The Intricate Interplay between Mechanisms Underlying Aging and Cancer

: Age is the major risk factor in the incidence of cancer, a hyperplastic disease associated with aging. Here, we discuss the complex interplay between mechanisms underlying aging and cancer as a reciprocal relationship. This relationship progresses with organismal age, follows the history of cell proliferation and senescence, is driven by common or antagonistic causes underlying aging and cancer in an age-dependent fashion, and is maintained via age-related convergent and divergent mechanisms. We summarize our knowledge of these mechanisms, outline the most important unanswered questions and suggest directions for future research. ? 114, 115, 125, 275-279 shape of the various signaling circuitry characteristic of

Aging and Disease • Volume 6, Number 1, February 2015 57 the resulting decline in their proliferation and mobilization to renewable and differentiated tissues not only suppresses cancer but also operates as a cellnonautonomous pro-aging mechanism by compromising tissue repair and regeneration, and thereby impairing tissue homeostasis [2,3,7,8,11,12,22]. The complexity of the interplay between mechanisms underlying aging and cancer is further underscored by the findings implying that in old organisms: (1) paracrine activities of the senescent noncancerous cells in a renewable tissue enable their interactions with mitotically active non-cancerous cells (in the same tissue or in other tissues) as well as with premalignant and tumor cells (in the same tissue or within the tumor microenvironment); and (2) these numerous interactions exhibit pleiotropic effects on aging and cancer, either beneficial or deleterious for the health of the organism [3-5, 7, 26, 28-30]. In fact, the cell cycle arrest at an early stage of the senescence program in stem/progenitor somatic cells and in division-competent cells within renewable tissues is followed by stepwise changes in chromatin organization and gene expressionwhich in turn alter secretion pattern of interleukins, inflammatory cytokines, chemokines, growth factors, insoluble protein components of the extracellular matrix, extracellular proteases, as well as such non-protein soluble compounds as reactive oxygen species (ROS), nitric oxide and prostaglandin E2 [4,5,7,19,[28][29][30][31][32][33]. Over time, cells at an advanced stage of the senescence program progress through several consecutive steps of developing a senescence-associated secretory phenotype (SASP) also called senescence-messaging secretome (SMS) [5,7,28,31,32]. Paracrine activities of various SASP components affect distant non-cancerous, premalignant and tumor cells through cell-nonautonomous mechanisms that underlie such diverse effects as: (1) tissue repair and regeneration; (2) wound healing; (3) cell senescencebased suppression of tumor growth; (4) disruption of structure and function of normal tissues and the resulting acceleration of age-related degenerative diseases; (5) lowlevel chronic inflammation; (6) immune clearance of noncancerous, premalignant and tumor cells; (7) excessive proliferation of division-competent non-cancerous, premalignant and malignant cells; (8) enhanced cell migration and tissue invasion; (9) tissue-specific changes in cell differentiation; and (10) promotion of tumor progression [3-5, 7, 22, 26, 28-30, 34-36].
In sum, it seems that in old organisms aging and cancer may have common or differing causes and coalescent or divergent mechanisms. Such dualistic relationship between aging and cancer in old organisms is due to: (1) the antagonistically pleiotropic effects and complex temporal organization of the cellular senescence program executed in excessively stressed or damaged non-cancerous cells; and (2) the ability of cancer cells to accelerate aging of the senescent non-cancerous cells within the tumor microenvironment, thus extracting from these non-cancerous cells certain growth-promoting nutrients that fuel tumor progression.
In this review, we discuss the intricate interplay between aging and cancer as a balance between coalescent and divergent mechanisms underlying them. We focus on the current knowledge of how this delicate balance is: (1) impacted by organismal age; (2) influenced by the proliferative history of cells; (3) affected by the temporal organization of the cellular senescence process; and (4) impinged on by the antagonistically pleiotropic effects of senescent cells on aging-and cancer-related processes.

