Fig 1.
MED19 overexpression confers a growth advantage and enables androgen-independent growth of LNCaP cells.
LNCaP cells stably overexpressing MED19 (MED19 LNCaP cells) and LNCaP cells expressing the empty lentiviral vector (control LNCaP cells) were cultured in media depleted of androgens by addition of FBS charcoal-stripped of steroids (A, B, C) or media containing androgens by addition of standard FBS (D, E, F). A) and D) Proliferation was measured over 7 days and is expressed as fold change in relative fluorescent units (RFU), normalized to Day 0. B) and E) Colony formation was evaluated by culturing MED19 LNCaP cells and control LNCaP cells at low density for 11 days, fixing and staining with crystal violet, and quantifying the number of colonies per field (n = 3). C) and F) Spheroid formation was evaluated by culturing cells on low attachment plates for 10 days and quantifying average spheroid area. Experiments were performed in biological triplicate, with representative results shown. *p < 0.05; **p < 0.01; and ***p < 0.001. ns = not significant.
Fig 2.
MED19 overexpression promotes proliferation in vitro in non-malignant RWPE-1 cells and in vitro and in vivo in mouse prostate stem cells.
A) RWPE-1, B) RWPE-2, or C) mouse stem cells (MSC) expressing a constitutively active myristoylated AKT, and stably overexpressing MED19 (MED19 RWPE-1/RWPE-2/MSC) or control empty vector (control RWPE-1/RWPE-2/MSC), were cultured in their standard media. A-C) Proliferation was measured over 5 days for the MSC, which have a rapid doubling time, or 7 days for RWPE-1 cells and RWPE-2 cells, and is expressed as fold change in relative fluorescent units (RFU) normalized to Day 0. Experiments were performed in biological duplicate, with representative results shown. D) MED19 MSC or control MSC were injected into the flanks of Nu/J mice and tumor volume was measured over 95 days (2 mice per group). Representative images of tumors taken at time of sacrifice are shown. *p < 0.05; **p < 0.01; and ***p < 0.001. ns = not significant.
Fig 3.
MED19 LNCaP cells depend on AR transcriptional activity for androgen-independent growth and do not have altered expression of AR.
A) RNA was extracted from control LNCaP and MED19 LNCaP cells cultured under androgen deprivation for 3 days, and AR mRNA measured by qPCR (fold change mRNA expression normalized to RPL19, with AR mRNA expression in control LNCaP cells set as “1”) ns = not significant. B) Total protein lysate was collected and probed for AR protein levels by western blot. ERK1/2 was used as a loading control. C) and D) MED19 LNCaP cells were treated with enzalutamide (0–80 μM) in C) androgen-depleted media, or D) androgen-containing media, alongside control LNCaP cells. Proliferation was measured over 7 days. Percent cell growth at day 7 is normalized to vehicle treatment (0 μM, 100%). The IC50 from three experiments is shown. Experiments were performed in biological triplicate, with representative results shown.
Fig 4.
MED19 overexpression causes a selective shift in gene expression and in the AR cistrome under androgen deprivation and with R1881 treatment.
MED19 LNCaP cells and control LNCaP cells were cultured under androgen deprivation for 3 days, and cells were treated with ethanol vehicle or R1881 (10 nM for 16 h for RNA-seq and 100 nM for 4 h for ChIP-seq). RNA-seq with ribodepletion was performed in biological triplicate. ChIP-seq for FLAG-MED19, AR, and H3K27ac was performed in biological triplicate, with the exception of ChIP-seq for AR in control LNCaP cells + R1881, where one sample was excluded from the analyses because of low signal. A) (Left) Heatmap of differentially expressed genes with MED19 overexpression (fold change ≥1.5, p-adj ≤0.05) for androgen deprivation, associated with AR as the top regulatory transcription factor from ChEA. (Right) Number and overlap of occupancy sites under androgen deprivation for AR in control LNCaP cells and MED19 LNCaP cells (top) and for AR and MED19 in MED19 LNCaP cells (bottom). B) (Left) Heatmap of differentially expressed genes with MED19 overexpression (fold change ≥1.5, p-adj ≤0.05) for R1881 treatment, associated with AR as the top regulatory transcription factor from ChEA. (Right) Number and overlap of occupancy sites with R1881 treatment for AR in control LNCaP cells and MED19 LNCaP cells (top) and for AR and MED19 in MED19 LNCaP cells (bottom).
