Nrdp1-mediated regulation of ErbB3 expression by the androgen receptor in androgen-dependent but not castrate-resistant prostate cancer cells

Abstract

Patients with advanced prostate cancer (PCa) are initially susceptible to androgen withdrawal (AW), but ultimately develop resistance to this therapy (castration-resistant PCa, CRPC). Here, we show that AW can promote CRPC development by increasing the levels of the receptor tyrosine kinase ErbB3 in androgen-dependent PCa, resulting in AW-resistant cell cycle progression and increased androgen receptor (AR) transcriptional activity. CRPC cell lines and human PCa tissue overexpressed ErbB3, whereas downregulation of ErbB3 prevented CRPC cell growth. Investigation of the mechanism by which AW augments ErbB3, using normal prostate-derived pRNS-1-1 cells, and androgen-dependent PCa lines LNCaP, PC346C, and CWR22 mouse xenografts, revealed that the AR suppresses ErbB3 protein levels, whereas AW relieves this suppression, showing for the first time the negative regulation of ErbB3 by AR. We show that AR activation promotes ErbB3 degradation in androgen-dependent cells, and that this effect is mediated by AR-dependent transcriptional upregulation of neuregulin receptor degradation protein-1 (Nrdp1), an E3 ubiquitin ligase that targets ErbB3 for degradation but whose role in PCa has not been previously examined. Therefore, AW decreases Nrdp1 expression, promoting ErbB3 protein accumulation, and leading to AR-independent proliferation. However, in CRPC sublines of LNCaP and CWR22, which strongly overexpress the AR, ErbB3 levels remain elevated due to constitutive suppression of Nrdp1, which prevents AR regulation of Nrdp1. Our observations point to a model of CRPC development in which progression of PCa to castration resistance is associated with the inability of AR to transcriptionally regulate Nrdp1, and predict that inhibition of ErbB3 during AW may impair CRPC development.

Figure 1. ErbB3 levels increase with prostate cancer progression

(A) Tissue microarrays representing (i) benign prostate (n=36) (BENIGN), (ii) high grade PIN (n=21) (HGPIN) and (iii) localized tumors (n=65) (TUMOR) obtained by prostatectomy and (iv) prostatic tissues (n=68) (CRPC/MET) from warm autopsies of men who died of CRPC were immunostained with anti-ErbB3 antibody (brown staining) and counterstained with hematoxylin (blue staining). (Upper): sections stained with anti-ErbB3. Note the strong ErbB3 stain in the tumor containing regions whereas the benign tissue alongside stained only weakly. The antibody used did not stain the nuclei in these tissues (20X magnification). (Lower): box plots representing range of ErbB3 expression in benign prostate, HGPIN, localized tumors and metastatic and localized tissues from warm autopsies of men who died of CRPC. (B) Increased ErbB3 expression in prostate cancer cells compared to lines derived from normal prostate- (upper): Normal prostate derived RWPE-1 cells, androgen-dependent LNCaP and its CRPC subline LNCaP-AI cells compared to stable LNCaP sublines overexpressing ErbB3 (LNCaP-ErbB3-1 and -2), (lower): ErbB3 expression in pRNS-1-1 cells derived from a normal prostate which upon culture lost the expression of the AR, and AR-null PC-3 and DU-145 cells. pRNS-1-1 cells were transfected with vector only, or mutant AR(T877A) or AR(K580R).

Figure 2. Overexpression of ErbB3 leads to increased AR transcriptional activity and cell proliferation

(A) (Left) Increased AR levels in stable LNCaP cell lines overexpressing HER2 or ErbB3 (LNCaP-HER2 and LNCaP-ErbB3) overexpressing plasmids (pcDNA3-HER2 or pcDNA3-ErbB3) in LNCaP cells. Overexpression of HER2 and ErbB3 was confirmed by Western blotting. (Right): Increased AR transcriptional activity in both LNCaP-HER2 and LNCaP-ErbB3 compared to LNCaP. Cells were cultured in CSS-containing medium and AR transcriptional activity measured on a PSA promoter by luciferase assay. LNCaP-AI cells were used as positive control. (B) ErbB3 overexpression stimulates proliferation in LNCaP cells. MTT assay showed increased growth rates of LNCaP-ErbB3-2 cells vs LNCaP in medium containing FBS vs. CSS. Data represents mean ± S.D. of three independent readings. (C) Downregulation of ErbB3 by siRNA suppressed cell growth as estimated through MTT assay in cells cultured in FBS or CSS, in the presence or absence of 10 μM bicalutamide. Data represents mean ± S.D. of three independent readings. (Inset) Immunoblot showing ErbB3 downregulation by siRNA.

Figure 3. Androgen receptor negatively regulates ErbB3 expression in androgen-dependent PCa cells

(A, left panels) Increased ErbB3 expression with decreasing AR levels in LNCaP cells cultured in CSS-containing medium over a period of 5 days. (A, right) Stimulation of ErbB3 expression in LNCaP cells following AR downregulation with two different AR siRNA duplexes (hAR1 and hAR2). (B) Increasing AR levels in parental AR-null pRNS1-1 cells transfected with vector alone or with increasing amounts of pAR0 revealed decreasing ErbB3 expression whereas EGFR levels were not altered. (C) Expression of AR(T877A) but not vector alone in pRNS1-1 cells caused increased expression of ErbB3. (inset) Treatment of pRNS-1-1 cells stably expressing AR(T877A) with 10 nM DHT suppressed ErbB3 levels whereas 5-day culture in CSS-containing medium stimulated ErbB3. (D) Decreased ErbB3 half-life upon expression of wild-type AR. pRNS1-1 cells stably transfected with vector only or wtAR were treated with 100 μg/ml cycloheximide for the indicated times. (Upper) Lysates were blotted with anti-ErbB3 or β-actin, and the bands were quantitated. (Lower) ErbB3 half-life was calculated by fitting the data to a single exponential. Results indicate that in pRNS-1-1 cells expressing vector alone, ErbB3 levels were stabilized whereas in those expressing wtAR, ErbB3 half-life was ~4 hrs.

