Advanced prostate cancer will often progress to a lethal, castration-resistant state. We previously demonstrated that IKKε expression correlated with the aggressiveness of prostate cancer disease. Here, we address the potential of IKKε as a therapeutic target in prostate cancer. We examined cell fate decisions (proliferation, cell death, and senescence) in IKKε-depleted PC-3 cells, which exhibited delayed cell proliferation and a senescent phenotype, but did not undergo cell death. Using IKKε/TBK1 inhibitors, BX795 and Amlexanox, we measured their effects on cell fate decisions in androgen-sensitive prostate cancer and androgen-independent prostate cancer cell lines. Cell-cycle analyses revealed a G2–M cell-cycle arrest and a higher proportion of cells with 8N DNA content in androgen-independent prostate cancer cells only. Androgen-independent prostate cancer cells also displayed increased senescence-associated (SA)-β-galactosidase activity; increased γH2AX foci; genomic instability; and altered p15, p16, and p21 expression. In our mouse model, IKKε inhibitors also decreased tumor growth of androgen-independent prostate cancer xenografts but not 22Rv1 androgen-sensitive prostate cancer xenografts. Our study suggests that targeting IKKε with BX795 or Amlexanox in androgen-independent prostate cancer cells induces a senescence phenotype and demonstrates in vivo antitumor activity. These results strengthen the potential of exploiting IKKε as a therapeutic target.
This article is featured in Highlights of This Issue, p. 395
Prostate cancer is the most frequently diagnosed cancer in North American men. Androgen deprivation therapy (ADT) is the standard of care for advanced prostate cancer. However, most patients with hormone-sensitive prostate cancer (HSPC) eventually progress to the more aggressive and lethal form of the disease, castration-resistant prostate cancer (CRPC). Mutations and aberrant expression of the androgen receptor (AR; refs. 1–4) as well as non–AR–related pathways are associated with the development of CRPC. However, the molecular events that lead to this progression are still largely unknown and likely multifactorial. A better understanding of these mechanisms is needed to identify new therapeutic targets and treatment strategies for prostate cancer that is no longer responsive to ADT in order to improve patient outcomes.
IKKε is a member of the IKK family and associates with the TANK-binding kinase (TBK1) to induce an immune response that involves Toll-like receptor 3 (TLR3)- and TLR4-mediated IFN production through the phosphorylation of IFN regulatory factors (IRFs), IRF3 and IRF7 (5). IKKε overexpression in prostate, breast, ovarian, and lung cancers can result in the deregulation of its different targets (6–8). Péant and colleagues showed that IKKε overexpression results in phosphorylation and nuclear translocation of the transcription factor CAAT/enhancer binding protein (C/EBP)-β, leading to transcription of IL6 and contributing to inflammation, cell proliferation, progression to CRPC, and the development of bone metastases (9, 10). In addition, CRPC as modelled by androgen-independent prostate cancer cell lines is associated with the constitutive expression of IKKε that is dependent on NF-κB signaling (11). It was shown that AR activation induces the expression of I-κBα, which in, turn, inhibits the NF-κB canonic pathway resulting in the failure to constitutively express IKKε in androgen-sensitive prostate cancer cell lines (11). Péant and colleagues also showed that depletion of IKKε expression decreased androgen-independent prostate cancer cell proliferation and tumor volume in a mouse model (12). Furthermore, IKKε expression directly correlates with aggressive prostate cancer disease in patients (9). These findings show that IKKε has an oncogenic role in prostate cancer and suggest its potential as a therapeutic target for CRPC.
Recent studies have shown that C/EBP-β participates in the development of senescence in prostate cancer cells in response to ADT, known as androgen deprivation—induced senescence (ADIS). Cellular senescence is a state of stable cell-cycle arrest induced by stress. In cancer, senescence plays a dual role that is either antitumorigenic by stimulating the immune system (13, 14) or protumorigenic by inducing apoptosis resistance or the senescence-associated secretory phenotype (SASP; ref. 15). Burton and colleagues demonstrated that prostate cancer cells in ADIS were enhanced in protumoral activities that involved prosurvival mechanisms and chemoresistance (10). In addition, Pernicova and colleagues showed a relationship between the inhibition of AR signaling and formation of tumor-promoting senescent cells through the downregulation of the S-phase kinase-associated protein 2 (Skp2; ref. 16).
To elucidate the role of IKKε in the progression of prostate cancer, we treated prostate cancer cell lines with the IKKε/TBK1 inhibitors, BX795, and Amlexanox, and examined the resulting cell fate decisions, such as senescence and cell death. BX795 and Amlexanox blocked IKKε activity in prostate cancer cell lines, independent of TBK1 and reduced cell proliferation with features of senescence mediated by IKKε in androgen-independent prostate cancer cells. At a high concentration, BX795 induced greater genomic instability and mitotic catastrophe that led to senescence and cell death. Using preclinical xenograft models, we showed that BX795 or Amlexanox treatment slowed the progression of tumor growth in androgen-independent prostate cancer xenografts, demonstrating a potential in targeting IKKε for novel androgen-independent prostate cancer interventions and treatment.
Materials and Methods
Cell lines and cell culture
PC cell lines PC-3, DU145, LNCaP, and 22Rv1 and the normal prostate cell line PZ-HPV-7 were purchased from the ATCC (CRL1435, ATCC HTB-81, ATCC CRL-174, and ATCC CRL-250, respectively). The prostate cancer cell line C4–2B was kindly provided by Dr. Martin Gleave (Vancouver Prostate Center). Cells were cultured in RPMI 1640 (Wisent Inc.) supplemented with 10% FBS, 100-μg/mL gentamicin, and 0.25-μg/mL amphotericin B (Invitrogen). Prior to beginning the study, cell lines were validated by short tandem repeat DNA profiling.
