Mitogen-activated protein kinase p38 and retinoblastoma protein signalling is required for DNA damage-mediated formation of senescence-associated heterochromatic foci in tumour cells

Authors


  • These authors contributed equally to this work.

Abstract

DNA-damaging agents are able to induce irreversible cell growth arrest and senescence in some types of tumour cells, thus contributing to the static feature of cancer. However, senescent tumour cells may re-enter the cell cycle, leading to tumour relapse. Understanding the mechanisms that control the viability of senescent cells may be critical for tumour suppression. Primary human fibroblasts undergoing oncogene-induced or replicative senescence are known to form senescence-associated heterochromatin foci (SAHF), which contribute to the stability of the senescent state. However, it is unclear whether SAHF formation is universal in tumour cells. We report that the DNA-damaging agents doxorubicin and 7-ethyl-10-hydroxycamptothecin were able to induce the formation of SAHF in some tumour cell types, and this induction was accompanied by activation of the retinoblastoma protein pathway. By contrast, tumour cells in which the retinoblastoma protein pathway could not be activated by doxorubicin or 7-ethyl-10-hydroxycamptothecin failed to form SAHF. In parallel, tumour cells with deficient retinoblastoma protein were also unable to form SAHF. In addition, we show that the mitogen-activated protein kinase p38 pathway was involved in tumour cell SAHF formation in response to doxorubicin and 7-ethyl-10-hydroxycamptothecin. Furthermore, HMG box transcription factor 1 (HBP1), a downstream target of the mitogen-activated protein kinase p38-mediated senescence pathway, was required for SAHF formation. Taken together, the results of the present study highlight the roles of the mitogen-activated protein kinase p38/retinoblastoma protein pathway in tumour cell SAHF formation in response to DNA-damaging agents, and provide new insights into the mechanisms of DNA damage-mediated tumour suppression.

Structured digital abstract

Abbreviations
BrdU

5-bromo-2′-deoxyuridine

ChIP

chromatin immunopreciptation

DAPI

4′,6-diamidino-2-phenylindole

Dox

doxorubicin

ERK

extracellular signal-regulated kinase

HBP1

HMG box transcription factor 1

HP1

heterochromatin-associated protein 1

HPV-16 E7

human papilloma virus-16 E7

HPV-16 E7

human papilloma virus-16 E7

JNK

Jun N-terminal protein kinase

MAPK

mitogen-activated protein kinase

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide

p38 MAPK

mitogen-activated protein kinase p38

PML

promyelocytic leukaemia

Rb

retinoblastoma protein

RNAi

RNA interference

SA-β-gal

senescence-associated β-galactosidase activity

siRNA

small interfering RNA

SN-38

7-ethyl-10-hydroxycamptothecin

Introduction

Cellular senescence is an irreversible cell cycle arrest that limits the proliferation of damaged cells and may act as a natural barrier to cancer progression. Several markers of cellular senescence have been identified, including lack of 5-bromo-2′-deoxyuridine (BrdU) incorporation, enhanced senescence-associated β-galactosidase activity (SA-β-gal), increased numbers of nuclear promyelocytic leukaemia (PML) bodies and persistent DNA damage-induced nuclear heterochromatinization [1-3]. The latter, manifested by the formation of senescence-associated heterochromatin foci (SAHF), stains intensely with the DNA dye 4′,6-diamidino-2-phenylindole (DAPI). SAHF are specialized domains of facultative heterochromatin that form in a number of human senescent somatic cells [2]. SAHF contain several molecular markers of transcriptionally silent heterochromatin, including a compact structure visible by DNA staining, hypoacetylated histones, methylation of histone H3 on lysine 9, and the presence of heterochromatin-associated protein 1 (HP1). A number of additional proteins are known to contribute to formation and/or maintenance of the SAHF, including histone chaperones HIRA and Asf1, high-mobility group A proteins, and histone variant macroH2A [4-6].

In normal somatic cells, SAHF can be induced by a pleiotropic group of stimuli, including oncogene activation, telomere dysfunction and agents that damage DNA or alter chromatin structure. At the molecular level, the p53 and the p16-retinoblastoma protein (Rb) tumour suppressor pathways serve as critical cell-cycle checkpoints that mediate both replicative and oncogene-induced senescence. The inactivation of these two pathways typically abolishes senescence in mouse and human cells, regardless of the initial senescence trigger [7-9]. The p53 pathway exerts its effects through the activation of downstream target genes, including the cell cycle inhibitor p21, whose expression is increased in senescent cells. Alternatively, the Rb pathway is activated by upregulation of p16 and inhibits cell proliferation through numerous downstream effectors. For example, phosphorylated Rb inhibits the E2F family of transcription factors, whose target genes are necessary for cell cycle progression through the S-phase [10]. The role of Rb and p53 in SAHF induction has been investigated. Ye et al. [4] showed that, in the process of SAHF formation, recruitment of HIRA to PML bodies did not require functional Rb or p53 tumour suppressor pathways; rather, the efficient formation of SAHF was dependent on active Rb and p53 pathways in WI38 cells. Another CDK inhibitor p16, albeit not universally essential for cellular senescence, was critical for the formation of SAHF [2].

