Seleno-aspirin compound AS-10 promotes histone acetylation ahead of suppressing androgen receptor transcription, G1 arrest, and apoptosis of prostate cancer cells

Abstract

Background

The novel selenium-aspirin compound AS-10 was recently reported by us with a cancer cell killing potency 3 orders of magnitude greater than aspirin in pancreatic cancer cell lines with caspase mediated apoptosis and a reasonable selectivity against maligant cells. Although we also observed its cytocidal activity against PC-3 and DU145 androgen receptor (AR)-negative and P53-null/mutant aggressive human prostate cancer (PCa) cell lines in NCI-60 screen, the potential involvement and targeting of AR and P53 pathways that are intact in early stage prostate carcinogenesis has not been examined, nor its primary molecular signaling after exposure.

Methods

Human LNCaP PCa cells with functional AR and intact P53 were used to examine their cell cycle and cell fate responses to AS-10 exposure and upstream molecular signaling events including histone acetylation as a known aspirin effect. The AR-positive 22Rv1 human PCa cells were used to validate key findings.

Results

In addition to confirming AS-10’s superior cytocidal potency than aspirin against all 4 PCa cell lines, we report a rapid (within 5 minutes) promotion of histone acetylation several hours ahead of the suppression of AR and prostate specific antigen (PSA, coded by KLK3 gene) in LNCaP and 22Rv1 cells. AS-10 decreased AR and KLK3 mRNA levels without impacting preexisting AR protein degradation or nuclear translocation in LNCaP cells. Sustained exposure to AS-10 arrested cells predominantly in G₁, and induced caspase-mediated apoptosis without necrosis. The death induced by AS-10 in LNCaP cells was attenuated by non-transcriptional activation of P53 protein or JNK cellular stress signaling and was mitigated modestly by glutathione-boosting antioxidant N-acetylcysteine. AS-10 synergized with histone deacetylase inhibitor SAHA to suppress AR/PSA abundance and kill LNCaP cells. RNA-seq confirmed AR suppression at the transcriptional level and suggested multiple oncogene, cyclin and CDK/CKI transcriptional actions to contribute to the cellular consequences.

Conclusions

AS-10 promotes histone acetylation as its probable primary mechanism of action to induce PCa cell cycle arrest and apoptosis, regardless of AR and P53 status. Nevertheless, the inhibition of AR signaling through mechanisms distinct from canonical AR antagonists may hold promise for combinatorial use with androgen deprivation therapy regimens or AR-axis targeting drugs.

1. Introduction

Prostate cancer (PCa) is the second leading cause of cancer mortality of men in the US, with an estimated diagnosis frequency of 1 in every 9 men in their life time ¹. The majority of PCa is characterized as adenocarcinomas. Androgen receptor (AR) signaling is crucial for the growth and survival of PCa cells in the cancer patients, even in the advanced castration resistant, heavily treated late stages ². Androgen deprivation therapy (ADT) has remained the standard of care for decades for recurrent cancer after localized surgery and/or radiation therapies ³. Next-generation AR targeting drugs, such as enzalutamide (Xtandi) and apalutamide (Erleada), and androgen synthesis blocker abiraterone acetate (Zytiga), and the microtubule targeting taxane drugs are now used in tandem in managing recurrent metastatic castration-resistant prostate cancer (mCRPC) patients ^{4, 5}. The latest FDA approved DNA repair inhibitor drugs for mCRPC applies to a small subset of cases with DNA repair defects ⁶.These hormonal and chemo therapies do not cure yet cause severe adverse side effects for the patients. The FDA-approved autologous cell-based immunotherapy, Sipuleucel-T, has shown modest survival benefit, however, the treatment remains highly patient individualized and cost prohibitive ⁷. Checkpoint inhibitor-based immunotherapy has so far been ineffective due to the “immune cold” nature of the PCa. Therefore, there is ample opportunity for novel agents with mechanisms of action distinct from current approved modalities that can be deployed in the primary prevention space to decrease the risk of developing PCa altogether, in the neo-adjuvant setting to reduce tumor burden before prostatectomy to improve surgical cure or in the adjuvant secondary interception context for reducing recurrent PCa following treatments of the localized malignancy.

Regular use of non-steroidal anti-inflammatory drugs (NSAIDs) has been linked to reduced cancer risk. Aspirin (acetyl salicylic acid, ASA) has been studied extensively in many clinical studies, supporting risk reduction of colorectal cancer (CRC) ⁸. Aspirin is FDA-approved for CRC chemoprevention. Because inflammation is known to contribute to PCa progression, there is plausibility for application of aspirin for risk reduction for PCa ⁹, but the results so far have been a mixed bag ^10–12. Nevertheless, a new aspirin trial is anticipated to shed more light on the chemopreventive benefits for PCa and 3 other solid organ cancers ¹³. However, the long-term use of aspirin has been associated with gastrointestinal bleeding side effects ¹⁴. Therefore, efforts are being made to enhance the selective cancer killing activity of NSAIDs and to improve the gastrointestinal safety margin.

Aiming at enhancing the cancer cytocidal activities of NSAIDs by incorporation of selenium (Se), Sharma lab synthesized AS-10, a cyclic selenazolidine ring substituted by two aspirin moieties (Fig. 1A)¹⁵. AS-10 was effective at inhibiting growth of different cancer cell lines including pancreatic ductal carcinoma cells (PDCA)¹⁵ and DU145 and PC-3 PCa cells in NCI-60 cell line screening panel (Fig S1).

