A virus-acquired host cytokine controls systemic aging by antagonizing apoptosis

Aging is characterized by degeneration of unique tissues. However, dissecting the interconnectedness of tissue aging remains a challenge. Here, we employ a muscle-specific DNA damage model in Drosophila to reveal secreted factors that influence systemic aging in distal tissues. Utilizing this model, we uncovered a cytokine—Diedel—that, when secreted from muscle or adipose, can attenuate age-related intestinal tissue degeneration by promoting proliferative homeostasis of stem cells. Diedel is both necessary and sufficient to limit tissue degeneration and regulate lifespan. Secreted homologs of Diedel are also found in viruses, having been acquired from host genomes. Focusing on potential mechanistic overlap between cellular aging and viral-host cell interactions, we found that Diedel is an inhibitor of apoptosis and can act as a systemic rheostat to modulate cell death during aging. These results highlight a key role for secreted antagonists of apoptosis in the systemic coordination of tissue aging.


Introduction
In most metazoans, aging is characterized by a drastic decline in cellular and tissue function. This ultimate decline in function is driven by a multitude of complex, age-related cellular changes. The cellular and tissue changes that promote aging can be defined by (i) primary causes of cellular damage (such as genomic instability/DNA damage and telomere attrition), (ii) antagonistic responses to damage (such as changes in mitochondria function, cell death regulation/senescence, and changes in nutrient sensing), and (iii) integrative responses that eventually promote tissue dysfunction (such as a decline in regenerative potential, described in detail in [1]). Adding to this complexity is the fact that tissues do not age in isolation, as aging also involves changes at the level of intertissue communication. The interconnectedness of primary, antagonistic, and integrative responses in the context of tissue aging is likely to involve both autonomous and systemic mechanisms. This is highlighted by the systemic control of age-related changes in tissue regeneration.
Attenuation in regenerative potential is a defining characteristic of aged tissues and a primary driver of tissue dysfunction. This decline in regenerative potential originates from agerelated changes in the regulation of stem cell function, which include stem cell exhaustion and/or deficient proliferation and differentiation, as well as stem cell hyperproliferation [1][2][3]. Collectively, a loss of proliferative homeostasis within stem and progenitor cells during aging promotes tissue dysfunction and degeneration. Heterochronic parabiosis and heterochronic blood exchange experiments have intricately demonstrated that the systemic environment plays an immense role in controlling age-related changes in proliferative homeostasis [2]. Systemic or secreted factors from "old blood" can have inhibitory effects on young tissue stem cells, and although less robust, "young blood" can reverse the decline of stem cell function in aged tissues [4][5][6].
The precise nature of communication between unique tissues with diverse functions, during aging, remains unclear. However, the systemic coordination of tissue aging and regeneration, or the ability of age-related changes in one tissue to alter aging in distal tissues, is constricted by both deleterious and beneficial consequences.
Antagonistic mechanisms that respond to primary causes of age-related cellular damage also govern proliferative homeostasis to control tissue regeneration. This includes altered mitochondrial function and associated stress response pathways (such as the endoplasmic reticulum stress response [ER stress], the unfolded protein response [UPR], mitochondrial DNA [mtDNA] stress, and autophagy), cell senescence, altered programmed cell death/apoptosis, and changes in nutrient sensing (through insulin-like growth factor [IGF], insulin, and target of rapamycin [TOR] signaling), [7][8][9][10][11][12][13]. Critically, many of these signaling mechanisms and stress pathways promote both normal tissue regeneration and age-related tissue degeneration, highlighting the pleiotropic nature of antagonistic responses to cellular aging. Thus, uncovering secreted factors (such as cytokines, growth factors, and other small molecules) that both enable organ cross-talk and precisely modulate antagonistic responses to affect proliferative homeostasis is likely to reveal fundamental mechanisms driving the systemic coordination of tissue aging and longevity.
The utilization of these antagonistic stress response mechanisms is not limited to cellautonomous/nonautonomous control of tissue function and regeneration. Similar signaling mechanisms play critical roles in pathogen-host cell interactions, especially those related to viruses. For example, host cells can employ mtDNA stress, ER/UPR stress, cell senescence, and/or apoptotic responses as antiviral mechanisms [14][15][16][17]. In contrast, many viruses have developed mechanisms to suppress or modify these host responses to facilitate viral replication and spread. The coevolution of viruses and hosts has also promoted viral hijacking of host-secreted factors (encoded virokines) in order to alter cell signaling and cell communication to favor virus-infected cells [18][19][20]. The obvious mechanistic overlap between cellular aging and pathogen-host cell interactions suggests that pathogens may contain vital information associated with age-related changes in intercellular or intertissue communication.
To this end, we show here that a secreted cytokine named Diedel, which has also been acquired by viruses [21,22], is a critical regulator of systemic tissue aging. Utilizing the genetic capacity of Drosophila, we designed a model in order to characterize age-related systemic interactions between distal tissues that influence proliferative homeostasis, tissue degeneration, and longevity (described in detail in the section Hormetic responses from muscle-specific DNA damage attenuate systemic tissue aging and enhance lifespan). Using this fly model, we uncovered that Diedel, when secreted from muscle or adipose, can attenuate age-related intestinal tissue degeneration by promoting proliferative homeostasis of stem cells and that Diedel is both necessary and sufficient to limit tissue degeneration and extend lifespan. Focusing on potential mechanistic overlap between cellular aging and viral-host cell interactions, we found that Diedel and its viral homolog act as functionally conserved inhibitors of apoptosis. For the virus, virokine Diedel can limit apoptosis to facilitate viral abundance and spread. For the host, cytokine Diedel acts as a systemic rheostat to limit apoptosis-induced degeneration of the aging intestine. These results highlight a key role for secreted antagonists of apoptosis, specifically those acquired by viral genomes, in the systemic coordination of tissue aging.

