5-Lipoxygenase regulates senescence-like growth arrest by promoting ROS-dependent p53 activation

Abstract

5-Lipoxygenase (5LO) is involved in the production of leukotrienes and reactive oxygen species (ROS) from arachidonic acid. Its strong activation has been associated with several diseases like cancer and neurodegeneration. Here we show that 5LO activity increases during senescence-like growth arrest induced by oncogenic ras or culture history in both human and mouse embryo fibroblasts. Overexpression of 5LO promotes senescence-like growth arrest via a p53/p21-dependent pathway, and this occurs independently of telomerase activity. 5LO stabilizes p53 through phosphorylation at Ser15 and increases expression of the p53-transcriptional target p21. This is achieved by regulating ROS production. Indeed, ROS are increased in 5LO-arrested cells. Antioxidants and a low oxygen environment prevent 5LO-induced growth arrest. Finally, 5LO inhibition reduces the growth arrest induced by oncogenic ras or culture history and these effects are neutralized by the addition of exogenous ROS. These data link the 5LO pathway to oxidative crises of primary fibroblast and suggest that the ability of 5LO to induce senescence-like growth arrest may be important in the pathogenesis of 5LO-associated disorders.

Figure 1

Oncogenic ras induces 5LO expression and activity. (A) WI38 fibroblasts were infected with empty virus as a control (B0) or Ha-RasV12-expressing retrovirus, or treated with 250 μM H₂O₂ for 2 h. Day 0 is the first day after selection for 4 days in 1 μg/ml puromycin. After 7 days, cells reduced [³H]thymidine uptake and were positive for SA-β-gal activity (data not shown). Cell lysates were then prepared and Western blot analyses of RasV12, 5LO and the senescence-related proteins p53 and p21 were performed. (B) Western blot analysis of 5LO in the indicated cells treated as described in the legend to panel A. Expression of actin was used as an internal control. (C) Ha-RasV12-infected or H₂O₂-treated WI38 cells were incubated with 10 μM calcium ionophore A23187. After 15 min at 37°C, the reaction was terminated, and the formation of AA metabolites (LTB4, ‘total LTA4-derived metabolites' and 5-HETE) was assayed as described in Materials and methods. P-values are also indicated, ANOVA, n=4.

Figure 2

5LO deficiency abolishes oncogenic ras-induced ROS production and stasis of MEFs. (A) Growth curves of wild-type or 5LO^−/− MEFs infected with control (B0) or Ha-RasV12 vectors. (B) DCF fluorescence intensity (left) and O₂⁻ recognition by lucigenin (right), assayed as described in Materials and methods, were used as indicators of ROS levels in both wild-type and 5LO^−/− MEFs infected with B0 or oncogenic ras. (C) 5LO^−/− MEFs infected with B0 or oncogenic ras were treated with 5 or 10 μM t-BTH for 2 days, that is, from 4 to 6 days after selection. SA-β-gal-stained cells were counted 7 days after selection. (D) Annexin–propidium iodide (PI) staining of B0- and Ha-RasV12-infected wild-type or 5LO^−/− MEFs. ^*P⩽0.05 versus control, ANOVA, n=3.

Figure 3

Comparison of the effects of different 5LO protein levels on ROS production and stasis in WI38 cells. (A, B) WI38 cells were infected with a pBabe-based vector containing empty (B0), GFP-tagged 5LO or its catalytically inactive mutant (H367Q). Infected cells were then selected or unselected with puromycin (for 4 days, 1 μg/ml). After 7 days, [³H]thymidine was added for 3 days and cells were subsequently stained for SA-β-gal (A) followed by autoradiography (B) as described in Materials and methods. About 500 cells were counted to determine the percentages of radiolabeled nuclei (% LN) and positive SA-β-gal cells (blue). A cell was only considered SA-β-gal positive when it was not radiolabeled. (C) ROS detection by DCF fluorescence intensity in WI38 cells after infection with GFP-tagged 5LO or H367Q in the presence or absence of puromycin. (D) Conditioned media and protein samples of equal size were then assayed for LTB4 formation (by HPLC procedures) and for 5LO protein expression using an anti-GFP as antibody (inset), respectively. ^*P⩽0.001 versus B0, ANOVA, n=3.