By contrast to normal somatic cells, most tumours exhibit extended or infinite life spans. However, in recent years, it has been accepted that cancer cells may also undergo senescence, and senescence induction could comprise an effective in vivo mechanism to restrain tumour progression by preventing cancer cell proliferation or by blocking the cells at risk of neoplastic transformation. In cancer cells, the presence of oncogenic mutations, chemotherapeutic drugs and oxidative stress can cause an acutely inducible, telomere-independent and stress-responsive form of cellular senescence, termed premature senescence [11]. DNA-damaging drugs, such as doxorubicin (Dox), 7-ethyl-10-hydroxycamptothecin (SN-38; the active metabolite of CPT-11) or etoposide, have been widely used for the treatment of a variety of malignancies [12, 13]. These drugs can induce cell senescence at concentrations significantly lower than those required for the induction of apoptosis. At clinically relevant concentrations, DNA-damaging drugs are able to induce irreversible cell growth arrest and senescence accompanied by upregulation of p53, p21 and p16INK4a protein.

Besides the p53/p21 and p16/Rb pathways, other pathways, such as the mitogen-activated protein kinase p38 (p38 MAPK) pathway, are also involved in DNA-damaging drug-induced senescence. Although the p38 MAPK pathway is usually linked to apoptosis, several studies have highlighted p38 MAPK signalling in growth arrest and premature senescence [14, 15]. We have previously shown that the p38 MAPK pathway is activated in the process of Dox-induced NCI-H460 cell senescence [16]. An active p38 MAPK pathway may be central to RAS-mediated senescence and tumour suppression. Among the downstream targets of the p38 MAPK pathway in premature senescence, HBP1  was found to be necessary for premature senescence by RAS-p38 MAPK [17]. HBP1 was first identified as a target of the Rb and p130 family members and was characterized as a transcriptional repressor and cell cycle inhibitor [15, 18-20].

Although there has been an increased understanding of heterochromatin biology in recent years, including SAHF formation, it is apparent that many critical questions in this respect remain to be answered. For example, what are the factors and pathways necessary in tumour cell SAHF formation? Does inactivation of p53, p16INK4a or Rb tumour suppressors (which are the most common events in human cancers) play a role? Is heterochromatin formation responsible for cellular senescence or is it a result of senescence? Can SAHF be used as a marker for tumour detection and therapy? In the present study, we show that the DNA-damaging agents Dox and SN-38 were able to induce SAHF formation in various types of tumour cells, and found that the Rb pathway was necessary for SAHF formation in tumour cells. In addition, we show that the p38 MAPK/HBP1 pathway played important roles in SAHF formation triggered by DNA-damaging agents. The findings of the present study provide new insights into the mechanisms of tumour cell SAHF formation.

Results

DNA-damaging agent-induced senescence-associated DNA foci formation in some tumour cell types

We first determined whether tumour cells are able to form SAHF-like human diploid fibroblast cells, such as IMR90 and WI38, when placed under senescence induction stress. We treated various tumour cell lines with the DNA-damaging agents Dox and SN-38 at concentrations that induce senescence for 1 + 7 days, and the formation of heterochromatin foci was observed in several tumour cell lines, including MDA-MB-231 breast cancer cells (Fig. 1A), high metastatic MDA-MB-231 cells (HM-MDA-MB-231; Fig. 1B), MDA-MB-435 breast cancer cells (Fig. 1C) and HepG2 cells (data not shown). The percentage of DNA foci-positive cells in these tumour cell lines was ~ 40% upon treatment with Dox or SN-38, whereas almost no DNA foci were detected in cells without treatment. As a positive control, SAHF formation was calculated in WI38 cells at day 8 after transduction of oncogene H-RasV12 lentiviral vector, and the proportion of SAHF positive cells reached ~ 70% (Fig. 1D). Moreover, H-RasV12 was also able to induce DNA foci formation in MDA-MB-231 cells, as visualized by DAPI staining (Fig. S1A).

Figure 1.

DNA foci accumulated in senescent tumour cells. MDA-MB-231(A), HM-MDA-MB-231(B) and MDA-MB-435 (C) cells were treated with 100 nm Dox or 100 nm SN-38 for 24 h, and then cultured in fresh medium without drugs for another 7 days, followed by DAPI staining. The DNA foci were counted and the percentage of cells with DNA foci was calculated. (D) WI38 cells were infected with control retrovirus (Con) or H-RasV12 for 8 days, before DNA foci detection.

The drug-induced senescence-associated DNA foci in tumour cells exhibited features of SAHF

We next aimed to confirm that the DNA foci in drug-treated tumour cells represent the SAHF found in normal fibroblast cells. We showed that, at the senescence-inducing concentrations of Dox (100 nm) or SN-38 (100 nm), MDA-MB-231 breast cancer cells exhibited phenotypic changes observed in senescent fibroblast cells, including intensified SA-β-gal staining (Fig. 2A) and high SA-β-gal activity (Fig. 2B). These concentrations of Dox (100 nm) or SN-38 (100 nm) did not bring about cleaved caspase 3 (Fig. S1B). For all of the cell lines used in the present study, we have carried out similar verifications and determined the proper doses of the agents that do not induce apoptosis but do induce senescence (data not shown). We further demonstrated that the molecular markers H3K9me3, HP1γ (Fig. 2C) and HP1β (data not shown), which are the regular signatures of SAHF, were co-localized with the DNA foci in senescent MDA-MB-231 and MDA-MB-435 cells, as revealed by confocal fluorescence microscopy (Fig. 2C). As a positive control, the old (senescent) WI38 cells and the H-RasV12-induced senescent WI38 cells displayed heterochromatin features equivalent to that observed in senescent MDA-MB-231 cells (Fig. 2C). In addition, MDA-MB-231 cells treated with Dox or SN-38 exhibited reduced BrdU incorporation compared to untreated cells (Fig. 2D). Accordingly, our western blotting revealed that the H3K9me3 level was significantly increased in MDA-MB-231 cells treated with Dox, although the HP1γ protein did not show apparent change (Fig. 2E). Similar results were obtained with another breast tumour cell line, MDA-MB-435. Treatment of MDA-MB-435 cells with senescence-inducing concentrations of Dox (100 nm) or SN-38 (25 nm) resulted in an increased intensity of SA-β-gal staining (Fig. S2A), high SA-β-gal activity (Fig. S2B), reduced BrdU incorporation (Fig S2C) and co-localization of H3K9me3 and HP1γ with the SAHF heterochromatin foci (Fig. 2C, middle panel). These data clearly indicate that the DNA-damaging agent-induced DNA foci observed in tumour cells possessed structural features and molecular signatures of SAHF similar to those observed in normal fibroblast cells.

