Exercise As a Therapy to Maintain Telomere Function and Prevent Cellular Senescence

Abstract

Exercise transiently impacts the expression, regulation, and activity of TERT/telomerase to maintain telomeres and protect the genome from insults. By protecting the telomeres (chromosome ends) and the genome, telomerase promotes cellular survival and prevents cellular senescence. By increasing cellular resiliency, via the actions of telomerase and TERT, exercise promotes healthy aging.

Summary for Table of Contents:

Exercise increases telomerase reverse transcriptase (TERT) expression and telomerase activity, thereby promoting telomere maintenance and preventing cellular senescence and aging phenotypes.

Introduction

Exercise increases median survival age and health span duration (1, 2). Consequently, determining the molecular mechanisms of how exercise ameliorates the aging process and increases health span is a primary point of inquiry for modern science. One key mechanism of aging is cellular senescence (3). Removal of senescent cells or prevention of cellular senescence attenuates aspects of the aging process and the progression of age-related diseases such as CVD and cancer (3). Cellular senescence can be induced by stress (e.g., DNA damage) and replicative senescence. Interestingly, both DNA damage and replication-based stresses can mechanistically converge on telomeres (3). For instance, progressive telomere shortening is a hallmark of aging and a driver of cellular senescence (4). In addition, telomere length independent DNA damage signaling can also induce cellular senescence (5). Telomeres are maintained by an enzyme, telomerase, and are protected by a series of telomeric DNA-binding proteins, shelterin (6). It is through telomerase and shelterin function that nuclear DNA-integrity is preserved and cellular senescence is avoided. As such, telomerase expression and repression of DNA damage signaling at the telomere by shelterin are powerful attenuators of the aging process.

It is our working hypothesis that exercise (a stimulus with wide-ranging effects) is mechanistically linked to maintained telomere length and prevention of DNA damage at telomeres by the activating telomerase, upregulating TERT mRNA and protein isoforms, and increased shelterin expression. This results in reduced cellular senescence with aging and improved human health. The purpose of this review is to highlight research examining the relation between exercise and telomere biology. We will discuss how exercise i) attenuates age-related telomere shortening ii) regulates telomerase, TERT expression, and TERT alternative splicing (AS) and iii) impacts shelterin expression.

What are telomeres and why are they important?

Telomeres are repeated DNA sequences (5’-TTAGGG_n-3’) located at the ends of linear chromosomes (7). A six-member protein complex called shelterin is associated with telomeric DNA that facilitates the formation of a loop structure (the t-loop) to protect chromosome ends from DNA damage detection and repair (6). The shelterin proteins are telomere repeat binding factors 1 and 2 (TRF1 and TRF2), protection of telomeres (POT1), repressor/activator protein 1 (RAP1), telomere repeat binding factor 1 interacting protein 2 (TIN2), and adrenocortical dysplasia homolog Shelterin Complex Subunit and Telomerase Recruitment Factor (ACD or TPP1, referred to as TPP1 from here on) (Fig. 1A). Progressive telomere shortening occurs in all dividing normal cells due to incomplete lagging strand DNA synthesis, oxidative damage, exonucleolytic processing, and other factors (Fig. 1B) (8). This telomere shortening phenomenon is an important dictator of mitotic potential; when telomeres reach a certain length and DNA damage-signaling occurs, cellular senescence is initiated. However, it is important to note that it is not mean telomere length that initiates cellular senescence. Rather cellular senescence is initiated when one or a few telomeres reach a critically short length and can no longer prevent DNA damage signaling (8). Activation of the DNA damage response at telomeres results in the formation of telomere-associated DNA damage response foci (TAFs) or telomere-induced DNA damage foci (TIFs), which are markers of cellular senescence in cells and tissues (5). Telomere/replication-induced cellular senescence is thought to be an important barrier to tumor formation in replicating cells of large long-lived mammals (i.e., Hayflick limit, (9–11)). Since cancer cells rely on continued proliferation, a telomere maintenance mechanism is critical to their survival (continued replication without maintaining telomeres in precancerous or cancerous cells would lead to cell death called crisis(12)). Therefore, somatic cells with critically short telomeres and telomeres with DNA damage signaling undergo cellular senescence to block tumor formation (13). Unfortunately, while telomere shortening is an important barrier to tumor formation, senescent cells secrete proinflammatory cytokines, known as the senescence-associated secretory phenotype (SASP) (14). SASP can be both promoted by the DNA damage response as well as promote DNA damage and TAFs/TIFs formation (5). Furthermore, the accumulation of senescent cells is a common feature of aging, while their clearance results in reduced age-related phenotypes and disease in a variety of contexts (3). Thus, maintenance of telomere length and reduction in DNA damage signaling at telomeres are critical to prevention of cellular senescence to slow aspects of the aging process. In addition to age-related cellular senescence and cancer, the biomedical relevance of telomere dysfunction underscores a myriad of pathologies including genetic diseases (i.e., short telomere syndromes or telomeropathies) (15), cardiovascular diseases, muscular dystrophy, neurodegenerative diseases, metabolic diseases (e.g., type II diabetes), lung diseases, kidney dysfunction, and skeletal health (5).

Figure 1. Telomere and telomerase biology and their role in cellular senescence.