BX795 was purchased from Invitrogen, and Amlexanox was purchased from Abcam. Drugs were dissolved in 100% DMSO and then further diluted in complete RPMI media for in vitro experiments. Drugs were added 24 hours after seeding.
Cloning, viruses, and infections
The lentiviral short hairpin RNA (shRNA) against IKKε (shIKKε.A: 5′-ATGGCTATCGTGTTGTGGGCA-3′; shIKKε.B: 5′-CGAAACATTAGCTCCTGCCCA-3′) and red fluorescent protein (RFP; control; shRFP: 5′-ACTACACCATCGTGGAACAGT-3′) were purchased from Dharmacon (lentiviral PLKO.1 vector with puromycin selection). Viruses were produced as previously described (17), and titers were adjusted to achieve approximately 90% infectivity. Selection of infected cells started 48 hours after seeding with puromycin (1 μg/mL), and clones were either used immediately or harvested, frozen, and stored at −80°C. The cells were left in culture for 24 hours after thawing before starting experiments.
IKKε-depleted PC-3 cells were plated at 1,000 cells per well in 6-well plates and incubated for 24 hours. After 6 days, cells were fixed with methanol and stained with a solution of 50% v/v methanol and 0.5% w/v methylene blue (Sigma-Aldrich). Colonies were counted under a stereomicroscope, and numbers were reported as a percentage of control. Each experiment was repeated three times.
IncuCyte cell proliferation assays
Cells were seeded at a density of 2,000 cells per well in 96-well plates. After 24 hours, cells were treated with BX795 or Amlexanox at different concentrations for 6 days. Cell proliferation was followed by phase contrast microscopy using an IncuCyte Zoom Live-Cell Imaging System (IncuCyte HD, Essen BioScience). Images were taken every 2 hours for 6 days. Proliferation growth curves were constructed using the cell confluence values established by confluence mask settings determined by the IncuCyte Zoom software.
siRNA targeting TBK1 was purchased from Dharmacon. Transient transfection of prostate cancer cells was performed using lipofectamine 2,000 (Invitrogen). Briefly, 400,000 cells were seeded into 6-well plates. After 24 hours, cells were transfected with 5-μmol/L siRNA for 4 hours. Before transfection, the siRNA was mixed with lipofectamine 2,000 for 20 minutes, according to the manufacturer's instructions. The transfection medium was replaced by RPMI supplemented with 10% FBS. Cells were collected and analyzed 48 hours later.
Cells were seeded at 500,000 cells in 60-mm plates. After 24 hours, cells were pretreated with 2.5-μmol/L BX795 for 2 hours before transfection with 10 μg/mL of the TLR3 agonist pIpC using the lipofectamine 2,000 protocol as described above. After 2 or 4 hours, cells were collected for Western blot experiments.
Western blot analysis
Proteins were obtained from whole cell lysates after a 30-minute incubation at 4°C in lysis buffer (Triton X-100, 50 mmol/L Tris pH 7.4, 150 mmol/L NaCl, 2 mmol/L EDTA, and 10% glycerol) supplemented with a protease inhibitor cocktail (Roche Applied Science) and phosphatase inhibitors (5 mmol/L NaF, 200 μmol/L NA3VO4, and 100 μmol/L PMSF). Protein concentrations were quantified by Bradford assay, and 30 μg of protein per well was separated in 4% to 15% SDS polyacrylamide gels, then transferred onto polyvinylidene difluoride (PVDF) membranes. The antibodies used in this study were anti-IKKε (1/2,000, Imgenex), anti–phospho-IRF-3 (1/3,000, Abcam), anti–ß-actin (1/2,000, Abcam), anti-TBK1 (1/1000, Cell Signaling Technology), and anti–IRF-3 (1/800, Santa Cruz).
Cell death and cell-cycle analyses by flow cytometry
Cells were plated at 80,000 cells per well in 6-well plates and incubated for 24 hours. After treatment with 0.75- or 2.5-μmol/L BX795, the supernatant and adherent cells were collected and pooled. For cell death experiments, cells were washed with PBS and incubated with Annexin V (BD Biosciences) for 25 minutes followed by the addition of DRAQ7 (Abcam) for 5 minutes. Cells were centrifuged, then resuspended in PBS. For cell-cycle experiments, cells were incubated with 100-μg/mL RNase A and 50-μg/mL propidium iodide (PI; Sigma Aldrich) for 30 minutes in the dark. Cell staining was analyzed using the Fortessa flow cytometer and the FlowJo 10 software (BD Biosciences).
The senescence-associated (SA)–β-galactosidase assay was performed as previously described (18). Briefly, 20,000 cells were seeded into 24-well plates. After 24 hours, cells were treated with 0.75 or 2.5 μmol/L BX795 for 6 days. Cells were fixed in formalin (10%), washed with PBS, and incubated overnight at 37°C with fresh staining solution (40 mmol/L citrate acid/sodium phosphate, 150 mmol/L NaCl, 2 mmol/L Mg2Cl.6H2O, 5 mmol/L K4FeCN6.3H2O, 5 mmol/L K3FeCN6, and 1 mg/mL X-Gal substrate). The next day, cells were washed with PBS and stained with DAPI. Pictures were taken with bright-field microscopy. Two wells per condition were counted, and positive blue cells were reported as a percentage of the total cell number.
