Nuclear receptor ERRα contributes to castration-resistant growth of prostate cancer via its regulation of intratumoral androgen biosynthesis

Abstract

Enhanced intratumoral androgen biosynthesis and persistent androgen receptor (AR) signaling are key factors responsible for the relapse growth of castration-resistant prostate cancer (CRPC). Residual intraprostatic androgens can be produced by de novo synthesis of androgens from cholesterol or conversion from adrenal androgens by steroidogenic enzymes expressed in prostate cancer cells via different steroidogenic pathways. However, the dysregulation of androgen biosynthetic enzymes in CRPC still remains poorly understood. This study aims to elucidate the role of the nuclear receptor, estrogen-related receptor alpha (ERRα, ESRRA), in the promotion of androgen biosynthesis in CRPC growth. Methods: ERRα expression in CRPC patients was analyzed using Gene Expression Omnibus (GEO) datasets and validated in established CRPC xenograft model. The roles of ERRα in the promotion of castration-resistant growth were elucidated by overexpression and knockdown studies and the intratumoral androgen levels were measured by UPLC-MS/MS. The effect of suppression of ERRα activity in the potentiation of sensitivity to androgen-deprivation was determined using an ERRα inverse agonist. Results: ERRα exhibited an increased expression in metastatic CRPC and CRPC xenograft model, could act to promote castration-resistant growth via direct transactivation of two key androgen synthesis enzymes CYP11A1 and AKR1C3, and hence enhance intraprostatic production of dihydrotestosterone (DHT) and activation of AR signaling in prostate cancer cells. Notably, inhibition of ERRα activity by an inverse agonist XCT790 could reduce the DHT production and suppress AR signaling in prostate cancer cells. Conclusion: Our study reveals a new role of ERRα in the intratumoral androgen biosynthesis in CRPC via its transcriptional control of steroidogenic enzymes, and also provides a novel insight that targeting ERRα could be a potential androgen-deprivation strategy for the management of CRPC.

Keywords: AKR1C3; CYP11A1; ERRα; castration resistance; intratumoral steroidogenesis; nuclear receptor; prostate cancer.

Figure 1

Up-regulation of ERRα is associated with CRPC. (A) ERRα immunohistochemistry. Representative micrographs of ERRα-immunostained tissue microarray spots of BPH and malignant lesions. A significant increase in malignant cells with positive and intense nuclear ERRα immunosignals was detected in adenocarcinoma lesions with higher Gleason scores. Magnification, ×40; scale bars, 250 μm. Insets show the enclosed areas at higher magnification. Magnification, ×200; scale bars, 50 μm. (B) ERRα immunoreactivity score analysis (ERRα‐IRS) performed on BPH and neoplastic prostatic tissues. Adenocarcinomas with higher Gleason scores (GS ≥ 7) showed significantly higher ERRα expression than BPH tissues. (C) Kaplan-Meier analysis of Grasso et al. 2012 study cohort (GSE35988) of CRPC revealed that higher ERRα mRNA expression was positively correlated with shortened overall survival of prostate cancer patients relapsed from hormone therapy.

Figure 2

ERRα enhances in vitro growth resistance to androgen-deprivation and antiandrogen in prostate cancer cells. (A-D) ERRα-overexpression analysis. (A) Immunoblot validation of LNCaP-ERRα transduced clones. (B-D) In vitro growth responses of LNCaP-ERRα cells under androgen-deprivation culture condition (CS-FBS) and Enzalutamide treatment assayed by MTT. (B) Both LNCaP-pBABE and LNCaP-ERRα transduced cells grew at comparable rates in culture with normal FBS until Day-7. (C and D) However, when being cultured with CS-FBS or Enzalutamide, LNCaP-ERRα cells grew at normal rates, in sharp contrast to LNCaP-pBABE cells that did not grow. (E-H) ERRα-knockdown analysis. (E) Immunoblot validation of shERRα-transduced clones. (F-H) In vitro growth responses of LNCaP-shERRα cells toward cultures with CS-FBS and Enzalutamide as assayed by MTT. The LNCaP-shERRα cells grew at slower rate than the LNCaP-shScramble cells and did not grow in culture conditions with CS-FBS and Enzalutamide. (I-J) XCT790 treatment of VCaP cells. (I) Immunoblot analysis showed that XCT790 treatment could significantly reduce or abolish the protein level of ERRα in VCaP cells. (J-L) XCT790 treatment could significantly suppress the in vitro growth of VCaP cells cultured with CS-FBS or Enzalutamide. *, P < 0.05; **, P < 0.01 versus vector control LNCaP-pBABE, LNCaP-shScramble cells or vehicle treatment.