Figure 2.

Drug-induced DNA foci in tumour cells formation exhibited the features of SAHF. (A, B) Dox- and SN-38-induced senescence in MDA-MB-231 cells. Cells were incubated with 100 nm Dox or 100 nm SN-38 for 24 h, subcultured to fresh medium for another 72 h, and then stained with SA-β-gal (A) or measured using flow cytometry by adding C12FDG (B). (C) Immunofluorescence microscopy of SAHF markers (H3K9me3 and HP1γ) in MDA-MB-231 and MDA-MB-435 cells after treatment with 100 nm Dox or 100 nm SN-38 for 24 h, and then cultured in fresh medium without drugs for another 7 days. Senescent WI38 cells (old or H-RasV12 induced) were used as a positive control. (D) MDA-MB-231 cells from (A) were pulse-labelled with 5′-BrdU for 24 h and then immunofluorescence stained to detect 5′-BrdU. A total of 100 cells were scored for the absence of 5′-BrdU incorporation. (E) MDA-MB-231 cells were treated with 100 nm Dox for the indicated number of days, and the protein level of H3K9me3 and HP1γ was detected by western blotting.

Some tumour cell types were unable to form SAHF when treated with DNA-damaging agents

Unlike the cells described above, we found that certain cancer cell types failed to embark upon such a heterochromatin chromatin change when induced with the DNA-damaging agents. For example, although treatment of the A549 human lung cancer cells with 200 nm Dox or 400 nm SN-38 induced cell senescence accompanied by intensified SA-β-gal staining (Fig. 3A), high SA-β-gal activity (Fig. 3B) and reduced BrdU incorporation (Fig. 3E), they were unable to form SAHF (Fig. 3G). Similarly, 100 nm Dox or 100 nm SN-38 induced senescence (Fig. 3C,D) and reduced BrdU incorporation (Fig. 3F) in human lung cancer NCI-H460 cells without the formation of SAHF (Fig. 3H). Even a prolonged presence of Dox or SN-38 for 3 weeks failed to induce A549 and NCI-H460 cells to form SAHF (data not shown). In addition, the oncogene H-RasV12 was unable to induce SAHF formation in A549 cells (Fig. S1C). Obviously, the ability of cancer cells to form SAHF upon senescence induction is cell lineage-dependent.

Figure 3.

Some types of tumour cell were unable to form SAHF. (A, B) Dox- and SN-38-induced senescence in A549 cells. Cells were incubated with 200 nm Dox or 400 nm SN-38 for 24 h, subcultured to fresh medium for another 72 h, and then stained for SA-β-Gal activity (A) or measured using flow cytometry by adding C12FDG (B). (C, D) Dox- and SN-38-induced senescence in NCI-H460 cells. Cells were incubated with 100 nm Dox or 100 nm SN-38 for 24 h, subcultured to fresh medium for another 72 h, and then stained for SA-β-Gal activity (C) or measured using flow cytometry by adding C12FDG (D). (E, F) A549 cells (E) from (A) and NCI-H460 cells (F) from (C) were pulse-labelled with 5′-BrdU for 24 h and then immunofluorescence stained to detect 5′-BrdU. A total of 100 cells were scored for the absence of 5′-BrdU incorporation. A549 (G) and NCI-H460 (H) cells were stained with DAPI and the percentage of cells with DAPI foci was calculated after treatment with senescent doses of Dox or SN-38 for 1 + 7 days.

The Rb pathway was necessary for SAHF formation in tumour cells

Next, we aimed to identify the intrinsic factor(s) or cell lineage background that determines the capacity of a particular cancer cell line to form SAHF under senescence induction conditions. On testing various cancer cell lines, we noted that the Rb deficient MDA-MB-468 breast cancer cells were unable to form SAHF (Fig. 4D) when induced by100 nm Dox or 100 nm SN-38, although they showed intensified SA-β-gal staining (Fig. 4A), relative high SA-β-gal activity (Fig. 4B) and reduced BrdU incorporation (Fig. 4C). In addition, in another Rb deficient breast cancer cell line, BT549, a wide range of doses of DNA-damaging agents (10–500 nm Dox or 10–200 nm SN-38) failed to trigger cell senescence (Fig. 4E) and SAHF formation (Fig. 4F).

Figure 4.

Tumour cells with deficient Rb failed to form SAHF. (A, B) Dox- and SN-38-induced senescence in MDA-MB-468 cells. Cells were incubated with 100 nm Dox or 100 nm SN-38 for 24 h, subcultured to fresh medium for another 72 h, and then stained for SA-β-Gal activity (A) or measured using flow cytometry by adding C12FDG (B). (C) MDA-MB-468 cells from (A) were pulse-labelled with 5′-BrdU for 24 h and then immunofluorescence stained to detect 5′-BrdU. A total of 100 cells were scored for the absence of 5′-BrdU incorporation. (D) MDA-MB-468 cells were stained with DAPI and the percentage of cells with DAPI foci was determined after treatment with Dox or SN-38 for 1 + 7 days. (E) Dox and SN-38 could not induce senescence in BT549 cells. Cells were incubated with 100 nm Dox or 20 nm SN-38 for 24 h, subcultured to fresh medium for another 96 h, and then stained for SA-β-Gal activity. (F) A549 cells from (E) were pulse-labelled with 5′-BrdU for 24 h and then immunofluorescence stained to detect 5′-BrdU. A total of 100 cells were scored for the absence of 5′-BrdU incorporation.

