Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
Contents lists available at ScienceDirect
Journal of Steroid Biochemistry and Molecular Biology journal homepage: www.elsevier.com/locate/jsbmb
DNA damage response (DDR) via NKX3.1 expression in prostate cells Burcu Erbaykent-Tepedelen a , Selda Karamil a , Ceren Gonen-Korkmaz b , Kemal S. Korkmaz a,∗ a b
Ege University, Faculty of Engineering, Department of Bioengineering, Cancer Biology Laboratory, Bornova, Izmir, Turkey Ege University, Faculty of Pharmacy, Department of Pharmacology, Bornova, Izmir, Turkey
a r t i c l e
i n f o
Article history: Received 23 August 2013 Received in revised form 3 January 2014 Accepted 6 January 2014 Available online 13 January 2014 Keywords: H2AX ATM phosphorylation Irinotecan (CPT-11) Doxorubicin Etoposide
a b s t r a c t It has been reported that NKX3.1 an androgen-regulated homeobox gene restricted to prostate and testicular tissues, encodes a homeobox protein, which transcriptionally regulates oxidative damage responses and enhances topoisomerase I re-ligation by a direct interaction with the ATM protein in prostate cells. In this study, we aimed to investigate the role of NKX3.1 in DNA doublestrand break (DSB) repair. We demonstrate that the DNA damage induced by CPT-11 (irinotecan, a topo I inhibitor), doxorubicin (a topo II inhibitor), and H2 O2 (a mediator of oxidative damage), but not by etoposide (another topo II inhibitor), is negatively influenced by NKX3.1 expression. We also examined ␥H2AX(S139) foci formation and observed that the overexpression of NKX3.1 resulted a remarkable decrease in the formation of ␥H2AX(S139) foci. Intriguingly, we observed in NKX3.1 silencing studies that the depletion of NKX3.1 correlated with a significant decrease in the levels of p-ATM(S1981) and ␥H2AX(S139) . The data imply that the DNA damage response (DDR) can be altered, perhaps via a decrease in the topoisomerase I re-ligation function; this is consistent with the physical association of NKX3.1 with DDR mediators upon treatment of both PC-3 and LNCaP cells with CPT-11. Furthermore, the depletion of NKX3.1 resulted in a G1/S progression via the facilitation of an increase in E2F stabilization concurrent with the suppressed DDR. Thus, the topoisomerase I inhibitor-mediated DNA damage enhanced the physical association of NKX3.1 with ␥H2AX(S139) on the chromatin in LNCaP cells, whereas NKX3.1 in the soluble fraction was associated with p-ATM(S1981) and RAD50 in these cells. Overall, the data suggest that androgens and NKX3.1 expression regulate the progression of the cell cycle and concurrently activate the DDR. Therefore, androgen withdrawal may augment the development of an error-prone phenotype and, subsequently, the loss of DNA damage control during prostate cancer progression. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction The NKX3.1 gene is transcriptionally upregulated by androgens in the prostate gland and in the androgen receptor (AR)-positive prostate tumor cell line, LNCaP [1,2]. It has been demonstrated that the loss of an Nkx3.1 allele results in epithelial dysplasia and benign hyperplasia in the rat prostate, and NKX3.1 expression was found to be inversely correlated with high-grade prostate tumors in humans [3,4]. The NKX3.1 protein contains conserved domains in its wild-type isoform (234 amino acids long) called
∗ Corresponding author. Tel.: +90 5383087613; fax: +90 2323884955. E-mail addresses: burcu [email protected] (B. Erbaykent-Tepedelen), [email protected] (S. Karamil), korkmaz [email protected] (C. Gonen-Korkmaz), ks [email protected] (K.S. Korkmaz). 0960-0760/$ – see front matter © 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jsbmb.2014.01.001
tinman [5] and homeobox domains, which mediate interactions with several nuclear proteins and DNA, respectively. Consistent with the tumor suppressor function of NKX3.1, this factor supposedly represses transcription together with the Groucho complex subsequent to the recruitment of HDAC1 to target promoters, while NKX3.1 contributes to the cell cycle and cell death machinery via an MDM2-dependent mechanism [6]. Furthermore, the proteosomal degradation of NKX3.1 by TOPORS, which directly interacts with and acts as a robust E3 ubiquitin ligase, has also been shown in prostate cancer cells. However, the knockdown of TOPORS using an siRNA approach increases the steady-state level of NKX3.1 and the half-life in LNCaP [7]. These data suggest that the loss of NKX3.1 expression is an important molecular alteration in the development of dysplastic factors in prostate cells, which presumably occurs subsequent to the loss of NKX3.1 function: to protect the cells from oxidative DNA damage [6,7]. Hence, the protective
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
role of NKX3.1 regarding the oxidative DNA damage in the cell cycle has not been fully investigated in prostate cancer development, and mechanistic studies are required. As cells are exposed to DNA-damaging agents, including ionizing radiation (IR) that produces ROS [8,9], chemicals and UV light, the incorporation of replication-blocking lesions (alkyl adducts, pyrimidine dimers and crosslinks) [10] creates a serious threat to cell viability and genomic stability. Because the cancer chemotherapeutic agents that block the activity of topoisomerases also produce DNA double-strand breaks (DSBs), these agents may block replication in the S phase or cause the formation of enzyme–DNA complexes upon DNA damage during any phase of the cell cycle, depending on the type of the insult [10,11]. Unrepaired or misrepaired DSBs may result in cell death or in the large-scale chromosomal alterations that promote genomic instability, which is a major hallmark of cancer cells [12]. It has been previously reported that three proteins with distinct structures and functions (MRE11, RAD50, and NBS1) are required for the recognition of DSBs [13–16]. NBS1 has several SQ motifs in its primary sequence that are putative targets of ATM and ATR kinases [9,17]. Additionally, NBS1 interacts with ␥H2AX(S139) (a well-accepted hallmark of DNA damage) [18], and mutations of NBS1 that were identified in sporadic and familial cases of prostate cancer are associated with an increased risk of progression [19,20]. Notably, ␥H2AX(S139) foci formation decreases after drug treatment, which was demonstrated in previous studies [21,22], suggesting that there could be a damage sensor in prostate cells to facilitate the DNA damage response (DDR) and increase the recognition complex (MRN) activity in the presence of androgens [23]. Further, in their recent study, Bowen et al. demonstrated that this might be mediated by the physical association of NKX3.1 with ATM in prostate cells [21]. They also mapped the interaction domains of these two proteins in in vitro pull-down assays. As we have reported previously, the expression of RAD50 was strongly downregulated in NKX3.1-expressing cells treated with topoisomerase inhibitors [23]. The interaction of RAD50 with NKX3.1 might also affect the functional MRN complex and/or the association of Mre11 and NBS1 with DSBs; further studies are required to determine whether these events are independent of ␥H2AX(S139) foci formation. NKX3.1 physically associates with topoisomerase I [4,24] and ATM [22,25] during replication and/or transcription and repair, respectively, and it was found to be negatively down-regulated either by the inflammatory response or by TOPORS specifically in prostate cancer [7]. The data consistently suggest that NKX3.1 might be a required factor for cellular responses. In this study, we demonstrate that the abundance of NKX3.1 and the association of this factor with the DSB repair machinery are important for regulating prostate cell growth and the timely progression of the cell cycle. We also examine the physical association of NKX3.1 in repair complexes to determine the importance of these interactions during DNA damage in prostate cancer. With these investigations, we also demonstrate that NKX3.1 significantly slows down the cell cycle at an intra-S checkpoint and initiates the recognition of DNA damage through the ATMmediated activation of MRN upon exposure to a topo I inhibitor, CPT-11, as well as to H2 O2 , but not to a topo II inhibitor, etoposide. Presumably, these effects are mediated via different pathways, resulting in the sensitivity of prostate epithelial cells to topo II inhibitors, whereas the response to topo I inhibitor- and H2 O2 mediated DNA damage is activated by the androgen responsive factor NKX3.1. Furthermore, our sub-cellular fractionation studies reveal that NKX3.1 facilitates the replicative senescence block in the intra-S phase upon DNA damage and thereby regulates the spatial accessibility of required factors, such as p-ATM(S1981) and RAD50, to the damage sites. The data imply that NKX3.1 is a DNA damage sensor expressed in prostate cells [26], maintains the
