Abstract
Recent studies suggest that glucocorticoid receptor (GR) activation can cause enzalutamide resistance in advanced prostate cancer (PCa) via functional bypass of androgen receptor (AR) signaling. However, the specific molecular mechanism(s) driving this process remain unknown. We have previously reported that the transcription factor TBX2 is over-expressed in castrate-resistant prostate cancer (CRPC). In this study, using human PCa and CRPC cell line models, we demonstrate that TBX2 downregulates AR and upregulates GR through direct transcriptional regulation. TBX2 also activated the GR via TBX2-GR protein-protein interactions. Together, TBX2-driven repression of AR and activation of GR resulted in enzalutamide resistance. Our laboratory findings are supported by clinical samples, which show a similar and consistent pattern of transcriptional activity among TBX2, AR and GR across patient cohorts. Notably, we report that SP2509, an allosteric inhibitor of the demethylase-independent function of LSD1 (a TBX2-interacting protein in the COREST complex) disrupts both TBX2-LSD1 and TBX2-GR protein-protein interactions, revealing a unique mode of SP2509 action in CRPC. Taken together, our study identifies the TBX2-driven AR- to GR- signaling switch as a molecular mechanism underlying enzalutamide resistance and provides key insights into a potential therapeutic approach for targeting this switch by disrupting TBX2-GR and TBX2-LSD1 protein-protein interactions.
Similar content being viewed by others
Introduction
Androgen deprivation therapy (ADT) that includes androgen receptor signaling inhibitors (ARSIs), such as enzalutamide, apalutamide, and darolutamide, forms the cornerstone of treatment of advanced prostate cancer (PCa) [1,2,3]. Following a period of response, patients inevitably develop resistance to ADT and the disease progresses to castration resistant prostate cancer (CRPC) [1,2,3,4]. In recent years, the use of more potent ADTs, like enzalutamide, has led to significant improvements in patient survival. However, the resulting increased selective pressure on the PCa cells leads to clinically unintended consequences on AR signaling, such as the activation of mechanisms that lead to bypassing AR signaling [5,6,7,8,9].
CRPC has been historically associated with heterogenous expression of the AR (AR positive and AR low/negative CRPC) [7, 10, 11]. Recent studies have demonstrated that the prevalence of AR low/negative CRPC has increased with the use of next generation ARSIs such that 20 to 40% of clinical CRPC is now AR low/negative [10]. The major challenge of AR negative CRPC subtypes is that they cannot be effectively treated with AR-targeted therapeutics. To date, only a few molecular mechanisms have been identified that result in the loss/downregulation of AR; there is a critical need to increase our understanding of AR regulation in the context of CRPC.
Most recently, it has been reported that enzalutamide resistance can be driven by the upregulation of the glucocorticoid receptor (GR) (NR3C1) [6, 8, 9, 12,13,14,15,16,17,18,19,20], a nuclear steroid receptor whose DNA binding domain shares a high level of similarity with the DNA binding domain of the AR [5, 6, 8, 9, 16]. The AR and GR share a 50% overlap in their cistromes and transcriptomes [5, 6, 8, 16, 21]. Together, this has led to the hypothesis that signaling bypass of the AR by GR drives CRPC [5,6,7,8,9, 22]. Unfortunately, more so than AR, the GR has pleiotropic effects in all nucleated cells [6], and therefore systemic clinical inhibition of GR results in deleterious side effects [23, 24]. Additionally, attempts at direct GR blockade in CRPC have failed to show beneficial effects, likely as GR antagonists counterproductively increase adrenal androgen production [25]. Greater elucidation of the molecular mechanisms driving AR-GR signaling bypass might lend insights on how to effect indirect modes of GR inhibition and identify novel targeted therapeutics for the clinical management of ARSI resistance.
Previous reports [26, 27] from our lab detailed the use of human xenograft models of disease-progression and human PCa samples to demonstrate that TBX2, a T-box family of transcription factor (TF) [28] is over-expressed during the progression from PCa to CRPC and to bone metastatic CRPC [26], drives bone metastatic progression in PCa [26], and mediates the propagation of the CRPC phenotype via both cell-intrinsic and exosome-mediated paracrine modes [27]. In accordance with our findings, a recent report confirmed TBX2 and GR as two of the four TFs that drive enzalutamide resistance [29]. Further, studies have demonstrated that TBX2 is part of the Co-repressor of RE1-Silencing Transcription Factor (CoREST) protein complex, that includes lysine-specific demethylase 1 (LSD1/KDM1A) and ZNF217, whose interaction is reported to mediate CRPC [30].
In the current study, we have discovered that TBX2 is a molecular switch that drives the AR to GR signaling bypass and thereby confers enzalutamide resistance. We genetically modulated TBX2 expression in cell line models of human prostate adenocarcinoma, CRPC, and enzalutamide-resistant CRPC to demonstrate that TBX2 downregulated AR and upregulated GR. Our study reveals that TBX2 represses AR via direct promoter binding while upregulating GR via TBX2-GR protein-protein interaction and direct GR promoter binding. Increase of TBX2 was associated with ARSI resistance that was reversed by knock-down of GR. Specifically, the relationships observed between TBX2, AR and GR in cell line models were corroborated by consistent associations of their TF activities in patient cohorts. Importantly, we found that SP2509, an allosteric inhibitor of the demethylase-independent function of LSD1, can disrupt both TBX2-LSD1 and TBX2-GR protein-protein interactions thereby uncovering a unique mode of SP2509 action in CRPC.
Taken together, our study unravels a new molecular model of CRPC wherein: 1) TBX2, LSD1 and GR proteins interface with each other, 2) pharmacological inhibition of LSD1 blocks the TBX2-driven AR-to-GR switch through disruption of the TBX2-GR and TBX2-LSD1 interactions, and 3) the relationships among the transcriptional activities of TBX2, AR and GR in patient cohorts may serve as a potential predictor of resistance to enzalutamide. Given that systemic inhibition of the GR causes deleterious side effects [15, 25], these results are a proof of principle for an indirect, targeted pharmacological approach to GR inhibition that has the potential to restore enzalutamide sensitivity in CRPC.
Results
An inverse relationship between TBX2 and AR
Our previous study used murine xenograft models of isogenic human PCa cell lines that mimic human PCa progression from adenocarcinoma (LNCaP) to CRPC (C4-2) to bone metastatic CRPC (C4-2B) to demonstrate an increase in TBX2 expression with the disease progression [26].
To investigate a potential relationship between TBX2 and AR, we used three distinct approaches to modulate TBX2 in human PCa cells: 1) dominant-negative (DN) based activity disruption of endogenous TBX2 using a dominant negative (TBX2DN) construct, 2) conversely, overexpressing (OE) TBX2 using the overexpression construct (TBX2OE), 3) knocking down TBX2 using an shRNA construct (shTBX2). Since previous reports, including our own studies, have shown that TBX2 predominantly acts as a repressor of gene expression, we postulated that the DN-based activity disruption of TBX2 using a protein variant with a deleted repressor domain would reveal targets being repressed by TBX2 [31]. We performed an unbiased RNA-seq analysis on PC3 human CRPC cells (GSE242282) in which DN-based activity disruption of TBX2 was performed i.e. PC3TBX2DN cells (Fig. 1A). Of note, parental PC3 cells do not express AR mRNA or protein [32]. We observed a dramatic (60-fold) increase in AR mRNA expression in PC3TBX2DN cells when compared with the PC3Neo control cells (Fig. 1A). Further, multiple AR target genes were also significantly upregulated in the RNA-seq (incl. TMPRSS2, PMEPA1, STEAP4) [33] (Fig. 1A). Upregulation of E-cadherin (CDH1) mRNA in PC3TBX2DN cells compared with the control PC3Neo cells (Fig. 1A) served to validate the DN approach since a previous report [34] showed that TBX2 directly represses CDH1 transcription.
A RNA-seq analysis in PC3TBX2DN compared with PC3Neo control showing a significant increase in the expression of AR (60-fold) and AR target genes; B q RT-PCR analysis showing increased AR mRNA expression in PC3TBX2DN and C4-2BTBX2DN when compared with the respective PC3Neo and C4-2BNeo controls, and conversely a decreased AR in LNCaPTBX2OE cells when compared with LNCaPNeo control; C Western blot analysis showing increased AR protein expression in PC3TBX2DN and C4-2BTBX2DN cells when compared with the respective PC3Neo and C4-2BNeo controls, and conversely decreased AR protein in LNCaPTBX2OE cells when compared with LNCaPNeo control; D qRT-PCR analysis showing increased mRNA expression of some of the known AR target genes in PC3TBX2DN and C4-2BTBX2DN cells when compared with the respective PC3Neo and C4-2BNeo controls, and conversely decreased expression of the AR target genes in LNCaPTBX2OE when compared with LNCaPNeo control; E Western blot analysis showing decreased TBX2 and increased AR protein expression in PC3shTBX2 cells when compared with the PC3NTSCR control cells; F Western blot analysis showing increased AR protein expression in DU145TBX2DN cells compared with the DU145Neo control cells. Data represent the average of triplicates ±S.D; Student’s unpaired 2-tailed t-tests were performed to compare the two groups **p ≤ 0.01; ***p ≤ 0.001; and ****p ≤ 0.0001.