Fig. 1.

AS-10 potently inhibited proliferation and survival of PCa cell lines of different androgen receptor status and its acetyl moeities were essential for the potency. (A) AS-10 chemical structure. (B) MTT viability assay for IC₅₀. LNCaP, DU 145 and PC3 cells were treated for 48 h with increasing concentrations of AS-10. IC₅₀ values were generated by non-linear regression using variable slope equation in Graphpad Prism software. Data were presented as mean ± Sem (n=3 independent expts, each with triplcate wells). (C) Immuno (Western) blot detection of cleaved PARP in whole cell lysate after PCa cell lines were treated with increasing concentrations of AS-10 for 24 h. β-actin was probed as a loading control. (D) Densitometric analysis of cPARP normalized to β-actin using Image J software. Data were represented mean ± SD, ** (p<0.003), *** (p<0.0001). (E) Chemical structure of AS-171 (an AS-10 analog without acetyl groups). (F) MTT viability assay for AS-171 exposed LNCaP, DU 145 and PC3 cells after 48h treatment with increasing concentrations (max 50 μM). IC₅₀ values were not reached.

Extending on the PDCA work with AS-10¹⁵, we report here, for the first time, structural activity relationship (SAR) comparison of AS-10 vs. its analog compound AS-171 without the acetyl moieties on PCa cell death responses for the crucial requirement of the acetyl groups. Besides DU145 and PC3 cells for confirming the growth suppression and apoptosis activity of AS-10 for robustness and generalizability, we selected LNCaP cells with functional AR, wild type P53 and inactive PTEN (high AKT) as the workhorse cell line for the in-depth studies in the current work to elucidate potential molecular mechanisms of action and cellular processes. The rationales for choosing the LNCaP cell line as target cells rest in the facts that AR signaling is crucial for all stages of PCa cell survival and proliferation in patients; that precancerous prostatic lesions and early stage PCa retain wild type P53; and that PTEN loss and AKT-mTOR axis activation is rather frequent and early in PCa development ^16–18. Another AR-positive PCa cell line CWR22Rv1 was used to validate key findings.

2. Materials and Methods

2.1. Selenium compounds, Antibodies and other reagents:

AS-10 was synthesized as reported recently¹⁵ by the reaction of O-acetylsalicyloyl chloride with 1,3-selenazolidin-2-imine hydrobromide in methylene chloride in the presence of triethylamine. AS-171 was prepared following a similar synthetic strategy by using benzoyl chloride in place of O-acetyl-salicyloyl chloride. The compounds were purified by silica gel column chromatography and characterized on the basis of NMR and MS spectroscopy.

Antibodies were ordered from the following commercial sources: Cell signaling for cPARP (9541L), β-Actin (3700s), β-Tubulin (2128), p-H2A.X (9718s), pAKT (4060s), AKT (9272s), cMyc (5605), P21Cip1 (2947s), P27Kip1(3686s), p-P53 (9284L), P53 (2527), Acetyl H3 (8173), H3 (4499), SirT1 (8469), SirT6 (12486), SirT3 (2627), Acetyl lysine (9411) and JNK (9252); Santa Cruz Biotechnology for Bax (sc-493) and MDM2 (sc-965); BD BioSciences for AR (554225); Dako for PSA (A0562). Chemicals were ordered from the following sources: Sigma-Aldrich for propidium iodide (P4170-10MG), ribonuclease A (RNase A, R6513-10MG) and N-Acetyl-L-cysteine (A2750-10G); Millipore for cycloheximide (239764-100MG).

2.2. Cell Culture:

LNCaP, PC3, CWR22RV1 and DU145 cell lines were purchased from ATCC. The LNCaP-AR-luc subline with retroviral integrated expression of wild type human AR coding cDNA under SV40 early promoter was generously provided by laboratory of Dr. Charles Sawyers (Memorial Sloan Kettering Cancer Center)¹⁹. This subline is more aggressive for tumor induction in immunodeficient mice and expresses only ~1% level of PSA as the parental LNCaP cells. They were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) at 37°C and 5% CO2.

2.3. Cell viability assay:

MTT assay for mitochondrial redox metabolism was used for estimating the cell viability/number after AS-10 treatment at different concentrations as reported earlier¹⁵. In brief, LNCaP cells (7500 cells/well), PC3 (6000 cells/well), and DU145 (6000 cells/well) were seeded into 96-well plate and allowed to adhere (PC-3 and DU145 cells for overnight, while LNCaP cells for 48 h) before treatment. Cells were treated with AS-10 (in triplicates) for 48 h at 37 °C and MTT solution (20 μL of 5.0 mg/mL solution) was added to each well for 2.5 h prior to termination and incubated at 37 °C before spectrophotometric readings¹⁵. IC₅₀ values were obtained via non-linear regression analysis using Graphpad Prism software.

2.4. Western immunoblot:

Whole-cell lysates of AS-10 treated PCa cells were prepared and blotted as before for pancreatic cancer cells¹⁵. When required, nuclear vs. cytosolic fractions of treated cells were prepared and blotted as reported before¹⁵.

2.5. Cell cycle analysis:

Flow cytometry-based protocol as described earlier¹⁵ was used to enumerate cells in different stages of cell cycle ²⁰. For asynchronous cell exposure, they were treated with AS-10 for different durations. For synchronized cell cycle analysis, the cells were serum-starved for 48 h and then treated with AS-10 in media with 10% FBS for different durations.