These results implicate the role of the Rb pathway in tumour cell SAHF formation. Further experiments were carried out to investigate whether Rb pathway is necessary for SAHF formation in DNA-damaging agent-induced senescent tumour cells. We first treated breast cancer MDA-MB-231 cells with 100 nm Dox or SN-38 for the indicated number of days, and found that Dox or SN-38 treatment resulted in an increased level of unphosphorylated Rb, the active form of Rb protein (Fig. 5A). Similarly, in another SAHF forming cell line, MDA-MB-435, Dox and SN-38 treatment also increased the Rb level (Fig. 5B). At the same time, the phospho-Rb (pSer807/811) level was found to decrease with drug treatment (Fig. S3A,B). These results coincide well with those obtained in WI-38 cells. In old or H-RasV12-induced senescent WI-38 cells, the Rb level was upregulated (Fig. S3C). By contrast, in A549 and NCI-H460 cells that are unable to form SAHF, although p21 was significantly upregulated upon drug treatment, the Rb protein level remained unchanged (Fig. 5C,D). To further address the role of Rb, we transfected the human papilloma virus-16 E7 (HPV-16 E7) to simultaneously inactivate all the three Rb pocket proteins in human breast cancer MDA-MB-231 cells. Notably, the transfection of HPV-16 E7 potently blocked SAHF formation induced by Dox in MDA-MB-231 cells (Fig. 5E). Furthermore, knockdown of Rb by a specific small interfering RNA (siRNA) prevented SAHF formation in response to Dox (Fig. 5F). The efficiency of Rb siRNA was verified by western blotting (Fig. S4A). Thus, these data demonstrate that the Rb pathway was necessary for SAHF formation induced by DNA-damaging agents in tumour cells.

Figure 5.

The Rb pathway was necessary for tumour cell SAHF formation. Western blots of the indicated protein levels in MDA-MB-231 (A), MDA-MB-435 (B) and NCI-H460 (D) cells after treatment with 100 nm Dox or SN-38 for the indicated number of days. (C) Western blots of the indicated protein levels from A549 cells after treatment with 200 nm Dox or 400 nm SN-38 for the indicated number of days. (E) MDA-MB-231 cells were infected with control (Con) or HPV-16 E7 plasmid, and then treated with 100 nm Dox for 1 + 7 days, and the percentage of cells with DAPI foci was counted. (F) MDA-MB-231 cells were infected with control interfering RNA (Con) or Rb siRNA, and then treated with 100 nm Dox for 1 + 7 days, and the percentage of cells with DAPI foci was counted.

The p38 MAPK pathway was required in DNA-damaging agent-induced SAHF formation in tumour cells

The p16 protein participates in the Rb pathway by inhibiting cyclin D-dependent kinases that would otherwise phosphorylate and inactivate Rb [21]. Because MDA-MB-231 cells are p16 deficient, we aimed to determine the signalling involved in activating the Rb pathway in the process of SAHF formation in tumour cells. It has been suggested that p38 MAPK converts a variety of stress stimuli into a common senescence signal [22]. Therefore, we investigated the role of the p38 MAPK pathway in tumour cell SAHF formation in response to DNA-damaging agents. Using a similar strategy to that employed for Rb, we showed that Dox or SN-38 treatment of breast cancer MDA-MB-231 cells significantly increased the active phosphorylated p38 protein level (Fig. 6A). As expected, in old or H-RasV12-induced senescent WI-38 cells, the phosphorylated p38 protein level was also upregulated (Fig. 6B). Additionally, we treated MDA-MB-231 cells with SB203580, which specifically inhibits the p38 MAPK pathway but not extracellular signal-regulated kinase (ERK) and Jun N-terminal protein kinase (JNK), and found that inhibition of p38 MAPK effectively blocked the formation of SAHF induced by Dox (Fig. 6C). The suppression efficiency of SB203580 was confirmed before use (Fig. S4B). Moreover, we also tested whether the other two MAPK pathways, ERK and JNK, were involved in the process of SAHF formation, and the results showed that p44/42 and phospho-JNK protein levels were not changed when treated with Dox or SN-38 (Fig. 6D). These results indicate that the p38 MAPK pathway was required in DNA-damaging agent-induced SAHF formation in tumour cells.

Figure 6.

p38 MAPK pathway was required in tumour cell SAHF formation. (A, D) Western blots of the indicated protein levels from MDA-MB-231 cells treated with 100 nm Dox or SN-38 for the indicated days. (B) Western blots of the indicated protein levels from young and senescent WI38 cells. (C) MDA-MB-231 cells were treated with or without p38 MAPK pathway inhibitor SB203580 (10 μm) for 2 h, then 100 nm Dox was added for 1 + 7 days. The percentage of cells with DAPI foci was determined.