Hormetic responses from muscle-specific DNA damage attenuate systemic tissue aging and enhance lifespan
In Drosophila, genetic manipulations in multiple individual tissues and unique cell types have been shown to impact longevity. In regard to systemic aging, muscle and fat body (adipose) appear to be critical effector/modulator tissues, and the brain can function as a control center though the regulation of secreted hormones [23][24][25][26]. The fly intestine is also central to lifespan. The Drosophila adult midgut (intestine) is regenerated by a population of intestinal stem cells (ISCs), and in a young midgut, ISCs divide rarely [27][28][29]. However, during aging, changes in activity of a multitude of stress response pathways leads to stem cell hyperproliferation/misdifferentiation and epithelial dysplasia [30]. This results in tissue degeneration, which ultimately limits lifespan [31].
In order to uncover novel systemic mechanisms and communication axes that govern age-related changes in proliferative homeostasis, we tested multiple genetic strategies in order to induce tissue-specific and age-dependent tissue damage/stress within thoracic skeletal muscle, with the aim to assay integrative responses to primary causes of aging in distal tissues. The muscle has emerged as a critical regulator of longevity in Drosophila, and muscle injury in mammals has been suggested to influence regeneration in distal organs [24,32,33]. We discovered that mild attenuation of DNA repair within skeletal muscle leads to systemic responses that limit age-related intestinal degeneration and extend lifespan. Using RNA interference (RNAi)-mediated targeting of the Mei-9-ERCC1 heterodimeric DNA repair nuclease (S1 Fig and [ 34,35]) specifically in adult thoracic skeletal muscle (using Act88FGal4, active in longitudinal indirect flight muscle but absent from other muscle found in the carcass and intestine [36]), we found that inhibiting DNA repair can dose-and age-dependently promote DNA damage, ultimately leading to muscle that presents hallmarks of tissue damage and premature aging (loss of proteostasis, Fig 1A and 1B). Attenuating Mei-9 alone (Act88FGal4>UAS-Mei-9 RNAi ) results in mild up-regulation of DNA damage and age-related increases in markers of muscle tissue damage/stress when compared thoracic muscle of young (10 d) Act88FG4>+(w 1118 ) controls, and flies with DNA repair attenuation specifically in thoracic muscle (mu-specific, Act88FG4>UAS-Mei-9 RNAi or Act88FG4>UAS-Mei-9 RNAi , UAS-ERCC1 RNAi ); assayed by phospho-H2aV immunostaining (red), counterstained with phalloidin ("Phall"; green, actin filaments) and DAPI (blue). Representative images shown. (B) Immunostaining to detect poly-ubiquitin protein (aggregates; "Poly-Ub.") in dissected longitudinal thoracic muscle from young (10 d) and old (30 d) flies, genotypes described above; anti-poly-ubiquitin (green), counterstained with to controls (Act88FGal4>+[w 1118 ], Fig 1A and 1B and S1C Fig). Despite these subtle changes in muscle aging, there was no indication of profound damage to the tissue, in regard to muscle structure, function, or feeding behavior (previously linked to muscle aging, Fig 1A  and 1B and S4A and S4B Fig). However, concurrently attenuating both Mei-9 and ERCC1 (Act88FGal4>UAS-Mei-9 RNAi , UAS-ERCC1 RNAi ) leads to robust increases in both DNA damage and apparent tissue damage, highlighted by a premature loss of proteostasis and disruption of myofibril muscle structure during aging (Fig 1A and 1B).
Beneficial effects of adaptive responses to low doses of cellular stress or damage (termed hormesis) have been linked to systemic aging, including stress associated with genomic instability [37,38]. We found that mild increases in muscle DNA damage (Act88FGal4>UAS-Mei-9 RNAi ) also increase mean adult lifespan and decrease age-specific mortality, although the impact on mortality rate was slightly reduced later in life (Fig 1C, S2A Fig and S1 Table). In order to confirm these longevity results, we generated a thoracic muscle-specific GeneSwitch driver (Act88FGS) for adult-specific and mifepristone (RU486)-inducible gene expression (S1E-S1L  Table). Muscle-specific DNA damage can also systemically attenuate intestinal tissue degeneration during aging. Inhibiting Mei-9 function in muscle limits age-related ISC hyperproliferation and epithelial dysplasia (Fig 1E and 1F). Utilizing a split-lacZ reconstitution assay to heritably label-select ISCs (clones), we also found that muscle-specific DNA damage limits clone growth (cells per clone) in aged midguts (Fig 1G and 1H), further highlighting a preservation of proliferative homeostasis. Female flies are used exclusively for all assays (except for survival analysis) because of sexual dimorphism related to midgut regeneration [39]. Alternatively, strong enrichment of muscle-specific DNA damage (Act88FGal4>UAS-Mei-9 RNAi , UAS-ERCC1 RNAi ) does not positively impact age-related tissue degeneration or lifespan (S4C and S4D Fig and S2 Table), and inhibiting DNA repair in other tissues, such as directly in the midgut, leads to accelerated aging (S4E- S4G Fig and S1 Table). These results highlight the dose dependency and tissue specificity associated with DNA damage-mediated hormesis.
Taken together, our data show that mild enhancement of DNA damage in muscle can drive systemic hormetic responses to promote proliferative homeostasis of distal regenerating tissues during aging and increase longevity.

