The role of DNA-PK in aging and energy metabolism (invited review)

Abstract

DNA-dependent protein kinase (DNA-PK) is a very large holoenzyme comprised of the p470 kDa DNA-PK catalytic subunit (DNA-PK_cs) and the Ku heterodimer consisting of the p86 (Ku 80) and p70 (Ku 70) subunits. It is best known for its non-homologous end joining (NHEJ) activity, which repairs double-strand DNA (dsDNA) breaks (DSBs). As expected, the absence of DNA-PK activity results in sensitivity to ionizing radiation, which generates DSBs and defect in lymphocyte development, which requires NHEJ of the V(D)J region in the immunoglobulin and T cell receptor loci. DNA-PK also has been reported to have functions seemingly unrelated to NHEJ. For example, DNA-PK responds to insulin signaling to facilitate the conversion of carbohydrates to fatty acids in the liver. More recent evidence indicates that DNA-PK activity increases with age in skeletal muscle, promoting mitochondrial loss and weight gain. These discoveries suggest that our understanding of DNA-PK is far from complete. As many excellent reviews have already been written about the role of DNA-PK in NHEJ, here we will review the non-NHEJ role of DNA-PK with a focus on its role in aging and energy metabolism.

Graphical abstract

DNA-dependent protein kinase (DNA-PK) is best known for its non-homologous end joining (NHEJ) activity, which repairs double-strand DNA breaks. Recent evidence indicates that DNA-PK activity increases with age in skeletal muscle promoting mitochondrial loss and weight gain. Here we will review the non-NHEJ role of DNA-PK with a focus on its role in aging and energy metabolism.

Introduction

Average life expectancy has increased dramatically, leading to a rapid graying of the population across the globe [1]. Although increased life expectancy in itself is a positive development, it comes with significant public health challenges, particularly increased frailty and risk for a diverse array of chronic diseases. Mounting evidence suggests that processes intrinsic to aging promote age-related diseases [2] but the pathway that links aging to the increase in disease risk is poorly understood.

Significant strides have been made in understanding the potential promoters of aging phenotypes and susceptibility to diseases [3, 4]. These age-related changes include mitochondrial dysfunction, reactive oxygen species (ROS), DNA damage and declines in protein quality control, cellular regenerative capacity and signaling control. These alterations disrupt metabolic and inflammatory homeostases, which create the tissue microenvironments permissive for development of diseases.

DNA damage is one well-known promoters of the aging phenotype [5]. Causes of DNA damage can range from single- and double-strand breaks produced by abortive topoisomerase actions, chemical reactions, to ionizing radiation to DNA-base mismatches introduced during DNA replication, or damage generated by chemicals or exposure to ultraviolet light [6]. DNA damage, which can cause cellular senescence or even death [7], is particularly consequential if it occurs in stem cells because it can diminish the regenerative capacity of organs and tissues [8].

To maintain genome stability, organisms have evolved mechanisms to sense the various forms of DNA damage and to initiate signaling pathways that repair them [9, 10]. As with most signaling pathways, the DNA damage signaling pathway is initiated by kinases [11]. One of the best characterized among these is DNA-PK, which is activated by dsDNA breaks [11, 12]. The existence of a kinase that is stimulated by dsDNA was first described by Ohtsuki et al. [13], although whether this activity corresponded to DNA-PK is not known. Subsequently, three groups independently detected a dsDNA binding kinase in a variety of mammalian cellular extracts that required dsDNA for its activity and called it DNA-PK [14–17]. Both Carter et al. [15] and Jackson et al. [16] observed that the enzyme functions efficiently in the presence of linear but not supercoiled DNA, indicating that DNA-PK is activated by dsDNA ends, which are absent in supercoiled DNA. Work by Lees-Miller et al. [18] indicated that a high resolution chromatographic step intended to purify the kinase dramatically reduced its catalytic activity. This curious observation was subsequently explained by the discovery that the catalytic subunit of DNA-PK (DNA-PKcs) required interaction with the DNA end-binding Ku heterodimer consisting of p86 (Ku 80) and p70 (Ku 70) for full catalytic activity [19, 20]. In 1995, cloning of the DNA-PKcs cDNA revealed that it is an approximately 470 kD polypeptide, with the kinase domain in the carboxy-terminal approximately 500 residues [21]. Interestingly, the kinase domain [21, 22] has homology to the phosphatidylinositol 3 (PI 3)-kinase family which phosphorylates inositol phospholipids [23], but DNA-PK is a pure serine/threonine protein kinase and has no lipid kinase activity [21, 24]. Ataxia-telangiectasia-mutated (ATM) [25, 26] and ATM- and Rad3-Related (ATR), two kinases that, together with DNA-PK, make up the triumvirate of DNA-damage sensing kinases also have homology to PI 3-kinase. Interestingly, while both ATM [25–28] and ATR [29, 30] are present in all eukaryotes, DNA-PKcs is only present in higher eukaryotes; vertebrates have clear homologs [31–34], but invertebrates do not.

