|Home | About | Journals | Submit | Contact Us | Français|
Early studies suggested androgen receptor (AR) splice variants might contribute to the progression of prostate cancer (PCa) into castration resistance. However, the therapeutic strategy to target these AR splice variants still remains unresolved. Through tissue survey of tumors from the same patients before and after castration resistance, we found that the expression of AR3, a major AR splice variant that lacks the AR ligand-binding domain, was substantially increased after castration resistance development. The currently used antiandrogen, Casodex, showed little growth suppression in CWR22Rv1 cells. Importantly, we found that AR degradation enhancer ASC-J9 could degrade both full-length (fAR) and AR3 in CWR22Rv1 cells as well as in C4-2 and C81 cells with addition of AR3. The consequences of such degradation of both fAR and AR3 might then result in the inhibition of AR transcriptional activity and cell growth in vitro. More importantly, suppression of AR3 specifically by short-hairpin AR3 or degradation of AR3 by ASC-J9 resulted in suppression of AR transcriptional activity and cell growth in CWR22Rv1-fARKD (fAR knockdown) cells in which DHT failed to induce, suggesting the importance of targeting AR3. Finally, we demonstrated the in vivo therapeutic effects of ASC-J9 by showing the inhibition of PCa growth using the xenografted model of CWR22Rv1 cells orthotopically implanted into castrated nude mice with undetectable serum testosterone. These results suggested that targeting both fAR- and AR3-mediated PCa growth by ASC-J9 may represent the novel therapeutic approach to suppress castration-resistant PCa. Successful clinical trials targeting both fAR and AR3 may help us to battle castration-resistant PCa in the future.
Prostate cancer (PCa) is currently the second leading cause of death in men in the United States . Androgen deprivation therapy (ADT) has been the standard treatment for patients with advanced PCa since Huggins and Hodges  reported the castration effect on PCa. ADT is initially effective to inhibit the growth of androgen-dependent PCa and suppresses tumor progression in most PCa patients; however, most patients treated with current ADT eventually progress with castration-resistant PCa (CRPC) within 1 to 2 years [3,4]. The mechanisms underlying castration-resistant androgen receptor (AR)-mediated signaling remain unclear, although several possible mechanisms have been proposed [5–11].
One proposed mechanism involves the AR splice variants, especially AR3 (also named as AR-V7) that lacks the portion of the ligand-binding domain (LBD) [8,9], which have been reported to transactivate AR-targeted genes in the absence of androgen [7–10,12]. Interestingly, a recent report from Watson et al.  indicated that such constitutively active AR splice variants (AR-V7) might require full-length AR (fAR). They demonstrated that the growth of LNCaP cells with AR-V7 overexpression was suppressed after MDV3100 (a new antiandrogen) treatment or using small interfering RNA to target fAR. These findings raised an interesting question as to whether those AR splice variants have any translational or clinical value to target.
We report here that AR3 might represent an important target to suppress owing to its roles at selective stage(s) of PCa progression. Furthermore, we demonstrated that AR degradation enhancer, ASC-J9, was able to degrade both fAR and AR3 that resulted in the suppression of AR-targeted genes expression and cell growth in several CRPC cells.
Human PCa cells CWR22Rv1, CWR22Rv1-fARKD (knockdown of fAR ), C4-2, and C81 were used. The antibodies for AR (N-20) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The antibody for AR-V7 was kindly provided by Dr Jun Luo . ASC-J9 (5-hydroxy-1,7-bis(3,4-dimethoxyphenyl)-1,4,6-heptatrien-3-one), also named as dimethylcurcumin, was a gift from AndroScience (San Diego, CA), and bicalutamide (Casodex) was purchased from AstraZeneca (Wilmington, DE). Plasmids containing AR3 complementary DNA and short hairpin RNAs specific for AR3 (shAR3) were kindly provided by Dr Yun Qiu . Human primary prostate tissues were collected from the same patients before ADT and after development to CRPC at Tohoku University Hospital (Japan), Miyagi Cancer Center (Japan), and Chang Gung Memorial Hospital (Taiwan). These patients underwent transrectal prostate needle biopsy or transurethral resection of the prostate. This study has been approved by the ethics committee of the three institutions (Tohoku University Hospital, Miyagi Cancer Center, and Chang Gung Memorial Hospital), and informed consent was obtained from each patient. The patients' characteristics (age, prostate-specific antigen [PSA] level, Gleason score, stage, and time to develop CRPC) and outcomes (sample harvest after progression to CRPC, survival time after ADT, and the current status of alive or death) are summarized in Table W1.
