Telomeres: beacons of autocrine and paracrine DNA damage during skin aging

ABSTRACT

Cellular senescence is an irreversible cell cycle arrest, which can be triggered by a number of stressors, including telomere damage. Among many other phenotypic changes, senescence is accompanied by increased secretion of pro-inflammatory molecules, also known as the senescence-associated secretory phenotype (SASP). It is thought that accumulation of senescent cells contributes to age-associated tissue dysfunction partly by inducing senescence in neighboring cells through mechanisms involving SASP factors. Here, we will review evidence suggesting that telomeres can become dysfunctional irrespectively of shortening, and that this may be a mechanism-driving senescence in post-mitotic or slow dividing cells. Furthermore, we review recent evidence that supports that senescent melanocytes induce paracrine telomere damage during skin aging, which may be the mechanism responsible for propagation of senescent cells. We propose that telomeres are sensors of imbalances in the cellular milieu and act as beacons of stress, contributing to autocrine and paracrine senescence.

Introduction

Cellular senescence is a state of irreversible cell-cycle arrest which is associated with a number of distinctive phenotypes, such as the development of a pro-inflammatory secretome also known as the senescence-associated secretory phenotype (SASP) [1]. Although senescence has been positively implicated in physiological processes such as wound healing and embryonic development [2, 3], accumulation of senescent cells has been causally linked to age-related tissue dysfunction and pathologies [4,5]. The negative effects of senescent cells are thought to be largely due to chronic exposure to SASP factors, which can induce paracrine senescence in healthy neighboring cells [6] and impair the regenerative capacity of cells in vivo [7]. Therefore, there is currently great interest in understanding the mechanisms involved in senescence and identifying interventions to target senescent cells as a therapy against age-related diseases.

One of the mechanisms that can trigger cells to undergo the senescence program is telomere dysfunction. Telomeres are found at the ends of linear chromosomes and are associated with a group of proteins collectively known as the shelterin complex. These proteins play a role in the establishment of a lariat-like structure known as the T-loop, which physically shields the exposed ends of chromosomes, preventing them from being recognized as double-stranded breaks (DSBs) [8]. However, with each round of cell division, telomeric repeats are lost, and this is believed to result in loss of shelterin components (i.e. telomeres become uncapped). When telomeres reach a critical length, the T-loop conformation can no longer be maintained and consequently, a DNA damage response (DDR) is activated at the exposed telomere ends, triggering senescence [9]. Other factors, such as oxidative stress, can also accelerate the rate of telomere shortening and contribute to senescence [10]. Although shortening is an important contributor to telomere dysfunction-induced senescence, a number of recent reports indicate that a persistent DDR can also be activated at telomeres in a length-independent manner both in senescence cells in vitro and in aging mammalian tissues [11–14].

In the remainder of this article, we will explore the relationship between telomere dysfunction, senescence, and aging with a specific focus on the skin.

The role of telomere dysfunction in melanocyte senescence and its role in skin aging

Skin aging occurs as a result of both intrinsic and extrinsic factors, which combined ultimately affect the structural integrity and function of the skin [15]. Aged skin displays a number of phenotypic changes, such as atrophy of the dermal and epidermal layers, flattening of the epidermal-dermal junction, loss of collagen and degeneration of elastic fiber network [16] (Figure 1). Moreover, the number of epidermal cells decreases and proliferation of basal keratinocytes is reduced, contributing to age-related epidermal thinning [16]. Senescent cells have been shown to accumulate in skin with age [17,18]; however, the role of senescent melanocytes in skin aging remained understudied.

Figure 1.

Senescent melanocytes contribute to skin aging characteristics by inducing paracrine telomere damage. Aged skin is characterized by a number of phenotypic changes, such as epidermal and dermal thinning, loss of subcutaneous fat, flattening of the epidermal-dermal junction, and disorganization and loss of collagen and elastin networks. These changes ultimately compromise the structural integrity and function of different skin compartments. In our recent-published study, we found that senescent melanocytes that accumulate in human epidermis with age can induce paracrine telomere damage in surrounding cells. This in turn impairs keratinocyte proliferation, contributing to age-associated epidermal thinning.

