GLUCOCORTICOID RECEPTOR PARTIALLY REPLACES ANDROGEN RECEPTOR SIGNALING IN
ENZALUTAMIDE-TREATED PROSTATE CANCER CELLS
Laura Helminen
Master of Science thesis Master´s Degree Programme in Biomedicine
University of Eastern Finland Faculty of Health Sciences School of Medicine
3.6.2020
University of Eastern Finland, Faculty of Health Sciences, School of Medicine Master´s Degree Programme in Biomedicine
Laura Helminen: Glucocorticoid receptor partially replaces androgen receptor signaling in enzalutamide-treated prostate cancer cells
Master of Science thesis: 36 pages + Supplementary tables Supervisors: postdoctoral researcher Ville Paakinaho, professor Jorma Palvimo
3.6.2020
Keywords: Glucocorticoid receptor, enzalutamide, chromatin, castration-resistant prostate cancer, antiandrogen, androgen receptor
Abstract
Prostate cancer is one of the most common cancer of men worldwide and it is driven by the androgen receptor (AR) signaling. Enzalutamide is a second-generation antiandrogen, which antagonizes AR and is used for the treatment of castration-resistant prostate cancer. Eventually, resistance to enzalutamide prevails. It has been shown that in enzalutamide resistance, the glucocorticoid receptor (GR) signaling is increased. GR can replace and activate AR target genes despite the AR blockade, but the exact mechanisms are yet unknown. The GR is an essential protein for life and therefore cannot be fully antagonized. Furthermore, GR activating synthetic glucocorticoids are often used as an adjunct therapy to relief side effects of the prostate cancer treatments. In this master’s thesis study, the more precise genome-wide mechanisms of GR in enzalutamide-treated VCaP prostate cancer cells were investigated. This was achieved by utilizing immunoblotting, RT-qPCR and deep sequencing-based chromatin accessibility, binding and RNA-profiling techniques. Prolonged enzalutamide treatment was found to increase protein levels of both AR and GR and affect cell growth-related pathways.
Genome-wide chromatin binding of AR was found to be substantially reduced in the presence of enzalutamide, whereas GR’s binding increased and new binding sites emerged. Most importantly, the results showed that GR replaces the chromatin binding of AR only partially, and the replacement occurred at chromatin sites found to be already pre-accessible. The results of the study increase the knowledge of enzalutamide resistance mechanisms, which could be utilized in development of prostate cancer pharmaceuticals. More research is needed to determine if the partial replacement itself is sufficient to confer resistance or if other factors have a role in it.
1
Introduction
Prostate is an organ that produces nutrient rich fluid for sperm. In males, besides the male phenotype, androgens, such as testosterone and especially 5α-dihydrotestosterone (DHT), are important for the development of the prostate and maintenance of its functions after puberty (Schiffer et al., 2018). Prostate cancer is one of the most common cancer of men worldwide and it is more common in developed countries (Center et al., 2012). The incidence rates of prostate cancer are increasing due to the aging of population and increased diagnosis, while death rates are decreasing due to the developed treatment options (Cronin et al., 2018).
Lifestyle factors, such as diet and smoking, susceptive genetics and hormones play roles in prostate cancer etiology and development, but the highest risk factor is age. The mean age at the time of diagnosis in the United States is close to 70 (Grozescu & Popa, 2017).
Prostate cancer is usually diagnosed with elevated levels of prostate-specific antigen (PSA) biomarker in blood, prostate biopsy and imaging the prostate (Grozescu & Popa, 2017). The PSA test has preponed diagnosis and improved life expectancy, but it also has led to overdiagnosis and overtreatment of the disease (Armstrong et al., 2017). As observed from the biopsies, prostate cancer is usually an adenocarcinoma (Humphrey, 2017). Biopsy sample can be scored and classified, by a grading scale called Gleason score, to help diagnosis and the treatment process. Gleason score, though given some criticism (Epstein et al., 2016), is based on the differentiation level of the tissue, where higher scores indicate more aggressiveness and worse prognosis (Gordetsky & Epstein, 2016).
Based on the status of the disease, identified cancers can be treated with surgery, radiotherapy and/or hormone therapy to reduce blood androgen and therefore PSA levels. Alternatively, cancer can be just actively monitored without therapy intervention. For most of the patients, prostate cancer is not lethal (Grozescu & Popa, 2017), though it can have a huge impact on patient’s quality of life (Holm et al., 2018). Nearly all men develop benign prostate hyperplasia by the age of 90. However, hyperplasia is not generally considered to be a precursor lesion to prostate cancer, even though symptoms like difficulty in urination are shared between the two (Aaron et al., 2016). Despite this, small latent prostate carcinoma was found around 10% of men in an autopsy (Breslow et al., 1977).
2 As indicated earlier, on a molecular level, androgens are important for the prostate development and maintenance. In general, androgen signaling regulates multiple cellular events of the normal and cancerous prostate, such as proliferation, apoptosis, migration, invasion and differentiation (Culig & Santer, 2014). The main endogenous androgen testosterone is synthesized from cholesterol mainly in the testes and in some extent in the adrenal glands.
Especially in the prostate tissue, DHT which has greater affinity towards androgen receptor and is therefore more potent androgen, is locally converted from testosterone by 5α-reductase (Schiffer et al., 2018).
Actions of endogenous and synthetic (such as methyltrienolone R1881) androgens in the cells are mediated through ~130 kDa size transcription factor androgen receptor (AR), which is expressed in many tissues, but at the highest levels in the reproductive organs. AR belongs to a hormone-dependent steroid receptor subfamily of nuclear receptor superfamily, and shares structural [N-terminal transactivation domain (NTD), central DNA-binding domain (DBD), hinge region and C-terminal ligand-binding domain (LBD)], functional and activating ligand similarities with the other members (glucocorticoid, estrogen, progesterone and mineralocorticoid receptor) of the subfamily (Figure 1) (Dai et al., 2017).
Figure 1. Similarity of the steroid receptors. (A) Phylogenetic tree of the steroid receptors showing the close evolutionary relationships. (B) Sequence homologies showing the similar domain structure (in different colors) of the steroid receptors. Dark green=NTD, light green=DBD, red=hinge region, grey=LBD, orange=additional C- terminal domain of ERα, ER=estrogen receptor, AR=androgen receptor, GR=glucocorticoid receptor, MR=mineralocorticoid receptor, PR=progesterone receptor, AF-1=transactivation region 1, AF-2=transactivation region 2. Numbers represent the length of the receptors in amino acids (Modified from Griekspoor et al., 2007).
In the blood stream, non-active androgens circulate attached to sex-hormone binding globulins.
Unbound, biologically active and lipophilic androgens diffuse through cell membrane and bind to the LBD of cytosolic AR that is in a complex with chaperone heat-shock proteins. In the
3 absence of androgens, while associated with the complex, AR remains transcriptionally inactive. After androgen binding to the LBD, conformational change induces AR dissociation, homodimerization and translocation into the nucleus. At the nucleus, AR binds to its target sequences in the promoter and distal enhancer areas of DNA called androgen responsive elements (AREs). This results in the recruitment of basal transcriptional machinery and coactivators or corepressors among others to the AR bound enhancers and promoters.
Subsequently, transcription of AR’s target genes is activated or repressed (Tan et al., 2015).
AR can also rapidly and non-genomically activate MAPK and Akt cell proliferation and survival kinase pathways (Leung & Sadar, 2017).
Androgen signaling is essential for normal prostate function and therefore also important for prostate cancer development and a target of hormone treatments (Culig & Santer, 2014). In prostate cancer, some part of the signaling pathway is dysregulated, usually being the AR protein and/or its function (Dai et al., 2017), which leads to aberrant and uncontrolled expression of AR target genes and malignant cell proliferation. In prostate cancer cases, typically at least one aberration is detected: AR point mutation, AR splice variant, AR gene amplification, AR upregulation, fusion genes, or crosstalk with other pathways (Ferraldeschi et al., 2015). Also, activation and amplification of an enhancer regulating the AR gene has also been observed (Takeda et al., 2018). These changes can lead to AR’s upregulation resulting in insensitivity to external signals. Although most of the genetic aberrations develop during tumorigenesis in prostate tissue, some prostate cancer development and treatment associated single-nucleotide polymorphisms (SNPs) can be seen in germline too (Hua et al., 2018, Jin et al., 2012, Sharifi et al., 2008).
