Regulation of carcinogenesis and mediation through Wnt/β-catenin signaling by 3,3′-diindolylmethane in an enzalutamide-resistant prostate cancer cell line

Enzalutamide (ENZ) is an important drug used to treat castration-resistant prostate cancer (CRPC), which inhibits androgen receptor (AR) signaling. Previous study showed that 3,3′-diindolylmethane (DIM) is an AR antagonist that also inhibits Wnt signaling and epithelial-mesenchymal transition (EMT). To investigate whether combined treatment with ENZ and DIM can overcome ENZ resistance by regulating Wnt signaling to inhibit AR signaling and EMT in ENZ-resistant prostate cancer cells, 22Rv1 cells were cultured in normal medium and treated with ENZ, DIM, and DIM with ENZ. Exposure of ENZ-resistant cells to both DIM and ENZ significantly inhibited cell proliferation without cytotoxicity and invasion in comparison with the control. DIM significantly increased the E-cadherin expression and inhibited the expressions of Vimentin and Fibronectin, subsequently inhibiting EMT. Co-treatment with ENZ and DIM significantly increased the expressions of GSK3β and APC and decreased the β-catenin protein expression, causing inhibition of Wnt signaling and AR expression, it also significantly decreased the AR-v7 expression and down-regulated AR signaling. Via suppression of Wnt and AR signaling, co-treatment increased the E-cadherin and decreased the Vimentin and Fibronectin RNA and protein expressions, then inhibited EMT. Co-treatment with DIM and ENZ regulated Wnt signaling to reduce not only the AR expression, but also the AR-v7 expression, indicating suppression of EMT that inhibits cancer cell proliferation, invasion and migration to ameliorate ENZ resistance.


DIM and ENZ treatment suppress cell proliferation and colony formation of PCa cells.
Exposure to a low dose of ENZ for 48 and 72 h had no suppressive effect on 22Rv1 cells (data not shown). It was confirmed that 22Rv1 cells exhibited ENZ resistance. Treatment of cells with DIM alone for 48 and 72 h caused significant dose-and time-dependent inhibition of cell proliferation, with 25% suppression following 72 h of exposure to 30 μM DIM. Therefore, in order to investigate the adjuvant effects of DIM on ENZ resistance, 22Rv1 cells were treated with DIM and ENZ simultaneously, and the results showed that treatment of 22Rv1 cells with 30 μM DIM alone for 48 and 72 h significantly decreased the cell viability ( Fig. 1; p < 0.05 vs. control), and the effect of co-treatment with DIM and 10 μM ENZ was similar to that of treatment with 40 μM ENZ. Co-exposure to DIM and ENZ resulted in significant inhibition of cell proliferation ( Fig. 1; p < 0.05 vs. DIM or ENZ alone).
In 22Rv1 cells, treatment with DIM alone significantly inhibited the clonogenic ability ( Fig. 2; p < 0.05 vs. control), while there was no suppression when cells were treated with 10 μM ENZ. In cells treated with a high concentration of ENZ, colony formation was significantly inhibited ( Fig. 2; p < 0.05 vs. control), and the inhibition was of the same magnitude as that caused by treatment with 30 μM DIM. Combined DIM and ENZ treatment significantly increased the inhibition of colony formation ( Fig. 2; p < 0.05 vs. control and 10 μM ENZ). The results of the present study indicated that ENZ and DIM may have a synergistic effect in suppressing cell proliferation and colony formation in PCa cells.

DIM potentiates the inhibition effect of ENZ on PCa cell invasion.