Cells were cultured and treated with or without ASC-J9 for 24 hours in 10% charcoal-dextran-stripped fetal bovine serum (CD-FBS) media. Cell lysates were harvested and subjected to Western blot analysis. Quantitative real-time polymerase chain reaction (qPCR) was performed in triplicate with a Bio-Rad iCycler system (Bio-Rad, Hercules, CA); and messenger RNA (mRNA) levels of PSA, TMPRSS2, FKBP5, and GAPDH were measured. Cells were transiently transfected with mouse mammary tumor virus luciferase reporter (MMTV-Luc) or ARE4-Luc plus pRL-TK as internal control. Luciferase activities were measured using GloMax 20/20 Luminometer (Promega, Madison, WI).
Cells were treated with vehicle, 1 nM dihydrotestosterone (DHT), 5 µM Casodex, and 5 or 10 µM ASC-J9 in 10% CD-FBS medium. The media were replenished every other day, and we followed the standard MTT assay protocol.
The paraffin-embedded tissue sections were stained with anti-ARV7 and counterstained with hematoxylin. These staining signals were manually evaluated by a pathologist (H.M.) blinded to patient identity. Each sample was classified with negative, weak, moderate, and strong expression based on intensity score and the percentage of immunoreactive cells .
Animal procedures were conducted in accordance with the protocol approved by the University of Rochester Committee on Animal Resources. CWR22Rv1 cells (1 x 106 cells per site) were injected into both anterior prostates (orthotopic) of castrated nude mouse after 2 weeks of implantation. The mice were randomly divided into two groups (four mice/eight tumors each group) and either received 75 mg/kg ASC-J9 intraperitoneal injection or vehicle control every other day. After 4 weeks of treatment, all mice were killed to examine the tumor growth. Body weights and mice activity were measured weekly.
All values are presented as mean ± SEM. Statistical analyses and comparisons among groups were performed using Student's t test (2-tailed). P < .05 was considered statistically significant.
We first surveyed AR3 expression in various castration-resistant human PCa cell lines and found that the expression amount of AR3 (80 kDa) was very low in C81 and C4-2 cells but was abundant in CWR22Rv1 cells (Figure 1A). The identification of AR3 splice variant was confirmed with specific AR-V7 antibody (data not shown). To further examine the effect of AR3-mediated PCa cell growth, we then stably infected AR3 into C4-2 and C81 cells (named C4-2/AR3 and C81/AR3), respectively (Figure 1B). We found that C4-2/AR3 cells showed elevated cell growth compared to C4-2/Vector in the androgen-free condition (Figure 1C) as well as in C81/AR3 cells (data not shown). Importantly, Casodex treatment resulted in little suppressive effects on CWR22Rv1 and C4-2/AR3 cells (Figure 1C). These data suggested that AR3 might play important roles for PCa progression at some selective stages and Casodex might have little impact to suppress AR3-mediated PCa cell growth.
To evaluate the changes of AR3 expression correlated with PCa progression, we surveyed six paired human PCa specimens before ADT and after development of castration resistance. Four specimens before ADT displayed benign prostate gland structure and AR3 expression was almost undetectable in benign prostate glands (Figure 2A, left). AR3 expression was weakly positive in PCa cells before ADT (Figure 2A, middle); however, there is a significant increase of AR3 expression after development into castration resistance (Figure 2, A and B, and Table W1), which is in agreement with previous studies [8,9].
Because Casodex failed to suppress AR3-mediated cell growth, we were interested to test the therapeutic effects of ASC-J9. ASC-J9 was identified as a new AR degradation enhancer that could promote AR degradation [15–17] by disrupting the interaction between AR and selective AR coregulators , mainly in prostate stromal and/or luminal epithelial cells, resulting in the suppression of AR transactivation (Chang et al., unpublished data). We also found ASC-J9 had little effect on other steroid receptors, such as glucocorticoid receptor, estrogen receptor α, and retinoid X receptor α . It is noteworthy to reiterate that the action of ASC-J9 is different from those currently available antiandrogens or a recently developed antiandrogen, MDV3100, which prevents androgens from binding to the LBD of AR. More importantly, mice treated with ASC-J9 retained normal sexual function and fertility .
