Ministry of Education Key Laboratory of Experimental Teratology and Institute of Molecular Medicine and Genetics, Shandong University School of Medicine, Jinan, Shandong, China
Department of Genetics, Rutgers University, Piscataway, NJ, USA
Changshun Shao, Institute of Molecular Medicine and Genetics, Shandong University School of Medicine, Jinan, Shandong 250012, China. Tel.: +86 531 8838 2043; fax: +86 531 8838 2502; e-mail:email@example.com
Mammalian cells may undergo permanent growth arrest/senescence when they incur excessive DNA damage. As a key player during DNA damage response (DDR), p53 transactivates an array of target genes that are involved in various cellular processes including the induction of cellular senescence. Chemokine receptor CXCR2 was previously reported to mediate replicative and oncogene-induced senescence in a DDR and p53-dependent manner. Here, we report that CXCR2 is upregulated in various types of cells in response to genotoxic or oxidative stress. Unexpectedly, we found that the upregulation of CXCR2 depends on the function of p53. Like other p53 target genes such as p21, CXCR2 is transactivated by p53. We identified a p53-binding site in the CXCR2 promoter that responds to changes in p53 functional status. Thus, CXCR2 may act downstream of p53. While the senescence-associated secretory phenotype (SASP) exhibits a kinetics that is distinct from that of CXCR2 expression and does not require p53, it reinforces senescence. We further showed that the cellular senescence caused by CXCR2 upregulation is mediated by p38 activation. Our results thus demonstrate CXCR2 as a critical mediator of cellular senescence downstream of p53 in response to DNA damage.
Cellular senescence is characterized by irreversible proliferative arrest and an array of morphological and functional alterations. Cellular senescence can be caused by oxidative stress, telomere dysfunction, DNA damage and oncogene activation (d'Adda di Fagagna, 2008). Cellular senescence may contribute to the decline in organ function and to the reduction in or loss of vitality of organisms (Wiemann et al., 2002; Baker et al., 2011). Moreover, through the acquisition of senescence-associated secretory phenotype (SASP), senescent cells may also promote the growth and aggressiveness of nearby precancerous or cancer cells (Coppe et al., 2008; Kuilman et al., 2008). Nevertheless, substantial evidence showed that cellular senescence also functions as an intrinsic defense mechanism that prevents tumor progression from oncogene-transformed premalignant cells (Campisi & d'Adda di Fagagna, 2007; Acosta & Gil, 2012). Thus, unraveling the mechanism of cellular senescence is critical for the understanding of the aging process and the development of malignancy.
Cellular senescence is mainly established and maintained by the p53 and p16–pRB tumor suppressor pathways (Coppe et al., 2008). Acting as a transcription factor, p53 regulates the expression of a broad array of genes that are involved in cell cycle checkpoints, apoptosis and senescence (Junttila & Evan, 2009; Vousden & Prives, 2009). In particular, DNA damage response (DDR) leading to p53 activation provides a first line of defense against cancerous transformation by preventing the proliferation of cells with severely damaged DNA (d'Adda di Fagagna, 2008). While cellular senescence has been shown to depend on p53 in many experimental settings (Pearson et al., 2000; Schmitt et al., 2002; Chen et al., 2005; Ventura et al., 2007), only a few p53 target genes, namely p21, PAI-1, PML, and E2F7, are known to mechanistically contribute to cellular senescence (Brugarolas et al., 1995; Brown et al., 1997; Kagawa et al., 1999; Wang et al., 1999; de Stanchina et al., 2004; Kortlever et al., 2006; Aksoy et al., 2012; Carvajal et al., 2012). Plasminogen activator inhibitor-1 (PAI-1) is found to be both necessary and sufficient for the induction of replicative senescence downstream of p53 (Kunz et al., 1995; Kortlever et al., 2006). PML and E2F7 appeared to be critical for the silencing of E2F target genes (Vernier et al., 2011; Aksoy et al., 2012; Carvajal et al., 2012). While the p53 target genes so far identified have been invaluable in understanding how senescence is triggered, it remains unclear whether there are additional p53 effectors that also contribute to cellular senescence.