Cytokine Diedel governs systemic tissue aging and lifespan
We next wanted to use this genetic model of hormesis, and putative muscle-gut communication axis, to uncover systemic factors/mechanisms that modulate age-related changes in proliferative homeostasis. To this end, we generated genome-wide expression profiles from phalloidin (red, actin filaments). Representative images shown. (C-D). Survival curves (lifespan, female flies) associated with mu-specific inhibition of Mei-9 using (C) the Act88FGal4 driver (compared to Act88FG4>+[w 1118 ] controls) or (D) a GeneSwitch inducible driver (Act88FGS, +RU486 compared with −RU486 [vehicle alone] sibling controls). (E) Quantification of mitoses per whole dissected midgut (assayed by anti-pH3 immunostaining) at indicted ages, genotypes described above; bars represent mean ± SE, n = 25-30. (F) Immunostaining of dissected intestines to assess epithelial integrity of posterior midguts at indicated ages, genotypes described above; pH3 (green), armadillo ("Arm"; membrane, red), and DAPI (blue). Representative images shown. (G-H) Lineage tracing from ISCs using FRT recombination of a split alpha-tubulin-lacZ transgene (in Act88FG4>+[w 1118 , controls] or Act88FG4>UAS-Mei-9 RNAi genetic background). (G) Changes in clone size (cell per clone form posterior midgut) at indicated ages; represented as box plot (median, red line), n = 25. (H) Representative images of lacZ clones from various genotypes at indicated ages, immunostaining of dissected midguts (posterior), anti-lacZ (green), and DAPI (blue). (I) Venn diagrams showing overlap of up-regulated genes (from dissected midguts) between Act88FG4>UAS-Mei-9 RNAi  Underlying data can be found in S1 Data. See also S1 and S2 Figs and S1 and S2 Tables. FRT, flippase recombination target; ISC, intestinal stem cell; pH3, phospho Histone H3; mu-specific, muscle-specific; RPKM, reads per kbp per million reads; RU486, mifepristone. dissected thoraces (enriched in indirect flight muscle) and intestines of young and old flies with muscle-specific attenuation of DNA repair (S4H and S4I Fig). In midguts, we found a significant overlap in age-related intestinal transcriptome changes (focusing on up-regulated gene expression) between Act88FGal4>UAS-Mei-9 RNAi flies and controls (Fig 1I), despite the strong rescue of intestinal epithelial dysplasia in animals with Mei-9 inhibition in muscle. Gene Ontology (GO) clustering analysis revealed that the reason for this overlap is due to overrepresentation of GO terms associated with innate immune responses. Chronic activation of innate immunity in the Drosophila intestine during aging can promote tissue degeneration [40], and age-related up-regulation of innate genes occurs in both genotypes (select genes highlighted in Fig 1J and confirmed by quantitative real-time PCR [qRT-PCR] in S4J and S4K Fig). These data suggest that the maintenance of proliferative homeostasis in flies with musclespecific DNA damage is not due to a general suppression of age-related systemic inflammatory responses and is likely achieved through more direct mechanisms.
We turned to thoracic/muscle transcriptomes in order to identify genes or secreted factors the might mediate muscle-gut communication during aging. We discovered a small subset of genes that are up-regulated uniquely in aged Act88FGal4>UAS-Mei-9 RNAi thorax/muscle samples (with minimal or no induction in controls samples, Fig 2A). One gene in particular, named Diedel, was of distinct interest because it encodes a small protein with a putative signal peptide, and a recent report has shown that Diedel is a secreted factor (cytokine) found in circulation (hemolymph, [21]). Gene expression profiles in Drosophila have previously linked Diedel transcriptional induction to sterile injury, septic infection, viral infection, and tissue damage [21,[41][42][43][44][45][46]. In our model, Diedel expression is strongly up-regulated in Act88F-Gal4>UAS-Mei-9 RNAi thorax/muscle (and also slightly up-regulated in controls) during aging ( Fig 2B). We thus asked whether Diedel expression from muscle could phenocopy the systemic hormetic responses induced by muscle-specific DNA repair inhibition.  Fig 2D). Furthermore, muscle-specific Diedel expression also inhibits age-related ISC hyperproliferation and epithelial dysplasia (Fig 2E and 2F) and increases mean lifespan (utilizing both Gal4 and GeneSwitch systems, Fig 2G and 2H, S2B and S3C Figs, S1 and S2 Tables).
These data thus suggest that Diedel is an important systemic factor governing tissue aging and longevity. Because this gene can be induced by a wide variety of cellular and tissue stress (some related to aging), we wanted to determine if Diedel is important for normal aging, independent of our hormetic model. Surprisingly, Diedel knock-down in muscle (using RNAimediated targeting of the Diedel, UAS-Die RNAi ) had no impact on systemic intestinal tissue aging (S5B and S5C Fig). We hypothesized that Diedel was either not required for normal aging or may be expressed in multiple tissues during aging. To answer this question, we generated Diedel reporter flies (using a 366-base-pair region located upstream of Diedel transcription start site, linked to red fluorescent protein [RFP]; Diedel P -RFP) to monitor Diedel expression changes in aging flies. Diedel reporter activity is off in young animals, but during aging, mild reporter activity is found in the muscle, while strong activation is observed in carcass fat body (muscle and fat body activation appear concurrently in the same animal, Fig 3A  and 3B). Populations show a significant age-dependent increase in Diedel P -RFP animals, reaching nearly 100% of surviving flies later in life ( Fig 3C). Similar affects are seen in our model of hormesis. DNA repair attenuation in muscle leads to reporter activation in the carcass fat body, although at a much accelerated rate over controls (S5D Fig), and age-related changes in Diedel gene expression are even stronger in fat body (Fig 3D and S5E Fig) when compared with the thorax/muscle (Fig 2B). Muscle tissue stress is thus likely to autonomously and nonautonomously (in fat body) promote Diedel expression and secretion [41].
Unlike the muscle, Diedel expression in fat body is required for normal systemic tissue aging and longevity. Diedel knock-down in fat body (using CGGal4 and PplGal4 coupled with RNAi) leads to accelerated aging of the intestine, in regard to ISC hyperproliferation and epithelial dysplasia (Fig 3E and 3F and S5F Fig). Similar to muscle, overexpression of Diedel in fat body also attenuates the age-related loss of proliferative homeostasis in the midgut ( Fig 3E) and does not impact feeding behavior (S5G Fig). Diedel is also required for normal longevity, as inhibiting Diedel in fat body or ubiquitously in the adult animal (using the TubGeneSwitch driver) shortens fly lifespan and increases age-specific mortality (Fig 3G and 3H, S3D Fig, S1 and S2 Tables). Ubiquitous or fat body-specific overexpression of Diedel also shortens lifespan, despite the attenuation of intestinal dysplasia (S1 and S2 Tables). Previous reports have (F) Immunostaining of dissected intestines to assess epithelial integrity of posterior midguts at indicated ages, genotypes described above; pH3 (green), armadillo ("Arm"; membrane, red), and DAPI (blue). Representative images shown. (G-H) Survival curves (lifespan, female flies) associated with mu-specific expression of Diedel using (G) the Act88FGal4 driver (compared to Act88FG4>+[w 1118 ] controls) or (H) a GeneSwitch inducible driver (Act88FGS, +RU486 compared to −RU486 [vehicle alone] sibling controls). Underlying data can be found in S1 Data. See also S1 and S2 Figs and S1 and S2 Tables. Ecc15, E. carotovora carotovora 15; FRT, flippase recombination target; ISC, intestinal stem cell; mu-specific, muscle-specific; qRT-PCR, quantitative real-time PCR; pH3, phospho Histone H3; RPKM, reads per kbp per million reads; RU486, mifepristone.  shown that the balance between normal regeneration and dysplasia in this tissue determines longevity [31]. Furthermore, Diedel overexpression using the muscle-specific GeneSwitch driver does not obviously impact mortality rates, notwithstanding the influence on mean lifespan (S3C Fig). We thus hypothesized that strong overexpression of Diedel might negatively impact normal regeneration and variably impact longevity. In support of this hypothesis, we found that levels of Diedel expression (using the TubGS driver and various doses of RU486) can positively or negatively influence longevity and that ultimately, Diedel can dose-dependently extend mean lifespan and impact age-specific mortality (Fig 3I and 3J, S3E and S3F Fig  and S3 Table).
In summary, these data show that the secreted factor Diedel is a critical regulator of systemic aging. As a critical regulator of tissue communication during aging, this cytokine is both necessary and sufficient to prevent the loss of proliferative homeostasis in the aging intestine, as well as influence longevity.

