Senescence and NFκB: A trojan horse in tumors?

The induction of senescence in tumor cells impairs transformation and promotes an anticancer immune response resulting from the production by senescent cells of cytokines and chemokines, an aspect known as "senescence-associated secretory phenotype" (SASP). Here we discuss recent findings regarding the role of NFκB in the modulation of the SASP and the consequent anticancer immune response.

Cellular senescence constitutes an irreversible cell cycle arrest that occurs in normal cells in response to various adverse conditions, including telomere shortening in replicative senescence as well as telomere dysfunction in oncogene-induced senescence (OIS). In addition, premature (or accelerated) senescence can be readily elicited in tumor cells by the administration of sublethal concentrations of conventional anticancer drugs.

The acquisition of a senescent phenotype in cells from different tissues (e.g., fibroblast, epithelial or endothelial cells) is accompanied by an increased secretion of specific cytokines, chemokines and other factors, which affect both senescent and neighboring, non-senescent cells via autocrine/paracrine mechanism. The secretion of these factors has been referred to as “senescence-associated secretory phenotype” (SASP). Notably, the SASP occurs not only in vitro, but also in vivo, in response to oncogene activation or DNA-damaging chemotherapy.

The induction of cell senescence has been associated with various anticancer effects. For instance, OIS acts as a very effective barrier to cell transformation, in both mouse models and in human cancers. In addition, the induction of an accelerated senescent phenotype by chemotherapeutic agents critically contributes to the efficacy of anticancer therapy. The SASP has been implicated in several of the tumor suppressive effects of senescence. Kuilman and colleagues identified interleukin (IL)-6 and IL-8 as essential mediators of oncogenic BRAF-induced senescence. In response to oncogenic stress, the activation of the transcription factor C/EBPβ stimulates IL-6 and IL-8 secretion. Interference with either of these cytokines, or with their receptors, resulted in the bypass of the senescence. Interestingly, IL-6 depletion also reduced C/EBPβ levels, highlighting the existence of a positive feedback loop that acts to sustain the senescence-associated cell cycle arrest. Membrane bound IL-1α has also been shown to act as a positive regulator of IL-6 and IL-8 expression in senescent human fibroblasts. Indeed, the autocrine stimulation of senescent cells with IL-1α induces the constitutive activation of NFκB, which in turn transactivates IL-1α, resulting in another positive feedback loop. The inhibition of IL-1α signaling limited both NFκB and C/EBPβ activity, as well as the secretion of IL-6 and IL-8. Finally, a positive feedback loop involving the chemokine receptor CXCR2 and its ligands, namely IL-8 and GRO-α, has been shown to act to reinforce replicative senescence in primary human fibroblasts. NFκB and C/EBPβ coordinately regulate the expression of CXCR2 ligands during OIS. Hence, several components involved in the SASP operate within positive feedback loops that reinforce the growth arrest associated with cell senescence.

Besides contributing to the autocrine loops that reinforce the senescent phenotype, the SASP is also responsible—at least in part—for a crosstalk between senescent cells and cells of the immune system. This crosstalk mediates additional anticancer function of senescence, as it is important for the recruitment of immune cells to the tumor site as well as for clearance of senescent cells. Such an immunomodulatory function of the SASP has been described in normal senescent cells as well as in senescent tumor cells., For instance, the restoration of endogenous p53 expression in hepatocarcinoma and sarcoma leads to induction of a senescence program that triggers tumor clearance in vivo, through the activation of an innate immune response. The molecular mechanisms mediating such anticancer immune responses have not yet been fully elucidated, although SASP components including several immune modulators are likely to be involved., In addition, it has been shown that senescent myeloma cells express natural killer (NK) cell-activating ligands on their surface.

We recently investigated the crosstalk between prematurely senescent tumor cells and cells of the innate immune system, as well as the potential involvement of death receptors in the recognition and clearance of premature senescent carcinoma cells by the immune system. We evaluated the expression of FAS, DR4 and DR5 on proliferating and senescent (in response to chemotherapy) induced lung adenocarcinoma A549 and breast carcinoma MCF7 cells. We showed that the induction of senescence in both tumor cell lines is associated with an increase in FAS expression. Such upregulation of FAS in senescent cells sensitized cells to the induction of apoptosis.

Inflammatory cytokines have previously been implicated in the modulation of FAS expression in different cell types. For instance, the administration of both tumor necrosis factor α (TNFα) and interferon γ (IFNγ) has been shown to trigger FAS upregulation in murine fibroblasts. IFNγ combined with IL-1β also increased FAS expression on human thyrocytes, resulting in enhanced susceptibility to FAS ligand (FASL)-mediated cell death. Finally, human endometrial stromal cells are normally resistant to FAS-dependent apoptosis, but become readily sensitive upon the administration of TNFα and IFNγ, which cause FAS upregulation. These observations prompted us to investigate whether increased FAS expression in senescent tumor cells was dependent on the SASP. Analyses of media conditioned by senescent MCF-7 cells highlighted an increased secretion of several cytokines and chemokines. Among these cytokines, we identified TNFα and IFNγ as mediating the upregulation of FAS in our experimental system. Accordingly, the treatment of proliferating cancer cell lines with TNFα and IFNγ resulted in FAS upregulation, while interfering with the TNFα and IFNγ signaling pathways decreased the amount of FAS expressed on the surface senescent tumor cells. Hence, SASP components are likely to promote the recognition of senescent carcinoma cells and their elimination by FASL-positive immune cells.

