Cisplatin modulates B-cell translocation gene 2 to attenuate cell proliferation of prostate carcinoma cells in both p53-dependent and p53-independent pathways

Cisplatin is a widely used anti-cancer drug. The B-cell translocation gene 2 (BTG2) is involved in the cell cycle transition regulation. We evaluated the cisplatin effects on prostate cancer cell proliferation and the expressions of BTG2, p53, androgen receptor (AR) and prostate specific antigen (PSA) in prostate carcinoma, p53 wild-type LNCaP or p53-null PC-3, cells. Cisplatin treatments attenuated cell prostate cancer cell growth through inducing Go/G1 cell cycle arrest in lower concentration and apoptosis at higher dosage. Cisplatin treatments enhanced p53 and BTG2 expression, repressed AR and PSA expression, and blocked the activation of androgen on the PSA secretion in LNCaP cells. BTG2 knockdown in LNCaP cells attenuated cisplatin-mediated growth inhibition. Cisplatin enhanced BTG2 gene expression dependent on the DNA fragment located within -173 to -82 upstream of BTG2 translation initiation site in prostate cancer cells. Mutation of the p53 response element from GGGCAGAGCCC to GGGCACC or mutation of the NFκB response element from GGAAAGTCC to GGAAAGGAA by site-directed mutagenesis abolished the stimulation of cisplatin on the BTG2 promoter activity in LNCaP or PC-3 cells, respectively. Our results indicated that cisplatin attenuates prostate cancer cell proliferation partly mediated by upregulation of BTG2 through the p53-dependent pathway or p53-independent NFκB pathway.

Cisplatin is a widely used anti-cancer drug. The B-cell translocation gene 2 (BTG2) is involved in the cell cycle transition regulation. We evaluated the cisplatin effects on prostate cancer cell proliferation and the expressions of BTG2, p53, androgen receptor (AR) and prostate specific antigen (PSA) in prostate carcinoma, p53 wild-type LNCaP or p53-null PC-3, cells. Cisplatin treatments attenuated cell prostate cancer cell growth through inducing Go/G1 cell cycle arrest in lower concentration and apoptosis at higher dosage. Cisplatin treatments enhanced p53 and BTG2 expression, repressed AR and PSA expression, and blocked the activation of androgen on the PSA secretion in LNCaP cells. BTG2 knockdown in LNCaP cells attenuated cisplatin-mediated growth inhibition. Cisplatin enhanced BTG2 gene expression dependent on the DNA fragment located within -173 to -82 upstream of BTG2 translation initiation site in prostate cancer cells. Mutation of the p53 response element from GGGCAGAGCCC to GGGCACC or mutation of the NFkB response element from GGAAAGTCC to GGAAAGGAA by site-directed mutagenesis abolished the stimulation of cisplatin on the BTG2 promoter activity in LNCaP or PC-3 cells, respectively. Our results indicated that cisplatin attenuates prostate cancer cell proliferation partly mediated by upregulation of BTG2 through the p53-dependent pathway or p53-independent NFkB pathway. P rostate cancer, ranking as the second most common solid tumor for men in United States, has caused 28,170 patients dying of this disease in 2012 1 . Although with the improvement in measurement technique of detection biomarker prostate-specific antigen (PSA) for prostate cancer, leading to the early diagnosis of prostate cancer more likely, the high risk prostate cancer patients still have high recurrence rate and distant metastasis 2,3 . Even under multimodal approaches, 10-25% prostate cancer patients still die of metastatic disease 3 .
Cisplatin, a neutral inorganic and square planar complex, functions through binding with DNA to form adduct to induce unique specific cellular responses, mainly culminating in apoptosis induction 4 . Since the application of cisplatin in clinical trial to treat cancer, cisplatin has brought a substantial impact on cancer treatment and changed the therapeutic regimens for a number of cancers including prostate 5,6 . Although the clinical benefits brought by cisplatin usage is obvious, the exact mechanism of how cisplatin exerts its antitumor effect is still not very clear, although the main mechanism is recognized as activation of p53 7 .