Apoptosis is induced when androgen levels in the cells decrease below certain threshold (Medh
& Thompson, 2000). In prostate cancer treatment, this can be achieved by surgical or chemical (GnRH antagonist to block testosterone production) castration or blocking the AR action with antiandrogens. Antiandrogens are antagonists that compete with androgens and bind to AR, inhibit it from binding to DNA and therefore repress its actions (Gillatt, 2006). Almost all patients initially benefit from hormonal treatments and blood PSA levels decrease, though side effects, such as flushing and sexual dysfunction, affect the patients’ quality of life (Tucci et al., 2018). Eventually within few years, tumor adapts to the treatments, becomes resistant and starts to develop and send metastasis, especially to bones (Logothetis et al., 2018). This state of the disease is called (metastatic) castration-resistant prostate cancer ((m)CRPC), because it
4 progresses despite the castration levels of androgens. Even more (genetic) aberrations are seen in the CRPC state (Grasso et al., 2012). CRPC patients usually have only months to live, since the current drugs only slow down the cancer progression (Chandrasekar et al., 2015).
Enzalutamide (ENZ) is a relatively new and widely used second-generation nonsteroidal antiandrogen. It is meant for both metastatic and non-metastatic CRPC and it antagonizes AR better than the first-generation antiandrogens, such as bicalutamide. ENZ prolongs CRPC patient survival by months and causes less side effects compared to the previous drugs (Tran et al., 2009). Another commonly used drug for CRPC is abiraterone. This drug is a CYP17A1 enzyme inhibitor that inhibits androgen synthesis in the testes and especially in the adrenal glands. A metabolite of abiraterone also acts as an antiandrogen (Attard et al., 2005).
Unfortunately, within months, resistance to ENZ (and abiraterone) prevails. ENZ (and abiraterone’s metabolite) bind to the LBD of AR, why for example AR splice variant 7 (AR- V7) lacking the LBD can mediate drug resistance (Antonarakis et al., 2014). That and other previously described aberrations in AR signaling can only partially explain ENZ resistance.
Additionally, upregulation of glucocorticoid receptor signaling has been revealed to play a role in the development of ENZ insensitivity (Arora et al., 2013, Li et al., 2017).
Glucocorticoids, like androgens, are cholesterol derived steroid hormones that are synthesized in the zona fasciculata of the adrenal gland in a circadian manner and in response to emotional and physical stress. Glucocorticoids are essential for life, since they are involved in many important physiological processes, including stress and inflammatory reactions, glucose metabolism and cell homeostasis (Ramamoorthy & Cidlowski, 2016). The response to glucocorticoids varies between cells and tissues and is very dynamic in the regulation of transcriptional repression and activation (John et al., 2009).
The most important human glucocorticoid is cortisol, which acts through ubiquitously expressed, ~100 kDa size glucocorticoid receptor (GR). The GR exists in two main isoforms, GRα being the active GR and GRß being formed by alternative splicing. GRα binds cortisol and synthetic glucocorticoids, whereas GRß does not (Narayanan et al., 2016). Thus, GRβ acts mainly as a dominant negative to the GRα’s function (Lewis-Tuffin & Cidlowski, 2006).
Synthetic glucocorticoids, such as dexamethasone, are used as a drug for variety of inflammatory diseases such as asthma and arthritis, as well as for the treatment of hematological cancers (Ramamoorthy & Cidlowski, 2016). The GR and therefore its activity
5 and protein-protein interactions are heavily influenced by post-translational modifications such as phosphorylation and sumoylation (Weikum et al., 2017).
The glucocorticoid signaling is very similar to androgen signaling. The GR also shares similar basic structure with AR (Figure 1). In the bloodstream, non-active cortisol is bound to corticosteroid-binding globulin. The bioavailability of cortisol is also regulated by 11β- hydroxysteroid dehydrogenases (11β-HSD) that convert cortisol into inactive cortisone (11β- HSD2) and back (11β-HSD1) (Li et al., 2017). In the absence of glucocorticoids, GR resides in the cytoplasm in a complex with heat-shock proteins, FKBP5 immunophilins and non- receptor tyrosine kinases (Kassi & Moutsatsou, 2011). Like AR, GR remains transcriptionally inactive when bound to the complex. Unbound, biologically active and lipophilic glucocorticoids diffuse through cell membrane and bind to the LBD of cytosolic GR. After conformational change, GR dissociates from the protein complex and translocates into the nucleus. There it binds mainly as a homodimer to its target sequences in promoters and mainly in the distal enhancer areas of DNA called glucocorticoid responsive elements (GREs) (Oakley
& Cidlowski, 2013). Higher oligomerization of GR, such as tetramerization, has also been reported (Presman & Hager, 2017). Receptor binding to GRE results in the recruitment of basal transcriptional machinery and coactivators or corepressors among others to the GR bound enhancers and promoters. Subsequently, transcription of GR’s target genes is activated or repressed. GR can also affect transcription by binding to other transcription factors, such as STAT (tethering), interacting with other transcription factors such as AP-1 (composite binding sites) or via rapid, non-genomic activation of kinase pathways (Figure 2) (Cain & Cidlowski, 2015).
6
Figure 2. Genomic (direct, tethering, composite) and non-genomic (kinase) GR signaling pathways induced by glucocorticoids. The response can be activating or repressing. nGRE=negative GRE (Modified from Cain &
Cidlowski, 2015).
Due to their wide and various actions in the body, glucocorticoids are used to treat different symptoms and diseases. The most common indication for their use is their ability to reduce inflammation. It is thought that this happens mostly through GR-mediated inhibition of pro- inflammatory AP-1 and NF-κB function (Escoter-Torres et al., 2019). In hematological malignancies, glucocorticoids are used for their antiproliferative and pro-apoptotic effects (Pufall, 2015). Related to prostate cancer, glucocorticoids are often used to relief the side effects, such as nausea, of the primary treatments (Hu & Chen, 2017) and due to their inhibitory effects on adrenal androgen production (Kassi & Moutsatsou, 2011). In the early stages of prostate cancer, glucocorticoids suppress both tumor angiogenesis (Yano et al., 2006a) and lymphangiogenesis (Yano et al., 2006b). With the abiraterone treatment, glucocorticoid supplementation is needed, because the drug also inhibits the production of glucocorticoids in the adrenal glands (Ryan et al., 2015). However, this is not trivial, as the glucocorticoid homeostasis is very sensitive. Prolonged use of glucocorticoids has many unfavorable side- effects such as osteoporosis, obesity and cardiovascular disease (Oray et al., 2016). In the context of prostate cancer, one of the unwanted effects with the use of glucocorticoids in the treatment regimen is also the observed GR upregulation in ENZ resistance. Due to this, GR activating glucocorticoid therapy could be detrimental for a prostate cancer patient (Arora et al., 2013).
7 As indicated earlier, AR and GR signaling pathways have been revealed to be very similar and also partly overlapping, which contributes to the AR target gene activation by GR in ENZ- resistant prostate cancer (Arora et al., 2013). In this context, aberrations in AR signaling are usually genetic, whereas for GR they are usually non-genetic, such as epigenetic (Arora et al., 2013, Shah et al., 2017). For example, L701H and T877A point mutations in LBD of AR promote its promiscuity towards glucocorticoids (Krishnan et al., 2002, Zhao et al., 2000), and F876L point mutation in the same domain turns ENZ’s function from antagonist to agonist (Pollock et al., 2016). AR and GR have very conserved domains. This is especially evident with the receptor’s zinc fingers in the DBD that bind to the ARE or GRE in the DNA, with the consensus sequence being 5′-AGAACAnnnTGTTCT-3′ for both receptors (Denayer et al., 2010). Approximately half of the AR binding events overlap with that of GR, meaning that the GR is capable of modulating many genes in the AR pathway and maintain the resistant phenotype even though the AR is blocked with antiandrogens such as ENZ (Arora et al., 2013, Sahu et al., 2013). Furthermore, AR and GR can share several interacting coactivator and corepressor proteins (Lempiainen et al., 2017).
GR is expressed in normal prostate tissue, but its expression is substantially reduced in primary prostate cancer (Yemelyanov et al., 2007). However, GR mRNA and protein levels have been seen to be increased in both experimental and clinical CRPC samples obtained after ENZ treatment, though not back to the levels seen in benign tissue (Li et al., 2017, Shah et al., 2017).