Cytokine and virokine Diedel antagonize apoptosis
We next wanted to explore potential mechanisms by which cytokine Diedel regulates systemic tissue regeneration that could can impact aging. A previous study showed that ubiquitous repression of this cytokine (using a mutant) leads to up-regulation of nuclear factor kappa B (NFkB; Drosophila Relish) target gene expression, highlighting that Diedel may directly influence NFkB transcriptional activation function [21]. While we observed some Diedel-dependent phenotypes associated with attenuation of NFkB function in the midgut (S5H and S5I Fig), NFkB, however, is not required for bacterial-induced regeneration [47], and presumably homeostatic clone formation as well. Thus, we hypothesized that this cytokine likely impacts age-related tissue regeneration through alternative/parallel or partially overlapping mechanisms. In order to sort through the numerous signaling pathways and stress responses that modulate regeneration in the fly intestine, which could be influenced by Diedel, we turned to Diedel homologs for guidance. Although there are no obvious homologs in vertebrate or mammalian genomes, Lamiable and colleagues previously highlighted that homologs of Diedel are found in unrelated families of DNA viruses that primarily infect lepidoptera (S6A Fig and [21,48]). During evolution, DNA viruses have acquired an array of cellular genes from various host genomes. Termed virokines, these often-secreted proteins may act as a mimetics or antagonists of their host homologs, altering cell signaling to promote viral infection [18][19][20]49]. Due to the previously mentioned overlap in host signaling mechanisms used to either (i) govern stress-related tissue regeneration or (ii) control antiviral responses (Fig 4A), we decided to explore virokine Diedel function in order to inform on the host's gene's role in influencing tissue degeneration during aging. First, we generated upstream activating sequence (UAS)dependent transgenic flies capable of expressing either a full-length virokine Diedel (ORF121 from Spodoptera frugiperda Ascovirus 1a [SfAV-1a]) or a recombinant version in which the putative signal peptide from the viral protein was replaced with the Drosophila signal peptide sequence (REC-ORF121, Fig 4B). Overexpressing either ORF121 or REC-ORF121 from fat body strongly attenuated midgut regenerative responses induced by bacteria (Fig 4C and 4D), while overexpressing REC-ORF121 from muscle inhibited regenerative responses during infection and aging (Fig 4E). ORF121 expression in muscle did not impact these responses (Fig 4E), suggesting that the full-length virokine may not be readily secreted from all types of host cells. These data suggest that the viral homolog of Diedel may mimic function of the host gene and thus can inform on its function.
Virokines have been shown to influence various host responses in order to limit antiviral mechanisms and promote infection, including programmed cell death (including apoptosis), inflammation, and growth-factor signaling ( Fig 4F and [50][51][52][53][54]). We next investigated virokine Diedel function in relation to viral infection. Utilizing a previously described in vitro infection model of Drosophila S2 cells with vesicular stomatitis virus (VSV, [55,56]), we assayed the ability of secreted virokine Diedel to impact viral infection. Transfection of expression plasmids containing tagged host Diedel or REC-ORF121 in S2 cells revealed that these proteins are readily secreted into media (S6B and S6C Fig). However, as predicted from in vivo experiments (Fig 4E), tagged full-length ORF121 secretion from these cells is less efficient (ORF121-conditioned media, S4B and S4C Fig). Still, we found that conditioned media containing secreted ORF121 markedly increased the abundance of VSV-G and VSV-M RNA in infected S2 cells (monitored by qRT-PCR of viral RNA, Fig 4G and 4H). These results suggest that Diedel may increase viral replication on a per-cell basis or inhibit apoptosis, the latter of which is an important host response to viral infection to minimize both viral replication and spread. To explore a role for Diedel in limiting apoptosis, we employed an immunofluorescence approach to coordinately monitor viral abundance and TUNEL positivity. In control experiments, many VSV-GFP-positive S2 cells were observed undergoing apoptosis (monitored by TUNEL-positive nuclei, Fig 4I). However, ORF121 promoted striking increases in the amount of cellular VSV-GFP but an absence of TUNEL staining. These data suggest that secreted virokine Diedel can antagonize apoptosis to increase viral abundance and likely promote spread.
We thus further investigated the role of host cytokine Diedel in the regulation of apoptosis. Still utilizing S2 cells, we assayed the ability of secreted Diedel (Die V5 -conditioned media, S6B- S6E [7,13]. Dronc, Hid, Rpr, and Grim can be transcriptionally induced [57,58]. We found that Diedel strongly attenuated UV-induced Dronc expression but not Hid, Rpr, Grim (only Rpr is shown), or Puckered (Puc, a classical JNK target gene) expression (Fig 5A-5C), at least suggesting that Diedel does not impact JNK-mediated apoptosis. Furthermore, Diedel is transcriptionally induced after UV-treatment in S2 cells, and Diedel inhibition (using double-stranded RNA [dsRNA]) significantly enhances UV-induced Dronc expression (Fig 5D and 5E). This cytokine also attenuates the temporal cleavage (activation) of Cas-3 after UV-treatment (visualized by western blots utilizing a mammalian cleaved Cas-3 antibody that is a marker for both effector caspase activity and Dronc activity in Drosophila [59], Fig 5F and 5G). Immunostains of S2 cells further revealed that secreted Diedel almost completely blocks the UV-induced activation/induction of effectors caspases (Cas-3 and Dcp-1, Fig 5H and 5I). We also monitored end points of apoptotic pathway activation (5 h after UV treatment) and found that Diedel can potently block DNA fragmentation assayed by TUNEL staining and gel electrophoresis (Fig 5J and 5K). Diedel purified from conditioned media has similar effects ( Fig 5L). Diedel did not influence NFkB target gene expression at steady state or acutely in response to UV in S2 cells, suggesting that NFkB is not required for these apoptotic responses (S6F-S6J Fig).