Total RNA extraction, reverse transcription, and qPCR
Cells were seeded at 100,000 cells per well in 6-well plates, incubated for 24 hours, and treated with 0.75 or 2.5 μmol/L BX795 for 6 days. RNA was extracted with the RNeasy Mini Kit (Qiagen) according to the manufacturer's protocol. RNA concentrations were measured by NanoDrop (Thermo Fisher Scientific) and assessed for purity. For reverse transcription experiments, the Quantitect Reverse Transcription kit (Qiagen) was used with the following conditions: 2 minutes at 42°C, 15 minutes at 42°C, and 3 minutes at 95°C.
qRT-PCR was performed using the SYBR Select Master Mix (Applied Biosystems) in the Applied 63 BioSystems Step One Plus system (UDG activation at 50°C/2 minutes, followed by AmpliTaq activation plus denaturation cycle at 95°C/2 minutes, followed by 40 cycles at 95°C/15 seconds, 60°C/1 minute, and 72°C/30 seconds). Target gene expression was normalized to TATA-box binding protein (TBP) levels. For each sample, the fold change value was expressed as 2–ΔΔCT: ΔΔCT = (CTtarget gene − CTTBP) control − (CTtarget gene − CTTBP) treated. Primers are listed in Supplementary Table S1.
γH2AX foci detection by immunofluorescence
Cells were seeded at 15,000 cells per well onto coverslips in 24-well plates. After 24 hours, cells were treated with 0.75 or 2.5 μmol/L BX795 for 6 days. Then, cells were washed with PBS, fixed with formalin, washed with PBS, permeabilized in PBS containing 0.25% Triton for 20 minutes, and blocked with PBS containing 1% BSA for 30 minutes. Cells were incubated with primary anti-γH2AX antibody (Millipore) for 2 hours. After washing with PBS, cells were incubated with Alexa Fluor Cy5 goat anti-mouse IgG secondary antibody (Life Technologies) for 1 hour. After washing with PBS and water, coverslips were mounted onto slides using ProLong Gold Anti-fade Mountant with DAPI (Life Technologies Inc.). Images (mosaics 4 × 4) were obtained using a Zeiss AxioObserver Z1 fluorescent microscope (Carl Zeiss). Automated analysis software from Zeiss (AxioVision, Carl Zeiss) was used to count and calculate the average number of foci per nucleus.
Cells were seeded at 15,000 cells per well onto coverslips in 24-well plates and incubated for 24 hours. IKKε-depleted cells or BX795-treated cells were incubated with 10 μmol/L of 5-ethynyl-2′-deoxyuridine (EdU; Invitrogen) overnight at 37°C. After washing with PBS, cells were fixed with formalin for 10 minutes. EdU staining was measured using the reaction solution containing 100 mmol/L Tris pH 8.5, vitamin C, CuSO4, and Alexa Fluor 647 dye. Cells were incubated with staining solution for 30 minutes in the dark. Next, the cells were washed with PBS, and coverslips were mounted onto slides with ProLong Gold Antifade Mountant with DAPI. Images (mosaics 4 × 4) were acquired with the Zeiss AxioObserver Z1 microscope and were analyzed with ImageJ software to count the total cell number and positive cells.
Murine xenograft model
All animal experiments complied with all relevant ethical regulations for animal testing and research at Centre de recherche du CHUM (CRCHUM). All experiments were conducted with approval from our institutional committee on animal care [Comité Institutionel de Protection des Animaux (CIPA)] under protocol number C14011AMMs. Experiments were carried out in 8-week-old male NODRag1null, IL2rgnull, NOD rag-γ (NRG) mice (from our in-house mouse colony). DU145, PC-3, and 22Rv1 cells were prepared at 1 million cells/200 μL in PBS–matrigel (v/v) and subcutaneously injected into the mouse flank. Mice were weighed once a week, and tumor volumes were measured twice a week. Tumor volumes were calculated using the following equation: V (mm3) = a × b × h, where a is the largest diameter, b is the perpendicular diameter, and h is the height. When xenograft tumor volumes reached between 200 and 300 mm3, mice were randomized into three groups: control (carrier: PBS-5% DMSO), BX795 (500 μg/kg), or Amlexanox (25 mg/kg). The carrier, BX795, and Amlexanox were administered by intraperitoneal injection every day. Mice were sacrificed when the xenograft tumor volumes reached between 2,000 and 2,500 mm3.
Immunofluorescence on xenograft tissue
Tumors were harvested, formalin-fixed, and paraffin-embedded. According to the manufacturer’s instructions, 4-μm-thick tissue sections were cut with a microtome and stained using the automated Ventana Discovery XT staining system (Ventana Medical Systems). Antigen retrieval was performed in Cell Conditioning 1 solution, and slides were incubated with anti-Geminin (Proteintech Group) antibodies in PBS at 37°C for 60 minutes. A negative control slide using PBS instead of the primary antibody was prepared in parallel. On the bench, slides were incubated for 20 minutes with blocking solution (Dako, Agilent Technologies) followed by washing with PBS and by incubation with the secondary antibody (anti-rabbit Cy5; Life Technologies Inc.) in PBS at room temperature for 45 minutes. Finally, after washing with PBS, slides were incubated for 15 minutes at room temperature with a 0.1% (w/v) solution of Sudan Black in 70% ethanol to quench tissue autofluorescence. Following a PBS wash, slides were mounted using ProLong Gold Antifade Mountant with DAPI and stored at 4°C. Images were taken with a Zeiss microscope and quantification was performed with ImageJ software.
Data were expressed as the mean ± SEM. Statistical analyses were performed using GraphPad Prism 6 software (GraphPad Software). Statistical significance was determined using Student t test or one-way ANOVA and was set at P = 0.05. All in vitro experiments were repeated at least three times, and the number of mice is indicated for each cohort.