AR directly represses GR expression via tissue-specific negative ARE located near the GR locus (Shah et al., 2017), and when AR is blocked with an antiandrogen like ENZ, GR mRNA and protein levels can rise in the absence of AR-mediated repression. Also, GR’s function in prostate can switch from normally being anti-proliferative to pro-cell survival after AR blockade (Isikbay et al., 2014).
Gene regulation is a complex system involving genetic and epigenetic aspects ranging from transcription factors, coregulators and chromatin modifiers at the beginning and post- transcriptional modifications at the end of the processes (Iglesias-Platas & Monk, 2016, Pope
& Medzhitov, 2018). To study these transcriptional mechanisms, such as those of AR and GR, in the cells on a genome-wide level, several methods have been developed. One of the most established methods is immunoblotting, where gel electrophoresis separated proteins of interest are detected using specific antibodies (Towbin et al., 1979). More recently, new deep sequencing methods have been developed such as chromatin immunoprecipitation coupled
8 with high-throughput sequencing (ChIP-seq), assay for transposase-accessible chromatin coupled with high-throughput sequencing (ATAC-seq) and high-throughput sequencing of RNA (RNA-seq). Using crosslinking and immunoprecipitation with specific antibodies, ChIP- seq detects protein-DNA (and to some extend protein-protein) interactions and is used to identify the DNA sites and sequences proteins of interest bind to. Besides global binding sites of transcription factors, it can be used to map the occurrence of epigenetic marks, such as histone modifications (Mundade et al., 2014). ATAC-seq assesses genome-wide accessible chromatin sites using hyperactive Tn5 transposase enzyme (Buenrostro et al., 2015). The openness of chromatin region is usually associated with transcriptional activity. Sequencing all or a subset (such as mRNAs) of RNAs, RNA-seq has the ability to reveal the expression of the whole transcriptome, which can then be used for example in pathway enrichment analysis (Hrdlickova et al., 2017). Reverse transcription quantitative PCR (RT-qPCR) is a method that also in general assays mRNA, but only known genes with specific primers (Taylor, S. C. et al., 2019). All the above-mentioned methods describe the same phenomenon from a different point-of-view. They all have their limitations, why they are needed to support each other’s information (Zhang, L. et al., 2018). Transcription factors’ binding and coregulators’ actions are fast and dynamic, why several methods are needed to reveal their cooperative actions. Also, majority of the promoter-enhancer interactions, especially of GR, are not located to the nearest transcription start site (TSS) (D'Ippolito et al., 2018, Hakim et al., 2009, Zhang, Y. et al., 2013).
Although much is known about the AR-GR interactions in prostate cancer, the more precise genome-wide mechanisms of GR overtake in ENZ resistance are still unknown. In this master’s thesis study, those mechanisms of GR were investigated using previously mentioned genome- wide techniques in ENZ-treated VCaP prostate cancer cells. As expected, prolonged ENZ treatment was found to increase mRNA and protein levels of GR, and to some extend those of AR, and affect cell growth related pathways. Chromatin binding of AR was found to be substantially reduced in ENZ treatment, whereas GR’s binding increased and new binding sites emerged. Most importantly, the results revealed that GR replaces the chromatin binding of AR only partially and the replacing occurred in already open chromatin binding sites. The results of the study help to better understand the mechanisms of ENZ resistance and its development, and to further help to improve better pharmaceuticals and treatments for prostate cancer patients.
9
Materials and methods
Cell culture
VCaP (ATCC) bone metastasis prostate cancer cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Gibco, 41965-039) containing 10% of U.S. origin Fetal Bovine Serum (FBS) (GE Healthcare, 11531831) and 1% penicillin-streptomycin (Gibco, 15140122), and incubated in humidified, 5% CO2 and 37 °C atmosphere. Enzalutamide treated cells were cultured in normal growth medium supplemented with 10 µM enzalutamide (ENZ) (Orion, ORM-0016678) for approximately 21 days (ENZ treatment ranged between 17-24 days) changing the medium every 3-4 days. For RT-qPCR, RNA-seq and some of the immunoblot, ChIP-seq and ATAC-seq samples, growth medium was changed into or replaced with experiment medium (growth medium with 10% FBS replaced with 10% charcoal-stripped South America origin FBS (Gibco, 10270106)). In the experiments, 100 nM dexamethasone (Dex) (Sigma-Aldrich, D4902) and 100 nM 5α-dihydrotestosterone (DHT) (Steraloids, A2570- 000) were used to activate GR and AR, while ethanol was used as a vehicle. Cells were used for a maximum of 30 passages.
Immunoblotting
To assess ENZ’s effects on GR and AR protein levels, immunoblot was performed. Cells were grown in either normal or experiment medium supplemented with 10 µM ENZ for approximately 21 days (ENZ treatment ranged between 17-24 days). Untreated cells grown in normal or experiment medium were used as controls. Proteins were isolated and separated according to the general protocol (Towbin et al., 1979). Briefly, cells were collected in [TBS + protease inhibitor cocktail (PIC (Sigma, #11836145001)) + 20 mM NEM] solution, and centrifuged pellets were resuspended in [SDS + PIC + 10 mM NEM]. After sonication and boiling, proteins were separated with 10% SDS-PAGE gel and transferred onto nitrocellulose membrane. Membranes were blocked with 5% milk and incubated with primary antibodies at 4 °C overnight. The used antibodies were anti-GR (1:1000, Cell Signaling Technology,
#12041) and anti-AR (1:10 000, K183, noncommercial rabbit serum (Karvonen et al., 1997)).
An anti-α tubulin (1:3000, Santa Cruz Biotechnology, sc-5286) was used as a control for sample loading and signal quantification. Anti-rabbit (1:10 000, Invitrogen, G-21234) was used as a secondary antibody for GR and AR, and anti-mouse (1:10 000, Zymed Laboratories, 81- 6520) for tubulin. Secondary antibodies were incubated at room temperature for 45 minutes.
10 Protein bands were detected using Pierce ECL Western Blotting Substrate kit (Thermo Scientific, 32106) and ChemiDoc Imager (Bio-Rad). Protein level quantification was calculated using ImageJ (National Institutes of Health).
ChIP-seq
To assess ENZ’s effects on GR and AR binding on chromatin, ChIP-seq was performed. Cells were grown in normal or experiment medium, with or without ENZ for 22-23 days. ChIP-seq experiments were done as previously described (Toropainen et al., 2015). Briefly, one-hour prior collection, cells were treated with 100 nM Dex (GR immunoprecipitation) or 100 nM DHT (AR immunoprecipitation). Proteins were crosslinked with 1% formaldehyde and collected in Farham Lysis Buffer [5 mM PIPES + 85 mM KCl + 0,5% IGEPAL + PIC].
Centrifuged pellets were resuspended in RIPA Buffer [PBS + 1% IGEPAL + 0,5% sodium deoxycholate + 0,1% SDS + PIC]. After sonication (30x 30 s ON/30 s OFF with Diagenode Bioruptor), samples were immunoprecipitated at 4 °C overnight. The used antibodies were anti-GR (Cell Signaling Technology, D6H2L, 12,5 µl per sample) and anti-AR (K183, home- made rabbit serum (Karvonen et al., 1997), 3 µl per sample) coupled to Invitrogen Dynabeads Protein G. Eluted [1% SDS + 0,1 M NaHCO3] samples were reverse crosslinked with 20 µg Proteinase K (Thermo Scientific, EO0492) at 65 °C overnight after which they were purified with Monarch PCR & DNA Cleanup Kit (5 μg) (New England BioLabs, T1030). In purification, technical replicates were pooled. DNA concentration was determined with Qubit Fluorometer (Invitrogen). Positive (FKBP5-(-3) and PSA enhancer, primers in Supplementary table 1) and negative (RHO, primers in Supplementary table 1) binding regions were verified with qPCR (Roche LightCycler 480 Instrument II). Samples were quantitated by normalizing to input values and untreated cells. ChIP-seq library was generated using NEBNext Ultra II DNA Library Prep Kit for Illumina (New England BioLabs, E703) according to instructions.
Analysis for size selection and quality was done with Agilent 2100 Bioanalyzer and its DNA 1000 Analysis kit (5067-1504). Based on the Bioanalyzer results, the best libraries were pooled and sequenced with Illumina NextSeq 500 (75SE). Sequencing was performed in the EMBL Genomics Core Facility (Heidelberg, Germany).