Validation of the RNA-seq results at the mRNA and protein levels revealed increased AR expression upon DN-based activity disruption of TBX2 (see PC3TBX2DN, C4-2BTBX2DN and DU145TBX2DN) versus respective controls (PC3Neo, C4-2BNeo and DU145Neo) [Fig. 1B, C, F and (Supplementary Fig. S1A)]. Next, using the converse approach, we tested the effects of TBX2 OE in LNCaP human PCa cells (Supplementary Fig. S1B). Of note LNCaP cells express low endogenous levels of TBX2 [26]. We observed decreased AR expression in LNCaPTBX2OE cells when compared to LNCaPNeo controls (Fig. 1B, C). In addition, certain known AR downstream targets (TMPRSS2, NKX3.1 and FKBP5 [35, 36] demonstrated a negative association with TBX2 expression in LNCaP, C4-2B and PC3 cells (Fig. 1D). As additional validation, shRNA mediated knock down of TBX2 in PC3 cells (PC3shTBX2) (Fig. 1E, Supplementary Fig. S1C) resulted in upregulation of AR compared to PC3NTSCR controls (Fig. 1E, Supplementary Fig. S1C). Further, while the DN and shRNA approaches of TBX2 modulation resulted in upregulation of CDH1, the converse approach of TBX2 OE resulted in CDH1 downregulation (Supplementary Fig. S1D, E) thereby providing validation of these multiple complementary genetic modulation approaches for TBX2. Together, these results demonstrated that TBX2 is inversely associated with AR in human PCa cells.
TBX2 directly represses AR transcription
As the genetic modulation studies showed a negative association between TBX2 and the AR and AR target genes, we investigated if TBX2 repressed AR via direct transcriptional regulation. Based on the consensus T-Box binding sequences that have been previously reported [37], in silico analysis revealed multiple sites on the AR promoter region to which TBX2 could potentially bind (Fig. 2A).
A In silico analysis of the AR promoter showing predicted TBX2 binding sites on the AR promoter; B Chromatin immunoprecipitation (ChIP) demonstrating TBX2 binding at the −82 bp and −3598 bp upstream of the transcription start (TSS) site on the AR promoter; C Quantitation of the ChIP assay using q-PCR analysis shown in B demonstrating a 5-fold and 10-fold enrichment of TBX2 binding normalized to the input on the −82 bp and −3598 bp binding sites on the AR promoter. One-way ANOVA was performed (n = 3), ***p < 0.001; ****p < 0.0001; D ChIP demonstrating decreased TBX2 binding on the AR promoter in PC3TBX2DN cells compared with the PC3Neo control cells; E Quantitation of the ChIP in D. One-way ANOVA was performed (n = 3), ***p < 0.001; ****p < 0.0001; F JASPAR based depiction of the two consensus T-box binding sequences on the AR promoter. Arrows indicate mutated base pairs in the mutant site1 (−82 bp) and mutant site2 (−3598 bp) on the AR promoter; G Quantitation of luciferase reporter assay in PC3 wildtype cells, showing the relative activity of the luciferase reporter gene at each of the two mutant sites: mutant site1 (−82 bp) and mutant site2 (−3598 bp). Data represent the average of triplicates ±S.D; Student’s unpaired 2-tailed t-tests were performed to compare the two groups **p ≤ 0.01; ***p ≤ 0.001; and ****p ≤ 0.0001; H Schematic representing transcriptional repression of AR by TBX2 through directly binding to the AR promoter.
Chromatin immunoprecipitation (ChIP) followed by PCR and quantitative real-time PCR (q-PCR) analyses in C4-2B cells revealed a 5-fold and 10-fold enrichment of TBX2 binding at −82 bp and −3598 bp upstream of the AR transcription start site (TSS) respectively, indicating TBX2 binding at these sites (Fig. 2A depicted as red highlighted boxes, 2B, C). Furthermore, our results revealed a marked decrease in TBX2 binding to the AR promoter in PC3TBX2DN cells compared to PC3Neo controls, demonstrating that TBX2 repression on the AR promoter is lifted when TBX2 is inhibited (Fig. 2D, E). To confirm AR transcriptional repression by TBX2, we performed site directed mutagenesis (SDM) experiments using base substitution at these two AR promoter binding sites (i.e., −82 bp and −3598 bp), an approach which has previously been reported [38] to abolish TBX2’s binding to the DNA (Fig. 2F). The mutated AR-luciferase plasmids were transfected into wildtype PC3 cells and luciferase reporter assays were performed. These experiments revealed a significant increase in the AR-relative luciferase activity for each of the mutant sites (Fig. 2G). Taken together, these results showed that: a) TBX2 directly binds to the AR promoter, and b) mutating TBX2 binding sites on the AR promoter increased AR transcription strongly demonstrating that TBX2 directly repressed AR transcription (Fig. 2H).
TBX2 directly represses AR coregulator GATA2
Transcription factor GATA2 is a crucial coregulator of AR signaling and necessary for optimal transcriptional activation of the AR [39,40,41,42]. As DN-based activity disruption of endogenous TBX2 increased GATA2 mRNA and protein levels in PC3 and C4-2B cells (Fig. 3A–D) and overexpression of TBX2 (Supplementary Fig. S1B) decreased GATA2 mRNA and protein levels in LNCaP cells (Fig. 3B–D), we investigated if TBX2 was a direct transcriptional repressor of GATA2. In silico analysis identified multiple, potential TBX2 binding sites on the GATA2 promoter (Fig. 3E). ChIP assay in C4-2B cells, followed by PCR and q-PCR analyses, demonstrated 8-fold and 12-fold enrichments of TBX2 at −510 bp and −460 bp upstream of the GATA2 transcription start site (TSS), respectively, indicating TBX2 binding (Fig. 3E depicted as red highlighted boxes, Fig. 3F, G). Additionally, our results showed: 1) a marked reduction in TBX2 binding to the GATA2 promoter in PC3TBX2DN cells compared to PC3Neo controls, and 2) a marked increase in TBX2 binding to the GATA2 promoter in LNCaPTBX2OE cells compared to LNCaPNeo control cells, thereby validating through complementary approaches that TBX2 represses the GATA2 promoter (Supplementary Fig. S2A, B). GATA2 binds at −4.6 kb upstream of the TSS of the AR to activate its transcription [42]. ChIP analysis revealed that DN-based activity disruption of TBX2 in PC3 cells (PC3TBX2DN) resulted in a 10-fold enhancement of GATA2 occupancy at the −4.6 kb upstream binding site on the AR (Fig. 3H). Taken together, these results indicate that TBX2 represses AR transcription in a bimodal manner: a) directly by repression of the AR promoter and, b) indirectly, by repression of AR co-activator, GATA2 (Fig. 3I).
A RNA-seq analysis in PC3TBX2DN cells compared with PC3Neo control showing a significant increase in the expression of GATA2 (4.4-fold, P value 8.3E-13); B qRT-PCR analysis showing increased GATA2 mRNA expression in PC3TBX2DN and C4-2BTBX2DN when compared with the respective PC3Neo and C4-2BNeo controls, and conversely decreased GATA2 expression in LNCaPTBX2OE cells when compared with LNCaPNeo control; C Western blot analysis showing increased GATA2 protein expression in PC3TBX2DN cells when compared with the respective PC3Neo control cells, and conversely decreased GATA2 protein in LNCaPTBX2OE cells when compared with LNCaPNeo control cells; D Quantitation of the Western blot in C; E In silico analysis of the GATA2 promoter showing the predicted TBX2 binding sites; F, G ChIP assay followed by q-PCR showing 8-fold and 12-fold enrichments of TBX2 occupancy on the GATA2 promoter at the −460 bp and −510 bp respectively normalized to input in C4-2B cells. Data represent the average of triplicates ± S.D; Student’s unpaired 2-tailed t-tests were performed to compare the two groups **p ≤ 0.01; ***p ≤ 0.001; and ****p ≤ 0.0001; H ChIP q-PCR in PC3TBX2DN cells showing significant enrichment (10-fold) of GATA2 on the AR promoter at −4.6kb upstream of TSS normalized to input. One-way ANOVA was performed where (n = 3), ***p < 0.001;****, p < 0.0001; I Schematic representing transcriptional repression of GATA2 by TBX2 through direct binding to the GATA2 promoter.