Figure 4. Negative regulation of ErbB3 by AR is mediated by Nrdp1

(A) (Left) Expression of the ubiquitin E3 ligase Nrdp1 corresponds to a decrease in ErbB3 expression. LNCaP cells were transiently transfected with plasmids encoding Flag vector, Flag-Nrdp1 or pSuper-Nrdp1 shRNA for 48 hours, after which the lysate was collected and blotted with anti-Nrdp1, anti-ErbB3 or, α-Tubulin as control. (Middle) Nrdp1 levels correspond to AR expression. Cultured of LNCaP cells over 2-days in CSS show a decrease in Nrdp1 with lower AR. (Right): Expression of AR in pRNS-1-1 cells stimulate Nrdp1. Parental pRNS1-1 cells were transiently transfected with vector alone or pAR0. Cell lysate was collected after 48 hours and blotted with anti-Nrdp1 and anti-AR. (B) The AR positively regulates Nrdp1 and negatively regulates ErbB3 levels in androgen-dependent PC-346C cells. Note the expression of a lower molecular weight AR product in these cells. (C) Decreased half-life of ErbB3 induced by AR expression is mediated by Nrdp1. pRNS1-1 cells stably transfected with wtAR were further transfected with pSuper-scramble or pSuper-Nrdp1 RNAI plasmids for 24 hours. ErbB3 half-life in pRNS-1-1/wtAR cells expressing control RNAi remained <4 hrs whereas in the absence of Nrdp1 expression, the half-life of ErbB3 increased to >24 hrs.

Figure 5. AR regulation of ErbB3 and Nrdp1 is not seen in CRPC cells with high AR transcriptional activity

(A) Expression of EGFR, HER2, ErbB3 and AR were compared in androgen-dependent LNCaP cells, and its CRPC sublines LNCaP-AI and C4-2 cells by Western blotting, and confirms overexpression of ErbB3 and AR in the CRPC lines. (B) Stimulation of AR expression with increasing DHT increases Nrdp1 and decreases ErbB3 in LNCaP but not LNCaP-AI. (upper), LNCaP and LNCaP AI cells were cultured in CSS-containing medium overnight, then the indicated concentrations of DHT were added for another 48 hours. Cell lysates were immunoblotted with anti-ErbB3, anti-Nrdp1, anti-AR and anti-tubulin antibodies. (Middle) AR increases Nrdp1 transcription in LNCaP but not LNCaP AI cells as determined from mRNA levels by RT-PCR after treatment with increasing doses of DHT. (Lower) This is despite increased AR transcriptional activity on a human PSA promoter as determined by luciferase assay upon DHT stimulation in LNCaP-AI cells. (C) AR fails to regulate Nrdp1 and ErbB3 levels in CRPC cells due to decreased Nrdp1 levels caused by ErbB3 overexpression. (Upper) Suppression of Nrdp1 expression in LNCaP AI, C4-2 and LNCaP-ErbB3 cells compared to LNCaP. Cells were cultured in FBS medium up to 75% confluence and collected for western blotting. (lower) DHT-induced suppression of ErbB3 levels are seen in LNCaP cells is mediated by high levels of Nrdp1. This effect was not seen when LNCaP cells were transfected with Nrdp1 siRNA in CSS; after 24 hours, and the indicated concentration of DHT was added for another 48 hours. Cell lysate was collected and blotted with anti-ErbB3 and anti-Nrdp1.

Figure 6. Effect of AR activity on ErbB3 and Nrdp1 expression seen in castration sensitive CWR22 tumors but not in CRPC cell lines derived from recurrent tumors from castrated mice

(A, B) CWR22 xenograft tumors were established by subcutaneous injections of cell suspensions (2.5 × 10⁶ cells in Matrigel; 1:1, v/v) bilaterally into the flanks of 4–5-week-old nu/nu athymic male mice (n=10) previously implanted with sustained-release testosterone pellets. When palpable tumors were observed, animals were treated with (i) vehicle (peanut oil) or (ii) the AR antagonist bicalutamide (n=5/group). After two weeks on this treatment, the mice were euthanized, tumors were harvested and divided into sections that were paraffin-embedded snap-frozen in liquid nitrogen. Paraffin-embedded tumors were analyzed by immunohistochemistry for ErbB3 whereas frozen tumors were excised, lysed and protein levels were determined by Western blotting. (A) Western blotting revealed that ErbB3 levels increased while Nrdp1 levels decreased with bicalutamide treatment. (B) Immunohistochemistry demonstrating that ErbB3 levels increased in tumors extracted from the bicalutamide-fed mice. Brown staining denotes ErbB3 while blue represents counterstaining with hematoxylin. Negative control showed no ErbB3 staining, the xenograft from the vehicle-fed mouse shown showed lower ErbB3 expression (scored +1) compared to the bicalutamide-treated one (scored +3). (C) In contrast, CWR22Rv1, a CRPC cell line derived from a CWR22 relapsed tumor grown in a castrated mouse, failed to respond to bicalutamide and in these cells, Nrdp1 and ErbB3 were not androgen regulated. (D) Scheme describing androgen regulation of Nrdp1 transcription in androgen-dependent but not androgen-independent cells.