The B-cell translocation gene 2 (BTG2), belonging to antiproliferative APRO family proteins 8 featuring highly conservative domains, the BTG-Box A (Y 50 -N 71 ) and BTG-Box B (L 97 -E 115 ), is located mainly in the cytoplasm and functions in a variety of important cellular responses 9 . In terms of cancer cells, BTG2 acts as a tumor suppressor gene in a number of cancers and is activated mainly by p53 dependent pathway subsequent to DNA damage 10,11 . The p53 independent BTG2 expression is also possible through the PKC-d pathway in p53-null cancer cells 12 . Regarding prostate cancer, our previous studies have indicated that ectopic overexpression of BTG2 in PC-3 cells, a p53-null prostate cancer cell line, was able to inhibit cancer cell proliferation 13 .
Our previous study has shown that topoisomerase inhibitors could repress prostate cancer cell growth and induce BTG2 expressions in a p53 dependent manner 14 . Our objectives for this study are to determine the effects of cisplatin on prostate cancer cell growth, the AR and PSA expression, as well as the regulatory mechanisms of cisplatin on the gene expression of BTG2 in prostate cancer cells.

Results
After different concentrations of cisplatin treatment (0-80 mM) for 24 or 48 hours, cell proliferation of LNCaP cells were measured by 3 H-thymidine incorporation assay (Fig 1a). Our results indicated LNCaP cell proliferation was inhibited by 24 hours of cisplatin treatment in a dose-dependent manner, with 41% and 50% decreases noted when treated with 40 and 80 mM cisplatin, respectively. 48 hours cisplatin treatment clearly showed more prominent cell proliferation inhibition in LNCaP cells under the concentrations from 5 to 80 mM. Results from flow cytometric analysis of LNCaP cells revealed that 40 mM of cisplatin treatment induced 15% increase in G1 phase cell together with a decrease in S phase cells in LNCaP cells after 24 hours incubation, indicating 40 mM cisplatin treatment induced G1/S arrest in LNCaP cells. Further, since 80 mM of cisplatin increased the sub-G1 fraction of cells by 7-10%, it clearly indicated high dose of cisplatin was able to induced LNCaP cell apoptosis ( Figure 1b). This is also supported by the immunoblot assay revealing that treatment with 80 mM of cisplatin induced the expression of cleaved form of PARP in LNCaP cells (Figure 1c).
The results from immunoblot assay demonstrated that upregulation of BTG2 and p53 expression was noted when LNCaP cells were treated with different concentrations of cisplatin (Figure 2a). Quantitative analysis of the immunoblot results revealed a 4.5-fold and 2.4-fold increases of p53 and BTG2 protein levels were observed after 40 or 20 mM cisplatin treatments (Figure 2b). Of note, instead of increasing steadily, 80 mM cisplatin decreased the BTG2 expression as compared to 40 mM cisplatin in LNCaP cells (Figure 2b). Results of RT-qPCR indicated that cisplatin treatments induced BTG2 mRNA levels in dosage-dependent manner (Figure 2c). The reason behind this is since this concentration of cisplatin induced LNCaP cells apoptosis, which, in turn, degraded BTG2 protein via ubiquitinproteasome system, which has been proved in our previous study 14 . Treatment of MG132, a proteasome inhibitor, increased BTG2 pro-    Results of immunoblot assays revealed that cisplatin blocked not only PSA but also AR expressions in a dose-dependent manner (Figure 4a). Quantitative analysis of immunoblot assays revealed that 80 mM of cisplatin decreased 63% and 54% of PSA and AR protein levels, respectively (Figure 4b). The ELISA assay indicated that cisplatin (40 mM) blocked the secretion of PSA in LNCaP cells. In addition, cisplatin also significantly blocked the stimulation of R1881 (1 nM) on PSA secretion (Figure 4c).
Further transient BTG2 gene expression analyzed by reporter assays with 59-deletion and site-mutation of p53 response element from GGGAAAGTCC to GGAGTCC within BTG2 promoter area showed that the effect of cisplatin on BTG2 gene expression not only depended on the p53 response element but also on the DNA fragment within 2173 to 282 upstream of BTG2 translation initiation site in LNCaP cells (Figure 5a). Similar results were observed in PC-3 cells with the difference that site-mutation of p53 response element within BTG2 promoter area failed to abolish cisplatin-mediated BTG2 expressions (Figure 5b).
