Disruption of a Sirt1 Dependent Autophagy Checkpoint in the Prostate Results in Prostatic Intraepithelial Neoplasia Lesion Formation

Abstract

The Sirtuin family of proteins (SIRTs) encode a group of evolutionarily conserved, NAD-dependent histone deacetylases, involved in many biological pathways. SIRT1, the human homolog of the yeast Silent Information Regulator 2 (Sir2) gene, deacetylates histones, p300, p53, and the androgen receptor. Autophagy is required for the degradation of damaged organelles and long-lived proteins, as well as for the development of glands such as the breast and prostate. Herein, homozygous deletion of the Sirt1 gene in mice resulted in prostatic intraepithelial neoplasia (PIN) associated with reduced autophagy. Genome-wide gene expression analysis of Sirt1^-/- prostates demonstrated that endogenous Sirt1 repressed androgen responsive gene expression and induced autophagy in the prostate. Sirt1 induction of autophagy occurred at the level of autophagosome maturation and completion in cultured prostate cancer cells. These studies provide novel evidence for a checkpoint function of Sirt1 in the development of prostatic intraepithelial neoplasia and further highlight a role for SIRT1 as a tumor suppressor in the prostate.

Introduction

The sirtuin family consists of NAD⁺-dependent histone deacetylases (HDACs), conserved from archaeobacteria to eukaryotes, and classified as class III HDACs (1,2). The enzymatic activity of sirtuins is NAD⁺ dependent and the Sirtuin family members (SIRT1-SIRT7), convey diverse functions (2). SIRT1 is the mammalian ortholog of the Sir2 gene, an important regulator of aging in Saccharomyces Cerevisiae, Caenorhabditis Elegans, and Drosophila Melanogaster (3). The role of SIRT1 in cellular growth control is complex and cell-type specific. In vitro, SIRT1 inhibits p53, Bax, Ku70, FOXO, and the retinoblastoma (Rb) protein (4,5), which may be anticipated to promote cell proliferation. Reduction in SIRT1 activity induced cell growth arrest and apoptosis in breast, lung, and colon cancer cells (5-7). Inhibition of SIRT1 with Sirtinol induced growth arrest in MCF7 and H1299 cells (5). In contrast, several in vivo studies suggest that SIRT1 may function as a tumor suppressor as Sirt1^-/- mice show an impaired DNA damage response, evident by increased genomic instability and tumorigenesis (8). Additional studies from Sirt1^-/- and transgenic mice are consistent with a role for Sirt1 in tumor suppression as Sirt1 was shown to suppress intestinal tumorigenesis and colon cancer (9).

Androgen receptor (AR) expression and activity are key determinants of prostate cancer onset and progression. Of potential importance to prostate biology and function, SIRT1 deacetylates the histone acetyltransferase (HAT) p300 and the AR. SIRT1 transduction of AR-expressing prostate cancer cells (LNCaP) decreased cell proliferation and blocked contact-independent growth (10). The AR colocalizes with SIRT1 in a nuclear sub-compartment, where SIRT1 binds to and deacetylates the AR, thereby inhibiting its activity (1,11). Histone acetyltransferases (p300, CBP/PCAF, Tip60) acetylate the AR at a conserved motif in response to dihydroxytestosterone (DHT), thereby stimulating the growth and anti-apoptotic functions of the AR. The AR lysine residues targeted by acetylation (K630, K632, K633) are well conserved between species and serve as substrates for SIRT1-mediated deacetylation (12,13), resulting in inhibition of ligand-induced AR activity (14).

Prostate cancer proceeds via morphological changes transitioning from the development of prostatic intraepithelial neoplasia (PIN), invasive adenocarcinoma, and metastasis. The pathognomonic features of PIN include changes in nuclear morphology such as enlargement of the nucleus and nucleolus. Molecular genetic dissection in the mouse demonstrated that forced expression of c-Myc (15), Akt, or deletion of Pten (16) leads to PIN and/or prostate adenocarcinoma.