Telomeres are specialized DNA structures at the end of chromosomes. Shelterin is a protein complex that coordinates telomere structure formation. Shelterin consists of six proteins including TRF1, TRF2, RAP1, TIN2, POT1 and TPP1. TRF1 and TRF2 bind to double-stranded DNA, while POT1 binds to single-stranded DNA. RAP1 interacts with TRF2, and TIN2 interacts with TRF1 and TRF2. In telomere-looping structure, TPP1 interacts with TIN2 and POT1 (A). When telomeres are long and intact, the telomeres are functional and inhibit the DNA damage response that induces replicative senescence in cells. However, when telomeres become critically short by cell division and/or DNA damage, the DNA damage response (DDR) is induced and cells enter cellular senescence (i.e., irreversible cessation of cell division). Telomeric DNA damage-induced DNA damage response can also be telomere length independent and result in telomere associated DNA damage foci (TAFs) or telomere damage induced foci (TIFs) which can also induce cellular senescence, even in post-mitotic cell types such as skeletal muscle and cardiac muscle (B). Telomerase is a reverse transcriptase enzyme that can synthesize telomere repeats. The two minimal essential components for telomerase activity are the protein catalytic component, telomerase reverse transcript (TERT), and the RNA template (telomerase RNA component, TERC or TR) (C). Abbreviation: TRF1 (Telomere Repeat binding Factor 1), TRF2 (Telomere Repeat binding Factor 2), RAP1 (Repressor / Activator Protein 1), POT1 (Protection of Telomere 1), TPP1 (adrenocortical dysplasia homolog shelterin complex subunit and Telomerase Recruitment Factor), TIN2 (telomere repeat binding factor 1 interacting protein 2). Created with BioRender.com.

Telomerase maintains and elongates telomeres

Telomerase is a ribonucleoprotein enzyme that maintains and elongates telomeres. It consists of two core elements: telomerase reverse transcriptase (TERT, protein catalytic subunit) and telomerase RNA component (TERC, RNA template for telomere synthesis; Fig. 1C) (16). Telomerase activity is present during fetal development, but its activity is tightly repressed in adult cells and tissues which helps form the tumor suppressive mechanism of telomere shortening (17). In many types of cancers, telomerase maintains telomeres and allows for continued growth of cancer cells. However, robust evidence from in vitro and in vivo studies have indicated that TERT expression and telomerase activity alone (i.e., without disruption of tumor suppressors) are not oncogenic (8, 18). In fact, gene therapy studies indicate that transient delivery of the TERT gene to reconstitute telomerase activity resulted in rejuvenation and reduced cellular senescence (19, 20). Most importantly to the topic of this review, the expression of TERT and telomerase activity can be induced by various physiological stimuli such as growth factors, hormones, and physical exercise (21–23).

Since most of the studies performed in the telomere and exercise field have been performed in rodent models, it is important to note a few differences between rodent telomere biology and human telomere biology. Telomeres are longer in most rodent strains besides CAST/EiJ mice (24). While telomeres do shorten in rodents due to the end replication problem and various other stressors (i.e., DNA damage), the long telomeres of rodents are rarely limiting for cellular replication and are hypothesized to act as oxidative stress sponges in rodent rather than a limit for cell division (25). According to Peto’s paradox (26), larger organisms should have a higher rate of cancers due to increased cell divisions needed to achieve the larger size. With more cycles of DNA replication there is an increased risk of mutations and thus increased risk for cancer-promotion. However, this is not the case, as it has been observed that larger organisms tend to be protected from cancer, while most smaller organisms are cancer prone (26). While several mechanisms, such as increased copies of tumor suppressor genes (tumor protein 53, p53), have been observed to be related to how larger organisms are protected from cancer, telomerase and telomere biology are also important in this process as well. Telomerase activity is reduced with larger body size (larger organisms such as humans tightly regulate telomerase, while smaller organisms like mice have detectable telomerase in most somatic tissues), leading to the idea that telomerase suppression and telomere shortening is integral to tumor suppression in larger organisms (27). Another aspect of this idea is the difference in life course and reproductive strategies between mice and humans. Since the life course strategy of rodents is to promote early growth and reproductive maturity to promote species propagation in lieu of avoiding cancer development, telomere length and telomerase would therefore not be regulated like larger-long lived species such as humans (28). The life course strategy of humans is to grow slowly, reproduce later in life and prevent cancer to allow for enough time to raise a human child to sexual maturity/independence (28). Telomerase and TERT are detectable in most rodent tissues at low levels, but undetectable in most human tissues. Mouse Tert is expressed mainly as the telomerase active TERT while human TERT is alternatively spliced with telomerase active TERT being a minority isoform(29). These differences are important to consider regarding how exercise impacts telomere biology in mice and men.

Exercise maintains telomeres in various tissues

Two seminal studies highlighted that moderately active adults across the age-span tended to have longer telomeres (30, 31). Cherkas et al. 2008 measured leukocyte telomere length and collected self-reported physical activity levels in a large cohort of individuals, including twins. These data indicated a positive linear relation between telomere length and physical activity level (30). Similarly, Ludlow et al. 2008 reported a positive linear relation between peripheral blood mononuclear cells (PBMC) telomere length and exercise energy expenditure up to moderate levels of physical activity (31). However, a subgroup of self-identified as “Master’s Athletes” that performed greater than 8 hours per week of exercise had shorter telomeres compared to age-matched moderately active participants (31). This observation led us to conclude that the relation between exercise and immune cell telomere length could have an “inverted U” shape. There appears to be a “Goldilocks level” of physical activity/exercise that results in telomere maintenance; too much and too little exercise result in accelerated telomere shortening. Several other reports have replicated this “inverted U” finding (32), while others have either observed a linear relationship between exercise amount (33, 34), or no relation between telomere length and exercise amount (35, 36). Potential explanations for these discrepant results could be recall bias of physical activity, degraded DNA samples, and other physiological or psychological variables that were not statistically controlled for, such as psychological stress or cortisol levels (37). Regardless, these inconsistent results point to the need for more carefully controlled studies and the need for standardized reporting of exercise variables and telomere length measurements.