Effect of IKKε knockdown on prostate cancer cell fate decisions
To study the role of IKKε in prostate cancer cell fate decisions, we first verified the reported differences in IKKε expression among our prostate cancer cell lines: androgen-independent prostate cancer cells (PC-3, DU145, and C4–2B) demonstrated IKKε overexpression (Fig. 1A) in comparison with androgen-sensitive prostate cancer cells (LNCaP, 22Rv1) (11). TBK1, the IKKε binding partner in the IFN signaling pathway, was equally expressed in androgen-independent prostate cancer and androgen-sensitive prostate cancer cell lines (Fig. 1A). Although inhibition of IKKε expression decreases PC-3 cell proliferation in vitro (12), TBK1 knockdown with siRNA did not affect the proliferation of androgen-independent prostate cancer cell lines (Supplementary Fig. S1). To further understand the IKKε-mediated decrease in cell proliferation, we performed IKKε depletion in PC-3 cells using lentiviral shRNAs (Fig. 1B). Cell-cycle progression was followed using a colony formation assay (Fig. 1C) and short EdU pulse-labeling assays (Fig. 1D and E). Inhibition of IKKε expression with either shIKKε.A or shIKKε.B significantly decreased colony formation (Fig. 1C), and reduced EdU incorporation with a 20% decrease in DNA synthesis (Fig. 1D and E).
We then evaluated the rate of cell death induced by IKKε depletion in PC-3 cells and observed a slight increase in cell death (<5%) with shIKKε.A, whereas no increase was observed with shIKKε.B (Fig. 1F; Supplementary Fig. S2). Inhibition of IKKε also induced a 10.5-fold increase in SA-β-galactosidase activity (Fig. 1G and H), a specific marker of cellular senescence. These results suggest that lack of IKKε activity induces cell fate decisions that include moderate cell-cycle arrest and a partial senescence phenotype.
Effect of IKKε/TBK1 inhibitor BX795 on prostate cancer cell proliferation
To further characterize prostate cancer cell fate decisions in response to IKKε inhibition, we used a known inhibitor of the IKKε/TBK1 complex, BX795 (5, 19). First, to detect IKKε activity, cells were treated with the TLR3 agonist pIpC, which leads to the phosphorylation of IRF-3 by the IKKε/TBK1 complex and activates genes in the IFN pathway (Supplementary Fig. S3; ref. 20). After 2 and 4 hours of stimulation with pIpC, IRF-3 phosphorylation was increased in androgen-independent prostate cancer (PC-3, DU145, and C4–2B) and androgen-sensitive prostate cancer (LNCaP and 22Rv1) cell lines. However, phosphorylation was impaired in all cell lines pretreated with 2.5 μmol/L of BX795 (Supplementary Fig. S3), demonstrating that BX795 blocked the IKKε/TBK1 complex activity.
We then examined whether BX795 decreased prostate cancer cell proliferation as efficiently as IKKε depletion in PC-3 cells. Real-time imaging of proliferation assays confirmed a concentration-dependent inhibition of androgen-independent prostate cancer cell proliferation (Fig. 2A) when concentrations of BX795 were over 750 nmol/L. At the highest concentration (10 μmol/L), proliferation was blocked for all androgen-independent prostate cancer cell lines, and cell numbers even decreased for DU145 and C4–2B (Fig. 2A). At 100 nmol/L, a slight increase in C4–2B cell proliferation was noted. In contrast, the effect of BX795 on androgen-sensitive prostate cancer cell proliferation (LNCaP and 22Rv1) appeared limited (Fig. 2B). Although proliferation of LNCaP decreased (0.8-fold) at concentrations greater than 1 μmol/L, cells were more resistant to treatment than the androgen-independent prostate cancer cell lines (Fig. 2B and C) and growth was not arrested, even at 10 μmol/L. Proliferation of 22Rv1 cells was only affected when BX795 concentrations reached 10 μmol/L (Fig. 2B). Using the AUC, the IC50 of BX795 was calculated for each cell line and revealed that androgen-independent prostate cancer cells had a significantly lower IC50 (1.3 μmol/L, 2.2 μmol/L, and 4.2 μmol/L for PC-3, DU145, and C4–2B, respectively) than androgen-sensitive prostate cancer cells (12.3 μmol/L and 10.1 μmol/L for LNCaP and 22Rv1, respectively; Fig. 2C).
In addition, we examined whether BX795 influences the proliferation of a normal prostate cell line, PZ-HPV-7. We monitored IKKε expression by Western blot. As expected, we do not observe constitutive expression of IKKε in PZ-HPV-7 cells (Supplementary Fig. S4). Moreover, real-time imaging assays of proliferation showed that BX795 does not have a significant effect on PZ-HPV-7 proliferation (Supplementary Fig. S4).
These results show that BX795 has a powerful antiproliferative effect on androgen-independent prostate cancer cells compared with androgen-sensitive prostate cancer cells. To characterize this effect, two BX795 concentrations were selected for further study: 0.75 and 2.5 μmol/L.
BX795 treatment induces a senescence-like phenotype in androgen-independent prostate cancer cells
Because IKKε depletion induced a senescence phenotype in PC-3 cells (Fig. 1), we examined whether BX795 could also induce senescence in androgen-independent prostate cancer cells. Using flow cytometry, cell-cycle progression was analyzed after 6 days of treatment. No effect was observed with 0.75 μmol/L BX795 in all cell lines except for a low but significant G2–M phase accumulation in PC-3 cells (Fig. 3A and B; Supplementary Fig. S4A and S4B). However, at 2.5 μmol/L BX795, the cell-cycle distribution was significantly altered in all cell lines. Moreover, under these conditions, a subpopulation of cells with abnormal DNA content (8N) appeared (Fig. 3A; Supplementary Fig. S5A) in all AnRPC cell lines, with DU145 showing the highest levels. Even with a delayed G2–M phase, this 8N subpopulation was not observed in androgen-sensitive prostate cancer cell lines (Fig. 3B; Supplementary Fig. S5B).