Fig 5.
MED19 overexpression alters the response to androgens.
MED19 LNCaP cells and control LNCaP cells were cultured under androgen deprivation for 3 days, and cells were treated with ethanol vehicle or R1881 (10 nM for 16 h for RNA-seq and 100 nM for 4 h for ChIP-seq). RNA-seq with ribodepletion was performed in biological triplicate. ChIP-seq for FLAG-MED19, AR, and H3K27ac was performed in biological triplicate, with the exception of ChIP-seq for AR in control LNCaP cells + R1881, where one sample was excluded from the analyses because of low signal. A) (Left) Heatmap of differentially expressed genes (fold change ≥1.5, p-adj ≤0.05) for control LNCaP cells treated with R1881 vs. vehicle, associated with AR as the top regulatory transcription factor from ChEA. (Right) Number and overlap of occupancy sites for R1881 vs. vehicle treatment for AR in control LNCaP cells. B) (Left) Heatmap of differentially expressed genes (fold change ≥1.5, p-adj ≤0.05) for MED19 LNCaP cells treated with R1881 vs. vehicle, associated with AR as the top regulatory transcription factor from ChEA. (Right) Number and overlap of occupancy sites for R1881 vs. vehicle treatment for AR (top) and for MED19 (bottom) in MED19 LNCaP cells.
Fig 6.
MED19 occupies gene targets like LRRTM3 under androgen deprivation and alters mRNA expression, AR occupancy, and H3K27 acetylation.
MED19 LNCaP cells and control LNCaP cells were cultured under androgen deprivation for 3 days and treated with ethanol vehicle (shown) or R1881 (10 nM for 16 h for RNA-seq and 100 nM for 4 h for ChIP-seq; shown in S11 Fig). RNA-seq with ribodepletion was performed in biological triplicate. ChIP-seq for FLAG-MED19, AR, and H3K27ac was performed in biological triplicate, with the exception of ChIP-seq for AR in control LNCaP cells + R1881, where one sample was excluded from the analyses because of low signal. A) Fold change mRNA expression from RNA-seq, and qPCR validation of upregulation of LRRTM3 mRNA expression under androgen deprivation (performed in biological triplicate with representative results shown; fold change expression normalized to RPL19 with LRRTM3 mRNA expression in control LNCaP cells set as “1”). ***p < 0.001. B) ChIP-seq tracks (representative results) for FLAG-MED19, AR, and H3K27ac under androgen deprivation shown for intronic regions of LRRTM3. Fold change (up (+) or down (-)) in occupancy scores for MED19 LNCaP cells compared to control LNCaP cells shown for each peak (see S6 Table for all occupancy scores). ++ indicates positive occupancy score in MED19 LNCaP cells and a score of zero in control LNCaP cells;—indicates an occupancy score of zero in MED19 LNCaP cells and a positive score in control LNCaP cells.
Fig 7.
MED19 promotes AR occupancy and H3K27 acetylation at MAOA under androgen deprivation, which controls androgen-independent growth.