In young and adult organisms, cell-autonomous mechanisms that eliminate aetiologies of aging also suppress tumor formation
The emergence and accumulation of stressed, damaged and dysfunctional macromolecules and organelles in mitotically active cells within renewable issues of young and adult organisms are known to have both the pro-aging and pro-cancer potentials [1][2][3][4][5][6][7][8]. Therefore, cellautonomous mechanisms that in young and adult organisms eliminate these common to aging and cancer aetiologies are expected to decelerate both the aging process and tumor formation [1-3, 5, 6, 8-10, 13, 44-46]. Because in young and adult organisms aging and cancer are likely to have common aetiologies and coalescent cellautonomous mechanisms, it is conceivable that in these organisms: (1) genetic interventions that accelerate the build-up of stress, damage and dysfunction in mitotically active cells by targeting some of such mechanisms may exhibit both pro-aging and pro-cancer effects; and (2) genetic, pharmacological and/or dietary interventions that decelerate the accumulation of stress, damage and dysfunction in mitotically active cells by affecting some of such mechanisms may elicit both anti-aging and anticancer effects [1-3, 6, 10, 13, 44-46].
Aging and Disease • Volume 6, Number 1, February 2015 58 A body of evidence in support of the common aetiologies and coalescent cell-autonomous mechanisms for aging and cancer in young and adult organisms has been extensively reviewed over the last few years [5][6][7][8][44][45][46][47][48][49][50][51][52][53]. In brief, the following three lines of evidence support this commonly accepted concept on the relationship between aging and cancer in mitotically active cells within renewable issues of such organisms.
Second, some protein components and downstream targets of the signaling network orchestrating cellular processes that in young and adult organisms affect the common to aging and cancer aetiologies have been shown to accelerate cellular aging and function as oncogenes. These protein components and downstream targets include PI3K, Ras, Raf, Akt, Ras homolog enriched in brain (Rheb), the eukaryotic translation initiation factor 4E (eIF4E), the hypoxia-inducible factor 1-α (HIF1-α), mitochondrial succinate dehydrogenase subunits SDHB, SDHC and SDHD, the mitochondrial succinate dehydrogenase assembly factor 2 (SDHAF2), and mitochondrial fumarate hydratase FH [46, 48-50, 59, 71, 72]. In contrast, other protein components and downstream targets of this signaling network have been shown to decelerate cellular aging and to act as tumor suppressors; they include PTEN, LKB1, the tuberous sclerosis proteins 1 and 2 (TSC1 and TSC2), the Von Hippel-Lindau tumor suppressor protein VHL, Ras inhibitors NF1 and NF2, and mitochondrial isocitrate dehydrogenases IDH1 and IDH2 [46, 48-50, 59, 71, 72].
Third, certain pharmacological interventions, a caloric restriction (CR) diet and some dietary restriction (DR) regimens exhibit both anti-aging and anti-cancer effects by specifically altering information flow along the PI3K/PTEN/Akt/mTOR, Ras/Raf/MEK/ERK/mTOR and/or LKB1/AMPK/mTOR signaling pathways as well as by modulating some of the downstream targets of these pathways. As it has been mentioned above in this section, the PI3K/PTEN/Akt/mTOR, Ras/Raf/MEK/ ERK/mTOR and LKB1/AMPK/mTOR pathways are integrated into a signaling network orchestrating cellular processes that in young and adult organisms affect the common to aging and cancer aetiologies. The cell-autonomous mechanisms that underlie both anti-aging and anti-cancer effects of such pharmacological interventions, CR and DR have been comprehensively discussed elsewhere [9, 44-46, 48-50, 62-69, 71-85].

In old organisms, multiple mechanisms underlying a multistep cellular senescence program impose antagonistically pleiotropic effects on aging and cancer
In response to excessive intracellular and extracellular stresses, mitotically active stem/progenitor somatic cells and division-competent cells within renewable tissues enter a senescence program that is initiated by an irreversible cell cycle arrest [3,4,7,8,24,28,32,86]. Some senescence-inducing stresses are coupled to cell division; because these stresses reflect the replicative history of division-competent cells, they function as biological clocks counting the finite number of cell divisions progressing with the chronological age of an organism [4, 7, 12-14, 22, 87-91]. Other stresses triggering cellular senescence do not relate directly to the replicative age of cells, and thus may not operate as "replicometers" or "mitotic clocks" set to count the progression of cell divisions with organismal chronological age [4,7,22,25,90]. The various stresses triggering cellular senescence generate certain intracellular signals modulating a distinct set of senescence-inducing signaling pathways [1,3,5,7,12,22,86]. These pathways are integrated into circuits that orchestrate a cellular senescence program progressing through several spatially and temporally distinct steps [1,3,5,22,86,[92][93][94][95][96]. Multiple mechanisms underlying the spatiotemporal organization of this program impose antagonistically pleiotropic effects on aging and cancer, as outlined in this section.