ATAC-seq
To assess ENZ’s effects on chromatin accessibility, ATAC-seq was performed. Cells were grown in either normal growth medium or changed into experiment medium two days prior the
11 experiment and treated with 10 µM ENZ for 23 days or left untreated to use as a control. Cells were treated with 100 nM Dex or vehicle (EtOH) for an hour prior collection. The ATAC-seq experiment was performed as previously described (Paakinaho et al., 2019). Briefly, the cell nuclei were isolated using Buffer A [15 mM Tris-HCl + 15 mM NaCl + 60 mM KCl + 1 mM EDTA + 0,5 mM EGTA + 0,5 mM spermidine + PIC] lysis buffer and Buffer A + 0,04%
IGEPAL, and resuspended in ATAC-RSB Buffer [10 mM NaCl + 10 mM Tris-HCl + 3 mM MgCl2]. 100 000 nuclei from each sample were used in transposition reaction. The reaction was done using Nextera DNA Flex Library Preparation Kit (Illumina, 20018705). DNA was purified with Monarch PCR & DNA Cleanup Kit (5 μg) (New England BioLabs, T1030). DNA fragments were amplified using PCR (Biometra T3 Thermocycler and Roche LightCycler 480 Instrument II), and samples were integrated with barcoded primers (custom Nextera PCR primers 1 and 2 (Buenrostro et al., 2013)). Amplified fragments were size selected (150 bp- 800 bp) using SPRIselect beads (Beckman Coulter, B23318). qPCR (Roche LightCycler 480 Instrument II) was used to verify positive (FKBP5-(-3), primers in supplementary table 2) and negative (RHO, primers in supplementary table 2) accessible sites. Analysis for size selection and quality was done with Agilent 2100 Bioanalyzer and its High Sensitivity DNA Analysis kit (5067-4626). Based on the Bioanalyzer results, the best libraries were pooled sent for sequencing with Illumina NextSeq 500 (40PE). Sequencing was performed in the EMBL Genomics Core Facility (Heidelberg, Germany).
RT-qPCR
To preliminary assess ENZ’s effects on a transcriptional level for known AR and GR target genes, RT-qPCR was performed. Untreated, 11-day-ENZ-treated and 18-day-ENZ-treated cells were divided onto 6-well plates, 800 000 cells per well, five days before RNA isolation (resulting in 0, 16 and 23 days of ENZ treatment). Medium was changed into experiment medium (±ENZ) two days before isolation. 24 hours before isolation, half of each treatment’s wells were supplemented with either 100 nM Dex, 100 nM DHT or left untreated, and 18 hours before isolation the same was repeated for the other half. Total RNA was extracted using TRIzol reagent (Invitrogen, 15596018). Quality and quantity of isolated RNA were determined with NanoDrop One (Thermo Scientific). 2 µg of total RNA from each sample was used for reverse transcription into cDNA with First Strand cDNA Synthesis Kit (Roche, 04897030001).
Genes studied were AR, GR, PSA/KLK3, SGK1, and each sample was quantitated by normalizing the sample values to RPL13A housekeeping gene and to untreated cells (primers
12 in Supplementary table 3). The qPCR was performed using Roche’s LightCycler 480 Instrument II. The relative gene expression was calculated using the 2-(ΔΔCt) method, where ΔΔCt is ΔCt(treatment)-ΔCt(vehicle) and ΔCt is Ct(target gene)-Ct(RPL13A). Ct presents the cycle at which the threshold was crossed.
RNA-seq
To study the effects of ENZ on a wider, whole-transcriptomics level compared to RT-qPCR, RNA-seq was performed. Samples were the same that were collected for the RT-qPCR but using only the 0 h and 18 h Dex treated normal, 16-day-ENZ and 23-day-ENZ samples. RNA was further DNase treated and extracted with RNeasy Plus Mini kit (Qiagen, 74134) after which the quality and quantity were determined with NanoDrop One (Thermo Scientific).
Proper RNA integrity number (RIN) was ensured using Agilent 2100 Bioanalyzer and its RNA 6000 Nano kit (5067-1511). All samples had RIN >9, indicating high integrity of RNA. RNA- seq library was generated using NEBNext Poly(A) mRNA Magnetic Isolation Module (New England BioLabs, E7490) and NEBNext Ultra II Directional RNA Library Prep with Sample Purification Beads Kit (New England BioLabs, E7765) according to manufacturer’s protocol.
Library quality was assessed with Agilent 2100 Bioanalyzer and its DNA 1000 Analysis kit (5067-1504). Pooled libraries were sequenced with Illumina NextSeq 500 (75SE). Sequencing was performed in the EMBL Genomics Core Facility (Heidelberg, Germany).
Data analysis
ChIP-seq data analysis was performed as previously described (Paakinaho et al., 2014, Toropainen et al., 2015). Publicly available FOXA1 ChIP-seq sequencing dataset was also used (Toropainen et al., 2015).
For ATAC-seq data, after filtering the low quality reads as with ChIP-seq data analysis, paired- end samples were aligned to hg38 genome using Bowtie2 (Langmead & Salzberg, 2012).
Alignment was performed with end-to-end sensitive mode allowing no mismatches. Around 30% of the reads were mapped to mitochondrial DNA and from the two biological replicate samples, at least 40 million unique non-mitochondrial reads were obtained for each condition.
Downstream data analysis was performed using HOMER (Heinz et al., 2010). Peaks in each GR dataset were called using default parameters on findPeaks with style factor, false discovery rate (FDR) <0.01, >25 tags, >4-fold over control sample and local background. ChIP input sample from VCaP cells was used as a control. DESeq2 (Love et al., 2014) through
13 getDiffrentialPeaksReplicates.pl was used to isolate differential binding peaks (ENZ-UN, ENZ-UP) (FDR <0.1, fold change >2) between the non-ENZ and ENZ treated cells. Aggregate plots and heatmaps were generated with 10 bp or 20 bp bins surrounding ±1 kb area around the center of the peak. All plots were normalized to 10 million mapped reads and further to local tag density, tags per bp per site, whereas box plots and scatter plots represented log2 tag counts.
Motif scores were calculated using AnnotatePeaks.pl and box plots represent log-odd motif scores. De novo motif searches were performed using findMotifsGenome.pl with the following parameters: 200 bp peak size window, strings with 2 mismatches, binomial distribution to score motif p-values, and 50 000 background regions. Statistical significance in the box plots was determined with One-way ANOVA with Bonferroni post hoc test.
RNA-seq data was aligned to hg38 genome using STAR2.7 (Dobin et al., 2013) with default settings and max 10 mismatches and max 10 multi-mapped reads. Differentially expressed genes were then analyzed with DESeq2 (Love et al., 2014) through HOMER (Heinz et al., 2010) for all comparisons. Total count per gene was calculated using transcripts per million (TPM) normalization and other than protein coding genes were filtered out as outliers. Genes with TPM >0.5 at least in one sample in any treatment were considered as expressed.
Differentially expressed genes were then defined as FDR <0.05 and log2 fold change ±0.5 between non-treated and dexamethasone treated samples. Heatmap displaying Z-scores for each replicate in each condition were drawn using heatmap.2 in gplots R package (R project).
Differentially expressed gene clusters were used in pathway analysis in Metascape (Zhou et al., 2019) with default settings. Lastly, gene set enrichment analysis (GSEA) was performed using fgsea 3.11 (Bioconductor) with default settings.
Results
The expression of GR is prominently increased by prolonged enzalutamide treatment of VCaP cells
First, ENZ’s effects on GR and AR expression levels was examined. After treating the cells with 10 µM of ENZ for approximately 23 days, the average protein levels of GR had increased by ~20% as quantified from immunoblots (Figure 3A and B). AR levels were also increased by ~30%. Furthermore, an ~80 kDa size AR band was evident in the immunoblots (Figure 3A).
This could represent AR-V7. In concordance, a rise in GR and AR mRNA levels was seen in
14 RT-qPCR (Figure 3C). However, the rise in the transcript levels of GR was clearly more prominent upon ENZ treatment than that of AR mRNA levels. Unbiased RNA-seq analyses indicated that GR mRNA levels are significantly increased in ENZ treated cells, whereas AR mRNA levels remain unchanged, though AR is much more prominently expressed than GR (Figure 3D and E).