As mentioned above, DNA-PK repairs DSBs, and as a result, cell lines carrying a mutation in either Ku or DNA-PKcs are sensitive to agents that induce DSBs such as ionizing radiation [35–42] (Fig. 1). DNA-PK also repairs stalled DNA replication forks [43, 44], consistent with the observation that DNA-PK deficient cells are sensitive to DNA replication fork stalling agents [45]. In addition to the DNA repair process itself, DNA-PK activates checkpoint in response to replication fork stress [43, 44, 46]. The primary mode of DSB repair by DNA-PK is NHEJ, a low-fidelity mode of repair in which the DNA ends are directly ligated without using a homologous template [47]. Programmed DSBs created during V(D)J recombination and class switching recombination in lymphocytes are joined by NHEJ [48]. Evidence that DNA-PK mediates NHEJ came from humans, animals and cell lines. Severe combined immune deficiency (SCID) mice which carry a leaky mutation in DNA-PKcs, have impaired lymphocyte development [35, 49–51]. The SCID phenotype has also been described in humans [34], dogs [52] and horses [33] carrying a mutation in DNA-PKcs.

Fig. 1.

DNA-PK function in genetic stability and cell growth. Structure of DNA-PK complexed with DSBs is shown in the center [54]. Three functions of DNA-PK shown are: DSB repair, V(D)J recombination in lymphocytes via NHEJ and replication fork stress repair and checkpoint (top), binding to and protecting telomeres (lower left) and promoting cell division and growth (lower right).

As a DNA DSB sensor and mediator of NHEJ, DNA-PK has some unexpected properties. For example, DNA-PK is very abundant: it is estimated that HeLa cells contain approximately 100,000 copies of DNA-PKcs per cell [53], far in excess of what is probably needed for NHEJ as only one DNA-PKcs binds to each DSB end [54]. In addition, DNA-PKcs is present not only in the nucleus but also in the cytoplasm [55]. These features suggest that DNA-PK may have additional functions unrelated to NHEJ; indeed, there is emerging evidence that DNA-PK also regulates aging and metabolism. Here, we will focus on this new development and discuss the questions and implications that these novel functions raise about DNA-PK, aging and metabolism.

DNA-damage and aging

The concept that DNA-damage is one of the drivers of the aging process is supported by numerous examples of systemic premature aging syndromes that occur due to a mutation in genes important for maintaining genomic integrity. Although there are a number of progeroid syndromes, perhaps the one that best recapitulates adult aging in an accelerated fashion is the Werner syndrome, which is caused by deficiency of WRN, a member of the RECQ family of DNA helicases [56]. Other syndromes with partial features of premature aging include Ataxia-telangiectasia, dyskeratosis congenita, Cockayne syndrome, Fanconi’s anemia and trichothiodystrophy [57, 58]. Since individuals with these syndromes have had the defects since the embryonic development, it is difficult to know what contribution, if any, that DNA damage occurring during development has on aging as an adult. However, the observation that long-term survivors of DNA-damaging chemotherapy exhibit multiple features of accelerated aging, both at the physiological and molecular levels [59], is consistent with the notion that DNA damage in adults can promote aging.