We then evaluated the apoptosis levels of cells after 6 days of BX795 treatment by flow cytometry. At 0.75 μmol/L BX795, no increase of cell death was observed for all prostate cancer cell lines (Supplementary Fig. S5). At 2.5 μmol/L BX795, apoptosis (cumulative AnnexinV/DRAQ7-positive cells) was significantly increased by 2.8-, 5.3-, 7-, 4.3-, and 4.2-fold for PC-3, DU145, C4–2B, LNCaP, and 22Rv1 cells, respectively (Supplementary Fig. S6).
Real-time imaging revealed changes in morphology of androgen-independent prostate cancer cells treated after BX795 treatment. We confirmed cell enlargement using flow cytometry analysis of cell size (forward scatter) and cell granularity (side scatter) in two distinct cell populations: population 1 (normal size) and population 2 (enlarged cells; Fig. 3C; Supplementary Fig. S5D and S5E). We observed a significant increase in the ratio of enlarged versus normal-sized cells in 2.5 μmol/L BX795-treated androgen-independent prostate cancer cells at day 6 (Fig. 3C; Supplementary Fig. S5D). Moreover, this trend was also observed in 2.5 μmol/L BX795-treated LNCaP cells (Fig. 3C; Supplementary Fig. S5E). This cell enlargement is consistent with a senescence phenotype.
Following 6 days of treatment with 0.75 μmol/L BX795, β-galactosidase–positive cells were observed in 20% of PC-3 and DU145 cells, and 30% of C4–2B cells (Fig. 3D and E). Treatment with 2.5 μmol/L BX795 resulted in higher levels of positive cells with 40% for PC-3 and DU145 cells and 70% for C4–2B (Fig. 3D and E). In androgen-sensitive prostate cancer cells (LNCaP and 22Rv1), β-galactosidase–positive cells were not detected after 6 days with either 0.75 or 2.5 μmol/L BX795 (Supplementary Fig. S5C).
To confirm the senescence-like phenotype induced in androgen-independent prostate cancer cells in response to BX795 treatment, EdU pulse-labeling was used to follow SA proliferation arrest, a key hallmark of senescence. Treatment with 0.75 μmol/L BX795 in androgen-independent prostate cancer showed a significant decrease in DNA synthesis (short 8 h EdU pulse) after 6 days, but not a complete proliferation arrest as observed with a longer 24-hour EdU pulse (Fig. 3F; Supplementary Fig. S7). In contrast, 2.5 μmol/L BX795 induced stronger arrests of DNA synthesis in androgen-independent prostate cancer cells with short or long pulses; only 5% (PC-3), 12% (DU145), and 5% (C4–2B) of cells were positive after a long EdU pulse (Fig. 3F; Supplementary Fig. S7). Compared with androgen-independent prostate cancer cells untreated or treated with 0.75 μmol/L, treatment with 2.5 μmol/L BX795 decreased cell-cycle progression at 16-, 6-, and 3.5-fold for PC-3, DU145, and C4–2B, respectively (Fig. 3F; Supplementary Fig. S7). In androgen-sensitive prostate cancer cells, 0.75 μmol/L BX795 did not alter DNA synthesis (short EdU pulse) or the cell cycle (long EdU pulse) in 22Rv1 cells (Fig. 3G; Supplementary Fig. S7). In LNCaP cells, the same treatment only affected DNA synthesis. With 2.5 μmol/L BX795 treatment, DNA synthesis decreased by two- and ninefold for LNCaP and 22Rv1, respectively. Also, this treatment altered cell-cycle progression with a decrease of two- and 5.5-fold for LNCaP and 22Rv1, respectively (Fig. 3G; Supplementary Fig. S7).
These results indicate that treatment with 0.75 μmol/L BX795 was capable of inducing a senescence-like phenotype in androgen-independent prostate cancer cell lines, whereas a higher concentration of 2.5 μmol/L induced abnormal DNA (8N) content and cell death. In androgen-sensitive prostate cancer cells, both BX795 concentrations did not induce a senescence phenotype, but only 2.5 μmol/L BX795 decreased cell-cycle progression and induced cell death.
BX795-induced senescence features are associated with DNA damage and cyclin-dependent kinase inhibitor overexpression
Increased DNA damage is often associated with the induction of senescence (21). To determine whether the senescence phenotype induced by BX795 was associated with DNA damage, we assessed the presence and accumulation of γH2AX foci, a marker of DNA double-strand breaks (22), by immunofluorescence (IF) after 6 days of treatment. At 0.75 μmol/L BX795, the number of γH2AX foci increased fourfold in PC-3 and DU145 cells (Fig. 4A and B), but no increase was observed in C4–2B cells (Fig. 4B; Supplementary Fig. S8A). At 2.5 μmol/L BX795, the number of γH2AX foci increased substantially in all androgen-independent prostate cancer cells (10-, 17-, and sevenfold in PC-3, DU145, and C4–2B cells, respectively). In contrast, neither concentration of BX795 induced any increase of γH2AX foci in androgen-sensitive prostate cancer cells (LNCaP and 22Rv1) after 6 days (Fig. 4B; Supplementary Fig. S8B).
