Loss of the candidate tumor suppressor BTG3 triggers acute cellular senescence via the ERK–JMJD3–p16INK4a signaling axis


The B-cell translocation gene 3 (BTG3) is a member of the antiproliferative BTG gene family and a downstream target of p53. BTG3 also binds and inhibits E2F1. Although it connects functionally two major growth-regulatory pathways, the physiological role of BTG3 remains largely uncharacterized. Here, we present evidence that loss of BTG3 in normal cells induced cellular senescence, which was correlated with enhanced ERK–AP1 signaling and elevated expression of the histone H3K27me3 demethylase JMJD3/KDM6B, leading to acute induction of p16INK4a. Importantly, we also found that BTG3 expression is specifically downregulated in prostate cancer, thus providing a physiological link with human cancers. Our data suggest that BTG3 may have a fail-safe role against tumorigenic progression.


The B-cell translocation gene 3 (BTG3) is a member of the B-cell translocation gene/transducer of ErbB2 antiproliferative protein family, which also includes BTG1, BTG2/PC3/Tis21, BTG4, Tob1 and Tob2 (Matsuda et al., 2001). The members of this protein family are characterized by a conserved N-terminal domain containing box A and box B signature motifs and a variable C-terminal domain. Sequence comparison indicates that Tob1 and Tob2, as well as BTG1 and BTG2, are highly similar, whereas BTG3 and BTG4, although similar to each other, are more distantly related to the other members of the BTG family (Winkler, 2009).

Overexpression of B-cell translocation gene/transducer of ErbB2 proteins is associated with inhibition of cell-cycle progression, which is mostly mediated by the conserved N-terminal domain. For example, BTG2 inhibits G1-to-S progression through the downregulation of cyclin D1 and cyclin E (Tirone, 2001), and BTG3 binds and inhibits E2F1, a transcription factor that is important for S-phase entry (Ou et al., 2007). Both BTG2 and BTG3 are p53 transcriptional targets, thus linking this protein family with the stress response (Guardavaccaro et al., 2000; Ou et al., 2007). Less is known regarding the functions of the structurally diverse C terminus compared with the N-terminal domain. The C terminus of BTG1 and BTG2 interacts with the protein arginine methyltransferase (Lin et al., 1996). Phosphorylation and ubiquitylation of the C-terminal domain have also been reported (Winkler, 2009).

Unexpected cell proliferation often triggers cellular senescence, as is observed after oncogene activation or deletion/inactivation of tumor suppressors (Collado and Serrano, 2010). This is believed to be mediated by the activation of the p16INK4a–RB pathway and of the ARF–p53 axis (Coppé et al., 2008). In addition, it was shown recently that persistent DNA damage also induces senescence (Rodier et al., 2009). The senescence phenotype is accompanied by widespread changes in protein and gene expression, as well as in secretion (Coppé et al., 2010). Accumulating evidence indicates that these alterations may have roles in the oncogenic development of cancers, serving either as a barrier to tumorigenesis or as a promoter of carcinogenesis (Bartkova et al., 2005, 2006; Coppé et al., 2010). In support of this notion, senescence and senescence-associated secretion have been observed in precancerous lesions (Chen et al., 2005; Bartkova et al., 2006; Majumder et al., 2008) and, in some cases, are accompanied by activation of DNA-damage signaling (Bartkova et al., 2005, 2006; Di Micco et al., 2006).

We demonstrated previously that BTG3 acts downstream of p53 during the stress response to bind and inhibit E2F1, therefore preventing entry into the S phase and also maintaining the G2/M arrest (Ou et al., 2007). However, several issues remain unaddressed; for example, the biological consequence of the loss of BTG3 in normal cells is largely uncharacterized. Here, we delineated further the effects of the loss of BTG3 in normal human fibroblasts, demonstrated its impact on senescence, and determined the underlying molecular basis of this phenomenon.


BTG3 depletion in normal human fibroblasts led to cellular senescence

Overexpression of the candidate tumor suppressor BTG3 suppresses cell growth; conversely, its deficiency enhances cell proliferation in cancer cells (Ou et al., 2007). To explore its regulatory role in normal cells, we first depleted BTG3 expression using two different small interfering RNAs (siRNAs) in normal human fibroblast IMR90 cells (Figure 1a). In contrast to what was observed previously in HCT116 cells (Ou et al., 2007), loss of BTG3 expression in IMR90 cells resulted in blunted cell proliferation, as demonstrated by a significant decrease of cells in the S phase, a moderate increase in the G1 population (Figure 1b), a marked reduction in BrdU incorporation (Figures 1c and d) and RB phosphorylation (Figure 1e). The slowed proliferation remained for at least for 2 weeks (Figure 1f), indicating that this is not a transient growth arrest. Additionally, BTG3-knockdown cells displayed the enlarged, flattened morphology that is typical of senescent cells (Supplementary Figure S1a). The phenotype was confirmed by staining the cells for senescence-associated β-galactosidase (SA-β-Gal) activity: we observed an increase in SA-β-Gal staining of at least sixfold 48 h after transfection with the BTG3-2 siRNA (Figure 1g), which was further enhanced 1 week after transfection (Supplementary Figure S1b). Induction of senescence was also apparent in another normal human fibroblast cell line (WI-38) after BTG3 downregulation (Figure 1h). Notably, upregulation of the G1 Cdk inhibitor p16INK4a, which is a hallmark of replicative and oncogene-induced senescence, was evident at the RNA and protein levels (Figures 1a and 2a). Together, these results indicate that BTG3 depletion in normal human fibroblasts induces growth arrest and a senescent phenotype, similar to what is observed during oncogenic stress.