Diedel systemically governs apoptosis during aging
Apoptosis is also required for normal tissue regeneration, but misregulation of apoptotic responses is increasingly linked to tissue aging [61,62]. Apoptosis has thus emerged as a critical antagonistic response to primary causes of cellular damage during aging and must be precisely modulated to ensure proper tissue architecture and epithelial function. In the Drosophila intestine, apoptotic responses in enterocytes are required for both homeostatic regeneration/ tissue growth and pathogen-induced regeneration through the production of mitogenic signals [63][64][65][66]. These paracrine signals drive stem cell proliferation and differentiation. Critically, overactivation of apoptotic responses in enterocytes promotes hyperproliferation and tissue degeneration [63]. Thus, we next wanted to determine if secreted Diedel could influence apoptosis in vivo in the aging midgut. Similar to our in vitro findings, Diedel expression in muscle (Act88FGal4>UAS-Diedel) inhibits age-related increases in intestinal Dronc expression but not JNK-target genes (Fig 6A). We also found a marked increase in apoptosis and cell death in the aging midgut (assayed by TUNEL staining and Cas-3 immunostaining), and muscle-specific Diedel expression significantly blocks these age-related increase in apoptosis (Fig 6B-6D and S7A-S7C Fig), which correlates with an attenuation of ISC hyperproliferation and tissue dysplasia ( Fig 2E). Alternatively, Diedel knock-down in fat body (using CGGal4 and PplGal4 coupled with RNAi) leads to hastened induction of apoptosis and cell death in the aging midgut (Fig 6E and S7D-S7H Fig), which correlates with accelerated aging of the intestine ( Fig  3E). Furthermore, secreted Diedel can inhibit cytokine (unpaired 3 [Upd3])-mediated Janus kinase (Jak)/signal transducer and activator of transcription proteins (Stat) signaling, which in turn regulates stress-and age-related stem cell proliferation/midgut regeneration (reviewed in [67]). Both in vitro and in vivo, Diedel indirectly attenuates Jak/Stat activation by repressing stress-induced Upd3 transcriptional up-regulation (S8A- S8J Fig). We also confirmed that apoptotic responses can drive Upd3 induction and Jak/Stat activation (S6E Fig). In summary, these data show that cytokine Diedel can systemically modulate apoptosis, and subsequently control regenerative signals, during tissue aging.
Combined with our other findings, our data suggest that Diedel governs apoptosis to ultimately control proliferative homeostasis, tissue degeneration, and aging ( Fig 6F).