2.6. Annexin V apoptosis assay:

Different stages of apoptosis in AS-10 treated PCa cells were evaluated via Muse Annexin V & Dead Cell kit (Millipore, Catalog No, MCH100105), using Muse Cell Analyzer (EMD Millipore, Billerica, MA, USA) as reported earlier¹⁵. Both floating and adherent cells were collected together and analyzed per manufacturer’s protocol.

2.7. Caspase 3/7 activity assay:

Muse Caspase-3/7 Assay kit (EMD Millipore) was used as reported earlier ¹⁵ to detect caspase 3/7 activation in AS-10 treated LNCaP cells. Both floating and adherent cells were collected together and analyzed per manufacturer’s protocol.

2.8. AR degradation assay with cycloheximide:

To examine whether AS-10 promoted AR protein degradation, LNCaP cells were treated with either AS-10 (5 μM) or DMSO in the presence of protein translation inhibitor cycloheximide (15 μg/mL) for 3, 6 and 12 h. AR level was monitored using Western blot.

2.9. Androgen receptor cytosol to nuclear translocation:

To examine the effect of AS-10 on AR nuclear translocation, LNCaP cells were cultured in phenol red–free medium supplemented with 5% charcoal stripped FBS for 48 h. The cells were pretreated with 5 μM AS-10 for 4 h and were stimulated with 1 nM mibolerone (a stable androgen analog) for 2 h in the continued presence of AS-10. Nuclear and cytosolic fractions were prepared as before for pancreatic cancer cells for Western blot¹⁵.

2.10. ROS measurements:

2.10.1. DCFDA staining for ROS detection:

Intracellular ROS generation was measured using 6-carboxy-2, 7-dichlorodihydrofluorescein diacetate (DCFDA) staining method as previously described¹⁵. DCFDA preferentially reacts with hydrogen peroxide and hydroxyl radicals.

2.10.2. Muse® Oxidative Stress Kit:

The superoxide levels in LNCaP were measured using the Muse Oxidative Stress Kit (EMD Millipore) as per the manufacturer’s protocol. Hydrogen peroxide was used to treat cells as positive oxidative stress control.

2.11. qRT-PCR:

LNCaP (1 × 10⁶) cells were seeded per 10-cm dish for 48 h. The cells were treated with DMSO (vehicle) or AS-10 (5 μM) for 3, 6 and 12 h and subjected to Trizol based RNA extraction. RNA extraction and cDNA prep was performed by Penn State College of Medicine Genome Sciences Core facility. Resulting cDNA was used to perform the qRT-PCR using TAQman primers. AR and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA were quantified using TaqMan Assay Kits (Applied Biosystems) according to the manufacturer’s instructions. TaqMan Assay kits Hs02576345_m1, Hs02786624_g1, and Hs00171172_m1 were used for quantifying KLK3 (PSA gene), GAPDH and AR levels, respectively.

2.12. RNA seq:

RNA samples from LNCaP cells treated with DMSO, 5 μM, and 10 μM AS-10 for 3, 6 and 12 h in duplicate were prepared, quality-checked, and processed for RNA-seq by Genome Sciences Core, as previously reported in our pancreatic cancer paper ¹⁵. Genes with fold change > 2.0 and P<0.05 were considered as differentially expressed for the further bioinformatic analysis. Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, www.qiagen.com/ingenuity) was used to identify top biological networks and canonical central nodes formed among the differentially expressed genes. Expression of other key network molecules were selectively extracted for temporal trend changes.

2.13. Statistical analysis:

Numerical data were presented as mean ± SD of independent experiments (at least 3). Statistical analysis was performed using GraphdPad Prism 8 software (La Jolla, CA, USA). Differences among the treated groups were analyzed by ANOVA. For comparison involving only two groups, the student t-test was used. p < 0.05 was considered statistically significant. All experiments have been repeated independently at least 3 times, except RNA seq profiling was carried out once, in duplicate for each sample.

3. Results

3.1. AS-10 Potently Inhibited Proliferation And Survival Of Prostate Cancer Cell Lines Regardless Of AR Status.

In NCI-60 cell line screening work, AS-10 not only inhibited the growth of, but also killed off AR-negative DU145 and PC3 PCa cells, leading to cell loss to below the cell number at start of exposure (Fig. S1). We cross validated the NCI-60 screening outcomes for these two PCa cell lines using MTT estimation of metabolically live cells (Fig. 1B, 48 h exposure) and by Western immunoblot detection of the cleavage of PARP as an indicator of caspase-mediated apoptosis (Fig. 1C,D, 24 h exposure). As shown in Fig 1B, AS-10 inhibited the growth of all three PCa cell lines with an IC₅₀ in range of 0.5 to 2.5 μM, with DU145 cell line being most sensitive. Fig 1C–D data supported the involvement of an induction of caspase-mediated cell death by AS-10 within the exposure range of 2.5–10 μM in DU145 cells and 5–10 μM in PC3 cells in concentration-dependent manners. For comparison, aspirin has been reported to inhibit the growth of PCa cells, and cancer cells of other organ sites at exposure concentrations in the 5–10 mM range ²¹. In the AR-positive LNCaP PCa cells, as shown in Fig. 1B–D, AS-10 decreased live LNCaP cell number estimated by the MTT colorimetric assay with an IC₅₀ of 2.3 μM after 48 h exposure. Western blot detection of cleavage of PARP showed LNCaP cells with apoptotic sensitivity between the DU145 cells and PC3 cells (Fig. 1C–D). Another AR-positive PCa line 22Rv1 showed a similar apoptotic death response by cPARP detection (Fig. S7A).