DNA-PK and aging

The first hint that DNA-PK may affect aging comes from the observation that DNA-PK plays a role in telomere maintenance. The Ku subunits can bind to telomeres [60, 61] and mouse embryonic fibroblasts (MEFs) deficient in Ku70, 80 or DNA-PKcs have a higher number of telomeric fusions, indicating that DNA-PK is important for telomere capping [62–66] (Fig. 1). Interestingly, bone marrow cells from DNA-PKcs−/− mice had no telomeric fusions. This difference between bone marrow cells and MEFs may be a reflection of different levels of extra- cellular oxygen, the source of ROS which can generate DSBs. Internal organs are likely to have lower oxygen levels than room air, the condition under which MEFs are grown, and as a result, MEFs may produce more ROS and DSBs.

DNA-PK also plays a role in the regulation of telomere length, which is controlled by the opposing effects of telomere synthesis by telomerase and telomere shortening. Inactivation of DNA-PK in the human cell line HCT116 shortened telomeres [67], indicating that DNA-PK plays a role in telomere length maintenance. Consistent with this, first generation (G1) DNA-PKcs−/− mice, telomeres were slightly shorter at older age compared with wild type mice [68, 69], but G3-G4 DNA-PKcs−/− mice had significantly shorter telomeres even at young age [70]. To determine whether the role of DNA-PKcs in telomere length control was telomerase-dependent, Espejel et al. studied telomere lengths in telomerase-deficient mice. They found that mice doubly deficient in DNA-PKcs and telomerase displayed an accelerated rate of telomere shortening when compared to mice deficient only in telomerase [70]. Whether it is due to telomere shortening or other factors, DNA-PKcs−/− mice also showed signs of accelerated aging. In keeping with the critical function of DNA-PKcs in immunity, DNA-PKcs−/− mice had higher rates of infection, including pneumonia and hepatitis and had a shorter life span [68, 69].

Like DNA-PKcs−/− mice, Ku 70−/− and Ku80−/− mice also exhibit growth retardation and accelerated aging phenotypes [71–73]. However, unlike mice doubly deficient in DNA-PKcs and telomerase, mice doubly deficient in Ku 80 and telomerase showed a similar rate of telomere shortening when compared to mice deficient in telomerase only [68]. Since the full DNA-dependent catalytic activity of DNA-PKcs requires the Ku 70/80 heterodimer, it is not clear whether the function of DNA-PKcs at telomeres is as a kinase or as a scaffold protein. These findings also suggest that the accelerated aging phenotype in DNA-PKcs−/− mice may not be directly related to telomere shortening.

Unexpectedly, Scid mice have telomeres that are 1.5-2 fold longer than those from wild-type mice [74]. Although the aging process in Scid mice has not been studied as extensively as that in DNA-PKcs−/− mice, Scid mice do not appear to exhibit significantly accelerated aging or shortened lifespan. Since the Scid mice carry a leaky mutation of DNA-PKcs, the discrepancy in telomere length and life span between Scid mice and DNA-PKcs−/− suggests that the effect of DNA-PKcs on telomere length and aging may not be simply dose-dependent and may even be indirect.

The potential role of DNA-PK in aging in humans has been reported in the context of the premature aging syndrome Hutchison-Gilford progeria syndrome (HGPS), which is caused by a single point mutation in the lamin A gene [75, 76]. The accumulation of mutated lamin A decreases expression of nuclear DNA-PKcs and Ku70/Ku80 in HGPS fibroblasts. Liu et al. found that decreasing the expression of DNA-PKcs reduced the proliferation of primary vascular smooth muscle cells and that fibroblasts isolated from normally aging individuals have reduced levels of DNA-PKcs/Ku70/Ku80 and also that decreased expression of DNA-PK is associated with reduced proliferative capacity [77] (Fig. 1).