Senescent cells have been observed in different cell types in the skin. For instance, it has been reported that with age, senescent fibroblasts accumulate mostly in the dermis [17,19]. On the other hand, we and others observed [20,21] that the senescence marker p16^INK4A was only expressed by melanocytes in the epidermis of aged human skin. This finding led us to investigate the mechanisms underlying melanocyte senescence and the contribution of senescent melanocytes to skin aging characteristics. This observation was intriguing, particularly if one considers that melanocytes divide very little throughout lifespan and constitute 5% to 10% of the cells in the basal layer of the epidermis. It was therefore unlikely that melanocyte senescence occurred due to replicative exhaustion and telomere shortening.

In our recently published study [20], we extensively characterized melanocyte senescence in human skin in vivo and found that they express a number of senescence markers such as increased p16^INK4A and loss of HMGB1 with age. Moreover, by performing immuno-FISH analysis combining immunofluorescence against DNA damage proteins and a telomere-specific PNA probe, we found that the frequency of dysfunctional telomeres was significantly higher in melanocytes in aged skin. Interestingly, we did not find evidence of age-dependent telomere shortening in these cells. In fact, our data show that telomeres co-localizing with DNA damage proteins were not preferentially shorter, suggesting that telomere shortening is not the main cause of telomere dysfunction in aged melanocytes.

We also demonstrated that conditioned medium from senescent melanocytes was sufficient to induce telomere dysfunction and reduce proliferation of dermal fibroblasts in vitro, suggesting that SASP components secreted by senescent melanocytes mediate such adverse paracrine effects. Mechanistically, we identified that the SASP component, IP-10, secreted by senescent melanocytes triggers CXCR3-dependent mitochondrial reactive oxygen species (ROS) production in neighboring cells, consequently inducing paracrine telomere damage. While ROS can also lead to damage in other regions of the genome, the significance of paracrine telomere damage is that when damage occurs at telomere regions it induces a persistent DNA damage response due to inhibition of DNA repair mechanisms [11,13,14].

Finally, we utilized 3D human epidermal equivalents, which consisted of either young or senescent melanocytes cultured with keratinocytes, ultimately forming a highly differentiated model of human epidermis. Consistent with our previous findings, keratinocytes in the presence of senescent melanocytes displayed increased numbers of dysfunctional telomeres, increased markers of senescence and decreased proliferation. As a result, epidermal atrophy, a skin-aging characteristic, was observed in 3D models containing senescent melanocytes. Interestingly, clearance of senescent melanocytes by treatment with the senolytic drug ABT737 prevented paracrine senescence and epidermal thinning. Similar results were obtained by reducing mitochondrial ROS production with the mitochondrial-targeted antioxidant MitoQ, suggesting a role for oxidative stress in mediating paracrine telomere dysfunction and senescence. Overall, we demonstrated that senescent melanocytes contribute causally to age-related epidermal atrophy by inducing paracrine telomere damage and senescence in surrounding keratinocytes (Figure 1).

This work suggests that therapies targeting senescent cells may be effective in improving aging phenotypes in the skin. Consistently, elimination of senescent cells using the senolytic drug ABT737 rescued proliferation of hair follicle stem cells in a mouse model where senescence can be induced in the basal layer of the epidermis [22]. Senostatic drugs such as rapamycin, which has been shown to reduce both the SASP and mitochondrial ROS [23,24] have been shown to reduce senescence markers in skin and improve several parameters of skin aging in humans [25].

Our study supports the concept that telomeres act as molecular sensors of stress, a process that may occur irrespectively of their length. In the next section, we will explore and discuss some of the literature which supports this hypothesis.

Telomeres as molecular sensors of stress

Telomeres serve as protective structures at the ends of chromosomes, and so perhaps not surprisingly, they are extremely sensitive to intrinsic and extrinsic stressors that affect the health status of the cell. For example, mild oxidative stress has been shown to accelerate telomere shortening and induce cellular senescence, which could be counteracted by the mitochondrial-targeted antioxidant, MitoQ [26]. It has been proposed that such susceptibility to oxidative stress is due to the high guanine content of telomeres, which can easily undergo oxidative modifications [27]. In fact, studies have shown that mild oxidative stress induces single-stranded breaks preferentially at telomeres, resulting in incomplete replication of telomeric DNA and thus telomere shortening [28,29]. Consistent with this, a recent study demonstrated that inducing 8-oxoG, a common oxidative DNA lesion, specifically at telomeres caused replication stress and subsequently loss of telomeric repeats and reduced proliferation, indicating that oxidative stress is a major contributor to telomere dysfunction [30]. Furthermore, damage at telomeres is less efficiently repaired when compared to the bulk of the genome, mainly because proteins of the shelterin complex inhibit components of the DNA repair machinery in an effort to prevent telomere fusions [31,32]. Non-homologous end joining (NHEJ) is an important mechanism of repair of double-stranded breaks and studies have reported that components of the shelterin, such as TRF2 and its binding partner RAP1 are necessary to prevent telomeric DNA fusions by inhibiting DNA-PK and ligase-IV mediated NHEJ [33]. As a consequence, telomeric lesions are extremely long-lived, and in fact have been shown to persist for several months both in vitro and in vivo [13 14,]. In addition, our group has demonstrated that the majority of long-lived DNA damage foci are located at telomeres in cells undergoing stress-induced senescence [13]. Interestingly, we and others have found that expression of telomerase is insufficient to rescue stress-induced telomere dysfunction and senescence [13,14].

