Jun dimerization protein 2 controls hypoxia‐induced replicative senescence via both the p16Ink4a‐pRb and Arf‐p53 pathways

The main regulators of replicative senescence in mice are p16Ink4a and Arf, inhibitors of cell cycle progression. Jun dimerization protein 2 (JDP2)‐deficient mouse embryonic fibroblasts are resistant to replicative senescence through recruitment of the Polycomb repressive complexes 1 and 2 to the promoter of the gene that encodes p16Ink4a and inhibits the methylation of lysine 27 of the histone H3 locus. However, whether or not JDP2 is able to regulate the chromatin signaling of either p16Ink4a‐pRb or Arf‐p53, or both, in response to oxidative stress remains elusive. Thus, this study sought to clarify this point. We demonstrated that the introduction of JDP2 leads to upregulation of p16Ink4a and Arf and decreases cell proliferation in the presence of environmental (20% O2), but not in low (3% O2) oxygen. JDP2‐mediated growth suppression was inhibited by the downregulation of both p16Ink4a and Arf. Conversely, the forced expression of p16Ink4a or Arf inhibited cell growth even in the absence of JDP2. The downregulation of both the p53 and pRb pathways, but not each individually, was sufficient to block JDP2‐dependent growth inhibition. These data suggest that JDP2 induces p16Ink4a and Arf by mediating signals from oxidative stress, resulting in cell cycle arrest via both the p16Ink4a‐pRb and Arf‐p53 pathways.

The main regulators of replicative senescence in mice are p16 Ink4a and Arf, inhibitors of cell cycle progression. Jun dimerization protein 2 (JDP2)-deficient mouse embryonic fibroblasts are resistant to replicative senescence through recruitment of the Polycomb repressive complexes 1 and 2 to the promoter of the gene that encodes p16 Ink4a and inhibits the methylation of lysine 27 of the histone H3 locus. However, whether or not JDP2 is able to regulate the chromatin signaling of either p16 Ink4a -pRb or Arf-p53, or both, in response to oxidative stress remains elusive. Thus, this study sought to clarify this point. We demonstrated that the introduction of JDP2 leads to upregulation of p16 Ink4a and Arf and decreases cell proliferation in the presence of environmental (20% O 2 ), but not in low (3% O 2 ) oxygen. JDP2mediated growth suppression was inhibited by the downregulation of both p16 Ink4a and Arf. Conversely, the forced expression of p16 Ink4a or Arf inhibited cell growth even in the absence of JDP2. The downregulation of both the p53 and pRb pathways, but not each individually, was sufficient to block JDP2-dependent growth inhibition. These data suggest that JDP2 induces p16 Ink4a and Arf by mediating signals from oxidative stress, resulting in cell cycle arrest via both the p16 Ink4a -pRb and Arf-p53 pathways.
After several weeks in cell culture, primary cells stop proliferating and enter an irreversible growth arrest stage called replicative senescence. Replicative senescence involves processes that include the accumulation of oxidative stress, genotoxic stress, and telomere shortening. These stimuli finally induce the expression of p16 Ink4a and Arf. In the case of cultured mouse fibroblasts, cells undergo senescence even though they have long telomeres and high telomerase activity [1]. The maintenance of telomere length remains important, as telomerase deficiency shortens the life span of mice and leads to premature aging [2][3][4]. However, it seems likely that factors such as oxidative stress, rather than telomere shortening, contribute to the activation of the Ink4a/Arf locus in the case of mouse fibroblasts, based on the observation that mouse embryonic fibroblasts (MEFs) cultured in low (3%) oxygen can proliferate for longer periods without senescence [5]. Transcription from the Ink4a/Arf locus is under complex control; p16 Ink4a and Arf respond independently to positive and negative signals, and the entire locus is epigenetically regulated. In young proliferating primary cells, the locus is transcriptionally silenced by the trimethylation of lysine 27 of histone H3 (H3K27). By contrast, the expression of p16 Ink4a and Arf increases in aged and senescent cells as a result of the loss of H3K27 trimethylation [6]. The methylation of H3K27 and the silencing of the p16 Ink4a /Arf locus are mediated by the Polycomb repressive complexes 1 (PRC1) and 2 (PRC2). PRC1 and PRC2 form a complex at the Ink4a/Arf locus in young cells, and dissociate from this locus in aged cells. A proposed molecular mechanism of PRC-mediated gene silencing is that Ezh2, a catalytic subunit of PRC2, trimethylates histone H3K27 [7], which acts as a binding site for the CBX subunit of PRC1 [8,9]. The Ring1B and Bmi1 subunits of PRC1 ubiquitylate histone H2A [10,11], resulting in compaction of chromatin and inhibition of the elongation of RNA polymerase II [12]. In aged and stressed cells, H3K27 trimethylation markers are lost because of the H3K27-specific demethylase, Jumonji domain containing 3C [13,14], and PRC1 dissociates from the p16 Ink4a locus, resulting in transcriptional activation by activators that include Ets1 and/or Ets2 [15]. p16 Ink4a binds cdk4/6 and changes its conformation, which prevents the phosphorylation of pRb by the cdk4/6-cyclin D complex. Thus, pRb-bound E2F fails to activate genes that are essential for cell cycle progression, such as cyclin E1. Arf is an inhibitor of mouse double minute 2 homolog, which, in turn, mediates the degradation of p53. Therefore, the expression of Arf stabilizes p53 and activates its downstream cell cycle inhibitors, including p21 Cip/Waf1 . In brief, p16 Ink4a and Arf inhibit cell proliferation via pRb-and p53-dependent pathways, respectively.
In this study, we showed that overexpression of JDP2 led to the upregulation of p16 Ink4a and Arf. We downregulated p16 Ink4a and Arf using a lentiviral shRNA, to assess whether the inhibition of cell proliferation might be dependent on p16 Ink4a and Arf, and found that the overexpression of JDP2 failed to induce cell cycle arrest in the p16 Ink4a /Arf-deficient condition. Interestingly, JDP2 did not inhibit cell cycle progression in hypoxia (3% O 2 ), but did so at environmental oxygen concentration (20% O 2 ). In addition, we demonstrated that the downregulation of both p53 and pRb, but not of each of them individually, suppressed JDP2-dependent growth inhibition. Taken together, these results suggest that JDP2 mediates signals that arise from environmental oxidative stress, induces the expression of p16 Ink4a and Arf, and inhibits cell cycle progression via both the p53 and pRb pathways.