In order to examine whether combined treatment with DIM and ENZ resulted in stronger inhibition of invasion in PCa cells, 22Rv1 cells were treated with DIM and ENZ alone and together for 24 h. The results illustrated that both DIM and ENZ significantly inhibited cell invasion ( Fig. 3B; p < 0.05 vs. control), but DIM had a better inhibition effect ( Fig. 3B; p < 0.05 vs. 10 and 40 μM ENZ). In addition, exposure to DIM and ENZ in combination significantly increased the efficacy of ENZ in suppressing cell invasion ( Fig. 3B; p < 0.05 vs. 10 and 40 μM ENZ). The results of the present study indicated that DIM may potentiate the efficacy of ENZ with regards to invasion of 22Rv1 cells. DIM and ENZ downregulate AR and AR-v7 expressions. Aberrant AR signaling is considered a critical mechanism in CRPC. Moreover, a mechanism of cross-talk between Wnt and AR signaling has been identified. Therefore, to elucidate whether AR signaling is inhibited when Wnt signaling is inhibited, western blot analysis was employed to detect the expression of AR. After 48 h of treatment, 40 μM ENZ significantly increased the expression of AR ( Fig. 7A; p < 0.05 vs. control); however, combined treatment with DIM and 40 μM ENZ significantly decreased the level of AR ( Fig. 7A; p < 0.05 vs. 40 μM ENZ). After 72 h of treatment, exposure to DIM and 40 μM ENZ significantly downregulated the AR expression ( Fig. 7B; p < 0.05 vs. control).
Besides AR overexpression, a high AR-v7 expression has been found in ENZ-resistant PCa, and AR-v7 may also activate AR signaling. Therefore, to examine the effect of DIM on AR-v7 expression, western blot analysis was used to detect the protein expression of AR-v7 in 22Rv1 cells. After 48 h of treatment, exposure to ENZ alone or in combination with DIM significantly decreased the AR-v7 expression ( Fig. 7C; p < 0.05 vs. control). Compared with treatment of 22Rv1 cells with 10 μM ENZ alone, co-treatment with ENZ and DIM significantly downregulated the level of AR-v7 ( Fig. 7C; p < 0.05 vs. 10 μM ENZ). After 72 h of treatment, 40 μM ENZ and combined treatment with DIM and ENZ significantly decreased the AR-v7 expression ( Fig. 7D; p < 0.05 vs. control). The results indicated that co-treatment with DIM and ENZ may inhibit AR signaling by downregulating the expressions of AR and AR-v7.  www.nature.com/scientificreports/ After treatment for 72 h, exposure to DIM alone and combined treatment with DIM and ENZ significantly upregulated the expression of CDH1, but cells treated with ENZ alone exhibited a significantly decreased level of CDH1 ( Fig. 8B; p < 0.05 vs. control). Exposure to DIM or ENZ alone, and co-treatment with DIM and ENZ, significantly inhibited the FN1 expression, but only ENZ treatment and combined treatment decreased the VIM expression ( Fig. 8B; p < 0.05 vs. control). Compared with treatment of 22Rv1 cells with 10 μM ENZ alone, cotreatment with ENZ and DIM significantly upregulated the level of CDH1 and downregulated the levels of FN1 and VIM ( Fig. 8B; p < 0.05 vs. 10 μM ENZ).
Furthermore, western blot analysis was employed to detect the protein expressions of E-cadherin (E-cad), Fibronectin (FIN) and Vimentin (VIM). After treatment for 48 h (Fig. 9A), exposure to DIM or 10 μM ENZ alone, and combined treatment with DIM and 40 μM ENZ, significantly upregulated the expression of E-cad ( After treatment for 72 h (Fig. 10A), exposure to DIM and 10 μM ENZ alone and combined treatment with DIM and ENZ significantly increased the expression of E-cad ( Fig. 10C; p < 0.05 vs. control). Exposure to DIM and 40 μM ENZ alone, and co-treatment with DIM and ENZ, significantly inhibited the FIN expression, but only combined treatment decreased the VIM expression (Fig. 10B,D; p < 0.05 vs. control). Compared with treatment of 22Rv1 cells with 10 μM ENZ alone, co-treatment with ENZ and DIM significantly decreased the expressions  The results showed that ENZ and DIM treatment may not enhance cell cytotoxicity. According to these results, we attempted to outline a possible model to elucidate the role of DIM in an ENZresistant prostate cancer cell line (Fig. 13). The cell treatment with 10 μM ENZ is as a model of ENZ resistant control group. Our results showed that co-treatment with DIM and ENZ resulted in the strongest inhibition of cell proliferation, colony formation, invasion and migration. Furthermore, co-treatment with DIM and ENZ significantly increased the AR and AR-v7 protein expressions to inhibit AR signaling (vs. control and 10 μM ENZ). In addition, treatment with DIM and ENZ for 72 h significantly increased the expressions of APC and GSK3β, and decreased the expression of β-catenin (vs. control), which indicated that co-treatment suppressed Wnt signaling. Both signaling pathways are related to the progression of EMT, and the results showed that cotreatment significantly increased the expression of E-cadherin and decreased the expressions of Fibronectin and Vimentin to suppress EMT (vs. control and 10 μM ENZ).