The role of SIRT1 in regulating prostate gland formation and androgen signaling in vivo was previously unknown. SIRT1 is expressed in several cell types in the prostate gland including basal cells, luminal cells, and stromal cells. Given the evidence that SIRT1 functions as a tissue-specific regulator of cellular growth and that SIRT1 inhibits tumor cell line growth in nude mice, we sought to determine the role of endogenous Sirt1 in regulating prostate gland development. Genome-wide expression profiling of Sirt1^-/- mice prostates and their littermate controls identified a molecular, genetic signature regulated by endogenous Sirt1. This signature highlights the ability of Sirt1 to inhibit androgen signaling and apoptosis in the prostate, while promoting autophagy. The Sirt1^-/- prostates demonstrated epithelial hyperplasia and prostatic intraepithelial neoplasia (PIN) suggesting that Sirt1 promotes autophagy and inhibits prostate epithelial cell proliferation in vivo.

Materials and Methods

Gross Anatomical Analysis

Sirt1^-/- mice and littermate controls aged 2-3 months were euthanized by CO₂ asphyxiation and subsequently weighed and measured for both mass and length. Animals were dissected with the following organs being removed: ventrodorsolateral prostate, anterior prostate, seminal vesicles, testes, epididymus, vas deferens, kidneys, liver, spleen, and pancreas. Portions of each organ were fixed in 4% paraformaldehyde to be used for sectioning and Hematoxylin and Eosin (H&E) staining. Ki67 staining was performed as previously described (17).

Transgenic Mice and Genotyping

All Sirt1 transgenic mice used were provided by Dr. Michael McBurney and have been previously described (18). The appropriate institutional committee approved protocols were employed when working with these mice. RT-PCR analysis to confirm the genotype of the mice used was conducted through genotyping with oligonucleotides directed toward exon 5 of Sirt1, generating a 132 bp amplimer (nucleotide sequences were forward: 5′ AATATATCCCGGACAGTTCCAGCC 3′ and reverse: 5′ ATCCTTTGGATTCCTGCAACCTGC 3′).

Microarray and Pathway Analysis

5μg of total RNA derived from Sirt1^+/+ and Sirt1^-/- ventrodorsolateral prostates (Trizol Reagent, Invitrogen, Carlsbad, CA) was reverse transcribed using Superscript III First-Strand Synthesis System (Invitrogen) using an HPLC purified T7-dT24 primer (Sigma Genosys, St. Louis, MO) containing the T7 polymerase promoter sequence. Single stranded cDNA was converted to double stranded cDNA using DNA polymerase I (Promega, Madison, WI) and purified by cDNA spin column purification using a GeneChip Sample Cleanup Module (Affymetrix, Santa Clara, CA). Double stranded cDNA was used as a template to generate biotinylated cDNA using Bioarray HighYield RNA Transcription Labeling Kit (Enzo, New York, NY). cRNA (15μg) was fractionated to produce fragments of between 35-200 bp using 5× fragmentation buffer provided in the Cleanup Module. Samples were hybridized to a mouse 430A 2.0 microarray (Affymetrix) representing approximately 14,000 well-characterized genes. Hybridization and washing steps were carried out in accordance with Affymetrix protocols for eukaryotic arrays. Arrays were scanned at 570nm with an Affymetrix confocal scanner.

Analysis of arrays was performed using the R statistics package (19) and the limma library (20) of the Bioconductor software package. Arrays were normalized using robust multiarray analysis (RMA), while a P-value of 0.05 and a fold change of at least 2 fold was applied as criteria for statistically differentially expressed genes. Gene Ontology (GO) analysis of gene functions was analyzed using “Webgestalt” (http://bioinfo.vanderbilt.edu/webgestalt). Genes were clustered using hierarchical clustering with “complete” agglomeration. Each cluster was further analyzed based on the known function of the genes contained in the cluster. Expression profiles are displayed using Treeview (21). Pathway analysis was performed using the following databases: Kyoto Encyclopedia of Genes and Genomes (KEGG), BioCarta, Gene Set Enrichment Analysis (GSEA), Analysis of Sample Set Enrichment Scores (ASSESS) and the Database for Annotation, Visualization and Integrated Discovery (DAVID).

