Crystal structures of the ATPase domains of four human Hsp70 isoforms: HSPA1L/Hsp70-hom, HSPA2/Hsp70-2, HSPA6/Hsp70B', and HSPA5/BiP/GRP78. PloS one 5 (1):e8625. 91. Buttigliero C, et al. (2015) Understanding and overcoming the mechanisms of primary and acquired resistance to abiraterone and enzalutamide in castration resistant prostate cancer. Cancer treatment reviews 41 (10):884-892. 92. Rodriguez-Gonzalez A, et al. (2008) Targeting steroid hormone receptors for ubiquitination and degradation in breast and prostate cancer. Oncogene 27 (57):7201- 7211. 93. Gong Y, et al. (2012) Nuclear export signal of androgen receptor (NESAR) regulation of androgen receptor level in human prostate cell lines via ubiquitination and proteasome- dependent degradation. Endocrinology, United States), Vol 153, pp 5716-5725. 94. Lin HK, Wang L, Hu YC, Altuwaijri S, & Chang C (2002) Phosphorylation-dependent ubiquitylation and degradation of androgen receptor by Akt require Mdm2 E3 ligase. The EMBO journal 21 (15):4037-4048. 95. Ermakova SP, et al. (2006) (-)-Epigallocatechin gallate overcomes resistance to etoposide-induced cell death by targeting the molecular chaperone glucose-regulated protein 78. Cancer research 66 (18):9260-9269.

96. Bhattacharjee R, Devi A, & Mishra S (2015) Molecular docking and molecular dynamics studies reveal structural basis of inhibition and selectivity of inhibitors EGCG and OSU- 03012 toward glucose regulated protein-78 (GRP78) overexpressed in glioblastoma. Journal of molecular modeling 21 (10):272. 97. Wang AM, et al. (2013) Activation of Hsp70 reduces neurotoxicity by promoting polyglutamine protein degradation. Nature chemical biology 9 (2):112-118. 98. Rees I, Lee S, Kim H, & Tsai FT (2006) The E3 ubiquitin ligase CHIP binds the androgen receptor in a phosphorylation-dependent manner. Biochimica et biophysica acta 1764 (6):1073-1079. 99. Segnitz B & Gehring U (1997) The function of steroid hormone receptors is inhibited by the hsp90-specific compound geldanamycin. The Journal of biological chemistry 272 (30):18694-18701. 100. Vanaja DK, Mitchell SH, Toft DO, & Young CY (2002) Effect of geldanamycin on androgen receptor function and stability. Cell stress & chaperones 7 (1):55-64. 101. Centenera MM, et al. (2015) Co-targeting AR and HSP90 suppresses prostate cancer cell growth and prevents resistance mechanisms. Endocrine-related cancer 22 (5):805-818. 102. He B, et al. (2004) An androgen receptor NH2-terminal conserved motif interacts with the COOH terminus of the Hsp70-interacting protein (CHIP). The Journal of biological chemistry 279 (29):30643-30653. 103. Sedarous M, et al. (2003) Calpains mediate p53 activation and neuronal death evoked by DNA damage. The Journal of biological chemistry 278 (28):26031-26038. 104. Del Bello B, Moretti D, Gamberucci A, & Maellaro E (2007) Cross-talk between calpain and caspase-3/-7 in cisplatin-induced apoptosis of melanoma cells: a major role of calpain inhibition in cell death protection and p53 status. Oncogene 26 (19):2717-2726. 105. Karimian A, Ahmadi Y, & Yousefi B (2016) Multiple functions of p21 in cell cycle, apoptosis and transcriptional regulation after DNA damage. DNA repair 42:63-71. 106. Liu T, et al. (2012) P21-activated kinase 6 (PAK6) inhibits prostate cancer growth via phosphorylation of androgen receptor and tumorigenic E3 ligase murine double minute- 2 (Mdm2). The Journal of biological chemistry).

100