5-Ethynyl-2 0 -deoxyuridine (EdU) incorporation assay
MEFs cultured in a 3% O 2 /5% CO 2 incubator were infected with the lentivirus for shRNA and/or that for gene expression at an multiplicity of infection (MOI) of 3. Two days later, the infected cells were selected in the presence of 10 lgÁmL À1 of puromycin (for the shRNA expression vector) and/or 10 lgÁmL À1 of blasticidin (for the protein expression vector) and incubated for at least 2 days in the presence of 3% O 2 in order to eliminate uninfected cells.

AlamarBlue assay
Cell proliferation was examined by the methods with Ala-marBlue reagent assays according to the manufacturers' instructions. The cells in the 96-well plate were cultured for   Table S1. Error bars in each graph represent the SD of the data. 1 or 4 days in the presence of 100 mL of Dulbecco's modified Eagle's medium supplemented with 10% FBS. Subsequently, 10 mL per well of AlamarBlue solution (DAL1100; Thermo Fisher Scientific Inc.) was added and further incubated for 2 h. The absorbance of each well was measured at 595 nm, and the cell numbers in the well were calculated. The standard deviation (SD) for the calculation of cell number was generated by pilot experiments using known samples. The division times (fold) from days 11 to 14 are shown in Fig. S2.

Statistical analysis
The quantitative variables are presented as the mean AE SD. The significance of differences was determined using an unpaired two-tailed Student's t-test. Differences with a P value < 0.05 were considered significant.