Figure 1

BTG3-knockdown cells exhibit senescence-associated β-galactosidase (SA-β-Gal) activity. (a) Downregulation of BTG3 by two BTG3-targeting siRNAs, BTG3-1 and BTG3-2, as assessed by RT-PCR. Cells were collected at 18 and 48 h after siRNA transfection. Note that expression of p16INK4a was elevated upon BTG3 depletion. SC, scramble control. (b) Cell cycle analysis of IMR90 cells transfected with either control or BTG3-2 siRNA. The mean±s.d. from three experiments is shown. (c, d) Incorporation of BrdU was reduced in BTG3 knockdown IMR90 cells. BrdU incorporation assay was performed and the cells were stained with anti-BrdU antibody before examination by fluorescence microscopy. Quantification from three independent experiments is shown in (d). (e) RB is hypophosphorylated in BTG3-depleted IMR90 cells. (f) Growth curve of siRNA-transfected IMR90 cells in a period of 2 weeks following transfection. (g, h) Senescence-associated β-galactosidase (SA-β-Gal) assay. Normal human fibroblasts IMR90 (g) or WI-38 (h) were transfected with either control siRNA (SC) or BTG3-2 siRNA. Cells were fixed at 48 h after transfection and stained for SA-β-Gal activity. Positively stained cells were counted and mean±s.d. of three independent experiments is shown. Insets are immunoblots showing the extent of BTG3 downregulation.

Figure 2

BTG3-downregulated cells display hallmarks of cellular senescence. (a) p16INK4a was upregulated in BTG3-knockdown IMR90 cells at the protein level as assessed using immunoblotting. (b) Increased MMP-1 and PAI-1 expression in BTG3-downregulated cells. RNA was prepared from IMR90 cells 48 h after siRNA transfection and RT–PCR was performed using specific primers. (c) Quantification of results shown in (b). The mean±s.d. of three independent experiments is displayed. (d) BTG3-downregulated IMR90 cells display senescence-associated heterochromatic foci. Cells were fixed 2 days after siRNA transfection and processed for immunofluorescence microscopy after staining with antibodies specific for BTG3, HP1-γ and histone H3 K9 trimethylation (H3K9me3). Nuclei were counterstained with 4,6-diamidino-2-phenylindole (DAPI). (e) Quantification of foci shown in (d). The mean±s.d. of three independent experiments is displayed.

BTG3-downregulated cells displayed characteristic senescence-associated markers

To confirm the senescence phenotype, other senescence markers were examined. In addition to the p16INK4a gene, the expression of other senescence-associated genes, such as MMP-1 and PAI-1, was also elevated as assessed using RT–PCR (Figures 2b and c). Cellular senescence is accompanied by facultative epigenetic changes that contribute to the repression of many proliferation-promoting genes (Narita et al., 2003). Therefore, we used immunofluorescence microscopy to examine whether these changes existed in BTG3-downregulated cells. Our results showed that senescence-associated heterochromatic foci were evident in BTG3-knockdown cells after 4,6-diamidino-2-phenylindole staining compared with the control (Figures 2d and e). Furthermore, foci of heterochromatin-associated markers, such as HP1-γ and histone H3 K9 trimethylation (H3K9me3), were also markedly increased in BTG3-depleted cells (Figures 2d and e).

Cellular senescence can also be triggered by DNA damage as a result of oncogenic activation (Mallette and Ferbeyre, 2007). To determine whether DNA damage occurred in BTG3-knockdown cells, possible γ-H2AX and phospho-ATM (pS1981) foci were sought. In contrast to the robust γ-H2AX and phospho-ATM (pS1981) staining observed in X-ray-irradiated cells, the γ-H2AX and phospho-ATM (pS1981) foci in BTG3-depleted cells were barely detectable, and almost none were detected in the control-siRNA-transfected cells (Supplementary Figures S2a and b). Furthermore, the levels of p53 and p21 were not significantly altered by BTG3 downregulation (Supplementary Figure S2c). Additionally, stress generated by oncogenic ras overexpression, which has been shown to induce DNA damage and senescence in normal human cells, also did not appear to alter BTG3 levels (Supplementary Figure S2d), suggesting oncogenic stress and BTG3 loss may contribute to cellular senescence in a parallel fashion. These data indicate that acute loss of BTG3 triggers a senescence response that is minimally attributable to DNA damage and is probably a result of elevated p16INK4a expression.