By screening for small RNAs that extend the lifespan of primary fibroblasts, Acosta et al. recently identified chemokine receptor CXCR2 as a key mediator of replicative and oncogene-induced senescence (Acosta et al., 2008). CXCR2 is upregulated in fibroblasts undergoing replicative and oncogene-induced senescence. CXCR2 ligands, which are subjected to regulation by transcriptional factors NF-κB and C/EBPβ, are also upregulated and reinforce cellular senescence. In the present study, we investigated the role of CXCR2 in cellular senescence induced by genotoxic insults. We found that CXCR2 was significantly upregulated in response to genotoxic or oxidative stress and played a critical role in the stress-induced senescence. In addition, we found that the CXCR2 upregulation was dependent on p53 function. p53 was shown to increase CXCR2 expression at transcriptional level by binding to CXCR2 promoter. Furthermore, we demonstrated that p38MAPK phosphorylation acted downstream CXCR2 in mediating DNA damage-induced senescence.
Cellular senescence induced by genotoxic stress is dependent on CXCR2 upregulation
CXCR2 was shown to be upregulated in replicative senescence and oncogene-induced senescence (Acosta et al., 2008). To determine the role of CXCR2 in senescence caused by acute DNA damage, we first measured the expression of CXCR2 in U2OS osteosarcoma cells in response to ionizing radiation (IR). We observed that IR caused a significant increase in the expression of CXCR2, at both mRNA and protein levels, in a dose-dependent manner (Fig. 1A). Cellular senescence, as determined by staining of SA-β-gal and BrdU incorporation, was evident at 120 h after exposure to 10 Gy of IR (Fig. 1B). To determine whether CXCR2 upregulation is required for IR-induced senescence, we applied a CXCR2 inhibitor SB225002 to U2OS cells before their exposure to IR. As shown in Fig. 1(B), application of SB225002 significantly attenuated the senescence caused by IR. It also relieved the growth arrest associated with senescence. The role of CXCR2 in promoting senescence was further confirmed by using small interfering RNAs specifically targeting CXCR2. Clearly, the alleviation of senescence by CXCR2 RNAi was dependent on the efficiency of RNAi (Fig. 1C,D). Together, these results indicate that CXCR2 can be induced by IR and that the activation of CXCR2 is required for IR-induced senescence.
Next, we tested whether CXCR2 can also be induced by other types of stress, by using agents that are known to cause replication stress (hydroxyurea and aphidicolin), or oxidative stress (H2O2), and whether DNA damage-induced CXCR2 upregulation and senescence also apply to other types of cancer cells. Like IR, hydroxyurea, aphidicolin, and H2O2, all induced CXCR2 expression in U2OS cells and caused cellular senescence (Fig. 2A and S1A). However, while IR, hydroxyurea, and aphidicolin all induced CXCR2 upregulation and senescence in Lewis lung carcinoma (LLC) cells, H2O2 appeared to be less effective (Figs 2B and S1B). Consistent with the lack of induction of CXCR2 and senescence by H2O2, whereas hydroxyurea, aphidicolin, and IR all inflicted remarkable DNA damage, no significant increase in the level of DSBs was detected in H2O2-treated LLC cells (Figs 2C and S1D), suggesting that the degree of CXCR2 induction, as well as that of cellular senescence, was correlated with the level of DNA damage.
It is important to verify whether CXCR2 upregulation also plays a role in genotoxic stress-induced senescence in normal cells. To this end, we treated normal human fibroblasts (NHF) with IR, hydroxyurea, aphidicolin, and H2O2, respectively. All the tested agents were able to induce senescence in NHF, though the degree to which CXCR2 was upregulated varied with the agents (Figs 2D and S1D). The hydroxyurea applied appeared to be least effective in inducing CXCR2. Accordingly, the percentage of SA-β-gal positive cells induced by hydroxyurea was lower than by the other three treatments. We also measured the expression of several other markers associated with senescence and found that p21, p16, and PAI-1 were all significantly increased in the senescent cells (Fig. S1E). As expected, lamin B1 showed the opposite trend. To test whether CXCR2 upregulation is required for H2O2-induced senescence in NHF, we established CXCR2 knockdown NHF cells with pGIPZ-shCXCR2 lentivirus. As shown in Fig. 2(E,F), H2O2-induced senescence was significantly attenuated in CXCR2 knockdown cells. Together, our data suggest that the CXCR2 upregulation is probably a general response to genotoxic stress.
