Dysregulation of cholesterol homeostasis in human prostate cancer through loss of ABCA1

Abstract

Recent epidemiologic data show that low serum cholesterol level as well as statin use is associated with a decreased risk of developing aggressive or advanced prostate cancer, suggesting a role for cholesterol in aggressive prostate cancer development. Intracellular cholesterol promotes prostate cancer progression as a substrate for de novo androgen synthesis and through regulation of AKT signaling. By conducting next-generation sequencing-based DNA methylome analysis, we have discovered marked hypermethylation at the promoter of the major cellular cholesterol efflux transporter, ABCA1, in LNCaP prostate cancer cells. ABCA1 promoter hypermethylation renders the promoter unresponsive to transactivation and leads to elevated cholesterol levels in LNCaP. ABCA1 promoter hypermethylation is enriched in intermediate- to high-grade prostate cancers and not detectable in benign prostate. Remarkably, ABCA1 downregulation is evident in all prostate cancers examined, and expression levels are inversely correlated with Gleason grade. Our results suggest that cancer-specific ABCA1 hypermethylation and loss of protein expression direct high intracellular cholesterol levels and hence contribute to an environment conducive to tumor progression.

Figure 1. DNA methylation analysis of ABCA1 5’ regulatory sequences in prostate cell lines

(A) UCSC genome browser snapshot displaying the DNA methylation sequencing signals in PrEC, LNCaP, and DU 145 cells at the ABCA1 promoter region (UCSC Hg18, chr9:106,728,482–106,730,800). (B) Bisulfite sequencing validation in PrEC and LNCaP cells. Black circles represent methylated CpG sites, and white circles represent unmethylated CpG sites. (C) Methylation-specific PCR (MSP) results in PrEC and LNCaP cells either mock treated (DMSO), or treated with 10µM T0901317 for 24 hours (T), 5 µM 5-aza-2’-deoxycytidine for 7 days (A), or a combination of 5-aza-2’-deoxycytidine and T0901317 (A+T).

Figure 2. Effects of promoter DNA methylation on basal expression and inducibility of ABCA1

(A) Relative mRNA expression of ABCA1 in prostate cells. LNCaP cells were treated identically as in Figure 1. Data are represented as mean ± SEM from triplicate experiments and * indicates p < 0.05. (B) Western blot analysis of ABCA1 and β actin (ACTB) in LNCaP and DU 145 cells. LNCaP cells were treated identically as in Figure 1. (C) Luciferase reporter assay of the unmethylated and the methylated ABCA1 promoter in DU 145 cells. The cells containing the indicated reporter construct were either mock treated (DMSO) or treated with 10µM T0901317 (T) or 5 µM 5-aza-2’-deoxycytidine (A) for 24 hours. Data are represented as mean ± SEM from triplicate experiments. All pair-wise comparisons were statistically significant (p < 0.05) except for between the two methylated treatment groups.

Figure 3. Functional analysis of ABCA1 expression in prostate cancer cells

(A) Representative fields of filipin staining for LNCaP and DU145. (B) Total cellular cholesterol content for LNCaP and DU 145 cells. LNCaP cells were treated identically as in Figure 1 while DU 145 cells were either mock treated (DMSO) or treated with 10 µM T0901317. (C–F) Cellular cholesterol efflux to APOA1 and HDL in LNCaP (C and D respectively) and DU 145 (E and F respectively) cells. Data are represented as mean ± SEM from triplicate experiments, and * indicates p < 0.05.

Figure 4. Analysis of ABCA1 promoter methylation and expression in radical prostatectomy specimens

(A) Methylation-specific PCR (MSP) was performed using MSP1 primer set (Supplementary Table S1) on bisulfite converted genomic DNA extracted from benign prostates (1–9), GS 6 (10–16), GS 7 (17–26), and GS 8–10 (27–42) prostate cancer specimens. LNCaP DNA was included as a positive control. Samples 18, 28, and 33 did not yield sufficient bisulfite converted DNA to produce successful PCR results and therefore were excluded from further analysis. (B) ABCA1 immunohistochemistry on benign prostate, GS 6, GS 7, and GS 8 prostate cancers. (C) Box plots of H-scores for ABCA1 staining in benign prostatic tissues (n=8), high grade prostatic intraepithelial neoplasia (HGPIN) (n=27), Gleason pattern 3 (n=13), 4 (n=24), and 5 (n=14) tumors. The box shows the first quartile, median, and third quartile values. The whiskers show the minimum and maximum values. For Gleason pattern 5, outlier values, defined as three times the interquartile range, are present, and the whiskers denote 1.5 times the interquartile range with outliers plotted as individual black circles. The difference among the H-scores of Gleason patterns 3, 4, and 5 is statistically significant (p = 0.0017; Kruskal-Wallis test). (D) Box plots of percentages of cancer cells expressing ABCA1. Graphical representation is identical to Figure 4c. The difference among the percentages of cells staining positive for ABCA1 in Gleason patterns 3, 4, and 5 is statistically significant (p = 0.0013; Kruskal-Wallis test).