Senescent cells as a source of inflammatory factors for tumor progression

Cellular senescence was first identified as a process that limits the proliferation of human cells in culture [26]. These early experiments showed that cultured human fibroblasts gradually lose proliferative capacity until all cells in the culture cease division. Much of this growth arrest is known to occur because most human cells do not express telomerase. Consequently, with each cell cycle, telomeres shorten and eventually fail, generating a persistent DNA damage signal that permanently arrests growth [27]. It is now known, however, that cellular senescence occurs in culture and in vivo as a response to excessive intracellular or extracellular stress (see Fig. 1). The senescence program locks the cell into a cell cycle arrest that prevents the damage from affecting the next cell generation, thereby preventing potential malignant transformation [24]. Senescent cells have been shown to accumulate over the life span of rodents, non-human primates, and humans and are found primarily in renewable tissues [15, 28].

Many types of stresses can provoke cellular senescence [5, 29] (see Fig. 1). These stresses include dysfunctional telomeres resulting from repeated cell division (replicative senescence) or other telomeric damage. They also include oxidative stress resulting from mitochondrial deterioration or other causes, severe or irreparable DNA damage from external sources and disrupted chromatin organization due to DNA replication or damage (genotoxic stress), and the expression of certain oncogenes (oncogene-induced senescence) [10, 14, 30‐35]. Stresses that cause cellular senescence can be induced by external or internal chemical or physical insults encountered during the course of the life span, during therapeutic interventions (for example, X-irradiation or chemotherapy), or as a consequence of endogenous processes such as oxidative respiration or mitogenic signals. External mitogenic signals, for example growth-related oncogene alpha (GROα) secretion by tumor cells in close to normal cells [36] or circulating angiotensin II [37, 38], have also been shown to induce cellular senescence. All somatic cells that have the ability to divide can potentially undergo senescence. Regardless of the disparate mechanisms of senescent-inducing stresses, the senescence program is activated once a cell has sensed a critical level of damage or dysfunction. Senescent cells are detected in culture and in vivo by a variety of markers, including the senescence-associated β-galactosidase (SA-βgal) [28], p16^INK4a [39], telomere-associated DNA damage foci [40], senescence-associated heterochromatin foci [41], and several other molecules; all of these markers have some limitations, and thus must be used in combination with each other, as well as proliferation markers (absent in senescent cells) [42].

Cellular senescence is accompanied by a striking increase in the secreted levels of >40 factors involved in intercellular signaling [22, 25, 43]. This phenotype has been termed the “senescence-associated secretory phenotype”, or SASP [22]. The SASP has many of the paracrine effects one would expect from a pro-inflammatory stimulus, which can be deleterious if left unchecked (see Fig. 2). For example, senescent cells promote the proliferation and tumorigenesis of epithelial cells [18], stimulate angiogenesis [44], trigger an epithelial to mesenchymal transition, accelerate the invasion of transformed cells [22], and increase the growth of xenograft tumors in vivo [19]. Further, the SASP has been shown to occur after treatment of cancer patients with DNA damaging chemotherapy [22].

The SASP includes several families of soluble and insoluble factors. These factors can affect surrounding cells by activating various cell surface receptors and corresponding signal transduction pathways that may lead to multiple pathologies, including cancer. SASP factors can globally be divided into the following major categories: soluble signaling factors (interleukins, chemokines, and growth factors), secreted proteases, and secreted insoluble components. SASP proteases can have three major effects: shedding of membrane-associated proteins resulting in soluble versions of membrane-bound receptors, cleavage/degradation of signaling molecules, and degradation or processing of the extracellular matrix. These activities provide potent mechanisms by which senescent cells can modify the tissue microenvironment. In the following sections, we discuss the SASP subsets and some of their known paracrine effects on nearby cells with an emphasis on their ability to facilitate cancer progression.

Overall, the gene expression profiles (mRNA) of senescent cells determined by transcript analyses resemble the profiles of secreted proteins determined by antibody arrays [22, 23]. This finding suggests that the secretory phenotype of senescent cells is at least in part regulated at the transcriptional level. However, because the changes in gene expression that occurs at senescence are so widespread, the transcriptional control may well be at the level of chromatin organization, rather than changes in specific transcription factors. In fact, dramatic chromatin alterations are known to occur at senescence [72‐75]. Support for the idea that gene expression specific to the senescence program may be partially attributed to larger changes in chromatin conformation is suggested by the physical clustering of genes that comprise the SASP [23].

The expression of many SASP components depends on the transcription factors NF-κB and C/EBPβ, which have increased activity at senescence [47, 76] (A. Freund and J. Campisi, unpublished data). Knockdown of C/EBPβ diminishes the expression of both IL-6 and IL-8, which are among the most upregulated cytokines at senescence [47], and NF-κB knockdown significantly decreases the levels of numerous (75%) SASP factors (A. Freund and J. Campisi, unpublished data).

A surprising recent finding was that the DNA damage response (DDR) is required for the increased secretion of a subset of SASP factors, including IL-6 and IL-8 [77]. The DDR is a signal amplification cascade that senses DNA damage, inducing cell cycle arrest and initiating DNA damage repair. If the extent of DNA damage is irreparable, the cell undergoes cellular senescence but maintains a chronic, low level DDR [78]. It is this persistent DDR that is necessary for a robust SASP; depletion of upstream components of the DDR cascade, specifically ATM, NBS1, or CHK2, prevents an increase in secreted protein levels of IL-6, IL-8, and the GRO family, among others.