27
fidelity of DNA damage regulation, its disruption results in serious malformations. 2. Methods 2.1. Cell propagation and androgen treatment PC-3, DU145 and LNCaP cells (American Type Culture Collection, Manassas, US) (passage numbers 10–20) were propagated in RPMI 1640 (Invitrogen, UK) medium supplemented with 5 or 10% heat inactivated FBS (Invitrogen, UK). Regarding the androgen administration, LNCaP cells were placed under serum starvation in 2% charcoal-treated serum (CT-FBS)-containing medium for 48 h and then 0.5% CT-FBS-containing RPMI 1640 for an additional 24 h. Finally, cells still in starvation were treated with R1881 (10 nM) for 4 and/or 24 h. 2.2. Treatment with DNA-damaging agents CPT-11 (Campto® ) and doxorubicin (Adriblastina® ) were kindly provided by Dr. M. Esassolak, Ege Univ., Med Faculty, Dept. of Oncology, Izmir, Turkey. Etoposide and H2 O2 were purchased from Sigma (UK). The treatments were performed for 0.5, 4 and 24 h where appropriate. 2.3. Antibodies The antibodies against NKX3.1, DNA-PK, pCHK2(T68) , p-p53(S20) , p53, ATM, p-ATM(S1981) , KU70, RAD50 and cyclin D1 were purchased from Santa Cruz, Inc. (Germany). The anti-p-H2AX(S139) monoclonal antibody and the anti-AR antibody were purchased from Upstate-Millipore (US). The anti--actin Ab was obtained from Sigma (UK). The anti-mouse and anti-rabbit Alexafluor 488-, 405- and 594-conjugated antibodies were purchased from Invitrogen (US). 2.4. Chromatin-associated protein separation For the protein extraction, the cells were lysed in 500 l EBC lysis buffer (50 mm Tris–Cl, pH 8.0, 120 mM NaCl, 0.5% Nonidet P-40 (v/v), 1 mm NaF, 0.2 mm Na3 VO4 ) freshly supplemented with 1 mM PMSF and a protease and phosphatase inhibitor mixture (EDTA-free cocktail) (Roche Applied Science). After centrifugation (16,000 × g, 15 min, 4 ◦ C), the supernatants were collected as soluble fractions. The pellets were further subjected to the extraction of chromatinassociated proteins. Briefly, the pellets were resuspended in 300 l EBC buffer supplemented with an additional 0.3 M NaCl. The suspension was cleared with a sonicator (Bandelin) for 8 rounds of 10 s “on” with intervals of 2 s “off” on ice, and the samples were then incubated on ice for another 10 min and centrifuged again (16,000 × g, 15 min, 4 ◦ C) to obtain the clear lysate as the chromatin associated-protein fraction. The protein concentrations in all the fractions were measured with the BCA assay. Then, the lysates were subjected to SDS-PAGE and western blotting. 2.5. Immunoblotting Cells collected in PBS were lysed with ice-cold lysis buffer (1% Nonidet P-40, 50 mM Tris–Cl, pH 7.4, 0.25% Na-deoxycholate, 150 mM NaCl, 1 mM EDTA) with NaF, Na3 VO4 (1 mM each) and complete protease and phosphatase inhibitors (Roche, Germany), unless otherwise indicated. The western blots were carried out by separating the proteins on 6–15% SDS-PAGE gels and immobilizing the proteins onto PVDF membranes (Amersham, UK and/or Santa Cruz, Germany) by a wet transfer. Briefly, the membranes were blocked with TBS-T buffer (Tris–Borate–Saline solution containing
28
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
0.1% Tween 20) containing either 5% skim milk (w/v) or 1% BSA for the phospho-antibodies. The primary and secondary antibody incubations were carried out in TBS-T containing 0.5% dry milk at RT for 1 h or at 4 ◦ C o/n. The membranes were developed with 2 ml ECL plus reagent (Amersham, UK) for 5 min and were photographed using Kodak X-ray film in a dark room. 2.6. Immunoprecipitation One milligram of protein lysate was used for each IP experiment. After 1 h of pre-clearing with 40 l protein A/G Sepharose beads (Santa Cruz), the supernatants (pre-cleared lysates) were divided into two groups, and the lysates were incubated o/n with either specific antibody or non-immune serum. The next day, protein A/G Sepharose beads were added to the pre-cleared lysates, and the samples were incubated for an additional 2 h to collect the antibody-specific complexes. Washes were carried out (twice each) with low and high salt buffers, and the precipitates were denatured in Laemmli (gel loading) buffer at 95 ◦ C for 5 min. The samples were loaded onto SDS-PAGE gels and then transferred onto PVDF membranes. The immunoblotting protocol was then applied. 2.7. Immunofluorescence labeling and microscopy For the detection of ␥-H2AX(S139) and NKX3.1, PC-3 cells were grown on coverslips, rinsed in PBS and fixed in 4% paraformaldehyde for 1 h at RT. After rinsing with PBS, the cells were permeabilized with 0.2% Triton X-100-containing PBS and were blocked with 1% BSA in PBS for 5 min before primary antibody incubation. The rinsed cells were labeled with the appropriate antibodies for 1 h and washed twice with PBS. The secondary antibody incubation was performed at RT for 20 min using Alexafluor 488/594 antibodies. The cells were washed four times with PBS, mounted with 0.5 g/ml DAPI containing 30% glycerol in PBS, and immediately analyzed with a Leica DMIL fluorescence microscope (Leica, Germany). 2.8. Flow cytometry and M30 analysis The cell cycle distribution of either LNCaP cells after R1881 administration or PC-3 cells following NKX3.1 transfection were performed with a FACSCanto (BD Biosciences, USA) and the FacsDiva 5.0.3. analysis software after treatment with CPT-11 and doxorubicin. Briefly, following transfections and/or treatments, the cells were collected into PBS and were fixed by the dropwise addition of cold (−20 ◦ C) 100% ethanol. The fixed cells were subsequently analyzed using M30 and PI staining, and the populations were plotted versus cell cycle. 2.9. siRNA mediated knockdown of NKX3.1 Cells were transfected with NKX3.1-specific or scrambled siRNA using the Dharmafect transfection reagent, which was purchased from Millipore (Germany). Before harvesting, siNKX3.1-specific- or scrambled siRNA-transfected cells at 24 h were either treated with CPT-11 or not and harvested after 2 h. The protein lysates were isolated and separated via SDS-PAGE for expression analysis. 2.10. Image analysis Images were analyzed with Image J software. Briefly, expression intensity measurements were performed for each cell using pre-defined boundaries of equal area and the following formulas: mean intensity(cell) = mean intensity(cell) − mean background; mean foci = total number of foci(cell) /number of cell counted. The data values are presented as error bars (±SD) and the differences