MED19 LNCaP cells and control LNCaP cells were cultured under androgen deprivation for 3 days and treated with ethanol vehicle (shown) or R1881 (10 nM 16 h for RNA-seq and 100 nM for 4 h for ChIP-seq; shown in S13 Fig). RNA-seq with ribodepletion was performed in biological triplicate. ChIP-seq for FLAG-MED19, AR, and H3K27ac was performed in biological triplicate, with the exception of ChIP-seq for AR in control LNCaP cells + R1881, where one sample was excluded from the analyses because of low signal. A) Fold change mRNA expression from RNA-seq, and qPCR validation of upregulation of MAOA mRNA expression under androgen deprivation (performed in biological triplicate with representative results shown; fold change expression normalized to RPL19 with MAOA mRNA expression in control LNCaP cells set as “1”). B) ChIP-seq tracks (representative results) for FLAG-MED19, AR, and H3K27ac under androgen deprivation shown for promoter region of MAOA. Fold change (up (+) or down (-)) in occupancy scores for MED19 LNCaP cells compared to control LNCaP cells shown for each peak (see S6 Table for all occupancy scores). ++ indicates positive occupancy score in MED19 LNCaP cells and a score of zero in control LNCaP cells;—indicates an occupancy score of zero in MED19 LNCaP cells and a positive score in control LNCaP cells. C) ChIP-qPCR for FLAG-MED19, AR, and H3K27ac at the MAOA promoter overlapping with published ARE. Experiment was performed in duplicate, with representative results shown. D) MAOA was depleted by siRNA and proliferation of MED19 LNCaP cells in androgen-depleted media was evaluated after 7 days, normalized to proliferation with scrambled siRNA treatment (negative control, light grey). KIF11 knockdown is included as a positive control (black). Experiment was performed in biological duplicate, with representative results shown. E) Validation of MAOA knockdown, with MAOA depleted by siRNA and mRNA expression of MAOA in MED19 LNCaP cells cultured in androgen-depleted media measured after 5 days (fold change expression normalized to RPL19, with MAOA mRNA expression with scrambled siRNA treatment set as “1”). *p < 0.05; **p < 0.01; and ***p < 0.001.
Fig 8.
ELK1 is enriched at sites of AR and MED19 occupancy unique to MED19 LNCaP cells under androgen deprivation, occupies the MAOA promoter, and controls MAOA expression and androgen-independent growth.
MED19 LNCaP cells and control LNCaP cells were cultured under androgen deprivation for 3 days and treated with ethanol vehicle. RNA-seq with ribodepletion was performed in biological triplicate. ChIP-seq for FLAG-MED19, AR, and H3K27ac was performed in biological triplicate, with the exception of ChIP-seq for AR in control LNCaP cells + R1881, where one sample was excluded from the analyses because of low signal. A) The top 10 enriched transcription factor motifs under androgen deprivation associated with sites of AR and MED19 occupancy in MED19 LNCaP cells where AR is present only in MED19 LNCaP cells are shown, with ELK1 as the top associated transcription factor (labeled diagram of occupancy sites, right). B) ELK1 knockdown is the top hit from Transcription Factor Perturbation from Enrichr associated with genes upregulated by MED19 overexpression under androgen deprivation from the RNA-seq study; AR knockdown is also strongly associated. Fold downregulation of MAOA mRNA with associated ELK1 knockdown (GSE 34589) or associated AR knockdown (GSE 22483) is shown. C) ELK1 was depleted by siRNA and mRNA expression of MAOA in MED19 LNCaP cells cultured in androgen-depleted media was measured after 5 days (fold change expression normalized to RPL19, with MAOA mRNA expression with scrambled siRNA treatment set as “1”). Experiment was performed in biological triplicate, with representative results shown. D) ELK1 was depleted by siRNA and proliferation of MED19 LNCaP cells in androgen-depleted media was evaluated after 7 days, normalized to proliferation with scrambled siRNA (negative control, light grey). KIF11 knockdown is included as a positive control (black). Experiment was performed in biological duplicate, with representative results shown. E) Validation of ELK1 knockdown, with ELK1 mRNA measured as in C (with ELK1 mRNA expression with scrambled siRNA treatment set as “1”). F) ChIP-qPCR for ELK1 at the MAOA promoter overlapping with published ARE. *p < 0.05; **p < 0.01; and ***p < 0.001.
Fig 9.
Model of MED19 driving androgen-independent growth by cooperating with ELK1 to promote AR occupancy and H3K27 acetylation at the MAOA promoter.
A) Under low androgen and low MED19, AR occupancy is low at the MAOA promoter, MAOA is weakly expressed, and cells are growth-inhibited. B) When MED19 is upregulated, MED19 in Mediator cooperates with ELK1 to recruit and stabilize AR via its N-terminal domain (NTD) at the MAOA promoter, also recruiting Pol II and HATs, upregulating MAOA expression, and driving androgen-independent growth.