Beyond survey-based physical activity research presented above, the gold-standard measure of physical fitness, maximal oxygen consumption (VO_2max) has been associated with telomere length in several studies. For instance, measuring telomere length between athletes and non-athletes finds a positive association between VO_2max and telomere length. Endurance trained athletes who have a greater VO_2max have been observed to have longer telomeres compared to age-matched controls in a variety of human tissues ((38)(skeletal muscle),(39) (leukocytes), (40)(leukocytes)). Implementing an endurance exercise training program has also been observed to increase VO_2max and maintain telomere length as well in a variety of human tissues ((41)(leukocytes), (42)(skeletal muscle), (43)(leukocytes)). On the other hand, two other studies have found no association between VO_2max and leukocyte telomere length (44, 45). Several factors could have led to these mixed findings such as training age (how many years of training), biological age of the subjects, training intensity (elite performance versus health and fitness), tissue type analyzed (i.e., immune cells versus skeletal muscle), and telomere length measurement techniques employed (PCR, southern blot, fluorescence in situ hybridization). Nonetheless, several meta-analyses of telomere length and exercise related variables have recently been published (35, 36, 46, 47), highlighting that there appears to be a moderately strong relationship between exercise and telomere length. Most studies relating exercise to telomere length have only measured mean telomere length, but considering that one or a few very short telomeres lead to cellular senescence, new methods that measure the quantity of shortest telomeres could produce additional insights. In summary, current literature supports that exercise most likely attenuates age-related telomere shortening, but the exact dose of exercise and impacted tissues needs further elucidation and research. Additionally, more functional outcomes such as cellular senescence markers (beta-galactosidase, p53, p21 and p16 expression, TAFs/TIFs)(48), indicators of DNA damage, expression of DNA repair proteins/enzymes, and their associations with telomere length and exercise should be documented.

Following up on our epidemiological/cross-sectional study in humans (31), we evaluated the impact of long-term exercise (44 weeks) on telomere dynamics in a wild-derived short telomere mouse model (CAST/EiJ) over the course of one year (49). We measured telomere length in skeletal muscle (gastrocnemius), cardiac muscle (ventricles), and liver from three groups of mice; sedentary young (8 weeks of age), sedentary old (52 weeks of age) and older active mice (52 weeks of age with access to running wheel). Liver, heart, and skeletal muscle were examined for telomere length. In both liver and cardiac muscle, we observed that one-year-old sedentary mice had the shortest telomere length and that the one-year-old exercised mice were not different from the sedentary young animals. In contrast in skeletal muscle, the young animals and the sedentary one-year old animals had similar telomere lengths, while the one-year-old exercised group exhibited significant telomere shortening, indicating a tissue-specific response to the exercise stimulus. Similar to these observations, Collins et al. 2003 observed overtrained athletes had shorter skeletal muscle telomeres compared to age-and-training matched healthy athletes (50). In a follow up study by the same group in healthy athletes, training volume and years spent training were also associated with shorter skeletal muscle telomeres (51). These authors speculated that shorter telomeres in the more active individuals could be due to the addition of muscle fiber nuclei with shorter telomeres following satellite cell replication to repair exercise-induced muscle fiber injuries (52). Alternatively, rather than satellite cell replication, which is unlikely to occur in response to voluntary wheel running (53) or typical endurance running training in humans, exercise-induced oxidative stress could have impacted telomere length and telomere dynamics in the skeletal muscles. We followed this rationale because the nucleotide guanine (G), which telomeres are rich in (5’-TTAGGG), is very sensitive to oxidative damage forming 8-oxoguanine (8-oxoG). Oxidized guanine has been shown to accumulate in telomeric regions, is associated with induction of cellular senescence phenotypes, and increases in aging humans (54, 55). Thus, we hypothesized the oxidative stress induced by bouts of exercise, especially the initial bouts of exercise prior to antioxidant system adaptations, could be particularly damaging to telomeres. We followed up on this hypothesis by pursuing studies in isolated muscle fiber cultures in two strains of mice (56). Here, we treated isolated flexor digitorum brevis muscle fibers with a common oxidizing agent (hydrogen peroxide) for five days and subsequently analyzed telomere length. In muscle fibers from both mouse strains, we observed oxidant-induced telomere shortening that was partially rescued by the antioxidant N-acetyl-cysteine. These data indicate that telomere damage-induced shortening could occur by exercise-induced oxidative stress rather than a cell replicative mechanism for telomere shortening in post-mitotic muscle fibers. Further research into this potential mechanism is needed, but it is important to highlight several recent lines of evidence that indicate the importance of telomeres in senescence of long-lived post-mitotic cells such as skeletal muscle (57) and cardiac muscle (58). These apparent discrepancies between telomere shortening induced by exercise and the healthy benefits of exercise need further study.

