High-Throughput Screen for Inhibitors of Androgen Receptor-RUNX2 Transcriptional Regulation in Prostate Cancer S
Total Page:16
File Type:pdf, Size:1020Kb
Supplemental material to this article can be found at: http://jpet.aspetjournals.org/content/suppl/2016/08/23/jpet.116.234567.DC1 1521-0103/359/2/256–261$25.00 http://dx.doi.org/10.1124/jpet.116.234567 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS J Pharmacol Exp Ther 359:256–261, November 2016 Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics High-Throughput Screen for Inhibitors of Androgen Receptor-RUNX2 Transcriptional Regulation in Prostate Cancer s Winston Vuong, Ben Y. Tew, Gillian H. Little, Baruch Frenkel, and Jeremy O. Jones Beckman Research Institute, City of Hope, Duarte, California (W.V., B.Y.T., J.O.J.); Keck School of Medicine, University of Southern California, Los Angeles, California (G.H.L., B.F.) Received June 23, 2016; accepted August 12, 2016 Downloaded from ABSTRACT Runt-related transcription factor 2 (RUNX2) plays a critical role in We therefore established a phenotypic cell-based screening prostate cancer progression. RUNX2 interacts with the androgen assay for compounds that could inhibit AR-RUNX2 synergistic receptor (AR) and modulates its transcriptional activity in a locus- activity either directly or indirectly. This assay was used to screen specific manner. RUNX2 and AR synergistically stimulate a 880 compounds as a proof of concept, resulting in identification subset of genes, including the pro-oncogene snail family zinc of several compounds that disrupted the synergistic stimulation jpet.aspetjournals.org finger 2 (SNAI2). AR-RUNX2 signaling cooperatively induces of genes. Further investigation suggested the involvement of invasiveness of prostate cancer cells via SNAI2; and coexpres- epidermal growth factor receptor (EGFR) signaling in AR/RUNX2 sion of AR, RUNX2, and SNAI2 in prostate cancer biopsy synergistic activity. Our assay is amenable to high-throughput samples predicts disease recurrence. Competitive inhibition of screening and can be used to identify inhibitors of the AR-RUNX2 AR alone could not disrupt the synergistic activation of SNAI2. interaction in prostate cancer cells. Introduction invasiveness and tissue destruction (Akech et al., 2010; at ASPET Journals on September 25, 2021 Baniwal et al., 2010). Runt-related transcription factor 2 (RUNX2) is a member of Androgen receptor (AR) signaling regulates prostatic ho- the mammalian RUNX family of transcription factors, which meostasis; however, its activity becomes oncogenic during have well-established roles in development and cancer (Blyth prostate carcinogenesis to promote cellular proliferation, et al., 2005; Ito, 2008). RUNX2, best known for its roles in survival, and aerobic glycolysis (Wang et al., 2009; Massie skeletal development (Komori et al., 1997), has been impli- et al., 2011). These changes, mediated by AR coactivators and cated in the evolution of breast and prostate cancer (PCa) collaborating transcription factors, are associated with re- metastasis, primarily through promotion of epithelial- distribution of AR across the genome and alterations to its mesenchymal transition (Pratap et al., 2008; Akech et al., transcriptional regulation (Massie et al., 2007; Wang et al., 2010; Baniwal et al., 2010). RUNX2 immunoreactivity is 2007; Jia et al., 2008; Baniwal et al., 2009; Sahu et al., 2011). higher in human PCa than in normal prostate epithelium RUNX2 has been implicated in AR signaling in physiologic and correlates with cancer grade and stage (Chua et al., 2009; and pathologic contexts, but the underlying mechanisms have Yun et al., 2012). Likewise, RUNX2 phosphorylation corre- not been fully defined (Kawate et al., 2007; Baniwal et al., lates with aggressive PCa (Ge et al., 2016). RUNX2 expression 2009; Frenkel et al., 2010). In PCa and other cell types, also increases during PCa development in a mouse PTEN physical interaction between AR and the RUNX2 DNA- conditional knockout model (Lim et al., 2010). RUNX2 activity binding domains was found to inhibit the ability of RUNX2 in PCa is negatively regulated by the tumor suppressor PTEN, to recruit and activate target genes (Kawate et al., 2007; which is often lost in PCa, again suggesting the importance of Baniwal et al., 2009; Frenkel et al., 2010; Baniwal et al., 2012). activating RUNX2 signaling as PCa progresses (Zhang et al., The reciprocal interaction, those of RUNX2 on AR, led to 2011). Furthermore, manipulation of RUNX2 in tissue cul- conflicting results indicating either inhibition (McCarthy ture and xenograft mouse models of PCa metastasis alters et al., 2003; Kawate et al., 2007) or stimulation (van der Deen et al., 2010; Baniwal et al., 2012) of AR activity. Recently, it was shown that RUNX2 modulates AR activity This work was solely