Dietary pterostilbene is a novel MTA1-targeted chemopreventive and therapeutic agent in prostate cancer

Abstract

Overexpression of the epigenetic modifier metastasis-associated protein 1 (MTA1) is associated with aggressive human prostate cancer. The purpose of this study was to determine MTA1- targeted chemopreventive and therapeutic efficacy of pterostilbene, a natural potent analog of resveratrol, in pre-clinical models of prostate cancer. Here, we show that high levels of MTA1 expression in Pten-loss prostate cooperate with key oncogenes, including c-Myc and Akt among others, to promote prostate cancer progression. Loss-of-function studies using human prostate cancer cells indicated direct involvement of MTA1 in inducing inflammation and epithelial-to-mesenchymal transition. Importantly, pharmacological inhibition of MTA1 by pterostilbene resulted in decreased proliferation and angiogenesis and increased apoptosis. This restrained prostatic intraepithelial neoplasia (PIN) formation in prostate-specific Pten heterozygous mice and reduced tumor development and progression in prostate-specific Pten-null mice. Our findings highlight MTA1 as a key upstream regulator of prostate tumorigenesis and cancer progression. More significantly, it offers pre-clinical proof for pterostilbene as a promising lead natural agent for MTA1-targeted chemopreventive and therapeutic strategy to curb prostate cancer.

Keywords: MTA1; chemoprevention; prostate cancer; pterostilbene; therapy.

Figure 1. MTA1 promotes the Pten loss-driven prostate tumorigenesis and cancer progression

(A) Comparison of H & E and IHC of MTA1, p-Akt and PTEN in the prostates from 10-month-old Pten^+/f mice and Cre-negative normal prostate (NP) controls. Scale bars, 100 μm. (B) Immunoblots of MTA1, PTEN, p-Akt, Akt, AR, c-Myc, CyclinD1, TGFβ1, Notch2, Ets2, and Hsp90 and (C) qRT-PCR analysis of MTA1, Akt1, c-Myc, Ets2 and Hsp90 mRNA levels in the prostate tissues from 10-month-old Pten^+/f mice compared to NP controls. (D) Comparison of H & E, MTA1, and p-Akt staining in the prostate tissues from 10-week-old Pten^f/f mice and NP controls. Scale bars, 100 μm. (E) Immunoblots of MTA1, p-Akt, Akt, AR, NF-κB (p65), IL-1β, Hsp90, E-cadherin (E-cad) and Vimentin in the prostate tissues from Pten^f/f mice compared to NP controls, isolated at the ages mentioned. Hsp70 was used as a loading control. qRT-PCR data represent the mean ± SEM (n = 3), *p < 0.05 (two-tailed, two-sample t-test).

Figure 2. MTA1 directly regulates key molecular drivers of tumor promotion

(A) Immunoblots of MTA1, NF-κB (p65), IL-1β, Hsp90, E-cadherin (E-cad), Vimentin, c-Myc, Cyclin D1, Notch2, and Ets2 in LNCaP (left) and DU145 (right) cells expressing (EV) and silenced for MTA1 (shMTA1). (B) qRT-PCR of MTA1, Ets2, Akt1, Notch2, c-Myc, Cyclin D1 and Hsp90 mRNA levels in LNCaP (top) and DU145 (bottom) EV and shMTA1 cells. (C) Immunoblot of MTA1, p-Akt and Akt in LNCaP EV and shMTA1 cells. (D) Immunoblot of p-Akt, Akt, c-Myc and MTA1 and (E) qRT-PCR of MTA1 mRNA levels in PC3M cells treated with vehicle (DMSO) and LY (LY294002). (F) Proposed mechanism involved in Pten loss-induced upregulation of MTA1, exhibiting the MTA1-Akt and MTA1-c-Myc feed-forward signaling loops (blue arrows), putative Akt-MTA1 link (dotted arrow). β-actin was used as a loading control. qRT-PCR data represent the mean ± SEM (n = 3), *p < 0.05; **p < 0.01 (two-tailed, two-sample t-test).

Figure 3. GEO analyses for correlation of MTA1 with PTEN, AKT1 and AR

GSE41967 study of human prostate tissues (n = 639) [38] was used. Scatter plot depicting (A) strong negative correlation between MTA1 and PTEN (r = −0.349, whole cohort), which becomes stronger with increased Gleason score (r = −0.299, Gleason < 7; r = −0.348, Gleason = 7; r = −0.433, Gleason > 7, p < 0.001); (B) positive MTA1 correlation with AKT1 expression (r = 0.499, whole cohort, p < 0.001); and no correlation between MTA1 and AR (r = 0.021, whole cohort. p = 0.592). p values were calculated using two-tailed one-sample z-test for a correlation coefficient.