CXCR2 induction and SASP are differentially regulated
Senescent cells are known to be capable of producing cytokines, chemokines, and other soluble factors, a phenomenon known as SASP (Coppe et al., 2008; Kuilman & Peeper, 2009). Interestingly, CXCR2 binds to some of the secreted factors, including IL-8, in the secretome of the senescent cells. Furthermore, the CXCR2 upregulation and the development of SASP were reported to be coordinated during oncogene-induced senescence (Acosta et al., 2008). To test whether CXCR2 upregulation and SASP are correlated during senescence induced by genotoxic stress, we measured the expression levels of CXCR2 and two representative SASP factors (IL-6, IL-8) in U2OS cells after they were exposed to IR. Both IL-6 and IL-8 showed a dose-dependent upregulation on day 5 (Fig. S2A). We then exposed U2OS cells to 10 Gy of IR and determined the expression levels of CXCR2, IL-6, and IL-8 at different time points postirradiation. As shown in Fig. 3(A,B), the transcript level of CXCR2 increased 10-fold at 24 h, peaked at 72 h, then dropped at 120 h, and the protein level of CXCR2 showed a similar expression pattern. In contrast, the expression levels of IL-6 and IL-8 remained unchanged at 24 h after irradiation, exhibited a modest increase at 72 h, and a sharp increase at 120 h (Fig. 3C). Therefore, the induction of CXCR2 follows a different kinetics from that of SASP. These results suggest that although SASP and the induction of CXCR2 are both associated with senescence, the onset of SASP is preceded by the upregulation of CXCR2 during senescence.
Because NF-κB plays a key role in the development of SASP, we next tested whether NF-κB also controls the transcription of CXCR2. Pretreatment of cells with BAY11-7082, an NF-κB inhibitor, significantly suppressed IR-induced expression of IL-6 and IL-8, but had no effect on IR-induced CXCR2 upregulation (Fig. 3D). Another NF-κB inhibitor CAPE similarly blocked the IR-induced expressions of IL-6 and IL-8. The CXCR2 expression, on the other hand, was enhanced by CAPE (Fig. 3E), probably due to a genotoxic effect of CAPE itself. Moreover, knockdown of p65, an NF-κB subunit, had no obvious effect on CXCR2 expression (Fig. 3F). These data indicate that while NF-κB controls the transcription of SASP factors IL-6 and IL-8, it is not required for the transcription of CXCR2. Nevertheless, NF-κB appeared to also play a critical role in the induction of senescence, as p65 knockdown could significantly attenuate senescence and growth arrest caused by IR (Fig. S2B). Together, these results suggest that CXCR2 induction and SASP are differentially regulated during IR-induced senescence.
p53 function is required for CXCR2 induction in response to genotoxic stress
Because p53 is usually activated in response to DNA damage and plays a key role in driving cellular senescence, we next asked whether the induction of CXCR2 during IR-induced senescence is controlled by p53. We observed that the MDM2 antagonist nutlin-3, which promotes the stabilization of p53, significantly increased the transcription level of CXCR2 in U2OS cells (Fig. 4A). In accordance with the elevated expression of CXCR2 in nutlin-3 treated cells, SA-β-gal positive cells were significantly induced in U2OS cells and NHF cells (Fig. S3A). Interestingly, the kinetics of CXCR2 transcription is analogous to that of p21, a well-established p53 target (Fig. S3B,C). We next tested the role of p53 in inducing CXCR2 expression by using p53-deficient osteosarcoma line Saos-2. While treatment of 10 Gy of IR triggered a remarkable increase in CXCR2 expression in U2OS cells, it had no noticeable effect on the expression of CXCR2 in Saos-2 cells (Fig. 4B). Correspondingly, whereas 70% of U2OS cells displayed positive staining for SA-β-gal, only 20% of Saos-2 cells displayed such staining after exposure to 10 Gy of IR (Fig. S3D). Because U2OS and Saos-2 lines are not isogenic, we next knocked down p53 in U2OS cells with a pLKO1-Puro-shp53 lentiviral vector and evaluated the expression of CXCR2 in U2OS-shp53 cells in response to doxorubicin. As shown in Fig. 4(C), CXCR2 upregulation caused by doxorubicin was significantly attenuated in U2OS-shp53 cells when compared to U2OS-shNeg cells, while the upregulation of IL-6, one of the SASP factors, was more pronounced in U2OS-shp53 cells. We next tested whether CXCR2 upregulation also depends on p53 in other cancer cells. We knocked down p53 in MCF7 breast cancer cells by using lentivirus and exposed them to doxorubicin. We found that the CXCR2 upregulation was significantly attenuated in MCF7-shp53 cells when compared to MCF7-shNeg cells (Fig. 4D).