We then determined the effects of ASC-J9 on the expression of fAR and AR3 in C81, C4-2, CWR22Rv1, and CWR22Rv1-fARKD cells. We found that ASC-J9 was able to degrade fAR and AR3 in a dose-dependent manner in various human PCa cells (Figure 3A). The nuclear and cytoplasmic extracts from ASC-J9-treated CWR22Rv1 cells were collected to determine compartmentalized fAR/AR3 expression levels. The results showed that 1 nM DHT could promote fAR nuclear translocation and cotreatment of ASC-J9 could inhibit such nuclear translocation (Figure W1A). AR3 predominantly located in the nuclear compartment and ASC-J9 could also inhibit its amounts (Figure W1A).
In contrast, MDV3100 failed to degrade fAR and AR3 in CWR22Rv1 cells (Figure W2A). Furthermore, we examined the mRNA of fAR and AR3 in CWR22Rv1 cells after ASC-J9 treatment and found that there are few changes compared with vehicle- or DHT-treated group (Figure W2B), suggesting that ASC-J9 mainly targets AR protein degradation.
Next, we examined the effects of ASC-J9 on fAR- and AR3-mediated targeted gene expression. One nanomolar of DHT, the androgen concentration of human prostate tissue after ADT , could induce the expression of AR-targeted genes (PSA, TMPRSS2, and FKBP5) in C81, C4-2, and CWR22Rv1 cells but not in CWR22Rv1-fARKD cells (Figure 3B). Addition of ASC-J9 could then suppress 1 nM DHT-induced AR-targeted gene expression in these three cell lines (Figure 3B). Moreover, ASC-J9 could also effectively suppress AR-targeted genes in CWR22Rv1-fARKD cells (Figure 3B), suggesting that ASC-J9 could also degrade AR3 resulting in suppressed AR3-mediated transcription activity. Interestingly, we also found that ASC-J9 was able to inhibit AR3 target genes Akt1  and c-Myc  in CWR22Rv1 and CWR22Rv1-fARKD cells (Figure W3A). The qPCR primer sequences of AR canonical-targeted genes (PSA, TMPRSS2, and FKBP5), AR3-targeted genes (Akt1 and c-Myc), AR, and AR3 are provided in Figure W4. Using MMTV and ARE4 luciferase assays, AR transactivational activity was suppressed by ASC-J9 treatment in the presence of 1 nM DHT (Figure 3C).
We also assessed the growth effects of C81, C4-2, and CWR22Rv1 cells with 5 or 10 µM ASC-J9 treatments. The results showed that 1 nM DHT can promote cell growth and ASC-J9 significantly suppressed the DHT-induced cell growth in all three PCa cell lines (Figure 3D). We also observed that ASC-J9 displayed better growth suppression than MDV3100 did in CWR22Rv1 cells (Figure W2C). More importantly, treatment with 5 µM ASC-J9 did not inhibit the cell growth of AR-negative cells (PC-3 and DU-145) (Chang et al., unpublished observations, 2012).
In contrast, although 1 nM DHT failed to induce cell growth in CWR22Rv1-fARKD cells, ASC-J9 could still significantly suppress its growth (Figure 3D, right). The potential mechanisms underlying the antiproliferative effects of ASC-J9 were determined using Western blot analysis to detect proliferation marker (proliferating cell nuclear antigen) and cyclin-dependent kinase inhibitors p21 and p27, suggesting that ASC-J9 may use the up-regulation of p21 and p27 to inhibit the cell growth in CWR22Rv1 cells (Figure 3E). The results were further confirmed using CWR22Rv1-fARKD cells with ASC-J9 treatment (Figure W5A), suggesting that targeting fAR either by shRNA or ASC-J9 could upregulate p27 expression leading to suppressed AR-mediated cell growth, yet degradation of AR3 in CWR22Rv1-fARKD cells did not display the significant changes of p27 expression (Figure W5A).
Together, results from Figures 3 and W3A suggested that, although 1 nM DHT failed to induce AR-mediated targeted genes expression and cell growth in CWR22Rv1-fARKD cells, the existence of AR3 and those AR3-targeted genes could still play important roles to maintain cell growth in CWR22Rv1-fARKD cells. Addition of ASC-J9 to degrade AR3, which resulted in the suppression of those AR3-targeted genes and cell growth in CWR22Rv1-fARKD cells, not only pointed out the important contribution of AR3 in PCa progression but also clearly showed that ASC-J9 treatment could suppress both fAR- and AR3-mediated transcriptional activity and growth in PCa cells.
