The Early Effects of Rapid Androgen Deprivation on Human Prostate Cancer

The androgen receptor (AR) is the dominant growth factor in prostate cancer (PCa). Therefore, understanding how ARs regulate the human transcriptome is of paramount importance. The early effects of castration on human PCa have not previously been studied 27 patients medically castrated with degarelix 7 d before radical prostatectomy. We used mass spectrometry, immunohistochemistry, and gene expression array (validated by reverse transcription-polymerase chain reaction) to compare resected tumour with matched, controlled, untreated PCa tissue. All patients had levels of serum androgen, with reduced levels of intraprostatic androgen at prostatectomy. We observed differential expression of known androgen-regulated genes (TMPRSS2, KLK3, CAMKK2, FKBP5). We identified 749 genes downregulated and 908 genes upregulated following castration. AR regulation of α-methylacyl-CoA racemase expression and three other genes (FAM129A, RAB27A, and KIAA0101) was confirmed. Upregulation of oestrogen receptor 1 (ESR1) expression was observed in malignant epithelia and was associated with differential expression of ESR1-regulated genes and correlated with proliferation (Ki-67 expression). Patient summary This first-in-man study defines the rapid gene expression changes taking place in prostate cancer (PCa) following castration. Expression levels of the genes that the androgen receptor regulates are predictive of treatment outcome. Upregulation of oestrogen receptor 1 is a mechanism by which PCa cells may survive despite castration.

The androgen receptor (AR) is the dominant growth factor in prostate cancer (PCa). Therefore, understanding how ARs regulate the human transcriptome is of paramount importance. The early effects of castration on human PCa have not previously been studied 27 patients medically castrated with degarelix 7 d before radical prostatectomy. We used mass spectrometry, immunohistochemistry, and gene expression array (validated by reverse transcription-polymerase chain reaction) to compare resected tumour with matched, controlled, untreated PCa tissue. All patients had levels of serum androgen, with reduced levels of intraprostatic androgen at prostatectomy. We observed differential expression of known androgen-regulated genes (TMPRSS2, KLK3, CAMKK2, FKBP5). We identified 749 genes downregulated and 908 genes upregulated following castration. AR regulation of a-methylacyl-CoA racemase expression and three other genes (FAM129A, RAB27A, and KIAA0101) was confirmed. Upregulation of oestrogen receptor 1 (ESR1) expression was observed in malignant epithelia and was associated with differential expression of ESR1-regulated genes and correlated with proliferation (Ki-67 expression). Patient summary: This first-in-man study defines the rapid gene expression changes taking place in prostate cancer (PCa) following castration. Expression levels of the genes that the androgen receptor regulates are predictive of treatment outcome. Upregulation of oestrogen receptor 1 is a mechanism by which PCa cells may survive despite castration. Prostate cancer (PCa) is the second most common cause of cancer death in men in the developed world [1]. The androgen receptor (AR) controls PCa growth. Previously, studies of the long-term effects (>3 mo) of medical castration have demonstrated differing transcriptional response following luteinising hormone-releasing hormone (LHRH) analogue and antiandrogens [2] and have implicated Wnt/ b-catenin signalling in castration-resistant PCa [3]. We used a new drug called degarelix to rapidly decrease testosterone levels and inform the early in vivo response of human PCa to castration.

1.
Patients and methods with similar ischemic time before sampling as described [5]. Serum samples were obtained at 8 AM prior to surgery.

Assay of androgen levels in body fluids and tissue
Fresh frozen prostate cores were homogenised, and steroids were extracted and quantified by liquid chromatography coupled with tandem mass spectrometry.

Accession numbers
Study data are deposited in the National Center for Biotechnology Information Gene Expression Omnibus under accession number GSE72920. Expression data were validated for known AR-regulated genes (KLK3, FASN), genes found to be AR regulated in our expression data with known AR binding sites in promoter regions [6] (FAM129A, KIAA0101, RAB27A) and genes of biological importance (AMACR, ESR1, cyclin D1) by reverse transcription-polymerase chain reaction (RT-PCR) and immunohistochemistry (IHC).