Further transient gene expression assays with the NFkB reporter vector indicated that cisplatin treatments blocked the NFkB activity in a dose-dependent manner in PC-3 cells (Figure 6a).Transient overexpression IkBa induced BTG2 promoter activity while overexpression MAP3K14, one kind of NFkB-inducing kinase (NIK), downregulated BTG2 promoter activity (Figure 6b). A putative NFkB response element was found within 2102 to 293 bp upstream of the translation starting point of the BTG2 gene. Transient overexpression IkBa (Figure 6c) or cisplatin (20 mM) (Figure 6d) treatment did not significantly enhance BTG2 promoter activity when the NFkB response element was mutated from GGAAAGTCC to GGAAAGGAA. Collectively, our result suggested that cisplatin modulated BTG2 expressions in PC-3 cells through the NFkB pathway.

Discussion
Cisplatin, one of the potent antitumor agents, has been shown to have significant clinical activity against a variety of solid tumors and, thus, been listed as the standard regimens for some cancers treatment. The mechanisms by which cisplatin exerts cytotoxic mode of action mainly lie in the cisplatin interaction with DNA to form unique DNA adducts, which are usually intra strand crosslink adducts 15 . These DNA adducts would further modulate the expressions of several signal transduction pathways concerning the cell survival, such as p53, p73, ATR and MAPK. Among others, the activation of p53 is deemed as the major effect of cisplatin treatment to repress cancer cell growth 4 . However, cisplatin-based treatments often induces sever toxic side-effects which causes drug discontinuation and limit therapeutic efficacy 16 . Therefore, studies in recent reports tried to combine the targeting delivery or gene therapy to enhance therapeutic effect and safety of cisplatin on the prostate cancer in vivo 6,17,18 .
As shown in Figure 1a and 3a, cisplatin (0-80 mM) treatment for 24 or 48 hours repressed LNCaP and PC-3 cell proliferation dosedependently. The lower concentrations of cisplatin (0-40 mM) treatment mainly induced cell cycle arrest at G0/G1 phase in both LNCaP and PC-3 cells as indicated by increased G1 phase cell percentage (Figure 1b and 3b). At higher dose (80 mM), cisplatin exerted apoptosis induction as demonstrated by increased sub-G1 phase cells and expression of c-PARP in prostate cancer cells (Figure 1b, 1c and 3b). Our results clearly indicated that cisplatin is a potent antitumor drug against prostate cancer in vitro.
p53, a well-known tumor suppressor gene, has been shown to have interaction with varied key transcription factors to involve in several important biological functions, such as DNA repair, apoptosis, cell cycle, apoptosis, senescence and angiogenesis. Thus, p53 plays an important role against cancer onset and progression, including prostate cancer 15 and is the target to be induced for a number of chemotherapy drugs. Our previous studies have shown that topoisomerase inhibitors could induce p53 expression in LNCaP cells 19,20 . As shown in Figure 2a and 2b, cisplatin treatment (from 0-80 mM) induced p53 expression in LNCaP cell, a p53 wild type prostate cancer cell, as determined by western blot. Our results were in line with previous studies indicating that cisplatin treatments could induce p53 activity in prostate cancer 17,21 .
BTG2, originally known as one kind of the early growth response genes 22 , was first isolated from 3T3 fibroblasts cells 23 . So far, BTG2 has been demonstrated to involve in varied vital cellular functions, such as in cell cycle regulation, in which BTG2 has been found to be able to repress cell cycle progression at G1/S or G2/M phases and function as a pan cell cycle modulator in a cell-or tissue-specific manner 9 . Thus, BTG2 has been deemed as a tumor suppressor gene in a variety of cancers, including gastric, breast, bladder, and prostate cancer 13,24-28 . Our results indicated that pre-apoptosis dosage of cisplatin (#40 mM) treatments upregulated BTG2 gene expression determining by immublotting, RT-qPCR, and transient gene expression assays in both LNCaP and PC-3 cells (Figure 2a, 2b, 2c, 2e, 3c, 3d, 3e). However, cisplatin treatment at apoptosis-inducing dosage (80 mM) decreased the BTG2 protein levels but the protein level was restored by cotreated with the proteasome inhbitior, MG132 (Figure 2d). These results are consistent with other reports that BTG2 protein is sensitive to the ubiquitin-proteasom system induced by genotoxic agents 14,29 .
Given that our previous study has shown that ectopic overexpression of BTG2 in PC-3 cells could inhibit cancer cell proliferation 13 , to further verify BTG2 role in cisplatin-induced inhibition on prostate cancer proliferation, we knocked down BTG2 expression by shRNA in LNCaP cells. The less proliferation inhibition noted in LNBTG2si cells by cisplatin clearly indicates that cisplatin-mediated LNCaP cell growth inhibition is partly dependent on BTG2 pathway (Figure 2f).