Discussion
The 22Rv1 cell line is a CRPC cell line that expresses AR and has the highest expression of AR-v7 as compared with other CRPC cell lines 37,38 . In comparison with these other cell lines, 22Rv1 cells have higher levels of expression of AR and AR-v7, the other cells barely expressing AR-v7 at all. In the present study, we aimed to investigate whether inhibition of AR and AR-v7 expressions by DIM is a main cause of enzalutamide resistance. Therefore, we employed 22Rv1 cells as the main model in this study.
The results of previous studies and the present study showed that treatment of 22Rv1 cells with ENZ (0-40 μM) did not significantly inhibit cell proliferation, proving that 22Rv1 is ENZ-resistant 39,40 . Therefore, in this study, we used 22Rv1 cells to ascertain whether DIM treatment could improve ENZ resistance. According to previous studies, DIM suppresses cell proliferation in various cancers, including prostate cancer, breast cancer and ovarian cancer 41,42 . Exposure of LNCaP, DU145, PC-3, C4-2B and 22Rv1 PCa cell lines to DIM or BR-DIM was found to reduce the cell viability [43][44][45][46] . Moreover, BR-DIM inhibited PCa cell invasion and proliferation by www.nature.com/scientificreports/ decreasing platelet-derived growth factor expression and activity 47 . In addition, Admad et al. 48 showed that treatment of PC-3 cells with 25 μM BR-DIM downregulated uPA by decreasing the expressions of VEGF and MMP-9 to inhibit cell proliferation, invasion and migration. The results of this study showed that DIM had significant anticancer effects on 22Rv1 cells, and combined treatment with DIM and ENZ had stronger effects in terms of inhibiting cancer cell proliferation, colony formation, invasion and migration. DIM re-sensitizes ENZ-resistant cells to ENZ treatment, proving that it may be able to be developed as an adjuvant therapy for PCa. In the clinical treatment of PCa, repressing AR signaling is a critical strategy 6,7 . ENZ is one of the newer and potentially more effective AR-targeting agents 49 ; however, the response to ENZ is not permanent, and almost all patients eventually develop resistance to ENZ, demonstrating that targeting AR signaling alone is not sufficient for CRPC therapy 6,10 . Lee et al. 50 revealed that Wnt signaling might be active in PCa cells after ADT, partly due to enforcement of interaction between β-catenin and TCF. In androgen-independent PCa cell lines, the combination of a GSK3β inhibitor or APC knockdown with ENZ increases growth inhibition through repressing both AR signaling and Wnt signaling 50 . In addition, overexpressions of AR and CTNNB1 (β-catenin) were found in ENZresistant PCa cells, and the combination of β-catenin inhibitor ICG001 with ENZ improved ENZ resistance 22 . A previous study indicated that exposure of LNCaP and C4-2B PCa cells to 50 μM BR-DIM decreased the phosphorylation of GSK3β, and then inhibited nucleus translocation of β-catenin; in addition, it caused FOXO3a to bind to the p27 kip1 promoter rather than the AR promoter 51 . Leem et al. 52 showed that treatment of colon cancer cells with DIM downregulated the expressions of β-catenin, Myc and FOS, which are related to cancer prevention and prognosis. In gastric cancer cells, treatment with different concentrations of DIM produces contrary results. A high level of DIM (30 μM) resulted in inhibition of cell proliferation, but a low level (1 and 10 μM) activated Wnt4 signaling, which enhanced the progression of gastric cancer 53 . In the present study, we decided to use a high level of DIM to avoid any adverse influence, and our results showed that DIM did not significantly inhibit Wnt signaling, but combined treatment with DIM and ENZ significantly inhibited Wnt signaling through upregulating the expressions of APC and GSK3β and downregulating the β-catenin expression.