TBX2 is associated with increased GR expression in PCa
AR and GR belong to the steroid receptor superfamily with highly conserved DNA binding domains [5, 6, 8, 16]. Recent evidence supports GR upregulation as a clinically unintended consequence of blocking AR in PCa, particularly when using 2nd generation ADTs, such as enzalutamide [5, 6, 8, 9, 16]. Studies have also identified the significant overlaps in the cistromes and transcriptomes of AR and GR as the likely molecular basis for the re-establishment of an AR-like response to enzalutamide therapy [5, 6, 8, 9, 21]. In order to interrogate the relationship between TBX2 and GR, we used three aforementioned approaches to modulate TBX2 in human PCa cells, i.e., the DN, shRNA and the OE approaches. DN-based activity disruption of TBX2 in the C4-2B and 22Rv1 human CRPC cell lines significantly reduced GR expression (Fig. 4A, B); conversely, TBX2 overexpression (OE) (Supplementary Fig. S1B) in LNCaPTBX2OE human PCa cells resulted in increased GR expression at both the mRNA (Fig. 4A) and protein levels as compared to control LNCaPNeo (Fig. 4C). Further, downregulation of GR was observed in C4-2BshTBX2 cells compared to C4-2BNTSCR controls (Supplementary Fig. S2C). Taken together, our data supports that TBX2 is positively associated with GR expression in CRPC.
A qRT-PCR analysis showing decreased GR mRNA expression in C4-2BTBX2DN and 22Rv1TBX2DN when compared with the respective C4-2BNeo and 22Rv1Neo control cells, and conversely an increased GR expression in LNCaPTBX2OE cells when compared with LNCaPNeo control cells; Data represent the average of triplicates ±S.D; Student’s unpaired 2-tailed t-tests were performed to compare the two groups **p ≤ 0.01; ***p ≤ 0.001; and ****p ≤ 0.0001; B Western blot analysis and its quantification showing decreased GR protein expression in C4-2BTBX2DN cells when compared with the respective C4-2BNeo control cells; C Western blot analysis and its quantification showing increased GR protein in LNCaPTBX2OE when compared with LNCaPNeo control cells.
TBX2 activates GR through transcriptional activation and protein-protein interactions
As our results supported a positive correlation between TBX2 and GR expression in cell-line models of human PCa/CRPC, we investigated if TBX2 could upregulate the GR at the transcriptional level. Interestingly, T-box family members have been shown to either activate or repress gene expression [43]. Our previous report [26] described the ability of TBX2 to activate gene transcription of WNT3A. In silico analysis of the GR promoter revealed the presence of T-box consensus binding sites (Fig. 5A). We performed ChIP in C4-2B cells followed by q-PCR analyses which revealed the presence of two TBX2 binding sites on the GR promoter, at −710 bp and −4854 bp upstream of the TSS, with significant enrichments of TBX2 binding of 12-fold and 8-fold, respectively (Fig. 5A depicted as red highlighted boxes, 5B, C). These results support that TBX2 directly activates GR transcription. Since 1) we observed a greater level of GR downregulation at the protein level compared to the mRNA level upon DN-based activity disruption of TBX2 in CRPC cells (Fig. 4B), and 2) previous reports have shown that some TFs can physically interact with proteins leading to their activation [44] - we investigated if TBX2 might also interact with the GR through a protein-protein interaction. A co-immunoprecipitation (CoIP) assay using C4-2B, DU145, and LNCaP enzalutamide-resistant (LNCaPEnzaR) human CRPC cell lines revealed an interaction between TBX2 and GR (Fig. 5D). Together, these results support that TBX2 regulates the GR via dual modes: through direct transcriptional activation and via protein-protein interactions (Fig. 5F).
A In silico analysis of the GR promoter showing potential TBX2 binding sites; B, C ChIP assay in C4-2B cells followed by q-PCR analysis showing TBX2 binding on the GR promoter with a significant enrichment of 12-fold and 8-fold respectively at −710 bp and −4854 bp upstream of the TSS; Data represent the average of triplicates ± S.D; Student’s unpaired 2-tailed t-tests were performed to compare the two groups **p ≤ 0.01; ***p ≤ 0.001; and ****p ≤ 0.0001; D Western blot analysis of the CoIP in C4-2B, DU145 and LNCaPEnzaR cells showing TBX2 protein-protein interactions with the GR. The whole cell extract (WCE) was immunoprecipitated with an anti-TBX2 antibody with species matched IgG (control) and samples were immunoblotted for anti-GR antibody; E Western blot analysis of CoIP in 22Rv1 cells showing no interaction of TBX2 with ARV7 or AR. The WCE was immunoprecipitated with an anti-TBX2 antibody with species matched IgG (control) and samples were immunoblotted for anti-ARV7 or anti-AR antibodies; F Schematic representing transcriptional activation of GR by TBX2 through directly binding to the GR promoter, and TBX2-GR protein-protein interactions.
Administration of enzalutamide is currently standard of care for advanced PCa patients since almost all patients develop resistance to first line ADT. The generation of AR splice variants (AR-Vs) has been attributed in large part due to the insensitivity of CRPC to enzalutamide treatment [10]. We therefore investigated whether TBX2 interacted with the ARV7 or the AR proteins. In contrast to the GR, CoIP experiments failed to show any interaction between TBX2 and the AR or variant ARV7 protein (Fig. 5E).
TBX2 driven switch from AR to GR signaling confers enzalutamide resistance
We next investigated if the AR to GR switch driven by TBX2 can confer enzalutamide resistance. Genetic modulation was used for DN-based activity disruption of TBX2 in the PC3 and C4-2B human CRPC cells (PC3TBX2DN and C4-2BTBX2DN). Conversely, TBX2 was over-expressed in enzalutamide-sensitive LNCaP human PCa cells. Western blot analysis revealed increased expression of TBX2 and GR in LNCaPEnzaR cells compared to LNCaP parental control cells (Supplementary Fig. S3A–C). Further cell viability assay results also revealed that TBX2 is associated with enzalutamide resistance (Supplementary Fig. S4A). Of note, PC3 cells have been reported to be resistant to enzalutamide treatment due to the absence of AR expression [45, 46]. DN-based activity disruption of TBX2 in PC3 cells (PC3TBX2DN) resulted in strong AR expression (as shown previously in Fig. 1C) and a restoration of enzalutamide sensitivity (Supplementary Fig. S4A). Conversely, overexpression of TBX2 in LNCaPTBX2OE cells decreased sensitivity to enzalutamide when compared to controls (Supplementary Fig. S4A). [NOTE: as shown previously, overexpression of TBX2 in LNCaPTBX2OE cells reduced AR expression compared to controls (Fig. 1C)]. We then investigated if siRNA knockdown of elevated GR levels (due to increased TBX2 levels) would increase enzalutamide sensitivity in LNCaPTBX2OE cells (resistant to enzalutamide treatment). Results revealed GR knockdown restored enzalutamide sensitivity to LNCaPTBX2OE cells compared to controls (Fig. 6A–C).
A q RT-PCR analysis showing decreased GR mRNA levels in LNCaPTBX2OE cells that were transfected with si-GR compared with si-controls, with or without enzalutamide treatment; B Cell viability assay showing increased enzalutamide sensitivity in LNCaPTBX2OE cells that were transfected with si-GR compared with si-controls. One-way ANOVA was performed where (n = 3), ***p < 0.001;****, p < 0.0001; C Schematic representing that GR knockdown in LNCaPTBX2OE cells leads to reversal to enzalutamide sensitivity; D, E Prostatospheres (PSs) formed from C4-2BNTSCR cells showing no significant changes in the number or size of PSs formed upon enzalutamide treatment either in the presence or absence of dexamethasone. One-way ANOVA was performed where (n = 3), ***p < 0.001;****, p < 0.0001; F, G Prostatospheres (PSs) formed from C4-2BshTBX2 cells showing a significant decrease in the number of PSs formed upon enzalutamide treatment, and a significant increase in the size and number of PSs formed upon dexamethasone treatment which failed to respond to continued enzalutamide treatment. One-way ANOVA was performed where (n = 3), ***p < 0.001;****, p < 0.0001; H Schematic representing that activation of GR in C4-2BshTBX2 cells leads to reversal to enzalutamide resistance.
We also tested if the GR agonist, dexamethasone, could reinstate enzalutamide resistance in C4-2B cells in which GR expression had been reduced by shRNA knockdown of TBX2 (C4-2BshTBX2 cells) or using the TBX2 DN construct (C4-2BTBX2DN). Prostatospheres (PSs) formed by control cells (C4-2BNTSCR or C4-2BNeo), which express elevated levels of TBX2 and GR, were unresponsive to enzalutamide treatment with or without dexamethasone (Fig. 6D, E, Supplementary Fig. S4C). In contrast, PSs formed by C4-2BshTBX2 or C4-2BTBX2DN cells displayed a reduction in both the number and size of spheres formed compared to C4-2BNTSCR or C4-2BNeo controls (Supplementary Figs. S4B, S4D). Enzalutamide treatment of C4-2BshTBX2 cell (Fig. 6F, G) or C4-2BTBX2DN cells (Supplementary Fig. S4D) reduced the number of spheres formed. However, dexamethasone treatment of C4-2BshTBX2 cells (Fig. 6F, G, H) or C4-2BTBX2DN cells (Supplementary Fig. S4D) increased both number and size of PSs even in the presence of enzalutamide suggesting that increased GR activity, even in the context of reduced TBX2, can confer enzalutamide resistance. Together, these results support that increased GR activity as a result of increased TBX2 significantly contributes to enzalutamide resistance (Fig. 6C, H), i.e. the TBX2-driven switch from AR- to GR- steroid signaling confers resistance to enzalutamide in PCa.