Results and Discussion
Comparative cell proliferation and expression of p16 Ink4a and Arf in the normoxic and hypoxic conditions To evaluate the effect of JDP2 on cell proliferation at environmental oxygen concentration (20% O 2 ), we generated JDP2 stable transformants by infecting Jdp2 À/À MEFs with a lentiviral JDP2 expression vector. After selecting the transformants based on blasticidin resistance in the low oxygen condition (3% O 2 ), A B Fig. 5. Effect of JDP2 on the trimethylation of histone H3K27 at the p16 Ink4a /Arf locus. (A) Schematic diagram showing the details of the assay conditions. Jdp2 À/À MEF cultured in the low oxygen concentration were infected with CSII or CSII-JDP2 and the transformants were selected using blasticidin for 3 days. The cells were cultured for 1 additional week in the environmental or low oxygen concentrations and their lysates were used in the ChIP assay. (B) Trimethylation of histone H3K27 at the p16 Ink4a /Arf locus. The ChIP assay was performed using the cell lysates, which were treated as indicated in (A). The DNA fragments bound to the trimethyl-H3K27 were immunoprecipitated using an antitrimethyl-H3K27 antibody. The amount of DNA corresponding to the p16 Ink4a /Arf locus was assessed by real-time RT-PCR using specific primers. Each value was normalized to the amount of input DNA. Error bars in each graph represent the SD of the data. the cells were further cultured in 20% or 3% O 2 for 1 week (Fig. 1A). The percentage of growing cells was analyzed by the EdU incorporation assay. The growth of JDP2 transformants was inhibited compared with that of the control (empty vector) transformants after 1 week of cell culture in 20% O 2 (P = 0.016; Fig. 1B,C). By contrast, no significant difference in proliferation between JDP2 and the empty vector transformants was observed when the cells were cultured in 3% O 2 (P = 0.32; Fig. 1B,C). To confirm these results of EdU incorporation using an alternative approach, we also performed an Ala-marBlue's assay. We compared the growth rates of the JDP2 and empty vector transformants from day 11 to day 14 after exposure to environmental O 2 concentration. The result also showed that the cell growth in the presence of JDP2 was inhibited compared with that of the empty vector control under the environmental O 2 concentration, but not under low oxygen concentration (Fig. S2). We analyzed the expression of the mRNA for p16 Ink4a and Arf by real-time RT-PCR. We found that the expression of p16 Ink4a and Arf was elevated in the JDP2 transformants compared with that of the empty vector transformants cultured in 20% O 2 (Fig. 1D), but not in 3% O 2 (Fig. 1D). The balancing of the effect of p16 Ink4a and Arf decided the rate of proliferation in this case. Taken together, these data indicate that both the overexpression of JDP2 and environmental oxidative stress are required for the induction of JDP2-dependent growth inhibition.
JDP2 plays a role in an oxidative stressdependent inhibition of cell growth Next, we examined whether endogenous JDP2 functions as an oxidative stress-dependent inhibitor of cell growth because our previous study demonstrated that endogenous JDP2 was elevated in 20% O 2 in comparison with that in 3% O 2 [24]. We generated transformants in which JDP2 was downregulated by infecting Wt MEFs with a lentiviral shRNA vector for JDP2 (shJDP2; Fig. 2A). After 1 week of cell culture in 20% O 2 , the shJDP2 transformants exhibited a higher rate of proliferation (P = 5.4 9 10 À4 ; Fig. 2B,C). Conversely, after 1 day of culture in 20% O 2 with a prior 6-day treatment of cells under 3% O 2 , no significant differences in cell proliferation were observed (P = 0.82). Similar results were obtained from the EdU assay using different shRNA against JDP2 (TRCN0000081973; Fig. S1). The result of the experiment using this alternative shJDP2 demonstrated that, after 27 days of cell culture in 20% O 2 , the higher rate of cell growth of shJDP2 transformants was more significant (P = 1 9 10 À4 ) than that of 7 days (P = 2.8 9 10 À3 ) and no significant change was observed after 1 day (P = 0.71; Fig. S3). These data suggest that the effect of shJDP2 was specific. The EdU-positive cells were increased by this treatment as compared with that of 1-week culture of cells in 20% O 2 , and the endogenous expression of JDP did not reach the control level of plko transformants after 1 week's culture of cells in 20% O 2 (Fig. 2C,D). These data indicate that a high level of endogenous JDP2 induces the inhibition of cell proliferation in response to the oxidative stress that accumulated for 1-week culture.
We analyzed the expression of the mRNA for p16 Ink4a and Arf and observed a lower expression of both mRNA in the shJDP2 transformants after 1 week of cell culture in the environmental oxygen concentration (Fig. 2D). By contrast, no significant differences in the expression of p16 Ink4a were observed between the shJDP2 and empty vector transformants after 1 day of cell culture. However, Arf level in shJDP2 transformants was slightly lower than that in the control vector transformants. Both the p16 Ink4a and Arf mRNA levels were lower in 1-day normoxia condition than that in 7 days (Fig. 2D), which reflected the difference in EdU-positive cells between the two conditions (Fig. 2C).

