Ultrastructural analysis of prostate cancer tissue provides insights into androgen-dependent adaptations to membrane contact site establishment

Abstract

Membrane trafficking and organelle contact sites are important for regulating cell metabolism and survival; processes often deregulated in cancer. Prostate cancer is the second leading cause of cancer-related death in men in the developed world. While early-stage disease is curable by surgery or radiotherapy there is an unmet need to identify prognostic biomarkers, markers to treatment response and new therapeutic targets in intermediate-late stage disease. This study explored the morphology of organelles and membrane contact sites in tumor tissue from normal, low and intermediate histological grade groups. The morphology of organelles in secretory prostate epithelial cells; including Golgi apparatus, ER, lysosomes; was similar in prostate tissue samples across a range of Gleason scores. Mitochondrial morphology was not dramatically altered, but the number of membrane contacts with the ER notably increased with disease progression. A three-fold increase of tight mitochondria-ER membrane contact sites was observed in the intermediate Gleason score group compared to normal tissue. To investigate whether these changes were concurrent with an increased androgen signaling in the tissue, we investigated whether an anti-androgen used in the clinic to treat advanced prostate cancer (enzalutamide) could reverse the phenotype. Patient-derived explant tissues with an intermediate Gleason score were cultured ex vivo in the presence or absence of enzalutamide and the number of ER-mitochondria contacts were quantified for each matched pair of tissues. Enzalutamide treated tissue showed a significant reduction in the number and length of mitochondria-ER contact sites, suggesting a novel androgen-dependent regulation of these membrane contact sites. This study provides evidence for the first time that prostate epithelial cells undergo adaptations in membrane contact sites between mitochondria and the ER during prostate cancer progression. These adaptations are androgen-dependent and provide evidence for a novel hormone-regulated mechanism that support establishment and extension of MAMs. Future studies will determine whether these changes are required to maintain pro-proliferative signaling and metabolic changes that support prostate cancer cell viability.

Keywords: androgen-deprivation; lipophagy; membrane-contact site; mitochondria-associated membrane; patient-derived tumor explant.

Figure 1

Analysis of the ultrastructure of secretory epithelial cells in human prostate tissue. (A) Glandular structures were identified in semi-thin sections of prostate tissue by light microscopy. Arrows point to glands. The sample was further trimmed down to regions containing glands and ultrathin section were prepared for TEM microscopy where the same structures were imaged at high resolution allowing secretory epithelial cells to be identified based on their morphology (B) Electron micrograph showing a gland at 100x magnification. Scale bar 20 µm. (C) Electron micrograph showing the secretory epithelia within a gland in the prostate. Scale bar, 5 µm. (D) Intracellular organelles in luminal epithelial cells were imaged at 8,000-20,000x magnification. Lipid droplet, LD; ER, endoplasmic reticulum; m, mitochondrion. Scale bar, 500 nm.

Figure 2

Ultrastructure of the Golgi apparatus in normal versus PCa tissue. (A–D) The Golgi apparatus in a secretory epithelial cell in a gland from a normal (A), intermediate grade (B, D) and a low grade (C) prostate cancer tissue region had a normal appearance with stacked cisternae and a classic ribbon structure. The Golgi ribbon, pseudo colored in blue, measured up to 6 µm in length (B). L, lysosome; cv, condensing vacuoles. Scale bars, 500 nm (A, D), 2µm (B) and 1µm (C).

Figure 3

Ultrastructure of organelles and their membrane contacts in PCa tissue. (A) Electron micrograph illustrating the presence of ER sheets close to the nucleus (n) of secretory prostate epithelial cells. Scale bar, 2 µm. (B) ER tubules form a network throughout the cytoplasm in intermediate grade PCa tissue. Scale bar, 500 nm. (C) In intermediate grade PCa tissue lipid droplets were often found in contact with the ER. In this image an ER whorl is forming a membrane contact site with a lipid droplet (LD). Scale bar, 500 nm. (D) An electron micrograph from intermediate grade PCa tissue showing a lipid droplet with striations indicating a liquid-phase separation of lipids. Scale bar, 500 nm. (E) Electron micrograph illustrating membrane contact sites between lipid droplets, mitochondria (m) and (ER) in a luminal epithelial cell from a PCa tissue of intermediate grade. Scale bar, 500 nm. (F) A micrograph showing an example of a lysosome engulfing small lipid droplets (arrows) in normal tissue. Scale bar, 500 nm. (G) Electron micrograph showing two lysosomes (arrows) engaged in lipophagy at a large lipid droplet (LD) in PCa tissue of intermediate grade. Scale bar, 1µm. (H) Electron micrograph showing membrane contact sites between a lipid droplet, mitochondria, peroxisome (Px) and ER in an intermediate grade PCa tissue. Scale bar, 1000 nm.

Figure 4

Ultrastructure of ER-mitochondria contacts in PCa tissue. (A-C) Ultrastructure of MAMs (magenta pseudo color) and wrappER-associated mitochondria (WAM) (green pseudo color) in luminal epithelial cells of normal, low and intermediate grade prostate cancer. Scale bar, 500nm. (D, E) Bar graphs showing the quantification of MAMs from normal (N=3), low grade (N=3) and intermediate-to-high grade (N=3) using 20 images per sample and >70 MAMs per group for length measurement. (F, G) Bar graphs showing the quantification of abundance and length of WAM contacts in the samples described in D. The length of WAM contacts were measured in a minimum of 150 mitochondria per group. All data are represented as mean ± sem; one-way ANOVA test. ****p<0.0001, ***p<0.001. ns, not significant.

Figure 5

Quantification of MAMs from patient-derived tissue explants treated with the anti-androgen enzalutamide. (A-C) Electron micrographs of MAMs in DMSO-treated (A), enzalutamide-treated (B) and Perhexiline-treated (C) explant tissue sections from PCa tissue with an intermediate-to-high Gleason score. Magenta arrows point to MAMs. Scale bar, 500 nm. (D-E) Analysis of enzalutamide-treated patient-derived tissue explant samples demonstrate a negative regulation of abundance and length of MAMs in three independent patient-derived tissue samples compared to paired DMSO-treated control samples. (D) Bar graph showing the percentage of MAM contacts among the total number of mitochondria. N=3, n≥20 images per sample. (E) Quantification of MAM length from electron micrographs (n=20 images) showed a significant difference between paired control-treated (DMSO) and enzalutamide-treated patient-derived tissue (N=3 clinical samples). We quantified the length of 183 control and 92 enzalutamide-treated MAMs. (F) Quantification of MAMs in Perhexiline-treated (C) tissue compared to a DMSO control from one patient-derived sample. The analysis included representative images; n=16 control and 20 treated. (G) Analysis of the length of MAMs in paired explant tissue samples treated with perhexiline. We quantified the length of 34 DMSO- and 49 perhexiline-treated MAMs. (H) Autophagic vacuoles in a perhexiline-treated explant tissue sample. Scale bar, 1 µm. All data are represented as mean ± sem; unpaired t-test. ****p<0.0001, **p<0.01. ns, not significant.