Previous studies have demonstrated a crucial role for NFκB in the induction of FAS expression., Since an increased activity of NFκB has been associated with cellular senescence, we investigated the ability of NFκB to modulate the SASP in carcinoma cells pushed into senescence by the administration of chemotherapeutic drugs. To this aim, we knocked-down the expression of RelA, the main member of the NFκB family, in MCF-7 and A549 cells by RNA interference, and then analyzed media conditioned from such cells (or cells receiving a scamble siRNA) upon the induction of senescence. Interestingly, many SASP components were significantly reduced in RelA-depleted cells. The essential role of NFκB in modulating the SASP has recently been substantiated in response to diverse senescence-inducing stimuli. Notably, the reduction of RelA signaling in senescent MCF-7 and A549 cells inhibited the secretion of both TNFα and IFNγ, as well as the surface expression of FAS. These data demonstrate the existence of a RelA-dependent autocrine loop that controls FAS expression by senescent tumor cells. Moreover, these findings suggest a novel anticancer function of the SASP: the sensitization of senescent tumor cells to FAS-mediated apoptosis (Fig. 1).

Figure 1. The induction of senescence promotes FAS upregulation via NFκB as well as the release of chemokines (blue diamonds) that recruit cells of the innate immune system. In addition, senescence results in the secretion of cytokines (red and yellow dots) that act locally within autocrine and paracrine loops, hence sustaining the senescent phenotype and promoting the killing of senescent cells by the immune system.

Hence, SASP components form a complicated network that is involved in reinforcing the senescence-associated growth arrest as well as in promoting the recognition and clearance of senescence cells by the immune system. These two biological outcomes of cell senescence are likely to be induced in vivo, in different settings. For instance, senescent melanocytes in nevi are maintained in a growth-arrested state for years, while senescent carcinoma and sarcoma cells are readily cleared by the innate immune system., It is tempting to speculate that the interplay of distinct transcription factors, including but perhaps not limited to NFκB, C/EBPβ and STAT1 might induce specific SASP components that regulate the outcome of senescence. Further investigation is required to assess how distinct secretory profiles are induced in different physiopathological and experimental settings.

Several lines of evidence point to NFκB as a critical regulator of the SASP. By using in silico approaches, different groups have identified multiple NFκB-controlled genes that are upregulated in senescent cells., However, the precise role of NFκB in the induction of senescence and in the maintenance of the senescent phenotype is not yet clear, similar to the functional role of NFκB in promoting growth arrest and aging as well as cell growth and cancer. Experimental evidence suggests that NFκB is important for the induction of a growth arrest. For example, Penzo, et al. reported a transient proliferative block in murine fibroblast upon enforced NFκB activation. Accordingly, Rovillain and colleagues found that different NFκB target genes are up- or downregulated upon senescent growth arrest, and that interfering with the activation of NFκB bypasses growth arrest. Conversely, our data strongly suggest that NFκB does not play a major role in mediating the induction of cell senescence and the associated growth arrest upon DNA damage, as NFκB-depleted cells readily undergo senescence in response to DNA-damaging agents. Rather NFκB seems to be important for the maintenance of the senescent phenotype. In fact, Rovillain, et al. reported that blocking NFκB activity (by the ectopic expression of a superrepressor form of the inhibitory subunit IκBα) overcomes growth arrest and increases the number of growing colonies. Accordingly, in our experimental system NFκB-deficient senescent cells escape from the senescence-associated cell cycle arrest (Crescenzi and Leonardi, unpublished data). This effect, however, is seemingly not due to a direct control of cell cycle progression by NFκB, but perhaps to the increased genomic instability observed in the absence of functional NFκB., In addition, as secreted proteins play a critical role in the maintenance of some forms of OIS,, it is also possible that quantitative and/or qualitative differences in the SASP (for instance, a loss of IL-8 secretion) might account for the increased escape of NFκB-deficient cells from the induction of senescence.

We and others have demonstrated that the secretion of some SASP components depends upon the transcriptional activity of NFκB.,, Such cytokines and chemokines are largely pro-inflammatory. However, at odds from cancer cells that take advantage of the inflammatory microenvironment and of the recruitment of inflammatory cells, senescent cells are cleared as a consequence of the SASP. Indeed, senescent cells upregulate FAS expression and recruit inflammatory cells to mediate their own clearance. These effects appear to be mediated by NFκB via an autocrine loop requiring TNFα and IFNγ. Interestingly, a similar autocrine loop involving TNFα has also been demonstrated to mediate the killing of cells in response to high doses of etoposide. It is possible that NFκB activation is necessary for the establishment of the SASP, and that its functional outcome, that is pro-tumor vs. antitumor, depends upon the microenvironment and/or subtle differences in the cytokine composition of the SASP. Indeed, cells lacking p53 are known to secrete markedly higher levels of most SASP factors, suggesting that one of the ways whereby p53 suppresses tumorigenesis is by modulating the inflammatory microenvironment.

From this perspective, senescence and NFκB may be considered a Trojan horse that we may harness against cancer with a dual advantage: (1) senescence may be triggered with a relatively low dose of anticancer agents (which results in less severe side effects) and (2) senescent cells—via the SASP—activate the immune system (hence resulting, at least in some scenarios, in improved antitumor effects).