We hypothesized that this increased DNA damage was associated with genomic instability, which is characterized by an accumulation of micronuclei and multinuclei structures (23). At 0.75 μmol/L BX795, 20% of PC-3 and DU145 cells and 10% of C4–2B cells were positive for micronuclei and multinuclei structures, indicating genomic instability (Fig. 4C; Supplementary Fig. S9). At 2.5 μmol/L BX795, an increase in genomic instability was demonstrated by 80%, 95%, and 75% of PC-3, DU145, and C4–2B cells, respectively (Fig. 4C; Supplementary Fig. S9). Micronuclei and multinuclei structures were not detected in androgen-sensitive prostate cancer cells LNCaP and 22Rv1 at either BX795 concentrations (Fig. 4C).
Alterations in cell-cycle progression (Fig. 3A and B) implicate cell-cycle regulators such as cyclin-dependant kinase inhibitors (CDKis). CDKis block cell-cycle progression by inhibiting cyclin and CDK interactions, and their overexpression is also a marker of senescence (24). CDKi expression was followed by qRT-PCR after 6 days of BX795 treatment and showed differences between cell lines (Fig. 4D and E). PC-3 cells treated with 2.5 μmol/L BX795 showed a significant increase in CDKN2B (p15) and CDKN1A (p21) expression levels compared with the untreated control (Fig. 4D). The same BX795 concentration in DU145 cells showed a significant increase in p15 and p16 expression levels (Fig. 4D). In C4–2B cells, p21, p16, and p15 were significantly overexpressed compared with the control (Fig. 4D). For androgen-sensitive prostate cancer cells (LNCaP and 22Rv1), only p21 expression was increased at 2.5 μmol/L BX795 (Fig. 4E), whereas p16 and p15 expression levels did not change after BX795 treatments (Fig. 4E).
IKKε inhibition delayed tumor growth in a mouse prostate cancer xenograft model
To ensure the activity we observed with BX795 was specific to IKKε, we evaluated the effect of BX795 on IKKε-depleted PC-3 cells by clonogenic and SA-β-galactosidase assays. The results suggest that at a dose of 0.75 μmol/L BX795 inhibition appears to be IKKε-specific. In contrast, at 2.5 μmol/L BX795 (Supplementary Fig. S10), we did note off-target activity. To ensure specificity in vivo, we used an FDA-approved IKKε inhibitor, Amlexanox, to evaluate the impact of IKKε inhibition on tumor growth in a mouse model. First, we validated its activity in vitro to ensure that Amlexanox induced the same cell fate decisions observed with BX795 in androgen-independent prostate cancer and androgen-sensitive prostate cancer cells. Androgen-independent prostate cancer cell proliferation significantly decreased at 50, 100, and 200 μmol/L Amlexanox for PC-3, DU145, and C4–2B cells, respectively (Supplementary Fig. S11A), whereas androgen-sensitive prostate cancer proliferation (LNCaP and 22Rv1) was not altered by similar Amlexanox treatment (Supplementary Fig. S11B). In addition, 100 μmol/L Amlexanox significantly increased SA-β-galactosidase activity in PC-3 and DU145 cells (Supplementary Fig. S11C and S11D). At 200 μmol/L Amlexanox, SA-β-galactosidase activity was increased for all androgen-independent prostate cancer cell lines (Supplementary Fig. S11C and S11D). In contrast, androgen-sensitive prostate cancer cells did not demonstrate any SA-β-galactosidase activity at either Amlexanox concentration (Supplementary Fig. S11E). These results confirm that Amlexanox treatment had a comparable effect to BX795 in androgen-independent prostate cancer cells: decreased cell proliferation and increased senescence phenotype.
To investigate the potential effects of BX795 or Amlexanox on prostate cancer tumor growth in a mouse model, subcutaneous androgen-independent prostate cancer xenografts using PC-3 and DU145 cell lines and an androgen-sensitive prostate cancer xenograft using 22Rv1 cells were established. About 2 or 3 weeks after prostate cancer cell injection, mice were treated with a daily dose of 500 μg/kg BX795 or 25 mg/kg Amlexanox for 4 weeks (or with the carrier as a control). There was no observable effect of BX795 or Amlexanox on the weight or behavior of the mice (Supplementary Fig. S12). PC-3 and DU145 tumor growth were significantly reduced by BX795 and Amlexanox treatments compared with the control group (Fig. 5A and B; Supplementary Fig. S13). In comparison, BX795 and Amlexanox did not affect the 22Rv1 tumor growth, which remained comparable with the control group. To validate the antiproliferative effects of BX795 and Amlexanox in our xenograft model, Geminin (GMNN) staining was performed on the harvested xenograft tissue and analyzed by IF, as described by Wohlschlegel and colleagues (25). BX795 treatment altered the cell cycle in androgen-independent Geminin (GMNN) xenografts, indicated by the proliferation-associated GMNN staining that showed a decrease by 1.4- and 1.7-fold in PC-3 and DU145 xenografts, respectively. Similarly, GMNN staining decreased by 1.6- and twofold in Amlexanox-treated PC-3 and DU145 xenografts, respectively (Fig. 5D–G). In 22Rv1 xenografts, GMNN expression was not altered by BX795 as the number of positive cells were similar in both treatment and control groups (Fig. 5H and I), and Amlexanox treatment showed only a slight increase (1.2-fold) in GMNN staining (Fig. 5H and I).