HBP1 was involved in SAHF formation

HBP1 is a downstream target of the p38 MAPK pathway, and has been implicated in mediating senescence signals. We then investigated the role of HBP1 in SAHF formation. We showed that both HBP1 mRNA and protein levels were upregulated in MDA-MB-231 cells treated with Dox (Fig. 7A). Inhibition of p38 MAPK by SB203580 effectively downregulated HBP1 expression induced by Dox (Fig. 7B). In addition, senescent (old) WI38 cells expressed a relatively higher HBP1 protein (Fig. S4C). By contrast, in A549 cells, which are unable to form SAHF, although the p38 MAPK pathway was activated, the HBP1 protein level remained constant when they were treated with Dox (Fig. 7C). Similarly, H-RasV12-transfected senescent MDA-MB-231 and WI38 cells exhibited high HBP1 protein expression compared to that transfected with control vector, although the HBP1 expression in A549 cells did not change after H-RasV12 transfection (Fig. 7D). Furthermore, suppression of HBP1 expression by RNAi (Fig. S4D) efficiently blocked the formation of SAHF induced by Dox (Fig. 7E). These data clearly demonstrate that HBP1 was involved in SAHF formation.

Figure 7.

HBP1 was involved in SAHF formation. (A) The mRNA and protein levels of HBP1 were detected by real-time RT-PCR and western blotting. (B) MDA-MB-231 cells were treated with or without p38 MAPK pathway inhibitor SB203580 for 2 h, and then 100 nm Dox was added for 48 h. The mRNA level of HBP1 was detected by real-time RT-PCR. (C) The expression of the indicated protein was determined by western blotting in A549 cells after 200 nm Dox treatment for the indicated number of days. (D) The indicated protein levels from MDA-MB-231, A549 and WI38 cells after infection with control (Con) or H-RasV12 retrovirus for 8 days. (E) MDA-MB-231 cells were infected with control (Con) or HBP1 short hairpin RNA plasmid, then treated with 100 nm Dox for 1 + 7 days, and the percentage of cells with DAPI foci was counted. (F) Rb and HBP1 existed in the same complex. MDA-MB-231 cells were treated with 100 nm Dox for 1 + 3 days, whole-cell extracts were prepared and immunoprecipitated with anti-IgG, anti-HBP1 or anti-Rb antibodies, and incubated with protein A agarose beads. Immunoprecipitates were subjected to immunoblotting analysis with anti-Rb and anti-HBP1 sera. Western blots of the whole-cell extracts were detected with anti-Rb and anti-HBP1 sera (Input). (G) HBP1 regulated the incorporation of Rb on the target genes of E2F. MDA-MB-231 cells were transfected with HBP1 siRNA vector for 48 h, then treated with or without 100 nm Dox for another 24 h and harvested for ChIP assays. Samples were immunoprecipitated with anti-Rb serum, and the precipitated DNA fragments were amplified using real-time PCR. Each pair of primer was used to amplify the DNA fragments in the immunoprecipitated samples or the non-immunoprecipitated (input) samples; the ratio was further normalized to the levels in cells transduced with empty vector and without drug treatment.

Previous studies have demonstrated that HBP1 transcriptional repressor is a target of the Rb family in differentiated cells [18]. HBP1 can block muscle cell differentiation via binding to Rb and/or p130 to inhibit MyoD family members [20]. It is therefore rational to speculate that HBP1 may also participate in mediating SAHF assembly by incorporating with Rb. We first used chromatin immunopreciptation assays to assess the relationship between HBP1 and Rb under Dox induction. We found that HBP1 and Rb proteins existed in the same complex (Fig. 7F). Chromatin immunopreciptation (ChIP) assays showed that HBP1 was recruited to the promoter of E2F1 targets when treated with 100 nm Dox (Fig. S5A). In addition, HBP1 siRNA weakened the incorporation of Rb onto E2F target genes, such as CCNA2, MCM3 and PCNA (Fig. 7G). These results indicate that HBP1 cooperated with Rb to regulate the binding of Rb onto target genes of E2F.

Discussion

Significantly, the present study establishes that, similar to normal human fibroblast cells, the DNA-damaging agent-induced senescent tumour cells are also able to form SAHF. However, SAHF formation is not a general phenomenon among tumour cells; rather, its occurrence is largely cell lineage-dependent. In an attempt to investigate the intrinsic events that distinguish this differential potentiality of tumour cells, we have concluded that the active Rb pathway is necessary because, among the tumour cells tested, we failed to induce SAHF formation in both cell lines where the Rb pathway cannot be activated (Fig. 3) and Rb deficient tumour cell lines (Fig. 4). We also found that the p38 MAPK pathway played a critical role in DNA-damaging agent-induced SAHF formation (Fig. 6). Furthermore, we determined that HBP1, a downstream target of the p38 MAPK pathway, played an indispensable part in tumour cell SAHF formation (Fig. 7), probably through mediating the incorporation of Rb protein onto the promoters of the E2F target genes that are key regulators of cell cycle progression and cell proliferation. Apparently, these data provided the first insights into the mechanisms that control the formation of SAHF in tumour cells.