DNA-PK promotes insulin-stimulated fatty acid synthesis

Storage of nutrients in the form of adipose tissues provides a steady energy supply despite fluctuating food intake (feast-famine cycle). After food intake, excess carbohydrates are converted to fatty acids for storage as triacylglycerol. Lipogenesis is tightly regulated by enzymes involved in fatty acid and triglyceride synthesis, such as fatty acid synthase (FAS) [78, 79] and mitochondrial glycerol-3-phosphate acyltransferase [80]. Their expressions are low during fasting and high during feeding [81]. Regulation of these enzymes occurs mainly at the transcriptional level [82–84] by the transcription factor Upstream Stimulatory Factor (USF)-1/2 heterodimer [85–87]. Although it was generally believed that the feeding/fasting cycle is relayed to USF via the insulin signaling, which increases with feeding, the precise mechanism of how insulin signaling is relayed to USF and to these lipogenic gene promoters, was unknown until Wong et al. discovered that insulin activates USF and increases FAS expression through DNA-PK [88]. Insulin begins the cascade by activating protein phosphatase 1 (PP1), which then dephosphorylates and activates DNA-PK. Activation of DNA-PK also requires DNA DSBs, and indeed Wong et al. observed signs of DNA DSBs in the FAS promoter region after insulin treatment: DNA ends capable of being labeled by biotin-UTP and binding of topoisomerase IIβ, which can cleave DS DNA. Activated DNA-PK phosphorylates Ser262 in USF, which promotes recruitment of and acetylation by coactivator p300/CBP-associated factor. In Scid mice, USF-1 phosphorylation and acetylation in response to insulin is reduced, blunting transcriptional activation of FAS and de novo lipogenesis after carbohydrate feeding [88] (Fig. 2).

Fig. 2.

DNA-PK promotes fatty acid synthesis in the liver after feeding. Carbohydrate intake stimulates insulin secretion, which activates DNA-PK via PP1. DNA-PK phosphorylates Ser262 of transcription factor USF, which results in the recruitment of p300 acetyl transferase and acetylation-mediated activation of USF. Activated USF increases transcription of FAS and conversion of carbohydrate to fatty acid.

DNA-PK inhibits AMPK at older age

Aging is associated with many common diseases such as cardiovascular diseases, cancer, type 2 diabetes and neurodegenerative diseases. Obesity increases the risk for all of them. One typically gains approximately 30 pounds from his/her 20s to 50s, which translates to roughly one pound gain per year. In the US, over 70% of adults are either overweight or obese, and the percent of people who are obese is rising. This weight gain occurs even though food intake typically decreases with aging, indicating that the metabolic rate decreases with aging. One potential cause for the decline in metabolic rate with age may be the loss of mitochondria, which metabolizes nutrients to generate energy and heat [89–91]. Loss of mitochondria in skeletal muscle may also explain, at least in part, the decline in physical fitness during aging.

Why does aging lead to mitochondrial loss in skeletal muscle? Accumulating evidence indicates that the aging-associated decline in the activity of AMP-activated protein kinase (AMPK), a key regulator of mitochondrial function and energy balance [92–94], plays an important role in the aging-associated decrease in mitochondrial function and insulin response [95–98]. Activation of AMPK has healthful effects including stimulation of glucose uptake, fat oxidation, energy production and mitochondrial biogenesis [94, 99]. Metformin, the first line drug for type 2 diabetes, acts in part by activating AMPK [100]. Increased AMPK activity decreases visceral fat [101] and increases mitochondrial biogenesis [102] and energy production in skeletal muscle resulting in improved physical fitness. On the other hand, AMPK-deficiency in skeletal muscle leads to mitochondrial loss, impaired glucose uptake and exercise intolerance [103]. AMPK activity declines with age, and this appears to the mediated by elevated intra-mitochondrial ROS because reducing ROS in skeletal muscle by targeting the antioxidant catalase to mitochondria prevents loss of AMPK activity and mitochondrial function at older age [97]. However, the molecular mechanism by which aging and ROS decrease AMPK activity in skeletal muscle was poorly understood.