BTG3 downregulation elevated p16INK4a expression through AP1-mediated transcriptional activation of JMJD3/KDM6B

The p16INK4a locus is silenced by histone H3 K27 trimethylation (H3K27me3) in actively proliferating human fibroblasts (Bracken et al., 2007; Kotake et al., 2007), which is derepressed in oncogenic stress-induced senescent cells by the demethylase JMJD3/KDM6B, leading to increased expression of p16INK4a in these cells (Agger et al., 2009; Barradas et al., 2009). In an attempt to decipher the molecular basis of the cellular senescence induced by the loss of BTG3, we compared the expression of JMJD3 in control and BTG3-downregulated cells. Using RT–PCR, we found that the expression of JMJD3 was increased in BTG3-knockdown cells (Figure 3a). This was, at least in part, mediated by the JMJD3 promoter, as a 400 bp proximal promoter region (–2.4 to –2.0 kb relative to the macrophage-specific transcription start site; JM-P2; Figure 3b) conferred enhanced activity in a reporter assay after depletion of BTG3 (Figure 3c). In comparison, the activity of a 1 kb distal promoter region (–3.4 to –2.4 kb; JM-P1) was not significantly altered (Figure 3c), which demonstrates the specificity of the effect.

Figure 3

BTG3 downregulation leads to upregulation of JMJD3 by enhanced ERK-AP1 signaling. (a) Upregulation of JMJD3 in BTG3-knockdown IMR90 cells, as demonstrated by RT–PCR. The lower panel depicts the quantification of the results of three independent experiments. (b) Schematic diagram depicting the JMJD3 promoter upstream of the macrophage-specific transcription start site (MF-TSS). P1 (–3.4 to –2.4 kb) and P2 (–2.4 to –2.0 kb) represent the two promoter regions tested using the luciferase reporter assay. m1 and m2 are two P2 mutants with indicated substitutions in the putative AP1 sites. (c) Activity of the luciferase reporter driven by JMJD3 P1 or P2 in control or in BTG3-depleted IMR90 cells. (d) Luciferase reporter activity driven by wild-type, m1 or m2 P2 promoter in control or BTG3-downregulated IMR90 cells. (e) Increased phospho-ERK (p-ERK) after BTG3 knockdown. Levels at 18 and 24 h after siRNA transfection are shown. (f) Increased binding of c-Jun to the JMJD3 promoter in BTG3-ablated IMR90 cells. Chromatin immunoprecipitation (ChIP) assays were performed using anti-c-Jun antibody or control rabbit IgG (R-IgG). Binding of c-Jun to the JMJD3 promoter was detected by PCR using primers corresponding to the P2 region. Actin promoter binding was used as a negative control. The amounts of immunoprecipitated c-Jun were determined by immunoblotting (WB). (g) Reduced histone H3 K27 trimethylation on the p16INK4a promoter upon depletion of BTG3. ChIP was performed using anti-H3K27me3 antibody or control R-IgG. The Cyclin D1 (CCND1) promoter was used as a control.

It was speculated that oncogenic BRAF induces the expression of JMJD3 via the AP1 transcription factor (Agger et al., 2009). To determine if AP1 was involved in senescence induced by BTG3 depletion, we scanned the JMJD3 promoter for probable AP1-binding site. Two conserved AP1 sites were found in the JMJD3 proximal promoter region (Figure 3b); mutation of either site dampened the activity of the promoter severely and at the same time abolished the difference between control and BTG3-depleted cells (Figure 3d), suggesting that AP1 has a role in the upregulation of JMJD3 in BTG3-knockdown cells. In support of this contention, upregulation of phospho-ERK, which is an upstream activator of AP1, was observed 24 h after transfection of the BTG3 siRNA (Figure 3e). In addition, binding of c-Jun, which is a component of AP1, to the endogenous JMJD3 proximal promoter was also increased after abrogation of the expression of BTG3 (Figure 3f). Furthermore, the H3K27me3 mark on the p16 promoter was reduced in BTG3 knockdown cells, despite that the overall H3K27me3 in cells was not significantly altered (Figure 3g).

Taken together, these data indicate that loss of BTG3 may trigger the upregulation of the ERK–AP1 signaling axis, leading to increased JMJD3 expression, which in turn derepresses p16INK4a and results in cellular senescence.

Inhibition of ERK signaling abrogated the induction of JMJD3/KDM6B and reduced the senescence phenotype in BTG3-downregulated cells

To confirm the involvement of the ERK signaling pathway, we treated BTG3-knockdown IMR90 cells with the ERK-specific inhibitor PD98059. As a result, a significant reduction in senescence was observed when ERK was inhibited (Figure 4a). In contrast, two other unrelated drugs, SB202190 (p38 inhibitor) and SB202474 (negative control), had no apparent effect, indicating the specificity of the treatment (Figure 4a). Notably, the senescence phenotype was not completely suppressed by the ERK inhibitor, suggesting that additional pathway may also be involved.