in mean values between groups were analyzed by two-tailed Student’s t test. p < 0.05 was considered statistically significant. 3. Results In this study, DNA damage was induced using topoisomerase inhibitors, CPT-11, doxorubicin, and etoposide, in addition to H2 O2 , and variations in the DNA damage response were studied in prostate tumor cell lines. The formation of ␥H2AX(S139) foci as an initial response to DNA damage was analyzed in PC-3 and LNCaP cells, and an increase or loss of phosphorylation was quantitated. Furthermore, the physical association of the NKX3.1 protein within repair complexes and the contribution of this factor to growth properties and apoptosis were also investigated in depletion and/or overexpression studies. 3.1. CPT-11-mediated H2AX foci formation is lost in NKX3.1-expressing cells To investigate the putative role of NKX3.1 in repair processes, PC-3 and DU145 cells were transfected with HM-NKX3.1 for the expression of wild-type NKX3.1 (234 aa), and H2AX(S139) phosphorylation and foci formation were then studied. While the control (untreated and untransfected) cells did not exhibit visible ␥H2AX(S139) foci without CPT-11 treatment, the drug treatments resulted in high levels of uniformly distributed foci (in nearly 70% of the cells) in the two cell lines studied. However, the cells with ectopic NKX3.1 expression exhibited a reduced number of ␥H2AX(S139) foci (Fig. 1A and Suppl. Fig. 1A). To examine variations in the expression of the mediators of DNA damage, PC-3 cells were transfected with either an HM-vector or HM-NKX3.1 and were examined during CPT-11-treatment time courses. We immediately observed that the phosphorylation of ATM(S1981) , NBS1(S343) , and H2AX(S139) and the expression of KU70 and DNA-PKcs were decreased by the ectopic expression of NKX3.1 compared to the controls (Fig. 1B). The data suggest that the expression of NKX3.1 contributes to the DNA damage response, most likely through the ATM pathway. Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jsbmb. 2014.01.001. Because the transfected cells represent a fraction of the overall cellular population, we were unable to distinguish the cells with different levels of NKX3.1 expression and ␥H2AX(S139) foci (Fig. 1A, subsection h) depending on the phase of the cell cycle. Therefore, we grouped the PC-3 cells as previously described [15] into four different levels of ␥H2AX(S139) expression with and without NKX3.1 and then counted the number of nuclei with foci. While foci were barely detectable without treatment, the percent of cells with clear foci increased to 58% with 24 h of CPT-11 treatment (Fig. 1C). The number of foci in cells with a low level of ␥H2AX(S139) expression (no ␥H2AX(S139) foci, p < 0.01) was quantitated, and the percent of cells with foci was significantly different in the PC-3 (p < 0.01) (Fig. 1D) and the DU145 cells (p < 0.01) (Suppl. Fig. 1B) compared to the controls. For PC-3 cells, the control- (n = 8) and vectortransfected (n = 66) cells exhibited remarkably high but close mean values, and the analysis of replicates did not reveal significant difference (p > 0.05), despite the CPT-11-induced ␥H2AX(S139) foci formation. 3.2. Doxorubicin- but not etoposide-mediated foci formation is altered in NKX3.1-expressing cells To investigate the putative role of NKX3.1 in topo II inhibitormediated DNA damage, cells were treated for 0.5 and 4 h with
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
29
Fig. 1. Topo I inhibitor-mediated foci formation is suppressed by NKX3.1 expression. (A) The untreated and untransfected control PC-3 cells exhibit no significant labeling of ␥H2AX(S139) and NKX3.1 except for the DAPI-stained nuclei (a–c), whereas the CPT-11-treated cells exhibit significant ␥H2AX(S139) foci formation (d–f). When cells are transfected with NKX3.1 and concurrently treated with drug (g–i), the level of ␥H2AX(S139) is significantly suppressed (the dashed lines represent the nuclei with NKX3.1 expression). The cells transfected with the HM-vector and treated with CPT-11 also exhibit intense foci formation (j–l) and are shown as controls. (B) Ectopic NKX3.1 expression in PC-3 cells influences the expression of DDR factors. When CPT-11 is administered over a time course, p-ATM(S1981) , p-NBS1(S343) , ␥H2AX(S139) and RAD50 levels are decreased remarkably in NKX3.1-expressing cells compared to controls. The native levels of DNA-PKcs and KU70 expression are also dissimilar but are not significantly relevant compared to the controls. -Actin was used for equality control. (C) The ␥H2AX(S139) intensity was given as percentages (n ≥ 200 for each group) in PC-3 cells. (D) Accordingly, PC-3 cells exhibit significantly different numbers of foci in the controls, the vector-treated and the NKX3.1-expressing cells (n ≥ 40 from each group, p < 0.01).
doxorubicin or etoposide. The treatment induced uniform foci formation in nearly all the nuclei, while the DNA content did not correlate with the intensity or number of ␥H2AX(S139) foci (Fig. 2A and B). However, when NKX3.1 was ectopically expressed before the treatments, it was observed that the ␥H2AX(S139) foci formation was partially reduced in some nuclei and the remaining foci were intense, although the size distribution was variable compared to the control cells (Fig. 2). Therefore, we divided the cells into groups according to the number of foci and analyzed the groups as previously described [23]. It was thus determined that doxorubicin- (only in the 20–50 foci range) but not etoposideinduced ␥H2AX(S139) foci formation was suppressed by NKX3.1 expression (Fig. 2C). The data suggest that NKX3.1 has no influence on the response to topo II inhibitor-mediated DNA damage. 3.3. Androgen decreases H2 O2 -mediated DNA damage in LNCaP cells Because doxorubicin generates reactive oxygen species (ROS) [27], to dissect the ability of doxorubicin to induce DNA damage through topo II inhibition and oxidative damage, H2 O2 -mediated foci formation and the effect of androgens on the oxidative DNA damage were examined in prostate cancer cells. First, LNCaP cells were treated with androgen (R1881), and then, time courses with H2 O2 were performed (Fig. 3A). Following an 8- to 10-fold upregulation of NKX3.1 expression, we investigated the levels of the following: (1) the expression of damage recognition factors; (2)
the androgen receptor status; and (3) the DNA damage response in these cells. When we examined the p-ATM(S1981) , pCHK2(T68) , p-p53(S20) and ␥H2AX(S139) levels, we observed that they were maintained at lower levels with androgen administration (Fig. 3A). Because the p-ATM(S1981) and ␥H2AX(S139) levels were higher in the control cells than the androgen-treated LNCaPs, the data led us to postulate that prolonged activation of the DDR might occur on androgen depletion due to a persistent failure to complete the repair process. Therefore, H2 O2 -mediated foci formation was also examined and quantitated with immunofluorescence labeling in the PC-3 cells (Fig. 3B and C). When the H2 O2 treatments were performed for 0.5 and 4 h, almost all the nuclei in the PC-3 cells had diffuse but non-uniform foci, and the ectopically expressed NKX3.1 partially reduced ␥H2AX(S139) foci formation in these cells. Also spontaneous- and H2 O2 -mediated ␥H2AX(S139) foci formation was low under the influence of androgens in LNCaP cells (Fig. 3D and E). The data suggest that either the damage or the response to the damage was low under the influence of NKX3.1 as well as androgens. Thus, significant differences were observed between the NKX3.1-expressing cells and the controls regarding the H2 O2 treatments, but no considerable differences were observed after the doxorubicin or etoposide treatments (Figs. 2C and 3). Then, we hypothesized that DSB formation might be a consequence of the collision of transcription and repair complexes and arise in the presence of topoisomerase inhibitors. Therefore, the observations directed us to investigate the physical association of NKX3.1 with the DDR mediators.