We speculate that exercise is unlikely to induce telomere shortening that would result in telomere-induced DNA damage senescence. One idea is that telomeres provide a “buffer zone” prior to telomeres reach a critically short length and triggering cellular senescence. Telomere length follows a normal distribution, meaning that a shorter mean telomere length is not necessarily indicative of a greater number of critically short telomeres that could induce cellular senescence (59). To further this hypothesis, measurement of the percentage of short telomeres in muscle fiber nuclei and elucidation if exercise results in fewer TAFs/TIFs compared to the number of TAFs/TIFs in sedentary muscles could have important implications on triggering senescence in skeletal muscles. This could indicate that even if exercise shortens mean telomere length in skeletal muscle, there could be fewer critically short telomeres or damaged telomeres with DNA damage signaling in muscles from trained individuals compared to sedentary individuals, who may have longer mean telomere length but a greater percentage of critically short or damaged telomeres with DNA damage signaling thus leading to greater senescence in sedentary muscles.

Effects of exercise on telomerase activity – early studies provide clues

Measuring telomerase activity can provide interesting insights about exercise how exercise may delay or prevent telomere shortening. Many of the initial telomere biology and exercise reports focused mainly on telomere length and did not measure telomerase activity, most likely since measuring telomerase activity is challenging due to its low abundance in adult somatic cells and problematic PCR-based techniques. However, certain adult cell types such as immune cells (i.e., T cells) can express TERT and telomerase when physiologically perturbed (60). There is also strong evidence that exercise is a potent immune perturbation resulting in immune cell proliferation (61). In fact, previous studies have indicated increased telomerase activity in immune cells following exercise interventions. In 2009, Werner et al. showed that older athletes had higher levels of PBMC telomerase activity compared to both young and aged sedentary controls, while having similar telomerase activity compared to younger athletes (62). More recently, Werner et. al., 2019 showed that both acute aerobic exercise and aerobic exercise training increased telomerase activity in immune cells of study participants (43). However, whether the telomerase activity is upregulated transiently after exercise or basal telomerase activity becomes chronically higher in trained individuals even in resting state remains still unclear because the studies refenced above either were specifically designed to measure the acute effects of exercise or did not report the conditions under which the cell samples were collected. As a result, it is unknown if the participants could have exercised within a short period of time prior to the sample collections. Similarly in humans, Hagman et al. first illustrated that regular exercise training (soccer) significantly increased telomerase activity in mononuclear cells of young athletes compared to the age-matched sedentary cohort (63). In contrast to the above data, when we measured telomerase activity in PBMCs from older adults with a wide variety of physical activity levels, it was similar between sedentary and more physically active older individuals (31). In our samples, subjects came to the laboratory after an overnight fast to have their blood drawn with no other constraints on their activity. Further research should build upon these initial reports and implement more rigorous sampling criteria into their experimental design to elucidate the effects of prior exercise and during of exercise effects on telomerase activity in PBMCs. A recent meta-analysis and systematic review was performed on this topic by Denham and Sellami and indicated that both acute exercise and long-term exercise training increase telomerase activity in various tissues of mammals (63). However, having constantly increased levels of telomerase activity in cells may not be beneficial. For instance, if a cell has a checkpoint mutation or other precancerous lesion, having unregulated telomerase activity could promote cancerous formation. Telomerase activity needs to be tightly regulated in an on-off fashion to restrain uncontrolled proliferation. Since this type of telomerase regulation is observed in antigen-stimulated immune cells (60), it is therefore important to understand if telomerase regulation (i.e., ability to turn telomerase on and off again) is improved with exercise or not and not just whether telomerase activity is higher or lower. In addition, careful documentation of the cell type telomerase was measured in, duration, intensity and frequency of exercise, age of the participants, and timing after the last bout of exercise are necessary to elucidate the impact of exercise on telomerase. Clearly, additional rigorous work is needed to understand the impact of exercise on telomerase activity, particularly in immune cells of humans of different ages, following different types of training, and at different time points following exercise bouts.

Acute aerobic exercise and aerobic exercise training increases TERT expression

TERT gene expression is correlated with telomerase enzyme activity (64). Also, TERT protein and mRNA levels are extremely low, even in telomerase positive cells (65). The best estimates in telomerase activity-positive cancer cells are 1 – 40 molecules per cell of TERT mRNA (66) and 200 – 600 TERT protein molecules per cell (65). Further, reliable and robust detection of TERT with commercially available antibodies in tissue or cell lysates from non-cancerous human somatic tissues is challenging due to the abundance of TERT (65). For these reasons, many studies only measure TERT mRNA (via RT-PCR methods) and equate this with an increase in telomerase enzyme activity.

Despite these challenges several groups have documented changes in TERT mRNA levels following exercise. We previously documented that 44 weeks of exercise training in mice (CAST/EiJ mice) induced an increase of mouse telomerase reverse transcriptase (mTert) mRNA levels in skeletal muscles when compared to sedentary animals (49). In a subsequent study, we found that an acute bout of treadmill exercise (~60 – 70% of maximal speed) elicited an increase in mTert mRNA levels, concurrent with an increase of telomerase enzyme activity in the hearts of female C57BL/6 mice (67). Studies by Werner et al. similarly indicated that three weeks of voluntary wheel running was sufficient to increase protein levels of mouse TERT in rodent mononuclear cells, aortic endothelial cells, aortas, and cardiac tissue (62, 68). Werner’s team further showed that endothelial nitric oxide synthase and insulin-like growth factor-1 signaling were important for the exercise induction of mouse TERT protein expression and telomerase activity. Since these studies, several groups have utilized human primary blood samples and determined that an acute bout of aerobic exercise increased human TERT mRNA expression (21, 69).