LC3 Immunofluorescence

LNCaP cells (American Type Culture Collection, Manassas, VA) transfected with either pcDNA3 or pcDNA3-SIRT1 (Lipofectin, Invitrogen, Carlsbad, CA) were plated on Poly-L-Lysine (Sigma, St. Louis, MO)-coated chamber slides (Nunc, Rochester, NY) in either complete media (CM: RPMI, 10% FBS, 1% Penicillin-Streptomycin, 1% L-Glutamine) or serum free media (SFM: RPMI, 1% Penicillin-Streptomycin, 1% L-Glutamine). After 24hrs, half of the cells plated with SFM were treated with 10nM DHT (SFMA) for 24hrs. All other cells were treated with vehicle (ethanol) control. After DHT treatment, media was removed and cells were washed twice with PBS. Next, cells were fixed with 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) for 15min at room temperature, followed by a 5min PBS wash and a 10min 1× PBS + 0.1% Triton wash. Cells were then incubated at room temperature for 1hr in 1% PBS-BSA+0.05% Triton. Afterward, cells were incubated at 4°C overnight with gentle shaking in primary antibody (LC3-II, Abgent, San Diego, CA, 1:50) prepared in PBS+0.05% Triton. Next, cells were washed with PBS-BSA three times, followed by treatment with secondary antibody in PBS+0.05% Triton (Alexa-Fluor 488 goat anti-rabbit, Invitrogen, 1:1000) for 1hr at room temperature. Cells were then washed with PBS three times, mounted (Prolong Gold antifade reagent with DAPI, Invitrogen), coverslipped, and examined by confocal microscopy (Zeiss LSM 510 META Confocal Microscope System, Germany).

Western Blot and siRNA Knockdown

Western blotting was performed as previously described (10) using 50μg of whole cell lysates prepared following cell treatments. Antibodies used included anti-LC3 (Novus, Littleton, CO, 1:500), anti-SIRT1 (Santa Cruz: H-300, Santa Cruz, CA, 1:1000), anti-GAPDH (Santa Cruz: FL-335, 1:1000), anti-ATG4c (Abcam: ab-75056, Cambridge, MA, 1:500), anti-Beclin1 (Santa Cruz: H-300, 1:500), anti-AR (Millipore: PG-21, Billerica, MA, 1:1000), and anti-GDIα (22). Knockdown experiments were performed as previously described (23), (24). siRNAs used included SIRT1 siRNA (Qiagen: SI00098434, Valencia, CA) and a validated, non-silencing control siRNA (Qiagen: 10227281). Quantitation and normalization of immunoblotting results was performed as previously described (25).

Results

Sirt1^-/- Mice Develop Prostatic Intraepithelial Neoplasia

To examine the role of Sirt1 in the development of androgen-responsive tissues such as the prostate, Sirt1^-/- mice were examined. Genotyping was conducted as previously described (18) and analysis of Sirt1 mRNA abundance conducted by RT-PCR of prostate tissue confirmed no detectable Sirt1 mRNA in Sirt1^-/- prostates (Fig. 1A). Morphological analysis of Sirt1^+/+and Sirt1^-/- mice showed that average body length and weight were decreased in the Sirt1^-/- mice (Fig. 1B,C). Individual organ weights were measured and normalized to the animal's body weight (Fig. 1D). The ventrodorsolateral (VDL) and anterior prostates and the seminal vesicles were significantly decreased in size when normalized to total body weight (Fig. 1D). In contrast, the weight of the intestine and pancreas was unchanged (Fig. S1A). The epididymal fat pad was increased in size, with a modest, but significant reduction in the normalized weight of the liver and kidney in the Sirt1^-/- mice (Fig. S1A). Collectively, these studies demonstrate that endogenous Sirt1 plays a role in the development of androgen-responsive tissues.

Figure 1. Sirt1 Deletion Alters Androgen-Responsive Tissue Development.