30
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
Fig. 2. Topo II inhibitor-mediated foci formation (at 0.5 or 4 h) is not influenced by NKX3.1 expression. (A) Doxorubicin and (B) etoposide-treated cells exhibited significant numbers of ␥H2AX(S139) foci (p < 0.05), while the ectopic expression of NKX3.1 did not influence the number of foci (Kruskal–Wallis, p > 0.05). (C) ␥H2AX(S139) foci-containing cells (due to varying intensity) were counted and separated into groups (four) to plot as percentages (2 independent experiments with n ≥ 100 for each group). The variations in the number of foci in each group are also shown as histogram plots in the control and NKX3.1-expressing groups.
3.4. NKX3.1 physically associates with cellular H2AX(S139) , RAD50, p-ATM(S1981) and DNA-PK, but not with ATM in PC-3 cells To test the possibility that there is a physical interaction of NKX3.1 with the DDR proteins, thereby regulating the initiation and level of foci formation upon DNA damage, the direct association of NKX3.1 with ␥H2AX(S139) , DNA-PKcs, ATM, p-ATM(S1981) and RAD50 was examined. The ectopic (in PC-3 and DU145 cells) and native, androgen-inducible (in LNCaP cells), expression of NKX3.1 was used in repair response studies (Fig. 4). Additionally, whether these interactions were altered by the administration of CPT-11 was also examined. First, NKX3.1 was ectopically expressed in PC3 (Fig. 4A–C) and DU145 cells (Fig. 4D) using the HM-NKX3.1 wt construct, and cellular ␥H2AX(S139) , RAD50, DNA-PKcs, ATM, and p-ATM(S1981) were then immunoprecipitated. Next, the NKX3.1 protein was immunoblotted with anti-NKX3.1 antibodies, and it was observed that NKX3.1 interacted with ␥H2AX(S139) and RAD50 somewhat, but this interaction was significantly increased by CPT11 treatment in PC-3 cells (Fig. 4A and C). When an IP was performed to detect p-ATM(S1981) in the absence and presence of CPT-11, we found that NKX3.1 also interacted with p-ATM(S1981) , consistent with the previous interactions, and this association increased with CPT-11 treatment. To reduce the possibility of false positives for subsequent ATM precipitations, p-ATM(S1981) was sequentially precipitated three times to remove most of it from the lysate. After this removal, we could not detect the interaction
of NKX3.1 with ATM (Fig. 4B). Furthermore, we also found that NKX3.1 physically associated with DNA-PKcs and that the interaction was reduced in CPT-11-treated DU145 cells. This is consistent with the level of DNA-PKcs expression obtained with western blot analysis (Fig. 4D). A reciprocal IP of DNA-PKcs was also performed (Fig. 4D). In addition, NKX3.1 was also precipitated from LNCaP cells before and after CPT-11-mediated DNA damage. It was observed that the increased expression level of NKX3.1 induced by androgen was accompanied by an increased physical association with p-ATM(S1981) (Fig. 4E). Furthermore, when the cells were transfected with the cDNA encoding the 234 aa full-length NKX3.1 (wt) or the mutant 63-Cter (99 aa) without the homeobox domain and treated with CPT-11 (Fig. 4F), we found that p-ATM(S1981) and ␥H2AX(S139) co-precipitated with NKX3.1 wt but not with the 63-Cter mutant, regardless of CPT-11 treatment (Fig. 4H). Taken together, NKX3.1 interacts with p-ATM(S1981) , and this physical association increases with CPT-11 treatment but does not change with androgen treatment; the wild-type NKX3.1 interacts with repair complexes, and this interaction requires the nuclear localization of NKX3.1 in prostate cells (Fig. 4F–H). Moreover, to gain additional insight into understanding the distribution of NKX3.1 and the co-localization of this protein with DNA damage mediators, we performed sub-cellular fractionation combined with western blot analysis in LNCaP cells. We observed that both native NKX3.1 and overexpressed NKX3.1 are in the soluble fraction along with ATM and RAD50, whereas NKX3.1 is
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
31
Fig. 3. Androgen protects prostate cells from DNA damage. (A) Androgen administration remarkably increases the AR, cyclin D1 and NKX3.1 expression levels in LNCaP cells. While the p-ATM(S1981) , p-CHK2(T68) , p-p53(S20) and ␥H2AX(S139) levels are decreased in androgen- and H2 O2 -treated cells, the expression levels of CHK2, p53, DNA-PKcs, Ku70, RAD50 and -actin are not significantly altered. Notably, the cellular ATM expression level remains approximately 2-fold depleted and is unchanged by androgen. (B) The H2 O2 -treated cells display a number of ␥H2AX(S139) foci, and (C) the ectopic expression of NKX3.1 in PC-3 cells suppresses the formation of these foci (2 independent experiments with n ≥ 100 for each group, p < 0.01). (D and E) Immunofluorescence data shows that androgen significantly (p < 0.001) suppresses the H2 O2 -induced ␥H2AX(S139) foci formation in LNCaP cells (n ≥ 20 for each group). The variations in number of foci in each group are shown as histogram plots in the control and treated groups.
only observed in the chromatin fraction when it is overexpressed, regardless of CPT-11 treatment. Consistently, the data suggest that a physical interaction of NKX3.1 with H2AX occurs at the chromatin level, and NKX3.1 interacts with ATM as well as RAD50 in soluble nuclear fraction (Fig. 4J). 3.5. NKX3.1 increases cyclin D1 expression We also observed that the ectopic expression of NKX3.1 resulted in relatively small sized nuclei in our studies, indicating that the NKX3.1 protein might trigger histone condensation as well as
decreased ␥H2AX(S139) foci formation (Fig. 5A). This observation led us to investigate whether the cells were under the control of a G1/S checkpoint and arrested at the G1 or S phase prior to DNA synthesis when NKX3.1 was expressed; therefore, cyclin D1 expression was examined. Cyclin D1 expression levels were measured in PC-3 cells treated with compounds (Fig. 5A) for 0.5 and 4 h, in addition to untreated and/or untransfected controls, using the image J software to measure the immunofluorescent intensity as an indication of cellular cyclin D1 expression level. We found that the average cyclin D1 expression level increased 2-fold over that of the controls with a high variance in the NKX3.1-expressing cells (Fig. 5B–E). Thus,
32
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
Fig. 4. NKX3.1 physically associates with DDR factors. (A) The ␥H2AX(S139) physically associates with NKX3.1. (B) p-ATM(S1981) , but not ATM, also co-precipitates with NKX3.1 in PC-3 cells. (C) Also the cellular RAD50 and (D) DNA-PK associate with ectopic NKX3.1 in PC-3 and in DU145 cells respectively. Reciprocal IPs were also performed to confirm the DNA-PK interaction with NKX3.1. (E) The cellular p-ATM(S1981) associates with native NKX3.1 in LNCaP cells regardless of androgen administration. (F) Schema demonstrates that the NKX3.1 wt construct encodes the 234 aa protein with the intact homeobox domain, whereas the 63-Cter construct 99 amino acid peptide without the homeobox domain. (G) The 63-Cter deletion mutant exhibited ␥H2AX(S139) foci formation upon CPT-11-mediated DNA damage, whereas the NKX3.1 wt did not. (H) The p-ATM and ␥H2AX(S139) precipitations demonstrate that the association with NKX3.1 wt (28 kD) but not 63-Cter (12 kD) occurs regardless of drug treatment. -Actin blots were also performed to show the similarity of sample inputs. (J) Additionally, NKX3.1 is localized in both the soluble and chromatin fractions, whereas the native ATM and RAD50 are soluble in the nucleus, while ␥H2AX(S139) and H2AX are in the chromatin fraction.