Like many kinases, AMPK [104] and its upstream activator kinase LKB1 [105] are folded in part by chaperone protein HSP90, which is unique among chaperone proteins in that it binds to and folds metastable proteins and only a small fraction of the total proteome. Most of the HSP90 client proteins are involved in signal transduction, including kinases and steroid hormone receptors [106]. HSP90 is composed of two isoforms: HSP90α, which is stress induced and is not essential for cellular survival, and HSP90β, which is constitutively expressed and is essential for cellular survival. Inhibiting HSP90 activity can result in misfolding of clients, which can lead to their degradation. Although HSP90α and HSP90β are 85% identical in protein sequence and have highly overlapping functions, they appear to have some distinct roles. It appears that HSP90α is particularly important for folding LKB1 and AMPK because knocking down HSP90α with siRNA decreases LKB1 and AMPK levels; whether HSP90β plays any role in folding LKB1 and/or AMPK is not known [107].

The connection between DNA-PK and the functions of HSP90α clients such as LKB1 and AMPK was made when it was discovered that DNA-PK phosphorylates Thr5,7 (T5,7) in HSP90α both in vitro [17] and in vivo [107–109]. Thr5,7 phosphorylation decreases HSP90α-client interaction and presumably the folding of these clients [107]. Consistent with the increase in genetic breaks [107, 110, 111] and DNA-PK activity [107] at older age, T5,7 phosphorylation increased with age in skeletal muscle of both rhesus monkeys and mice. (Fig. 3). In line with this concept, AMPK activity is decreased in skeletal muscle of older WT mice but not of older Scid mice compared to young mice. Muscle-specific knockout of DNA-PKcs also decreased T5,7 phosphorylation and prevented aging-associated loss of AMPK activity in skeletal muscle, indicating that the effect of DNA-PK on AMPK activity is cell autonomous and is unrelated to its immune function [107]. Interestingly, in tissue culture cells, AMPK activation by glucose deprivation is decreased in the absence of DNA-PK [112], suggesting that the role of DNA-PK in AMPK regulation is complicated.

Fig. 3.

Aging promotes DNA-PK-mediated suppression of AMPK in skeletal muscle. DSBs and/or ROS, both of which increase with age, likely activate DNA-PK. Activated DNA-PK phosphorylates Thr5,7 in HSP90α which decreases the activity (folding) of its client proteins LKB1/AMPK. AMPK activity increases mitochondrial biogenesis, quality control and function (e.g. fat oxidation) and glucose uptake and decreases ROS, protecting against obesity and improving glucose tolerance and physical fitness.

DNA-PK drives mitochondrial loss at older age

As AMPK plays a critical role in mitochondrial biogenesis, increasing DNA-PK activity at older age may promote mitochondrial loss in skeletal muscle. Indeed, there is a strong inverse correlation between mitochondrial content and DNA-PK activity in skeletal muscle of middle-aged rhesus monkeys [107], and mitochondrial loss does not occur in older SCID mice. In skeletal muscle, activation of AMPK promotes a switch from glycolytic fibers to oxidative fibers [101, 113, 114]. Consistent with DNA-PK being a negative regulator of AMPK, older SCID muscle had more oxidative fibers than WT muscle as did the muscle from the mice treated with DNA-PK inhibitor NU7441. Glycolytic fibers generate lactic acid, a major source of muscle fatigue [115, 116]; and switching to oxidative fibers confers fatigue-resistance [101, 117, 118]. Consistent with this, older SCID mice have lower serum lactic acid levels and have reduced aging-associated decline in exercise capacity on treadmill running compared to WT mice.

Higher percentage of oxidative fibers in SCID muscle is consistent with higher AMPK activity in SCID muscle, but it conflicts with the observation that it is the glycolytic fiber, not the oxidative fiber, that is preferentially lost in the elderly in a process called sarcopenia. This seeming contradiction most likely reflects the complex nature of sarcopenia, a process marked by overall declines in size and in number of glycolytic skeletal muscle fibers and an infiltration of fibrous and adipose tissue into the skeletal muscle [119]. In addition, satellite cells, the skeletal muscle precursor cells that are activated and trigger skeletal muscle repair and regeneration in response to the stress of heavy muscle use or injury [120], also undergo aging-related changes, including reduction of satellite cell content, particularly in the glycolytic skeletal muscle fibers [121].

