.
F U L L T E X T S O U R C E ( . p d f ) : Cell
Cellular senescence is a cell state implicated in various physiological processes and a wide spectrum of age-related diseases. Recently, interest in therapeutically targeting senescence to improve healthy aging and age-related disease, otherwise known as senotherapy, has been growing rapidly. Thus, the accurate detection of senescent cells, especially in vivo, is essential. Here, we present a consensus from the International Cell Senescence Association (ICSA), defining and discussing key cellular and molecular features of senescence and offering recommendations on how to use them as biomarkers. We also present a resource tool to facilitate the identification of genes linked with senescence, SeneQuest (available at http://Senequest.net).Lastly, we propose an algorithm to accurately assess and quantify senescence, both in cultured cells and in vivo.
Cellular Senescence: Walking a Line between Life and Death
Cell states link both physiological and stress signals to tissue homeostasis and organismal health. In both cases, the outcomes vary and are determined by the signal characteristics (type, magnitude, and duration), spatiotemporal parameters (where and when), and cellular capacity to respond (Gorgoulis et al., 2018). In the case of potentially damaging stress, damage is reversed and the structural and functional integrity of cells restored. Alternatively, damage can be irreversible, and cells activate death mechanisms mainly to restrict the impact on tissue degeneration. Between these extremes, cells can acquire other states, often associated with survival but also with permanent structural and functional changes. An example is the non-proliferative but viable state, distinct from G0 quiescence and terminal differentiation, termed cellular senescence (Rodier and Campisi, 2011). Formally described in 1961 by Hayflick and colleagues, cellular senescence, derived from the latin word senex meaning ‘‘old’’ (Hayflick and Moorhead, 1961), was originally observed in normal diploid cells that ceased to proliferate after a finite number of divisions (Hayflick limit), later attributed to telomere shortening (see section ‘‘CellCycle Arrest’’). Cellular senescence has since been identified as a response to numerous stressors, including exposure to genotoxic agents, nutrient deprivation, hypoxia, mitochondrial dysfunction, and oncogene activation (Table 1). Over the last decade, improved experimental tools and the development of reporterablation mouse models have significantly advanced our knowledge about causes and phenotypic consequences of senescent cells. However, specific markers and a consensus on the definition of what constitutes senescent cells are lacking. Further, although a link to organismal aging is clear, aging and senescence are not synonymous (Rodier and Campisi, 2011). Indeed, cells can undergo senescence, regardless of organismal age, due to myriad signals including those independent of telomere shortening. Consequently, senescent cells are detected at any life stage from embryogenesis, where they contribute to tissue development, to adulthood, where they prevent the propagation of damaged cells and contribute to tissue repair and tumor suppression. Thus, cellular senescence might be an example of evolutionary antagonistic pleiotropy or a cellular program with beneficial and detrimental effects. Here, we clarify the nature of cellular senescence by: (1) presenting key features of senescent cells, (2) providing a comprehensive definition of senescence, (3) suggesting means to identify senescent cells, and (4) delineating the role of senescent cells in physiological and pathological processes, that altogether pave the way for developing new therapeutic strategies.
Definition and Characteristics of Cellular Senescence
Cellular senescence is a cell state triggered by stressful insults and certain physiological processes, characterized by a prolonged and generally irreversible cell-cycle arrest with secretory features, macromolecular damage, and altered metabolism (Figure 1). These features can be inter-dependent (Figure 1) but for clarity are described here separately.