the percent of cells that expressed cyclin D1 with a mean value >40 (relative intensity measurement) was 26% in controls and increased to 34% with NKX3.1 expression (data not shown). Moreover, when the cells with a mean value of expression intensity >20 and
Contents lists available at ScienceDirect
Journal of Steroid Biochemistry and Molecular Biology journal homepage: www.elsevier.com/locate/jsbmb
DNA damage response (DDR) via NKX3.1 expression in prostate cells Burcu Erbaykent-Tepedelen a , Selda Karamil a , Ceren Gonen-Korkmaz b , Kemal S. Korkmaz a,∗ a b
Ege University, Faculty of Engineering, Department of Bioengineering, Cancer Biology Laboratory, Bornova, Izmir, Turkey Ege University, Faculty of Pharmacy, Department of Pharmacology, Bornova, Izmir, Turkey
a r t i c l e
i n f o
Article history: Received 23 August 2013 Received in revised form 3 January 2014 Accepted 6 January 2014 Available online 13 January 2014 Keywords: H2AX ATM phosphorylation Irinotecan (CPT-11) Doxorubicin Etoposide
a b s t r a c t It has been reported that NKX3.1 an androgen-regulated homeobox gene restricted to prostate and testicular tissues, encodes a homeobox protein, which transcriptionally regulates oxidative damage responses and enhances topoisomerase I re-ligation by a direct interaction with the ATM protein in prostate cells. In this study, we aimed to investigate the role of NKX3.1 in DNA doublestrand break (DSB) repair. We demonstrate that the DNA damage induced by CPT-11 (irinotecan, a topo I inhibitor), doxorubicin (a topo II inhibitor), and H2 O2 (a mediator of oxidative damage), but not by etoposide (another topo II inhibitor), is negatively influenced by NKX3.1 expression. We also examined ␥H2AX(S139) foci formation and observed that the overexpression of NKX3.1 resulted a remarkable decrease in the formation of ␥H2AX(S139) foci. Intriguingly, we observed in NKX3.1 silencing studies that the depletion of NKX3.1 correlated with a significant decrease in the levels of p-ATM(S1981) and ␥H2AX(S139) . The data imply that the DNA damage response (DDR) can be altered, perhaps via a decrease in the topoisomerase I re-ligation function; this is consistent with the physical association of NKX3.1 with DDR mediators upon treatment of both PC-3 and LNCaP cells with CPT-11. Furthermore, the depletion of NKX3.1 resulted in a G1/S progression via the facilitation of an increase in E2F stabilization concurrent with the suppressed DDR. Thus, the topoisomerase I inhibitor-mediated DNA damage enhanced the physical association of NKX3.1 with ␥H2AX(S139) on the chromatin in LNCaP cells, whereas NKX3.1 in the soluble fraction was associated with p-ATM(S1981) and RAD50 in these cells. Overall, the data suggest that androgens and NKX3.1 expression regulate the progression of the cell cycle and concurrently activate the DDR. Therefore, androgen withdrawal may augment the development of an error-prone phenotype and, subsequently, the loss of DNA damage control during prostate cancer progression. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction The NKX3.1 gene is transcriptionally upregulated by androgens in the prostate gland and in the androgen receptor (AR)-positive prostate tumor cell line, LNCaP [1,2]. It has been demonstrated that the loss of an Nkx3.1 allele results in epithelial dysplasia and benign hyperplasia in the rat prostate, and NKX3.1 expression was found to be inversely correlated with high-grade prostate tumors in humans [3,4]. The NKX3.1 protein contains conserved domains in its wild-type isoform (234 amino acids long) called
∗ Corresponding author. Tel.: +90 5383087613; fax: +90 2323884955. E-mail addresses: burcu [email protected] (B. Erbaykent-Tepedelen), [email protected] (S. Karamil), korkmaz [email protected] (C. Gonen-Korkmaz), ks [email protected] (K.S. Korkmaz). 0960-0760/$ – see front matter © 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jsbmb.2014.01.001
tinman [5] and homeobox domains, which mediate interactions with several nuclear proteins and DNA, respectively. Consistent with the tumor suppressor function of NKX3.1, this factor supposedly represses transcription together with the Groucho complex subsequent to the recruitment of HDAC1 to target promoters, while NKX3.1 contributes to the cell cycle and cell death machinery via an MDM2-dependent mechanism [6]. Furthermore, the proteosomal degradation of NKX3.1 by TOPORS, which directly interacts with and acts as a robust E3 ubiquitin ligase, has also been shown in prostate cancer cells. However, the knockdown of TOPORS using an siRNA approach increases the steady-state level of NKX3.1 and the half-life in LNCaP [7]. These data suggest that the loss of NKX3.1 expression is an important molecular alteration in the development of dysplastic factors in prostate cells, which presumably occurs subsequent to the loss of NKX3.1 function: to protect the cells from oxidative DNA damage [6,7]. Hence, the protective
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
role of NKX3.1 regarding the oxidative DNA damage in the cell cycle has not been fully investigated in prostate cancer development, and mechanistic studies are required. As cells are exposed to DNA-damaging agents, including ionizing radiation (IR) that produces ROS [8,9], chemicals and UV light, the incorporation of replication-blocking lesions (alkyl adducts, pyrimidine dimers and crosslinks) [10] creates a serious threat to cell viability and genomic stability. Because the cancer chemotherapeutic agents that block the activity of topoisomerases also produce DNA double-strand breaks (DSBs), these agents may block replication in the S phase or cause the formation of enzyme–DNA complexes upon DNA damage during any phase of the cell cycle, depending on the type of the insult [10,11]. Unrepaired or misrepaired DSBs may result in cell death or in the large-scale chromosomal alterations that promote genomic instability, which is a major hallmark of cancer cells [12]. It has been previously reported that three proteins with distinct structures and functions (MRE11, RAD50, and NBS1) are required for the recognition of DSBs [13–16]. NBS1 has several SQ motifs in its primary sequence that are putative targets of ATM and ATR kinases [9,17]. Additionally, NBS1 interacts with ␥H2AX(S139) (a well-accepted hallmark of DNA damage) [18], and mutations of NBS1 that were identified in sporadic and familial cases of prostate cancer are associated with an increased risk of progression [19,20]. Notably, ␥H2AX(S139) foci formation decreases after drug treatment, which was demonstrated in previous studies [21,22], suggesting that there could be a damage sensor in prostate cells to facilitate the DNA damage response (DDR) and increase the recognition complex (MRN) activity in the presence of androgens [23]. Further, in their recent study, Bowen et al. demonstrated that this might be mediated by the physical association of NKX3.1 with ATM in prostate cells [21]. They also mapped the interaction domains of these two proteins in in vitro pull-down assays. As we have reported previously, the expression of RAD50 was strongly downregulated in NKX3.1-expressing cells treated with topoisomerase inhibitors [23]. The interaction of RAD50 with NKX3.1 might also affect the functional MRN complex and/or the association of Mre11 and NBS1 with DSBs; further studies are required to determine whether these events are independent of ␥H2AX(S139) foci formation. NKX3.1 physically associates with topoisomerase I [4,24] and ATM [22,25] during replication and/or transcription and repair, respectively, and it was found to be negatively down-regulated either by the inflammatory response or by TOPORS specifically in prostate cancer [7]. The data consistently suggest that NKX3.1 might be a required factor for cellular responses. In this study, we demonstrate that the abundance of NKX3.1 and the association of this factor with the DSB repair machinery are important for regulating prostate cell growth and the timely progression of the cell cycle. We also examine the physical association of NKX3.1 in repair complexes to determine the importance of these interactions during DNA damage in prostate cancer. With these investigations, we also demonstrate that NKX3.1 significantly slows down the cell cycle at an intra-S checkpoint and initiates the recognition of DNA damage through the ATMmediated activation of MRN upon exposure to a topo I inhibitor, CPT-11, as well as to H2 O2 , but not to a topo II inhibitor, etoposide. Presumably, these effects are mediated via different pathways, resulting in the sensitivity of prostate epithelial cells to topo II inhibitors, whereas the response to topo I inhibitor- and H2 O2 mediated DNA damage is activated by the androgen responsive factor NKX3.1. Furthermore, our sub-cellular fractionation studies reveal that NKX3.1 facilitates the replicative senescence block in the intra-S phase upon DNA damage and thereby regulates the spatial accessibility of required factors, such as p-ATM(S1981) and RAD50, to the damage sites. The data imply that NKX3.1 is a DNA damage sensor expressed in prostate cells [26], maintains the