Figure 4

Inhibition of ERK or depletion of p16INK4a abrogates the senescence phenotype in BTG3-depleted cells. (a) Inhibition of ERK, but not p38, mitigated the senescence phenotype. IMR90 cells were transfected with control or BTG3-2 siRNA and treated with PD98059 (ERK inhibitor, 10 μM), SB202190 (p38 inhibitor, 5 μM), or SB202474 (negative control, 5 μM) for 24 h. The drugs were then removed and the cells were cultured for an additional 20–24 h before fixation for β-Gal staining. For immunoblotting (right panel), cells were treated with the inhibitors for 12 h before collection. (b) Induction of JMJD3 and p16INK4a was diminished after ERK inhibition. RNA was prepared from cells treated as described in (a) and analyzed using RT–PCR. (c) Downregulation of p16INK4a and p53 by specific siRNAs in IMR90 cells. (d) Senescence induced by BTG3 depletion was suppressed by co-depletion of p16INK4a. SA-β-Gal assays were performed on cells transfected with indicated siRNAs. Mean±s.d. from three independent experiments is shown.

The expression of JMJD3 in the presence of the ERK inhibitor was also examined using RT–PCR. Similar to the effect on senescence, the specific inhibitor PD98059 blunted the induction of JMJD3 and p16INK4a (Figure 4b), demonstrating the requirement of active ERK in the signaling cascade that leads to the expression of JMJD3 and p16INK4a.

To determine if elevated p16INK4a expression is the major contributing factor in the senescence driven by loss of BTG3, we simultaneously deplete BTG3 and p16 using siRNAs (Figure 4c). As a result, the senescence phenotype was completely suppressed as assessed by SA-β-Gal staining (Figure 4d), indicating the increase in p16INK4a expression has a major role in the senescence observed in our study. Interestingly, simultaneous depletion of p53 yielded partial suppression (Figure 4d), suggesting that the contribution from the p53 pathway cannot be completely discounted.

BTG3 is downregulated in prostate cancer

As the loss of tumor-suppressor genes such as NF1, VHL and PTEN triggers senescence, which serves as a fail-safe barrier to oncogenic progression (Chen et al., 2005; Courtois-Cox et al., 2006; Young et al., 2008), we wondered whether BTG3 has a similar role in human cancers. To explore this possibility, we first determined whether the loss of BTG3 occurs in human cancers. We screened panels of cDNA arrays derived from human cancer specimens of various types and stages for expression of BTG3 using real-time qPCR. Preliminary results using a limited number of samples indicated that BTG3 was downregulated in later stages of kidney and prostate cancers (Supplementary Figure S3). Notably, we did not find reduction of BTG3 expression in lung cancer as previously reported (Yoneda et al., 2009). We then performed additional analyses using an expanded cohort of prostate cancer samples, which confirmed the downregulation of BTG3 (Figure 5a). Significantly, this trend was not shared by another member of the BTG family BTG2, suggesting the functional disparity between these two proteins in prostate (Figure 5a).

Figure 5

Downregulation of BTG3 in prostate cancer. (a) Progressive decrease of BTG3 but not BTG2 expression in clinical prostate cancer specimens. PCR arrays containing cDNA derived from clinical specimens of various stages of prostate cancer were assayed by qPCR using primers specific for BTG2 or BTG3. The averages of three independent experiments are shown. Sample size is denoted by ‘n’. (b) IHC staining of prostate cancer specimens. Tissue sections of paraffin-embedded prostate specimens were stained with anti-BTG3 antibody and counterstained with hematoxylin. Paired tumor and non-tumor regions from 3 representative cases are shown. Scale bars represent 100 μm. (c) qPCR analysis of RNA prepared from tumor and neighboring non-tumor tissues captured using LCM. The averages of paired normal/tumor tissues from four different cases are shown.

We also examined the expression of BTG3 by immunohistochemical staining of prostate sections from clinical specimens. Of the ten cases examined, seven revealed strong BTG3 staining in the nuclei of basal cells in the histologically normal region, but not in the tumor region (Figure 5b, cases no. 1 and no. 3). The other three showed weak nuclear positivity for BTG3 in the basal cells of non-tumor parts (Figure 5b, case no. 2).

The difference in the expression of BTG3 was revealed further by qPCR analysis of the tumor and the adjacent non-tumor tissues captured separately by laser microdissection (LCM), which further demonstrated the clear reduction of BTG3 expression in the carcinoma regions compared with the neighboring non-tumor regions (Figure 5c).


The data presented in our present study revealed previously uncharacterized roles of BTG3, that is, antagonization of ERK signaling, and induction of JMJD3 and cellular senescence after its loss in normal human cells. These results, together with previous findings by Agger et al. (2009) and Barradas et al. (2009), who demonstrated the role of JMJD3 in derepressing p16INK4a in oncogene-induced senescence, support a model in which BTG3 depletion induces senescence by enhancing the expression of p16INK4a at the mRNA and protein levels, which is, at least in part, a result of increased expression of JMJD3, the histone H3K27me3 demethylase, driven by enhanced ERK signaling and AP1 activity (Figure 6). The model is significant especially in light of the accumulating evidence implicating cellular senescence (which is often detected in precancerous lesions) as a barrier to malignant transformation (Bartkova et al., 2005, 2006; Chen et al., 2005; Courtois-Cox et al., 2006). In support, we have also found reduced expression of BTG3 in prostate cancer (Figure 5) and in kidney cancer (Supplementary Figure S3). How BTG3 is involved in keeping the ERK activation at bay is still unclear though it has been reported to negatively regulate the Src tyrosine kinase activity in PC12 cells (Rahmani, 2006). It remains to be determined whether direct physical interaction with BTG3 is involved in the inhibition of ERK or of any of its upstream regulators.