Although the telomerase regulation field has long upheld that TERT is regulated only by transcription, recent evidence has challenged this idea. In 2012, Hrdlickova et al. demonstrated the presence of TERT transcript isoforms (via alternative RNA splicing (AS)) in normal fibroblasts thus challenging the common thought paradigm that TERT is only regulated by transcription (57). In 2016, Kim and Ludlow et al. demonstrated that in vitro aged human fibroblasts had increased TERT transcripts and epigenetic changes to the TERT promoter making it more permissive to transcriptional activation (i.e., more TERT transcripts) despite the aged fibroblasts lacking telomerase activity (70). These data strongly indicate that post-transcriptional modification (i.e., AS) of the TERT mRNAs is a key regulatory feature that dictates the functional outcome of TERT transcripts. Consequently, the groundwork was laid for new insights into TERT regulation and the role of AS following exercise.

TERT AS isoform generation

TERT is a 16-exon, 15-intron gene spanning a 42-kilobase region located on chromosome 5 (5p15.33) in humans (Fig. 2). The critical protein domains of TERT are the telomerase N-terminal domain (TEN) mainly encoded by exon 1, the telomere RNA binding domain (TRBD) encoded by exons 2 and 3, the reverse transcriptase domain (RT) encoded by exons 4 to 13, and the C-terminal extension domain (CTE) encoded by exons 14–16. Additionally, we show key intronic elements in Fig. 2 that our group has identified as being important for splicing regulation in cancer cells (71).

Figure 2. Cartoon of the TERT gene, protein domains and isoforms.

Chromosomal location, size, 16 exons, 15 introns, and cis-elements of mouse Tert (mTert) and human TERT (hTERT) are shown in panel A. mTert has significantly shorter introns and lacks intronic cis-elements (Alu elements and VNTRs) compared to hTERT. These intronic elements result in human specific regulation of TERT compared to mouse Tert (A). Two major TERT isoforms and four functional protein domains are shown. Only full-length TERT with all 16 exons can be translated into active telomerase enzyme that can synthesize de novo telomere repeats. Stop codon of full-length TERT is in exon 16. In minus beta TERT exons 7 and 8 are spliced out by alternative splicing which induces a frameshifting event and a premature stop codon located in exon 10 (B). Abbreviation: VNTR = variable number tandem repeat, TEN (telomerase N-terminal), TRBD (telomere RNA binding domain), RT (reverse transcriptase), CTE (C-terminal extension). Created with BioRender.com.

For context, splicing is the process by which non-coding information (introns) is removed from a primary RNA transcript (which contains all exon and introns of a gene) and the remaining coding information (exons) is recombined in sequential order (full-length (FL) protein-coding isoform). In the case of TERT, this corresponds to exons 1–16 (Fig. 2). Alternative splicing (AS), on the other hand, occurs when exons, and even parts of introns, are spliced together in an order different from the sequential FL isoform. AS can therefore generate multiple transcripts, known as isoforms, from the same gene. In fact, up to 95% of genes in humans generate more than two transcript isoforms (72). AS is a highly conserved processes and more than 80% of mouse genes are also AS (73). AS generates proteome diversity and contributes to how humans utilize ~20,000 genes to produce more than ~60,000 proteins (74).

During the initial cloning of human TERT mRNA, it was observed that several transcript isoforms existed (75). These initial studies resulted in the discovery of AS of the TERT RT domain, rendering the TERT transcripts and proteins unable to maintain telomeres (76–78). To date, only one transcript of TERT is known to confer telomerase activity and that is the full-length transcript containing all 16 exons in sequence. The most widely studied AS event is the skipping of exons 7 and 8 generating the minus beta TERT isoform. It is important to note that TERT AS is known to be regulated in a physiologically relevant system. For instance, during human fetal development TERT is transcribed and alternatively spliced to generate both full-length and minus beta transcript isoforms prior to week 12. In a tissue-dependent fashion, telomerase and full length TERT are extinguished between weeks 12 – 18 and only the minus beta TERT isoform remains (79). These data indicate that TERT can be transcribed while not forming active telomerase and demonstrate that AS of human TERT is a regulated process.

Since these initial studies, 21 TERT AS isoforms have been identified (80). Some isoforms have known functions such as minus beta (81, 82), INS3, and INS4 (83). However, some isoforms are degraded by RNA quality-control mechanisms, and other isoforms have not had their potential functions studied (80). Thus, it is important to consider AS when investigating TERT mRNA following exercise. No human studies have investigated AS of TERT following exercise. My laboratory has extensively studied TERT mRNA AS regulation in cancer cells (29, 64, 71) and we have recently published the first report describing how an acute aerobic exercise bout impacts TERT AS and transcriptional level changes (84) which may have important function implications for how telomere biology impacts cellular senescence as described below.

Exercise effects on TERT AS isoform generation

As our laboratory highlights, the post-transcriptional regulation of mTert is different than hTERT (29). We recently utilized long-read RNA sequencing of targeted amplicons (PCR-targeted TERT amplification) to compare the AS isoforms of TERT in human cancer cells (HeLa) against mouse immortal cells (NIH 3T3). The findings of our targeted long-read sequencing analysis revealed, for the first time, that the majority (>85%) of mTert transcripts were spliced to the telomerase activity-coding FL mTert, while most hTERT transcripts were AS to non-telomerase-coding hTERTs (>60%). These differences in mouse and human TERT regulation force researchers to carefully interpret their results from exercise interventions on TERT and telomerase when investigating both cells/tissues for either humans or rodents.