27
fidelity of DNA damage regulation, its disruption results in serious malformations. 2. Methods 2.1. Cell propagation and androgen treatment PC-3, DU145 and LNCaP cells (American Type Culture Collection, Manassas, US) (passage numbers 10–20) were propagated in RPMI 1640 (Invitrogen, UK) medium supplemented with 5 or 10% heat inactivated FBS (Invitrogen, UK). Regarding the androgen administration, LNCaP cells were placed under serum starvation in 2% charcoal-treated serum (CT-FBS)-containing medium for 48 h and then 0.5% CT-FBS-containing RPMI 1640 for an additional 24 h. Finally, cells still in starvation were treated with R1881 (10 nM) for 4 and/or 24 h. 2.2. Treatment with DNA-damaging agents CPT-11 (Campto® ) and doxorubicin (Adriblastina® ) were kindly provided by Dr. M. Esassolak, Ege Univ., Med Faculty, Dept. of Oncology, Izmir, Turkey. Etoposide and H2 O2 were purchased from Sigma (UK). The treatments were performed for 0.5, 4 and 24 h where appropriate. 2.3. Antibodies The antibodies against NKX3.1, DNA-PK, pCHK2(T68) , p-p53(S20) , p53, ATM, p-ATM(S1981) , KU70, RAD50 and cyclin D1 were purchased from Santa Cruz, Inc. (Germany). The anti-p-H2AX(S139) monoclonal antibody and the anti-AR antibody were purchased from Upstate-Millipore (US). The anti--actin Ab was obtained from Sigma (UK). The anti-mouse and anti-rabbit Alexafluor 488-, 405- and 594-conjugated antibodies were purchased from Invitrogen (US). 2.4. Chromatin-associated protein separation For the protein extraction, the cells were lysed in 500 l EBC lysis buffer (50 mm Tris–Cl, pH 8.0, 120 mM NaCl, 0.5% Nonidet P-40 (v/v), 1 mm NaF, 0.2 mm Na3 VO4 ) freshly supplemented with 1 mM PMSF and a protease and phosphatase inhibitor mixture (EDTA-free cocktail) (Roche Applied Science). After centrifugation (16,000 × g, 15 min, 4 ◦ C), the supernatants were collected as soluble fractions. The pellets were further subjected to the extraction of chromatinassociated proteins. Briefly, the pellets were resuspended in 300 l EBC buffer supplemented with an additional 0.3 M NaCl. The suspension was cleared with a sonicator (Bandelin) for 8 rounds of 10 s “on” with intervals of 2 s “off” on ice, and the samples were then incubated on ice for another 10 min and centrifuged again (16,000 × g, 15 min, 4 ◦ C) to obtain the clear lysate as the chromatin associated-protein fraction. The protein concentrations in all the fractions were measured with the BCA assay. Then, the lysates were subjected to SDS-PAGE and western blotting. 2.5. Immunoblotting Cells collected in PBS were lysed with ice-cold lysis buffer (1% Nonidet P-40, 50 mM Tris–Cl, pH 7.4, 0.25% Na-deoxycholate, 150 mM NaCl, 1 mM EDTA) with NaF, Na3 VO4 (1 mM each) and complete protease and phosphatase inhibitors (Roche, Germany), unless otherwise indicated. The western blots were carried out by separating the proteins on 6–15% SDS-PAGE gels and immobilizing the proteins onto PVDF membranes (Amersham, UK and/or Santa Cruz, Germany) by a wet transfer. Briefly, the membranes were blocked with TBS-T buffer (Tris–Borate–Saline solution containing
28
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
0.1% Tween 20) containing either 5% skim milk (w/v) or 1% BSA for the phospho-antibodies. The primary and secondary antibody incubations were carried out in TBS-T containing 0.5% dry milk at RT for 1 h or at 4 ◦ C o/n. The membranes were developed with 2 ml ECL plus reagent (Amersham, UK) for 5 min and were photographed using Kodak X-ray film in a dark room. 2.6. Immunoprecipitation One milligram of protein lysate was used for each IP experiment. After 1 h of pre-clearing with 40 l protein A/G Sepharose beads (Santa Cruz), the supernatants (pre-cleared lysates) were divided into two groups, and the lysates were incubated o/n with either specific antibody or non-immune serum. The next day, protein A/G Sepharose beads were added to the pre-cleared lysates, and the samples were incubated for an additional 2 h to collect the antibody-specific complexes. Washes were carried out (twice each) with low and high salt buffers, and the precipitates were denatured in Laemmli (gel loading) buffer at 95 ◦ C for 5 min. The samples were loaded onto SDS-PAGE gels and then transferred onto PVDF membranes. The immunoblotting protocol was then applied. 2.7. Immunofluorescence labeling and microscopy For the detection of ␥-H2AX(S139) and NKX3.1, PC-3 cells were grown on coverslips, rinsed in PBS and fixed in 4% paraformaldehyde for 1 h at RT. After rinsing with PBS, the cells were permeabilized with 0.2% Triton X-100-containing PBS and were blocked with 1% BSA in PBS for 5 min before primary antibody incubation. The rinsed cells were labeled with the appropriate antibodies for 1 h and washed twice with PBS. The secondary antibody incubation was performed at RT for 20 min using Alexafluor 488/594 antibodies. The cells were washed four times with PBS, mounted with 0.5 g/ml DAPI containing 30% glycerol in PBS, and immediately analyzed with a Leica DMIL fluorescence microscope (Leica, Germany). 2.8. Flow cytometry and M30 analysis The cell cycle distribution of either LNCaP cells after R1881 administration or PC-3 cells following NKX3.1 transfection were performed with a FACSCanto (BD Biosciences, USA) and the FacsDiva 5.0.3. analysis software after treatment with CPT-11 and doxorubicin. Briefly, following transfections and/or treatments, the cells were collected into PBS and were fixed by the dropwise addition of cold (−20 ◦ C) 100% ethanol. The fixed cells were subsequently analyzed using M30 and PI staining, and the populations were plotted versus cell cycle. 2.9. siRNA mediated knockdown of NKX3.1 Cells were transfected with NKX3.1-specific or scrambled siRNA using the Dharmafect transfection reagent, which was purchased from Millipore (Germany). Before harvesting, siNKX3.1-specific- or scrambled siRNA-transfected cells at 24 h were either treated with CPT-11 or not and harvested after 2 h. The protein lysates were isolated and separated via SDS-PAGE for expression analysis. 2.10. Image analysis Images were analyzed with Image J software. Briefly, expression intensity measurements were performed for each cell using pre-defined boundaries of equal area and the following formulas: mean intensity(cell) = mean intensity(cell) − mean background; mean foci = total number of foci(cell) /number of cell counted. The data values are presented as error bars (±SD) and the differences
