anscriptomic and proteomic modifications 2583244 indicative of an only partial transformed state, accompanied by a very slight EMT. Yet they can develop into many small, disseminated skin hyperplasias and UPF 1069 carcinomas when xenografted into nude mice, indicating that they have actually undergone profound neoplastic changes. In the present study, we establish that the generation of these cells and the modifications they undergo are influenced by the protumorigenic switch of the aging microenvironment. We show that when exposed to a senescent fibroblast secretome, keratinocytes escape from senescence more frequently, and that the postsenescent keratinocytes display a stronger EMT, including loss of E-cadherin, upregulation of vimentin and TWIST-1, a scattered phenotype, and acquisition of a migratory capacity, these being very early events in carcinogenesis. In our previous study, we noticed that hyperplasias and carcinomas arising from xenografted PSE-NHEKs developed mostly in aged mice. We now assume that they probably developed only after the aged microenvironment was established. Moreover, the skin hyperplasias and carcinomas did not develop at the site of the graft, but were highly disseminated, in accordance with very early acquisition of a migratory capacity by PSE-NHEKs. It is noteworthy that in vitro, PSE-NHEKs appear more sensitive to the pro-EMT effect of the senescent fibroblast secretome when the NHEKs are exposed to this secretome from the beginning of the culture. This suggests that the senescent microenvironment may contribute to the early transformation of keratinocytes. We show in this study that the secretome of dermal fibroblasts is similar to that described for other cell models. Among the cytokines and growth factors of this secretome known to have protumoral effects, IL-1, -6, and -8 appear to contribute very little or only indirectly to keratinocyte migration in vitro. We therefore did not privilege them as potential inducers of the proEMT effect on PSE-NHEKs. HGF/SF has been implicated in keratinocyte migration during wound healing, and its receptor, c-Met is essential to epithelial cell scattering. Although the synthesis and secretion of HGF/SF are enhanced in senescent NHDFs, this factor can induce PSE-NHEK migration only very modestly, in agreement with the loss of C-Met 
in these cells. This, by the way, supports the view that PSE-NHEKs have undergone an EMT, since c-Met is normally expressed only by epithelial cells. The TGF-b1 pathway is known to be involved in EMT induction, partly through induction of TWIST-1 synthesis. Senescent NHDFs show increased levels of the latent and active forms of TGF-b1, and although PSENHEKs show the same level TGF-b as young NHEKs, recombinant TGF-b1 failed to induce PSE-NHEK migration directly. It thus seems that neither HGF/SF nor TGF-b is responsible for the migration-promoting effect of SF-CM on PSENHEKs. MMPs are also an established component of the senescent fibroblast secretome. Here we 1432690 confirm this with with NHDFs, particularly as regards MMP-1 and MMP-2. Our results of in-gel and in situ zymography demonstrate the activity of the secreted MMPs, associated with reduced TIMP-1 expression. MMPs are widely known to promote late-phase cancer cell invasion through their proteinolytic action on extracellular matrix components. Their involvement in earlier phases of cancer initiation and their ability to trigger EMT have also been demonstrated. As components of the SASP, MMPs have been l