Promoting longevity by maintaining metabolic and proliferative homeostasis

Aging is associated with a widespread loss of function at all levels of biological organization: molecular, cellular, tissue and organismic (Greer and Brunet, 2008; Murphy and Partridge, 2008; Schumacher et al., 2008; Broughton and Partridge, 2009; Campisi and Sedivy, 2009; Hayes, 2010; Richardson, 2009; Schumacher, 2009; Kapahi, 2010; Katewa and Kapahi, 2010; Liu and Rando, 2011; Rando and Chang, 2012). The proximal cause of this loss is the progressive decline of processes that maintain biological order and complexity, and thus ensure homeostasis in the young. Studies using model organisms have generated significant insight into the genetic factors and environmental conditions that influence, cause and prevent this age-related decline (Guarente and Kenyon, 2000; Tissenbaum and Guarente, 2002; Kenyon, 2005; Greer and Brunet, 2008; Katewa and Kapahi, 2010). It is clear that a combination of stress-induced damage to biological macromolecules, side effects of energy metabolism, deregulation of processes ensuring cellular and tissue integrity, and loss of coordination of systemic physiological control mechanisms are involved in the age-related loss of homeostasis. At the same time, a number of conserved genetic mechanisms preventing or counteracting this decline have evolved. Understanding the integration of these diverse processes, and thus gaining comprehensive insight into the causes and mechanisms of the aging process, remains a challenge. In recent years, however, the study of aging has progressed from the identification of single-gene mutations that influence lifespan toward more complex integrative analysis of the aging process. This work has also been facilitated by genetically accessible model systems and includes detailed analysis of the interaction between environmental factors and genetics in the control of longevity, as well as assessment of tissue-specific contributions to lifespan regulation (Guarente and Kenyon, 2000; Tissenbaum and Guarente, 2002; Kenyon, 2005; Greer and Brunet, 2008; Mair and Dillin, 2008; Broughton and Partridge, 2009; Panowski and Dillin, 2009; Toivonen and Partridge, 2009; Taylor and Dillin, 2011; Conboy and Rando, 2012). An important emerging insight is the complex role that various regulatory processes play during aging: signaling interactions required for the maintenance of cellular, tissue or organismal integrity in the young can cause dysfunction in the aging organism (Kahn and Olsen, 2010; Karpac and Jasper, 2009; Toivonen and Partridge, 2009; Kapahi, 2010; Biteau et al., 2011b). This ‘antagonistic pleiotropy’ of biological regulation is an important cause for many age-related metabolic and proliferative diseases, and its causes and immediate consequences are only beginning to be understood.

A particularly successful model in the characterization and elucidation of these regulatory interactions is Drosophila melanogaster, which has been used to characterize signaling mechanisms that control cellular damage responses and metabolic activity; coordinate tissue responses to damage, infection and inflammation; regulate tissue regeneration; and promote longevity in a variety of environmental conditions (Tatar et al., 2003; Lemaitre and Hoffmann, 2007; Leopold and Perrimon, 2007; Broughton and Partridge, 2009; Karpac and Jasper, 2009; Kapahi et al., 2010; Biteau et al., 2011a; Biteau et al., 2011b; Jones and Rando, 2011). Classic studies have established genetic interventions that can promote longevity, including a reduction in insulin signaling activity, increased expression of antioxidant proteins and protein chaperones, and activation of stress response signaling pathways (Parkes et al., 1998; Clancy et al., 2001; Tatar et al., 2001; Wang et al., 2003; Sun et al., 2004; Tower, 2011). These studies have been complemented in recent years with work addressing the complex interactions between tissue damage, metabolic regulation, innate immune processes and regenerative decline (Leopold and Perrimon, 2007; Biteau et al., 2011a). A comprehensive overview of causes and consequences of aging is emerging from this work, and new insight into possible therapeutic interventions against age-related diseases, is being generated.