in mean values between groups were analyzed by two-tailed Student’s t test. p < 0.05 was considered statistically significant. 3. Results In this study, DNA damage was induced using topoisomerase inhibitors, CPT-11, doxorubicin, and etoposide, in addition to H2 O2 , and variations in the DNA damage response were studied in prostate tumor cell lines. The formation of ␥H2AX(S139) foci as an initial response to DNA damage was analyzed in PC-3 and LNCaP cells, and an increase or loss of phosphorylation was quantitated. Furthermore, the physical association of the NKX3.1 protein within repair complexes and the contribution of this factor to growth properties and apoptosis were also investigated in depletion and/or overexpression studies. 3.1. CPT-11-mediated H2AX foci formation is lost in NKX3.1-expressing cells To investigate the putative role of NKX3.1 in repair processes, PC-3 and DU145 cells were transfected with HM-NKX3.1 for the expression of wild-type NKX3.1 (234 aa), and H2AX(S139) phosphorylation and foci formation were then studied. While the control (untreated and untransfected) cells did not exhibit visible ␥H2AX(S139) foci without CPT-11 treatment, the drug treatments resulted in high levels of uniformly distributed foci (in nearly 70% of the cells) in the two cell lines studied. However, the cells with ectopic NKX3.1 expression exhibited a reduced number of ␥H2AX(S139) foci (Fig. 1A and Suppl. Fig. 1A). To examine variations in the expression of the mediators of DNA damage, PC-3 cells were transfected with either an HM-vector or HM-NKX3.1 and were examined during CPT-11-treatment time courses. We immediately observed that the phosphorylation of ATM(S1981) , NBS1(S343) , and H2AX(S139) and the expression of KU70 and DNA-PKcs were decreased by the ectopic expression of NKX3.1 compared to the controls (Fig. 1B). The data suggest that the expression of NKX3.1 contributes to the DNA damage response, most likely through the ATM pathway. Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jsbmb. 2014.01.001. Because the transfected cells represent a fraction of the overall cellular population, we were unable to distinguish the cells with different levels of NKX3.1 expression and ␥H2AX(S139) foci (Fig. 1A, subsection h) depending on the phase of the cell cycle. Therefore, we grouped the PC-3 cells as previously described [15] into four different levels of ␥H2AX(S139) expression with and without NKX3.1 and then counted the number of nuclei with foci. While foci were barely detectable without treatment, the percent of cells with clear foci increased to 58% with 24 h of CPT-11 treatment (Fig. 1C). The number of foci in cells with a low level of ␥H2AX(S139) expression (no ␥H2AX(S139) foci, p < 0.01) was quantitated, and the percent of cells with foci was significantly different in the PC-3 (p < 0.01) (Fig. 1D) and the DU145 cells (p < 0.01) (Suppl. Fig. 1B) compared to the controls. For PC-3 cells, the control- (n = 8) and vectortransfected (n = 66) cells exhibited remarkably high but close mean values, and the analysis of replicates did not reveal significant difference (p > 0.05), despite the CPT-11-induced ␥H2AX(S139) foci formation. 3.2. Doxorubicin- but not etoposide-mediated foci formation is altered in NKX3.1-expressing cells To investigate the putative role of NKX3.1 in topo II inhibitormediated DNA damage, cells were treated for 0.5 and 4 h with
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
29
Fig. 1. Topo I inhibitor-mediated foci formation is suppressed by NKX3.1 expression. (A) The untreated and untransfected control PC-3 cells exhibit no significant labeling of ␥H2AX(S139) and NKX3.1 except for the DAPI-stained nuclei (a–c), whereas the CPT-11-treated cells exhibit significant ␥H2AX(S139) foci formation (d–f). When cells are transfected with NKX3.1 and concurrently treated with drug (g–i), the level of ␥H2AX(S139) is significantly suppressed (the dashed lines represent the nuclei with NKX3.1 expression). The cells transfected with the HM-vector and treated with CPT-11 also exhibit intense foci formation (j–l) and are shown as controls. (B) Ectopic NKX3.1 expression in PC-3 cells influences the expression of DDR factors. When CPT-11 is administered over a time course, p-ATM(S1981) , p-NBS1(S343) , ␥H2AX(S139) and RAD50 levels are decreased remarkably in NKX3.1-expressing cells compared to controls. The native levels of DNA-PKcs and KU70 expression are also dissimilar but are not significantly relevant compared to the controls. -Actin was used for equality control. (C) The ␥H2AX(S139) intensity was given as percentages (n ≥ 200 for each group) in PC-3 cells. (D) Accordingly, PC-3 cells exhibit significantly different numbers of foci in the controls, the vector-treated and the NKX3.1-expressing cells (n ≥ 40 from each group, p < 0.01).
doxorubicin or etoposide. The treatment induced uniform foci formation in nearly all the nuclei, while the DNA content did not correlate with the intensity or number of ␥H2AX(S139) foci (Fig. 2A and B). However, when NKX3.1 was ectopically expressed before the treatments, it was observed that the ␥H2AX(S139) foci formation was partially reduced in some nuclei and the remaining foci were intense, although the size distribution was variable compared to the control cells (Fig. 2). Therefore, we divided the cells into groups according to the number of foci and analyzed the groups as previously described [23]. It was thus determined that doxorubicin- (only in the 20–50 foci range) but not etoposideinduced ␥H2AX(S139) foci formation was suppressed by NKX3.1 expression (Fig. 2C). The data suggest that NKX3.1 has no influence on the response to topo II inhibitor-mediated DNA damage. 3.3. Androgen decreases H2 O2 -mediated DNA damage in LNCaP cells Because doxorubicin generates reactive oxygen species (ROS) [27], to dissect the ability of doxorubicin to induce DNA damage through topo II inhibition and oxidative damage, H2 O2 -mediated foci formation and the effect of androgens on the oxidative DNA damage were examined in prostate cancer cells. First, LNCaP cells were treated with androgen (R1881), and then, time courses with H2 O2 were performed (Fig. 3A). Following an 8- to 10-fold upregulation of NKX3.1 expression, we investigated the levels of the following: (1) the expression of damage recognition factors; (2)
the androgen receptor status; and (3) the DNA damage response in these cells. When we examined the p-ATM(S1981) , pCHK2(T68) , p-p53(S20) and ␥H2AX(S139) levels, we observed that they were maintained at lower levels with androgen administration (Fig. 3A). Because the p-ATM(S1981) and ␥H2AX(S139) levels were higher in the control cells than the androgen-treated LNCaPs, the data led us to postulate that prolonged activation of the DDR might occur on androgen depletion due to a persistent failure to complete the repair process. Therefore, H2 O2 -mediated foci formation was also examined and quantitated with immunofluorescence labeling in the PC-3 cells (Fig. 3B and C). When the H2 O2 treatments were performed for 0.5 and 4 h, almost all the nuclei in the PC-3 cells had diffuse but non-uniform foci, and the ectopically expressed NKX3.1 partially reduced ␥H2AX(S139) foci formation in these cells. Also spontaneous- and H2 O2 -mediated ␥H2AX(S139) foci formation was low under the influence of androgens in LNCaP cells (Fig. 3D and E). The data suggest that either the damage or the response to the damage was low under the influence of NKX3.1 as well as androgens. Thus, significant differences were observed between the NKX3.1-expressing cells and the controls regarding the H2 O2 treatments, but no considerable differences were observed after the doxorubicin or etoposide treatments (Figs. 2C and 3). Then, we hypothesized that DSB formation might be a consequence of the collision of transcription and repair complexes and arise in the presence of topoisomerase inhibitors. Therefore, the observations directed us to investigate the physical association of NKX3.1 with the DDR mediators.
30
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
Fig. 2. Topo II inhibitor-mediated foci formation (at 0.5 or 4 h) is not influenced by NKX3.1 expression. (A) Doxorubicin and (B) etoposide-treated cells exhibited significant numbers of ␥H2AX(S139) foci (p < 0.05), while the ectopic expression of NKX3.1 did not influence the number of foci (Kruskal–Wallis, p > 0.05). (C) ␥H2AX(S139) foci-containing cells (due to varying intensity) were counted and separated into groups (four) to plot as percentages (2 independent experiments with n ≥ 100 for each group). The variations in the number of foci in each group are also shown as histogram plots in the control and NKX3.1-expressing groups.