The AS-10 analog compound AS-171 without the acetyl moieties (see Fig. 1E for structure) was not active within the exposure range up to 50 μM in all 3 PCa cell lines and with projected IC₅₀ greater than 50 μM (Fig. 1F). Therefore the acetyl groups were crucial for the striking potency of AS-10 for cancer cell growth suppression and death induction.

3.2. AS-10 Induced Predominantly G₁ Arrest and Caspase-mediated Apoptosis wihout Necrosis

To delineate the extent of apoptosis vs. necrosis induced by AS-10, we treated LNCaP cells with increasing concentrations of AS-10 for 24 h and subjected them to annexin V binding assay for externalized phosphatidylserine opsinizing the apoptotic cells (Fig. 2A) and the caspase-3/7 activity assay (Fig. 2B), combined with cell membrane integrity test by 7-ADD dye exclusion. AS-10 treatment increased the total annexin V-positive (Fig. 2A) and caspase-3/7 positive (Fig. 2B) apoptotic metrics (i.e., early stage, staining in lower right sector, plus later stage, staining upper right sector, Fig. S2) in a concentration-dependent manner without an increase in necrotic cells that stained for 7-AAD only (upper left sector, Fig. S2). The apoptosis induction was further supported by increased sub-G₁ fraction in flow cytometry analyses of cell cycle distribution (Fig. 2C) in a time-dependent manner with a threshold AS-10 exposure concentration > 2.5 μM. These independent assays for different aspects of apoptosis supported the ability of AS-10 above certain exposure threshold to induce caspase-mediated apoptosis with minimal necrosis, if any.

Figure 2.

Characterization of apoptosis and cell cycle metrics in LNCaP cells treated with AS-10 via flow cytometry (A-C) and associated signaling proteins of interest by immunoblot (D-F). (A) Annexin V staining and (B) Caspase 3/7 activity assay for LNCaP cells exposed to AS-10 at indicated concentrations for 24 h. Data were presented as the means ± SD. ** (p<0.003), *** (p<0.0001). For graphic data output plots refer to Fig. S2. (C) Cell cycle arrest actions of AS-10 in asynchronous LNCaP cells treated for indicated durations. Representative histograms and quantification of different cell cycle phases at different time points (Mean ± SD). ** (p < 0.03) *** (p < 0.001). Sub-G₁ fraction reflected apoptotic cells. For cell cycle impact of AS-10 on synchronized cells, please see Fig. S3. (D-F) Represntative Western blot images of biomarker proteins related to cell proliferation and cell death in whole cell lysates after AS-10 exposure for 24 h. β-Actin was used as loading control. Densitometric quantification data are in Fig. S4.

To examine the effects of AS-10 on cell cycle progression, we treated asynchronous LNCaP cells with 2.5 and 5 μM of AS-10 for 24, 36 and 48 h and subjected the treated cells to flow cytometeric analysis. By 24 h of AS-10 exposure at 2.5 μM, S phase cells almost disappeared with increases in G₁ and G₂/M cells (Fig. 2C). Over 36 h and 48 h, the abundance of S phase cells recovered somewhat while that for the G₂/M cells held steady and elevated (Fig. 2C). Exposure to AS-10 at 5 μM caused a significant increase of sub-G₁ apoptotic population over time, amounting to nearly half the total cells by 48 h. The appearance of extra S-phase peaks perhaps suggested sub-G₂/M dying cells from 36 h onward (Fig. 2C).

To more specifically probe the cell cycle arresting action of AS-10, we deprived LNCaP cells of FBS for 48 h and then refed the complete medium in the absence or presence of increasing concentrations of AS-10 for 12, 24 or 36 h. Cell cycle distribution analysis showed no cell transited to S phase within the first 12 h (data not shown) and a progression of cells to S by 24 h, reaching G₂/M phase at 36 h (Fig. S3). AS-10 at 2.5 μM exposure slowed G₁/S transition and at 5 μM completely blocked the movement of cells from G₁ to S phase (Fig. S3).

3.3. Survey Of Select Proteins Involved In Apoptosis And Cell Cycle Arrest Induced By AS-10

Western blot analyses of asynchronous LNCaP cells exposed to AS-10 for 24 h showed a concentration-dependent increase of cPARP (Fig. 1C, 2D) in an inverse association with the abundance of c-Myc protein (Fig. 2E), which is a transcription factor controlling early genes in G₁ cell cycle progression and is an over-expressed oncoprotein in a sizable proportion of human PCa. The early transcriptional impact of AS-10 on c-Myc gene expression was latter confirmed by RNAseq profiling.