(A) RT-PCR using ventrodorsolateral prostate RNA from Sirt1^+/+ and Sirt1^-/- mice. (B) Mean body length, (C) mass, and (D) select tissue weights normalized to total mass (see also Fig. S1A). (E) Hematoxylin and Eosin (H&E) staining of prostate glands from Sirt1^+/+and Sirt1^-/- mice (see also Fid. S1B). (F) Ki67 staining of Sirt1^+/+ and Sirt1^-/- prostates. All data are mean ± standard error of the mean (SEM) and represent n=5/genotype. P values were determined by student's t-test (*P<0.05, **P<0.01, ***P<0.001, † denotes no statistical significance).

Histological analysis of the Sirt1^-/- mice tissues was conducted (Fig. 1E,S1B). Prostates were smaller in Sirt1^-/- mice compared to Sirt1^+/+ littermate controls with a larger stromal layer (Fig. 1D,E). The Sirt1^-/- prostates exhibited a morphological phenotype similar to that commonly observed within prostatic intraepithelial neoplasia (PIN) lesions, accompanied by increased cellularity and nuclear atypia (Fig. 1E), a phenotype confirmed to remain without further disease progression in 7 month old Sirt1^-/- mice (data not shown). Sirt1^-/- prostates showed glandular hyperplasia and an increased Ki67 proliferative index (Fig.1E,F). The epididymis of the Sirt1^-/- mice was hypoplastic, while the seminal vesicles were smaller in size and demonstrated a large stromal layer surrounding the epithelium (Fig. S1B). The seminiferous tubules were dilated with a significant reduction in the number of mature sperm within the Sirt1^-/- mice (Fig. S1B).

Sirt1 Governs Prostate Autophagy, Apoptosis, and Cell Proliferation Signaling Pathways

To examine the gene expression profile regulated by endogenous Sirt1, ventrodorsolateral prostate RNA from Sirt1^+/+ and Sirt1^-/- mice was compared. 498 genes were differentially regulated by Sirt1, with 146 mRNAs altered greater than two-fold (Fig. 2A). The biological pathways associated with this altered gene expression signature analyzed using Biocarta and the Kyoto Encyclopedia of Genes and Genomes (KEGG) demonstrated that endogenous Sirt1 regulates pathways governing the cell cycle, cell adhesion, diabetes, and immune function (Tables S1,S2). Comparisons made with published datasets of androgen-responsive genes identified in prostate cancer cell lines or within the prostate gland (Tables S4A,B), displayed an overlap between Sirt1-regulated and androgen-regulated genes as 12.45% of the genes regulated by Sirt1 were also previously reported to be androgen-responsive. Several genes induced by DHT in vitro correspond to genes repressed by Sirt1 in the prostate and similarly Sirt1 enhanced expression of several genes repressed by DHT. For other genes, directionality of expression changes by DHT and Sirt1 were concordant (Tables S3-S4B). The highest proportion of genes coinciding with our data was found with in vivo data (22/498, 4.42%) rather than cultured prostate cancer cell lines (26). This finding suggests that androgen signaling may differ between the in vivo and cultured cell environment.

Figure 2. Sirt1 Governs Vital Biological Pathways in the Prostate.

(A) Treeview of microarray data generated from Sirt1^+/+ and Sirt1^-/- ventrodorsolateral prostates. (B) Heatmap displaying pathways affected by Sirt1 in the prostate as determined by the Analysis of Sample Set Enrichment Scores (ASSESS) database (see also Fig. S1C, Tables S1-S4). (C) Graphical schematic highlighting Sirt1 affected pathways in the prostate as revealed by the Database for Annotation, Visualization and Integrated Discovery (DAVID). Also illustrated is Sirt1's ability to affect a subset of autophagy related genes (Atg's) involved in autophagosome maturation and completion.