Until recently, a commonly accepted assumption was that cancer cells, unlike normal cells, cannot undergo senescence. This paradigm has now been overset by recent studies demonstrating that not only genetic manipulation, but also radiation or drug treatment can induce the cessation of cancer cell proliferation, leading to a senescence-like phenotype [23, 24]. SAHF were first explicitly described by Narita et al. [2] who discovered that, when stained with DAPI, normal human cells exhibited a relatively even, diffuse distribution of DNA throughout the nucleus. Normally, in DAPI-stained senescent human cells, SAHF appear as ~ 30–50 bright, punctuate DNA foci, and these foci are more resistant to nuclease digestion than chromatin from growing cells [2]. Of normal human cells, WI38 and IMR90 fibroblasts and primary human melanocytes form pronounced SAHF, whereas BJ fibroblasts form less marked SAHF [25]. Nevertheless, senescent mouse cells have not been shown to accumulate domains of facultative heterochromatin as pronounced as the punctuate SAHF observed in human cells, although some mouse cells exhibit a general increase in the amount of nuclear heterochromatin, as judged by histone modification [26]. Although SAHF formation has been described as a widespread phenomenon in many senescent normal human cells, whether SAHF is a common feature of senescent tumour cells remains unanswered. Based on studies using the haematopoietic system, Braig et al. [26] proposed that heterochromatin formation may be a tumour-suppressive mechanism. Whether SAHF could be used as a reliable indicator for tumour suppression requires more extensive study. The results obtained in the present study suggest a discriminating and cell lineage-dependent occurrence of SAHF among different tumour cell types. Typically, the percentage of SAHF formation in tumour cells is lower than that in normal cells, even under prolonged treatment (Fig. 1). We consider that tumour cells escape the apoptosis and senescence programmes more readily and are less sensitive to such senescence inducers. In normal human fibroblasts, SAHF formation induced by different stimuli is cell-type-restricted. For example, H-RasV12 induces SAHF formation in MRC5, BJ and HEKn cells, whereas etoposide and Dox could not induce SAHF formation in BJ and HEKn cells [27]. Similarly, our data revealed that the ratio of SAHF formation triggered by expressing the oncogene Ras in human breast cancer MDA-MB-231 cells was higher than that induced by DNA-damaging agents (Fig. 1A and Fig. S1A), indicating a stronger potent of oncogene in induction of SAHF than DNA-damaging agents. Nevertheless, Ras still could not induce SAHF formation in tumour cells that were not inducible by DNA-damaging agents (e.g. the A549 cells) (Fig. S1B). Presumably, this reflects a variable sensitivity of tumour cells to DNA-damaging agents in that tumour cells that are more sensitive to drugs for SAHF formation may better maintain the senescence status and hence be more stably suppressed for growing. By contrast, tumour cells that lack the ability to form SAHF under DNA-damaging stress may be less stable and may re-enter the cell cycle leading to tumour relapse. The results of the present study indicate that heterochromatin formation is a result of senescence but is not responsible for cellular senescence, at least in tumour cells.

The observation in the present study of the discriminating ability of tumour cells with respect to SAHF formation prompted us to investigate the nature of this distinction. To date, the molecular mechanisms underlying the formation of SAHF have not been fully understood. The formation of SAHF is a multistep process involving many factors and events. As important regulators for cellular senescence, the roles of Rb and p53 in SAHF formation have been valued in normal human fibroblast cells. Moreover, HIRA and ASF1 histone chaperones cooperate with the Rb and/or p53 tumour suppressor proteins to drive the formation of SAHF, acting in parallel, interdependent pathways. Translocation of HIRA to the PML nuclear bodies, a presumed upstream step in the SAHF assembly process, occurs in the absence of functional Rb and p53. By contrast, a downstream step, the formation of SAHF by the HIRA/ASF1a histone chaperon complex, requires the activity of both Rb and p53 [4]. Furthermore, Rb recruits the histone methylase SUVAR39H to chromatin to irreversibly silence the expression of E2F target genes such as cyclin A [26]. On the other hand, the expression of a dominate negative form of p53 that abolishes p21 activation does not affect RasV12-induced SAHF formation, suggesting that the p53-p21 pathway is not indispensable in SAHF formation [2]. Also, shRNA-mediated knockdown of p400, an SWI/SNF family of chromatin remodelling protein, results in p53- and p21-dependent (but Rb-independent) premature senescence and SAHF formation in IMR-90 fibroblasts [28]. Taken together, these data indicate that SAHF can be formed through the p53-p21 pathway independent of Rb. Furthermore, the Asf1a-induced SAHF formation is efficiently blocked by an SV40 large T antigen mutant that inactivates p53 but not Rb [4]. p16 is well known as a vital regulator of Rb. It modulates the Rb pathway by inhibiting cyclin D-dependent kinases that would otherwise phosphorylate and inactivate Rb. Ectopic overexpression of p16 induces SAHF formation in certain cellular contexts. It was implicated that the RasV12-induced SAHF formation was significantly impaired in IMR-90 human fibroblasts that expressed short hairpin RNAs against Rb or p16 [2]. Therefore, the question arises as to whether these factors are also necessary for tumour cell SAHF formation. We have demonsrated that the human breast cancer cell MDA-MB-231 is among those that can be induced to form SAHF. However, the gene for p16 is deficient in MDA-MB-231 cells, which implies that p16 is not necessary in tumour cell SAHF formation. In addition, MDA-MB-231 is also a p53 mutant cell line, which suggests that the presence of wild-type p53 is not a prerequisite for SAHF formation in tumour cells. In the present study, by investigating the roles of the Rb pathway in SAHF formation in MDA-MB-231 cells, we found that the Rb pathway was activated during SAHF formation triggered either by DNA-damaging agents or by Ras oncogene. By contrast, in tumour cells that could not form SAHF, such as A549 and NCI-H460, the Rb pathway was not activated by DNA-damaging agents, although they induced cellular senescence (Fig. 3 and 5). Similarly, in Rb-deficient tumour cells, such as MDA-MB-468 and BT549, DNA-damaging agents also failed to initiate SAHF formation (Fig. 4). Thus, our results further highlight the critical role of Rb in SAHF formation of tumour cells, whereas p53 and p16 were not necessary.