Could a small molecule inhibitor of DNA-PK be an effective fitness enhancer? Indeed, feeding obese and middle-aged WT mice NU7441 increased the exercise capacity of mice by approximately 60% and 40%, respectively [107]. The metabolic effects of NU7441 were mediated by AMPK as NU7441 did not increase mitochondrial biogenesis in skeletal muscle of or exercise capacity in AMPKα2 KO mice, indicating that the fitness-enhancing effect of NU7441 requires AMPK.

Inhibiting DNA-PK improves glucose metabolism

Incidences of disease associated with dysregulation of metabolism increase with aging. For example, among adults aged 18-44, 4% have diabetes (diagnosed and undiagnosed) whereas among adults 65 years or older, 25% have diabetes [122]. One of the characteristics of both type 2 diabetic and obese insulin-resistant nondiabetic individuals is decreased fat oxidation in skeletal muscle [123], which is consistent with impaired mitochondrial oxidative capacity in muscle [124, 125]. Even insulin-resistant offspring of type 2 diabetic parents have impaired mitochondrial activity [126–128]. The decline in mitochondrial content and function also occurs with aging in healthy humans and rats and may contribute to insulin resistance at older age [89–91]. These findings have led some to conclude that the primary defect is muscle mitochondrial dysfunction, resulting in elevated intramyocellular fatty acid metabolites and insulin resistance [129, 130]. However, small increases in palmitoyl carnitine can decrease ATP synthesis in mitochondria [131], which suggests that increased muscle lipid content resulting from increased fatty acid delivery could also be the primary defect that results in mitochondrial dysfunction and insulin resistance. Regardless of the nature of the primary defect, inhibiting DNA-PK may improve insulin sensitivity and protect against type 2 diabetes by activating AMPK and thereby increasing mitochondrial function. Indeed, in mice fed with a high fat diet, the inclusion of NU7441 in the food significantly decreased weight gain [107]. Moreover, NU7441 increased both insulin sensitivity and glucose tolerance primarily by increasing glucose uptake in skeletal muscle and adipose tissue. Therefore, increased DNA-PK activity at older age may contribute to the development of insulin resistance and ultimately, type 2 diabetes (Fig. 3).

Calorie restriction and aerobic fitness are associated with decreased DNA-PK activity

A calorie restricted diet can activate AMPK, increase mitochondrial biogenesis, protect against a diverse array of diseases and disease risk factors such as insulin resistance and metabolic syndrome [132, 133] and extend life span compared to an ad libitum diet [132, 133]. Similarly, rats selectively bred for high intrinsic exercise capacity [134] have higher mitochondrial biogenesis and are protected from a number of diseases and have a longer lifespan compared to rats bred for low intrinsic exercise capacity [135]. Interestingly, both calorie-restricted middle age rhesus monkeys and rats with high intrinsic exercise capacity have decreased DNA-PK activity and HSP90α phosphorylation in muscle. These findings suggest that decreased DNA-PK activity may contribute to the metabolic benefits produced by both calorie restriction and intrinsic aerobic fitness and may warrant further studies.

DISCUSSION

DNA-PK is largely known for its function in NHEJ [11, 48], but more recent evidence indicates that DNA-PK has functions far beyond NHEJ. It is this complex repertoire of functions that can result in seemingly paradoxical observations. Given its function in telomere maintenance and genetic stability, it was not at all surprising that DNA-PK−/− mice undergo accelerated aging [68, 69]. Therefore, it was unexpected that the SCID mutation or treatment with a DNA-PK inhibitor protected against aging-associated decline in metabolism and physical fitness in mice. The most likely explanation for this paradox is that neither the SCID mutation nor the DNA-PK inhibitor inhibit DNA-PK completely, but incomplete loss of DNA-PK activity may be sufficient to decrease its metabolic function while at the same time providing some protective function against the level of DNA DSBs generated naturally. It is also possible that the results may depend on the cell type in DNA-PK−/− or SCID mice.