Higher order structural analysis of the expression data was conducted using gene set enrichment analysis (GSEA) and Analysis of Sample Set Enrichment Scores (ASSESS) to identify discrete biological functions associated with cell signaling in our array. GSEA utilizes gene expression levels as a means to establish a set rank list of significantly regulated genes within compared sample populations. ASSESS refers to a measure of enrichment of the gene sets across multiple established and annotated gene sets (27). Supervised clustering was employed and identified a set of biological pathways in which the expression profiles were altered upon deletion of Sirt1. Such pathways included cell motility, chemotaxis, cell adhesion, and cell cycle (Fig. 2B,S1C). Additional pathway and enrichment analysis was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) database. These analyses confirmed similar biological pathways governed by endogenous Sirt1 in the prostate including autophagy, apoptosis, and cell proliferation (Fig. 2C). Sirt1 regulated the expression of Atg4, Atg7, and Atg8 in the prostate. These genes are required for late-stage autophagy including phagophore elongation, maturation, and autophagosome vesicle formation (Fig. 2C). This biological role for Sirt1 in the promotion of autophagy was further confirmed by ASSESS pathway analysis, as a defined set of genes up-regulated following rapamycin treatment and mTOR (mechanistic target of rapamycin) inhibition (known autophagic stimulants) were also promoted by Sirt1 in the prostate (Fig. S1C).

SIRT1 Inhibits AR-Dependent Repression of Autophagy

In view of our finding that Sirt1 induced gene expression pathways associated with autophagy, we conducted tissue culture-based experiments to determine the effects of SIRT1 on autophagy in prostate cancer cells. Formation of autophagosomes can be assayed by the sub-cellular distribution of LC3 (microtubule-associated protein 1 light chain 3 alpha), the mammalian homologue of the Atg8 gene. A hallmark of mammalian autophagy is the conversion of LC3-I to LC3-II via proteolytic cleavage and lipidation. This modification of LC3 is essential for the formation of autophagosomes and for the completion of macroautophagy. Additionally, the LC3-GFP fusion protein can be used as a surrogate measure of autophagy through the formation of “LC3 punctae or dots” (28).

To validate our microarray results, AR-expressing prostate cancer cells (LNCaP) were treated with either control siRNA or siRNA directed against SIRT1. Whole cell lysates were subjected to immunoblotting for SIRT1 and autophagy markers, ATG4c, LC3-I, and LC3-II. Following SIRT1 knockdown, the abundance of these proteins decreased suggesting that SIRT1 promotes autophagy in prostate cancer cells (Fig. 3A). Western blots performed using protein isolated from the ventrodorsolateral prostates of Sirt1^+/+ and Sirt1^-/- mice further confirmed a decrease in the abundance of autophagy markers (Atg4c, Atg6 (Beclin1), and Atg8 (LC3)) and an increase in AR abundance in the Sirt1^-/- prostate (Fig. S3A). Next, LNCaP cells were transfected with either an expression vector encoding SIRT1 or a control vector (pcDNA3). After transfection, cells were cultured in the presence (CM) or absence of serum (SFM, autophagy stimulant) and treated with either vehicle or DHT and subjected to immunofluorescence (IF) for either LC3-II (Fig. 3B-D) or IgG control (Fig. S2A). Our results indicated that serum deprivation induced autophagy when compared with complete media (Fig. 3B,C) and that 10nM DHT (SFMA) inhibited serum withdrawal-induced autophagy (Fig. 3C,D). Furthermore, transfection of cells with SIRT1 enhanced autophagy and abrogated DHT-mediated inhibition of autophagy (Fig. 3B-D). Cells were counted and graded based on abundance of LC3-II positive staining. The number of LC3-II positive cells and severity of autophagic response per cell (defined according to number of autophagosomes present per cell) was increased following SIRT1 transfection, regardless of serum or androgen presence (Fig. 3E,F). SIRT1 Western blot analysis following transfection confirms that the levels of SIRT1 post-transfection fall within physiological levels as previously described (29). Cells were also co-transfected with GFP to confirm transfection efficiency (>80% GFP positive cells post-transfection) (Fig. 3G).

Figure 3. SIRT1 Antagonizes AR-Mediated Inhibition of Autophagy.