Correlation analyses of the transcriptional activities of TBX2, AR, and GR demonstrate consistent associations across different PCa patient cohorts
To investigate the relationships among TBX2, AR and GR in PCa patient cohorts, we performed pairwise correlative analyses between the transcriptional activities of these three TFs. A previous report [47] has demonstrated that compared to gene expression data, calculating the transcriptional regulatory activities of TFs using the bioinformatics-based Virtual Interference of Protein-activity by Enriched Regulon analysis (VIPER) method provides an accurate assessment of TF protein activities. We utilized two different CRPC cohorts - SU2C East Coast cohort (n = 280) and SU2C West Coast cohort (n = 83) - to estimate transcriptional regulatory activity levels of TBX2, AR, and GR (see methods). Spearman correlation analysis of TF activity levels revealed consistent and highly significant correlations in both cohorts: 1) a negative correlation between TBX2 and AR (Fig. 7A, D), 2) a positive correlation between TBX2 and GR (Fig. 7B, E), and 3) a negative correlation between AR and GR (Fig. 7C, F). Further, we observed a similar correlation in untreated, hormone-naïve primary PCa patients from the Swedish Watchful Waiting cohort (n = 281, Fig. 7G–I), suggesting that the relationships among TBX2, AR and GR could serve as a potential predictor of intrinsic (primary) resistance to enzalutamide in primary tumors. In summary, the relationships among TBX2, AR and GR proteins revealed by our transcriptional regulation experiments along with genetic manipulation of TBX2 in multiple human PCa/CRPC cell-line models: 1) prostate adenocarcinoma (LNCaP), 2) CRPC (C4-2B, PC3 and DU145), and 3) enzalutamide-resistant CRPC (LNCaPEnzaR) - were in direct agreement with the consistent associations among the transcriptional/protein activities of these TFs across different patient cohorts of PCa and CRPC. These findings underscore the translational relevance of our studies.
Spearman correlation analysis was conducted using the activity levels of transcription factors TBX2, AR, and GR (NR3C1) in the following cohorts: (A–C) SU2C East Coast cohort (n = 280); (D–F) SU2C West Coast cohort (n = 83); and (G–I) Swedish Watchful Waiting cohort (n = 281). Activity levels of transcriptional regulators were estimated using VIPER analysis. Rho and p-values for the Spearman correlation analysis are indicated, and the fitted linear regression line is shown in red.
SP2509, an allosteric LSD1 inhibitor, disrupts TBX2-GR and TBX2-LSD1 protein-protein interactions
Having identified that TBX2 novelly interacts with GR, we attempted to pharmacologically disrupt the TBX2-GR interaction. Previous studies in breast cancer cell lines identified an interaction between TBX2 and LSD1 as a part of the COREST complex [30]. It was also reported that this interaction could be pharmacologically disrupted using the LSD1 inhibitor, SP2509 [30]. Significantly, disruption of the TBX2-LSD1 interaction resulted in the inhibition of TBX2 downstream targets [30]. Further, and of relevance to our studies, it has also been reported that LSD1 is upregulated in CRPC and is associated with enzalutamide resistance [48, 49] (Supplementary Fig. S3A–C). Therefore, we tested whether SP2509 might also disrupt TBX2-GR interaction. After optimizing the dosing of SP2509, we performed CoIP in SP2509 treated C4-2B and LNCaPEnzaR human CRPC cells. We observed that SP2509 disrupted the interaction between TBX2 and GR in addition to disrupting the TBX2-LSD1 interaction compared to controls (Fig. 8A, C). Further, we also observed that protein expression of TBX2, GR and LSD1 were decreased in SP2509-treated C4-2B and LNCaPEnzaR cells compared to controls (Fig. 8B, D). To evaluate the effects of genetic modulation of TBX2 in the context of enzalutamide resistance, we performed an sh-RNA-mediated knockdown of TBX2 in LNCaPEnzaR cells to generate LNCaPEnzaR shTBX2 cells. LNCaPEnzaR cells express elevated levels of TBX2, GR and LSD1 compared to LNCaP cells (Supplementary Fig. S3A–C). We then assessed the effects of knocking down TBX2 in LNCaPEnzaR on GR expression, and potential proliferation effects in the presence of SP2509. Western blot analyses showed that: a) GR levels were markedly reduced in LNCaPEnzaR shTBX2 cells compared to controls, and b) SP2509 showed a greater efficacy in reducing the protein levels of both TBX2 and GR in LNCaPEnzaR shTBX2 cells compared to LNCaPEnzaR NTSCR control cells (Fig. 8E). In agreement with the Western blot analyses, SP2509 displayed a significantly greater efficacy in reducing cell proliferation in LNCaPEnzaR shTBX2 cells compared to LNCaPEnzaR NTSCR control cells (Fig. 8F). Together, these results suggest that: a) TBX2 and GR protein levels may have a cumulative effect on enzalutamide resistance, and b) in accordance with previous reports [50] demonstrating that protein function is intimately linked to interaction with other proteins, the stability of TBX2 and GR proteins can be affected by SP2509. In summary, these results provide key insights into a potential therapeutic modality for targeting the AR- to GR- signaling switch via disruption of TBX2-GR and TBX2-LSD1 protein-protein interactions.
A Western blot analyses of the CoIP in SP2509-treated (1 µM) C4-2B cells showing disruption of TBX2-GR and TBX2-LSD1 protein-protein interactions. Samples were immunoprecipitated with an anti-TBX2 antibody using species matched control IgG and immunoblotted for anti-GR and anti-LSD1 antibodies; B Western blot analyses of SP2509-treated (1 µM) C4-2B cells showing reduced expression of GR, TBX2, and LSD1; C Western blot analyses of the CoIP in SP2509-treated (2 µM) LNCaPEnzaR cells showing disruption of TBX2-GR and TBX2-LSD1 protein-protein interactions. Samples were immunoprecipitated with an anti-TBX2 antibody using species matched control IgG and immunoblotted for anti-GR and anti-LSD1 antibodies; D Western blot analyses of SP2509-treated (2 µM) LNCaPEnzaR cells showing reduced expression of GR, TBX2, and LSD1; E Western blot analyses of SP2509- (2 µM) or DMSO-treated LNCaPEnzaR shTBX2 cells or the control LNCaPEnzaR NTSCR cells showing a greater reduction of TBX2 and GR levels in LNCaPEnzaR shTBX2 compared with LNCaPEnzaR NTSCR control cells; F Cell proliferation assay in LNCaPEnzaR shTBX2 cells and LNCaPEnzaR NTSCR control cells at 72 and 96 h in the presence of enzalutamide (20 µM) and SP2509 (1 µM and 2 µM) showing significantly reduced proliferation in LNCaPEnzaR shTBX2 cells compared with LNCaPEnzaR NTSCR control cells.
Discussion
Recent evidence has demonstrated that acquired resistance to androgen deprivation therapy in PCa is often due to loss/downregulation of the AR and upregulation of the GR [5,6,7,8,9, 16]. The GR, a steroid receptor akin to the AR, shares a significant homology to the AR DNA binding domain, including an overlapping target gene set [6, 8, 16, 21, 22]. This has led to the hypothesis that 2nd generation ARSIs, such as enzalutamide, by unknown mechanism(s), unintendedly select for the functional bypass of AR signaling via increased GR activity [5,6,7,8,9]. Previous reports have shown that GR is negatively regulated by the AR [6, 12, 13].
In an effort to identify drivers of PCa progression, in a previous study, we determined that TBX2, a developmental master regulator, is over-expressed in CRPC [26] and drives bone metastatic progression [26]. A recent report confirmed that TBX2 and GR are two of the four TFs that drive enzalutamide resistance in advanced PCa [29].