The study of DNA-PK in non-NHEJ functions such as aging and metabolism is in its early stages and a number of unanswered questions remain. First, what other proteins does DNA-PK phosphorylate in the aging-related pathway and what do they do? One such candidate is Ser473 of AKT, which is required for AKT activation. This site can be phosphorylated by mTORC2 [136] in response to growth signals or by DNA-PK [137–143] in response to DNA breaks. Both insulin and insulin-like growth factor 1 pathways promote aging, and they do so in part by activating AKT [144]. Decreasing the activity of mTOR, which is activated by AKT, increases longevity [145–148]. Aging is also associated with inflammation [149], which is a common denominator of most chronic diseases, including metabolic syndrome [150] and type 2 diabetes [151]. NF-κB, the master transcription factor for inflammation [152], can be activated by AKT [153]. DNA-PK has also been reported to phosphorylate and activate NF-κB [154], but evidence contradicting this has also been reported [155]. Therefore, the DNA-PK-AKT pathway, may contribute to the aging phenotype, not only by increasing mTOR activity, but also by increasing inflammation as evidenced by the role of DNA-PK in inflammation-driven diseases such as asthma [156, 157]. Second, why would nature use a DNA break sensor such as DNA-PK for non-NHEJ functions? This raises the question as to whether DNA-PK is really sensing chromosomal breaks for the aging-related pathways. Although both chromosomal breaks and DNA-PK activity increase with aging [107, 111], we cannot conclude that the increase in DNA-PK activity is mediated solely by chromosomal breaks as DNA-PK is also activated by ROS independently of chromosomal breaks [158]. This may be relevant because ROS also increases with aging [97] and mitochondria, which are regulated by DNA-PK (via AMPK), are the main source of ROS production [159]. This relationship may hint at some type of feedback mechanism, but more work is needed to demonstrate this feedback relationship. The published study of the role of DNA-PK in metabolism and aging has been largely limited to the liver and skeletal muscle, and it is not known whether the role of DNA-PK in metabolism and aging extends to other tissues as well.

HSP90 is considered to be a capacitor of phenotypic variation [160, 161]: by folding metastable proteins and therefore acting as a buffer, HSP90 allows mutations or polymorphisms (inherited or acquired) to accumulate unseen, unaffected by the pressures of natural selection. In this context, HSP90α T5,7 phosphorylation might decrease the ability of HSP90α to provide the necessary buffering at older age. As a consequence, HSP90 chaperone function may be decreased and mutations or polymorphisms may be more exposed at older age. In this scenario, aging and aging-associated diseases may partly be a manifestation of the exposure of these mutations or polymorphisms.

Finally, could a DNA-PK inhibitor be useful for delaying aging and/or ameliorating the risks for aging-associated diseases? Even though a DNA-PK inhibitor may inhibit DNA-PK incompletely, the function of DNA-PK in NHEJ and telomere maintenance may be affected by the inhibitor, potentially increasing the risk for cancer. Therefore, a safer approach may be to decrease the generation of the trigger that activates DNA-PK at older age, but this requires further investigation into the nature of the in vivo trigger(s) of DNA-PK.

Abbreviations

DNA-PK

DNA-dependent protein kinase

DNA-PKcs

DNA-PK catalytic subunit

dsDNA

double-strand DNA

DSB

double-strand break

ATM

ataxia-telangiectasia-mutated

ATR

ATM- and Rad3-Related

NHEJ

non-homologous end joining

SCID

severe combined immune deficiency

AMPK

AMP-activated protein kinase

LKB1

liver kinase B1

MEFs

mouse embryonic fibroblasts

HGPS

Hutchison-Gilford progeria syndrome

FAS

fatty acid synthase

ROS

reactive oxygen species

USF1

upstream stimulatory factor

PP1

protein phosphatase 1