To study hTERT expression, AS, and regulation in tissues difficult to obtain from humans, we utilized a transgenic mouse model expressing the human TERT gene (hTERT) including the 5’ and 3’ regulatory sequences, exons and introns (85). This mouse model, commonly referred to as the hTERT bacterial artificial chromosome (hTERT BAC) mouse, served as an in vivo model to study the gene expression regulation of hTERT in tissues following environmental exposures. Importantly, this mouse model recapitulated the expression pattern of hTERT as observed in human tissues (i.e., repression of hTERT in somatic cells besides the testis and thymus) (84, 85). Also, the splicing patterns of hTERT were preserved indicating that the mouse splicing machinery recognized the human gene-splicing elements like the human splicing machinery does. However, several limitations should be mentioned with this mouse model. First, hTERT proteins cannot form active telomerase complexes with mouse telomerase RNA (86), thus no telomerase activity changes will be formed by changes in hTERT expression in the mouse tissues/cells. Second, telomere length regulation still occurs by endogenous mouse Tert (mTert) and mouse telomerase, limiting our ability to measure telomere length changes by interventions.

Recently, my group utilized the hTERT BAC mouse model and performed an aerobic (treadmill running) exercise bout to begin understanding how exercise impacts AS of hTERT mRNAs. The mice ran on a treadmill for 30 minutes at 60% maximum speed and tissues were collected immediately following, 1-, 24-, 48-, and 72-hours following the exercise bout (i.e., in recovery) and compared to control animals that were exposed to a motionless treadmill for 30-minutes and their tissues were immediately harvested (84). We measured FL hTERT (exons 7/8 included) and minus beta TERT (skipping of exons 7/8 indicated by 6/9 junction primers) by RT-PCR in skeletal muscle, cardiac muscle, and the brain. In comparison to the no exercise condition, we observed an overall increase in hTERT mRNA at 1-hour post-exercise in the skeletal muscle samples. Splicing ratios were also altered in the skeletal muscle with a significant increase in FL during recovery (72-hours). In contrast, the cardiac muscle (left ventricle), had a different AS pattern but a similar increase in hTERT transcripts. The percentage of FL hTERT coding transcripts in comparison to total hTERT decreased in recovery from exercise (1-, 24-, and 48-hours). Minimal changes were observed in the brain, but this does not rule out that specific regions of the brain may be better targets for future analysis (i.e., hippocampus, nucleus acumbens, or caudate). These tissue-specific responses led to the hypothesis that hTERT and its isoforms likely play specific roles in certain tissues and contexts. While further research will be needed to elucidate the functions of these TERT isoforms in cardiac versus skeletal muscle, these data support our hypothesis that TERT and telomerase are a part of the exercise induced adaptations.

Impact of exercise on shelterin

Another area of exercise and telomere biology that has been investigated is how shelterin components are impacted by exercise. Werner et al. 2008 first documented that exercise (voluntary wheel running for three weeks) impacted Terf1 (TRF1) and Terf2 (TRF2) shelterin mRNA and protein expression in cardiac tissues (68). We followed this up with tissues collected from one-year-old mice, which underwent 48 hours of exercise washout following 44 weeks of voluntary wheel running. In this research study we analyzed expression of shelterin components at the mRNA and protein levels (for Terf1 (TRF1) and Terf2 (TRF2)) in skeletal muscles, liver, and cardiac muscle (left ventricle). In the liver tissue, we observed few changes in the expression of shelterin components (49) indicating that neither age nor exercise had a striking impact on this tissue. In cardiac muscle, however, shelterin changes were more pronounced, while age overall resulted in reduced shelterin expression that was attenuated in exercised animals (54). This indicates that exercise (voluntary wheel running) results in favorable remodeling of telomere biology, likely resulting in part of the exercise-induced cardioprotective phenotype. In the skeletal muscle, we observed muscle-specific effects between the extensor digitorum longus (EDL) and plantaris muscle for expression of Terf1 (TRF1) and Terf2 (TRF2). The plantaris muscle from older sedentary animals had increased expression of TRF1 compared to the older exercised animals and young animals, while minimal changes in TRF1 levels were observed in the EDL. Interestingly, overexpression of TRF1 in dividing cells results in progressive telomere shortening (87). Therefore, exercise may help maintain telomeres by attenuating an age-related increase in TRF1 protein. The impact of exercise on shelterin expression was intriguing and we pursued mechanistic follow-up studies to investigate some of the changes we observed with the training model.

We wanted to understand the initial signaling mechanisms that could be driving the adaptive process associated with changes in telomere biology induced by exercise. To do this, we utilized an acute exercise model in mice, specifically a bout of treadmill running followed by analysis of tissues in recovery from exercise to explore how these two variables impacted expression of shelterin components. We investigated shelterin expression in three groups of mice; a sedentary control group, and two groups that underwent an intensity controlled (60% of maximal running speed on the treadmill) 30-minute bout of exercise and had tissues collected either immediately following exercise or 1-hour post exercise (88). We observed that Terf1 mRNA levels were significantly reduced immediately following exercise in the plantaris muscle, recapitulating our training-induced effects. Next, we wanted to determine what signaling mechanisms were induced following exercise and pursued studies of the mitogen activated protein kinase (MAPK) pathway (p38 MAPK, ERK, and JNK). We observed an increase of p38 MAPK activation (phosphorylation) was correlated with changes in Terf1 mRNA levels. This prompted us to pursue mechanistic studies in C2C12 mouse myotubes. We treated myotubes with calcium ionophore which increased p38 MAPK phosphorylation and reduced levels of Terf1 mRNAs, recapitulating our exercise phenotypes. We speculated that the increased intracellular calcium-associated with muscle contractions during exercise results in increased activation (phosphorylation) of p38 MAPK and reduced mRNA levels of Terf1 in rodent skeletal muscle. Since TRF1 is a negative regulator of telomere length, we then speculated that the reduced levels of TRF1 may be compensating for the shorter telomeres observed in our exercise and telomere length studies of skeletal muscles. The reduction in TRF1 may be a mechanism that skeletal muscle uses to prevent or slow further telomere shortening. With these intriguing results in skeletal muscle, we next investigated the impact of exercise on cardiac tissues (left ventricle).