Triggers, sensors, signaling pathways and circuits of the cellular senescence program
A replicative mode of cellular senescence can be triggered by the following two kinds of intrinsic stresses that are coupled to cell division: (1) the gradual loss of telomeric DNA elements at S phases of successive mitotic cell divisions leading to telomeric dysfunction and causing a form of cellular senescence known as telomere-initiated senescence; and (2) the steady rise in the expression of the INK4a/ARF locus leading to a progressing with the proliferative history of cells accumulation of the p16INK4a and p14ARF tumor suppressor proteins [1, 11-13, 22, 86-89, 91]. Some stresses can trigger a mode of cellular senescence known as premature or stress-induced senescence; these stresses include: (1) an accumulation of unrepaired damage to chromosomal DNA and the resulting genomic damage at non-telomeric sites; (2) chromatin remodeling resulting in heterochromatin foci formation and large-scale chromatin condensation; (3) oncogene overexpression/ activation or tumor suppressor gene inactivation, all causing a so-called oncogeneinduced form of cellular senescence; (4) an enhanced expression of cell proliferation activators that create robust mitogenic signals; (5) an excessive proliferation of dysfunctional mitochondria, which results in ROS accumulation and oxidative stress; (6) autophagy induction; and (7) changes in expression patterns of numerous microRNAs [3,4,7,12,22,23,86,[97][98][99][100][101][102][103][104][105][106][107][108][109][110][111][112][113][114][115][116]. Because these stresses are not coupled to cell division (and thus do not relate directly to the replicative age of cells, sometimes being called cell-extrinsic stresses), they are unlikely to function as molecular chronometers that count the number of successive mitotic cell divisions progressing with organismal chronological age [4,12,22,86,90]. When the extent of cellular stress, damage and dysfunction inflicted by a combined action of various triggers of replicative and premature senescence reaches a threshold level, mitotically active cells respond by activating a multistep senescence program that is initiated by an irreversible cell cycle arrest [3,4,7,22,24,28,86]. The cell cycle arrest and the ensuing downstream events of the cellular senescence program are orchestrated by complex circuits integrating several signaling pathways and networks, including: (1) the p14ARF/p53 and p16INK4a/pRB tumor suppressor pathways, two master regulator pathways of senescence that are activated by various cell-intrinsic and cell-extrinsic stresses in a parallel-(in human fibroblasts) or linear (in mouse fibroblasts) fashion; (2) the Ras/Raf/MEK/ERK/mTOR oncogene signaling pathway; (3) the PI3K/PTEN/Akt/mTOR nutrient-sensing signaling pathway; (4) the Wnt/HIRA/ASF1a/UBN1 chromatin remodeling pathway; and (5) the C/EBPβ-and NFκBgoverned senescence secretome transcriptional network [3, 4, 5, 12-14, 22, 33, 44, 45, 86, 117-138]. These signaling pathways: (1) transmit signals generated by sensor and effector proteins in response to individual senescence triggers (cell-intrinsic or cell-extrinsic) or to their combinations; (2) are linked via a series of connections; (3) are integrated into circuits by the p14ARF/p53 and p16INK4a/pRB master regulator pathways of senescence; (4) govern the spatiotemporal organization of the multistep cellular senescence program; and (5) elicit various hallmark features of the senescent phenotype [3,5,12,13,22,44,86,124]. A detailed description of the signaling circuitry characteristic of the cellular senescence program is beyond the scope of this review; the recent significant progress in this area has been comprehensively summarized elsewhere [3,5,12,22,44,86,124].