PSA, an androgen-upregulated protein, is well-used as a sensitive biomarker to detect the recurrence of prostate cancer 30,31 . PSA also has been shown to enhance prostate cancer cell invasion and facilitate refractory prostate cancer progression 32,33 . AR plays a vital role in androgen-dependent prostate cancer growth. Since p53 has been shown to be involved in negative regulation of AR activity 14,34 , we, therefore, also evaluated the AR expression and PSA expression in AR1 LNCaP cells after cisplatin treatment. As shown in Figure 4a (Figure 4c). Taken together, our results are in agreement with previous study which also showed that cisplatin upregulated p53 expression and blocked AR and PSA expression in LNCaP cells 21 .
Since the BTG2 has been regarded as the corepressor for AR in prostate carcinoma cells 35 , cisplatin may decrease the PSA expression p53-indirectly by upregulation of BTG2. Although we demonstrated cisplatin repressed AR and PSA expression in p53 wild-type LNCaP cells, however, we still can not rule out the possibility that the unidentified p53-independent pathway may also involve in the blocking effect of cisplatin on AR activity 21 .
Regarding BTG2 induction, two pathways have been demonstrated to involve in, which are p53 dependent 10,11 , or p53 independent PKC-d pathway 12 . In this study, we demonstrated for the first time that cisplatin treatment induced BTG2 expression in both p53wild LNCaP cells and p53-null PC-3 cells. Since previous reports indicated that a putative p53 response element (RRRCWWGY-YYN (0-13) RRRCWWGYYY) is located within human BTG2 promoter area and site-mutation of p53 response element abolished p53-dependent BTG2 induction in prostate cancer cells 13,14,36 , to further investigate the mechanisms whereby cisplatin modulates BTG2 expression in p53-wild LNCaP cells and p53-null PC-3 cells, we therefore applied 59-deltion method and site-directed mutation of p53 response element, combined with BTG2 reporter assay, to further study. As shown in Fig. 5a and 5b, cisplatin enhanced BTG2 gene expression dependent on the DNA fragment located within 2173 to 282 upstream of BTG2 translation initiation site in both LNCaP and PC-3 cells. Mutation of the p53 response element from GGGCAGAGCCC to GGGCACC by site-directed mutagenesis abolished the stimulation of cisplatin on the BTG2 promoter activity in LNCaP cells with no influence in PC-3 cells. Based on our result, we thus concluded that cisplatin upregulates BTG2 expression in prostate cancer cells through both p53-dependent and -independent pathways.
The IKK/NF-kB signaling plays an important role as its aberrant expression is usually linked to human cancers [37][38][39] . Amid the latent state, NF-kBs are sequestered in the cytosol due to binding with IkB (the NFkB inhibitor). As stimulated, IkB is phosphorylated by activated IKK (IkB kinase) and degraded by proteasome, leading to the release of NF-kB and subsequent nuclear translocation to modulate gene expression 40 .
Since previous studies have suggested that a putative NFkB response element (59-GGGRNWYYCC-39) is located at 59-flanking region of BTG2 gene 13,36 , we further evaluated whether cisplatin modulates BTG2 through the NFkB pathway in PC-3 cells. Our www.nature.com/scientificreports results of transient gene expression assays showed that cisplatin blocked NFkB activity in PC-3 cells (Figure 6a), in agreement with the previous report showing that cisplatin treatment perturbed the interaction between NFkB protein and NFkB response element 41 . As we blocked or induced NFkB pathway by transient overexpression of IkB or MAP3K14, BTG2 expressions in PC-3 cells were upregulated or downregulated as determined by BTG2 reporter assay, indicating NFkB is a deactivator of BTG2 expression in PC-3 cells (Figure 6b). These results are in agreement with previous study showing curcumin upregulated BTG2 expression by blocking the NFkB activity in U937 cells, a human leukemic monocyte lymphoma cell line 42 although other early study indicated that NFkB is the activator of BTG2 expression in T47D cells, a human ductal breast epithelial tumor cell line 43 . Further transient gene expression assays showed cisplatin treatment lost the activation on BTG2 promoter activity when the NFkB response element was mutated from GGAAA-GTCC to GGAAAGGAA in the BTG2 reporter vector, confirming that cisplatin upregulated BTG2 expression through the NFkB pathway in PC-3 cells.
