Beta-catenin represses protein kinase D1 gene expression by non-canonical pathway through MYC/MAX transcription complex in prostate cancer

Abstract

Down regulation of Protein Kinase D1 (PrKD1), a novel serine threonine kinase, in prostate, gastric, breast and colon cancers in humans leads to disease progression. While the down regulation of PrKD1 by DNA methylation in gastric cancer and by nuclear beta-catenin in colon cancer has been shown, the regulatory mechanisms in other cancers are unknown. Because we had demonstrated that PrKD1 is the only known kinase to phosphorylate threonine 120 (T120) of beta-catenin in prostate cancer resulting in increased nuclear beta-catenin, we explored the role of beta-catenin in gene regulation of PrKD1. An initial CHIP assay identified potential binding sites for beta-catenin in and downstream of PrKD1 promoter and sequencing confirmed recruitment of beta-catenin to a 166 base pairs sequence upstream of exon 2. Co-transfection studies with PrKD1-promoter-reporter suggested that beta-catenin represses PrKD1 promoter. Efforts to identify transcription factors that mediate the co-repressor effects of beta-catenin identified recruitment of both MYC and its obligate heterodimer MAX to the same binding site as beta-catenin on the PrKD1 promoter site. Moreover, treatment with MYC inhibitor rescued the co-repressor effect of beta-catenin on PrKD1 gene expression. Prostate specific knock out of PrKD1 in transgenic mice demonstrated increased nuclear expression of beta-catenin validating the in vitro studies. Functional studies showed that nuclear translocation of beta-catenin as a consequence of PrKD1 down regulation, increases AR transcriptional activity with attendant downstream effects on androgen responsive genes. In silico human gene expression analysis confirmed the down regulation of PrKD1 in metastatic prostate cancer correlated inversely with the expression of MAX, but not MYC, and positively with MXD1, a competing heterodimer of MAX, suggesting that the dimerization of MAX with either MYC or MXD1 regulates PrKD1 gene expression. The study has identified a novel auto-repressive loop that perpetuates PrKD1 down regulation through beta-catenin/MYC/MAX protein complex.

Keywords: MAX; MYC; beta-catenin; prostate cancer; protein kinase D1.

Figure 1. Presence of the beta-catenin and MYC/MAX binding site at the PrKD1 gene promoter

(A) Chromatin immunoprecipitation (CHIP) for beta-catenin on LNCaP cells. Cross-linked DNA/protein complex of LNCaP cells was pulled down with beta-catenin antibody or normal IgG antibody (as control). DNA was isolated from the pulled down complex and the enrichment of PrKD1 promoter sequence in the complex were quantified by qPCR, which shows significantly higher recruitment of beta-catenin to PrKD1 promoter region compared to IgG control. (B) CHIP assay following pull down with MYC and MAX antibodies showed both MYC and MAX are recruited to the same PrKD1 binding site for beta-catenin. (C)Transcription factor (TF) profiling using a 48-TF array. LNCaP nuclear extract pulled down with beta-catenin was mixed with either probe and PrKD1 promoter sequence or probe alone. The luminescence of the 48-TF array plate quantified and compared between the two samples. The bar graph shows the luminescence intensity for streptavidin. The TFs that regulates PrKD1 will bind to PrKD1 sequence and not to the corresponding probes showing lower intensity of bars compared to samples without PrKD1 sequence. TF binding is considered significant when the intensity for TF in the samples without PrKD1 is ≥1.5 fold compared to samples containing PrKD1 sequence bound to the PrKD1 promoter.

Figure 2. PrKD1 down regulation or representative condition (beta-catenin T120 mutation) increases nuclear beta-catenin in vitro and in vivo

(A) The western blot using nuclear fraction of the transfected cells and PCNA as loading control for nuclear protein confirming transfection efficiency. The numbers under beta-catenin bands show the relative density of the bands, calculated as the ratio of each sample density for beta-catenin to the density of the same sample for PCNA. The results for wild type beta-catenin (WT) and T120 mutant (T120 mu) were normalized against empty vector (E). Increased beta-catenin nuclear localization in PrKD1 knocked out (KO) mice (B) compared to litter mate control (C). In B and C panels representative samples of beta-catenin immunofluorescent staining in mouse prostate tissue demonstrating the co- localization of beta-catenin (red, left panel) and nuclear DAPI (blue, middle panel) results in purple color (right panel) under x63 magnification of the confocal microscope. In PrKD1 KO mice (B), beta-catenin shows more nuclear translocation compared to littermate controls (C). In each set of pictures, the left panel: beta-catenin/AF594; middle panel: DAPI nuclear staining and right panel: merged. Scale bar shows 30μM.