Figure 3. AR and GR expression is induced by prolonged enzalutamide treatment in VCaP cells. (A) Immunoblot of control and 22-day ENZ-treated cells showing rise in GR and AR protein levels and a possible emergence of AR-V7. (B) Quantification of AR and GR protein levels from immunoblot on average 23-day ENZ treated cells showed ~30% increase in AR and ~20% in GR protein levels. N=4 for 22–24-day ENZ treated cells for both AR and GR. (C) Quantification of AR and GR mRNA levels from RT-qPCR of 23-day ENZ-treated cells showed ~50% increase in AR and ~250% in GR mRNA levels. N=3. (D) TPM values of GR (NR3C1) mRNA levels from RNA-seq with 0-day, 16-day and 24-day ENZ treated cells showed a rise in GR mRNA with ENZ treatment. P-values are calculated using One-way ANOVA with Bonferroni post hoc test (n=3). (E) TPM values of AR mRNA levels from RNA-seq of 0-day, 16-day and 24-day ENZ-treated cells showed no change. All bar graphs represent mean and standard deviation.
15 GR partially replaces AR chromatin binding after prolonged enzalutamide exposure
To observe ENZ-induced changes in AR and GR chromatin binding, ChIP-seq was performed.
After treating the cells with 10 µM ENZ for 22-23 days, AR binding triggered by 1-hour exposure of DHT was substantially reduced (Figure 4A, Figure 5A). The decrease of AR binding was uniformly observed at all ~50 000 receptor binding sites. For GR, its binding induced by 1-hour Dex exposure was substantially increased and new binding sites had emerged (Figure 4B, Figure 5B and C). GR binding sites were divided into ENZ-UN and ENZ- UP regions based on the statistical comparison and fold change (see Materials and methods section for details). ENZ-UN region consist of ~4800 binding sites where upon ENZ treatment, GR binding does not significantly change, whereas ~12 000 binding sites at ENZ-UP region, ENZ significantly increases GR binding.
Figure 4. Enzalutamide treatment reduces AR and increases GR chromatin binding. (A) Heatmap of AR binding sites in DHT ENZ conditions by ChIP-seq (red color). (B) Heatmap of GR binding sites in Dex ENZ conditions by ChIP-seq (blue color), and heatmap of chromatin accessibility at GR binding sites in Dex ENZ conditions by ATAC-seq (purple color). GR binding sites are divided into ENZ-UN (log2 fold change <1) and ENZ-UP (log2 fold change >1) regions based on GR binding ENZ. Each heatmap represents ±1 kb around the center of the receptor peak. Binding intensity scale is noted below on a linear scale. All heatmaps are normalized to a total of 10 million reads and further to tags per site per bp. ENZ-UN=log2 fold change <1 after ENZ treatment, ENZ-UP=log2 fold change >1 after ENZ treatment. FDR <0.1.
A B
16
Figure 5. GR chromatin binding is increased upon enzalutamide treatment not only at ENZ-UP but to some extend also at ENZ-UN sites. Histograms of AR (A) and GR (B, C) chromatin binding after cognate hormone
ENZ treatment. Each histogram represents ±1 kb around the center of the receptor peak. All histograms are normalized to a total of 10 million reads and further to tags per site per bp. ENZ-UN=log2 fold change <1 after ENZ treatment, ENZ-UP=log2 fold change >1 after ENZ treatment. FDR <0.1.
Comparison of receptor binding at GR-bound sites (Figure 6) indicated that for majority of GR binding sites (~84.5%), ENZ treatment results in an increase of GR binding, while at the same time AR binding decreases. However, since AR binds vastly more chromatin sites (~50 000) without ENZ treatment compared to GR (~16 600) after ENZ treatment (Figure 4A and B), the GR only partially replaces the AR chromatin binding in ENZ-treated VCaP cells.
Figure 6. Majority of the sites with increased GR binding after enzalutamide treatment display decreased AR binding. Scatter plot indicates log2 AR tag change (ENZ+DHT/DHT) in the x-axis, and log2 GR tag change (ENZ+Dex/Dex) on the y-axis at GR binding sites. Different effect of AR and GR binding is indicated with color coding and with arrows. Percentage of GR binding sites in each quadrant is shown. Scatter plot log2 tag counts are normalized to 10 million reads.
A B C
17 Chromatin binding of GR in enzalutamide-treated cells occurs at pre-accessible sites that are enriched with FOXA1 motifs
In order to evaluate the influence of ENZ on chromatin accessibility at GR binding sites, ATAC-seq was performed. Activation of GR did not increase but decrease the chromatin accessibility in non-ENZ-treated conditions (Figure 4B, Figure 7). After 23-day exposure of 10 µM ENZ, chromatin accessibility was significantly reduced in both ENZ-UN and ENZ-UP clusters (Figure 4B, Figure 7). In the presence of ENZ, GR’s function changed into being able to open chromatin when activated with Dex (Figure 7). Interestingly, regardless of chromatin accessibility changes, GR mostly binds to already open chromatin sites (Figure 4B).
Figure 7. Chromatin accessibility is reduced in enzalutamide-treated cells, but activated GR is then able to open chromatin. Box plot displaying log2 ATAC tag count at GR-binding sites in indicated conditions. All box plot comparisons are normalized to total of 10 million reads. P-values are calculated using One-way ANOVA with Bonferroni post hoc test. ENZ-UN=log2 fold change <1 after ENZ treatment, ENZ-UP=log2 fold change >1 after ENZ treatment.
Subsequently, de novo motif analysis was performed to observe additional transcription factors that might play a role at GR binding sites. Analyses revealed AR, GR (ARE/GRE, NR3C) as well as transcription factor FOXA1 and HOXB13 motifs to be the most enriched at GR-binding sites (Figure 8A). FOXA1 motifs were the most highly enriched with ~63% of the sites harboring the motif. There was no apparent change in FOXA1 and HOXB13 motifs between ENZ-UN and ENZ-UP, but ARE and GRE motifs were more enriched in ENZ-UP regions (Figure 8B). Furthermore, FOXA1 ChIP-seq data from VCaP cells (treated without ENZ, Dex or DHT) indicated, that FOXA1 binds onto both ENZ-UN and ENZ-UP regions ((Toropainen et al., 2015) Figure 8C). These results suggest that FOXA1 might have a role in the sites where GR replaces AR.
18
Figure 8. The most enriched motifs in GR-binding sites are AR/GR as well as transcription factors FOXA1 and HOXB13. (A) De novo motif analysis scores of top three motifs between ENZ-UN and ENZ-UP regions.
Enriched de novo motifs are displayed on the left. P-value, % of sites with motif, and % of background (bg) sites with motif are indicated next to the motifs. (B) Box plot display log-odd motif scores of most enriched motifs between ENZ-UN and ENZ-UP regions. (C) Histogram of FOXA1 chromatin binding without any treatment on ENZ-UN and ENZ-UP regions. Histogram represents ±1 kb around the center of the receptor peak and is normalized to a total of 10 million reads and further to tags per site per bp. ENZ-UN=log2 fold change <1 after ENZ treatment, ENZ-UP=log2 fold change >1 after ENZ treatment.
GR’s transcriptional regulatory potential is increased in enzalutamide-treated VCaP cells Lastly, to study the ENZ induced changes in transcriptomics, RNA-seq was performed. When the cells were treated with 10 µM ENZ for 16-24 days, significantly more genes (~600) became Dex-regulated (after 18 h of Dex exposure) compared to non-ENZ treated cells (~100 genes) (Figure 9, Figure 10). This was observed in both Dex upregulated (DexUP, from 58 to 335) as well as Dex downregulated (DexDN, from 37 to 279) genes.
A
B C
19
Figure 9. The number of Dex-regulated genes is significantly increased in enzalutamide-treated VCaP cells.
Venn diagrams display the number of Dex-regulated unique and shared genes ENZ treatment. ENZ treatments of 16 and 24-day were combined into one group (ENZ+Dex) due to their similarity. Comparison of all Dex- regulated genes is shown on the left, and comparison of Dex upregulated (top) and Dex down-regulated (down) genes are shown on the right. Statistical definition of Dex-regulated genes in displayed in the figure. DexUP=Dex upregulated genes, DexDN=Dex down-regulated genes.
Pathway enrichment analysis of the Dex regulated clusters showed that ENZ induced and DexUP-regulated genes were related to steroid hormone and corticosteroid response, and prostate cancer pathways (red bar graph in Figure 10). ENZ-induced and DexDN-regulated genes were related to cell death and differentiation pathways (blue bar graph in Figure 10).
Dex-regulated gene expression patterns between 16 and 24-day ENZ-treated samples did not show clear differences (Figure 10).