However, the DDR cannot be the sole regulator of the SASP because the DDR is activated immediately after damage, whereas the SASP, like other aspects of the senescence phenotype such as SA-βgal activity, takes several days to develop. Additionally, a transient DDR—for example, caused by a low level of ionizing radiation that does not induce senescence—does not induce a SASP [77]. Thus, while the DDR is important, it is not sufficient; there must be other, slower events that cooperate with the DDR to induce the SASP. One such event is p38MAPK activation, which increases slowly after DNA damage, reaching peak levels after several days (A. Freund and J. Campisi, unpublished data).

Like many cytokine networks, the SASP also has an important positive-feedback component. IL-1α, a cytokine that regulates its own synthesis in an autocrine, receptor-mediated, positive-feedback loop via NF-κB [79, 80], is a key positive regulator of IL-6 and IL-8 expression at senescence [81]. IL-1α depletion by RNAi in senescent cells markedly reduced the extracellular protein levels of IL-6 and IL-8. Similar results were achieved using an IL-1 receptor (IL-1R) antagonist or neutralizing antibodies, demonstrating that sustained IL-1R stimulation by surface-bound IL-1α is required to maintain senescence-associated IL-6 and IL-8 extracellular levels [81].

MicroRNAs also play a role in SASP regulation. Thus far, two microRNAs, miR-146a and miR-146b (miR-146a/b), have been demonstrated to negatively regulate the senescence-associated secretion of IL-6 and IL-8 [82]. Senescent human fibroblasts with a strong SASP upregulate these microRNAs late in senescence, tuning down the secretion of inflammatory cytokines. IRAK1 is an established target of miR-146a/b and is indeed targeted in senescent cells, suggesting that these microRNAs downregulate the SASP by reducing NF-κB activity [83]. When miR-146a/b are overexpressed in human fibroblasts, IRAK1 levels decline, along with the enhanced secretion of senescence-associated IL-6/IL-8. In addition, blockage of IL-1 receptor signaling prevented upregulation of miR-146a/b, consistent with these microRNAs being part of an NF-κB feedback loop [83]. These data once again highlight the important role of NF-κB, and place miR146a/b as central players of IL-6 and IL-8 secretion within the SASP.

Although the SASP is at least partly regulated by the activation of transcription factors, the general gene expression profile acquired at senescence may be more attributable to the larger changes in chromatin conformation that senescent cells develop [72‐74, 84]. Consistent with this idea is the physical clustering of SASP genes such as matrix metalloproteinases (MMP-1, MMP-3, MMP-10, and MMP-12) as well as CXCL and CCL family members [23]. Moreover, recent work indicates the key role for the High Mobility Group Box 1 protein (HMGB1) in senescent cells.

HMGB1 is a member of the non-histone, chromatin-binding high mobility group family of proteins [85]. Unlike histones, HMGB1 loosely binds the minor grove of DNA, stabilizing nucleosome formation, thereby influencing the expression of certain genes [86‐88]. HMGB1 was shown to directly interact with p53, and recombinant HMGB1 protein enhanced p53 DNA binding in vitro [89].

The discovery that necrotic cells, but not apoptotic cells, passively release HMGB1 focused recent studies on the role of extracellular HMGB1 [90]. Initial studies suggested that extracellular recombinant HMGB1 promoted secretion of inflammatory cytokines (TNFα, IL-1β, and IL-6) by macrophages [91]. However, later work revealed that recombinant HMGB1 protein alone exhibited only weak activity, but synergized with the ability of bacterial components like lipopolysaccharides to stimulate inflammatory cytokine secretion [92, 93]. Further, HMGB1 can enhance the pro-inflammatory activity of cytokines [94]. These findings appear in conflict with those showing that anti-HMGB1 antibodies or antagonist reduced inflammation in arthritis models [95]. However, the general consensus now is that HMGB1 alone may not promote inflammation, but rather augment the inflammatory response in certain biological contexts.

While it was not observed that cells cultured with recombinant HMGB1 stimulated cytokine secretion, extracellular HMGB1 may enhance of the effect of the SASP (A. Davalos and J. Campisi, unpublished data). Similar to recent data showing that apoptotic cells release “find-me signals”, which recruit innate immune to clear the dying cells, senescent cells also secrete HMGB1, known to recruit and activate cells from the innate immune system [96]. This notion is consistent with two studies that suggest senescent cells may undergo clearance in vivo [13, 97].

Despite the obvious benefit of recruiting and activating innate immune cells to clear senescence cells, extracellular HMGB1 protein may promote cancer and metastatic progression. Experiments in cell culture showed that HMGB1 stimulated the migration of neuroblastoma cells, which was attenuated by an antibody against HMGB1 [98]. Further, in an in vitro angiogenesis assay, HMGB1 stimulated sprouting of endothelial [99]. When immunocompromised mice were injected with rat glioma cells, followed by injection with anti-RAGE (a HMGB1 receptor) or anti-HMGB1 antibodies, both antibodies individually reduced tumor volume, but maximal reduction occurred with both antibodies [100]. Thus, extracellular HMGB1 is able to provoke both beneficial and deleterious biological consequences. It may act to alert and identify senescent cells that require clearance. Failure to clear all senescent cells could leave sites of persistent secretion of inflammatory molecules.

Inflammation has been shown to initiate or enhance several age-related diseases, including atherosclerosis, Alzheimer’s, osteoarthritis, and cancer [101‐103]. Factors secreted by senescent cells can promote tumor development in vivo and malignant phenotypes such as proliferation and invasiveness in cell culture models. These effects have been observed with a number of cell types, including cells derived from breast [18, 19, 21, 22, 104], skin [105], prostate [22, 106], pancreas [107], and oro-pharyngial mucosa [108]. The effects of the complex SASP is, of course, dependent on the tissue context. In the following sections, we present the various behavioral changes cells can undergo when residing in the proximity of senescent cells and discuss how the senescent tissue microenvironment can facilitate tumor initiation and progression (see Fig. 2).