In addition to studying the impacts of ACS-J9 on the endogenous AR3 in CWR22Rv1 cells, we further validated AC-J9 can degrade AR3 and demonstrated this by ectopically delivering AR3 into C81 and C4-2 cells (Figure 4A). We found that the level of those AR-targeted gene expressions is elevated after adding AR3 and that 1 nM DHT treatment further potentiates the AR-targeted gene expressions (Figure 4B). As expected, ASC-J9 could degrade both endogenous fAR and overexpressed AR3 in both C81/AR3 and C4-2/AR3 cells (Figure 4A). This led to the suppression of 1 nM DHT-induced AR-targeted genes expression (Figure 4B) and cell growth in these AR3-overexpressed cells (Figure 4C).
To further confirm whether AR3 contributes to cell growth at castration conditions, the lentivirus carrying shAR3 or scrambled control was transduced into CWR22Rv1 and CWR22Rv1-fARKD cells to specifically suppress AR3 expressions (Figure 5A). As shown in Figures 5B and W6A, shAR3 could suppress AR- and AR3-targeted genes, respectively, as well as cell growth in CWR22Rv1 cells (Figure 5C). Importantly, we found that shAR3 could also suppress AR-targeted genes and cell growth in CWR22Rv1-fARKD cells. These results are similar to ASC-J9 treatment in CWR22Rv1-fARKD cells (Figure 3), suggesting that AR3 itself is important to maintain PCa cells growth in such castration conditions.
Collectively, these results of Figures 3 to to55 suggested that AR3 by itself might contribute to the AR-targeted gene expression and AR mediated PCa cell growth, and ASC-J9 could inhibit both fAR- and AR3-mediated gene expression and cell growth in various CRPC cells.
Based on the previously mentioned results, we further evaluated the effects of ASC-J9 in vivo, especially at the castration-resistant stage. We found that even with complete ADT in mice through surgical castration, the orthotopically implanted CWR22Rv1 cells with vehicle treatment were still able to grow into tumors of significant size (Figure 6A). However, mice treated with ASC-J9 had relative smaller tumors compared to those with vehicle injection (Figure 6A). We also found that the body weight was comparable in mice either from vehicle control or ASC-J9-treated (Figure 6B), suggesting little apparent adverse event of ASC-J9.
We performed immunostaining with anti-AR antibody on tumors and found that the AR intensity was reduced in ASC-J9-treated tumors (Figure 6C). Western blot analysis further confirmed that ASC-J9 could degrade both fAR and AR3 in these xenografted tumors in vivo (Figure 6D). In addition, Ki67 staining showed that ASC-J9-treated tumors had significantly decreased Ki67-positive cells (Figure 6, C and E [upper panels]). Finally, TUNEL assay also showed ASC-J9-treated tumors displayed increased apoptotic cells, which may also contribute to tumor growth suppression of these xenografts (Figure 6, C [arrows] and E [lower panel]).
Together, results from Figure 6A to E demonstrated that ASC-J9 treatment might represent the first therapeutic approach that could further suppress CRPC growth in complete androgen-deprived conditions through degradation of fAR and AR3.
CWR22 xenografted tumor was established from human primary prostate tumor in the patient with bone metastases [21,22]. CWR22Rv1 is a CRPC cell line derived from the CWR22R subline, which was isolated from the recurrent tumor of the androgen-dependent CWR22 cells in castrated mice [23,24]. The origin of CWR22Rv1 cells was different from other CRPC cell lines, such as C4-2 or C81, that were from metastatic lymph nodes of mice xenografted with human PCa cell line, LNCaP [25–28]. CWR22Rv1 expresses an exon 3-duplicated fAR in addition to the multiple spliced forms [7–9,29]. Recent studies  revealed a duplicated AR locus mediated by repetitive elements, which may account for the aberrantly spliced forms of AR. The truncated receptor has been suggested to be derived from two possible pathways, the aberrant splicing as is the case for AR3 and protease cleavage of exon 3-duplicated fAR [29,31]. A different CWR22-derived relapsed cell line, CWR22R1 , also displayed abundant truncated AR . Thus, in the evolution of CRPC such as with CWR22, receptor truncation typified by AR3 seems to be a dominant underlying mechanism. Although the detailed truncation points may be different, most the reported species lack the LBD, and are expected not to respond to androgen and antiandrogens. As such, drugs which target the LBD may not work as effectively. The present study is designed to investigate the efficacy of drugs that target degradation of both fAR and the truncated receptor.