20
Figure 10. Enzalutamide induced, Dex-upregulated genes are related to steroid response and prostate cancer pathways. Heatmap (left) of Dex regulated gene clusters in control (CTRL), 16-day (ENZ 16d) and 24- day (ENZ 24d) ENZ treated cells and Metascape pathway enrichment analysis (right) of the indicated clusters.
ENZ treatments of 16 and 24-day were combined into one group in the pathway enrichment analyses due to their similarity. The number of Dex-regulated genes is shown in parenthesis next to the heatmap. Heatmap color depicts row Z-scores with the scale shown on the bottom of the graph. Top pathways are displayed as bar graphs for the pathway enrichment analyses. Different colors depict different clusters with the name indicated on top of each graph. Scale depicts -log10 p-value with p-value=0.01 highlighted as dashed line. CTRL=control, DexUP=dexamethasone upregulated genes, DexDN=dexamethasone downregulated genes.
Next, genes whose regulation changes upon ENZ treatment regardless of Dex treatment were assessed. Whereas ENZ+Dex treatment changed the expression of ~600 genes (Figure 9), ENZ by itself affected over 3000 genes (Figure 11). When clustering the genes based on the ENZ treatment, pathway enrichment analysis showed changes especially in cell growth related pathways in both ENZ up- and downregulated genes. As a control for the effect of ENZ on AR target gene regulation, gene set enrichment analysis indicated that hallmark of androgen
21 response was significantly repressed in both 16 and 24-day ENZ-treated samples (Figure 11).
This, combined with the decrease in AR chromatin-binding, indicates that AR’s transcriptional activity is significantly decreased in the experimental model upon ENZ treatment as expected.
Figure 11. Enzalutamide has wide effects on basal gene regulation with significant repression of androgen response. Heatmap (left) of ENZ up- and downregulated gene clusters in control (CTRL), 16-day (ENZ 16d) and 24-day (ENZ 24d) ENZ treated cells, Metascape pathway enrichment analysis (up right) of ENZ up- and downregulated genes, and gene set enrichment analysis (GSEA) (down right) of hallmark of androgen response between control (CTRL) and 16-day (ENZ 16d) or 24-day (ENZ 24d) ENZ treatment. The number of ENZ changed genes is shown in parenthesis next to the heatmap. Heatmap color depicts row Z-scores with the scale shown on the bottom of the graph. Top pathways are displayed as bar graphs for the pathway enrichment analyses.
Different colors depict different clusters with the name indicated on top of each graph. Scale depicts -log10 p- value, with p-value=0.01 highlighted as dashed line. ENZ treatments of 16- and 24-day were combined into one group in the pathway enrichment analyses due to their similarity. CTRL=control, UP=enzalutamide upregulated genes, DN=enzalutamide down-regulated genes, NES=normalized enrichment score. FDR <0.01 in GSEA.
22 Serum/glucocorticoid-induced protein kinase 1 (SGK1) is an example of both AR and GR induced anti-apoptotic gene, whose action is important in prostate cancer development. SGK1 is not Dex-regulated in control conditions, but it becomes Dex induced in ENZ treated cells (Figure 12, Figure 13A). Furthermore, more GR binding is observed near SGK1 locus in ENZ+Dex compared to Dex alone conditions (Figure 12). These GR binding events occur at pre-accessible sites as observed from ATAC-seq data.
Figure 12. SGK1 locus harbored GR-binding sites, and the gene becomes Dex-regulated in enzalutamide- treated cells. An example genome browser track (left) of SGK1 locus (bolded name). Genome browser tracks display GR ChIP-seq (blue) and ATAC-seq (purple) from indicated conditions. The genomic location and 10 kb distance scale bar are shown on top of the graph. Scale of tracks are shown on the left of the graph. SGK1 mRNA levels from RNA-seq data (right) is displayed as TPM values. N=3. All bar graphs represent mean and standard deviation.
Figure 13. SGK1 becomes Dex induced after enzalutamide treatment, whereas Dex induction of PSA (KLK3) is retained. Fold changes of SGK1 (A) and PSA (KLK3) (B) mRNA levels from RT-qPCR at 0, 16-day and 23- day ENZ-treated cells exposed ±Dex. N=3. All bar graphs represent mean and standard deviation.
In comparison to SGK1, prostate cancer biomarker PSA (KLK3) displayed clear GR binding in both ENZ conditions (Figure 14). In addition, KLK3 showed increased Dex-induced expression in all conditions. However, the total KLK3 mRNA levels were clearly lower after ENZ treatment (Figure 14, Figure 13B).
A B
23
Figure 14. KLK3 locus binds GR and is Dex-regulated in enzalutamide-treated and non-treated VCaP cells.
An example genome browser track (left) of KLK3 locus (bolded name). Genome browser tracks display GR ChIP- seq (blue) and ATAC-seq (purple) from indicated conditions. The genomic location and 10 kb distance scale bar are shown on top of the graph. Scale of tracks are shown on the left of the graph. KLK3 mRNA levels from RNA- seq data (right) is displayed as TPM values. N=3. All bar graphs represent mean and standard deviation.
Discussion
Nuclear receptors, such as AR and GR, have been in the forefront of cancer research, since they are known to act as crucial regulators of the disease (Dhiman et al., 2018). Several antiandrogens have been developed for CRPC, one of the latest being enzalutamide (Tran et al., 2009) and darolutamide (Rice et al., 2019). However, the utilization of these treatments are not curative, but only able to provide extended patient survival (Chandrasekar et al., 2015).
Once resistance to these antiandrogens develops, the treatment options for CRPC are significantly less effective.
Upregulated GR signaling has been linked to enzalutamide resistance in prostate cancer (Arora et al., 2013), among other cancers (Dhiman et al., 2018). Since GR is essential for life and therefore cannot be fully antagonized, there is a need to better understand GR’s actions. In this master’s thesis study, the more precise genome-wide mechanisms of GR in VCaP prostate cancer cells treated with enzalutamide were investigated using immunoblotting, RT-qPCR, ChIP-seq, ATAC-seq and RNA-seq. Prolonged enzalutamide treatment was found to increase mRNA and protein levels of GR, and to some extend AR, and influence cell growth related pathways. Chromatin binding of AR was found to be substantially reduced, whereas GR’s binding increased and new binding sites emerged. Most importantly, the results showed that GR replaced the chromatin binding of AR partially and the replacement, based on the data
24 presented here, occurs in already open chromatin sites, which has not been previously indicated in the literature.
Enzalutamide resistance has been found to increase mRNA and protein levels of AR and GR in prostate cancer cell lines and patient samples, though not back to the levels seen in benign tissue (Arora et al., 2013, Chen et al., 2004, Li et al., 2017, Shah et al., 2017). Likewise, increase in AR and GR levels was also found in this study. Thus, the observation of GR level increase upon enzalutamide treatment agrees with published works (Li et al., 2017, Shah et al., 2017). Increased expression to amplify signal output is usually the primary mechanism towards a drug resistance (Chen et al., 2004). Other AR signaling related alterations include AR gene amplification, mutations, intratumoral androgen biosynthesis, AR enhancer amplification and AR variants (Antonarakis et al., 2014, Ku et al., 2019). Although, it should be noted that changed expression does not necessarily directly reflect the transcriptional activity of the receptor within the cells. AR directly represses GR expression via a tissue-specific negative ARE found near GR locus (Shah et al., 2017). When AR is blocked with an antiandrogen like enzalutamide, GR mRNA and protein levels can rise in the absence of AR-mediated repression.
Taken together, elevated expression of GR supports its overtake after prolonged enzalutamide treatment.
More genes were found to be Dex-regulated in enzalutamide-treated VCaP cells compared to non-enzalutamide treated cells, and enzalutamide especially affected cell growth related pathways. As an antiandrogen, enzalutamide represses actions of AR (Tran et al., 2009).
Therefore, the repression of androgen responsive genes shown by GSEA was expected. Due to the upregulation of GR levels, the number of Dex-regulated genes were significantly increased, many of which are presumably previously regulated by DHT/AR. Cholesterol synthesis pathway, taking part in AR- and GR-activating sterol biosynthesis, was seen to be upregulated by Dex in enzalutamide-treated cells. Therefore, inhibition of the pathway could be helpful to overcome enzalutamide resistance (Kong et al., 2018). Dex strongly induced for example known AR target genes SGK1 and PSA through increased binding of GR. Persistent PSA expression in patients responding poorly to enzalutamide could therefore be driven by GR (Arora et al., 2013). In DexUP-ENZ cluster, steroid hormone and prostate cancer pathways were the most enriched, indicating the role of GR in replacing AR in the regulation of these pathways. Taken together, suppression of androgen responsive genes and induction of Dex-
25 regulated cell growth-affecting genes support GR’s overtake after prolonged enzalutamide treatment.