Persistent DNA damage signaling is believed to be important for the initiation and maintenance of senescence [34], as DDR pathways converge into downstream effectors such as the cyclin-dependent kinase inhibitors p16 and p21, reinforcing the cell-cycle arrest [10]. Therefore, dysfunctional telomeres represent important sources of persistent DDR signaling that contributes to the induction and stabilization of senescence. Indeed, telomere damage accumulates in human fibroblasts during replicative [35], stress- [13,14], and oncogene-induced senescence in vitro [36], and in many tissues in mice, primates and humans in vivo [13,14,20,37], suggesting that telomere dysfunction may also contribute to the decline in tissue function that occurs with age. While evidence suggests that DSBs persist at telomere regions, it is still unclear if other types of lesions are also persistent at telomeric sites. For instance, cyclobutane pyrimidine dimers (CPD), which have been shown to interfere with DNA replication and transcription, seem to be more rapidly repaired at telomere regions than the bulk of the genome. This work suggests that nucleotide excision repair is important for preserving telomere integrity [38].

Consistent with a role for telomeres as stress sensors, population studies have suggested an association between psychological stress and shorter telomeres. It has been shown that both chronic and perceived stress are associated with reduced telomere length, lower telomerase activity and increased oxidative stress in peripheral blood mononuclear cells (PBMCs) from healthy women [39]. Consistent with this, a link between chronic stress and shorter telomeres in PBMCs has been reported in caregivers of Alzheimer’s disease patients who also displayed signs of impaired immune function [40]. Interestingly, some studies suggest that stresses experienced in early life may also affect the rate of telomere shortening. For example, it was shown that adults with a history of childhood abuse, such as physical and emotional neglect, had decreased telomere length in leukocytes in comparison to those who did not report maltreatment [41]. An association between prenatal stress and decreased leukocyte telomere length has also been observed, where offspring of mothers who experienced psychosocial stress during pregnancy had shorter telomeres in adult life [42]. Furthermore, a longitudinal study reported accelerated telomere shortening in buccal cells of children who experienced two or more types of violence between the ages of 5 and 10 y, suggesting a link between cumulative childhood stress and telomere maintenance [43]. These observations raise the possibility that stress-associated changes in telomere integrity in early life may contribute to the risk of disease development in adulthood. In fact, studies have shown that short telomeres are a risk factor for age-related diseases such as cancer, cardio-metabolic dysfunction, and diabetes [44–46]. Mechanistically, it is unclear how psychological stress accelerates the rate of telomere shortening. One possibility is that chronic stress induces a state of persistent physiological stress arousal and subsequently sustains high levels of stress hormones, such as cortisol [47]. Experimentally, cortisol has been shown to increase the susceptibility of neurons to oxidative stress [48] and to reduce telomerase activity in human T lymphocytes [49]. Therefore, it is possible that stress-induced cortisol contributes to accelerated telomere shortening by increasing oxidative stress and lowering telomerase activity, although further studies are needed to corroborate these associations.