Figure 4

Impact of ROS in 5LO-induced stasis. (A) O₂⁻, H₂O₂, NO₂⁻ and 8-isoprostane productions, analyzed as described in Materials and methods, were detected in WI38 cells containing an empty vector (B0), wild-type 5LO, its catalytically inactive mutant (H367Q) or oncogenic ras. ^*P⩽0.05 versus control vector (B0), ANOVA, n=3. (B, C) Growth curves and SA-β-gal activity of WI38 fibroblasts infected with the indicated retroviruses. Infected cells were cultured in 20 or 3% oxygen for the indicated period of time (B) or with 1 or 2 mM NAC as well as 0.5 or 1 μM AA861 for 10 days (C). ^*P⩽0.05 versus cells infected with 5LO in 20% oxygen, ANOVA, n=3.

Figure 5

5LO induces stasis engaging p53/p21 pathway. (A) Western blot analysis of GFP, p53, p21, PML, p16, pRb and actin in WI38 cells containing an empty vector (B0), GFP-tagged 5LO or its mutant H367Q. (B) Growth curves of p53^−/− or p21^−/− MEFs infected with B0 or 5LO vectors. ^*P⩽0.05 versus B0, paired t-test, n=3. (C) Immunoblotting of 5LO and p21 in both wild-type and p53^−/− MEFs infected with B0 or 5LO vectors. Expression of actin was used as an internal control.

Figure 6

5LO promotes p53 phosphorylation. (A) Western blot analysis of p21, p53, actin and p16 upon infection of wild-type and 5LO^−/− MEFs with control (B0) or Ha-RasV12 vectors. (B) The luciferase reporter gene plasmid p21-luc was transfected into wild-type and 5LO^−/− MEFs infected with B0 or Ha-RasV12 and cultured in the absence or presence of 1 μM AA861 (AA), 2 mM NAC (N) or 10 μM t-BTH (BH). At 24 h after transfection, the luciferase activity was measured. The activity in B0-infected wild-type MEFs is set at 100%. P-values are also indicated, paired t-test, n=3. (C) Wild-type and 5LO^−/− MEFs were infected with control (B0), Ha-RasV12 or 5LO. Cell lysates were prepared when Ha-RasV12- or 5LO-infected culture was terminally arrested, immunoprecipitated with anti-p53 antibody and analyzed by Western blot using anti-p53-phosphoSer15 and Ser46 antibodies. (D) Luciferase activity of p21-luc in p53^−/− MEFs untransfected (□) or transfected with 50 pg (▒) or 100 pg (▪) of wild-type p53 or p53S15A constructs in the presence of control (B0) or Ha-RasV12- or 5LO-expressing vectors. Values are normalized by cotransfected β-gal. (E) MEFs were infected with Ha-RasV12, 5LO or control (B0) and cultured with or without NAC (2 mM) or AA861 (1 μM). Cell lysates were prepared as described in the legend to panel C and protein levels of p53 and phosphoSer15 were detected by Western blotting (left). 5LO^−/− MEFs expressing control (B0) or Ha-RasV12 were untreated or treated with 10 μM t-BTH as described in the legend to Figure 2C. Protein levels of p53 and phosphoSer15 were detected by Western blotting.

Figure 7

5LO controls cellular lifespan of primary fibroblasts. (A) WI38 fibroblasts were serially subcultured from early passage until they acquired the senescent phenotype. Protein extracts were prepared at the indicated times: early passage, p14; midpassage, p28; late passage, p37; senescence, p44. Representative Western blots of 5LO, p53, Ser15 p53, p21 and pRb from three independent experiments are shown. (B) 5LO metabolites in conditioned media of early passage, midpassage, late passage and senescence cells were measured as described in Materials and methods. (C) DCF fluorescence intensity of WI38 fibroblasts subcultured from early passage (p14) until senescence (p44) in the presence of indicated concentrations of AA861. (D) Cell lysates were prepared at the early passage and senescence, immunoprecipitated with anti-p53 antibody and analyzed by Western blot using anti-p53-phosphoSer15 antibodies. (E, F) Growth curves of WI38 cells incubated with different concentrations of AA861 (E) or wild-type and 5LO^−/− MEFs (F). The number of population doublings (PDs) was determined over the indicated period of time. ^*P⩽0.05 versus controls, ANOVA, n=3.

Figure 8

Summary of stasis program involving 5LO-linked cascade in primary fibroblasts. For simplicity, only RasV12-mediated oncogenic activation and ‘culture shock' are shown, although a number of other effectors and/or stimuli are known to be capable of inducing stasis.