The complexity and spatiotemporal organization of the cellular senescence program
When mitotically active cells respond to excessive intracellular and extracellular stresses in culture or in vivo by entering a state of senescence, they undergo various morphological and functional changes to acquire a number of features (Tables 1 and 2). Some of these features are often observed not only in different types of cultured cells exposed to various triggers of either replicative or premature (stress-induced) mode of cellular senescence, but also in senescent cells derived from several organismal tissues. These features therefore are likely to serve as hallmarks of a state of cellular senescence and to be used as diagnostic biomarkers of cells entered such a state in different tissues. A body of evidence supports the view that these hallmarks/biomarkers of senescent cells may include: (1) cell enlargement and acquisition of a flat or spindle-like shape [5,22,139,140]; (2) cell cycle arrest (which is an irreversible process in vivo, but in culture can be reversed by certain genetic manipulations) [5, 7, 12, 97-99, 139, 141, 142]; (3) increased size and number of lysosomes, many of which are non-functional due to accumulation of lipofuscin-like indigestible molecular aggregates [140,[143][144][145][146][147][148]; (4) an elevated activity of senescenceassociated β-galactosidase (SA β-Gal) detectable at pH 6 [140,147,[149][150][151]; (5) an excessive proliferation of mitochondria that are elongated, interconnected to form an extensive network, aggregated, depolarized, dysfunctional, impaired in ATP synthesis, and produce excessive quantities of ROS [140,147,[152][153][154][155][156][157][158][159][160][161]; (6) a permanent establishment of DNA damage nuclear foci that are marked with a set of the DNA damage response (DDR) proteins and known as DNA segments with Aging and Disease • Volume 6, Number 1, February 2015 60 chromatin alterations reinforcing senescence (DNA-SCARS), DNA double-strands breaks (DSBs), senescence-associated DNA-damage foci (SDF), and telomere dysfunctioninduced foci (TIF) [4,12,22,25,111,140,[162][163][164][165][166][167][168]; (7) a formation of promyelocytic leukemia nuclear bodies (PML NBs) -also known as PML oncogenic domains (PODs) -that concentrate numerous DNA-binding proteins initiating heterochromatin establishment [103,124,[169][170][171]; (8) a PML NBs-instigated formation of senescence-associated heterochromatic foci (SAHF) that concentrate a set of heterochromatin-associated proteins [4,12,[172][173][174][175][176][177]; and 9) specific changes in cell transcriptome and the resulting stepwise development of a complex secretion pattern known as SASP and also called SMS [4,11,22,28,31,32,165] (Table 1). Importantly, none of the above features considered as probable hallmarks/biomarkers of senescent cells has been found to be common for all types cells entered a state of senescence in culture or in vivo. Therefore, only the simultaneous assessment of many of these features can identify cells entered such a state. Furthermore, it is conceivable that some of the features outlined in Table 2 can be elicited only in response to a particular trigger of a certain mode of cellular senescence and/or can be seen only in senescent cells confined to a specific tissue.  Expression of numerous SA-miRNAs is altered (either elevated or reduced) in cultured cells undergoing senescence caused by cell exposure to various triggers of either replicative or premature (stressinduced) mode of cellular senescence; many of these SA-miRNAs play essential roles in regulating senescence of cultured cells by targeting the signaling circuitry characteristic of the cellular senescence program; at least one of these SA-miRNAs, miR-22, can induce cellular senescence in vivo  ? ?
Recent findings strongly suggest that the numerous events characteristic of a state of cellular senescence (Tables 1 and 2) are organized into a multistep cellular senescence program. As outlined below in the section, the advancement of this program through spatially, temporally and mechanistically separable steps is orchestrated by complex circuits integrating several signaling pathways and networks (Figure 1).