supported by institutional funds. dx.doi.org/10.1124/jpet.116.234567. in PCa cells in a locus-dependent manner (Little et al., 2014). s This article has supplemental material available at jpet.aspetjournals.org. This work revealed several important facets of AR-RUNX2 ABBREVIATIONS: AR, androgen receptor; C4-2bRx2dox, doxycycline-inducible RUNX2 C4-2b cells; CSS, charcoal-dextran stripped serum; DHT, dihydrotestosterone; Dox, doxycycline; EGFR, epidermal growth factor receptor; MMTV, mouse mammary tumor virus; NSC659174, 1-chloro-6- [chloro(difluoro)methyl]-1,1,7,7,7-pentafluoro-2,6-dihydroxy-2-(trifluoromethyl)heptan-4-one; NSC689857, 1-adamantylmethyl 4-[(2,5- dihydroxyphenyl)methylamino]benzoate; PCa, prostate cancer; RT-PCR, reverse-transcription polymerase chain reaction; RUNX2, Runt-related transcription factor 2; siRNA, small-interfering RNA; SNAI2, snail family zinc finger 2; YFP, yellow fluorescent protein. 256 HTS for AR-RUNX2 Inhibitors 257 interaction that suggest a more nuanced RUNX2 influence on Library compounds were initially screened in duplicate between PCa progression. At most affected loci, RUNX2 antagonizes separate plates. Compounds selected as hits were screened in AR recruitment to, and stimulation of, target genes including quadruplicate for the dose–response validation screen. the NKX3.1 and TMPRSS2 tumor suppressors (type I genes). Hit Selection. Compounds were conditionally selected as hits In other cases, type IIa genes including the pro-oncogenes based upon binary categorization as either an attenuator or not, contingent on classification as minimally cytotoxic, on both duplicate snail family zinc finger 2 (SNAI2) and SRY-box 9 are synergis- plates. Attenuation was defined as having a lower YFP/mCherry tically up-regulated by RUNX2 and AR. Most importantly, the signal, Scompound, than 1.5 S.D. of the average YFP/mCherry signal of cooperative signaling appears to induce invasiveness of PCa via Dox 1 DHT–treated controls, Sreference (eq. 1). Compounds were SNAI2 on a physiologic level as coexpression of AR, RUNX2, determined to be minimally cytotoxic if their mCherry signals, and SNAI2 in PCa biopsies predicts disease recurrence. T , were greater than half of the average mCherry signal compound Based on these observations we developed a screen to identify across the plate, Tplate (eq. 2). Results from the hit selection were inhibitors of the AR-RUNX2 interactions and activation of plotted using Sʹ versus Tʹ, the normalized compound toxicity and prometastatic type IIa genes, especially SNAI2. Such com- modulation signals from both plates as shown in eqs. 3 and 4. Hits had ʹ ʹ pounds may have the potential to treat AR-RUNX2-driven PCa. an S lower than 1 and an R greater than 1. Conditions for hit selection were as follows: À Á S , S 2 1:5 SD (1) Materials and Methods compound reference Downloaded from dox Tplate Cell Culture. A modified human PCa cell subline, C4-2bRx , Tcompound . (2) dox 2 was used for all experiments (Baniwal et al., 2009). C4-2bRx cells were engineered from the LNCaP C4-2b cell line to conditionally Scompound 9 5 S ,1 (3) express RUNX2 at levels similar to the PC-3 PCa cell line and Sreference 2 1:5 SD osteoblast cells upon treatment with doxycycline (Dox). The cells were maintained in RPMI-1640 medium (Cellgro/Mediatech, Manassas, Tcompound 9 5 T .1 (4) VA) supplemented with penicillin–streptomycin and 10% fetal bovine Tplate jpet.aspetjournals.org 2 serum or 5% charcoal-dextran stripped serum (CSS, Gemini Bio- Products, West Sacramento, CA). After transfection or conditioning Quantitative Reverse-Transcription Polymerase Chain Re- alone in CSS, the cells were treated with distilled water, 0.25 mg/ml action. Total RNA was isolated from the treated C4-2bRxdox cells Dox, 1 nM dihydrotestosterone (DHT), Dox with DHT, or library using the GeneJet RNA purification kit (Thermo Fisher Scientific, compound with Dox and DHT. Waltham, MA). The isolated RNA was then reverse-transcribed with Transfection. One day before transfection, the cells were split into MMLV-reverse transcriptase (Invitrogen). Relative target-gene ex- a 10-cm plate for pooled transfection. On the next day, a dual- pression was then assessed by quantitative reverse-transcription transfection was prepared with a mouse mammary tumor virus yellow polymerase chain reaction (RT-PCR) with a SYBR green detection at ASPET Journals on September 25, 2021 fluorescent protein (MMTV-YFP) reporter construct (8 mg) and a dye (Invitrogen) and Rox reference dye (Invitrogen) on the StepOne pIRES-mCherry control plasmid (3 mg) using Lipofectamine LTX with Real Time PCR System (Applied Biosystems, Foster City, CA). Using Plus reagent (Invitrogen/Life Technologies, Carlsbad, CA). Transfection the DDCt relative quantification method, target gene readouts were complexes were prepared