In our current study, we report that TBX2 is the molecular driver of the AR- to GR- signaling switch in CRPC (Fig. 9A). Through transcriptional regulation experiments and various approaches to genetically modulate TBX2 expression in vitro using: 1) LNCaP, an AR-positive human prostate adenocarcinoma cell line; 2) C4-2B, an AR-positive human CRPC cell line; 3) PC3 and DU145, AR-negative human CRPC cell lines; and 4) LNCaPEnzaR, an AR-positive enzalutamide-resistant human CRPC cell line – our studies demonstrated that TBX2 downregulates AR while upregulating GR. Mechanistically, we found that TBX2 is a key regulator of AR expression in PCa by directly binding to and repressing the promoters of both AR and its co-regulator GATA2. Moreover, TBX2 upregulates GR by binding to the GR promoter and interacting with GR via protein-protein interactions. Among these two mechanisms, we observed that the upregulation of GR by TBX2 is more prominent at the protein level. Using complementary genetic approaches to modulate TBX2 expression, we showed that TBX2 drives the switch from AR to GR signaling thereby conferring enzalutamide resistance in CRPC. The clinical significance of these findings is underscored by the consistent pairwise correlations in transcriptional activity among TBX2, AR, and GR across two different CRPC patient cohorts. Moreover, these results are translationally significant, as similar correlations were observed in untreated, hormone-naïve prostate tumors. Notably, previous studies [8, 18,19,20, 51] have reported that: 1) GR is expressed at low levels in untreated hormone-naïve PCa, and 2) acute AR inhibition leads to GR upregulation in a subset of CRPC cells. Our findings, therefore, suggest that the relationships among TBX2, AR and GR could serve as predictive markers for identifying tumor subpopulations in untreated primary PCa that either have intrinsic resistance to enzalutamide or are likely to develop resistance over time.
A TBX2 acts as a switch to turn OFF AR on one hand and turn ON GR on the other to orchestrate resistance to enzalutamide in CRPC; B SP2509, an allosteric inhibitor of LSD1 could be utilized to disrupt both TBX2-GR and TBX2-LSD1 protein-protein interactions thereby reinstating sensitivity to ARSIs.
Transcription factors, including TBX2, have traditionally been considered “undruggable” targets [52]. However, recent reports [30] have highlighted that TBX2 target specificity is dependent to a large extent on its protein-protein interactions - providing crucial insights into therapeutic strategies for targeting TBX2 via disruption of its protein partners (Fig. 9B). These reports evidence that TBX2 is part of the CoREST protein complex and interacts with lysine-specific demethylase 1 (LSD1) [30]. LSD1 has canonical demethylase activity associated with AR co-activation [53]. However, LSD1 also promotes PCa survival phenotype through demethylase-independent function(s) [49] associated with AR independence or castrate levels of androgens [49]. SP2509 is a small molecule allosteric inhibitor of the demethylase-independent functions of LSD1. In CRPC cell lines, SP2509 bound a pocket distinct from that bound by LSD1 catalytic inhibitors and disrupted interaction of LSD1 with ZNF217, another member of the COREST complex, to reverse castrate resistance [49]. In breast cancer cell lines [30], SP2509 disrupted TBX2-LSD1 interactions resulting in de-repression of TBX2 targets. Also, in that study, the silencing of LSD1 and its protein partners, including TBX2, emulated the effect of TBX2 knock-down and resulted in de-repression of TBX2 targets. Together, these data support the potential clinical utility of specifically disrupting TBX2-LSD1 interactions. Importantly, various LSD1 inhibitors, including a clinical analog of SP2509, are currently in clinical trials for the treatment of cancers [54].
In summary, our study demonstrates that TBX2 is a molecular switch that controls the AR to GR signaling bypass to confer enzalutamide resistance. Further, our study provides insight into potential therapeutic modalities to effectively target and reverse the AR to GR signaling switch, such as via disruption of TBX2-LSD1 and/or TBX2-GR protein-protein CoREST interactions (Fig. 9B). The translational significance of our findings is underscored by the long-standing efforts to develop pharmaceuticals capable of re-sensitizing CRPC to ARSIs. Indeed, in this regard, detailed molecular characterization of TBX2-LSD1/TBX2-GR protein complex interactions (Fig. 9B) would enable more effective drug discovery efforts aimed at disrupting TBX2-driven, AR- to GR- signaling. And given that long-term direct systemic suppression of GR signaling can result in both significant toxicities and potentially deleterious effects on PCa treatment [15, 25, 55], our current findings may demonstrate an indirect and potent molecular mechanism of GR inhibition to restore enzalutamide sensitivity in CRPC while avoiding the harmful effects of direct GR ablation.
Materials and methods
Cell lines and growth conditions
LNCaP [56, 57], C4-2B, 22Rv1, PC3 and DU145 cell lines were obtained from the laboratory of Dr. Leland W. K. Chung, Uro-Oncology research program, Department of Medicine, Cedars-Sinai Medical Center, Los Angeles, CA. LNCaP enzalutamide-resistant (LNCaPEnzaR) cells were obtained from the laboratories of Drs. Robert J. Matusik and Renjie Jin, Vanderbilt University Medical Center (VUMC), Nashville, TN. All cells were maintained at 37 °C in humidified 5% CO2, ambient oxygen incubator using either Dulbecco’s Modified Eagle Medium (DMEM, Corning Cat # 10013CV) or Roswell Park Memorial Institute medium −1640 (RPMI-1640 Corning, Cat #10040CV) supplemented with 10% Fetal Bovine Serum (FBS, Gibco, Cat#16000044) and 1% penicillin/streptomycin (Corning, Cat#30-002-Cl). The cell lines used in this study were evaluated for mycoplasma contamination in-house.
RNA isolation, cDNA synthesis and quantitative Real Time RT-PCR (qRT-PCR)
Total RNA was isolated from PCa cells using the RNAse easy mini kit by following the manufacture’s protocol (Qiagen Inc, Valencia, CA, USA). This was followed by cDNA synthesis using 2 µg of RNA to transcribe into cDNA using the high-capacity RNA to cDNA kit (Applied Biosystems, Carlsbad, CA, USA). qRT-PCR was performed to check the change in mRNA expression using powerup SYBR green master mix (#A25742, Life Technologies, Carlsbad, CA) and specific gene primers synthesized from IDT (IDT, Coralville, IA). Quantstudio 12k Flex qRT-PCR platform was used to perform the qRT-PCR. The relative amount of mRNA expression was normalized to β-actin. The list of primers used is provided in supplementary information (Supplementary Table S1).
Cell lysate preparation and Western blot analysis
Total protein lysates from PCa cells were isolated followed by protein quantification using the protein DC assay kit (Bio-rad,Cat# 5000-001) as described previously [27]. Equal amounts (65 µg) of protein were resolved onto 10% SDS-PAGE gel and electroblotted onto PVDF membrane (Sigma Aldrich, Cat# IPVH00005). The membranes were blocked with 5% BSA (Boston Bioproducts, Cat# P-753) in TBST (0.1% Tween) (Boston Bioproducts, BM-300) and incubated with the respective primary antibodies at 1:1000 dilution overnight at 4 °C. The list of the antibodies used is provided in the supplementary information (Supplementary Table S1). Subsequent to washing with TBST (0.1% Tween), the membranes were probed with HRP-conjugated secondary antibodies at 1:2000 dilution (rabbit/mouse, Cell Signaling Technology) for 1h at RT. The bands were visualized using west Pico chemiluminescent kit (#34580, Thermo Fisher Scientific, Rockford, IL), and Gel Doc System (Azure Biosystems, CA, USA and the Chemi-doc touch imaging system,Bio-Rad, CA, USA). β-Actin was used as the loading control.
Chromatin immunoprecipitation (ChIP) assay
JASPAR [58], an in silico analysis tool, in combination with published literature was used to predict the T-box consensus binding sequences on the AR, GATA2 and GR promoters [37]. Chromatin immunoprecipitation (ChIP) assays were performed as described previously [27] with minor modifications. Briefly, C4-2B cells were grown to approximately 80% confluency in 150 mm tissue culture plates. For crosslinking protein to the DNA, formaldehyde was added at a final concentration of 1% at room temperature (RT) and incubated for 8 min The reaction was quenched by glycine at a final concentration of 0.125M for 5 min at RT. The media was removed under a fume hood and the cells were washed 3X with cold PBS. Cells were harvested in PBS containing 1X protein phosphatase inhibitor (PPI) cocktail (Pierce, Rockford, IL, USA) and centrifuged briefly at 300 g for 2 min to remove the PBS. The cells were lysed using ChIP lysis buffer (50 mM Tris-HCl, pH8.0, 10 mM 0.5 M EDTA, and 1% SDS) supplemented with 1X PPI. The samples were sonicated for 11 cycles 30 secs on and off in a bioruptor (Diagenode Inc. Denville, NJ, USA). The soluble chromatin fraction was quantified and ChIP was set using 100 ug of chromatin material and incubated for 2 h at 4 °C with anti-TBX2 antibody (1-2 µg) (SCBT, sc-514291) and mouse IgG (control) (Cell signaling technology, Danvers, MA, USA). This was followed by incubating the chromatin-antibody mixture with 11 µl of Dynabeads protein A/G (Invitrogen, #300D, Carlsbad, CA, USA) for 2 h at 4 °C in a rotator. The ChIP products were washed with low salt buffer (20 mM Tris-HCL, pH 8.1; 0.1% SDS, 1% TritonX-100, 2 mM EDTA and 150 mM NaCl) followed by washes with high salt buffer (20 mM Tris-HCL, pH 8.1; 0.1% SDS, 1% TritonX-100, 2 mM EDTA and 500 mM NaCl) and TE buffer (10 mM Tris-HCl, pH8.0 and 1 mM EDTA). The samples were reverse crosslinked overnight using 5M NaCl at 65 °C, followed by RNAse A (20 mg/ml) (Sigma-aldrich) digestion for 2 h at 37 °C followed by proteinase K (20 mg/ml) (Sigma-aldrich) digestion for 2 h at 55 °C. The ChIP DNA was purified using PCR purification kit (Qiagen, Hilden, Germany) and amplified by q-PCR using specific primer pairs listed in the supplementary information (Supplementary Table S1).