Here, we review recent findings in the D. melanogaster model that are the basis for emerging concepts explaining the breakdown of homeostasis in the aging organism. Two specific areas will be discussed: signaling mechanisms regulating metabolic responses to tissue damage, and the control of proliferative homeostasis in the aging intestinal epithelium. The unifying concept emerging from these areas of research is the importance of signaling interactions between stress signaling pathways, innate immune signaling and growth signaling through the insulin and Tor signaling pathways in the maintenance of homeostasis in the young, and the deregulation of these interactions in the old organism. Because the signaling interactions discussed here are evolutionarily conserved, these paradigms provide important insight into the development of age-related metabolic and proliferative diseases, and may lead the way towards development of integrative therapies against such diseases.

The insulin signaling pathway is a central regulator of longevity in both invertebrates and vertebrates (Wolkow et al., 2000; Tatar et al., 2003; Kenyon, 2005; Greer and Brunet, 2008; Broughton and Partridge, 2009; Kenyon, 2010; Barzilai et al., 2012). Single-gene loss-of-function mutations in insulin signaling components extend lifespan in Caenorhabditis elegans and D. melanogaster, and loss of insulin signaling activity in specific tissues of mice can also promote longevity. Insight into the role of insulin signaling in regulating lifespan was first obtained from genetic screens in worms, in which mutations in the PI3K homologue AGE-1 were identified that extend lifespan by more than 100% (Friedman and Johnson, 1988). Further genetic studies identified loss-of-function mutations in the insulin receptor daf-2 and in the Akt and SGK kinases as long-lived (Kenyon et al., 1993; Kimura et al., 1997; Hertweck et al., 2004). Similar genetic studies in flies identified long-lived strains mutant in the insulin receptor and the insulin receptor substrate Chico (Clancy et al., 2001; Tatar et al., 2001). In mice, heterozygosity for the insulin-like growth factor (IGF) 1 receptor increases lifespan, while knockdown of the insulin receptor in adipose tissue or knockdown of IRS2 in the brain is sufficient to increase lifespan (Blüher et al., 2003; Holzenberger et al., 2003; Taguchi et al., 2007).

These longevity effects of reducing insulin signaling are caused by activation of the transcription factor Foxo (Forkhead box protein O), which is negatively regulated downstream of Akt by insulin signaling, and is essential for the lifespan extension observed in insulin loss-of-function conditions in both worms and flies (Kenyon et al., 1993; Giannakou et al., 2004; Hwangbo et al., 2004; Wang et al., 2005; Greer and Brunet, 2008). Foxo translocates to the nucleus and activates transcription when insulin signaling activity is low, inducing the expression of a number of stress defense and repair proteins in addition to metabolic enzymes and regulators (Murphy et al., 2003; Greer and Brunet, 2008; Biteau et al., 2011b). This activity of Foxo promotes cellular stress resistance and damage repair, and can regulate apoptosis (Brunet et al., 1999; Kops et al., 2002; Brunet et al., 2004; Wang et al., 2005; Luo et al., 2007; Demontis and Perrimon, 2010).

Through Foxo-induced gene expression, the lifespan extension observed by reduced IIS activity is thus linked to metabolic regulation, but also to Foxo-mediated cellular stress response and repair functions.

Metabolic regulation by Foxo has been described in several organisms, and is likely tied to its role in longevity. Adaptation to dietary conditions is required to maintain metabolic and energy homeostasis (Roberts and Rosenberg, 2006; Luca et al., 2010), and deregulation of adaptive responses because of modern (high sugar/high fat) diets or age-related dysfunction has been proposed as a cause for metabolic diseases (Odegaard and Chawla, 2013). This notion is supported by the increased incidence of metabolic diseases in the elderly (Roberts and Rosenberg, 2006).