Figure 4. Pterostilbene reduces PIN formation in Pten^+/f and blocks progression to adenocarcinoma in Pten^f/f mice

(A) Gross anatomy (top) and ex vivo images (middle) of urogenital system (UGS) and dissected prostate lobes (AP_R, right anterior; AP_L, left anterior, and DLV, dorso-latero-ventral) (bottom) of the representative prostates from 10-month-old Pten^+/f mice on phytoestrogen free AIN76 diet (Ctrl-Diet) and 100 mg/kg diet supplementation with pterostilbene (PTER-Diet). (B) Percentage of prostate glands from 10-month-old Pten^+/f mice on Ctrl− (n = 6) and PTER-Diet (n = 7) involved in high grade mouse PIN (mPIN). p < 0.001 (Fisher's exact test). (C) Comparison of H & E prostate histology and PTEN staining in representative 10-month-old mice with NP and Pten^+/f mice on Ctrl− and PTER-Diet. Scale bars, 100 μm. (D) Gross anatomy of the representative UGS from 10-week-old (top) and 33-week-old (middle) Pten^f/f mice treated with vehicle (DMSO) and 10 mg/kg bw PTER. Representative images of dissected prostate lobes of Pten^f/f mice (bottom). (E) Comparison of UGS weights of vehicle or PTER treated Pten^f/f mice, isolated at the indicated ages (n = 3/group). *p < 0.05 (two-tailed, two-sample t-test). (F) Incidence of mPIN, pre-invasive and invasive adenocarcinoma (AC) in Pten^f/f mice treated with vehicle (n = 19) and PTER (n = 18). p < 0.01 (Fisher's exact test). (G) Comparison of H & E, smooth muscle actin (SMA) and cytokeratin 8 (CK8) staining from representative 6-, 10- and 25-week-old mice with NP and vehicle or PTER treated Pten^f/f mice. Arrows indicate loss of SMA staining and CK8 positive luminal cells in the stroma of vehicle treated Pten^f/f mice as signs of invasiveness. Scale bars, 100 μm.

Figure 5. Inhibition of MTA1 and its associated signaling by pterostilbene (PTER) in Pten^+/f mice

(A) Immunoblots of MTA1, PTEN, p-Akt, Akt, AR, c-Myc, CyclinD1, TGFβ1, Notch2, Ets2, and Hsp90 of prostate tissues from representative 10-month-old Pten^+/f mice on Ctrl- and PTER-Diet. Hsp70 was a loading control. (B) Comparison of MTA1, p-Akt and AR IHC staining of the prostate sections from representative 10-month-old Pten^+/f mice on Ctrl- and PTER-Diet. NP, normal prostate. Scale bars, 100 μm. (C) qRT-PCR of PTEN and MTA1 mRNA levels in prostate tissues from 10-month-old Pten^+/f mice on Ctrl- and PTER-Diet. Data are mean ± SEM (n = 3), *p < 0.05; **p < 0.01 (two-tailed, two-sample t-test). (D) Comparative analysis of MTA1 binding in the prostate tissues of Pten^+/f mice on Ctrl- and PTER-Diet. Representative MTA1 ChIP-Seq tracks for Pten, Akt1, c-Myc, CyclinD1, Notch2, Ets2 and Hsp90 gene loci at 10 kb resolution are shown. (E) Quantitation of MTA1, p-Akt/Akt and AR expression in prostate lobes of 10 month-old Pten^+/f mice on Ctrl- and PTER-Diet (see F). Data represent the mean ± SEM (n = 3), *p < 0.05; **p < 0.01 (two-tailed, two-sample t-test). (F) Immunoblots of MTA1, p-Akt, Akt and AR in the dissected prostatic lobes (AP_R, right anterior; AP_L, left anterior, and DLV, dorso-latero-ventral) from 10-month-old Pten^+/f mice on Ctrl- and PTER-Diet. Hsp70 was used as a loading control.