In the hearts of these same mice, we measured shelterin protein levels and mRNA levels of DNA repair and response genes (Chek2 and Ku80/Xrcc5) (67). In contrast to skeletal muscle, we observed increased TRF1 and TRF2 protein and mRNA levels, greater expression of DNA-repair and -response genes (Chek2 and Ku80/Xrcc5) and greater protein content of phosphorylated p38 MAPK immediately post-exercise compared with both controls and 1-hour post-exercise animals. These data provide insights into how physiological stressors remodel the heart tissue and how an early adaptive response mediated by exercise may be maintaining telomere length and/or stabilize the heart genome through the upregulation of telomere-protective genes.

Future research directions

As discussed in the previous paragraphs, the evidence currently suggests that exercise impacts telomere length, telomerase activity, TERT mRNA expression and shelterin proteins and mRNA expression. Given that the evidence is still preliminary and additional investigations are required for robust conclusions and recommendations to be made. While many research directions could be imagined, we outline a few of the main goals and challenges that future studies could consider addressing.

1. Measurement of individual telomere lengths with cell type identification. It is clear from the pre-clinical animal studies that longer telomeres and enhanced telomerase and TERT expression are protective against many age-related pathologies. Likewise, long-term exercise training reduces the risk of all-cause mortality and appears to impact telomere biology in a variety of tissues. Our hypothetical graph (Fig. 3A) depicts our working model for how exercise could maintain telomere length and prevent cellular senescence. However, one should have a healthy skepticism of mean telomere length measurements that have been performed to date in tissues from exercised mammals. The induction of cellular senescence phenotypes is dependent only on a few critically short telomeres. Single telomere length measurements are therefore vital for advancing the field forward since minimal telomere length is more physiologically relevant than average telomere length. Measuring telomere length via the telomere shortest length analysis (TeSLA) (89), from single cells/nuclei following flow cytometry for cell surface indicators to determine the percentage of short telomeres in specific cell types would be a strong start in determining the tissue specific effects of exercise on telomere length regulation.

2. How does exercise impact telomere length independent DNA damage signaling in post-mitotic tissues such as skeletal muscle and cardiac muscle? Recent evidence indicates that length independent DNA damage to telomeres also induces premature senescence (54). One could measure TAFs/TIFs and correlate these outcomes to senescence markers, such as beta-galactosidase, p53, p21, or p16 expression with and without exercise. These types of studies would lend robust support to the hypothesis that part of the exercise adaptive process is by maintenance of telomeres in cells and prevention of the cellular senescence program from being induced (48). Combining future directions 1 and 2 in skeletal muscle one could help to resolve the conundrum of some data indicating that exercise may shorten mean telomere length. By first determining the percentage of short telomeres in skeletal muscle fiber nuclei and then measuring TAFs/TIFs to determine the number of telomeres with DNA damage signaling, one could determine if exercise is protective of telomere function by prevention of TAFs/TIFs in post-mitotic skeletal muscle despite exercise-induced telomere shortening. This would provide solid evidence that exercise prevents the onset of cellular senescence in aging skeletal muscle and that telomere function is more important than length in post-mitotic tissues.

3. What dose of exercise and when does exercise need to be initiated during the life course to protect telomeres and prevent cellular senescence? Establishment of the above types of methods could allow one could delve into what the dose (frequency, intensity and duration) of aerobic exercise is required for telomeres to protect nuclear DNA, prevent NHEJ (non-homologous end joining), and DNA damage responses. It is also unclear whether exercise prevents DNA damage along the telomeres (TIFs or TAFs). Answering whether life-long exercise is required, or if an individual could start exercising at age 50-years and still adapt to improve telomere maintenance, prevent cellular senescence, and improved functional outcomes (independent living, grip strength, cardiovascular disease risk, 6-meter walk test, chair stands, etc.) is a high priority question worthy of pursuit.

4. Confirmation of splicing/AS of TERT following exercise in human samples. The hTERT AS phenotypes observed following acute aerobic exercise in the hTERT BAC mouse model are ripe for further investigation in humans. Understanding which hTERT transcripts are increased following exercise would help investigators determine which phenotypes should be measured and further pursued.

5. Mechanisms of TERT expression and AS regulation during and following exercise. TERT is regulated at the epigenetic levels (e.g., DNA methylation and histone modification), as well as at the transcriptional level by a host of transcription factors, and by AS. Understanding this mechanism is important because tight regulation of telomerase is cancer protective while constitutively high levels of telomerase in the presence of oncogenic mutations could be cancer permissive. We speculate that exercise promotes tight regulation and transient activation of TERT and telomerase rather than a constantly high level of telomerase. This is a relatively untapped area of inquire in the exercise and telomere biology field, but we offer some potential insights. Namely, several transcription factors are activated by an acute bout of aerobic exercise, such as c-MYC (90), which is known to activate the TERT promoter (91, 92). Reactive oxygen species (ROS) also impacts the genome by causing oxidative damage the nucleobases, including the sensitive telomeric G-rich repeats. Therefore, we suggest a model for how exercise-induced c-MYC upregulation in concert with exercise-induced ROS could be key signaling molecules linking aerobic exercise to transient induction of TERT expression (Fig. 3B). Our data, and that of others, indicates that exercise is a post-transcriptional gene-regulating stimulus. We propose that exercise impacts the expression and activity of splicing factors and RNA-binding proteins alike, modifies their cis-element binding preferences, and subsequently impacts the AS of many transcripts including hTERT (Fig. 3B). The functional outcomes of the choice to splice or to alternatively splice TERT is the topic of our last potential future direction.