Cell-Cycle Arrest
One common feature of senescent cells is an essentially irreversible cell-cycle arrest that can be an alarm response instigated by deleterious stimuli or aberrant proliferation. This cell-cycle withdrawal differs from quiescence and terminal differentiation (He and Sharpless, 2017). Quiescence is a temporary arrest state with proliferation re-instated by appropriate stimuli; terminal differentiation is the acquisition of specific cellular functions accompanied by a durable cell-cycle arrest mediated by pathways distinct from those of cellular senescence (Figure 2). In turn, senescent cells acquire a new phenotype. Although the senescence cell-cycle arrest is generally irreversible, cell-cycle re-entry can occur under certain circumstances, particularly in tumor cells (Galanos et al., 2016; Milanovic et al., 2018; Patel et al., 2016; Saleh et al., 2019) (Figure 2). In mammalian cells, the retinoblastoma (RB) family and p53 proteins are important for establishing senescent cell-cycle arrest (Rodier and Campisi, 2011). RB1 and its family members p107 (RBL1) and p130 (RBL2) are phosphorylated by specific cyclin-dependent kinases (CDKs; CDK4, CDK6, CDK2). This phosphorylation reduces the ability of the RB family members to repress E2F family transcription factor activity, which is required for cell-cycle progression (Sharpless and Sherr, 2015). In senescent cells, however, the CDK2 inhibitor p21WAF1/Cip1 (CDKN1A) and CDK4/6 inhibitor p16INK4A (CDKN2A) accumulate. This accumulation results in persistent activation of RB family proteins, inhibition of E2F transactivation, and consequent cell-cycle arrest, which, in time, cannot be reversed by subsequent inactivation of RB family proteins or p53 (Beause´ jour et al., 2003). This persistence is enforced by heterochromatinization of E2F target genes (Salama et al., 2014), the effects of cytokines secreted by senescent cells (Rodier and Campisi, 2011), and/or enduring reactive oxygen species (ROS) production (Takahashi et al., 2006). Notably, in senescent murine cells, ARF—an alternate reading frame protein of the p16INK4a gene locus that activates p53—also has an important role in regulating cell-cycle arrest (Sharpless and Sherr, 2015). Additional features of the senescent cell-cycle arrest include ribosome biogenesis defects and derepression of retrotransposons (De Cecco et al., 2019; Lessard et al., 2018). However, currently no specific marker of the senescent cell-cycle arrest has been identified (Hernandez-Segura et al., 2017). For example, RB and p53 activation also occurs in other forms of cell-cycle arrest (Rodier and Campisi, 2011). Even p16INK4A, which is considered more specific to senescence, is expressed in certain non-senescent cells (Sharpless and Sherr, 2015) and is not expressed by all senescent cells (Hernandez-Segura et al., 2017). Thus, detecting a senescence-associated cell-cycle arrest requires quantification of multiple factors and features.
Secretion
Senescent cells secrete a plethora of factors, including pro-inflammatory cytokines and chemokines, growth modulators, angiogenic factors, and matrix metalloproteinases (MMPs), collectively termed the senescent associated secretory phenotype (SASP) or senescence messaging secretome (SMS) (Figure 1; Table 2) (Coppe´ et al., 2010; Kuilman and Peeper, 2009). The SASP constitutes a hallmark of senescent cells and mediates many of their patho-physiological effects. For example, the SASP reinforces and spreads senescence in autocrine and paracrine fashions (Acosta et al., 2013; Coppe´ et al., 2010; Kuilman and Peeper, 2009) and activates immune responses that eliminate senescent cells (Krizhanovsky et al., 2008a; Mun˜ oz-Espı´n and Serrano, 2014). SASP factors mediate developmental senescence (Mun˜ oz-Espı´n et al., 2013; Storer et al., 2013), wound healing (Demaria et al., 2014), and tissue plasticity (Mosteiro et al., 2016) and also contribute to persistent chronic inflammation (known as inflammaging) (Franceschi and Campisi, 2014). Thus, the SASP can explain some of the deleterious, pro-aging effects of senescent cells. Further, the SASP can recruit immature immune-suppressive myeloid cells to prostate and liver tumors (Di Mitri et al., 2014; Eggert et al., 2016) and stimulate tumorigenesis by driving angiogenesis and metastasis (Coppe´ et al., 2010).
While the senescent cell-cycle arrest is regulated by the p53 and p16INK4A/Rb tumor suppressor pathways, the SASP is controlled by enhancer remodeling and activation of transcription factors, such as NF-kB, C/EBPb, GATA4 (Ito et al., 2017; Kang et al., 2015; Kuilman and Peeper, 2009; Salama et al., 2014), mammalian target of rapamycin (mTOR) and p38MAPK signaling pathways (Freund et al., 2011; Ito et al., 2017; Kuilman and Peeper, 2009). Upstream signals triggering SASP activation are multiple and differ depending on the senescence inducer but include DNA damage, cytoplasmic chromatin fragments (CCFs) that trigger a type 1 interferon response, and damage-associated molecular patterns (DAMPs) that activate the inflammasome (Acosta et al., 2013; Davalos et al., 2013; Li and Chen, 2018).