Androgen-independent prostate cancer cell lines (PC-3, DU145, and C4–2B) have characteristics close to CRPC tumors in patients. These cell lines show a constitutive activity of NF-κB and overexpress IKKε. Although androgen-sensitive prostate cancer cell lines (LNCaP, 22Rv1) are similar to HSPC tumors as express AR and respond to androgen signaling. We have previously shown that IKKε expression correlates with the aggressiveness of prostate cancer in patients (9) and that IKKε expression is inducible in androgen-sensitive prostate cancer cell lines, while it is constitutively overexpressed in androgen-independent prostate cancer cells (12). In this study, we show that C4–2B cells, an androgen-independent prostate cancer cell line derived from the IKKε-inducible androgen-sensitive prostate cancer LNCaP cell line (26), display IKKε overexpression. These results support the premise that IKKε expression is a key factor in prostate cancer progression. We also show that IKKε depletion slows proliferation and induces moderate SA-β-galactosidase activity in PC-3 cells. Inhibition of the IKKε/TBK1 complex by BX795 or Amlexanox decreased proliferation in androgen-independent prostate cancer cell lines and induced features of senescence accompanied by induction of DNA damage and genomic instability in androgen-independent prostate cancer cells. Unlike true senescence, this effect appears to be reversible, a characteristic that has previously been identified in cancer cells (17). In contrast, these inhibitors had a limited antiproliferative effect on androgen-sensitive prostate cancer cell lines, and BX795 did not induce a senescence phenotype in these cells. Finally, treatment with BX795 and Amlexanox delayed androgen-independent prostate cancer tumor growth in mouse models but had no effect on 22Rv1 (androgen-sensitive prostate cancer) xenograft growth. The IKKε inhibitors seem only efficient against androgen-independent prostate cancer cell lines. Although there are many drugs that target HSPC tumors, in the CRPC setting, there are few therapeutic strategies and none that are curative, highlighting the importance of targeting this stage of the disease.
A variety of cancers, including lung, breast, melanoma, pancreatic, bladder, and colon cancers, demonstrate overexpression of TBK1, which is the binding partner of IKKε (27–30). TBK1 plays an important role in oncogenic transformation in melanoma, breast, and lung cancers (31–33) and may also play a role in drug resistance. For example, TBK1 overexpression in breast cancer is correlated with tamoxifen resistance (33) and is associated with resistance to BRAF (31) or MEK inhibitors (28) in melanoma. TBK1 inhibition by BX795 increases the sensitivity of breast cancer cells to tamoxifen treatment (33) and has an antiproliferative effect in human oral squamous cell carcinoma (34). BX795 also inhibits TBK1 enhancement of bladder cancer cell proliferation and migration (35). Moreover, IKKε/TBK1 inhibition by Amlexanox inhibits melanoma proliferation (36) and potentiates the antitumor and antimetastatic effects of docetaxel in breast cancer (37). Although BX795 and Amlexanox reduced androgen-independent prostate cancer tumor size and in vitro cell growth, we showed that this was not specifically associated with TBK1 because TBK1 knockdown had no impact on androgen-independent prostate cancer cell proliferation (Supplementary Fig. S1). In contrast, IKKε depletion significantly decreased androgen-independent prostate cancer proliferation [(12) and Fig. 1]. Based on these results, we suggest that the effect of BX795 and Amlexanox in androgen-independent prostate cancer cell lines were not due to TBK1 inhibition but is principally related to IKKε inhibition.
BX795 was originally developed as a PDK1 inhibitor but functions more effectively as an inhibitor of the IKKε/TBK1 complex (19). Although previous studies have reported that BX795 had multiple off-target effects, such as the Aurora, AKT (34), TAK1, JNK, or MKK4 (38), our results showed that BX795 can show high specificity to inhibit IKKε activity at appropriate doses. We observe that BX795 induces a phenotype of senescence accompanied by genomic instability only in androgen-independent prostate cancer cells (PC-3, DU145, and C4–2B cells) for which IKKε expression is constitutive. In the LNCaP and 22Rv1 lines, where IKKε expression is only inducible, we do not observe cell senescence, genomic instability, or DNA damage. Furthermore, we showed in our study that BX795 treatment replicated the effect of IKKε depletion by shRNA on PC-3 cells (i.e., increase of SA-β-galactosidase activity and cell growth arrest). In fact, BX795 treatment at low (0.75 μmol/L) and high (2.5 μmol/L) concentrations produced two different phenotypes in androgen-independent prostate cancer cells. At 0.75 μmol/L, BX795 induced a moderate senescence in androgen-independent prostate cancer cells, accompanied by slower cell growth with an increase of γH2AX foci and about 20% of cells showing micronuclei and multi-nuclei structures. At 2.5 μmol/L of BX795, growth arrest and a senescence phenotype were observed in a greater proportion of androgen-independent prostate cancer cells of which about 80% of cells showed genomic instability and a significant accumulation of DNA damage with an increase of cell death. Because IKKε depletion and 0.75 μmol/L BX795 treatment induce a comparable, moderate senescence in PC-3 cells, and a higher BX795 concentration induces a stronger senescence phenotype with genomic instability, we suggest that increasing BX795 intensifies the effect of inducing a senescence phenotype in an IKKε-dependent manner. In support of this notion, our specificity experiments show that 0.75 μmol/L BX795 is a specific dose inhibiting IKKε activity. Likewise, Amlexanox, an IKKε inhibitor with better specificity for IKKε than BX795, also induces a senescence phenotype similar to BX795. Amlexanox is an FDA-approved drug for the treatment of aphthous ulcers and shows the same in vivo antitumor effect as BX795. Notably, IKKε inhibitors have no effect on the proliferation or induction of senescence in androgen-sensitive prostate cancer cells, which do not constitutively express IKKε. Therefore, these observations support the hypothesis that BX795 and Amlexanox induce a senescence phenotype via IKKε in androgen-independent prostate cancer cell lines.