Besides the Rb and p53 suppressor pathways, a number of other pathways are also implicated in the process of SAHF formation. For example, the repression of Wnt2 occurs early in senescence independent of the Rb and p53 tumour suppressor proteins, and drives the re-localization of HIRA to PML nuclear bodies, followed by the onset of senescence and the formation of SAHF, likely through GSK3β-mediated phosphorylation of HIRA [29]. Martin et al. [30] showed that changes in the organization of heterochromatin in NIH3T3 cells contributed to the cell cycle arrest induced by prolonged activation of the Raf/ERK pathway. Accordingly, we investigated which pathway is needed to activate Rb pathway in the p16-deficient MDA-MB-231 cell line. We focused on the p38 MAPK pathway because it has been suggested to convert a variety of stress stimuli into a common senescence signal. The results obtained confirmed that the p38 MAPK pathway was required in Dox-driven SAHF formation in MDA-MB-231 cells (Fig. 6C). Meanwhile, the other two MAPK pathways, ERK and JNK, were not activated in the process of SAHF formation (Fig. 6D). We further showed that inhibition of the p38 MAPK pathway reversed the enrichment of Rb on E2F1 targets in MDA-MB-231 cells (Fig. S5B), suggesting that p38 MAPK may regulate the Rb pathway in the induction of SAHF formation. To support this assumption, we focused on a downstream target of the p38 pathway, HBP1. HBP1 has been reported to mediate senescence signals by cooperating with Rb [31]. Knockdown of HBP1 by RNA interference prevents oncogene-induced senescence without affecting the p38 phosphorylation levels [17]. We showed in the present study that, during SAHF formation in MDA-MB-231 cells, both HBP1 mRNA and protein levels were upregulated, and inhibition of the p38 pathway restored HBP1 levels (Fig. 7A,B). Furthermore, knockdown of HBP1 significantly blocked SAHF formation in MDA-MB-231 treated with Dox (Fig. 7E). These results indicate that, in tumour cells, HBP1 is a responsive target of the p38 pathway and plays a pivotal role in SAHF assembly. Further investigations into the molecular action of HBP1 revealed that HBP1 may cooperate with Rb to mediate the binding of Rb onto E2F target genes in the process of SAHF assembling (Fig. 7F,G). Besides p16, HBP1 can also repress the p21 promoter and inhibit the induction of p21 expression by E2F [32]. HBP1 may also integrate into the control of other cell cycle regulators in the process of SAHF formation. A more detailed mechanistic study will be helpful with respect to unravelling the function of HBP1 in SAHF formation.

Collectively, the data reported in the present study establish that SAHF formation accompanies the cellular senescence induced by DNA-damaging agents in some tumour cell types, although this is not a common phenomenon in all tumour cell contexts. The Rb pathway is necessary for SAHF formation in tumour cells. The p38 MAPK/HBP1 pathway also plays an important role in SAHF formation. Notably, the present study evokes a potential possibility for the design of new strategies that aim to induce SAHF in tumour cells using senescence agents. If this can be realized, relatively low concentrations of DNA-damaging agents would be sufficient to maintain the stable status of senescence of tumour cells, and thus suppress tumour progression.

Materials and methods

Cell lines and drug treatment

All the cell lines (WI38, MDA-MB-231, MDA-MB-435, A549, NCI-H460, MDA-MB-468 and BT549) were obtained from the ATCC (Manassas, VA, USA) and cultured in accordance with protocols provided by the ATCC. Dox and SN-38 were purchased from Sigma (St Louis, MO, USA). To induce senescence, cells were seeded at a density of 1 × 104·cm−2 24 h before Dox or SN-38 treatment. Cells were cultured in the presence of indicated doses of Dox or SN-38 for 1 day and then cultured in fresh medium without drugs for additional days as indicated, represented as 1 + days. The p38 MAPK pathway inhibitor SB203580 (Sigma) was dissolved in dimethylsulfoxide to a stock concentration of 10 mm. MDA-MB-231 cells were treated with 10 μm SB203580 for indicated days. Control cells were treated with an equivalent volume of dimethylsulfoxide.

DAPI staining

For SAHF detection, cells were plated on cover slips. After being treated with DNA-damaging drugs for indicated days, cells were fixed in 1% formaldehyde in NaCl/Pi for 10 min at 37 °C and permeabilized with 0.2% Triton X-100 in NaCl/Pi for 10 min at 4 °C. Cells were then washed twice in NaCl/Pi and incubated with 500 nm DAPI (Invitrogen, Paisley, UK). Cells were documented using a charge-coupled device camera attached to a fluorescence microscope.

Cellular immunofluorescence

Cells were fixed in 1% formaldehyde in culture medium for 10 min at 37 °C and permeabilized with 0.2% Triton X-100 in NaCl/Pi for 10 min at 4 °C. Cells were washed twice in NaCl/Pi and blocked for 1 h with 5% BSA in NaCl/Pi, and then incubated with primary antibodies for 1 h at room temperature, washed three times in NaCl/Pi and incubated with secondary antibodies for 45 min at room temperature. Photographs were taken using a confocal microscope. The primary antibodies were rabbit anti-H3K9me3 (Millipore, Billerica, MA, USA) and mouse anti-HP1γ (Millipore).

Detection of SA-β-gal

SA-β-gal staining was performed as described previously [1]. Briefly, cells were fixed in 2% formaldehyde, 0.2% glutaraldehyde in NaCl/Pi, washed and exposed overnight at 37 °C to a solution containing 1 mg·mL−1 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside, 5 mm potassium ferrocyanide, 5 mm potassium ferricyanide, 150 mm NaCl, 2 mm MgCl2 and 0.1 m phosphate buffer (pH 6.0). Cells were observed under a fluorescence microscope. All the experiments were repeated three times, and one of the representative results is shown.