We then validated the role of p53-CXCR2 cascade in stress-induced senescence in normal human cells. In NHF-shp53 cells in which p53 was knocked down, the CXCR2 upregulation induced by H2O2 was again significantly attenuated (Fig. 4E). Accordingly, H2O2-induced senescence was significantly reduced in NHF-shp53 cells when compared to NHF-shNeg (Fig. S3F). Measurement of Cxcr2 expression in mouse skin fibroblasts (MSFs) further confirmed the role of p53 in upregulating CXCR2. While 10 Gy of IR remarkably increased Cxcr2 and Mdm2 expressions in p53 wild-type MSFs, it hardly changed Cxcr2 and Mdm2 expressions in p53−/− MSFs (Fig. 4F). It should be noted that while the basal transcript levels of p21 and MDM2 were reduced in p53 knockdown or deficient cells, the CXCR2 transcription was not decreased in p53 knockdown or deficient cells (Fig. 4B–F). Interestingly, basal transcription of E2F7, a recently identified p53 target gene, was also independent of p53 function (Carvajal et al., 2012). Together, our data suggest that genotoxic stress-induced CXCR2 upregulation is p53-dependent.
CXCR2 is a transcriptional target of p53
The results shown above suggest that CXCR2 may act as a transcriptional target of p53. The p53 target genes usually contain defined p53 responsive elements (REs), which physically bind to p53 protein in supporting transactivation (Riley et al., 2008). To explore if CXCR2 gene contains p53 RE, we analyzed the human CXCR2 promoter sequence by a p53Scan program (http://www.ncmls.eu/bioinfo/p53scan/), which employs an algorithm that identifies consensus p53-binding motif. The analysis revealed a putative p53RE in the CXCR2 promoter, from −80 to −61, and this putative p53RE has a high degree of homology with the canonical p53RE (Fig. 5A).
To test whether this sequence is functional in mediating p53-dependent upregulation of CXCR2 expression, we cloned two fragments, 1412 bp and 958 bp, respectively, in the CXCR2 promoter region into the pGL3 luciferase expression vector. The 1412-bp fragment, containing sequence from −1075 to 336, included the putative p53RE, while the 958-bp fragment, containing sequence from −1075 to −118, lacked the putative p53RE (Fig. 5A). We observed that nutlin-3 treatment increased the luciferase activity of the CXCR2-LUC-1412 reporter in a dose-dependent manner, while PFT-α treatment had an opposite effect (Fig. 5B,C). IR also increased CXCR2-LUC-1412 luciferase reporter activity (Fig. 5D). Furthermore, IR-induced CXCR2-LUC-1412 luciferase activity only in p53 wild-type U2OS, but not in p53-null Saos-2 (Fig. 5E). The dependence of the CXCR2-LUC-1412 luciferase activity on p53 function was further validated in U2OS-shp53 and U2OS-shNeg cells. The reporter activity could be induced by IR in U2OS-shNeg cells, but not in U2OS-shp53 cells (Fig. 5F). In addition, the basal CXCR2-LUC-1412 reporter activity was much lower in U2OS-shp53 cells than in U2OS-shNeg cells (Fig. 5F). These results demonstrated that the CXCR2 promoter activity is p53-dependent. Moreover, while irradiation increased the activity of CXCR2-LUC-1412 reporter by > 2.5-fold, it had no significant effect on that of CXCR2-LUC-958 reporter (Fig. 5G). More strikingly, whereas nutlin-3 treatment increased the CXCR2-LUC-1412 activity by > 30-fold, it only increased the CXCR2-LUC-958 activity by 5-fold (Fig. 5H). We further tested the function of the putative p53RE in CXCR2 promoter by substituting some of the nucleotides in the luciferase reporter. CXCR2-LUC-205-WT contains part of the CXCR2 promoter that encompasses the putative p53RE (Fig. 5I). CXCR2-LUC-205-MT carries the substitutions in the putative p53RE. As shown in Fig. 5(J), only CXCR2-LUC-205-WT, but not CXCR2-LUC-205-MT, showed a strong response to nutlin-3 treatment. These results suggest that the p53RE in CXCR2 promoter is critical in mediating the transactivation by p53.