Discussion
Overall, our results reveal that primary causes of cellular aging can have tissue-autonomous and systemic effects on integrative responses that drive age-related phenotypes, highlighting a Actin), bars represent mean ± SE, n = 3. (H-I) Images of (H) cleaved Cas-3 ("Cas 3") or (I) cleaved Dcp-1 ("Dcp1") immunostaining in S2 Cells exposed to UV (100 mJ/cm 2 ) and treated with mock or Die V5 conditioned media ("CM"). (H) Cas-3 or (I) Dcp-1 (green), phalloidin ("Phall"; membrane, red), DAPI (blue). Bottom-right panel shows (5× zoom) Cas-3+ or Dcp-1+ cells in mock-conditioned media and UV-treated cells. Representative images shown. (J) Images of TUNEL immunostaining in S2 Cells exposed to UV (100 mJ/cm 2 ) and treated with mock or Die V5 conditioned media ("media"); TUNEL (nuclear, green), DAPI (blue), phalloidin (membrane, red). Bottom panels show TUNEL+ nuclei (white) with quantification; represents mean percent of TUNEL-positive nuclei (to total cells) ± SEM, n = 20-30. (K-L) Genomic DNA samples were isolated from S2 cells after UV exposition and treatment with (K) mock or Die V5 conditioned media (2 of 3 independent samples shown) or (L) purified Die V5 -tagged protein and separated by agarose gel electrophoresis. (M-N) Images of (M) cleaved Cas-3 ("Cas 3") or (N) TUNEL immunostaining in MEFs exposed to UV (100 mJ/cm 2 ) and treated with mock or Die V5 CM. (M) Cas-3 or (N) TUNEL (nuclear, green), phalloidin (actin filaments, red), DAPI (blue). Bottom-right panel shows isolated (white) Cas-3+ or TUNEL+ cells in mock-conditioned media and UVtreated cells. Representative images shown. Underlying data can be found in S1 Data. See also S4 and S6 Figs  systemic coordination in tissue aging. Furthermore, we uncovered a systemic signaling mechanism (mediated by the cytokine Diedel) that is capable of promoting communication between muscle/adipose and the regenerating midgut during aging in Drosophila. This cytokine is both necessary and sufficient to limit age-related tissue degeneration and extend lifespan, and our data highlight Diedel as a critical regulator of aging and longevity.
Diedel appears to impact tissue aging through the attenuation of apoptosis. Apoptosis is the major programmed cell death pathway used to remove unneeded or harmful cells during development or tissue maintenance/regeneration. Increasing evidence also suggests that apoptosis may be a driving force of tissue aging, albeit extremely pleiotropic. Too little can promote cancer or age-related (and harmful) cell senescence, while too much can promote tissue atrophy or hyperplastic overgrowth through compensatory proliferation [61,62]. The loss of proper apoptotic responses during aging is thus likely to be influenced by changes in both proand antiapoptotic mechanisms. An abundance of precisely regulated mammalian cytokines act as critical proapoptotic signals to control various programmed cell death pathways, especially those related to immune cell regulation (reviewed in [60]). Diedel emerges as an intriguing antagonist of these responses, as it is a secreted factor under tight transcriptional control (similar to other stress response genes and cytokines), allowing it to function as a systemic rheostat to properly balance apoptosis.
While Drosophila Diedel is able to attenuate activity of the apoptotic pathway in both fly and mammalian cells, the underlying mechanism remains unclear. Apoptosis is governed by both intrinsic and extrinsic pathways. Extrinsic receptor-induced signaling (triggered at the cell surface by ligands such as TNFα and IL-1β) and the release of intrinsic apoptogenic factors (triggered by changes in the permeabilization of mitochondrial membranes) ultimately converge on activation of downstream caspases [68][69][70]. JNK, through the induction of mitochondrial-localized IAP antagonists, plays a crucial role in the intrinsic activation of apoptosis, especially in Drosophila [7,58,71]. Our data suggest that secreted Diedel does not influence JNK activation after cell damage, implying that this cytokine attenuates apoptosis through an extrinsic pathway and putative unknown receptor. Diedel's potentially unique role in regulating apoptosis may influence other signaling pathways as well, such as immune/NFkB responses [21], known to share common control elements [72]. Furthermore, Diedel may also directly influence NFkB in certain cell types (39) and thus impact apoptosis through governing NFkB activity. A more detailed spatiotemporal analysis of Diedel molecular function is required to understand how various signaling pathways intersect to control outputs. The exact fate of Diedel-treated cells also needs further examination, as it remains unclear if these cells are just temporally blocked from cell death or display more permanent features that could be beneficial, detrimental, or likely both. Exploring this protein's unique function is likely to inform on how secreted antagonists of apoptosis precisely modulate cell death in concert with extrinsic and intrinsic activators. Furthermore, uncovering the underlying mechanism by which Diedel influences apoptotic responses will also shed light on the complex spatiotemporal role of programmed cell death pathways in aging phenotypes.
Our data show that Diedel expression increases with age in Drosophila and that the cytokine limits age-related induction of apoptosis and concurrent loss of proliferative homeostasis in the aging intestine. However, Diedel can be induced in response to various stressors, suggesting a broad role in regulating apoptosis beyond just that associated with midgut regeneration. Our data also imply that Diedel, and antiapoptotic mechanisms in general, may be critical for beneficial hormetic responses. Age-related hormetic effects induced by various antagonistic stress responses appear to have diverse systemic components, which are likely to exert broad effects on cellular targets [32,37,[73][74][75]. Because change in the regulation of apoptosis is a general feature of many aging metazoans, Diedel may represent an important secreted factor that systemically communicates various types of cell damage/stress between tissues during aging, generally eliciting beneficial effects.
Diedel's potentially broad role in regulating apoptosis is perhaps best highlighted by its viral counterparts. Diedel, like other host genes from various species-including humans-has been hijacked by viral genomes. During evolution, DNA viruses have acquired a broad array of genes from hosts in order to shift virus-host cell interactions toward survival of virusinfected cells [18,19,21,53,76,77]. Encoded virokines often act as mimetics or antagonists of their cellular homologs, and we find that virokine Diedel can also limit apoptosis during viral infection, as well as attenuate age-related tissue degeneration in vivo. These data show that viral genomes may contain vital information associated with underlying mechanisms that govern aging and can perhaps be exploited to uncover novel modulators and potential therapeutics related to tissue aging. Using Diedel as an example, secreted virokines have been proposed as putative immunomodulatory therapeutics [19,78], and there is a long history of targeting cell death (apoptosis) pathways for drug development related to cancer and other age-associated diseases [79,80]. Finally, while most virus-acquired host genes are linked to evolution of pathogenic virus-host cell interactions, mutualistic relationships also exist in which viruses are beneficial to their hosts [81,82]. Thus, secreted viral proteins emerging from symbiotic or "good" viruses might also have the capability to influence host-tissue homeostasis during stress or aging, similar to mutualistic microorganisms.