A decrease in level of pro-survival AKT phosphorylation in AS-10 treated cells coincided with substantial apoptotic cPARP (e.g., 5 μM) and p-H2A.X, which marked DNA double strand breaks (Fig. 2D, E). Temporally, c-PARP appeared in synchrony with the increase in p-H2A.X levels in AS-10 exposed cells (Fig. S5B), indicating a strong association of DNA double strand breaks with caspase-mediated apoptosis known to involve nucleosomal fragmentation. The observed AS-10 induction of p-H2A.X marker prompted us to consider an involvement of P53 genome guardian tumor suppressor pathway in cell cycle arrests and apoptosis. The rationales were rooted in the activation of this best-known DNA damage repair pathway by ionization radiation and DNA damaging drugs; and that some Se compounds, such as sodium selenite ²² and ISC-4 ²³, have been shown to induce ROS, P53 activation and ROS-P53 mediated apoptosis. Likewise, aspirin had been shown to induce P53 signaling ²⁴.

We examined the expression of P53 protein itself and its Ser15 phosphorylation (Fig. 2F), the abundance of P53 canonical targets such as the cell cycle inhibitor protein P21Cip1(Fig. 2F) and Bax which is a mitochondrial targeting proapoptotic protein as well as MDM-2 (ubiquitin ligase), which is both a P53 target gene and a feedback negative regulator of P53 through proteosomal degradation (Fig. S5A), in LNCaP cells after 24 h exposure to increasing concentrations of AS-10. Neither P53 abundance nor that of Bax was changed; however, P53 Ser15 phosphorylation was increased, tracking with cPARP and p-H2A.X (Fig. 2D–F and Fig. S5A). P21Cip1 protein expression showed a concentration-dependent increase starting at 1 μM and peaking at 5 μM and crushed as massive apoptosis occurred at 10 μM (Fig. 2D and 2F). The expression of P27Kip1, which is generally considered P53-independent, followed a similar concentration-dependence pattern (Fig. 2F). The abundance of MDM-2 was decreased at the apoptotic concentrations of AS-10 exposure (Fig. S5A). Temporally, induction of P53 Ser15 phosphorylation trailed behind apoptotic PARP cleavage and p-H2A.X (Fig. S5B). Overall, alterations of P21Cip1, P27Kip1, c-Myc, pAKT, p-P53 and p-H2A.X were associated with AS-10 induction of cell cycle arrest and/or apoptosis. The functional significance of P53/P21 axis in these cellular processes and AR signaling were examined later by siRNA approach.

3.4. AS-10 Decreased AR Signaling Primarily At Endogenous Promoter Transcriptional Level

AS-10 exposure of LNCaP cells for 24 h decreased the cellular abundance of AR and PSA proteins in concentration dependent manners (Fig. 3A,B). Measurement of the secreted PSA in the spent medium showed a similar degree of reduction (Fig. S5C), ruling out the possibility of diminished cellular PSA abundance due to its enhanced discharge. The AR/PSA suppressing effects were also observed in 22Rv1 cells, in which both the full length AR and the well known splice variant AR-V7 (lower band, Fig. S7A) responded to AS-10 expsoure in similar concentration-dependent manners.

Figure 3.

AS-10 decresaed AR axis in LNCaP cells primarily at the transcriptional level. (A) Western blot analysis of AR and PSA in LNCaP cells treated with increasing concetrations of AS-10 for 24 h. β-actin was probed as a loading control. (B) Image J densitometric analysis of AR and PSA normalized to β-actin. * (p < 0.02) **(p < 0.004). (C) Western blot analysis of temporal changes of AR and PSA in whole cell lysates of LNCaP cells treated with AS-10 for different durations. β-actin was probed as a loading control. (D,E) Image J densitometric analysis of AR (D) and PSA (E) normalized to β-actin. *** (p < 0.004). (F) Lack of effect of AS-10 on pre-existing AR protein degradation after new protein synthesis was blocked by cycloheximide (20 μg/mL) in LNCaP cells treated with DMSO or AS-10 for indicated durations. β-actin as a loading control. (G,H) qRT-PCR assay for steady state mRNA levels of AR and PSA (KLK3 gene) in the LNCaP cells treated with AS-10 for different durations. Data for AR and KLK3 (PSA) were normalized to GAPDH mRNA. **(p < 0.009) **** (p < 0.0001).

Temporally, AS-10 (5 μM) suppressed the protein level of AR in LNCaP cells between 6 and 12 h and that of PSA between 3 and 6 h (Fig. 3C–E), ahead of or concurrently with the onset of apoptotic PARP cleavage (Fig. S5B). When new protein synthesis was blocked by cycloheximide, AS-10 treatment of LNCaP cells for up to 12 h did not accelerate the decline of pre-exisitng AR protein than that in DMSO control cells (Fig. 3F), indicating that AS-10 did not enhance post-translational turnover of pre-existing AR protein. Pretreatment of androgen-starved (charcoal striped FBS) LNCaP cells with AS-10 for 6 h did not inhibit the translocation of AR from the cytosol to the nucleus upon stimulation with a stable AR agonist mibolerone (Fig. S6A).

Next, we investigated whether AS-10 affected the steady state mRNA level of both AR and KLK3 (PSA) genes by real time quantitative reverse transcription PCR (qRT-PCR) (Fig. 3G,H). AS-10 treatment decreased the level of AR and KLK3 mRNA in a time dependent manner, as early as 3 h. The impact of AS-10 on KLK3 mRNA was to a greater extent than on AR. These transcriptional impacts were independently confirmed by RNAseq later.