Figure 3

ERRα overexpression promotes in vivo castration-resistant growth capacity and enhances intratumoral androgen levels in LNCaP-ERRα-derived xenograft tumors. (A) Images show the representative castrated SCID mice bearing the xenograft tumors formed by LNCaP-ERRα and LNCaP-pBABE cells and the dissected xenograft tumors formed by the corresponding cells after 1-month growth in castrated host. (B) Growth curve shows the growth responses of LNCaP-ERRα and LNCaP-vector clones in intact host mice for initial 9 weeks followed by 4-week growth in same hosts after castration. LNCaP-ERRα clones formed larger tumors than vector clones in intact mice before castration and continued to grow aggressively in castrated hosts, as compared to tumors formed by LNCaP-pBABE clones that became atrophied after host castration. (C) Measurement of wet weights of xenograft tumors. LNCaP-ERRα clones formed tumors heavier than their vector counterpart in castrated hosts. (D) Measurement of androgens in xenograft tumors by LC-MS/MS. Significant higher levels of testosterone and DHT were detected in LNCaP-ERRα-derived tumors as compared to its vector counterpart. Data are represented as mean ± SD (n = 5) and analyzed by Students' t-test. *, P < 0.05; **, P < 0.01 versus vector control LNCaP-pBABE cells.

Figure 4

ERRα-overexpression enhances whereas its knockdown or suppression attenuates the expressions of some key steroidogenic enzymes in prostate cancer cells. (A-C) qRT-PCR analysis. (A) LNCaP-ERRα clones expressed higher transcript levels of CYP11A1 and AKR1C3 as compared to empty vector LNCaP-pBABE clones. (B) ERRα-knockdown in LNCaP-shERRα cells. ERRα-knockdown could significantly suppress the mRNA levels of CYP11A1 and AKR1C3 in LNCaP-shERRα clones as compared to their shScramble clones. (C) XCT790 treatment. XCT790 treatment could significantly suppress the mRNA levels of CYP11A1 and AKT1C3 in VCaP cells. *, P < 0.05; **, P < 0.01 versus LNCaP-pBABE cells, LNCaP-shScramble cells or vehicle treatment. (D) Immunoblot analysis of AKR1C3 and CYP11A1 in LNCaP-ERRα cells, LNCaP-shERRα cells and XCT790-treated VCaP cells. ERRα-overexpression enhanced AKR1C3 and CYP11A1 expression in LNCaP-ERRα cells whereas ERRα-knockdown and XCT790 treatment suppressed AKR1C3 and CYP11A1 expression in LNCaP and VCaP cells, respectively. (E) Linear regression analysis between ERRα and AKR1C3 immunoreactivity scores (IRS). Results showed ERRα and AKR1C3 manifested a positive correlation in their IRS.

Figure 5

Direct transactivation of AKR1C3 and CYP11A1 genes by ERRα in prostate cancer cells. (A and B) Schematic diagrams depict the putative ERRα-binding sites (ERREs), as predicted by the online program MatInspector (https://www.genomatrix.de), located in the (A) AKR1C3 (P1-P5) and (B) CYP11A1 (P1-P7) gene promoter/regulatory regions. (C) ChIP-PCR assay of AKR1C3 and CYP11A1 gene promoters performed in VCaP prostate cancer cells. Results validated that two ERRE sites (P1 and P3) located respectively at 1.8 kb and 6.8 kb upstream of the transcription start site of AKR1C3 promoter (upper graph), and one ERRE site (P7) located at 4.3 kb upstream of the CYP11A1 promoter (lower graph), were enriched of ERRα. **, P < 0.01 versus non-immune IgG-treated DNA. (D-F) Luciferase reporter assay of AKR1C3-Luc and CYP11A1-Luc reporters performed in ERRα-transfected HEK293 cells. The AKR1C3-Luc and CYP11A1-Luc reporter constructs, containing inserts of ERRα promoter/regulatory regions, could be significantly transactivated by the transfected ERRα. Deletion of identified ERREs in the reporters (P1 and P3 in AKR1C3-Luc; P7 in CYP11A1-Luc) abolished or significantly reduced the ERRα-induced transactivation. (G) Schematic diagram shows the functional domains of wild-type ERRα protein and two of its truncated mutants generated. (H-J) Luciferase reporter assays of AKR1C3-Luc and CYP11A1-Luc reporters performed in HEK293 cells. The AKR1C3-Luc and CYP11A1-Luc reporters could be dose-dependently transactivated by ERRα and further potentiated by co-transfection with co-regulator PGC-1α (2×9), but not by ERRα-ΔAF2 and ERRα-ΔZF2 truncated mutants. *, P < 0.05; **, P < 0.01 versus empty vector pcDNA3.1. Data are presented as mean ± SD (n = 3) and analyzed by Students' t-test.