In addition, there is a connection between AR signaling and Wnt signaling. In normal prostate epithelial cells, androgen modulates the activation of AR signaling 19,28 . However, in PCa, the increased level of β-catenin leads to the formation of β-catenin-AR complexes, and the formation of β-catenin-TCF complexes is preferred under the condition of absence of androgen, which express target genes, including AR, c-myc, MMP7, and VEGF 19,28 . DIM has been proven to be a strong androgen antagonist that is a competitive inhibitor of DHT binding to AR 54 . DIM binds to AR directly, and then inhibits nucleus translocation of AR and transcription activation 54 . However, Palomera-Sanchez et al. 55 treated LNCaP cells with DIM for 48 h, and the results showed that DIM inhibited www.nature.com/scientificreports/ the protein expression of AR, but increased the mRNA expression of AR. Although the mRNA expression of AR increased, chromatin modifications at its promoter region had no significant effects 55 . In our study, DIM downregulated the protein expression of AR, but upregulated its mRNA expression (data not shown), consistent with previous studies. DIM treatment might cause AR turnover by inducing AR-protein instability. Several studies have revealed that AR overexpression and increased expressions of AR-Vs alleviate the efficacy of ENZ and induce progression of PCa 11,23,27,56,57 . Of the variants, AR-v7 is the most common in patients with ENZ resistance, and activates ligand-independent AR signaling [58][59][60] . Ong et al. 61 showed that exposure of PCa cells to BR-DIM for 96 h significantly downregulated the AR-v7 mRNA expression. However, DIM did not significantly decrease the level of AR-v7 protein expression in this study. Although the effects of DIM on AR-v7 expression were not significant, combined DIM and ENZ treatment significantly inhibited the AR-v7 expression. www.nature.com/scientificreports/ Recent studies have indicated that EMT is involved in PCa progression, migration and therapy resistance, but the effects of ADT on EMT are still unclear 62 . There are several mechanisms, including AR signaling and Wnt signaling, that activate the process of EMT 62 . Numerous molecular processes mediate the process of EMT, which is associated with loss of epithelial factors (e.g., E-cadherin and zona occludens protein-1) and gain of mesenchymal factors (e.g., N-cadherin, Fibronectin and Vimentin) 24 . In normal prostate tissue, canonical AR signaling suppresses EMT 24 ; however, a previous study showed that androgen deprivation increased the expressions of AR and AR-Vs, and upregulated EMT markers (ZEB1, Snail, Twist, N-cadherin and Vimentin), which caused therapy resistance 63 . Li et al. 63 reported that BR-DIM increased the E-cadherin expression and decreased the vimentin expression by inhibiting miR-92a expression, and then suppressed the process of EMT. In addition, miR-34a, miR-27b, miR-320m, miR-200 and let-7 were re-expressed by BR-DIM, which then repressed the expressions of AR, AR-Vs and EMT marker 45,61,64 . In the present study, DIM increased the E-cadherin expression and decreased the expressions of Fibronectin and Vimentin, then inhibited the process of EMT, which was consistent with the results of previous studies. Compared with treatment with ENZ alone, combined treatment with DIM and EMT resulted in stronger inhibition of EMT, which confirmed that DIM may mediate ENZ resistance by inhibiting EMT.
