With in vivo aging of the organism, upon time and doubling in in vitro culture, or following a stress, cells enter a specific state named senescence. At the molecular level, it is now well established that senescence can be the consequence either of telomere shortening, oxidative stress, accumulation of DNA damage, or activation of oncogenes such as Ras. Senescence is characterized by an increase in cell-size, an irreversible cell-cycle arrest involving p53/p21 and p16/Rb pathways, epigenetics changes, transcriptome, proteome and secretome changes, and an increase in autophagic activity. Our contributions in the field of senescence mainly arise from studies performed with primary cultures of normal human epidermal keratinocytes (NHEKs) of the basal layer and normal human dermal fibroblasts (NHDFs).

Senescence and post-senescence neoplastic emergence

The dominant paradigm in the field of senescence is that senescence is a cell-autonomous tumor-suppressor mechanism that the cell must bypass to become immortal and tumorigenic. As such, it is the first line of defense induced following an initial oncogenic activation. The experiments we performed with NHDFs confirmed this concept. Indeed, in in vitro culture, NHDFs undergo an exponential growth phase followed by a senescence plateau, very stable, associated with shortened telomeres which induce a permanent DNA Damage response (DDR) and an irreversible cell-cycle arrest as published by several teams for different types of human primary fibroblasts. In contrast, NHEKs also undergo an exponential growth phase followed by a senescence plateau, but this latter is not stable and leads to two different outcomes. The main outcome is cell death that affects almost all cells. The alternative outcome affects only about one cell on 10,000. They re-enter cell-cycle and generate clones of cells that re-proliferate (Fig 1A and B).

Fig_growthcurve Fig_morpho

Fig 1. Senescence and emergence of NHEKs. A, growth curve. B, cell morphologies observed by phase-contrast microscopy. Bar, 80 μm.

 We have established by transcriptomic analyses that these post-senescence (PS) cells are transformed and we have shown by xenograft assays in nude mice that they are tumorigenic.

We therefore call this phenomenon Post-Senescence Neoplastic Emergence (PSNE). Both senescent and PS NHEKs display still long telomeres, although the telomerase is not reactivated. Finally, and importantly, we have demonstrated that PS cells are not the progeny of already transformed cells, or stem cells, that could have been present in the original tissue explant, but do originate from the division of fully senescent mother cells by an atypical mechanism of budding mitosis called neosis (Fig 2; Gosselin et al. Cancer Research 2009; 69:7917).


Fig 2. Senescent cells produce emergent cell by budding. Panel 1 and 2 are general views. Panels 3 and 4 are optical cross-sections viewed with a confocal microscope. Arroheads marked emerging cells.


PSNE would not be specific of keratinocytes: a similar phenomenon was described in Normal Mammary Epithelial cells (Romanov et al. Nature 2001; 409:633). In addition, it may occur in vivo. Indeed, senescent cells are found in most benign tumors, most often in consequence of the activation of an oncogene, but are lost in the following malign tumor, suggesting that at least some cells have evade senescence and have re-proliferate to generate the tumor. Therefore, in contrast to senescence in fibroblasts, we have proposed that senescence in epithelial cells, at least keratinocytes and mammary epithelial cells, would be a transient state of stasis during which molecular events prone to malignant transformation could take place. All of our efforts during these last 5 years have focused on understanding the mechanisms of keratinocyte PSNE as an in vitro model of the earliest steps of carcinogenesis.

Our program has currently four components:

1-Post-senescence neoplastic emergence proceeds by escaping autophagic cell death

By observing NHEK cultures, we hypothesized that those senescent cells which do not produce PS cells die, suggesting that PSNE could necessitate cell-death escape. We have demonstrated that senescent NHEKs do not die by apoptosis but following an overactivation of autophagy (Gosselin et al. 2009 Am J Pathol. 2009; 174:423). In contrast, senescence in fibroblasts is very stable and senescent fibroblasts do not undergo massive cell death. We have shown that the NF-kB>MnSOD>H2O2 pathway we demonstrated before (Bernard et al. Cancer Res 2004; 64:472) to be involved in the occurrence of senescence in NHEKs was also responsible for the induction of the associated autophagic activity, through the induction of oxidative damages to mitochondria and nucleus (Deruy et al. Plos One 2010; 5:e12712) We are presently characterizing more precisely the crosstalk between oxidative stress and autophagy and their role in PSNE.

2-Post-senescence neoplastic emergence occurs in consequence of DNA single strand breaks

Oxidative stress and DDR signaling seem to be the common denominators of the senescence-inducing pathways. The oxidative stress occurs with the aging of the organism, or because of ionizing radiations or UV radiations, or following an oncogenic activation. We have shown that oxidative stress is the motor of PSNE. Indeed, treating young NHEKs or young NHDFs with ROS induces a premature senescence plateau followed by PSNE, and, conversely, treating presenescent NHEKs with antioxidants delays the occurrence of senescence and inhibits PSNE (Gosselin et al. Cancer Research 2009; 69:7917). Since NHDFs never spontaneously give rise to PSNE in contrast to NHEKs, we postulated that oxidative stress could operate through the generation of some mutagenic DNA damages that could be either specific or at least more numerous in senescent NHEKs than in senescent NHDFs. We are therefore analyzing DNA double-strand and single-strand breaks (DSBs and SSBs respectively) occurring at senescence of NHEKs versus NHDFs.

3-Senescent fibroblasts enhance early skin carcinogenic events via a paracrine MMP-PAR-1 axis.

Another well demonstrated paradigm in the field of senescence is that, despite being in a cell-autonomous tumor suppressor state, a senescent fibroblast produces a modified secretome -the senescence-associated secretory phenotype (SASP)- which has paracrine inflammatory and tumor-promoting effects on already pre-transformed epithelial cells, but not on normal epithelial cells. We took advantage of our model of PSNE to investigate whether the SASP could also be able to stimulate the very early phases of carcinogenesis. We demonstrated that a culture medium conditioned by senescent NHDFs was able to increase the PSNE frequency and to increase the epithelium-mesenchyme transition of the very first PS NHEKs. We highlighted matrix metalloproteinases 1 and 2 (MMP1 and MMP2) as the main components of the SASP responsible for these changes. We identified the Protease Activated Receptor 1 (PAR1) as a receptor specifically overexpressed in PS NHEKs, activated by MMP1 and MMP2 and transducing the epithelium-mesenchyme transition changes (Malaquin et al. Plos One 2013; 8:e6360). Our goal is to identify other key signaling molecules and mechanisms for further investigations.

4-Cellular senescence involves an intracrine prostaglandin E2 pathway in NHDFs

Lipid mediators such as prostaglandins could also be components of the protumorigenic SASP besides already described cytokines, growth factors and MMPs. We had shown a few years ago that COX-2, the limiting and inducible enzyme of the prostaglandins biosynthesis pathway takes part in senescence of NHDFs (Zdanov et al. Exp Cell Res 2007; 13:3046). We showed that PGE2 acting on its EP specific receptors is able to induce both the onset of senescence and the maintenance of the phenotype. PGE2 acts on senescence more via the pool of intracellular EP receptors than via those localized at the cell surface. We also provided evidence that ROS by-produced by the COX-2 enzymatic activity are negligible in front of the total senescence-associated oxidative stress. Taken together, these results suggest that COX-2 contributes to the establishment and maintenance of senescence of NHDFs via an independent-ROS and a dependent-PGE2/EPs intracrine pathway (Martien et al. BBA Lipids 2013; 1831:1217). In this project, we aim to investigate how this pathway induces senescence.