Figure 6

Model depicting the BTG3–ERK–JMJD3–p16INK4a signaling axis, which is enhanced after abrogation of BTG3 and acts to suppress the development of cancer in a fail-safe mechanism.

In addition to oncogenic stress, loss of tumor-suppressor genes may also trigger senescence. Significantly, inactivation of the neurofibromin gene NF1, which encodes a RasGAP, induces senescence in human fibroblasts; in addition, senescence was observed in human neurofibromas (Courtois-Cox et al., 2006). Acute loss of Pten in the mouse prostate induces senescence; accordingly, senescence was detected in early stages of human prostate cancer (Chen et al., 2005). Furthermore, loss of VHL causes an RB- and p27-dependent senescent response in cells and in the mouse (Young et al., 2008). The findings reported here bear a resemblance with these previous discoveries. Although our findings are not a direct demonstration, they provide further support to the role of BTG3 as a tumor suppressor.

Our discovery that the INK4a locus was regulated by BTG3 is interesting in that BTG3 appears to suppress the expression of the two products of the locus, p16INK4a and ARF, albeit by two separate pathways. In addition to the ERK–JMJD3–p16INK4a axis shown in this study, we demonstrated previously that BTG3 may prevent ARF expression by inhibiting E2F (Ou et al., 2007). There could be several implications of this function. On the one hand, BTG3 may prevent cells from going into permanent growth arrest, for example, after DNA damage. On the other hand, a fail-safe mechanism may be implemented after loss of BTG3 to avoid uncontrolled cell proliferation, as we demonstrated here that the loss of BTG3 activates the MAPK pathway. The choice and the regulation of these pathways may have important implications for the strategic approach in treating cancer.

One relevant question that arises from our study is whether inactivation of Rb or p53 would rescue the senescent phenotype induced by BTG3 loss and cause neoplastic transformation, as observed for oncogene-induced senescence. We tested this possibility in telomerase-immortalized BJ fibroblasts. However, the combination of BTG3 depletion and expression of large T antigen (which simultaneously inactivates Rb and p53) failed to transform these fibroblasts in a soft-agar assay (data not shown). This result suggests that loss of BTG3 is not functionally equivalent to oncogene activation in cellular transformation, although it does enhance cancer cell proliferation (Ou et al., 2007). This conclusion is also supported by our immunofluorescence data, which showed that unlike oncogene activation, loss of BTG3 did not trigger mass amount of DNA damage (Supplementary Figures S2a and b), nor did it significantly induce the p53 response (Supplementary Figure S2c). Apparently, much remains to be learned regarding the regulation of cell proliferation by BTG3.

Another question raised by our findings is whether loss of BTG3 expression correlates with gain of p16 INK4a expression in a clinical setting, especially in benign tumors. To address this issue, we have examined benign prostate tumor specimens in which large area of carcinoma have not developed. In one of the three cases we studied, we observed partial loss of BTG3 expression and gain of expression of p16INK4a in the benign region; in the carcinoma region, neither BTG3 nor p16 INK4a was detected in any of the three cases we examined. This observation, though with limited sample numbers, implicates the possible role of BTG3 as a fail-safe anticancer barrier, which upon its loss, induces senescence by upregulation of p16INK4a and further inactivation of the latter might trigger the onset of malignant cancers. Further investigation with larger cohort is warranted to solidify this conclusion.

Materials and methods

Cell culture and treatment

IMR90 and WI38 cells were maintained in Eagle's Minimum Essential Medium (HyClone, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA, USA), 100 U/ml penicillin, 100 μg/ml streptomycin (Invitrogen), 0.1 mM nonessential amino acids, 2 mM L-glutamine and 1 mM sodium pyruvate.

The MAP kinase inhibitor set was from Calbiochem (Gibbstown, NJ, USA).

Cell lysis and immunoblotting

Cell lysates were prepared as described (Ou et al., 2005). The antibodies used for immunoblotting were as follows: anti-actin and anti-α-tubulin from Sigma-Aldrich (St Louis, MO, USA), anti-c-Jun (no. 9165) from Cell Signaling (Danvers, MA, USA), anti-ERK1&2 (pTpY185/187) (no. 44680G) and anti-ERK1&2 (no. 44-654G) from Invitrogen, anti-Human Rb (no. 554136, BD Pharmingen, San Diego, CA, USA), anti-c-H-Ras (OP23T, Calbiochem), mouse anti-Histone H3(trimethyl-K27)(ab6002, abcam, Cambridge, MA, USA) and anti-p16INK4a (K0077-3, MBL, Nagoya, Japan).