Quantitative real-time polymerase chain reaction
We performed RT-PCR in triplicate using SYBR Green and the 7900HT

Results
We saw no differences in the tumour or patient characteristics of the study and control cohorts (Supplementary Table 1). We saw a rapid decrease in the serum testosterone at 7 d after degarelix administration in all treated patients (11.73 AE 5.08 vs 1.19 AE 0.63 nmol/l) ( Supplementary Fig. 1B), paralleled by a decrease in intratumoural androgens (Supplementary Fig. 1C). We identified 749 genes downregulated and 908 genes upregulated in response to castration. Expression levels of known AR-regulated genes-including TMPRSS2, FKBP5, KLK3, and FASN-were among those most strongly affected by treatment (Supplementary Table 2 and 3). We validated differential expression for eight genes by RT-PCR and IHC ( Fig. 1B and 1C). Degarelix treatment decreased nuclear expression of proliferation marker Ki-67 protein and cell cycle progression (expression of nuclear CCND1) (Fig. 1B and 1C). We observed and validated decreased expression of three genes with limited evidence of AR regulation but with AR binding sites in their promoter regions (RAB27A, KIAA0101, and FAM129A) following degarelix treatment (Fig. 1A-1C) 2.1.
ESR1 expression is upregulated in malignant epithelia, is pro-proliferative, and is associated with transcription of known

ESR1 target genes
Degarelix treatment upregulated ESR1 mRNA ( Fig. 2A). Using GSEA, we demonstrated enrichment of genes that had [ ( F i g . _ 1 ) T D $ F I G ] an ESR1 binding motif (Fig. 2B) [7] within the promoter region and genes known to be involved in ESR1 signalling (Fig. 2C) among those genes whose expression was upregulated with degarelix treatment (Fig. 2C).
IHC demonstrated expression of ESR1 by stromal cells but not epithelial cells (benign or malignant) in untreated samples, but 24% of degarelix-treated cancers stained positive for ESR1 in malignant epithelia compared [ ( F i g . _ 2 ) T D $ F I G ] (Bi) Using the Gene Set Enrichment Analysis (GSEA) to examine the distribution of known genes with the ESR1 binding motif, shown within their promoter regions, we found that these genes were enriched among genes that degarelix treatment upregulated when they were ranked by their statistical significance. The ESR1 binding motif analysed was described previously [7]. (Bii) This ESR1 binding motif closely matches the validated, experimentally derived ESR1 binding motif [10] shown here. (C) GSEA demonstrates enrichment of factors known to be involved with ESR1 signalling (from the National Cancer Institute BIOCARTA curated database) among genes differentially expressed in response to degarelix ranked by statistical significance. For this analysis, the degree to which the genes were enriched is defined by the running sum statistic called the normalised enrichment score, which was 2.036 (false discovery rate q-value of 0.027; p = 0.024). with 8% of untreated samples (Fig. 2E). Despite an overall decrease in the expression of the proliferation marker Ki-67 in malignant glands following degarelix treatment ( Fig. 1B and 1C); in those glands with increased nuclear ESR1 staining, proliferation was upregulated (Fig. 2F).

Discussion
In addition to identifying a large number of AR-regulated genes, we have confirmed AR regulation of three genes (RAB27A, FAM129, and KIAA0101) not proven to be AR regulated in vivo as well as AMACR, for which data were conflicting. ESR1 is essential for prostate carcinogenesis and implicated in PCa growth control [8]. Polymorphism of ESR1 is associated with PCa prognosis [9]. We observed rapid upregulation of ESR1 expression with castration ( Fig. 2A). Increased ESR1 mRNA expression following prolonged castration using LHRH analogues has been observed [2,3]. Upregulation of ESR1 expression may represent an intrinsic mechanism by which some malignant prostate epithelial cells proliferate despite castration.