Although our results demonstrated that DIM could inhibit both Wnt signaling and AR signaling, the inhibition effects were not as obvious as those reported in previous studies, which might be because most of those studies used BR-DIM, rather than DIM. DIM is stable under acidic conditions, but the administration of DIM to female Sprague-Dawley rats by the i.p. route has better effects than the oral route 65 . Therefore, in order to enhance bioavailability, a formulation of DIM (BR-DIM) was developed to contain small particles of DIM within a water-soluble matrix, containing α-tocopherol polyethylene glycol succinate and phosphatidyl choline, by solubility-enhancing microencapsulation technology 66 . Anderton et al. administered DIM and BR-DIM to mice and then analyzed the concentration of DIM in the plasma and liver, kidney and lung tissues. The results showed that BR-DIM exhibited an approximate 50% higher bioavailability than DIM 66 . Moreover, in clinical studies, adjuvant intervention in the early stage of PCa has been performed using BR-DIM, and was proven to be an effective method by which to improve the therapeutic outcome 34 .
In our study, DIM was effective in ameliorating carcinogenesis through modulating AR signaling, Wnt signaling and EMT in ENZ-resistant cells. The findings may have several implications for the use of DIM as a dietary supplement or a therapeutic agent. The results only relate to patients who have already exhibited ENZ resistance, and in future study, cells could be treated with ENZ for a long duration in order to build a model of long-term www.nature.com/scientificreports/ ENZ resistance to solve the clinical problem. Furthermore, transcriptional activation of ARE and TCF directly influences AR and Wnt signaling, and can therefore be analyzed to confirm the results in future study.

Conclusion
Combined treatment with DIM and ENZ significantly inhibited the proliferation, invasion and migration of 22Rv1 cells. DIM could regulate Wnt signaling to inhibit the recurrence of AR signaling by reducing the expressions of AR and AR-v7. In addition, inhibition of these two signaling pathways is related to suppression of the EMT process. Our results demonstrated that treatment with DIM, or a combination of DIM and ENZ, could potentially represent other ADTs.

Wound-healing.
A wound-healing assay was employed to evaluate cell migration using a culture insert (ibidi GmbH, Munich, Germany). First, the culture insert was placed on a 24-well plate, and 10 6 cells were seeded onto the insert. After the cells had grown in the medium until approximate confluence, the culture insert was removed and washed with PBS. Under varying conditions, different treatments were added to the well, and microscopy was employed to capture images at different time points. Images were analyzed using Image J (National Institutes of Health, Bethesda, MD, USA).  www.nature.com/scientificreports/ invasion assay, while uncoated chambers were used for the migration assay. Cells at densities of 2 × 10 5 and 10 5 for the invasion and migration assays, respectively, were resuspended in 300 μL serum-free medium with different treatments and then added to the upper chamber. Then, 800 μL medium containing 10% FBS were added to the lower chamber. After 24 h of incubation, cells were fixed with 1% paraformaldehyde at 4 °C for 24 h and then stained with Giemsa (Merck, Darmstadt, Germany) for 2 min. Images were captured using a microscope, and invaded cells were counted using Image J (National Institutes of Health) 16,67,68 . Quantitative polymerase chain reaction (qPCR). Cells were treated with different concentrations of DIM and ENZ for 48 and 72 h, and then total RNA was extracted using a RNeasy Plus Mini Kit (QIAGEN, Hilden, Germany). cDNA was generated from 2 μg of total RNA using SuperScript III Reverse Transcriptase (Invitrogen, Karlsruhe, Germany). qPCR analysis was performed using 2X SYBR Green I master on a LightCycler 480 system (Roche, Penzberg, Germany). The data were analyzed using the 2 −ΔΔCt method. The gene expression levels were normalized to GAPDH, and the results are presented as the relative fold change compared with the control. The sequences of the primers used are presented in Supplementary Table 1 16,67 .