Consistent with a transcription level action of AS-10 on the endogenous AR promoter, LNCaP-AR cells (provided by the laboratory of Dr. Charles Sawyers) with a stably-integrated retroviral expression vector under the SV-40 early promoter to drive exogenous wild type human AR protein expression ¹⁹ underwent apoptotic PARP cleavage with a similar sensitivity as LNCaP cells when exposed to increasing concentrations of AS-10 (Fig. S6B), but the total AR protein (endogenous mutant AR plus exogenous wild type AR) abundance was increased by AS-10 in a strong concentration-dependent manner (Fig. S6B). The viral SV-40 promoter has been shown to be stimulated by increased histone acetylation by the histone deacetylase (HDAC) inhibitors ²⁵. Taken together, the data indicated that AS-10 decreased AR abundance and signaling primarily at the endogenous promoter transcription level without an interference of ligand-activated AR translocation to the nucleus nor enhancing the degradation of pre-existing AR protein.

3.5. AS-10 Promoted Histone Acetylation And Synergized With HDAC Inhibitor SAHA To Induce Apoptosis And Suppress AR/PSA

Aspirin has been known to increase acetylation of histones and non-histone proteins such as AR and P53 protein ^{26, 27}. Given the crucial role of the acetyl moieties of AS-10 for its striking cancer cell growth arrest and death potency (Fig. 1), we investigated the ability of AS-10 to affect histone acetylation using Western blot detection with antibodies against acetylated H3 protein or acetylated lysine residues of histones and other proteins in LNCaP cells. As shown in Fig 4A, the increased H3 acetylation and Lys-acetylation was associated with a reduction of the abundance of sirtuin proteins (with HDAC activities) SirT6 and SirT1, not SirT3 (Fig. 4A,B). The increased H3 and Lys-acetylation signals can be detected as early as 5 minutes of AS-10 treatment in LNCaP cells and plateaued out around 1 h (Fig. 4C). The rapid acetylation action of AS-10 was also observed in 22Rv1 and PC-3 cells (Fig. S7B).

Figure 4.

AS-10 rapidly increased histone acetylation and synergized with histone deactylase inhibitor SAHA to induce apoptosis. (A,B) Western blot analysis of aceylated H3 and SirTs in whole cell lysates of LNCaP cells treated with increaing concetrations of AS-10 for 24 hr. (C) Time couse of histone acetylation induced by 5 μM AS-10. β-actin was probed as a loading control. (D) Chemical structure of SAHA/Varinostat. (E) Western blot analysis of AR, PSA and c-PARP in LNCaP cells treated for 48 h by AS-10 or SAHA each alone or their combination.

The case of histone acetylation promotion by AS-10 treatment was strengthened by a combination with a known HDAC inhibitor, SAHA, which was expected to increase histone lysinyl acetylation. Indeed, when LNCaP cells were treated with sub-apoptotic concentrations of AS-10 (1 and 2.5 μM), or SAHA and their combinations for 48 h (Fig. 5E), AS-10 alone did not induce cell death at either concentration, while SAHA-only exposed cells showed cell death (cPARP) at 10 μM. The combination of AS-10 potentiated the death response of SAHA as indicated by an increase in PARP cleavage that was more than additive. Whereas SAHA alone had a modest suppressing effect on AR abundance at up to 10 μM, the combination exacerbated the AR and PSA suppression effect of AS-10 in more than an additive manner. Together, the data indicated that AS-10 promoted histone acetylation far ahead of reduction of AR/PSA abundance and the other molecular and cellular responses as a probable primary mode of action to lead/contribute to the cell cycle arrest and cell fate consequences. AS-10 was able to promtoe histone acetylation regardless of the AR status as it induced rapid H3 acetylation in AR-negative PC3 cells (Fig. S7B) and Panc-1 pancreatic cancer cells (data not shown).

Figure 5.

RNA seq profiling of AS-10 induced transcriptional changes in LNCaP cells. (A) Top IPA-determined network in LNCaP cells affected by 5 μM AS-10 at 12 h. The network interactions among genes were related to cancer, cell growth, and proliferation. Central node molecules (AR, CDKN1A, KLK3, NKX3-1, and TP53) linking multiple interacting genes were enlarged. Red genes were significantly upregulated; green genes were downregulated; grey genes were not significantly changed. (B) Temproal changes of mRNA expression levels of representative genes from the top networks. Each was normalized to respective DMSO-exposed control (as 1) at each time point.

3.6. RNAseq Confirmed AS-10 Transcription Level Inhibition Of AR Axis, And Suggested Post-transcriptional P53 Regulation and JNK Stress Signaling

Given histone acetylation impacts chromatin packing and gene transcription, we next explored by RNA-Seq of the temporal impact of AS-10 exposure on global gene transcription as well as AR axis and genes associated with P53 and key cell cycle promoting (MYC, Cyclins, CDKs) and CKI genes (CDKNs) and cellular stress family genes. LNCaP cells were exposed to increasing concentrations of AS-10 or DMSO vehicle for 3, 6 and 12 h. Fig. 5A shows a graphic summary of mRNA level changes with AR and key target genes (KLK3 [PSA], NKX3-1) after 12 h AS-10 exposure. The mRNA level for AR gene was decreased as early as 3 h exposure, and progressively over time (Fig. 5B) and the KLK3 and NKX3-1 mRNA levels followed similar suppression patterns. These data independently confirmed the qRT-PCR data (Fig. 3G,H).