β-Galactosidase (β-Gal) staining

Cells were washed in PBS, fixed for 3–5 min at room temperature in 2% formaldehyde/0.2% glutaraldehyde, washed, and incubated at 37 °C (no CO2) with freshly prepared senescence-associated β-Gal (SA-β-Gal) stain solution (1 mg/ml 5-bromo-4-chloro-3-indolyl β-D-galactoside (X-Gal), 40 mM citric acid/sodium phosphate pH 6.0, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM NaCl and 2 mM MgCl2). Staining was evident at 2–4 h and maximal at 12–16 h. Cells were observed using a bright-field microscope at × 100–200 magnification.


For RT–PCR, total RNA prepared by TRIzol extraction (Invitrogen) was reverse-transcribed using the Super-Script III reverse transcriptase (Invitrogen), and was then amplified by PCR. The primer pairs used are listed in Supplementary Table S1.

qPCR was performed on TissueScan Cancer Survey Panels (CSRT101, OriGene, Rockville, MD, USA) and TissueScan Prostate Cancer quantitative PCR Panels (HPRT502, OriGene) using an ABI Prism 7500 thermocycler (Applied Biosystems Inc., Foster City, CA, USA). The primers used in this assay are listed in Supplementary Table S2.

BrdU incorporation

To analyze cell proliferation, IMR-90 cells were plated on coverslips. After overnight growth, the cells were transfected with siRNA for 48 h. BrdU (10 μM) was added 3 h before fixation with 4% paraformaldehyde/PBS for 15 min at room temperature. Following treatment with 2 M HCl for 20 min and neutralization with 0.1 M sodium borate (pH 8.5), the cells were permeabilized in 0.5% Triton X-100/PBS and then blocked in 5% BSA/PBS. Anti-BrdU (no. 555627, BD Pharmingen, San Diego, CA, USA) antibody was added and incubated for 1 h at room temperature followed by the secondary antibody DyLight 488-conjugated AffiniPure goat anti-mouse IgG antibody (115-485-006, Jackson ImmunoResearch, West Grove, PA, USA). DNA was counterstained with 4,6-diamidino-2-phenylindole. The images were obtained using a Carl Ziess Axiovert 200 M Live Cell Observation System (Carl Zeiss Inc., Thornwood, NY, USA).


IMR90 cells grown on cover slips were fixed in 4% formaldehyde/PBS and permeabilized in 0.5%. Triton X-100/PBS at room temperature for 10 min. Cells were then blocked in 5% BSA/PBS at room temperature for 15 min, followed by incubation with primary antibodies for 1 h at room temperature. The antibodies used for immunofluorescence were: mouse anti-HP1γ (H00011335-M01, Abnova, Taipei, Taiwan) and rabbit anti-trimethyl-histone H3 (Lys9) (no. 07-442, Upstate, Temecula, CA, USA). For BTG3 staining, cells were incubated with the anti-BTG3 antibody (Ou et al., 2007) overnight at 4 °C. An FITC-conjugated goat anti-rabbit IgG antibody and a TRITC-conjugated goat anti-mouse IgG antibody (Jackson ImmunoResearch) or a DyLight488-conjugated goat anti-mouse IgG antibody and a TRITC-conjugated goat anti-rabbit IgG antibody (Jackson ImmunoResearch) were used as secondary antibodies. DNA was counterstained with 4,6-diamidino-2-phenylindole. Images were obtained using a Carl Zeiss Axiovert 200 M Live Cell Observation System (Carl Zeiss Inc.) or a Bio-Rad Radiance-2100 Confocal Microscope (Bio-Rad, Hercules, CA, USA).

siRNA transfection

All siRNAs were synthesized and annealed by Sigma-Aldrich. Transfection of siRNA was performed using Oligofectamine (Invitrogen). The sequences targeted by the BTG3-1 and BTG3-2 siRNA are 5′-IndexTermGGCTAGTTCGAAAACATGA-3′ and 5′-IndexTermTTGAGAGGTTTGCTGAGAA-3′, respectively.

Plasmids and constructs

For the luciferase reporter assay, the JMJD3 promoter regions were PCR amplified from MRC5 genomic DNA and cloned between the SacI and XhoI sites (P1) or the KpnI and XhoI sites (P2) of the pGL3 promoter vector (Promega, Madison, WI, USA). The m1 and m2 mutants of the P2 promoter region were generated by PCR-coupled site-directed mutagenesis and confirmed by DNA sequencing.

Luciferase reporter assay

IMR90 cells were first transfected with control or BTG3-2 siRNA using Oligofectamine (Invitrogen). The following day, cells were transfected with the luciferase reporter plasmids using Lipofectamine 2000 (Invitrogen). For normalization purposes, pSV40-β-gal-expressing β-galactosidase was cotransfected as an internal control.