To determine whether p53 can physically bind to the CXCR2 promoter, we performed chromatin immunoprecipitation (ChIP) assay using a p53 polyclonal antibody to immunoprecipitate the p53-DNA complexes. CXCR2 promoter sequence encompassing the putative p53RE was detected in ChIP assay from U2OS cells treated with nutlin-3. As a positive control, p21 promoter sequence encompassing p53RE was also detected in ChIP assay from nutlin-3-treated U2OS cells. In contrast, the corresponding sequences could not be detected from nutlin-3-treated Saos-2 cells (Fig. 5K). Taken together, these results showed that CXCR2 transcription is directly activated by p53.
G protein-coupled receptor (GPCR) signaling, such as the one that is initiated from CXCR4 activation, can lead to p38 phosphorylation (Sun et al., 2002). We therefore speculated that induction of senescence associated with CXCR2 upregulation may be mediated by the p38 signaling pathway. Consistent with the previous report that p38 is critical in IR-induced senescence (Hong et al., 2010), we were able to confirm the role of p38 in this process. When a p38-specific inhibitor, SB203580, was applied to cells exposed to IR, senescence and growth arrest were significantly attenuated (Fig. 6A). The role of p38 during this process was further tested by using small interfering RNA specifically targeting p38 (Fig. 6B). As expected, senescence and growth arrest in cells treated by IR were similarly reduced by p38 knockdown (Fig. 6C).
We next tested whether the level of p38 phosphorylation corresponds to CXCR2 function. Phosphorylation of p38 was significantly increased in U2OS cells exposed to IR (Fig. 6D). Importantly, when cells were treated with CXCR2 inhibitor SB225002, the p38 phosphorylation induced by IR was greatly attenuated (Fig. 6E). Furthermore, when CXCR2 was knocked down by RNAi, p38 phosphorylation was also reduced (Fig. 6E). To complement the findings made with CXCR2 inhibitor and CXCR2 siRNAs, we constructed CXCR2 expression vector pcDNA3.1-CXCR2 (Fig. 6F) and transfected U2OS cells. Ectopic CXCR2 expression in U2OS cells led to an increased phosphorylation of p38, but had no effect on p53 protein level (Fig. 6F). These data suggest that CXCR2 acts upstream of p38 activation.
Because p53 acts upstream of CXCR2, as shown above, it is expected that p53 activation may also lead to p38 activation. Indeed, nutlin-3 treatment could increase p38 phosphorylation, while p53 inhibitor PFT-α attenuated the doxorubicin-induced p38 phosphorylation (Fig. 6G). p38 phosphorylation caused by doxorubicin was significantly attenuated in U2OS-shp53 cells when compared to U2OS-shNeg cells (Fig. 6H). These data indicate that p53 may contribute to the activation of p38, as reported previously (Kwon et al., 2002; Bragado et al., 2007). Together, our results support a notion that p53, CXCR2, and p38 form an axis in leading to cellular senescence in response to DNA damage.
It was previously reported that CXCR2 can mediate oncogene-induced senescence and replicative senescence (Acosta et al., 2008). We showed here that CXCR2 is upregulated in response to various types of genotoxic stress and plays a critical role in senescence induced by ionizing radiation and H2O2. Therefore, CXCR2 probably plays a general role in mediating cellular senescence. Importantly, we demonstrated that the upregulation of CXCR2 in response to DNA damage depends on p53. While p53 activation is well established to cause senescence in some cell types, the p53 transcriptional targets that mediate cellular senescence remain to be completely identified. The known p53 target genes involved in senescence include p21 (Brugarolas et al., 1995), promyelocytic leukemia gene (PML) (de Stanchina et al., 2004), plasminogen activator inhibitor 1 (PAI-1) (Kortlever et al., 2006) and E2F7 (Aksoy et al., 2012; Carvajal et al., 2012). Thus, CXCR2 may represent an additional p53 transcriptional target in mediating cellular senescence.