Drosophila stocks and culture
The following strains were obtained from the Bloomington Drosophila Stock Center: w1118, Act88FGal4 (38461). UAS-ERCC1RNAi (110419) was obtained from the Vienna Drosophila RNAi Center. UAS-MEI9RNAi was a gift from J. Sekelsky, CGGal4 was a gift from C. Thummel, PplGal4 was a gift from M. Pankratz, TubGeneSwitch was a gift from H. Jasper, NP1Gal4 was a gift from D. Ferrandon, and 10xSTAT-GFP was a gift from E. Bach. UAS-DiedelRNAi transgenic lines were generated with plasmid (dna571) from the Vienna Drosophila RNAi Center. UAS-Diedel, UAS-ORF121, UAS-REC-ORF121, Diedelp-RFP, Act88FGeneSwitch, and UAS-DiedelRNAi transgenic flies were generated for this study. NOTE: All flies used in this study were backcrossed ×10 into the w1118 background, with continued backcrossing every 6-8 mo to maintain isogenecity.
All flies were reared on standard yeast-and cornmeal-based diet at 25˚C and 65% humidity on a 12-h light/dark cycle, unless otherwise indicated. The standard lab diet (cornmeal-based) was made with the following protocol: 14 g Agar/ 165.4 g Malt Extract/ 41.4 g Dry yeast/ 78.2 g Cornmeal/ 4.7 ml propionic acid/ 3 g Methyl 4-Hydroxybenzoate/ 1.5 L water. For RU486 food, RU486 or vehicle (ethanol 80%) was mixed with food, resulting in a 200 uM concentration of RU486 in the food, unless otherwise indicated.
All analysis was exclusively done in female flies (with the exception of lifespan assays) because of sex-specific differences in midgut regeneration.

Generation of transgenic Drosophila
UAS-Diedel flies were generated by PCR amplification of adult Drosophila (w1118) cDNA, with specific primers, and then cloned into the pUASt plasmid. DiedelP-RFP flies were generated by PCR amplification of Diedel promoter sequence with specific primers from w1118 genomic DNA and clone in pB-RFP plasmid [83]. Act88FGeneSwitch flies were generated by PCR amplification of Act88F promoter (from w1118 flies, including TATAbox) and cloned into pP(wlo+hsinGs) plasmid without the hsp minimal promoter [84]. UAS-DiedelRNAi transgenic lines were generated with plasmid (dna571) from the Vienna Drosophila RNAi Center (in the pUASt plasmid).
Confocal images were collected using a Nikon Eclipse Ti confocal system (utilizing a single focal plane) and processed using the Nikon software and Adobe Photoshop.

Lifespan analysis
Thirty virgins were crossed to 10-15 males of the desired genotype for all survival experiments unless otherwise indicated (see below). After initial mating, parental flies were allowed 2 d to lay eggs before being removed from bottles. Progeny of crosses was collected for 3-4 d after initial eclosion. Collected progeny were allowed to mate for 2 d at room temperature. Progeny were then separated according to sex and genotype into cages (around 75-100 flies/cage) and aged at 25˚C with constant humidity (approximately 65%).
In order to normalize population densities for RU486 dose-dependency longevity experiments (Fig 3I and 3J), 75 virgin flies were mated with 50 males and allowed to lay eggs for less than 24 h on apple plates, supplemented with live yeast paste. Eggs were collected from the apple plate by washing in 1X PBS solution into 15 ml tubes. Then, 18-20 μL of egg suspension was dispensed into 200-mL bottles containing standard lab food. Two days after eclosion, once-mated flies were transferred to plastic cages at a density of 100-120 flies per cage.
For all independent populations, plastic cages (175-ml volume, 5-cm diameter from Greiner Bio-One) were used for lifespan experiments. Food, changed every 2-3 d, was provided in vials inserted into a foam plug (4.9 cm in diameter, 3-cm thick from Greiner Bio-One). For RU486-dependent experiments, RU486 or vehicle (ethanol, 80%) was mixed with food, resulting in a 200 μM concentration of RU486 in the food (unless otherwise indicated for dose response experiments). Dead flies were counted every 2-3 d. Survival data were analyzed using Prism statistical software.
All independent lifespan analyses for independent trials are included in the S2 Data.

Mortality estimation
Mortality rate was calculated as described in [85]. N0 is the number of individuals in the initial cohort, as well as Nx, the number alive at the start of each day. The probability of surviving from age x to age x + 1, given the individual is alive at the start of age x, is Px = Nx + 1/Nx. The age-specific rate of mortality is estimated as μx = −ln(Px). The mortality rate is plotted as ln (μx).

ISC lineage analysis
The previously developed marked lineage system was used to generate clones of lacZ-expressing cells [86]. In this system, flippase-mediated recombination fuses the α-tubulin promoter to the lacZ gene, allowing transcription of lacZ. In the absence of heat shock, X15 flies carry 2 inactive tubulin-promoter lacZ (X-15-29 and X-15-33). Upon heat shock induction of flippase in dividing cells, these transgenes can recombine through the flippase recombination target (FRT) region, generating an active lacZ transgene. Thereafter, cells carrying the recombined lacZ transgene, as well as its progeny, will be marked by constitutive production of ß-galactosidase. The size of the marked clone is a direct measure of the division rate. To induce somatic recombination, adult flies were heat-shocked at 10 d or 30 d of age at 37˚C for 60 min. Midguts were dissected 5 d after heat shock and stained with anti-LacZ antibody.