Chromatin immunoprecipitation and PCR amplification

IMR-90 cells were transfected with control or BTG3-2 siRNA using Oligofectamine for 24 h before chromatin immunoprecipitation was performed as described previously (Ou et al., 2005) using a rabbit anti-c-Jun antibody (no. 9165, Cell Signaling) or a rabbit anti-trimethyl-Histone H3(Lys27)(no. 07-449, Millipore, Billerica, MA, USA). The primers used in the assays are listed in Supplementary Table S3.


Prostate tissue sections were obtained from the National Taiwan University (NTU) Hospital using protocols approved by NTU (200812101R) and by Academia Sinica (AS-IRB01-08080). The tissue sections (5 μm in thickness) were deparaffinized in two changes of xylene and rehydrated through a series of graded ethanol solutions. Antigen retrieval was performed using 10 mM citrate buffer (pH 6.0) or a commercial antigen retrieval buffer (pH 6.0, S1700, Dako, Glostrup, Denmark). Slides were quenched in 3% hydrogen peroxide and blocked with TBST (50 mM Tris pH 7.6, 150 mM NaCl and 0.1% Tween 20) containing 5% normal goat serum. The primary antibody (rabbit anti-phospho-Akt-Ser473) was diluted in TBST/5% normal goat serum and incubated overnight at 4 °C. For BTG3 staining, the anti-BTG3 antibody was diluted and incubated in PBST (PBS with 0.05% Tween20)/5% normal goat serum for 1 h at room temperature. Detection was performed using a SuperPicTure Polymer Detection Kit (87-9663, Invitrogen) or an Envision System Labeled polymer-HRP kit (K4002, Dako).

For LCM, the formaldehyde-fixed paraffin-embedded tissue sections were deparaffinized, rehydrated and counterstained with hematoxylin. Normal epithelial cells and carcinoma cells were identified and captured separately from each section using the PixCell II LCM System (Arcturus, Mountain View, CA, USA). On average, a total of 150 mm2 of tissue was collected for each cell population using several CapSure HS LCM Caps. RNA was extracted and reverse transcribed using a ParadisePlus Whole Transcript Reverse Transcription kit (Arcturus), as instructed. The quality of the cDNA was confirmed by PCR and gel electrophoresis. Quantitative real-time PCR was then performed using an ABI Prism 7500 thermocycler (Applied Biosystems Inc.). The primers used in this assay are listed in Supplementary Table S2.


  1. Agger K, Cloos PAC, Rdkjaer L, Williams K, Andersen G, Christensen J et al. (2009). The H3K27me3 demethylase JMJD3 contributes to the activation of the INK4A-ARF locus in response to oncogene- and stress-induced senescence. Genes Dev 23: 1171–1176.

    CAS  Article  Google Scholar 

  2. Barradas M, Anderton E, Acosta JC, Li S, Banito A, Rodriguez-Niedenführ M et al. (2009). Histone demethylase JMJD3 contributes to epigenetic control of INK4a/ARF by oncogenic RAS. Genes Dev 23: 1177–1182.

    CAS  Article  Google Scholar 

  3. Bartkova J, Hořejši Z, Koed K, Krämer A, Tort F, Zieger K et al. (2005). DNA damage response as a candidate anti-cancer barrier in early human tumorigenesis. Nature 434: 864–870.

    CAS  Article  Google Scholar 

  4. Bartkova J, Rezaei N, Liontos M, Karakaidos P, Kletsas D, Issaeva N et al. (2006). Oncogene-induced senescence is part of the tumorigenesis barrier imposed by DNA damage checkpoints. Nature 444: 633–637.

    CAS  Article  Google Scholar 

  5. Bracken AP, Kleine-Kohlbrecher D, Dietrich N, Pasini D, Gargiulo G, Beekman C et al. (2007). The Polycomb group proteins bind throughout the INK4A–ARF locus and are disassociated in senescent cells. Genes Dev 21: 525–530.

    CAS  Article  Google Scholar 

  6. Chen Z, Trotman LC, Shaffer D, Lin H-K, Dotan ZA, Niki M et al. (2005). Crucial role of p53-dependent cellular senescence in supression of Pten-deficient tumorigenesis. Nature 436: 725–730.

    CAS  Article  Google Scholar 

  7. Collado M, Serrano M . (2010). Senescence in tumours: evidence from mice and humans. Nat Rev Cancer 10: 51–57.

    CAS  Article  Google Scholar 

  8. Coppé J-P, Desprez P-Y, Krtolica A, Campisi J . (2010). The senescence-associated secretory phenotype: the dark side of tumor suppression. Annu Rev Pathol Mech Dis 5: 99–118.

    Article  Google Scholar 

  9. Coppé J-P, Patil CK, Rodier F, Sun Y, Munoz DP, Goldstein J et al. (2008). Senescence-associated secretory phenotypes reveal cell-nonautonomous functions of oncogenic RAS and the p53 tumor suppressor. PLoS Biol 6: 2853–2868.

    Article  Google Scholar 

  10. Courtois-Cox S, Williams SMG, Reczek RE, Johnson BW, McGillicuddy LT, Johannessen CM et al. (2006). A negative feedback signaling network underlies oncogene-induced senescence. Cancer Cell 10: 459–472.