The SASP composition and strength varies substantially, depending on the duration of senescence, origin of the pro-senescence stimulus, and cell type (Childs et al., 2015). Further, single-cell RNA sequencing (scRNA-seq) reveals considerable cell-to-cell variability of SASP expression (Wiley et al., 2017). For example, transition from an early transforming growth factor b (TGF-b)-dependent secretome to a pro-inflammatory secretome is governed by fluctuation of Notch1 activity (Ito et al., 2017). Moreover, an interferon type 1 response occurs as a later event and is driven in part by derepression of LINE-1 retrotransposable elements (De Cecco et al., 2019). Senescent cells also communicate with their microenvironment through juxtacrine NOTCH/JAG1 signaling (Ito et al., 2017), release of ROS (Kuilman et al., 2010), cytoplasmic bridges (Video S1) (Biran et al., 2015), and extracellular vesicles, such as exosomes Figure 1. The Hallmarks of the Senescence Phenotype Senescent cells exhibit the following four interdependent hallmarks: (1) cell-cycle withdrawal, (2) macromolecular damage, (3) secretory phenotype (SASP), and (4) deregulated metabolism (see also text). (Takasugi et al., 2017). Overall, defining the senescent secretome in each biological context will help identify senescencebased molecular signatures.
Macromolecular Damage
- DNA Damage. The first molecular feature associated with senescence was telomere shortening, a result of the DNA end-replication problem, during serial passages (Shay and Wright, 2019). Telomeres are repetitive DNA structures found in terminal loops at chromosomal ends and stabilized by the Shelterin protein complex. This organization renders telomeres unrecognizable by the DNA damage response (DDR) and doublestrand DNA break (DSB) repair pathways (de Lange, 2018; Shay and Wright, 2019). Telomerase, the enzyme that maintains telomere length, is not expressed by most normal somatic (non-stem) cells but is expressed by most cancer cells that have overcome senescence. Moreover, telomerase activity reconstitution in normal cells leads to telomere elongation, extending their replicative lifespan in culture (Bodnar et al., 1998; Shay and Wright, 2019).
Telomere shortening during proliferation culminates in telomeric DNA loop destabilization and telomere uncapping, generating telomere dysfunction-induced foci (TIFs) that activate the DDR, eventually causing cell-cycle arrest. This response can also be elicited by inhibiting or altering genes involved in telomere maintenance (d’Adda di Fagagna, 2008). Another form of DNA damage, termed telomere-associated foci (TAFs), can exist at telomeres due to oxidative DNA damage at telomeric G-reach repeats, irrespective of telomere length or shelterin loss (de Lange, 2018; Shay and Wright, 2019).
Although half of the persistent DNA damage foci in senescent cells localize to telomeres, other stressful subcytotoxic insults can trigger senescence by inducing irreparable DNA damage (Figure 1). Numerous genotoxic agents, including radiation (ionizing and UV), pharmacological agents (e.g., certain chemotherapeutics), and oxidative stress trigger senescence. Moreover, activated oncogenes can induce senescence (known as OIS) as a tumor-suppressive response, restricting the uncontrolled proliferation of potentially oncogenic cells. OIS is often mediated by the tumor suppressors p16INK4A and ARF, both encoded by the CDKN2A locus, imposing a cell-cycle arrest (Kuilman et al., 2010; Serrano et al., 1997). However, the DDR also plays a major role in triggering OIS (Gorgoulis and Halazonetis, 2010; Gorgoulis et al., 2018; Halazonetis et al., 2008). In this case, the damage signal originates at collapsed replication forks as a result of oncogene-driven hyperproliferation. Recently, it was shown that the DDR and ARF pathways can act in concert during OIS with the former requiring a lower oncogenic load than the latter (Gorgoulis et al., 2018).
Senescent cells harbor persistent nuclear DNA damage foci termed DNA-SCARSs (DNA segments with chromatin alterations reinforcing senescence). DNA-SCARSs are distinct from transient damage foci; unlike transient foci, they specifically associate with promyelocytic leukemian (PML) nuclear bodies, lack the DNA repair proteins RPA and RAD51 as well as singlestranded DNA (ssDNA), and contain activated forms of the DDR mediators CHK2 and p53 (Rodier et al., 2011). DNASCARSs are dynamic structures with the potential to regulate multiple aspects of the senescent cells, including growth arrest and SASP (Rodier et al., 2011). However, as not all senescence-inducing stimuli generate a persistent DNA damage response, DNA-SCARSs are not a global feature of the senescent cells. CCFs are another type of DNA damage in senescent cells (Ivanov et al., 2013). These CCFs activate a proinflammatory response, mediated by the cGAS-cGAMP-STING pathway (Ivanov et al., 2013; Li and Chen, 2018), that can serve as another non-inclusive senescence-associated marker.