Alternatively, SA-β-gal activity was measured using flow cytometry, as described previously [33]. Briefly, living cells were incubated with a nonfluorescent substrate of SA-β-gal, C12FDG (Invitrogen) at 33 μm, and the level of fluorescent product was measured immediately by flow cytometry using cellquest Software, with 10 000 events being counted for each sample (FACSCalibur; Becton Dickinson, Franklin Lakes, NJ, USA).

BrdU incorporation assay

For DNA synthesis assay, BrdU (BD Pharmingen, San Diego, CA, USA) was added to the medium (10 nmol·mL−1) at the indicated time points and cells were cultured for an additional 24 h. Cells were then fixed, permeabilized and denatured. BrdU incorporation was measured by immunofluorescence using an anti-BrdU serum (BD Phaemingen) and the nuclei were stained with DAPI. The stained cells were visualized under a fluorescence microscope, and at least 200 cells were scored for BrdU incorporation. Each experiment was repeated at least three times, and the results are displayed as a percentage of BrdU-positive cells in histogram.

RT-PCR

For cDNA synthesis, 1 μg of total RNA was reverse transcribed using the RT-Systems supplied by Promega (Madison, WI, USA). Quantitative real-time RT-PCR was carried out on an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) and SYBR Green (Toyobo, New York, NY, USA) was used as a double-stranded DNA-specific fluorescent dye. The PCR primer sequences were HBP1: 5′-TGAAGGCTGTGATAATGAGGAAGAT-3′ (sense) and 5′-CATAGAAAGGGTGGTCCAGCTTA-3′ (antisense) [34]; β-actin: 5′-TCGTGCGTGACATTAAGGAG-3′ (sense) and 5′-ATGCCAGGGTACATGGTGGT-3′ (antisense) [16].

Western blotting

Western blotting analysis was carried out on 20 μg of whole-cell lysate using enhanced chemiluminescence (Thermo Scientific, Rockford, IL, USA). The antibodies used to probe the blots were: rabbit anti-H3K9me3 (Millipore), mouse anti-HP1γ (Millipore), rabbit anti-Rb (BD Pharmingen), rabbit anti-phospho-Rb (pSer807/811) (Sigma), rabbit anti-p21 (Santa Cruz Biotechnology, Dallas, TX, USA), rabbit anti-p38 (CST, Danvers, MA, USA), rabbit anti-phospho-p38 (Thr180/Tyr182) (CST), rabbit anti-p44/p42 (CST), rabbit anti-phospho-p44/p42 (CST), rabbit anti-SAPK/JNK (CST), rabbit anti-phospho-SAPK/JNK (CST), rabbit anti-HBP1 (Santa Cruz) and mouse anti-β-actin (Sigma).

Vectors and viral infections

The expression plasmid pcDNA3-HPV-16 E7 was kindly provided by Gonzalo de Prat-Gay (Instituto Fundación Leloir, Buenos Aires, Argentina). The lentiviral vector used was: expression vector pBabe-Puro-H-RasV12 (a gift from Scott Lowe, Cold Spring Harbor Laboratories, New York, NY, USA). The HBP1 short hairpin RNA plasmid was constructed into the pDSL-hpUGIP backbone. Packaging vectors were psPAX2 and pMD2.G. The sequences were: GATCCCCACTGTGAGTGCCACTTCTCTTCAAGAGAGAGAAGTGGCACTCACAGTTTTTTC [17]. Lentiviral gene transfer was performed as described previously [35].

siRNA transfection

The human Rb and control siRNA were purchased from Qiagen (Valencia, CA, USA). On the day before transfection, 5 × 105 MDA-MB-231 cells were plated in six-well plates. After 24 h in culture, 60 μL of 0.4 μm stock solution of siRNA duplexes was transfected into cells with the HiPerFect Transfection Reagent (Qiagen), in accordance with the manufacturer's instructions. After 24 h of incubation, the cells were treated with Dox and maintained in culture for the indicated times before analysis.

Co-immunoprecipitation assay

Total cell extracts from Dox-treated MDA-MB-231 cells were pre-cleared with salmon sperm DNA/protein A-agarose beads (Millipore). Rabbit anti-HBP1 was added for immunoprecipitation. The precipitates were then subjected to SDS/PAGE followed by transfer onto a poly(vinylidene difluoride) membrane and incubation with anti-Rb serum. Samples were detected using the Super Signal West Pico Chemiluminescent Substrate (Pierce, Rockford, IL, USA) detection method in accordance with the manufacturer's instructions.

ChIP assay

Immunoprecipitation was carried out with rabbit anti-Rb antibody (BD Pharmingen). Primer pairs for the CCNA2 promoter were 5′-TCCGAAGGCTGACTCTAAGC-3′ (sense) and 5′-TGGGCAACCCAAATGATAGT-3′ (antisense) [36]; for the MCM3 promoter were 5′-TCTTTGGCAGCGGGCAT-3′ (sense) and 5′- CGCAGCTCCACATCGTCC-3′ (antisense) [36]; and for the PCNA promoter were 5′-CTGGCTGCTGCGCGA-3′ (sense) and 5′- CACCACCCGCTTTGTGACT-3′ (antisense) [27].

Statistical analysis

Student's t-test was used to determine statistical significance (*P < 0.05; **P < 0.01). Error bars indicate the SDs.

Acknowledgements

This work was supported by grants received from the National Natural Science Foundation of China (Grant numbers 31100998, 31071149, 31170719, 91019011 and 31050015); the Programme for Introducing Talents to Universities (B07017); and the Fundamental Research Funds for the Central Universities (No. 11SSXT132).

Ancillary