Figure 6. Inhibition of MTA1 and its associated signaling by pterostilbene (PTER) in Pten^f/f mice

(A) Immunoblots of MTA1, p-Akt, Akt, AR, IL-1β, Hsp90 and E-cadherin (E-cad) in the prostate tissues of vehicle and PTER treated Pten^f/f mice, isolated at indicated ages. Hsp70 was a loading control. (B) Comparative analysis of MTA1 binding in the prostate tissues of Pten^+/f mice on Ctrl- and PTER-Diet. Representative MTA1 ChIP-Seq tracks for IL-1β, E-cadherin and Vimentin gene loci at 10 kb resolution are shown. (C) Comparison of IHC staining of MTA1, p-Akt and AR in the prostate sections with carcinoma lesions from representative 10-, 25- and 33-week old vehicle or PTER treated Pten^f/f mice and NP controls. Scale bars, 100 μm. (D) Immunoblots of MTA1, p-Akt, Akt and AR in the dissected prostate lobes from vehicle or PTER-treated 10-, 15-, 20- and 25-week-old Pten^f/f mice. Hsp70 was used as a loading control. For quantitation of MTA1, p-Akt/Akt and AR expression in prostate lobes of Pten^f/f mice at different ages (n = 3/group) see Supplementary Figure S5.

Figure 7. Pterostilbene significantly inhibits MTA1-dependent cell proliferation and induces MTA1-targeted apoptosis in Pten^+/f mice

(A) Representative Ki-67 (top) and cleaved Caspase-3 (bottom) staining of prostate tissues from 10-month-old Pten^+/f mice on Ctrl- and PTER-Diet. Scale bars, 100 μm (Ki-67) and 50 μm (cleaved Caspase-3). (B) Quantitation of Ki-67 (left) and cleaved Caspase-3 (right) positive cells of prostate tissues from 10-month-old Pten^+/f mice on Ctrl- and PTER-Diet (n = 5/group). (C) Immunoblots of total and cleaved Caspase-3 in prostate tissues of 10-month-old Pten^+/f mice on Ctrl- and PTER-Diet. (D) Immunoblots of p21 and p27 in the prostate tissues of 10-month-old Pten^+/f mice compared to NP controls (left) and mice on Ctrl- and PTER-Diet (right). (E) Comparative analysis of MTA1 binding in the prostate tissues of Pten^+/f mice on Ctrl- and PTER-Diet. Representative ChIP-Seq tracks for p21 and p27 gene loci at 10 kb resolution are shown. (F) Immunoblots of p21 and p27 in LNCaP (top) and DU145 (bottom) cells expressing (EV) and silenced for MTA1 (shMTA1). (G) qRT-PCR of p21 and p27 mRNA levels in cells expressing MTA1 and silenced for MTA1 (shMTA1). Data are mean ± SEM (n = 3), *p < 0.05; **p < 0.01 (two-tailed, two-sample t-test).

Figure 8. Pterostilbene significantly inhibits MTA1-dependent cell proliferation and angiogenesis and induces MTA1-targeted apoptosis in Pten^f/f mice

(A) Representative Ki-67 (top, each panel), CD31 (middle, each panel) and VEGF-C (bottom, each panel) staining of the prostate tissues from Pten^f/f mice treated with vehicle and PTER, at indicated ages. Arrows indicate vessels. Scale bars, 100 μm. (B) Quantitation of Ki-67 (top) and CD31 (bottom) positive cells of prostate tissues from mice treated with vehicle and PTER (n = 5/group). (C) Representative images and (D) Quantitation of cleaved Caspase-3 staining at the indicated ages of vehicle and PTER treated Pten^f/f mice (n = 5/group). Scale bars, 10 μm. Data are mean ± SEM (n = 3), *p < 0.05; **p < 0.01; ***p < 0.001 (two-tailed, two-sample t-test). (E) Immunoblots of MTA1, Ac-p53, p53 and Bak in the prostate tissues from vehicle and PTER treated Pten^f/f mice, isolated at the indicated ages. NP, normal prostate. Hsp70 was used as loading controls from prostate tissues. (F) Densitometry of the Ac-p53/p53 ratio from the representative blot. (G) Comparative analysis of MTA1 binding in the prostate tissues of Pten^+/f mice on Ctrl- and PTER-Diet. Representative MTA1 ChIP-Seq tracks for Vegf-c gene locus at 10 kb resolution are shown.

Figure 9. Schematic representation of MTA1-targeted effects of pterostilbene in prostate cancer

Significantly increased levels of MTA1, a key upstream epigenetic regulator, promote inflammation, tumorigenesis, EMT, angiogenesis, and survival signaling and repress apoptosis. Pterostilbene (PTER) targets MTA1 and MTA1-guided molecular drivers of tumor promotion, thereby blocking the Pten loss-driven prostate tumorigenesis and cancer progression.