Although telomere shortening is an important contributor to the initiation of senescence, a number of studies have now demonstrated that telomere dysfunction can occur irrespectively of length. Persistent DDR signaling has been observed at telomeres independently of shortening in response to genotoxic stress in human fibroblasts in vitro and mouse neurons in vivo [14]. Accumulation of DNA damage at longer telomeres has also been reported during the aging process in vivo. An age-dependent increase in telomere-associated damage, which occurred irrespectively of length, was shown in the gut and liver of mice [13,37], and in hippocampal neurons and liver of baboons [14]. In addition, length-independent telomere damage was reported in small and large airway epithelial cells in the lung of COPD [12] and bronchiectasis patients [50], respectively, which also showed increased markers of senescence, suggesting that telomere damage might contribute to tissue dysfunction by inducing senescence. Furthermore, length-independent telomeric damage has also been observed in cells undergoing replicative senescence [35] and in melanocytic nevi, which are a model of oncogene-induced senescence [36], demonstrating that telomeres that are not critically short can also contribute to persistent DDR signaling in senescence.

Recently, length-independent telomere damage has been suggested as a mechanism-driving senescence of slow dividing and/or post-mitotic cells. In our study, we observed that terminally differentiated human melanocytes, which have low proliferative capacity in vivo, accumulate dysfunctional telomeres with age without significant shortening [20]. Activation of DDR signaling at telomeres that are not critically short has also been shown in other non-dividing cells such as neurons [51] and cardiomyocytes [11]. In the latter, it was proposed that increased mitochondrial dysfunction and ROS contributed to persistent telomere-associated damage, consequently driving cardiomyocyte senescence [11]. Interestingly, it has been recently shown that ROS produced in response to mitochondrial dysfunction-induced DSBs primarily at telomeres, suggesting that mitochondria-derived ROS plays a major role in telomere dysfunction [52]. Therefore, in slow-diving or post-mitotic cells, where extensive loss of telomeric repeats is unlikely to occur, telomeres might act as signaling hubs, conveying the status of upstream cellular functions, so that when homeostasis is severely compromised, a DDR is activated at chromosome ends in a length-independent manner and senescence is initiated.

Increased oxidative stress may also contribute to damage at longer telomeres by inducing uncapping. It has been shown that oxidative damage disrupts the binding of certain shelterin proteins [53], which may contribute to telomere uncapping in a length-independent manner. Indeed, ROS has been shown to play an important role in the initiation and maintenance of the senescent phenotype [24,26,54–56]. Increased levels of ROS have been observed in replicative, stress-, and oncogene-induced senescence [26,55,57], and treatment with antioxidants or other interventions that reduce oxidative stress extends the replicative capacity of cells [26,58]. Nonetheless, studies have reported that shelterin proteins are retained in a fraction of dysfunctional telomeres in senescent cells [14,35,36], suggesting that uncapping is not a limiting factor for DDR activation at telomeres.

The relationship between ROS and telomere maintenance is likely to be more complex. Interestingly, it has been shown that under conditions of oxidative stress, telomerase shuttles to the mitochondria where it improves mitochondrial function and reduces ROS through yet unclear mechanisms [59 60,]. However, the relative impact of non-canonical vs. canonical functions of telomerase during aging and disease warrants further investigation.

The role of telomeres in bystander senescence induction

Increasing evidence suggests that senescent cells contribute to age-related tissue dysfunction by paracrine mechanisms that negatively affect surrounding cells [4,6,7,61]. Indeed, senescent cells have been shown to induce senescence in nearby cells [6,20,61]. Such paracrine proliferation arrest was dependent on p16 and p21, which are major senescence effectors, thus fully resembling the senescence response [6]. A number of studies also provide evidence that senescence can be transmitted to neighboring cells in vivo. For example, clusters of senescent hepatocytes have been shown in the liver of aged mice [61] and in a mouse model of chronic inflammation, which displays accelerated aging phenotypes [37]. Further supporting this notion, in a recent study where senescent cells were transplanted into the skeletal muscle and skin of immunocompromised mice, it was shown that markers of senescence were increased in dermal fibroblasts and myofibers in proximity of the transplanted senescent cells [62]. Another study showed that transplantation of a small number of senescent cells into young mice induced telomere-associated DNA damage foci and other senescent markers in normal host cells [63]. Importantly, this study showed that transplantation of senescent cells was sufficient to cause persistent physical dysfunction in young mice. Evidence of paracrine senescence has also been shown in humans, where stromal cells positive for p21 were observed in the vicinity of human colon sessile serrated adenomas (SSAs), pre-malignant lesions which result from BRAF mutations that trigger oncogene-induced senescence [6]. These results suggest that senescent cells are capable of spreading senescence to healthy, neighboring cells in vivo, contributing to the accumulation of senescent cells in tissues with age.