Figure 6

ERRα-overexpression can promote while its suppression can attenuate DHT production in prostate cancer cells. (A) Diagram shows the classical (canonical or front-door) and alternative (back-door) pathways of androgen biosynthesis. Androgens are synthesized from cholesterol via multiple enzymatic steps. CYP11A1 is responsible for converting cholesterol to pregnenolone by side chain cleavage of cholesterol. Pregnenolone is then converted to dehydroepiandrosterone (DHEA) and androstenedione by CYP17A1. The classical pathway for testosterone biosynthesis involves conversion of major adrenal DHEA and androstenediol to testosterone in the testis, followed by 5α-reduction of testosterone to dihydrotestosterone (DHT) by 5α-reductases (SRD5As). On the other hand, DHT biosynthesis via 5α-reduction of upstream steroids bypassing testosterone can be achieved by two other back-door pathways. In the primary backdoor pathway, 17OH-progesterone is 5α- and 3α-reduced by SRD5As and AKR1C2 before the 17,20-lyase reaction of CYP17A1, yielding the androsterone and then to androstanediol by HSD17Bs and AKR1C3. In the secondary backdoor pathway, androstenedione is converted to 5α-androstenedione (5α-Adione) by SRD5As and then to DHT by HSD17Bs and AKR1C3. Through these backdoor pathways, DHT is synthesized without using testosterone as intermediate. (B and C) UPLC-MS/MS measurement of DHT in LNCaP-ERRα cells and XCT790-treated VCaP cells. Results showed that upon supplement with 5α-Adione, LNCaP-ERRα cells contained significant higher level of DHT than the empty vector LNCaP-pBABE cells. On the other hand, suppression of ERRα activity by XCT790 could significantly reduce the DHT level in 5α-Adione-supplemented VCaP cells. (D and E) UPLC-MS/MS measurement of DHT in LNCaP-ERRα cells upon XCT790 treatment and shRNA-mediated knockdown of AKR1C3. Results showed that ERRα overexpression-induced increase of DHT biosynthesis could be abolished by either XCT790 treatment or AKR1C3 knockdown. *, P < 0.05; **, P < 0.01 versus empty vector pBABE or vehicle.

Figure 7

ERRα functions to activate AR signaling in prostate cancer cells. Cultured cells were first serum-starved in phenol red-free medium with 1% CS-FBS for 48 hours and then treated with androgen metabolite 5α-Adione (100 nmol/L) or vehicle control (0.1% ethanol) for 24 hours before mRNA or protein analyses. (A) Immunoblot analysis of AR and PSA in LNCaP-pBABE and LNCaP-ERRα transduced cells. Results showed that supplement with 5α-Adione could enhance the nuclear AR (with no change in total AR level) and PSA expression levels in both LNCaP-pBABE and LNCaP-ERRα cells, with significant higher levels in LNCaP-ERRα cells. (B) qRT-PCR analysis of ERRα and KLK3 (PSA) expression in 5α-Adione-supplemented LNCaP-pBABE and LNCaP-ERRα transduced cells. Supplement with 5α-Adione induced no change in mRNA levels of ERRα in both LNCaP-pBABE and LNCaP-ERRα cells. However, 5α-Adione supplement could significantly increase PSA mRNA levels in both LNCaP-pBABE and LNCaP-ERRα cells, with significant higher levels in LNCaP-ERRα cells. Supplement with R1881 induced increase of PSA mRNA levels in both LNCaP-pBABE and LNCaP-ERRα cells at same levels. (C) Immunoblot analysis of AR and PSA in VCaP cells treated with 5α-Adione and XCT790. Results showed that suppression of ERRα activity by XCT790 could significantly reduce or abolish the protein levels of nuclear AR, PSA and ERRα in VCaP cells with or without 5α-Adione supplement. (D) qRT-PCR analysis of ERRα and PSA expression in VCaP cells treated with 5α-Adione and XCT790. Results showed that XCT790 treatment did not affect the mRNA levels of ERRα, but significantly suppressed the mRNA levels of KLK3 in VCaP cells supplemented with 5α-Adione. Supplement with R1881 induced increase of PSA mRNA levels in VCaP cells upon treatment with XCT790 or vehicle at same levels. **, P < 0.01 versus empty vector pBABE transduced cells or vehicle.

Figure 8

ERRα inverse agonist C29 suppresses castration-resistant growth of prostate cancer in vivo. (A) Growth curve of VCaP-CRPC xenograft tumors upon C29 or vehicle treatment. Once the castration-relapsed VCaP-CRPC tumors regrew to sizes as that in pre-castration at 6-7 week post-castration, tumor-bearing mice were randomly assigned to intraperitoneal injections of C29 or vehicle for additional 3 weeks. Results showed that C29 could significantly retard the castration-relapsed growth of VCaP-CRPC tumors as compared to vehicle. (B) Images show the representative dissected VCaP-CRPC xenograft tumors upon 3-week treatment with C29 or vehicle in castrated host. (C) Measurement of DHT in VCaP-CRPC xenograft tumors by LC-MS/MS. Significant reduction of DHT levels was detected in tumors upon C29 treatment as compared vehicle. (D) Immunoblot analysis. VCaP-CRPC tumors upon C29 treatment expressed lower protein levels of ERRα and AKR1C3 as compared to vehicle. **, P < 0.01 versus vehicle. (E) Schematic diagram depicts the hypothetical role of ERRα in the promotion of intratumoral androgen biosynthesis and reactivation of AR signaling in CRPC via its direct transactivation of some key steroidogenic enzyme genes.