A negative feedback signaling network underlies oncogene-induced senescence

Abstract

Oncogene-induced senescence functions to limit tumor development. However, a complete understanding of the signals that trigger this type of senescence is currently lacking. We found that mutations affecting NF1, Raf, and Ras induce a global negative feedback response that potently suppresses Ras and/or its effectors. Moreover, these signals promote senescence by inhibiting the Ras/PI3K pathway, which can impact the senescence machinery through HDM2 and FOXO. This negative feedback program is regulated in part by RasGEFs, Sprouty proteins, RasGAPs, and MKPs. Moreover, these signals function in vivo in benign human tumors. Thus, the ultimate response to the aberrant activation of the Ras pathway is a multifaceted negative feedback signaling network that terminates the oncogenic signal and participates in the senescence response.

Figure 1

Loss of neurofibromin induces senescence in human fibroblasts, but immortalizes mouse embryonic fibroblasts A: Growth curve and immunoblot of normal human IMR90 fibroblasts infected with a control lentiral vector (V), or lentiviruses expressing NF1 shRNAs. At each cell passage cells were re-plated and SA-β-gal activity and morphology was assessed (bottom panels). B: SA-β-gal activity in response to p120RasGAP knock-down. C: Proliferative properties of wild-type and NF1^KD MEFs. Doxorubicin (DR) treatment (0.2 μg/ml for 18 hours) was used at passage 21, to assess p53 status.

Figure 2

Senescence triggered by neurofibromin-deficiency requires Rb and p53 pathways A: Immunoblots demonstrating that p53 and Rb are activated in response to NF1-loss. B: NF1 was inactivated in cells previously infected with an empty vector (pMKO.1), LgT, a dominant-negative p53 (p53DD), or a Large T Δ234-444 mutant. SA-β-gal-expression is shown.

Figure 3

Negative feedback suppression of the Ras pathway underlies cellular senescence A: Immunoblots of IMR90 lysates showing the dynamic effects of NF1-loss on pERK and pAKT levels. Intervening lanes are spliced out where indicated, but all time-courses represent data from the same immunoblot. B: Ras activation in NF1^KD versus control cells on days 2 and 6 (see Experimental procedures). The long exposure is shown to highlight the decrease in Ras-GTP levels between vector and NF1 knock-down cells at this latter time point. C: pAKT, Rb and p53 immunoblots of MEF lysates in response to NF1-deficiency. D: Immunoblots of lysates from BJ fibroblasts infected with retrovirus expressing the GRD of neurofibromin, 9 days post-infection. E: SA-β-gal staining of fibroblasts (IMR90 and BJ) infected with the GRD of neurofibromin. F: Immunoblots of cell lysates from BJ fibroblasts in response to NF1-loss, 6 days post-selection.

Figure 4

Aberrant Raf activation suppresses Ras via multiple distinct mechanisms A: SA-β-gal staining of IMR90 infected with Raf1-CAAX (left panel). BJ fibroblasts, expressing an inducible Raf construct, were treated for 30 or 90 minutes with tamoxifen (4-HT, 1 μM). Ras-GTP levels were assessed using a Ras pull-down (PD) assay. Control immunoblots on total cell lysates (TCL) are shown (middle panel). Ras-GTP levels after 24 hours of 4-HT (right panel). B: SOS immunoblots of BJ cells expressing the inducible Raf construct in the presence and absence of CIP (added to lysates) (top). p120RasGAP protein levels increase as pAKT levels decrease (bottom). C: BJ cells expressing the inducible Raf were treated for 42 hours with tamoxifen (1 μM) or ethanol. Real-time PCR was performed in triplicate on the listed genes and fold changes were calculated. D: Immunoblot analysis of lysates from C. E: Real-time PCR analysis of genes in response to NF1-loss, as compared to cycling, vector-expressing cells. F: SA-β-gal staining and immunoblots of BJ cells expressing exogenous Sprouty 2.

Figure 5

Inactivation of the PI3K pathway triggers senescence A: BJ fibroblasts, expressing an inducible form of Raf, were treated with 1 μM tamoxifen (+) or ethanol (−). Immunoblots demonstrate the dramatic decrease in pAKT levels. B: Percentage of SA-β-gal positive IMR90 cells after mock, DMSO, LY294002 and PD98059 treatments lasting 2 or 4 days. C: SA-β-gal staining of BJ cells treated with DMSO, LY294002, Wortmannin or expressing an HA-PTEN (left). Control immunoblots are shown (right).

Figure 6

Oncogenic insult regulates HDM2 and FOXO proteins A: BJ fibroblasts were treated for the indicated times with LY294002. Rb, p53 and p-AKT levels were assessed by immunoblot. B: Levels of p-HDM2 were assessed by immunobloting lysates from BJ cells treated with DMSO (D) or LY294002 (LY) for 24 hours, from cells expressing an inducible Raf construct 48 hours after induction, or in NF1^KD cells after 6 days. C: Luciferase activity of a FOXO reporter in the presence of exogenous FOXO1 or FOXO3, 72 hours after Raf induction in BJ cells. D: A phosphorylation mutant of FOXO1 (FKHR-AAA) induces senescence and Rb activation in human fibroblasts.

Figure 7

Senescence occurs in human neurofibromas in vivo A and B: SA-β-gal expression in dermal neurofibromas. C: A bright-field picture of a SA-β-gal stained section (blue) overlayed with a dark-field image of immunofluorescent neurofibromin staining (pink). Dermal neurofibroma sections, first stained for SA-β-gal, were then stained for: D: p16, E: AP2-α, F: Krox-20. G: p-ERK, H: p-AKT (top large image and middle small images), and FOXO1 (bottom small images). High magnification images of single cells demonstrate the nuclear localization of FOXO1 and an absence of nuclear p-ERK or p-AKT staining in all senescent cells. For FOXO1 analysis sections were counterstained with hematoxylin to demonstrate surrounding negative cells.

Figure 8

Revised model of oncogene-induced senescence A: In cells that are insensitive to oncogene-induced senescence, activation of the Ras pathway leads to hyper-activation of well known effector pathways which function to promote tumorigenesis. B: In sensitive cells the ultimate response to the aberrant activation of the Ras pathway is the initiation of a multi-faceted negative feedback signaling network designed to terminate the oncogenic signal. These signals are triggered by the Raf/MEK/ERK pathway and involve numerous transcriptional and post-translational events, including the suppression of Ras exchange factors, and the up-regulation of Sprouty proteins and RasGAPs, among others. These inhibitory signals, via the consequential suppression of PI3K, can activate Rb and p53 through multiple signals. Thus, we propose that oncogene-induced senescence is mediated by this negative feedback signaling program, that functions along with other known senescence regulators, such as p16 and ARF, to achieve a threshold senescence signal.