We also showed that the cellular senescence caused by the upregulation of CXCR2 appears to be mediated by p38 signaling pathway. p38 mitogen-activated protein kinases can be activated by multiple types of stress. p38 was shown to be essential for oncogene-induced senescence (Wang et al., 2002; Sun et al., 2007) and IR-induced senescence (Hong et al., 2010; Freund et al., 2011). In accordance with an earlier report (Hong et al., 2010), we also demonstrated that p38 kinase-specific inhibitor SB203580 or p38 knockdown significantly attenuated IR-induced senescence. Importantly, we showed that genetic depletion or pharmacological inhibition of CXCR2 could inhibit the phosphorylation of p38, whereas ectopic expression of CXCR2 could increase the p38 phosphorylation in U2OS cells. Taken together, our data suggest that the senescence induced by CXCR2 upregulation is mediated by p38. Thus, DNA damage-induced senescence involves a novel p53-CXCR2-p38 pathway (Fig. 6I).
The demonstration of a p53-CXCR2-p38 cascade in inducing cellular senescence in response to DNA damage is in contrast to the notion that p53 acts downstream of CXCR2 in inducing cellular senescence (Acosta et al., 2008). As shown in Fig. 6(F), although ectopic expression of CXCR2 could lead to increased p38 phosphorylation, it failed to activate p53. Further work is apparently needed to resolve the discrepancy between the two studies. Examination of the proteins that interact with CXCR2 may provide insights into the detailed mechanism by which DNA damage leads to cellular senescence along this p53-CXCR2 cascade.
Our findings have implications in our understanding of the role of CXCR2 in carcinogenesis and cancer therapy. On the one hand, it is reasonable to expect that cells that fail to upregulate CXCR2 may be impaired in cellular senescence and thus are more likely to proliferate and turn malignant. Consistent with this notion, Acosta et al. reported that CXCR2 was upregulated in preneoplastic lesions and was accompanied by senescence, whereas CXCR2 mutation or downregulation was observed in some types of cancer (Acosta et al., 2008). On the other hand, CXCR2 expression was also reported to be associated with poor prognosis in lung adenocarcinoma (Saintigny et al., 2013) and in ovarian cancer (Yang et al., 2010). In addition, CXCR2 expression in inflammatory cells contributes to the formation of tumor microenvironment that promotes tumor progression and eliminating CXCR2 function breaks the inflammatory microenvironment and thus suppresses the tumorigenesis (Jamieson et al., 2012). Senescence-associated cytokines secreted by senescent tumor cells are probably responsible for the survival of tumor cells after chemotherapy. Furthermore, contrary to the conventional belief that p53 activation may lead to fast tumor cell clearance, a recent study showed that cellular senescence mediated by wild-type p53 actually confers chemoresistance by impairing the apoptosis response (Jackson et al., 2012). Thus, CXCR2-mediated cellular senescence may have different consequences in carcinogenesis and cancer therapy depending on contexts.
Senescent cells usually produce an array of cytokines, chemokines, and other soluble factors, a phenotype that is believed to reinforce senescence (Kuilman et al., 2008). As shown in this study, and in accordance with previous reports, SASP is dependent on NF-κB (Acosta et al., 2008; Chien et al., 2011; Freund et al., 2011). Interestingly, one of the ligands for CXCR2, IL-8, was remarkably elevated during senescence. However, we found that the expression of IL-8 and other cytokines and the induction of CXCR2 followed very different kinetics. More importantly, the transcription of CXCR2 depends on p53, but not on NF-κB. These results suggest that SASP and CXCR2 expressions are not coupled. Nevertheless, NF-κB is also required in the senescence induced by IR. Therefore, some of the SASP factors, such as IL-8, probably contribute to the reinforcement of senescence. Therefore, although SASP upregulation and CXCR2 upregulation are not coupled at transcription level during senescence, they may act in concert in driving senescence.
In summary, our study shows that CXCR2 can be upregulated by genotoxic stimuli and that CXCR2 acts downstream of p53 in inducing cellular senescence. Our findings enrich our understanding of how p53 activation leads to cellular senescence in response to DNA damage.