3.4. NKX3.1 physically associates with cellular H2AX(S139) , RAD50, p-ATM(S1981) and DNA-PK, but not with ATM in PC-3 cells To test the possibility that there is a physical interaction of NKX3.1 with the DDR proteins, thereby regulating the initiation and level of foci formation upon DNA damage, the direct association of NKX3.1 with ␥H2AX(S139) , DNA-PKcs, ATM, p-ATM(S1981) and RAD50 was examined. The ectopic (in PC-3 and DU145 cells) and native, androgen-inducible (in LNCaP cells), expression of NKX3.1 was used in repair response studies (Fig. 4). Additionally, whether these interactions were altered by the administration of CPT-11 was also examined. First, NKX3.1 was ectopically expressed in PC3 (Fig. 4A–C) and DU145 cells (Fig. 4D) using the HM-NKX3.1 wt construct, and cellular ␥H2AX(S139) , RAD50, DNA-PKcs, ATM, and p-ATM(S1981) were then immunoprecipitated. Next, the NKX3.1 protein was immunoblotted with anti-NKX3.1 antibodies, and it was observed that NKX3.1 interacted with ␥H2AX(S139) and RAD50 somewhat, but this interaction was significantly increased by CPT11 treatment in PC-3 cells (Fig. 4A and C). When an IP was performed to detect p-ATM(S1981) in the absence and presence of CPT-11, we found that NKX3.1 also interacted with p-ATM(S1981) , consistent with the previous interactions, and this association increased with CPT-11 treatment. To reduce the possibility of false positives for subsequent ATM precipitations, p-ATM(S1981) was sequentially precipitated three times to remove most of it from the lysate. After this removal, we could not detect the interaction
of NKX3.1 with ATM (Fig. 4B). Furthermore, we also found that NKX3.1 physically associated with DNA-PKcs and that the interaction was reduced in CPT-11-treated DU145 cells. This is consistent with the level of DNA-PKcs expression obtained with western blot analysis (Fig. 4D). A reciprocal IP of DNA-PKcs was also performed (Fig. 4D). In addition, NKX3.1 was also precipitated from LNCaP cells before and after CPT-11-mediated DNA damage. It was observed that the increased expression level of NKX3.1 induced by androgen was accompanied by an increased physical association with p-ATM(S1981) (Fig. 4E). Furthermore, when the cells were transfected with the cDNA encoding the 234 aa full-length NKX3.1 (wt) or the mutant 63-Cter (99 aa) without the homeobox domain and treated with CPT-11 (Fig. 4F), we found that p-ATM(S1981) and ␥H2AX(S139) co-precipitated with NKX3.1 wt but not with the 63-Cter mutant, regardless of CPT-11 treatment (Fig. 4H). Taken together, NKX3.1 interacts with p-ATM(S1981) , and this physical association increases with CPT-11 treatment but does not change with androgen treatment; the wild-type NKX3.1 interacts with repair complexes, and this interaction requires the nuclear localization of NKX3.1 in prostate cells (Fig. 4F–H). Moreover, to gain additional insight into understanding the distribution of NKX3.1 and the co-localization of this protein with DNA damage mediators, we performed sub-cellular fractionation combined with western blot analysis in LNCaP cells. We observed that both native NKX3.1 and overexpressed NKX3.1 are in the soluble fraction along with ATM and RAD50, whereas NKX3.1 is
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
31
Fig. 3. Androgen protects prostate cells from DNA damage. (A) Androgen administration remarkably increases the AR, cyclin D1 and NKX3.1 expression levels in LNCaP cells. While the p-ATM(S1981) , p-CHK2(T68) , p-p53(S20) and ␥H2AX(S139) levels are decreased in androgen- and H2 O2 -treated cells, the expression levels of CHK2, p53, DNA-PKcs, Ku70, RAD50 and -actin are not significantly altered. Notably, the cellular ATM expression level remains approximately 2-fold depleted and is unchanged by androgen. (B) The H2 O2 -treated cells display a number of ␥H2AX(S139) foci, and (C) the ectopic expression of NKX3.1 in PC-3 cells suppresses the formation of these foci (2 independent experiments with n ≥ 100 for each group, p < 0.01). (D and E) Immunofluorescence data shows that androgen significantly (p < 0.001) suppresses the H2 O2 -induced ␥H2AX(S139) foci formation in LNCaP cells (n ≥ 20 for each group). The variations in number of foci in each group are shown as histogram plots in the control and treated groups.
only observed in the chromatin fraction when it is overexpressed, regardless of CPT-11 treatment. Consistently, the data suggest that a physical interaction of NKX3.1 with H2AX occurs at the chromatin level, and NKX3.1 interacts with ATM as well as RAD50 in soluble nuclear fraction (Fig. 4J). 3.5. NKX3.1 increases cyclin D1 expression We also observed that the ectopic expression of NKX3.1 resulted in relatively small sized nuclei in our studies, indicating that the NKX3.1 protein might trigger histone condensation as well as
decreased ␥H2AX(S139) foci formation (Fig. 5A). This observation led us to investigate whether the cells were under the control of a G1/S checkpoint and arrested at the G1 or S phase prior to DNA synthesis when NKX3.1 was expressed; therefore, cyclin D1 expression was examined. Cyclin D1 expression levels were measured in PC-3 cells treated with compounds (Fig. 5A) for 0.5 and 4 h, in addition to untreated and/or untransfected controls, using the image J software to measure the immunofluorescent intensity as an indication of cellular cyclin D1 expression level. We found that the average cyclin D1 expression level increased 2-fold over that of the controls with a high variance in the NKX3.1-expressing cells (Fig. 5B–E). Thus,
32
B. Erbaykent-Tepedelen et al. / Journal of Steroid Biochemistry & Molecular Biology 141 (2014) 26–36
Fig. 4. NKX3.1 physically associates with DDR factors. (A) The ␥H2AX(S139) physically associates with NKX3.1. (B) p-ATM(S1981) , but not ATM, also co-precipitates with NKX3.1 in PC-3 cells. (C) Also the cellular RAD50 and (D) DNA-PK associate with ectopic NKX3.1 in PC-3 and in DU145 cells respectively. Reciprocal IPs were also performed to confirm the DNA-PK interaction with NKX3.1. (E) The cellular p-ATM(S1981) associates with native NKX3.1 in LNCaP cells regardless of androgen administration. (F) Schema demonstrates that the NKX3.1 wt construct encodes the 234 aa protein with the intact homeobox domain, whereas the 63-Cter construct 99 amino acid peptide without the homeobox domain. (G) The 63-Cter deletion mutant exhibited ␥H2AX(S139) foci formation upon CPT-11-mediated DNA damage, whereas the NKX3.1 wt did not. (H) The p-ATM and ␥H2AX(S139) precipitations demonstrate that the association with NKX3.1 wt (28 kD) but not 63-Cter (12 kD) occurs regardless of drug treatment. -Actin blots were also performed to show the similarity of sample inputs. (J) Additionally, NKX3.1 is localized in both the soluble and chromatin fractions, whereas the native ATM and RAD50 are soluble in the nucleus, while ␥H2AX(S139) and H2AX are in the chromatin fraction.
the percent of cells that expressed cyclin D1 with a mean value >40 (relative intensity measurement) was 26% in controls and increased to 34% with NKX3.1 expression (data not shown). Moreover, when the cells with a mean value of expression intensity >20 and