    CAS  Article  Google Scholar 

  11. Di Micco R, Fumagalli M, Cicalese A, Piccinin S, Gasparini P, Luise C et al. (2006). Oncogene-induced senescence is a DNA damage response triggered by DNA hyper-replication. Nature 444: 638–642.

    CAS  Article  Google Scholar 

  12. Guardavaccaro K, Corrente G, Covone F, Micheli L, Dagnano I, Starace G et al. (2000). Arrest of G1-S progression by the p53-inducible gene PC3 is Rb dependent and relies on the inhibition of Cyclin D1 transcription. Mol Cell Biol 20: 1797–1815.

    CAS  Article  Google Scholar 

  13. Kotake Y, Cao R, Viatour P, Sage J, Zhang Y, Xiong Y . (2007). pRB family proteins are required for H3K27 trimethylation and Polycomb repression complexes binding to and silencing p16INK4a tumor suppressor gene. Genes Dev 21: 49–54.

    CAS  Article  Google Scholar 

  14. Lin W-J, Gary JD, Yang MC, Clarke S, Herschman HR . (1996). The mammalian immediate-early TIS21 protein and the leukemia-associated BTG1 protein interact with a protein-arginine N-methyltransferase. J Biol Chem 271: 15034–15044.

    CAS  Article  Google Scholar 

  15. Majumder PK, Grisanzio C, O'Connell F, Barry M, Brito JM, Xu Q et al. (2008). A prostatic intraepithelial neoplasia-dependent p27Kip1 checkpoint induces senescenc and inhibits cell proliferation and cancer progression. Cancer Cell 14: 146–155.

    CAS  Article  Google Scholar 

  16. Mallette FA, Ferbeyre G . (2007). The DNA damage signaling pathway connects oncogenic stress to cellular senescence. Cell Cycle 6: 1831–1836.

    CAS  Article  Google Scholar 

  17. Matsuda S, Rouault R-P, Magaud J-P, Berthet C . (2001). In search of a function for the TES21/PC3/BTG1/TOB family. FEBS Lett 497: 67–72.

    CAS  Article  Google Scholar 

  18. Narita M, Nuñez S, Heard E, Narita M, Lin AW, Heran SA et al. (2003). Rb-mediated heterochromatin formation and silencing of E2F target genes during cellular senescence. Cell 113: 703–716.

    CAS  Article  Google Scholar 

  19. Ou Y-H, Chung P-H, Hsu F-F, Sun T-P, Chang W-Y, Shieh S-Y . (2007). The candidate tumor suppressor BTG3 is a transcriptional target of p53 that inhibits E2F1. EMBO J 26: 3968–3980.

    CAS  Article  Google Scholar 

  20. Ou Y-H, Chung P-H, Sun T-P, Shieh S-Y . (2005). p53 C-terminal phosphorylation by CHK1 and CHK2 participates in the regulation of DNA damage-induced C-terminal acetylation. Mol Biol Cell 16: 1684–1695.

    CAS  Article  Google Scholar 

  21. Rahmani Z . (2006). APRO4 negatively regulates Src tyrosine kinase activity in PC12 cells. J Cell Sci 119: 646–658.

    CAS  Article  Google Scholar 

  22. Rodier F, Coppé J-P, Patil CK, Hoeijmakers WAM, Muñoz DP, Raza SR et al. (2009). Persistent DNA damage signaling triggers senescence-associated inflammatory cytokine secretion. Nat Cell Biol 11: 973–979.

    CAS  Article  Google Scholar 

  23. Tirone F . (2001). The gene PC3TIS21/BTG2, prototype member of the PC3/BTG/TOB family: regulator in control of cell growth, differentiation, and DNA repair? J Cell Physiol 187: 155–165.

    CAS  Article  Google Scholar 

  24. Winkler GS . (2009). The mammalian anti-proliferative BTG/Tob protein family. J Cell Physiol 222: 66–72.

    Article  Google Scholar 

  25. Yoneda M, Suzuki T, Nakamur T, Ajima R, Yoshida Y, Kakuta S et al. (2009). Deficiency of antiproliferative family protein Ana correlates with development of lung adenocarcinoma. Cancer Sci 100: 225–232.

    CAS  Article  Google Scholar 

  26. Young AP, Schlisio S, Minamishima YA, Zhang Q, Li L, Grisanzio C et al. (2008). VHL loss actuates a HIF-independent senescence programme mediated by Rb and p400. Nat Cell Biol 10: 361–367.

    CAS  Article  Google Scholar 

Download references


This work was supported by grants from the National Science Council of Taiwan and Academia Sinica to S-Y Shieh.

Author information



Corresponding author

Correspondence to S-Y Shieh.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lin, TY., Cheng, YC., Yang, HC. et al. Loss of the candidate tumor suppressor BTG3 triggers acute cellular senescence via the ERK–JMJD3–p16INK4a signaling axis. Oncogene 31, 3287–3297 (2012). https://doi.org/10.1038/onc.2011.491

Download citation


  • BTG3
  • ERK
  • cellular senescence

Further reading