(A) Western blot for SIRT1, ATG4c, and LC3 following SIRT1 siRNA treatment in LNCaP cells. All data are representative of triplicate experiments. Statistical significance was determined by Student's t-test (*P<0.05, ***P<0.001, † denotes no statistical significance). (B-D) LNCaP cells transfected with pcDNA3 or SIRT1 and cultured in complete serum media (CM), serum free media (SFM), or serum free media supplemented with 10nM DHT (SFMA) were subjected to LC3-II immunofluorescence (see also Figs. S2A-S2C). (E,F) Quantitation of autophagic cells based on LC3-II positivity (ANOVA, **P<0.01) and autophagic grade (Mann Whitney U, **P<0.01). All data are representative of triplicate experiments. Quantitation represents at least 100 cells counted and scored per treatment. (G) SIRT1 Western blot following co-transfection with pcDNA3 vector control or SIRT1 and GFP (micrographs confirm GFP-positive cells post-transfection).

Immunoblotting confirmed the induction of LC3-II by serum deprivation. SIRT1 transfection of LNCaP cells further enhanced the formation of LC3-II in both complete growth media and in serum-deprived media (Fig. 4A). SIRT1 induced the abundance of the autophagy-related proteins ATG5, ATG6 (Beclin1) and ATG9 (Data Not Shown). The SIRT1 antagonist Sirtinol enhanced DHT-mediated repression of autophagy (Fig. 4B). Collectively, these studies demonstrate that SIRT1 antagonizes DHT-mediated repression of autophagy, an effect that can be abolished by Sirtinol-mediated SIRT1 inhibition.

Figure 4. SIRT1 Promotes Autophagy in Prostate Cancer Cells.

(A,B) LC3 Western blots in LNCaP cells following SIRT1 over-expression or inhibition via Sirtinol (30μM) in various media conditions. (C) Western blot verifying SIRT1 abundance in LNCaP cells transduced with MSCV-IRES-GFP or MSCV-IRES-SIRT1-GFP retroviruses. (D,E) SIRT1 and LC3 dual-immunofluorescence of transduced LNCaPs in the presence or absence of autophagic stimuli (HBSS) (see also Fig. S7). All data are representative of triplicate experiments. Statistical significance was measured by ANOVA (*P<0.05, **P<0.01, ***P<0.0001, † denotes no statistical significance).

The formation of acidic, vesicular organelles (AVOs) in the cell cytoplasm, which can be stained by acridine orange (AO) (22), is an additional measure of autophagy. Punctate, cytoplasmic staining in the CY3 wavelength is indicative of autophagolysosome formation and autophagy. Fluorescence in the FITC channel is indicative of AO staining of DNA, while accumulation of red staining in the nucleolus of cells represents AO staining of RNA. Acridine orange staining demonstrated the presence of AVOs in the absence of serum and in the presence of SIRT1, the formation of which was inhibited by both Sirtinol and/or physiological concentrations of DHT (Figs. S2B,C,S3B,C).

An additional surrogate measure of AVO formation and active autophagy is staining with monodansylcadaverine (MDC) (30). Transfection of LNCaP cells with SIRT1 resulted in an increase in the intensity and abundance of MDC positive staining regardless of serum or androgen presence compared to pcDNA3 vector control, consistent with our prior results. Conversely, similar effects were blocked by treatment with the SIRT1 inhibitor, Sirtinol (Fig. S4A-C).

To examine further the mechanism by which SIRT1 regulates autophagy in prostate cancer cells, LNCaP cells were transduced with either MSCV-IRES-GFP or MSCV-IRES-SIRT1-GFP retroviruses and subsequently sorted by GFP expression (Fig. 4C). Cells were GFP sorted and stained with MDC or AO and analyzed by confocal microscopy. Cells transduced with SIRT1 exhibited heightened AVO formation in both the presence and absence of autophagic stimuli (serum) and androgens (Figs. S4D-F,S5,S6). Co-immunofluorescence for LC3, SIRT1 (Fig. 4D,E), GFP, and IgG controls (Figs. S7A-D) was also performed on these cells in the presence and absence of autophagic stimuli (Hank's Balanced Buffer Solution, HBSS, Gibco/Invitrogen). HBSS was shown to increase the autophagic response as evident by an increase in LC3 positive staining (Fig. 4D,E). This effect appears to be mediated at some level through SIRT1 as HBSS treatment stimulated SIRT1 expression in MSCV-IRES-GFP transduced cells (Fig. 4D). The abundance of LC3 staining and severity of the autophagic response was enhanced in MSCV-IRES-SIRT1-GFP transduced cells (Fig. 4E). Notably, HBSS treatment of these cells stimulated a shift of SIRT1 expression from a primarily nuclear distribution to a more uniform distribution throughout the nucleus and cytosol. This observation could be explained by the ability of HBSS to induce cellular stress, in turn triggering SIRT1 to transfer into the cytosol where it may deacetylate and activate late-stage autophagy related genes to finalize the induction of autophagy and protect the cell from stress.