TP53 mRNA level was not increased over the duration by AS-10 exposure, and, in fact, decreased steadily over 3 to 12 h exposure (Fig. 5B). The level of CDKN1A mRNA (P21Cip1) was not increased until 12 h of AS-10 exposure (Fig. 5B), when significant apoptotic PARP cleavage was detected (Fig. S5B). In addition, CDKN2B (P15INK4b) was significantly increased, while c-MYC and a number of cyclin genes were suppressed by AS-10, as early as 3 h of exposure. These data are consistent with multiple gene expression changes that could regulate the CDK activities, probably responsible for G₁ arrest. The increased mRNA levels of c-JUN (5 and 10 μM AS-10) and JUND (10 μM AS-10) suggest a potential involvement of JNK stress signaling in apoptosis execution by AS-10.

3.7. A Paradoxical Role Of P53 And JNK In AS-10 Induction Of Cell Death

To interrogate the roles of P53/P21Cip1 signaling in relationship to AR/PSA and apoptosis in AS-10-treated LNCaP cells, we knocked down P53 or P21Cip1 via siRNA apporach and then treated the cells with AS-10 for 24 h. As shown in Fig. 6A, AS-10 treatment of siCDKN1A (P21)-transfected cells increased apoptotic PARP cleavage compared to the response induced by AS-10 in mock-transfected cells. Silencing P53 increased AS-10 induced PARP cleavage even more than knocking down P21Cip1. Knocking down either protein did not affect the suppression of AR and PSA expression by AS-10. These data suggested that post-translational mild activation of P53 (phosphorylation) or one of its best known downstream target P21Cip1 in the AS-10 exposed LNCaP cells protected them against AS-10-induced cell death, rather than serving as a death mediator.

Figure 6.

Paradoxical role of P53 and JNK pathways in AS-10 induction of apoptotic death of LNCaP cells. (A) Western blot analysis of P21Cip1, P53, c-PARP, AR and PSA in LNCaP cells treated with AS-10 in presence or absence of transfected siRNA. LNCaP cells were transfected with 50 nM of si-p21 or si-p53 RNA for 48 hrs and treated with 5 μM of AS-10 for 24 h. β-actin was probed as a loading control. (B) Western blot assessment of cPARP, PSA, AR, and JNK in LNCaP cells treated with AS-10 with or without JNK inhibitor (SP600125) co-treatment for 24 h. β-actin was probed as a loading control. (C) Western blot analysis of p-H2AX, c-PARP, AR, PSA, P53 and P21Cip1 in LNCaP cells with or without pretreatment with NAC for 2 h followed by treatment with DMSO, AS-10 or ISC-4 for 24 h. β-actin was used as a loading control. (D) Reactive oxygen species (ROS) measured using the DCFDA assay in LNCaP cells treated with 5 μM AS-10 (6 h) ± pre-treatment with NAC 10 mM for 2 hr.

To address the role of JNK pathway, we treated LNCaP cells with or without JNK inhibitor SP600125 and AS-10. Hydrogen peroxide was included as a positive ROS stressor control. Whereas the impact of JNK inhibition on AR signaling was quite evident, and additive with AS-10, this pharmacological manipulation failed to avert apoptotic cPARP (Fig.6B). Therefore, the data did not support either P53 or JNK axis as a mediator of the apoptosis signaling induced by AS-10, rather they acted as cellular survival responses to death signaling.

3.8. N-Acetyl Cysteine (NAC) Attenuated AS-10-induced Cell Death

Select Se compounds, such as sodium selenite and ISC-4, have been shown to rapidly induce DCFDA-detectable ROS and DNA damage, followed by P53 activation and P53-mediated apoptosis ²³. Likewise, aspirin had been shown to also induce P53 signaling and acetylation ²⁷. Therefore, we investigated the ability of NAC (thought to increase glutathione synthesis) pretreatment of LNCaP cells to regulate AS-10 mediated cell death biomarkers (cPARP, p-H2A.X), P53/P21Cip1 and AR/PSA axis. As a positive control reference, NAC at 2 mM completely blocked ISC-4 induced P53 activation, apoptosis and AR and PSA suppression (Fig. 6C). We have shown earlier that ISC-4 rapidly induced ROS as the primary signal to trigger P53-mediated apoptosis and suppressed AR signaling²³. Much higher concentrations of NAC were needed for a concentration-dependent attenuation of the apoptotic PARP cleavage in the AS-10-exposed LNCaP cells, with a modest impact on PSA and minimal if any restoration effect on AR abundance. Exposure of LNCaP cells to AS-10 for 3 and 6 h (early time points) did not lead to DCFDA-detectable ROS signal (Fig. 6D and Fig S8). Such data implicated that the apoptosis attenuation activity of higher concentrations of NAC on AS-10 might be through either delayed ROS generation or a mechanism independent of ROS.

4. Discussion

Initial cancer cell cytotoxicity screening done with NCI 60 cancer cell line panel suggested that AS-10 could potentially inhibit the growth and kill numerous solid cancer cell lines including the AR-negative DU145 and PC3 PCa cell lines. Our current work confirmed the cell death inducing activity of AS-10 in these cell lines and extended into the LNCaP and 22Rv1 lines with AR/AR-V7 and wild type P53, features more relevant to prostate cancer biology and early stage disease than those represented by DU145 and PC3 lines. The growth arrest and cytocidal activities of AS-10 were substantial at 10 μM for all 4 cell lines, approximately 3 orders of magnitude more active than aspirin, which has been reported to inhibit growth of PCa, and other cancer cells at exposure concentrations in the 5–10 mM range ^28–30.

