Available online at www.sciencedirect.com Cellular senescence: hot or what? Gerard I Evan 1 and Fabrizio d’Adda di Fagagna 2 The phenomenon of replicative senescence was first observed more than 40 years ago by Hayflick who noted the inability of cultured human fibroblasts to proliferate indefinitely. The recent discovery that cellular senescence is triggered by many different activated oncogenes has led to the notion that senescence, like oncogene-induced apoptosis, serves as a critical and cell-autonomous tumor preventive mechanism. Both the DNA damage response and the ARF tumor suppressor have been mechanistically implicated in oncogene- induced senescence and the relative contributions of, and potential interactions between, these two pathways remain subjects of a lively debate. More recently, the discovery that cellular senescence can be bypassed during the epithelial– mesenchymal transition (EMT) that typically accompanies tumor progression, the observation that organ fibrosis is controlled by cellular senescence and, most noticeably, the mounting evidence linking cellular senescence to inflammation, make cellular senescence a still flaming hot subject after all these years. Addresses 1 Department of Pathology and Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, CA 94143-0502, USA 2 IFOM Foundation – FIRC Institute of Molecular Oncology Foundation, via Adamello 16, 20139 Milan, Italy Corresponding author: d’Adda di Fagagna, Fabrizio (fabrizio.dadda@ifom-ieo-campus.it) Current Opinion in Genetics & Development 2009, 19:25–31 This review comes from a themed issue on Genetic and cellular mechanisms of oncogenesis Edited by Julian Downward and William Hahn Available online 30th January 2009 0959-437X/$ – see front matter # 2008 Elsevier Ltd. All rights reserved. DOI 10.1016/j.gde.2008.11.009 Introduction Cellular senescence is characterized by the inability of cells to proliferate despite the presence of abundant nutrients and mitogens and by the maintenance of cell viability and metabolic activity [1]. The phenomenon was first noted by Hayflick [2  ,3] who demonstrated that normal human fibroblasts, when maintained in culture conditions that drove their continuous replication, even- tually loose proliferative capacity after a finite number of cell divisions. We now know that in many human fibro- blast cell lines this ‘Hayflick limit’ is the consequence of progressive shortening of telomeres (the ends of linear chromosomes) at each replicative round. When the length of one or more telomeres falls below a certain threshold, the exposed DNA end is recognized as a DNA double- strand break (DSB) by the DNA damage response (DDR) machinery [4  ,5  ] and triggers the canonical DNA damage checkpoint [6]. The DDR coordinates activation of the DNA repair machinery and a prompt replicative arrest, the so-called checkpoint, to allow the cell to repair the damage. Upon completion of DNA repair, DDR signaling attenuates and the cell re-enters cell cycle. However, especially severe or unresolvable/persistent DNA damage triggers permanent blockades to cell pro- liferation. Depending on cell types and context, this can involve either programmed cell death (apoptosis) or a form of viable but permanent proliferative arrest, dubbed replicative cellular senescence, marked by characteristic senescence-associated DNA damage foci (SDFs). Impor- tantly, suboptimal growth conditions may also cause a stress that ultimately leads to DNA damage accumu- lation, DDR activation and senescence [7,88]. Induction of senescence is also the frequent outcome of oncogenic mutations in normal cells and, like oncogene- induced apoptosis [8], is thought to have evolved as an important mechanism to limit tumor evolution ([9  ] and Table 1 for a list of senescence-inducing oncogenes). Oncogene-induced senescence (OIS) appears to be engaged by a variety of mechanisms, both cell-autonom- ous and cell-extrinsic, one of which is the proclivity for many oncogenic mutations to activate the DDR by alter- ing the DNA replication process [10  ,11  ]. Indeed, sev- eral reports have now described DDR activation in the earliest phases of tumor formation in both human samples and mice models, and evidence of strong selection for inactivating mutations in DDR genes in more advanced cancers (reviewed in [12] and [6,13]). Belt and braces: ARF and DDR At least in murine cells, most growth-deregulating onco- genic mutations are potent inducers of p19 ARF , the pro- duct of the alternate reading frame in the INK4a locus. p19 ARF is a potent activator of p53 through its inhibitory actions on Mdm2, the E3 ubiquitin ligase and transcrip- tional squelcher that maintains p53 function at a consti- tutively low level in normal, unstressed cells. p19 ARF expression is absent from virtually all normal adult and developing mouse tissues but is potently induced in tumors [14,15  ]. Although p19 ARF can cooperate with DNA damage to engage p53 [16], multiple studies suggest that p19 ARF plays no part in p53-mediated DNA damage responses in normal tissues in vivo www.sciencedirect.com Current Opinion in Genetics & Development 2009, 19:25–31