In this study, we used various PCa cell lines, including C81, C4-2, and CWR22Rv1 cells, as well as C81/AR3, C4-2/AR3, and CWR22Rv1-fARKD cells to study AR3 in vitro function. The expression level of fAR and AR3 in these PCa cells was different: in C81 and C4-2 cells, most AR expression was fAR (fAR AR3). In C81/AR3 and C4-2/AR3 cells, fAR expression level was equivalent to AR3 (fAR = AR3). In contrast, AR3 expression level was higher than fAR (fAR < AR3) in CWR22Rv1 cells, but most AR expression was AR3 in CWR22Rv1-fARKD cells (fAR AR3). By characterizing these various PCa cells with differential expression ratios of fAR to AR3, we concluded that AR3 plays a critical role to promote PCa cell growth at some selective stages, which is similar to early findings showing that overexpression of AR3 in LNCaP cells promoted the cell growth . Furthermore, our data showed DHT-induced AR-targeted genes were obviously enhanced with addition of AR3 in C81 and C4-2 cells, suggesting that AR3 might be able to cooperate with fAR to promote DHT-induced AR-targeted genes. This is in agreement with a previous report showing that constitutively active AR splice variants (AR-V7) required fAR to promote AR-targeted genes and cell growth in LNCaP cells in the presence of androgen .
However, our data also showed that cell growth of CWR22Rv1-fARKD was substantially increased compared to that of CWR22Rv1 cells under androgen-free conditions. Specifically knocking down AR3 in CWR22Rv1 and CWR22Rv1-fARKD cells resulted in significant growth inhibition, suggesting that AR3 might have dual roles: AR3 by itself might be able to promote PCa cell growth particularly in the absence of androgen at some selective PCa stages and the other role is that AR3 could cooperate with fAR to modulate AR-targeted genes and cell growth in the presence of androgen at many other PCa stages. These conclusions strengthened our central hypothesis and led us to believe that it is necessary to target both fAR and AR3 to have better therapeutic efficacy, especially at castration-resistant stages when AR3 expression is increased.
We demonstrated that Casodex failed to suppress the cell growth of CWR22Rv1 cells, which is in agreement with an early report showing that flutamide, another antiandrogen, had little suppressive effect on the transactivation of ARv567es (another AR splice variant with deletion of AR exon 5–7 regions) . In this present study, we provided a new therapeutic approach through using ASC-J9, which was able to degrade both fAR and AR3 leading to the suppression of their mediated targeted genes and cell growth in various CRPC cells in vitro and in vivo.
In conclusion, our data suggest that AR3 may have its essential roles to promote PCa growth at selective PCa stages and targeting both fAR- and AR3-mediated cell growth through ASC-J9 may become a new useful therapeutic strategy to target PCa cells, which express AR splice variants such as AR3.
Human PCa cell lines CWR22Rv1 and CWR22Rv1-fARKD were used. The antibodies for AR (N-20), PARP-1, α-tubulin, p27, and GAPDH were purchased from Santa Cruz Biotechnology. MDV3100 was purchased from SelleckChem (Houston, TX) and ASC-J9 (5-hydroxy-1,7-bis(3,4-dimethoxyphenyl)-1,4,6-heptatrien-3-one) is a gift from AndroScience. The NE-PER nuclear and cytoplasmic extraction reagents purchased from Thermo Scientific (Rockford, IL) were used to separate nuclear and cytoplasmic fractions.
Cells were treated with vehicle, 1 nM DHT, 10 µM ASC-J9, and 5 or 10 µM MDV3100 in 10% CD-FBS medium. The media were replenished every other day, and we followed the standard MTT assay protocols.
Cells were cultured and treated with or without 10 µM MDV3100 or 10 µM ASC-J9 for 24 hours in 10% CD-FBS medium. Cell lysates were harvested and subjected to Western blot analysis. Quantitative real-time PCR (qPCR) was performed in triplicate with a Bio-Rad iCycler system, and mRNA levels of fAR, AR3, Akt1, c-Myc, and GAPDH were measured.
The authors thank Karen Wolf for assistance in article preparation and Yun Qiu (University of Maryland) and Jun Luo (Johns Hopkins University) for their gifts of AR3 plasmid and AR-V7 antibody, respectively. The authors also thank Koji Mitsuzuka and Mitsuharu Sasaki (Tohoku University Graduate School of Medicine) for helping collect human PCa samples.
1ASC-J9 was patented by the University of Rochester, the University of North Carolina, and AndroScience Corp and then licensed to AndroScience Corp. Both the University of Rochester and Chawnshang Chang own royalties and equity in AndroScience Corp. This study was supported by the National Institutes of Health (CA122840 and CA127300), George Whipple Professorship Endowment, and National Science Council, Taiwan Department of Health Clinical Trial, and Research Center of Excellence grant DOH99-TD-B-111-004 (China Medical University, Taichung, Taiwan).