Cell culture and treatment of cells
Human osteosarcoma cell lines U2OS and Saos-2 were obtained from the American Type Culture Collection (ATCC). LLC and MCF7 cells were obtained from Shanghai Cell Bank, Chinese Academy of Sciences. Normal human fibroblasts were established from circumcized foreskin of a juvenile. The tissue collection was approved by the ethics committee of Shandong University Medical School, and informed consent was obtained from patient's parents. U2OS, Saos-2, MCF7, and NHF cells were cultured as described by (Liu et al., 2009) (Liu et al., 2011; Xu et al., 2011). Primary mouse skin fibroblasts (MSF) were prepared from C57/BL mice, as described by (Shao et al., 1999). Cells were irradiated using a cabinet X-ray system (Faxitron X-ray Corp., Wheeling, IL, USA) at a dose rate of 0.4 Gy min−1. Inhibitors or activators were added 2 h before radiation. SB225002 (CXCR2 inhibitor) and CAPE (NF-κB inhibitor) were acquired from Cayman (Ann Arbor, MI, USA). SB203580 (p38 inhibitor) was acquired from Sigma-Aldrich (Shanghai, China). BAY 11-7082 (NF-κB inhibitor), pifithrin-α (p53 inhibitor), and nutlin-3 (p53 activator) were acquired from Beyotime (Haimen, China).
Total RNA was extracted from cultured cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and was reverse-transcribed using reverse transcriptase (Thermo Scientific, Beijing, China). Quantitative real-time PCR analysis was performed on an ABI prism 7500 (Applied Biosystems, Foster City, CA, USA) using SYBR GREEN mix (TOYOBO, Osaka, Japan). Primers used for quantitative RT-PCR were listed in the Data S1.
Chromatin immunoprecipitation assays were performed essentially as described (Hu et al., 2012). The antibodies used in the ChIP assays were p53 antibody (sc-126, Santa Cruz Biotechnology, Santa Cruz, CA, USA) and mouse IgG antibody (sc-2025, Santa Cruz). The primers for detecting the putative p53 DNA-binding sequence in CXCR2 promotor and p21 promotor were listed in the Data S1.
CXCR2 reporter construction and luciferase reporter assay
CXCR2 reporter construction and luciferase reporter assay were performed as described in detail in the Data S1.
Stable and transient transfections
Transfections of U2OS cells were performed with Lipofectamine 2000(Invitrogen) as described in detail in the Data S1.
Preparation of lentiviral vectors and infection
Lentivirus sh-p53 (shp53-pLKO.1-puro, addgene) and shCXCR2 (pGIPZ-shCXCR2-puro, Thermo Scientific, Shanghai, China) were prepared in HEK293T cells packaged by pMD2G and psPAX2. For stable infection, 5 × 104 cells were plated in each well of the six-well plates along with 2 mL of medium without antibiotics. After overnight incubation, the medium was removed and replaced with 1 mL per well of medium containing lentivirus and 8 μg mL−1 polybrene. 24 h later, fresh medium containing 2 μg mL−1 puromycin was added to each well for stable infection.
Western blot were performed as described previously (Liu et al., 2009). The membranes were probed with the following primary antibodies: anti-CXCR2 antibody (555932, BD), anti-p53 antibody (sc-126, Santa Cruz), anti-β-actin antibody (sc-69879, Santa Cruz), anti-phospho-p65 antibody (3033S; Cell Signaling, Danvers, MA, USA), anti-p38 antibody (9212, Cell Signaling), and anti-phospho-p38 antibody (9215, Cell Signaling).
BrdU incorporation assay
BrdU incorporation assay was carried out 5 days posttreatment as described (Liu et al., 2009).
Cells were previously treated with irradiation or indicated chemicals. Five days later, SA-β-gal activity was evaluated using the ‘SA-β-gal Staining Kit’ (Beyotime) following the manufacturer's instructions.
Data represent the mean ± SD; statistical significance was assessed by a two-tailed Student's t-test.
This study was supported by National Basic Research Program of China (973 Program) Grants (2011CB966200; 2012CB944700), National Science Foundation Research Grants (81171968 and 30771231), Graduate Student Innovation Fund of Shandong University (10000080398131) and State Program of National Natural Science Foundation of China for Innovative Research Group (81021001).
Conflict of interest
The authors declare that they have no conflict of interest.
HG and CS conceived the project; HG performed the majority of the experiments and data analysis; BX, HH, QL, ZL, XD, ZW, YW, XZ, MZ, and YG provided technical assistance and discussion; HG and CS wrote the manuscript.