In vivo TUNEL (cell death) assay
Intact midguts were dissected (at indicated ages) in 1× PBS and fixed in 4% paraformaldehyde for 25 min at room temperature. Samples were washed with 1× PBT for 5 min. Cell death was detected by using In Situ Cell Death Detection Kit (Roche) according to manufacturer's instruction.

qRT-PCR analysis
Total RNA from dissected intact fly thorax, carcass (all of the eggs and intact intestines were removed) and midguts, or S2 cells and MEFs, were extracted using Trizol and complementary DNA synthesized using Superscript II (Invitrogen). RT-PCR was performed using SYBR Green (biorad), the Applied Biosystems StepOnePlus Real-Time PCR systems, and the primer pairs described in the extended experimental procedures (see S4 Table). Results are average ± standard error of at least 3 independent samples, and quantification of gene expression levels calculated using the ΔCt method and normalized to actin5C for all in vivo experiments or rpl32 for all S2 cell in vitro experiments (virus and UV-treatment) or GAPDH for all MEF in vitro experiments. The standard curve method was used to quantify dose-dependent Diedel mRNA.

RNA-seq analysis
Intact fly thorax and midguts were dissected (at indicated ages) in PBS. Total RNA was extracted using Trizol reagent and used as template to generate sample libraries for RNA sequencing (using the TruSeq Stranded Total RNA Library Prep Kit). Sample libraries were sequenced using the Illumina HiSeq 2500. Sequence cluster identification, quality prefiltering, base calling, and uncertainty assessment were done in real time using Illumina's HCS 2.2.58 and RTA 1.18.64 software with default parameter settings. Between 7 and 12 million (1X50) base pair reads were generated per library and mapped to the Drosophila genome (Release 6). Expression was recorded as reads per kbp per million reads (RPKM). GO clustering analysis was performed using FlyMine. FASTQ data files representing unique libraries are deposited in the NCBI Gene Expression Omnibus database (GSE111248).

Feeding behavior assay
The CAFE assay was performed as follows: Briefly, a single fly was transferred from standard food to vials filled with 5 mL of 1.5% agar that maintains internal humidity and serves as a water source. Flies were fed with 5% sucrose solution maintained in 5-μl capillaries (VWR, #53432-706). After 12 h of habituation, the old capillaries were replaced with a new one at the start of the assay. The amount of liquid food consumed was recorded after 24 h and corrected on the basis of the evaporation (typically <10% of ingested volumes) observed the identical vials without flies. Five flies were weighed in order to normalize samples.

Climbing assay
Twenty flies (at indicated ages) were placed into the empty vials, taped to the bottom, and then given 30 s to climb a distance of 6 cm. Flies that successfully climbed 6 cm or beyond in 30 s were counted. At least 100 total flies (5 cohorts) were used for each genotype tested.

Commensal quantification
To culture commensal bacteria, dissected midguts were homogenized in 100 μl PBS using pestles, and different dilutions of this suspension were plated onto mannitol agar plates. Plates were incubated at 30˚C for 36-48 h, and the number of colonies per plate was counted.

Cell culture conditions and UV treatment
Drosophila S2 cells (obtained from Drosophila Genomics Resource Center) were maintained in Schneider's Drosophila media supplemented with 10% FCS, 50 U/ml penicillin, and 50 μg/ml streptomycin at 25˚C. Primary MEFs were cultured in DMEM supplemented with 10% FCS, 50 U/ml penicillin, and 50 μg/ml streptomycin at 37˚C in 5% CO2. To induce apoptosis, the cell media was removed, and cells were treated with ultraviolet C light (UVC; dose as indicated).
To obtain dsRNA, sense and antisense strands were annealed by heating to 65˚C for 15 min and then allowed to cool to room temperature. The quality of the dsRNA was analyzed on an agarose gel and stored at −80˚C. For knock-down experiments, S2 cells were seeded in 6-well plates in 2 ml of medium the day before the RNAi procedure. The following day, medium was removed, and dsRNA (10 μg) was directly added to 1 ml of serum-free media and mixed vigorously. Cells were incubated for 1 h, followed by the addition of 2 ml of media supplemented with 10% FBS. The S2 cells were incubated for 2 d to induce posttranscriptional gene silencing. , and beta-Actin (Cell signaling, #4970). The immobilized proteins were further incubated with the corresponding HRP-conjugated secondary antibody, and protein bands were visualized using a commercial chemiluminescence detection kit (ECL,Amersham).

Viral infection assay
VSV-GFP [87] was propagated and assayed on BHK cells. In brief, confluent BHK cells were infected with VSV and incubated with virus until complete infection occurred (1 d). Cells were centrifuged at 2,000 rpm for 5 min, and supernatant was collected. Virus was concentrated by centrifuging at 25,000 g for 1.5 h at 4˚C. The pellet was resuspended in sterile PBS and stored at 80˚C. A standard plaque assay on BHK cells was performed on all purified virus stocks before use. S2 cell infections were performed in serum-free Schneider's media for 1 h using a multiplicity of infection (MOI) of 20 (as previously described, [55]); then, conditioned media containing serum was added for the duration of the experiment. Sixteen hours after infection, cells were processed for either RNA collection (and subsequent qRT-PCR, see S4 Table) or immunostaining (TUNEL assay).

In vitro TUNEL (cell death) assay
S2 cells or MEFS were fixed on a poly-L-lysine (VWR)-coated glass coverslip with 10% formaldehyde for 10 min and washed 2 times with PBS. Cell death was detected by using an In Situ Cell Death Detection Kit (Roche) according to manufacturer's protocol. Following washes in PBS, cells were stained with Alexa Fluor FITC phalloidin (Life Technologies) in PBST with Hoechst for 30 min at room temperature.

In vitro DNA fragmentation analysis
DNA ladder fragmentation was detected as described [88], with some modifications. Briefly, S2 cells (2 × 106) were seeded into 6-well plates. The next day, the cell media was removed, and cells were exposed to UV using Stratlinker1800. The cells were treated for 5 h with conditioned media with or without Die V5 . At the end of incubation, the DNA was extracted from S2 cells with 2 phenol/chloroform extractions and 1 chloroform extraction and finally precipitated with NaCL (300 Mm) and ethanol. The DNA was incubated for 1 h at 37˚C with RNAse A. The fragmented DNA was electrophoresed on 2% agarose gel.