Effects of knockdown of p16 Ink4a and Arf on cell growth
Based on these results, we hypothesized that JDP2 mediates the signal from oxidative stress and increases the expression of cell cycle inhibitors, including p16 Ink4a and Arf. To substantiate this hypothesis, we generated transformants in which p16 Ink4a and Arf were downregulated by shRNA and analyzed the effect of JDP2 on cell proliferation after 2 weeks of cell culture in 20% O 2 (Fig. 3A). In the presence of JDP2, the expression of the p16 Ink4a and Arf mRNA was increased (Fig. 3D), and cell proliferation was inhibited (P = 1.7 9 10 À4 ; Fig. 3B,C). In contrast, cell growth was promoted rather than inhibited (P = 1.2 9 10 À5 ) in the shp16 Ink4a/ Arf transformants, in which the mRNA for p16 Ink4a and Arf were downregulated.
These results suggest that JDP2 is an upstream positive regulator of p16 Ink4a and Arf and that the presence of one or both proteins is essential for effective JDP2-mediated growth inhibition. We also confirmed that the forced expression of p16 Ink4a and Arf inhibited cell proliferation in both Jdp2 À/À and Wt MEFs (Fig. S4A). Forced expression of p16 Ink4a did not increase the expression of JDP2 in Wt MEFs (Fig. S4B). These data suggest that JDP2 is neither a downstream factor in the signal transduction pathway nor a transcriptional target of p16 Ink4a . Interestingly, the forced expression of Arf decreased the expression of JDP2; thus, a negative feedback loop depending on the Arf-p53 pathway might exist.
Both p53 and pRb pathways are required for JDP2-dependent growth suppression We examined which pathway is essential for JDP2dependent cell cycle inhibition. As it is technically difficult to downregulate p16 Ink4a and Arf independently via shRNA, we downregulated their downstream targets, pRb and p53, respectively, via lentiviral shRNA and analyzed the effect of JDP2 on cell growth by cointroducing a JDP2 expression vector. The EdU incorporation assay was performed after 2 weeks of cell culture in 20% O 2 (Fig. 4A). The downregulation of pRb and p53 was confirmed by real-time RT-PCR (Fig. 4B). The results of the cell proliferation assay demonstrated that the downregulation of either p53 or pRb was not sufficient for the blockage of the JDP2dependent growth inhibition (Fig. 4C,D). However, a deficiency in both p53 and pRb significantly inhibited the growth arrest observed in the presence of JDP2. These data demonstrate that both the Arf-p53 and p16 Ink4a -pRb pathways are required for JDP2-dependent inhibition of cell proliferation.
JDP2 regulates the trimethylation of H3K27 at p16 Ink4a and Arf locus As the expression of p16 Ink4a and Arf is epigenetically regulated by histone methylation, we analyzed the effect of JDP2 on the methylation of H3K27 in the p16 Ink4a /Arf locus. We performed a ChIP assay using an anti-histone H3 trimethyl lysine 27 antibody and a cell extract of Jdp2 À/À MEFs infected with the JDP2 expression vector (CSII-JDP2) or the empty vector (CSII) and cultured cells for 1 week at the environmental or low oxygen concentration. The results of this experiment demonstrated that the extent of the trimethylation of lysine 27 in histone H3 at both the p16 Ink4a and Arf transcription regulatory sites was decreased in the presence of JDP2 under the environmental oxygen concentration (Fig. 5A,B). Under the low oxygen concentration, the trimethylation of lysine 27 remained at a higher level even in the presence of JDP2, suggesting that JDP2 affects the methylation of H3K27 only under the higher environmental oxygen conditions.

Conclusion
Based on our findings, we propose the following model (Fig. 6). In young undamaged cells, the expression of p16 Ink4a and Arf is silenced because of the trimethylation of histone H3K27. The expression of JDP2 does not play a role in the absence of a stress signal. In old cells, the oxidative stress signal mediated by JDP2 induces the demethylation of histone H3 on the transcription regulation sites of both p16 Ink4a and Arf, resulting in the activation of the pRb and p53 pathways and the cell cycle arrest. Our findings demonstrate that JDP2 is essential for replicative senescence and is involved in both the p16 Ink4a -pRb and Arf-p53 pathways for the inhibition of cell growth by mediating the signal from oxidative stress.

Supporting information
Additional Supporting Information may be found online in the supporting information tab for this article: Fig. S1. Evaluation of the inhibitory activity of the lentiviral shRNA expression vectors. Fig. S2. Effect of the forced expression of JDP2 on cell growth in Jdp2 À/À MEFs using an AlamarBlue assay. Fig. S3. Downregulation of JDP2 by shRNA targeting different site (TRCN0000081973) inhibited the growth arrest induced by oxidative stress. Fig. S4. Forced expression of p16 Ink4a or Arf inhibits cell proliferation even in the absence of JDP2. Table S1. Actual values of cell growth for Figs 1-4 are listed.