Figure 3. The unphosphorylated threonine 120 (T120) mutant beta-catenin represses PrKD1 gene expression

LNCaP prostate cancer cells transfected with wild type beta-catenin (WT- B-cat), T120 mutated beta-catenin (T120 mu) or empty vector (E). (A) The unphosphorylated T120 mutant beta-catenin represses PrKD1 promoter activity; LNCaP cells co-transfected with beta-catenin constructs (WT beta-catenin, T120-mutant beta-catenin and empty vector) and PrKD1/luciferase vector. The non-insert luciferase vector was used as control. Thirty six hours after transfection, 100 μl of the assay reagent was added to each well, the plates were incubated at room temperature for 30 min and read on an illuminometer. The Log2 ratio of the average signal from PrKD1 promoter/reporter transfected cells divided by average signal from non-targeting control was calculated. The results for WT-beta-catenin transfected cells and T120-mutated cells were normalized against empty vector, which show that the cells transfected with unphosphorylated-T120 mutant beta-catenin have significantly less active PrKD1 promoter activity. (B) PrKD1 gene expression is significantly lower in LNCaP cells transfected with T120 mutant beta-catenin compared to WT beta-catenin or vector control transfected cells. The transcriptional expression of PrKD1 was normalized against RNA18S using ddCT. P<0.05 value is considered significant. (C) PrKD1 promoter activity in LNCaP transfected cells in the presence and absence of MYC/MAX dimerization inhibitor demonstrating rescue of PrKD1 promoter activity by the inhibitor. (D-G) The effect of MYC/MAX dimerization inhibitor on high- and low- PrKD1 expressing LNCaP and C4-2 prostate cancer respectively. All experiments were done with or without treatment with 100uM of MYC/MAX dimerization inhibitor. (D) Quantification of PrKD1 transcriptional expression in the presence or absence of MYC inhibitor. Inhibition of MYC resulted in significant increase in the expression of PrKD1 in C4-2 cells. (E) MTS proliferation assay after 72 hours. (F) Migration wound healing assay for 48 hours. 20000 cells were seeded, after an overnight the wound was made using IncuCyte wound maker. Cells incubated in 37°C incubator in the presence or absence of MYC inhibitor. The plate was scanned every 2.5 h hours. Data collection started after an overnight culture. The graph shows the confluence of the cells for 48hours. (G) Matrigel invasion assay; the bar graph shows the number of invaded cells through matrigel coated inserts compared to non-coated control inserts. The invasion ability was significantly lower in C4-2 cells treated with MYC inhibitor.

Figure 4. The effect of unphosphorylated T120 mutant beta-catenin on androgen receptor (AR) activity

The LNCaP cells were transfected with wild type beta-catenin (WT- B-cat), T120 mutated beta-catenin (T120 mu) or empty vector (E), cultured in charcoal stripped medium with or without 1nM R1881. (A) The western blot using nuclear fraction of the transfected cells. The density of each band represented as a relative value normalized against loading control (PCNA). (B) AR transcriptional activity using ARE reporter gene in LNCaP transfected cells demonstrating significant activation of AR activity in T120 mutant beta catenin compared to WT-beta-catenin. (C) The expression of select androgen responsive genes in response to treatment with R1881. The expression in both WT and mutant beta-catenin transfected samples normalized against the expression in LNCaP cells transfected with empty vector. P values <0.05 is considered as significant. The expression of androgen responsive genes correlates directly with increased AR activity following transfection of T120 beta-catenin mutant compared to WT-beta-catenin control. (D) The increased expression of KLK3 (as representative of AR responsive genes) following transfection with T120 mutant beta-catenin was rescued by 10 μM MDV300 (Enzalutamide), suggesting that the T120 beta-catenin mutant effect is mediated by AR. The expression of genes was normalized against the expression of genes in LNCaP cells transfected with empty vector.

Figure 5. The expression of MAX reversely correlates with PrKD1 in human prostate cancer tissue

(A) Box plots of PrKD1, MYC, and MAX expression levels in benign (Ben), primary (Prim), and metastatic (Met) human prostate cancer tissue (microarray data set GSE8511). The box represents the interquartile range of data with various samples, and the line through that box represents the median of the distribution. The range is indicated by whiskers on the plot. Benign tumor, n=16; primary tumor, n=12; metastatic tumor, n=13. (B) Scattered correlation plots of PrKD1 versus MYC, MAX, and MXD1; and MAX versus MXD1 for all samples displayed in A.

Figure 6. Auto repressive loop of PrKD1 down regulation in prostate cancer

PrKD1 is down regulated in advanced prostate cancer, which decreases phosphorylation of beta-catenin at thereonine120 (T120) residue. The T120 unphosphorylated beta-catenin translocates to the nucleus resulting in increased nuclear beta-catenin activity. The nuclear beta-catenin in turn represses PrKD1 expression through MYC/MAX complex and thereby perpetuating down regulation of PrKD1, which is associated with prostate cancer progression. Increased nuclear beta-catenin resulted in increased AR coactivation. Down regulation of PrKD1 decreases phosphorylation of HSP27 at Serine 82 (S82) which facilitates and increases AR nuclear translocation (see discussion).