It is still unclear which specific factors contribute to paracrine senescence; however, it is likely that this will depend on the cell-type as well as the senescence-inducing mechanism. Acosta and colleagues identified that multiple SASP factors play a role in paracrine senescence in the context of oncogene-induced senescence [6]. Inhibition of VEGFR2/FLT3, TGFBR1, and CCR2 receptors was able to partially rescue paracrine senescence. In our recently published study, we showed that IP-10 released by senescent melanocytes plays a role in paracrine senescence; however, this does not exclude that other SASP factors may be involved [20]. Small extracellular vesicles have also been shown to mediate paracrine senescence, a process partially involving interferon-induced transmembrane protein 3 (IFITM3) [64]. The length of exposure to the SASP may also play a role in paracrine senescence. For instance, Ritschka et al. 2017 reported that keratinocytes transiently exposed to the SASP show increased regenerative capacity [7]. However, chronic exposure to the SASP led to the induction of senescence in these cells [7].

Although the mechanisms contributing to paracrine senescence have not been fully elucidated, it has been demonstrated that both direct cell-cell contact and SASP factors released by senescent cells can induce paracrine DNA and telomere damage in neighboring cells [6,20,61,65,66]. Mechanistically, we and others have observed that soluble factors secreted by senescent cells trigger an increase in ROS production in surrounding cells, consequently inducing bystander DNA damage formation [20,65,66]. One study showed that exposure to senescent cells generated the formation of large DNA damage foci in adjacent fibroblasts [61]. Large foci persist for longer than smaller ones [61], and in fact, the majority of long-lived foci have been shown to be located at telomeres [13]. It is therefore possible that dysfunctional telomeres in bystander cells provide persistent DDR signaling that contributes to paracrine senescence. In agreement with this, we demonstrated that senescent melanocytes induce paracrine telomere damage and limit the replication of keratinocytes in 3D epidermal equivalents [20]. These effects were rescued by treatment with the mitochondrial-targeted antioxidant MitoQ [20], suggesting that ROS-mediated paracrine telomere dysfunction plays an important role in the bystander effect of senescent cells. Another possibility is that ROS produced by senescent cells might leak out of membranes and damage telomeres directly [61]. One group has shown that paracrine DNA damage required direct cell-to-cell contact and was mediated by ROS derived from senescent cells [61]. In contrast, we and others [11,20,65] have observed that conditioned media from senescent cells is by itself capable of inducing telomeric dysfunction, suggesting that soluble factors also play a role in paracrine damage induction (Figure 2). It is important to note that Nelson et al. [61] measured the total number of DDR foci rather than telomeric DNA damage, which may explain the discrepancies in results. In summary, it is becoming increasingly evident that telomeres are not only molecular clocks, determining the number of replication cycles a cell is capable of undergoing. Instead, telomeres are extremely sensitive to intra- and extracellular imbalances, and also act as beacons, signaling changes in the microenvironment that might severely impair cellular function. By doing so, telomeres prevent proliferation of cells that have accumulated significant damage. However, as we age, accumulation of cells containing damaged telomeres can lead to detrimental effects, by activating senescence pathways, particularly the SASP, which can result in additional telomere damage in surrounding cells.

Figure 2.

Telomeres act as beacons of intra- and extracellular stresses. Telomeres are targets of intracellular oxidative stress, which increases in senescent cells due to mitochondrial dysfunction. In turn, accumulation of damage at telomeres provides persistent DNA damage signaling that reinforces the cell cycle arrest and senescence. SASP factors released by senescent cells can also induce paracrine telomere damage mediated by increased mitochondrial ROS generation in surrounding cells. Additionally, ROS can diffuse through the membrane of senescent cells and directly induce paracrine damage at telomeres. Recently, small extracellular vesicles (sEVs) derived from senescent cells have also been shown to induce paracrine senescence. We hypothesize that as well as acting as sensors of autocrine cellular homeostasis, telomeres also signal changes in the microenvironment, serving as beacons of paracrine damage.

Funding Statement

This work was supported by the Biotechnology and Biological Sciences Research Council [BB/L502066/1]; Biotechnology and Biological Sciences Research Council [BB/K017314/1]; Ted Nash Foundation [N/A].

Acknowledgments

SV and JFP would like to acknowledge the Ted Nash Long Life Foundation, Unilever, and BBSRC [grants BB/L502066/1; BB/K017314/1].

Disclosure statement

No potential conflict of interest was reported by the authors.