Loss of HIF1α induces premature senescence in wild-type MEFs. (A) Western and Northern analyses of Ad-empty-infected (Ad) or Ad-Cre-infected (Cre) HIF1α^fl/fl MEFs following 24 h incubation in either normoxia (Nx; 21% O₂) or hypoxia (Hyp; 2% O₂). The top panels show immunoblots for HIF1α and for α-tubulin as a loading control. Molecular weight markers in kilodaltons. The bottom panels show expression of the HIF1α target gene GLUT1, and the ethidium bromide-stained gel as a loading control. (B) Growth curves of Ad-empty-infected or Ad-Cre-infected HIF1α^fl/fl MEFs grown in either normoxia or hypoxia. (C) Representative photographs of SA-βgal staining of the cells in B after the last counting time point. (D) Quantitation of the SA-βgal-positive fraction of cells in C. “Ad Nx” and “Cre Nx” are statistically different by Student’s t-test (p = 0.003). Error bars indicate standard deviation.

MIF is a HIF1α-dependent hypoxia-inducible gene. (A) Northern analyses of Ad-empty-infected or Ad-Cre-infected HIF1α^fl/fl MEFs in normoxia (Nx) or 2% O₂ (Hyp). GLUT1 and MIF transcript expression are shown, as well as the ethidium bromide-stained gel as a loading control. (B) Western analyses of cells treated as in A. The immunoblots were probed with antibodies for MIF or α-tubulin as indicated. Molecular weight markers in kilodaltons. (C) Promoter analysis of the −70 to +63 region of the murine MIF promoter. Promoter fragments were cloned into pGL3promoter; a 5XHRE construct, including five copies of the human VEGF HRE, serves as a positive control. Constructs (identified by the letters A–G in the middle of the figure) are schematically represented in the bar diagram on the left, while their relative promoter activities in normoxia (black bars) and hypoxia (gray bars) are displayed on the right of the figure. Arrows at the bottom of the bar diagram indicate the regions that are essential for promoter activity. (D) Promoter regions determined to be required for promoter activity in C. Putative promoter-binding sites were identified by computer database analysis (http://www.genomatrix.de). Four nucleotide mutations in the core of each site are indicated. (E) Luciferase assays of construct C with specific mutations identified in D. Mutations of the four core nucleotides of each site were made alone or in combination, as indicated at the bottom of the bar graph. (F) ChIP analysis of the murine MIF promoter. Genomic DNA from the mouse keratinocyte cell line (Balb/MK) was fixed and immunoprecipitated with either IgG or an α-HIF1α antibody. Subsequently, PCR was performed with primers specific to the GLUT1 HRE, putative MIF HRE, or upstream MIF sequences (MIF control). Error bars indicate standard deviation.

MIF is necessary and sufficient to delay senescence downstream from HIF1α. (A) Northern blot demonstrating the knockdown of MIF using a retroviral shRNA in normoxic-treated (N) and hypoxic-treated (H) wild-type MEFs at day 0 and day 11 following selection. GLUT1 hybridization demonstrates that MIF knockdown does not markedly affect other HIF1 target genes. Ethidium bromide staining of the RNA demonstrates equal loading of the gel. (B) Western blot demonstrating the knockdown of MIF protein in normoxia (Nx) and hypoxia (Hyp). α-Tubulin is used as a loading control. Molecular weight in kilodaltons. (C) Growth curves of MIF knockdown and control (GFP)-infected cells grown in normoxia (Nx) or 2% O₂ (Hyp). (D) Quantification of SA-βgal staining of cells in B. “shMIF Nx” and “GFP Nx” are statistically different by Student’s t-test (p = 0.022). (E) Western analysis of stable ectopic expression of MIF in MEFs following Ad-empty (Ad) or Ad-Cre (Cre) infection. Immunoblots for HIF1α (arrow), MIF, and α-tubulin as a loading control are shown. pLPC refers to the empty vector backbone, compared with pLPCMIF, which expresses MIF. Molecular weight in kilodaltons. (F) Growth curves of MIF-overexpressing cells following infection with Ad-empty or Ad-Cre in normoxia. Error bars indicate standard deviation.

Loss of HIF1 activity sensitizes cells to radiation-induced senescence. (A) Growth curves of Ad-empty-infected or Ad-Cre-infected HIF1α^fl/fl MEFs grown in 2% O₂ following treatment with 0.5 Gy or 2 Gy of radiation. (B) Representative photographs of SA-βgal staining of the cells in A after the last time point. (C) Quantification of the SA-βgal-positive cells in B. “Ad 0.5 Gy” and “Cre 0.5 Gy” are statistically different by Student’s t-test (p = 0.0002); “Ad 2 Gy” and “Cre 2 Gy” are also statistically different by Student’s t-test (p = 0.0003). (D) Growth curves of GFP or shMIF MEFs grown in 2% O₂ following treatment with 0.5 Gy or 2 Gy of radiation. (E) Quantification of the SA-βgal-positive cells in D 13 d after irradiation. “GFP 0.5 Gy” and “shMIF 0.5 Gy”, and “GFP 2 Gy” and “shMIF 2 Gy” are both also statistically different by Student’s t-test (p < 0.0001). (F) Model for the role of HIF1α in modulating senescence through the regulation of MIF. Error bars indicate standard deviation.