Previously, we have shown that IKKε overexpression activates IL6 and IL8 secretion, whereas IKKε depletion inhibits their expression (12, 39). High concentrations of these cytokines may also exist in the SASP of senescent cells, including androgen-independent prostate cancer cells (40). This excess in IL6 and IL8 may enhance the growth of neighboring nonsenescent androgen-independent prostate cancer cells through activation of the androgen receptor in a STAT3-dependent manner leading to cell proliferation (41, 42). Moreover, the upregulation of IL6 and activation of STAT3 lead to crucial mechanisms of therapeutic resistance in cancer cells (43, 44). These observations suggest that IKKε may contribute to androgen-independent prostate cancer progression through IL6 and IL8 upregulation.
C/EBP-β is a key senescence regulator that represses and activates different senescence genes in a context-dependent manner (45) and induces the expression of SASP genes, including IL6 and IL8 (45). Numerous studies have shown that C/EBP-β is involved in diverse functions such as inflammation, cellular differentiation, cell survival, tumor invasiveness, and growth arrest (46–51). In prostate cancer, C/EBP-β is activated by IKKε and induces IL6 and IL8 secretion that is associated with inflammatory prostate cancer progression (12). C/EBP-β expression was shown to be high in patients with HSPC, and upon androgen deprivation, C/EBP-β levels rapidly increased in androgen-sensitive prostate cancer cell lines (9, 52). In addition, androgen deprivation induces a senescence phenotype (ADIS) associated with tumor-promoting traits that lead to prosurvival mechanisms (10, 16, 52, 53). Our observations and data from the literature suggest a mechanism of ADT resistance influenced by IKKε. The induction of ADIS leads to IL6 and IL8 secretion via C/EBP-β activation (40, 52). When ADT resistance sets in, IL6 and IL8 levels continue to be high, thereby contributing to prosurvival mechanisms, cell proliferation and ultimately castrate resistance. Because proliferating androgen-independent prostate cancer cells no longer benefit from the SASP, they would normally reduce IL6 and IL8 secretion. However, we propose that androgen-independent prostate cancer cells start to constitutively express IKKε to activate C/EBP-β, thus driving IL6/IL8 secretion independent of senescence. Conversely, IKKε inhibition by shRNA or inhibitors, decreases IL6 and IL8 secretions and pushes androgen-independent prostate cancer cells to return to a senescence phenotype.
In addition to its key function in proinflammatory cytokine secretion, C/EBP-β has a role in DNA damage resolution because it regulates the transcription of genes involved in DNA repair, such as BRCA1/2, ATM, or Rad51 (54–56). Importantly, IKKε also plays a role in protecting cells from DNA-damage–induced cell death (57). After genotoxic stress, IKKε translocates to the nucleus, leading to the formation of promyelocytic leukemia nuclear bodies (57). We have shown that IKKε translocates into the nucleus of androgen-independent prostate cancer cells (39). This prosurvival function of IKKε in the nucleus is dependent on the phosphorylation of IKKε itself and its targets, which include C/EBP-β (57). Our results showing DNA damage accumulation in response to BX795 treatment support the following hypothesis: nuclear IKKε activity increases C/EBP-β activation, which promotes DNA repair mechanisms that may allow the cancer cell to either be protected from ADIS and/or promote resistance to ADIS.
We have previously shown that IKKε depletion decreased cell proliferation in vitro and in vivo in PC-3 cells (12). In this study, we demonstrated that IKKε inhibitors induced a senescence phenotype and genomic instability in androgen-independent prostate cancer cell lines. We validated that IKKε inhibitors delayed androgen-independent prostate cancer tumor growth in mouse models. These observations add to previous evidence suggesting that IKKε allows prostate cancer cells to progress to a androgen-independent or castrate-resistant state by escaping ADIS, possibly through the activation of C/EBP-β, which induces DNA damage repair and IL6/IL8 secretion. Together, this supports the therapeutic potential of IKKε for the treatment of patients with CRPC, as well as for the prevention of early-onset prostate cancer progression in combination with ADT.
No disclosures were reported.
S. Gilbert: Conceptualization, validation, investigation, visualization, methodology, writing–original draft. B. Péant: Conceptualization, supervision, investigation, writing–original draft. N. Malaquin: Visualization, methodology. V. Tu: Methodology. H. Fleury: Visualization, methodology. K. Leclerc-Desaulniers: Methodology. F. Rodier: Visualization, methodology. A.-M. Mes-Masson: Supervision, funding acquisition, project administration, writing–review and editing. F. Saad: Supervision, funding acquisition, project administration, writing–review and editing.
We thank the animal facility of the Centre de recherche du CHUM (CRCHUM) for performing all animal experiments as well as the molecular pathology core facility and the Institut du cancer de Montréal (ICM) imaging/live imaging platform for performing IHC staining and scans, respectively. The authors also thank Christine Caron for assisting with the quantification of immunofluorescence signals. F. Rodier, A.-M. Mes-Masson, and F. Saad are researchers of the CRCHUM/ICM, which receives support from the Fonds de recherche du Québec - Santé (FRQS). F. Rodier was supported by FRQS junior II – senior career awards (33070, 281706). N. Malaquin was supported by a Cancer Research Society (CRS) operating grant to F. Rodier (23464). V. Tu was supported by a FRQS PhD fellowship and received ICM-Canderel excellence awards. S. Gilbert is supported by a studentship from the Canderel Fund of the ICM. F. Saad holds the Université de Montréal Raymond Garneau Chair for Prostate Cancer Research. This work was supported by the Raymond Garneau Chair in Prostate Cancer research and grants from the Canadian Urologic Oncology Group, Prostate Cancer Canada, ICM, and the Cancer Research Society.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).