Site directed mutagenesis (SDM)
The AR promoter to 4 Kb upstream of transcription start site (TSS) was cloned in pMCS-Cypridina Luc Vector (#16149) procured from Custom DNA Constructs, Islandia, NY. Site directed mutagenesis (SDM) was performed by mutation using base substitution on the TBX2 binding sites (−82bp, −3598bp) [38]. The mutagenic primers (Supplementary Table S1) were designed using QuickChange primer design program at www.agilent.com/genomics/qcpd according to the guidelines outlined in the manufacturer’s protocol (QuickChange Lightning Site-Directed Mutagenesis kit, Cat# 210518, Agilent Technologies). This was followed by mutant strand synthesis by thermal cycling. Dpn1 digestion was performed at 37 °C for 5 min to digest the parental (non-mutated strand). Then transformation was performed as outlined in the manufacturer’s protocol (QuickChange Lightning Site-Directed Mutagenesis kit, Cat# 210518, Agilent Technologies). Transformed plates were incubated at 37° for >16 h and mutant colonies were observed the following day. The mutant colonies were further grown in bulk in LB-amp broth and plasmid isolation was performed in accordance with the manufacturer’s protocol (Qiagen Miniprep, Cat#27106).
Luciferase reporter assay
Luciferase (luc) reporter assays were performed using PierceTM Cypridina-Firefly Luciferase Dual Assay Kit (Cat#16183). Approximately 0.3 × 106 cells/ well were seeded and 1 ug plasmid DNA was used for transfection of the mutated Cypridina luc reporter plasmid and red firefly luc control plasmid (cat#16156), the cells were incubated for 16–72 h at 37 °C in a 5% CO2, ambient oxygen incubator. Luc reporter assays were performed as outlined in the manufacturer’s protocol (PierceTM Cypridina-Firefly Luciferase Dual Assay Kit, Cat# 16183). The reading of Cypridina luc signal was normalized to control red firefly luc signal and the fold change in luc activity was calculated.
Plasmid isolation, transduction and genetic manipulation of TBX2 and GR using sh-RNA/ si-RNA
The custom constructs for sh-TBX2 and its non-targeting scrambled RNA duplex siRNA control (NTSCR) were procured from Sigma Aldrich (RNAi single clones, Millipore sigma, Burlington, MA). These lentiviral constructs along with packaging plasmid (pCMV delta R8.2 dvpr, Addgene#8455) and envelope plasmid (pCMV-VSV-G, Addgene#8454) were used for lipofectamine 2000-mediated transfection of approximately 0.8 × 106 293FT cells to produce lentiviral pseudo typed particles. The supernatant containing viral particles was collected by centrifugation and passed through 0.45 µm filter after 36 h of transfection. The viral particles were mixed with fresh media of the target cells and infection with viral particles was repeated three times and selection was performed with neomycin at 500 µg/ml concentration. The si-RNA mediated knockdown of GR was performed using human si-GR construct (SCBT # sc35505) and control si-RNA (sc37007) was purchased from Santa Cruz Biotechnology (TX, USA). The si-RNA transfection was performed in LNCaPTBX2OE cells following manufacturers protocol (Santa Cruz Biotechnology, TX, USA). Subsequent to transfection, the cells were incubated for 18–24 h after which fresh 1X normal growth serum was added along with enzalutamide treatment (20 µM) (MDV3100, Cat#915087-33-1, Selleck Chemical LLC) and the cells were assayed for cell viability after 72 h of treatment.
Cell viability assay
The Cell Proliferation Reagent WST-1 was used for the quantification of cell proliferation and viability. Cell viability assays were performed according to standard manufacturer’s protocol (Cat# 05015944001, Roche Applied Science, Indianapolis, IN). Approximately 5 × 103 cells/well were seeded in 96-well plates in five replicates. This was followed by the enzalutamide (MDV3100, Cat#915087-33-1, Selleck Chemical LLC) treatment of PC3Neo/PC3TBX2DN, C4-2BNeo/C4-2BTBX2DN, LNCaPNeo/LNCaPTBX2OE at 80 µM, 20 µM and 20 µM concentrations respectively. Cell proliferation was examined at 48, 96, and 72 h using WST-1 reagent (Cat# 05015944001, Roche Applied Science, Indianapolis, IN). Absorbance was measured at 450 nm before saturation on the iMARK plate reader (Bio-Rad). For background control, 1X WST-1 was added in empty wells in triplicate and absorbance was subtracted and relative percentage of cell proliferation was calculated.
Co-immunoprecipitation (Co-IP)
Co-immunoprecipitation (Co-IP) was performed as described previously [59] with minor modifications. To perform the Co-IP, C4-2B and 22Rv1 cells were grown to approximately 75% confluency. This was followed by treatment of the cells if required with SP2509/DMSO. The Co-IP lysis buffer was prepared in 1X PPI. The cells were washed with cold PBS and the lysis buffer (25 mM tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, pH 8.0, 1% NP-40, 5% glycerol) was added and the cells were harvested. The samples were incubated on ice for 20–30 min followed by centrifugation at 13,200 rpm at 4 °C for 15 min The supernatant was collected as cell lysate and this was followed by protein quantification using Bio-Rad DC kit (Bio-Rad). Immunoprecipitation was set with a starting protein concentration of 2 mg/ml and using desired antibody for 2 h at 4 °C, followed by precipitation of the protein-antibody complex with Dynabeads protein G (Invitrogen, #300D, Carlsbad, CA, USA) for 2 h in 4 °C on the rotator. The list of antibodies used is provided in the supplementary information (Supplementary Table S1). The co-immunoprecipitated samples were washed with lysis buffer and eluted in 1X gel loading buffer.
Correlative analyses of the transcriptional regulatory activities of TBX2, AR and GR in PCa patient cohorts
To investigate the association of activity levels of TBX2, AR, and GR, we utilized three independent patient cohorts: 1) Stand Up to Cancer (SU2C) East Coast cohort [60]; 2) SU2C West Coast cohort [61]; and 3) Swedish Watchful Waiting cohort [62]. [(SU2C East Coast cohort (n = 280)] was downloaded from dbGaP (phs000915.v2.p2). This cohort consists of metastatic CRPC (mCRPC) samples obtained from fresh-frozen needle biopsies, profiled on the Illumina HiSeq 2500. Sequence Read Archive (SRA) files were downloaded from dbGaP using the prefetch command and converted to FASTQ format using the fastq-dump command from the SRA Toolkit (v10.8.2). The resulting FASTQ files were aligned to the human reference genome (hg19) using STAR aligner with the --quantMode option to generate raw count files, which were subsequently normalized using the DESeq2 [63] package in R. SU2C West Coast cohort (n = 83) was accessed through dbGaP (phs001648.v2.p1) and GDC portal (https://portal.gdc.cancer.gov/projects/WCDT-MCRPC). This cohort consists of mCRPC samples collected through fresh-frozen, image-guided core needle biopsies, profiled on either the Illumina HiSeq 2500 or NextSeq 500 platforms. This cohort was downloaded and processed as above. Swedish Watchful Waiting cohort (n = 281) was downloaded from the Gene Expression Omnibus (GEO, GSE16560). This cohort consist of untreated, hormone-naïve primary PCa samples obtained through transurethral resection of the prostate (TURP) or adenoma enucleation, corresponding to stage T1a-b tumors. The samples were analyzed using the 6k cDNA-mediated annealing, selection, ligation, and extension (DASL) microarray platform. The data was normalized using the cubic spline algorithm [64]. To estimate the activity levels of transcriptional regulators, we employed VIPER (Virtual Inference of Protein-activity by Enriched Regulon analysis) [47] analysis. PCa-specific transcriptional regulatory network (interactome), which consists of transcriptional regulators (transcription factors and co-factors) and their corresponding transcriptional targets was obtained from a previous report [65]. For VIPER analysis, gene expression profiles were z-scored so that the mean expression for each gene is 0 and the standard deviation is 1. For each sample, a single-sample signature was then defined by ranking genes based on their z-scores, which served as a reference signature in the VIPER analysis. The transcriptional targets of each transcriptional regulator from the PCa-specific interactome were used as query gene sets to assess the activity of the transcriptional regulators in individual samples. In VIPER analysis, normalized enrichment scores (NES) are calculated for each transcriptional regulator based on the enrichment of their transcriptional targets in each single-sample signature, and define activity levels of transcriptional regulators. Single-sample VIPER analysis was performed using VIPER function from R VIPER package in Bioconductor [47]. The estimated activity levels for TBX2, AR, and GR were utilized for Spearman correlation analysis, using the Spearman r method from the scipy.stats module in Python. The resulting correlations were visualized using scatter plots or regression plots and were added using the seaborn module.