LNCaP cells stably expressing either GFP (LNCaP-GFP) or the LC3-GFP fusion protein (LNCaP-LC3-GFP) were established. These cells were sequentially transfected with either pcDNA3 or SIRT1, treated with either vehicle or DHT, and analyzed by confocal microscopy for the appearance of GFP-positive, LC3 punctae (Fig. 5A-D). SIRT1 induced the formation of LC3 punctae, which was reversed upon inhibition of SIRT1 with Sirtinol (Figs. 5A-D,6A,B).

Figure 5. SIRT1 Enhances LC3 Cleavage and Autophagy Regardless of Media Conditions.

(A,B) LNCaP cells stably expressing GFP (LNCaP-GFP) or a GFP-tagged, pre-cleaved version of LC3 (LNCaP-LC3-GFP) transfected with SIRT1 or pcDNA3 vector control. (C,D) LNCaP-GFP and LNCaP-LC3-GFP cells transfected with pcDNA3 or SIRT1 in the presence and absence of DHT. Quantitation of triplicate experiments was performed based on autophagic grade and graphed adjacent to micrographs (Mann Whitney U, **P<0.01, ***P<0.001, † denotes no statistical significance).

Figure 6. Chemical Inhibition of SIRT1 Abolishes Autophagy in Prostate Cancer Cells.

(A,B) LNCaP-GFP and LNCaP-LC3-GFP cells cultured in various media conditions in the presence and absence of Sirtinol (30μM) and analyzed by confocal microscopy for LC3 punctae. Statistical significance of triplicate experiments was measured by Mann Whitney U test (***P<0.001, † denotes no statistical significance).

Herein we demonstrate that endogenous Sirt1 inhibits murine prostate epithelial cell proliferation. Sirt1^-/- prostates demonstrate increased Ki67 staining, prostatic hyperplasia, and prostatic intraepithelial neoplasia. Sirt1 promoted autophagy in the prostate, functioning at the level of autophagosome maturation and vesicle formation (Fig. 7).

Figure 7. SIRT1 Promotes Autophagy-Mediated Prostate Development and Health.

(A) Proposed model of SIRT1 function in the prostate as it relates to autophagy and prostate homeostasis.

Discussion

Cancer is characterized by a series of transitions from normal to preneoplastic, invasive, and metastatic disease. The checkpoints to progression of early phenotypic transitions in this process are increasingly well understood. Cellular senescence, the angiogenic switch, and epithelial-mesenchymal transitions may represent barriers to transitions throughout the tumorigenic phenotype. The current studies identify a role for endogenous Sirt1 as a key regulator of normal prostate gland homeostasis in vivo, in part through its ability to modulate autophagic signaling pathways. These studies also highlight the repercussion of Sirt1 loss in the prostate gland, as deletion of Sirt1 results in the development of prostatic intraepithelial neoplasia (PIN). Herein, Sirt1^-/- prostates were reduced in size with features of PIN. The features of PIN observed in the Sirt1^-/- mice included cellular hyperplasia, increased Ki67 staining, hyperchromatic nuclei, and prominent nucleoli. These observations and the reduced size of the prostate are consistent with a role for Sirt1 in the regulation of organ development. Previously defined regulators of murine PIN (activation of the Akt pathway, c-myc, or p27^KIP1) were not observed in the Sirt1^-/- prostate, suggesting Sirt1 regulates PIN via an alternative mechanism. Genome-wide microarray and pathway analysis demonstrated that endogenous Sirt1 regulates pathways governing apoptosis, cell proliferation, Ras signaling, androgen signaling and autophagy in the prostate. The finding that Sirt1 inhibits Ras signaling is consistent with recent studies in which SIRT1 inhibited growth of Ras-transformed cells in nude mice (31) and inhibited LNCaP cell proliferation and colony formation (10). Endogenous Sirt1 repressed genes that promote cellular proliferation and cell cycle progression including cdc25, cyclin D2, cyclin D3, and cyclin B2. In addition, Sirt1 repressed several growth factors and pro-proliferative cytokines including CXCL9 and CCL5.