6. Exercise adaptations to telomere maintenance are driven by transient increases in expression of full-length TERT and TERT AS isoforms.

   1. Non-canonical functions of full-length TERT: Numerous studies mentioned above have tested the straightforward hypothesis that exercise maintains telomere length, which is likely driven by increased TERT protein levels and telomerase activity. However, as pointed out from dominant-negative TERT studies, TERT proteins have alternative or non-canonical roles in cell biology (93, 94). Some non-canonical functions of full-length TERT protein include DNA damage sensing and repair, enhancing mitochondrial function, transfer RNA synthesis, RNA-dependent DNA synthesis, WNT/Beta-catenin signaling, and cellular proliferation (Fig. 3B) (95). Through these non-canonical functions of TERT proteins, could TERT prevent TAFs/TIFs in post-mitotic tissues such as skeletal and cardiac muscles? It could be interesting to engineer a mouse or cell line to express a non-catalytically active full-length TERT and investigate how exercise or exercise-like perturbations impact TERT non-canonical functions that may play a role in exercise-induced adaptations including enhanced telomere function and protection from senescence.
   2. Identification and functions of exercise induced TERT AS isoforms: To date, only two isoforms of TERT (potential full length (exons 7/8 included) and minus beta (skipping of exons 7/8)) have been measured using RT-PCR methods following aerobic exercise. While these two isoforms of TERT are the most studied and two of the more abundant isoforms, there are also additional isoforms. Using more direct methods, such as targeted long-read RNA sequencing from tissues before and after aerobic exercise allowing the identification and quantification of TERT mRNA isoforms in full-length (rather than more traditional fragmented methods) would be very insightful for identifying exercise induced gene expression responses. From here an investigator could initiate studies into the protein-coding capacity of the identified mRNAs and subsequently determine if these proteins are exercise induced and delay senescence.

Figure 3. Hypothetical model of how aerobic exercise can attenuate the induction of cellular senescence and thus aspects of aging.

Hypothetical trajectory of telomere shortening determined by lifestyle. Telomere attrition is attenuated by a physically active lifestyle resulting in telomeres reaching a critically shortened length and thus inducing cellular senescence at an older age compared individuals with a sedentary lifestyle (A). Reactive oxygen species (ROS) and c-MYC expression are known to be increased by acute aerobic exercise. Transcription of TERT is upregulated by ROS and/or c-MYC expression. Epigenetic changes can be induced by exercise and are involved in the upregulation of TERT transcription, but they remain uninvestigated following exercise at the TERT loci. Following aerobic exercise, TERT isoforms including full-length and alternatively spliced isoforms will be generated. By transient upregulation of full-length TERT, telomerase activity would be increased and thus reduce telomere shortening compared to sedentary individuals. Telomere elongation is the canonical function of telomerase and only full-length TERT can generate telomerase that can elongate telomeres. However, non -canonical functions of TERT can be achieved by both full-length and alternatively spliced TERT isoforms. Regulation of TERT isoform expression could be related to the attenuation of aspects of aging, and here we show six different non-canonical functions of TERT that could be related to exercise induced TERT isoforms (both full-length and alternatively spliced variants). Overall, this hypothetical model indicates that exercise induces changes in telomere biology driven by TERT and telomerase that help maintain telomere length and prevent DNA damage response-induced cellular senescence and thus attenuate aspects of the aging process and promote health in physically active humans. Created with BioRender.com.

Conclusions

Currently, exercise is the most effective and powerful intervention we have in our “toolbox” to extend health in older humans. By understanding the mechanisms, by which exercise impacts telomere biology and, in turn, cellular senescence, critical insights into telomere biology could undeniably be made. Advancements in our understanding of telomere physiology could lay the groundwork for future targeted pharmacological treatments and be a powerful conductor of public health messaging surrounding physical activity. The literature summarized in this perspective review highlights that exercise promotes the expression of TERT, which in turn forms active telomerase and helps to maintain telomeres, thus preventing the accumulation of senescent cells with aging (as summarized in Fig. 3). Although more comprehensive and confirmatory studies are needed surrounding dosage and mechanism-specificity, the telomere maintaining, and anti-senescent effects of exercise have been demonstrated repeatedly by leading investigators and our lab. To increase the rigor of the presented evidence, we propose further mechanism-based investigations, the implementation of preclinical and human studies, and the extension of our fields findings towards prescribing exercise as an effective, safe, and implementable therapeutic strategy to promote healthy aging.

Key points:

• Telomere length maintenance and enhanced telomerase activity are exercise-induced adaptations.

• Exercise induces the gene expression of the catalytic subunit of telomerase, telomerase reverse transcriptase (TERT).

• TERT and alternatively spliced isoforms of TERT may have specific exercise-induced functions separate from telomere length maintenance.

• Attenuating cellular senescence via maintenance of telomeres is a key morbidity-reducing adaptation of exercise.