Statistical analyses
All experiments were performed at least three times in biological replicates and data expressed as mean ± SD. Wherever appropriate, data were subjected to unpaired two-tailed Student’s t test for comparison of two groups and one-way ANOVA for three groups or more. Differences were considered statically significant when *p ≤ 0.05. Statistical calculations were executed using GraphPad prism9 where *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.
Data availability
The transcriptome data analyzed in this current study has been deposited in the NCBI Gene Expression Omnibus database (accession number GSE242282).
References
Karantanos T, Corn PG, Thompson TC. Prostate cancer progression after androgen deprivation therapy: mechanisms of castrate resistance and novel therapeutic approaches. Oncogene. 2013;32:5501–11.
Boevé LM, Hulshof MC, Vis AN, Zwinderman AH, Twisk JW, Witjes WP, et al. Effect on survival of androgen deprivation therapy alone compared to androgen deprivation therapy combined with concurrent radiation therapy to the prostate in patients with primary bone metastatic prostate cancer in a prospective randomised clinical trial: data from the HORRAD trial. Eur Urol. 2019;75:410–8.
Schröder F, Crawford E, Axcrona K, Payne H, Keane T. Androgen deprivation therapy: past, present and future. BJU Int. 2012;109:1–12.
Chandrasekar T, Yang JC, Gao AC, Evans CP. Mechanisms of resistance in castration-resistant prostate cancer (CRPC). Transl Androl Urol. 2015;4:365.
Montgomery B, Cheng HH, Drechsler J, Mostaghel EA. Glucocorticoids and prostate cancer treatment: friend or foe? Asian J Androl. 2014;16:354.
Sakellakis M, Flores LJ. Is the glucocorticoid receptor a key player in prostate cancer?: A literature review. Medicine. 2022;101:e29716.
Watson PA, Arora VK, Sawyers CL. Emerging mechanisms of resistance to androgen receptor inhibitors in prostate cancer. Nat Rev Cancer. 2015;15:701–11.
Arora VK, Schenkein E, Murali R, Subudhi SK, Wongvipat J, Balbas MD, et al. Glucocorticoid receptor confers resistance to antiandrogens by bypassing androgen receptor blockade. Cell. 2013;155:1309–22.
Li J, Alyamani M, Zhang A, Chang KH, Berk M, Li Z et al. Aberrant corticosteroid metabolism in tumor cells enables GR takeover in enzalutamide resistant prostate cancer. Elife. 2017;6:e20183.
Vellky JE, Ricke WA. Development and prevalence of castration-resistant prostate cancer subtypes. Neoplasia. 2020;22:566–75.
Formaggio N, Rubin MA, Theurillat J-P. Loss and revival of androgen receptor signaling in advanced prostate cancer. Oncogene. 2021;40:1205–16.
Chen S-Y, Wang J, Yu G-Q, Liu W, Pearce D. Androgen and glucocorticoid receptor heterodimer formation: a possible mechanism for mutual inhibition of transcriptional activity. J Biol Chem. 1997;272:14087–92.
Xie N, Cheng H, Lin D, Liu L, Yang O, Jia L, et al. The expression of glucocorticoid receptor is negatively regulated by active androgen receptor signaling in prostate tumors. Int J Cancer. 2015;136:E27–38.
Valle S, Sharifi N. Targeting Glucocorticoid Metabolism in Prostate Cancer. Endocrinology. 2021;162:bqab132.
Serritella AV, Shevrin D, Heath EI, Wade JL, Martinez E, Anderson A, et al. Phase I/II trial of enzalutamide and mifepristone, a glucocorticoid receptor antagonist, for metastatic castration-resistant prostate cancer. Clin Cancer Res. 2022;28:1549–59.
Hirayama Y, Sadar MD. Does increased expression of glucocorticoid receptor support application of antagonists to this receptor for the treatment of castration resistant prostate cancer? AME Med J. 2018;3:66.
Shah N, Wang P, Wongvipat J, Karthaus WR, Abida W, Armenia J, et al. Regulation of the glucocorticoid receptor via a BET-dependent enhancer drives antiandrogen resistance in prostate cancer. Elife. 2017;6:27861.001.
Puhr M, Hoefer J, Eigentler A, Ploner C, Handle F, Schaefer G, et al. The Glucocorticoid Receptor Is a Key Player for Prostate Cancer Cell Survival and a Target for Improved Antiandrogen Therapy. Clin Cancer Res. 2018;24:927–38.
Szmulewitz RZ, Chung E, Al-Ahmadie H, Daniel S, Kocherginsky M, Razmaria A, et al. Serum/glucocorticoid-regulated kinase 1 expression in primary human prostate cancers. Prostate. 2012;72:157–64.
Isikbay M, Otto K, Kregel S, Kach J, Cai Y, Vander Griend DJ, et al. Glucocorticoid receptor activity contributes to resistance to androgen-targeted therapy in prostate cancer. Horm Cancer. 2014;5:72–89.
Sahu B, Laakso M, Pihlajamaa P, Ovaska K, Sinielnikov I, Hautaniemi S, et al. FoxA1 Specifies Unique Androgen and Glucocorticoid Receptor Binding Events in Prostate Cancer CellsAR and GR Cistromes in Prostate Cancer Cells. Cancer Res. 2013;73:1570–80.
Narayanan S, Srinivas S, Feldman D. Androgen–glucocorticoid interactions in the era of novel prostate cancer therapy. Nat Rev Urol. 2016;13:47–60.
Molitch ME. Glucocorticoid receptor blockers. Pituitary. 2022;25:733–6.
Kach J, Conzen SD, Szmulewitz RZ. Targeting the glucocorticoid receptor in breast and prostate cancers. Sci Transl Med. 2015;7:305ps319.
Taplin ME, Manola J, Oh WK, Kantoff PW, Bubley GJ, Smith M, et al. A phase II study of mifepristone (RU‐486) in castration‐resistant prostate cancer, with a correlative assessment of androgen‐related hormones. BJU Int. 2008;101:1084–9.
Nandana S, Tripathi M, Duan P, Chu CY, Mishra R, Liu C, et al. Bone Metastasis of Prostate Cancer Can Be Therapeutically Targeted at the TBX2-WNT Signaling Axis. Cancer Res. 2017;77:1331–44.
Patel GK, Dutta S, Syed MM, Ramachandran S, Sharma M, Rajamanickam V, et al. TBX2 Drives Neuroendocrine Prostate Cancer through Exosome-Mediated Repression of miR-200c-3p. Cancers. 2021;13:5020.
Abrahams A, Parker MI, Prince S. The T‐box transcription factor Tbx2: its role in development and possible implication in cancer. IUBMB Life. 2010;62:92–102.
Zhang Z, Zhou C, Li X, Barnes SD, Deng S, Hoover E, et al. Loss of CHD1 Promotes Heterogeneous Mechanisms of Resistance to AR-Targeted Therapy via Chromatin Dysregulation. Cancer Cell. 2020;37:584–98. e511.
McIntyre AJ, Angel CZ, Smith JS, Templeman A, Beattie K, Beattie S, et al. TBX2 acts as a potent transcriptional silencer of tumour suppressor genes through interaction with the CoREST complex to sustain the proliferation of breast cancers. Nucleic Acids Res. 2022;50:6154–73.
Vance KW, Carreira S, Brosch G, Goding CR. Tbx2 is overexpressed and plays an important role in maintaining proliferation and suppression of senescence in melanomas. Cancer Res. 2005;65:2260–8.
Tai S, Sun Y, Squires JM, Zhang H, Oh WK, Liang CZ, et al. PC3 is a cell line characteristic of prostatic small cell carcinoma. Prostate. 2011;71:1668–79.
Norris JD, Chang C-Y, Wittmann BM, Kunder RS, Cui H, Fan D, et al. The homeodomain protein HOXB13 regulates the cellular response to androgens. Mol Cell. 2009;36:405–16.
Wang B, Lindley LE, Fernandez-Vega V, Rieger ME, Sims AH, Briegel KJ. The T box transcription factor TBX2 promotes epithelial-mesenchymal transition and invasion of normal and malignant breast epithelial cells. PLoS One. 2012;7:e41355.
Jariwala U, Prescott J, Jia L, Barski A, Pregizer S, Cogan JP, et al. Identification of novel androgen receptor target genes in prostate cancer. Mol Cancer. 2007;6:1–15.
Tan PY, Chang CW, Chng KR, Wansa KSA, Sung W-K, Cheung E. Integration of regulatory networks by NKX3-1 promotes androgen-dependent prostate cancer survival. Mol Cell Biol. 2012;32:399–414.