Gene expression analysis further demonstrated that loss of endogenous Sirt1 inhibited autophagy. At a higher level of resolution, our studies demonstrated that SIRT1 antagonized DHT-mediated inhibition of autophagy in the prostate. Autophagy allows for degradation of proteins and organelles (32, 33) and is induced by nutrient withdrawal, rapamycin (inhibition of mTOR signaling), and hormone signaling (PPARγ, ERα/Tamoxifen) (22). Our findings are consistent with prior studies demonstrating that SIRT1 induces autophagy by deacetylating ATG5, 7, and 8 (34) and inhibits AR signaling via deacetylation of the AR (10). Comparisons with previously published studies identified an overlap of 12.45% between genes regulated by endogenous Sirt1 and those targeted by androgens in the prostate gland and in prostate cancer cells. These results are consistent with prior findings that Sirt1 inhibits ligand-dependent AR signaling and gene expression in vitro (10, 35).

Sirt1 deletion and the resulting inhibition of autophagy may have contributed to the reduction in prostate size as autophagy regulates normal gland development. Mutagenesis of the Dictyostelium ortholog of the yeast Atg 5,6,7,8 genes results in aberrant development (36). In Saccharomyces Cerevisiae, autophagy is essential for differentiation and sporulation (37), and in Caenorhabditis Elegans, bec-1 (ATG6/VPS30/Beclin1) is essential for dauer diapause formation (38). Mice defective in autophagy have shown defective tissue development as conditional Atg5 and Atg7 knockout mice demonstrate neuronal and hepatic abnormalities (39, 40). Autophagy has also been implicated in tissue remodeling associated with duct formation in breast and prostate cells in vitro and in 3D matrigel (41). Prostatic 3D cultures also form hollow, acinus-like structures that stain strongly for LC3 (42). Autophagy has pro-survival and pro-death functions (43). A pro-survival function for autophagy is evidenced by a loss of autophagy resulting in death of both C. Elegans and D. Melanogaster (37, 38). In contrast, autophagy can also be pro-death. Knockdown of Atg's inhibits non-apoptotic cell death (44). These studies suggest that physiological levels of autophagy are pro-survival, whereas an excessive level of autophagy promotes autophagic cell death (45).

The role of autophagy in cancer was proposed over 20 years ago (46). Autophagy appears to be essential for tumor suppression as well as for cell survival (47). Autophagy plays a pro-survival function for cancer cells during nutrient deprivation or when apoptotic pathways are compromised, a phenotype often accompanied by inflammation. In contrast, upon disruption of tumor suppressors, autophagy adopts a pro-death role with apoptotic pathways (48). In prostate, breast, ovarian, and lung cancer, loss of Beclin1 or inhibition of Beclin1 by the BCL-2 family of proteins causes defective autophagy, increased DNA damage, metabolic stress, and genomic instability. These cancers also display neoplastic changes and increased cell proliferation, unlike cells over-expressing Beclin1, which undergo apoptosis (49, 50). Loss of PTEN, p53, ATG4, ATG5, and MAP1LC31 (ATG8) are linked to tumorigenesis, while up-regulation of PI3K, AKT, BCL-2, and mTOR are associated with inhibition of autophagy and the promotion of tumorigenesis.

Prostate cancer onset and progression is correlated strongly with aging and SIRT1 function governs aging in multiple species. Further studies will be required to determine whether this checkpoint function of Sirt1 in regards to prostate growth is linked to its role in organismal aging.