Aging and Disease • Volume 6, Number 1, February 2015 63 Cells that entered a state of senescence in culture are resistant to apoptotic death elicited by certain proapoptotic stimuli [12,124,[196][197][198][199][200][201][202]. The resistance of these cells to apoptosis is due in part to elevated expression of a gene encoding the anti-apoptotic Bcl-2 protein [124,198,[203][204][205][206][207]. Transcription of this gene is activated by a phosphorylated form of the cAMP responsive element binding protein (CREB). Several senescence triggers increase the level of phosphorylated CREB by causing inactivation of the protein phosphatase PP2A (which dephosphorylates it) -thereby promoting transcription of the Bcl-2 gene known to be stimulated by phosphorylated CREB (Figure 1) [124,196,207,208]. Furthermore, the resistance of senescent cells to apoptosis can also be caused by the demonstrated ability of certain senescence triggers to repress transcription of a gene for the executioner caspases-3, perhaps by activating a yet-tobe-identified transcriptional repressor protein (Figure 1) [12,201]. Of note, the preferential recruitment of p53 to the promoters of certain cell-cycle arrest genes in cultured cells becoming senescent can weaken its ability to activate transcription of such pro-apoptotic genes as TNFRSF10b, TNFRSF6 and PUMA -thus also contributing to the resistance of senescent cells to apoptosis [12,209]. Formation of SAHF and the resulting global chromatin reorganization in cells entered a state of senescence is a multistep process [3,174,175,177,211]. It is initiated by chromosome condensation, which is driven by: (1) a pRB-orchestrated alteration of the cellcycle gene expression profile; (2) a replacement of histone H1 by HMGA (High Mobility Group A) proteins during an early step of chromatin condensation; and (3) a governed by the ASF1a/HIRA (anti-silencing function 1a/Histone Repressor A) protein complex generation of nucleosome-dense heterochromatin (Figure 1) [3, 172-177, 210, 212-214]. The ASF1a/HIRA complex is established after GSK3β (Glycogen Synthase Kinase 3 β) phosphorylates HIRA in the PML NBs; the activity of this histone chaperone is under negative control of the Wnt signaling pathway [3,177,215]. The associated with the condensed chromatin pool of histone H3 is then methylated to create binding sites for HP1γ (Heterochromatin Protein 1 γ; which is phosphorylated in the PML NBs) and the histone variant macroH2A, thereby finalizing the process of SAHF formation ( Figure 1) [3, 173-175, 177, 210].
The cell cycle arrest and stepwise epigenomic changes at early stages of the senescence program are followed by the establishment of a specific pattern of gene expression that results in secretion of numerous proteins and non-protein soluble compounds (Figure 1). Over time, senescent cells progress through a multistep process of developing a complex secretion pattern known as SASP and also called SMS (Table 1) [4,11,22,28,31,32,165]. An initial step in this process involves a transcriptional activation of genes encoding at least two early-response SASP/SMS proteins, specifically α and β isoforms of the multifunctional cytokine IL-1 (Figure 1). A cascade of the DNA damage response kinases ATM/CHK2 (ataxia telangiectasia mutated/checkpoint kinase 2) as well as a governed by yet-to-be-identified proteins chromatin remodeling activate transcription of these genes, whereas the p14ARF/p53 tumor suppressor pathway represses it [3,26,28,31,112,165,172,175,210,216]. A cell surface-bound pool of the IL-1α isoform then binds to its juxtaposed receptor IL-1αR, which in turn activates the downstream protein kinase IRAK1 to stimulate the transcriptional factors NFκB and C/EBPβ [4,22,26,30,112,120,123,[216][217][218]. These two factors activate transcription of genes encoding numerous lateresponse SASP/SMS proteins (including the proinflammatory cytokines IL-6 and IL-8 and their protein receptors, other pro-inflammatory cytokines and chemokines, growth factors, insoluble protein components of the extracellular matrix, and extracellular proteases) as well as the SA-miRNAs miR-146a and miR-146b ( Figure 1) [4,22,28,32,112,120,123]. The spatiotemporal organization of SASP/SMS is modulated by a positive transcriptional feedback loop involving NFκB and by a negative post-transcriptional feedback loop involving miR-146a/b (Figure 1) [4,30,112,219,220].

Pleiotropic effects of the multistep cellular senescence program on aging and cancer
Multiple mechanisms underlying the advancement of the cellular senescence program through temporally and spatially separable steps impose antagonistically pleiotropic effects on aging and cancer.
The growth regulator amphiregulin AREG is a procancer late-response SASP/SMS protein component that stimulates proliferation of premalignant epithelial cells, whereas the pro-cancer impacts of hepatocyte and fibroblast growth factors HGF and FGF are due to their stimulatory effects on pancreatic cancer cells migration and tissue invasion ( Figure 2) [4,25,28,29,226,229,230]. Such late-response SASP/SMS protein components as the vascular endothelial growth factor VEGF as well as chemokines IL-8, eotaxin and I-309 stimulate the migration and tissue invasion of endothelial cells -thereby promoting tumor-associated angiogenesis and facilitating cancer cell invasion and metastasis to distant sites [4,25,28,29,227,[231][232][233][234]. In contrast, the produced by senescent keratinocytes late-response SASP/SMS protein component maspin is a tumor suppressor that slows down angiogenesis by inhibiting the migration of endothelial cells ( Figure 2) [4,28,235,236].