Conlon FL, Fairclough L, Price BM, Casey ES, Smith J. Determinants of T box protein specificity. 2001. Development. 2001;128:3749–58.
Carreira S, Dexter T, Yavuzer U, Easty D, Goding C. Brachyury-related transcription factor Tbx2 and repression of the melanocyte-specific TRP-1 promoter. Mol Cell Biol. 1998;18:5099–108.
Rodriguez-Bravo V, Carceles-Cordon M, Hoshida Y, Cordon-Cardo C, Galsky MD, Domingo-Domenech J. The role of GATA2 in lethal prostate cancer aggressiveness. Nat Rev Urol. 2017;14:38–48.
Zhao JC, Fong K-W, Jin H-J, Yang YA, Kim J, Yu J. FOXA1 acts upstream of GATA2 and AR in hormonal regulation of gene expression. Oncogene. 2016;35:4335–44.
He B, Lanz RB, Fiskus W, Geng C, Yi P, Hartig SM, et al. GATA2 facilitates steroid receptor coactivator recruitment to the androgen receptor complex. Proc Natl Acad Sci. 2014;111:18261–6.
Wu D, Sunkel B, Chen Z, Liu X, Ye Z, Li Q, et al. Three-tiered role of the pioneer factor GATA2 in promoting androgen-dependent gene expression in prostate cancer. Nucleic Acids Res. 2014;42:3607–22.
Paxton C, Zhao H, Chin Y, Langner K, Reecy J. Murine Tbx2 contains domains that activate and repress gene transcription. Gene. 2002;283:117–24.
Liberati NT, Datto MB, Frederick JP, Shen X, Wong C, Rougier-Chapman EM, et al. Smads bind directly to the Jun family of AP-1 transcription factors. Proc Natl Acad Sci. 1999;96:4844–9.
Smith R, Liu M, Liby T, Bayani N, Bucher E, Chiotti K, et al. Enzalutamide response in a panel of prostate cancer cell lines reveals a role for glucocorticoid receptor in enzalutamide resistant disease. Sci Rep. 2020;10:1–13.
Barrado M, Blanco-Luquin I, Navarrete PA, Visus I, Guerrero-Setas D, Escors D, et al. Radiopotentiation of enzalutamide over human prostate cancer cells as assessed by real-time cell monitoring. Rep. Practical Oncol Radiother. 2019;24:221–6.
Alvarez MJ, Shen Y, Giorgi FM, Lachmann A, Ding BB, Ye BH, et al. Functional characterization of somatic mutations in cancer using network-based inference of protein activity. Nat Genet. 2016;48:838–47.
Ellis L, Loda M. LSD1: A single target to combat lineage plasticity in lethal prostate cancer. Proc Natl Acad Sci. 2018;115:4530–1.
Sehrawat A, Gao L, Wang Y, Bankhead A 3rd, McWeeney SK, King CJ, et al. LSD1 activates a lethal prostate cancer gene network independently of its demethylase function. Proc Natl Acad Sci USA. 2018;115:E4179–e4188.
Fields PA. Review: Protein function at thermal extremes: balancing stability and flexibility. Comp Biochem Physiol A Mol Integr Physiol. 2001;129:417–31.
Yemelyanov A, Bhalla P, Yang X, Ugolkov A, Iwadate K, Karseladze A, et al. Differential targeting of androgen and glucocorticoid receptors induces ER stress and apoptosis in prostate cancer cells: a novel therapeutic modality. Cell Cycle. 2012;11:395–406.
Crespo-García T, Cámara-Checa A, Dago M, Rubio-Alarcón M, Rapún J, Tamargo J, et al. Regulation of cardiac ion channels by transcription factors: Looking for new opportunities of druggable targets for the treatment of arrhythmias. Biochemical Pharmacol. 2022;204:115206.
Cai C, He HH, Gao S, Chen S, Yu Z, Gao Y, et al. Lysine-specific demethylase 1 has dual functions as a major regulator of androgen receptor transcriptional activity. Cell Rep. 2014;9:1618–27.
Welch D, Kahen E, Fridley B, Brohl AS, Cubitt CL, Reed DR. Small molecule inhibition of lysine-specific demethylase 1 (LSD1) and histone deacetylase (HDAC) alone and in combination in Ewing sarcoma cell lines. PloS One. 2019;14:e0222228.
Abida W, Hahn AW, Shore N, Agarwal N, Sieber P, Smith MR, et al. Phase I Study of ORIC-101, a Glucocorticoid Receptor Antagonist, in Combination with Enzalutamide in Patients with Metastatic Castration-resistant Prostate Cancer Progressing on Enzalutamide. Clin Cancer Res. 2024;30:1111–20.
Thalmann GN, Anezinis PE, Chang S-M, Zhau HE, Kim EE, Hopwood VL, et al. Androgen-independent cancer progression and bone metastasis in the LNCaP model of human prostate cancer. Cancer Res. 1994;54:2577–81.
Wu TT, Sikes RA, Cui Q, Thalmann GN, Kao C, Murphy CF, et al. Establishing human prostate cancer cell xenografts in bone: Induction of osteoblastic reaction by prostate‐specific antigen‐producing tumors in athymic and SCID/bg mice using LNCaP and lineage‐derived metastatic sublines. Int J Cancer. 1998;77:887–94.
Castro-Mondragon JA, Riudavets-Puig R, Rauluseviciute I, Berhanu Lemma R, Turchi L, Blanc-Mathieu R, et al. JASPAR 2022: the 9th release of the open-access database of transcription factor binding profiles. Nucleic Acids Res. 2022;50:D165–D173.
Sharma M, Molehin D, Castro-Piedras I, Martinez EG, Pruitt K. Acetylation of conserved DVL-1 lysines regulates its nuclear translocation and binding to gene promoters in triple-negative breast cancer. Sci Rep. 2019;9:16257.
Robinson D, Van Allen EM, Wu YM, Schultz N, Lonigro RJ, Mosquera JM, et al. Integrative clinical genomics of advanced prostate cancer. Cell. 2015;161:1215–28.
Quigley DA, Dang HX, Zhao SG, Lloyd P, Aggarwal R, Alumkal JJ, et al. Genomic Hallmarks and Structural Variation in Metastatic Prostate Cancer. Cell. 2018;174:758–69. e759.
Sboner A, Demichelis F, Calza S, Pawitan Y, Setlur SR, Hoshida Y, et al. Molecular sampling of prostate cancer: a dilemma for predicting disease progression. BMC Med Genomics. 2010;3:8.
Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11:R106.
Hoshida Y, Villanueva A, Kobayashi M, Peix J, Chiang DY, Camargo A, et al. Gene expression in fixed tissues and outcome in hepatocellular carcinoma. N Engl J Med. 2008;359:1995–2004.
Aytes A, Mitrofanova A, Lefebvre C, Alvarez MJ, Castillo-Martin M, Zheng T, et al. Cross-species regulatory network analysis identifies a synergistic interaction between FOXM1 and CENPF that drives prostate cancer malignancy. Cancer Cell. 2014;25:638–51.
Acknowledgements
The authors are grateful to Barry Maurer (TTUHSC), Gail Cornwall (TTUHSC), and Robert Matusik (VUMC) for their critical reading of the manuscript. The authors would like to thank Sambantham Shanmugam (TTUHSC) for his assistance with the luciferase reporter assay.
Funding
This research was funded by the following grants: NIH NLM R01LM013236, ACS RSG-21-023-01-TBG, DOD Data Science HT94252410346 (to AM); DoD-PCRP W81XWH-17-1-0353 and The CH Foundation (to MT); and DoD-PCRP W81XWH-16-1-0174, The CH Foundation, The Ted Nash Long Life Foundation, and the CPRIT-TREC RP210154 (to SN).
Author information
Authors and Affiliations
Contributions
SD, MT, and SN conceived the project and designed the study. The development of methodology was carried out by SD, GKP, MT, and SN. Data acquisition was conducted by SD, HK, GKP, DL, JW, MT, and SN. Data analysis and interpretation (e.g., statistical analysis, biostatistics, computational analysis) was performed by SD, GKP, MM, AM, MT, and SN. The manuscript was written, reviewed, and revised by SD, AM, MT, and SN. Study supervision was provided by MT and SN. Administrative, technical, or material support (such as data organization and database construction) was offered by SD, HK, MT, and SN.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethics approval and consent to participate
All methods were performed in accordance with the relevant guidelines and regulations. This study did not use any human subjects, vertebrate animals or identifiable images from human research patients.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Dutta, S., Khedmatgozar, H., Patel, G.K. et al. A TBX2-driven signaling switch from androgen receptor to glucocorticoid receptor confers therapeutic resistance in prostate cancer. Oncogene 44, 877–892 (2025). https://doi.org/10.1038/s41388-024-03252-5
Received:
Revised:
Accepted:
Published:
Issue date:
DOI: https://doi.org/10.1038/s41388-024-03252-5