It is noteworthy that the apoptotic activity of AS-10 was independent of P53 and AR status. The sensitivity to AS-10 induction of apoptosis could be related to the PTEN/AKT status in that DU145 cells responded most sensitively with wild type PTEN and low AKT, while PC-3 without PTEN and high AKT the least sensitive. While highly AKT active due to defective PTEN in LNCaP cells, the presence of wild type P53 even moderated the apoptosis outcome in the AS-10 exposed cells. In fact, the P53 signaling pathway played a pro-survival role in the AS-10 exposed LNCaP cells (Fig. 6A), in contrast to a typical pro-apoptotic role of P53 when the apoptosis is induced by DNA damaging drugs or ionizing radiation. Such an interpretation would be consistent with a minor role of ROS in AS-10-induced death signaling and therefore only a modest protection offered by NAC at much higher levels (Fig. 6C) than that was sufficient to block ROS driven P53-mediated apoptosis induced by ISC-4 ²³ (Fig. 6C).

In addition to cell cycle arrests at G₁ and a less extent at G₂/M phases and caspase-mediated apoptosis induction by AS-10, salient observations pertinent to the PCa targeting action centered on the ability of AS-10 to exert a rather rapid transcriptional suppression of AR and its canonical target KLK3 (PSA) (qRT-PCR, RNAseq) (Fig. 3G,H and Fig 5) before cell cycle arrest and cell death were reliably detectable (Fig. 2). We ruled out an accelerated AR protein degradation (Fig. 3F) or blockage of AR nuclear translocation by AS-10 when stimulated by an AR agonist analog (Fig. S6A). The specificity for the AR promoter targeting is further supported by even increased AR protein abundance in LNCaP-AR cells derived from LNCaP cells that had stably integrated a retroviral expression vector under the SV-40 early promoter to drive exogenous expression of the wild type human AR gene (Fig. S5B). The SV-40 promoter transcription has been shown to be activated by HDAC inhibitors ²⁵. The ability of AS-10 to down regulate AR mRNA from its own endogenous promoter and down stream signaling offers opportunity for combination with other next-gen AR targeting drugs to seek synergy through different targeting mechanisms to achieve PCa specific response.

Our study with LNCaP cells suggested a prominent role of AS-10 to rapidly promote the acetylation of histones (within 5 minutes) (Fig. 4C) to regulate transcriptional activities of genes important for cell cycle control and cell death (Fig. 5). Our findings were confirmed in other PCa cell lines such as PC3 and 22Rv1 (Fig. S7B) suggesting that AS-10 effects on histone acetlyation may be a primary event regardless of AR and P53 status. The synergy of AS-10 with HADC inhibitor drug SAHA to induce apoptosis was consistent with this supposition (Fig. 4E). Acetylation of AR has been well studied and the HADC inhibitor drug SAHA is known to suppress AR signaling ³¹. At the present time, little is known about how AS-10 promotes histone acetylation. Possibilities include acting as 1) an activator or acetyl substrate donor for histone acetyl-transferases (HAT), 2) a novel HDAC inhibitor or down regulator (suggested by Sirt T6 and SirT1), 3) direct chemical acetylation, or through other mechanisms.

In previous work we have described the lack of growth inhibitory and cytocidal activity by AS-10 on normal MEFs ¹⁵ at exposure concentrations efficacious in the PDAC cell lines, supporting a reasonable degree of selectivity towards malignant cancer cells. The knowledge with cell line models of cancers of different organ sites afford us opportunity to appreciate similarities for AS-10 induced changes as well as unique cellular and molecular features specific to each malignancy. While the G₁ arrest and caspase-mediated apoptosis are common cellualr outcomes among the PCa and Panc-1 cells, the findings in the three PDAC lines, all mutant P53, corroborated independence of AS-10 from P53 to induce apoptosis. The inhibitory activity of AS-10 in Panc-1 cells on NF-kB transcriptional factor signaling illustrated yet an another transcriptional factor inhibitory action. In contrast to the lack of AS-10 impact on androgen-stimulated AR nuclear translocation in LNCaP cells, AS-10 inhibited cytokine-stimulated nuclear translocation of NF-kB in the Panc-1 cells¹⁵.

5. Conclusions

AS-10 treatment rapidly promoted histone acetylation within minutes of exposure, preceding its suppression of AR promoter transcription and AR target genes without inhibiting AR nuclear translocation or increasing AR protein degradation in LNCaP cells. Sustained exposure to AS-10 arrested cell cycle predominantly at G₁/S boundary, and induced caspase-mediated apoptosis which was mitigated modestly by NAC and attenuated by the non-transcriptional activation of P53 that was pro-survival rather than apoptotic. The histone acetylation action synergized with HDAC inhibitor SAHA at killing the LNCaP cells and at suppressing AR/PSA expression. RNA-seq profiling confirmed the suppressing action against AR at transcriptional level and suggested multiple cyclin and CDK/CKI transcriptional actions to contribute to cell cycle arrests. Overall, the data suggest that AS-10 may be a promising drug lead for PCa, either as a single agent or in combination with AR-axis targeted drugs or HDAC inhibitors. Optimization of its delivery formulations and investigation of pharmacokinetics, metabolism, in vivo efficacy, PD targets and toxicity profile are warranted to facilitate translational research efforts from animal models to clinical trials.

Data Availability Statement:

The data that supports the findings of this study are available in the supplementary material of this article