- Protein Damage. Proteotoxicity is a hallmark of aging and cellular senescence (Kaushik and Cuervo, 2015). Hence, Figure 2. Cell-Cycle Withdrawal in Senescent, Quiescent, and Terminally Differentiated Cells Depicted are differences in cell-cycle-arrest reversibility, activated signals (see text), secretory functions, and macromolecular damage that allow discrimination between these cellular states. Macromolecular damage is a common feature of senescence. Secretion is another common feature of senescence and is sometimes (contextdependently) found in the differentiated state. Cell-cycle arrest is generally considered irreversible during senescence and terminal differentiation, although cell-cycle re-entry can occur under certain conditions. Green color: active and/or present, red color: inactive and/or absent. Arrows depict connections between the cellular states. damaged proteins help identify senescent cells (Figure 1). A prominent source of protein damage is ROS, which oxidize both methionine and cysteine residues and alter protein folding and function (Ho¨ hn et al., 2017). Many protein tyrosine phosphatases (PTPs) contain cysteine residues in their active sites that can be inactivated by oxidation. This inactivation can trigger senescence by hyperactivating ERK signaling, similar to the effect of activated oncogenes (Descheˆ nes-Simard et al., 2013). High phospho-ERK levels were detected in pre-neoplastic lesions, rich in senescent cells such as melanocytic nevi and benign prostatic hyperplasia (BPH) (Descheˆ nes-Simard et al., 2013), and are a characteristic of therapy-induced senescence (Haugstetter et al., 2010). The PTP oxidation pattern can be revealed by a monoclonal antibody that recognizes oxidized cysteine (Karisch et al., 2011).
ROS, in the presence of metals, can carbonylate proline, threonine, lysine, and arginine residues. Protein carbonylation exposes hydrophobic surfaces, leading to unfolding and aggregation, and protein carbonyl residues can be specifically detected using antibodies (Nystro¨ m, 2005). Moreover, carbonyl residues can react with amino groups to form Schiff bases, contributing to protein aggregation. Subsequent cross linking with sugars and lipids forms insoluble aggregates, termed lipofuscin from the Greek ‘‘lipo’’ meaning fat and ‘‘fuscus’’ meaning dark. Lipofuscin can be visualized in lysosomes by light microscopy or a histochemical method using a biotinylated Sudan Black B (SBB) analog (GL13) (Evangelou et al., 2017). The latter is emerging as another indicator of senescent cells in culture and in vivo (Evangelou et al., 2017; Gorgoulis et al., 2018; Myrianthopoulos et al., 2019). It should be noted that damage accumulation continues, even when cell division ceases, and can continue for months or even years.
Most protein oxidative damage is not reversible, and degradation by the ubiquitin proteasome system (UPS) or autophagy often eliminates these proteins. As the UPS (Descheˆ nes-Simard et al., 2013) and autophagy are active in senescent cells, they could prove to be useful in characterizing the senescent state (Ogrodnik et al., 2019a). Similarly, PML bodies act as sensors of ROS and oxidative damage (Niwa-Kawakita et al., 2017) and can also be non-exclusive biomarkers of cellular senescence (Vernier et al., 2011).
- Lipid Damage. Lipids are essential for cell membrane integrity, energy production, and signal transduction. Some age-related diseases are characterized by altered lipid metabolism, resulting in lipid profile changes (Ademowo et al., 2017). Although senescent cells are marked by changes in lipid metabolism, it is unclear how this contributes to the senescent phenotype (Figure 1).
Mitochondrial dysfunction during senescence can result in ROS-driven lipid damage, lipid deposits (Correia-Melo et al., 2016; Ogrodnik et al., 2017), and lipofuscin accumulation (Gorgoulis et al., 2018). Apart from oxidation, lipid-derived aldehyde modifications (e.g., 4-hydroxy-2-nonenal [4-HNE]) have been reported in senescent cells (Ademowo et al., 2017; Jurk et al., 2012).
Lipid accumulation in senescent cells can be visualized using various commercial dyes and assays (Ogrodnik et al., 2017) or immunostaining for lipid associated proteins such as perilipin 2 (Ogrodnik et al., 2017). Importantly, genetic or pharmacological clearance of senescent cells in obese and aging mice reduced lipid deposits in liver (Ogrodnik et al., 2017) and brain (Ogrodnik et al., 2019b).
.../...