Based on our results, we concluded that cisplatin is a potent drug to repress prostate cancer growth in vitro through induction of cell cycle arrest at G0/G1 and apoptosis. Cisplatin blocks AR and PSA expressions while increased p53 and BTG2 expressions. Cisplatinmediated growth inhibition in p53 wild-type LNCaP and p53 null  PC3 cells is partly dependent on induction of BTG2 expression, which is modulated in either p53-dependent (LNCaP cells) or NF-kB (PC-3 cells), p53-independent, pathway.

Methods
Materials, cell lines, and cell culture. LNCaP and PC-3 cell lines were obtained and maintained as described previously 30  Cell proliferation assay. Cell proliferation in response to cisplatin was measured using a 3 H-thymidine incorporation assay as described previously 21 or measured using a CyQUANT cell proliferation assay kit as described by manufacturer (Invitrogen, Carlsbad, CA, USA).
Flow cytometry. Cells were serum starved for 24 hours and then cultured in RPMI 1640 medium with 10% FCS and with or without different concentrations of drugs for another 24 hours. Cell cycle analysis was performed using the FACS-Calibur cytometer and CellQuestPro software (BD Biosciences, San Jose, CA, USA); the data were analyzed using ModFit LT Mac 3.0 software as described previously 20 .
Immunoblot assay. Equal quantities of cell extract were loaded onto a 12% sodium dodecyl sulfate polyacrylamide (SDS) gel and analyzed by the electrochemiluminescent detection system. The blotting membranes were probed with 15500 polyclonal BTG2 antiserum 12   PSA enzyme-linked immunosorbent assay. The PSA levels in the conditional media were measured by PSA enzyme linked immunosorbent assay (ELISA) and was adjusted by the concentration of protein in the whole cell extract as described previously 31 .
BTG2 knockdown. LNCaP cells were transduced with BTG2 small hairpin RNA lentiviral particles (Sc-43645-V; Santa Cruz Biotechnology) as described by the manufacturer. Two days after transduction, the cells (LN-BTG2si) were selected with 5 mg/ml puromycin dihydrochloride. The mock-transfected LNCaP cells (LN-COLsi) were transduced with control small hairpin RNA lentiviral particles (Sc-10808-V, Santa Cruz Biotechnology) and were clonally selected in the same manner as the BTG2-knockdown cells.
Reverse transcriptase-real-time polymerase chain reaction. Total RNA was isolated using Trizol reagent, and cDNA was synthesized and real-time polymerase chain reactions (qPCR) were performed using an ABI StepOne Plus Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) as described previously 20 . FAM dye-labeled TaqMan MGB probes and PCR primers for BTG2 (Hs00198887_ml) were purchased from Applied Biosystems. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Hs99999905_m1) was used as the internal positive control.
Expression vector constructs. The pCMV-IkBam expression vector which contains serine to alanine mutations at residues 32 and 36 of IkBa was purchased from Clontech (Moutain View, GA, USA). The MAP3K14 expression vector was constructed by cloning the MAP3K14 cDNA (MGC:45335; Invitrogen) after digestion with Eco RI and Xho I into the pcDNA3 expression vector (Invitrogen).
Report vector constructs and reporter assay. The NFkB reporter vector was purchased from Clontech. The reporter vectors containing 59-flanking region -1 to 2297, 21 to 2244, 21 to 2173, and 21 to 282 of the human BTG2 gene was cloned as modified from previously study 12 . The reporter vectors were constructed by mutation of p53 response element from GGGCAGAGCCC to GGGCACC and by mutation of NFkB response element from GGAAAGTCC to GGAAAGGAA of BTG2 gene. Cells were seeded onto 24-well plates at 1 3 10 4 cells/well 1 day prior to transfection. The cells were transiently transfected with 1 mg/well of reporter vectors as indicated and 0.5 mg/well of b-galactosidase expression vector (pCMVSPORTbgal; Life Technologies) as described previously 30 . The activities of luciferase and bgalactosidase were assayed as specified by the manufacturer instructions (Promega Bioscience, San Luis Obispo, CA, USA).
Statistical analysis. Results are expressed as the mean 6 S.E. of at least three independent replication of each experiment. Statistical significance was determined by one-way ANOVA and Student's t test using the SigmaStat program for Window version 2.03 (SPSS Inc, Chicago, IL, USA).