After three-week enzalutamide treatment, chromatin binding of AR substantially reduced.
Moreover, GR’s chromatin binding increased and new binding sites emerged. Enzalutamide was not able to fully inhibit AR binding, though significantly reduce it at all ~50 000 receptor- binding sites. This could partially be related to increased AR protein levels (Chen et al., 2004).
However, the decrease in AR binding is similar with what has been observed in VCaP cells in a previous study utilizing a shorter ENZ exposure (Asangani et al., 2014). At ~70% of the GR- binding sites (ENZ-UP) the receptor binding was significantly increased after enzalutamide treatment, also revealing new binding sites. Previous ChIP-seq analysis have shown that GR can bind to ~52% of all AR-binding sites in enzalutamide-resistant LREX’ cells grown in mice from LNCaP/AR cells (Arora et al., 2013). Our results show partial, ~32% replacement in already AR-expressing VCaP cells. Taken together, the reduced AR chromatin-binding and increased GR binding support its overtake after prolonged enzalutamide treatment.
De novo motif analysis of GR binding sites revealed AR and GR (NR3C), and transcription factor FOXA1 and HOXB13 motifs as the most enriched motifs. FOXA1 displayed the highest percentage of enrichment. There was no apparent change in FOXA1 and HOXB13 motif enrichment or motif score between ENZ-UN and ENZ-UP, but NR3C motifs were more enriched and had a higher motif score in ENZ-UP regions. Therefore, in enzalutamide-treated cells, GR preferably binds more onto ENZ-UP regions, and FOXA1 and HOXB13 potentially keep the chromatin open for the receptor binding (Hankey et al., 2020). Moreover, increased GR protein levels could allow elevated chromatin occupancy. Enzalutamide reduced the overall chromatin accessibility, which could be expected since the binding of AR as one of the most important transcription factors in prostate tissue is inhibited. GR mostly bound to already open chromatin sites. Although, in the presence of enzalutamide, activated GR’s function changed into being able to induce more openness of chromatin. GR is known to be able to bind to open, nucleosome depleted chromatin sites and closed sites enriched with a nucleosome (Swinstead et al., 2018). Though so-called pioneer factors, like FOXA1, are thought to facilitate other transcription factors’ binding, GR can also affect their binding (Swinstead et al., 2016). The FOXA1 motif at AR- and GR-binding sites was also seen as one of the most enriched in a previous study (Arora et al., 2013). Enrichment of FOXA1 indicates its possible important role in facilitating GR overtake. Indeed, FOXA1 binds to GR occupied sites in
26 untreated VCaP cells, though more research is needed to confirm its importance in enzalutamide-resistant prostate cancer cells.
Most importantly, the results showed that GR replaces the chromatin binding of AR partially (~32%), contrary to previously reported ~52% replacement (Arora et al., 2013). In most of the sites where AR binding decreased due to enzalutamide, GR binding increased, indicating a direct replacement. The results also reveal that the replacement occurs only in already open chromatin, since GR mostly binds to open chromatin areas in VCaP cells, whereas AR also binds to closed chromatin sites where GR does not bind to (data not shown). Furthermore, since chromatin is already open, factors such as FOXA1 and HOXB13 are most likely needed for its upkeep.
Taken together, this study showed that GR replaces the chromatin binding of AR at ~32% of sites, contrary to previously reported ~52% replacement. In addition, the replacement occurs in already open chromatin sites, which has not been previously indicated in any studies. This gives a more detailed mechanism behind GR overtake in enzalutamide resistance, which can be further utilized in developing new pharmaceuticals for prostate cancer. More research is needed to determine whether the GR only partially replaces the AR’s functions, whether the replacement is sufficient to lead to the resistance, and what is the role of others chromatin proteins, such as cofactors, in the replacement and the enzalutamide resistance. The results also need to be validated using data derived from patient samples, since several different mechanisms are possible in the development of drug resistance (Zhang, Z. et al., 2020).
Prostate cancer is a complex and heterogenous disease of which treatments affect both the emotional and physical quality of life of the patients and the people around them (Nead et al., 2017, Nelson et al., 2008, Taylor, L. G. et al., 2009). Therefore, there is a need for effective and personalized drugs. In addition to nuclear receptors, many other cofactors, such as histone modifying writers, readers and erasers are misregulated in cancers (Swinstead et al., 2018).
Identifying a tumor-specific GR downstream target, rather than systemic GR antagonism, could make it possible to regulate GR’s actions without its adverse effects (Li et al., 2017). Also, finding a factor that facilitates the GR overtake could be a target rather than GR itself.
Therefore, glucocorticoid administration could be continued for their beneficial effects.
Combining several drugs could also provide synergy (Groner et al., 2016). Currently, glucocorticoid administration after initial positive effects leads to a selection of prostate cancer
27 cells that have upregulated GR. This enhances the cell survival, leading to accelerated disease progression.
Acknowledgements
I wish to thank everyone who has helped me during this master’s thesis project. I thank the laboratory technicians for their practical guidance, and other members of the research group, my family and my friends for their general encouragement. Mostly, I wish to thank my supervisors, postdoctoral researcher Ville Paakinaho and professor Jorma Palvimo, for their superior scientific expertise and enthusiasm.
28
References
Aaron L, Franco OE & Hayward SW (2016) Review of Prostate Anatomy and Embryology and the Etiology of Benign Prostatic Hyperplasia. Urol Clin North Am 43: 279-288
Antonarakis ES, Lu C, Wang H, Luber B, Nakazawa M, Roeser JC, Chen Y, Mohammad TA, Chen Y, Fedor HL, Lotan TL, Zheng Q, De Marzo AM, Isaacs JT, Isaacs WB, Nadal R, Paller CJ, Denmeade SR, Carducci MA, Eisenberger MA et al (2014) AR-V7 and resistance to enzalutamide and abiraterone in prostate cancer. N Engl J Med 371: 1028-1038
Armstrong BK, Barry MJ, Frydenberg M, Gardiner RA, Haines I & Carter SM (2017) PSA testing for men at average risk of prostate cancer. Public Health Res Pract 27:
10.17061/phrp2731721
Arora VK, Schenkein E, Murali R, Subudhi SK, Wongvipat J, Balbas MD, Shah N, Cai L, Efstathiou E, Logothetis C, Zheng D & Sawyers CL (2013) Glucocorticoid receptor confers resistance to antiandrogens by bypassing androgen receptor blockade. Cell 155: 1309-1322 Asangani IA, Dommeti VL, Wang X, Malik R, Cieslik M, Yang R, Escara-Wilke J, Wilder- Romans K, Dhanireddy S, Engelke C, Iyer MK, Jing X, Wu YM, Cao X, Qin ZS, Wang S, Feng FY & Chinnaiyan AM (2014) Therapeutic targeting of BET bromodomain proteins in castration-resistant prostate cancer. Nature 510: 278-282
Attard G, Belldegrun AS & de Bono JS (2005) Selective blockade of androgenic steroid synthesis by novel lyase inhibitors as a therapeutic strategy for treating metastatic prostate cancer. BJU Int 96: 1241-1246
Breslow N, Chan CW, Dhom G, Drury RA, Franks LM, Gellei B, Lee YS, Lundberg S, Sparke B, Sternby NH & Tulinius H (1977) Latent carcinoma of prostate at autopsy in seven areas. The International Agency for Research on Cancer, Lyons, France. Int J Cancer 20:
680-688
Buenrostro JD, Giresi PG, Zaba LC, Chang HY & Greenleaf WJ (2013) Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat Methods 10: 1213-1218
Buenrostro JD, Wu B, Chang HY & Greenleaf WJ (2015) ATAC-seq: A Method for Assaying Chromatin Accessibility Genome-Wide. Curr Protoc Mol Biol 109: 21.29.1-9 Cain DW & Cidlowski JA (2015) Specificity and sensitivity of glucocorticoid signaling in health and disease. Best Pract Res Clin Endocrinol Metab 29: 545-556
29 Center MM, Jemal A, Lortet-Tieulent J, Ward E, Ferlay J, Brawley O & Bray F (2012)
International variation in prostate cancer incidence and mortality rates. Eur Urol 61: 1079- 1092
Chandrasekar T, Yang JC, Gao AC & Evans CP (2015) Mechanisms of resistance in castration-resistant prostate cancer (CRPC). Transl Androl Urol 4: 365-380
Chen CD, Welsbie DS, Tran C, Baek SH, Chen R, Vessella R, Rosenfeld MG & Sawyers CL (2004) Molecular determinants of resistance to antiandrogen therapy. Nat Med 10: 33-39 Cronin KA, Lake AJ, Scott S, Sherman RL, Noone AM, Howlader N, Henley SJ, Anderson RN, Firth AU, Ma J, Kohler BA & Jemal A (2018) Annual Report to the Nation on the Status of Cancer, part I: National cancer statistics. Cancer 124: 2785-2800
Culig Z & Santer FR (2014) Androgen receptor signaling in prostate cancer. Cancer Metastasis Rev 33: 413-427
Dai C, Heemers H & Sharifi N (2017) Androgen Signaling in Prostate Cancer. Cold Spring Harb Perspect Med 7: 10.1101/cshperspect.a030452
Denayer S, Helsen C, Thorrez L, Haelens A & Claessens F (2010) The rules of DNA recognition by the androgen receptor. Mol Endocrinol 24: 898-913
Dhiman VK, Bolt MJ & White KP (2018) Nuclear receptors in cancer - uncovering new and evolving roles through genomic analysis. Nat Rev Genet 19: 160-174
D'Ippolito AM, McDowell IC, Barrera A, Hong LK, Leichter SM, Bartelt LC, Vockley CM, Majoros WH, Safi A, Song L, Gersbach CA, Crawford GE & Reddy TE (2018) Pre-
established Chromatin Interactions Mediate the Genomic Response to Glucocorticoids. Cell Syst 7: 146-160.e7
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M &
Gingeras TR (2013) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29: 15-21 Epstein JI, Zelefsky MJ, Sjoberg DD, Nelson JB, Egevad L, Magi-Galluzzi C, Vickers AJ, Parwani AV, Reuter VE, Fine SW, Eastham JA, Wiklund P, Han M, Reddy CA, Ciezki JP, Nyberg T & Klein EA (2016) A Contemporary Prostate Cancer Grading System: A Validated Alternative to the Gleason Score. Eur Urol 69: 428-435
Escoter-Torres L, Caratti G, Mechtidou A, Tuckermann J, Uhlenhaut NH & Vettorazzi S (2019) Fighting the Fire: Mechanisms of Inflammatory Gene Regulation by the
Glucocorticoid Receptor. Front Immunol 10: 1859
30 Ferraldeschi R, Welti J, Luo J, Attard G & de Bono JS (2015) Targeting the androgen
receptor pathway in castration-resistant prostate cancer: progresses and prospects. Oncogene 34: 1745-1757
Gillatt D (2006) Antiandrogen treatments in locally advanced prostate cancer: are they all the same? J Cancer Res Clin Oncol 132 Suppl 1: S17-26
Gordetsky J & Epstein J (2016) Grading of prostatic adenocarcinoma: current state and prognostic implications. Diagn Pathol 11: 25-016-0478-2
Grasso CS, Wu YM, Robinson DR, Cao X, Dhanasekaran SM, Khan AP, Quist MJ, Jing X, Lonigro RJ, Brenner JC, Asangani IA, Ateeq B, Chun SY, Siddiqui J, Sam L, Anstett M, Mehra R, Prensner JR, Palanisamy N, Ryslik GA et al (2012) The mutational landscape of lethal castration-resistant prostate cancer. Nature 487: 239-243
Griekspoor A, Zwart W, Neefjes J & Michalides R (2007) Visualizing the action of steroid hormone receptors in living cells. Nucl Recept Signal 5: e003
Groner AC, Cato L, de Tribolet-Hardy J, Bernasocchi T, Janouskova H, Melchers D, Houtman R, Cato ACB, Tschopp P, Gu L, Corsinotti A, Zhong Q, Fankhauser C, Fritz C, Poyet C, Wagner U, Guo T, Aebersold R, Garraway LA, Wild PJ et al (2016) TRIM24 Is an Oncogenic Transcriptional Activator in Prostate Cancer. Cancer Cell 29: 846-858
Grozescu T & Popa F (2017) Prostate cancer between prognosis and adequate/proper therapy.
J Med Life 10: 5-12
Hakim O, John S, Ling JQ, Biddie SC, Hoffman AR & Hager GL (2009) Glucocorticoid receptor activation of the Ciz1-Lcn2 locus by long range interactions. J Biol Chem 284:
6048-6052
Hankey W, Chen Z & Wang Q (2020) Shaping Chromatin States in Prostate Cancer by Pioneer Transcription Factors. Cancer Res
Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, Cheng JX, Murre C, Singh H &
Glass CK (2010) Simple combinations of lineage-determining transcription factors prime cis- regulatory elements required for macrophage and B cell identities. Mol Cell 38: 576-589 Holm M, Doveson S, Lindqvist O, Wennman-Larsen A & Fransson P (2018) Quality of life in men with metastatic prostate cancer in their final years before death - a retrospective analysis of prospective data. BMC Palliat Care 17: 126-018-0381-6
Hrdlickova R, Toloue M & Tian B (2017) RNA-Seq methods for transcriptome analysis.
Wiley Interdiscip Rev RNA 8: 10.1002/wrna.1364. Epub 2016 May 19
31 Hu J & Chen Q (2017) The role of glucocorticoid receptor in prostate cancer progression:
from bench to bedside. Int Urol Nephrol 49: 369-380
Hua JT, Ahmed M, Guo H, Zhang Y, Chen S, Soares F, Lu J, Zhou S, Wang M, Li H, Larson NB, McDonnell SK, Patel PS, Liang Y, Yao CQ, van der Kwast T, Lupien M, Feng FY, Zoubeidi A, Tsao MS et al (2018) Risk SNP-Mediated Promoter-Enhancer Switching Drives Prostate Cancer through lncRNA PCAT19. Cell 174: 564-575.e18
Humphrey PA (2017) Histopathology of Prostate Cancer. Cold Spring Harb Perspect Med 7:
10.1101/cshperspect.a030411
Iglesias-Platas I & Monk D (2016) Nongenomic regulation of gene expression. Curr Opin Pediatr 28: 521-528
Isikbay M, Otto K, Kregel S, Kach J, Cai Y, Vander Griend DJ, Conzen SD & Szmulewitz RZ (2014) Glucocorticoid receptor activity contributes to resistance to androgen-targeted therapy in prostate cancer. Horm Cancer 5: 72-89
Jin G, Lu L, Cooney KA, Ray AM, Zuhlke KA, Lange EM, Cannon-Albright LA, Camp NJ, Teerlink CC, Fitzgerald LM, Stanford JL, Wiley KE, Isaacs SD, Walsh PC, Foulkes WD, Giles GG, Hopper JL, Severi G, Eeles R, Easton D et al (2012) Validation of prostate cancer risk-related loci identified from genome-wide association studies using family-based
association analysis: evidence from the International Consortium for Prostate Cancer Genetics (ICPCG). Hum Genet 131: 1095-1103
John S, Johnson TA, Sung MH, Biddie SC, Trump S, Koch-Paiz CA, Davis SR, Walker R, Meltzer PS & Hager GL (2009) Kinetic complexity of the global response to glucocorticoid receptor action. Endocrinology 150: 1766-1774
Karvonen U, Kallio PJ, Janne OA & Palvimo JJ (1997) Interaction of androgen receptors with androgen response element in intact cells. Roles of amino- and carboxyl-terminal regions and the ligand. J Biol Chem 272: 15973-15979
Kassi E & Moutsatsou P (2011) Glucocorticoid receptor signaling and prostate cancer.
Cancer Lett 302: 1-10
Kong Y, Cheng L, Mao F, Zhang Z, Zhang Y, Farah E, Bosler J, Bai Y, Ahmad N, Kuang S, Li L & Liu X (2018) Inhibition of cholesterol biosynthesis overcomes enzalutamide
resistance in castration-resistant prostate cancer (CRPC). J Biol Chem 293: 14328-14341 Krishnan AV, Zhao XY, Swami S, Brive L, Peehl DM, Ely KR & Feldman D (2002) A glucocorticoid-responsive mutant androgen receptor exhibits unique ligand specificity:
therapeutic implications for androgen-independent prostate cancer. Endocrinology 143: 1889- 1900