Matrix metalloproteinases MMP-1, -2, -3, -10, -12, -13 and -14 are late-response SASP/SMS protein components that exhibit several antagonistically pleiotropic effects on aging and cancer (either beneficial or detrimental for the health of the organism) in a senescence trigger-and tissue context-dependent manner. By proteolytically degrading the extracellular matrix (ECM) proteins that are secreted by hepatic stellate cells or fibroblasts to form a fibrotic scar after acute liver injury or during skin wound healing (respectively), several MMPs impose an anti-aging effect by resolving the fibrotic mass -thus maintaining tissue integrity by promoting its repair and regeneration ( Figure 2) [4,25,28,122,131,237]. However, by proteolytically degrading the ECM proteins in other tissue contexts, the MMPs have a detrimental impact on organismal health as they impose: (1) a pro-aging effect by compromising the unique physical, biochemical and biomechanical properties of the tissue surrounding senescent cells and ultimately impairing tissue architecture and function; and (2) a procancer effect by facilitating the migration of tumor cells through the ECM, promoting the invasiveness of tumor cells and ultimately enabling their metastasis to distant sites ( Figure 2) [4,25,28,29,[238][239][240][241][242][243][244][245][246].

Conclusions and perspectives
A growing body of evidence supports the view that the complex relationship between mechanisms underlying aging and cancer evolves with organismal chronological age. Significant progress has been made in defining cellautonomous and cell-nonautonomous mechanisms that in young and adult organisms simultaneously delays aging Aging and Disease • Volume 6, Number 1, February 2015 66 and suppress tumor formation. Furthermore, it is now well established that the intricate interplay between mechanisms underlying aging and cancer reflects the proliferative history of cells and is impacted by the progression of a cellular senescence program through temporally and spatially separable steps. Recent findings imply that the advancement of the multistep cellular senescence program imposes antagonistically pleiotropic effects on aging and cancer. Mechanisms underlying some of these effects have emerged.
Despite an important conceptual advance in our understanding of the complex interplay between mechanisms underlying aging and cancer, we are still lacking answers to the following fundamentally important questions.
Which of the numerous morphological and functional changes observed in various types of senescent cells in culture and in vivo (Tables 1 and 2) are universal hallmarks of a state of cellular senescence -and, thus, which of these changes can be used as diagnostic biomarkers of cells entered such a state in any tissue? Which of these features of senescent cells are, in contrast, characteristic only of a certain senescence trigger, mode of cellular senescence or tissue? The use of genome-wide expression analyses and/or antibody arrays designed to detect various SASP/SMS protein components could facilitate the identification of both universal and tissuespecific senescence biomarkers.
Given that the progression of the cellular senescence program through temporally and spatially separable steps imposes antagonistically pleiotropic effects on aging and cancer (Figures 1 and 2), what therapeutic interventions have a potential to be used not only for enhancing those effects that are anti-aging and/or anti-cancer but also for attenuating those effects that are pro-aging and/or procancer? Recent findings in mice engineered for a reversal of the cellular senescence state by a drug-inducible telomerase reactivation [261] or for a late-life immune clearance of senescent cells by their drug-inducible elimination [262] suggest that small chemicals can be used for: (1) a protein target-specific pharmacological enhancement of the beneficial for organismal healthspan effects imposed by the cellular senescence program; and/or (2) a protein target-specific pharmacological attenuation of the deleterious for organismal healthspan effects inflicted by this program [5,7,8,[22][23][24].
Another attractive direction for future research is a temporal separation of the exogenously accelerated progression of the cellular senescence program from the pharmacologically triggered attenuation of its SASP/SMS at an advanced stage of SASP/SMS developmentthereby limiting chronic inflammation, enabling tissue repair and stimulating a targeted immune clearance of those senescent cells that have developed a harmful for organismal health version of this extracellular molecular signature [5, 7, 8,