The role of insulin signaling in metabolic processes is well understood, but many questions about the link between metabolic regulation by insulin and lifespan remain. Understanding the complex network of tissue-specific functions of insulin/IGF signaling (IIS) in the control of metabolic homeostasis is likely to be required to fully grasp the causal relationship between the regulation of metabolic adaptation and longevity. IIS governs energy homeostasis systemically by mediating tissue interactions that control lipid and carbohydrate metabolism, stabilizing blood glucose levels and serving anabolic purposes (Saltiel and Kahn, 2001). The health consequences of reducing insulin signaling activity, however, are complex and tissue-specific. While loss of the insulin receptor in adipose tissue of mice can have positive effects on lifespan (Blüher et al., 2003), IIS inhibition and Foxo activation in other tissues (such as liver, brain and muscle) result in metabolic and degenerative diseases (Kido et al., 2000; Michael et al., 2000; Suzuki et al., 2010; Mihaylova et al., 2011). These phenotypes are reminiscent of many age-related metabolic pathologies in humans that are associated with chronic insulin resistance (Barzilai et al., 2012). In these conditions, Foxo activation contributes to the breakdown of lipid and glucose homeostasis, promoting diabetes (Samuel and Shulman, 2012; Eijkelenboom and Burgering, 2013).

Many of the metabolic consequences of altered IIS and Foxo signaling are caused by changes in glucose and lipid metabolism. The function of insulin signaling in regulating glucose uptake and glycogen synthesis in the muscle and liver is well understood (Saltiel and Kahn, 2001; Accili and Arden, 2004; Samuel and Shulman, 2012; Eijkelenboom and Burgering, 2013). At the same time, altered IIS activity is also associated with abnormal lipid metabolism (Kitamura et al., 2003; Teleman, 2010), yet the cell- and tissue-specific mechanisms by which IIS regulates lipid metabolism are only partially understood. In both Drosophila and mammals, Foxo regulates the transcription of lipases in adipose tissue required for the lipolysis of stored lipids (Vihervaara and Puig, 2008; Chakrabarti and Kandror, 2009; Wang et al., 2011), as well as the expression of enzymes and other transcription factors involved in lipid catabolism (Deng et al., 2012; Xu et al., 2012). Accordingly, changes in IIS/Foxo regulated lipases (such as adipose triglyceride lipase) and lipogenic transcription factors (such as SREBP-1c) have been linked to the dyslipidemia associated with type 2 diabetes and other metabolic syndromes (Shimomura et al., 2000; Schoenborn et al., 2006; Badin et al., 2011).

In addition to genes that promote metabolic homeostasis, Foxo also regulates genes encoding stress-response and repair proteins. The activation of Foxo in conditions of low insulin signaling can thus shift energy resources from anabolic functions to cellular preservation and repair. This central role as a switch between anabolism and repair is reflected in the importance of the C. elegans Foxo homologue Daf16 in both dauer formation (where developing worms enter developmental and metabolic arrest under nutrient deprivation or stress) and longevity (Kenyon et al., 1993; Giannakou et al., 2004; Hwangbo et al., 2004; Wang et al., 2005; Greer and Brunet, 2008).

Highlighting the importance of Foxo in regulating cellular stress responses, its activity is not only controlled by insulin signaling, but can also be stimulated by stress-response signaling pathways. A well-understood interaction between IIS and stress signaling is mediated by the Jun-N-terminal kinase (JNK) pathway, which regulates gene expression in response to a wide range of environmental stressors. JNK can promote nuclear translocation of Foxo independently of insulin signaling, providing an avenue to induce Foxo-mediated gene expression programs independently of the metabolic status of the cell (Hirosumi et al., 2002; Karpac and Jasper, 2009; Biteau et al., 2011b; Samuel and Shulman, 2012).

JNK activity is induced by a range of intrinsic and environmental insults (e.g. UV irradiation, reactive oxygen species, DNA damage, heat and inflammatory cytokines) through a MAP kinase cascade that is initiated by a number of stress-specific upstream kinases (JNK kinase kinases). JNK commonly regulates transcription factors, such as Foxo and AP-1, leading to changes in gene expression. Its activation has highly context-dependent consequences for the cell, ranging from apoptosis over morphogenetic changes to increased survival (Davis, 2000; Weston and Davis, 2002; Manning and Davis, 2003; Karin and Gallagher, 2005). Mirroring the lifespan extension observed in insulin signaling mutants, moderate activation of JNK signaling results in increased stress tolerance and extended lifespan in flies and worms. Flies heterozygous for the JNK phosphatase Puckered (puc), or in which the JNK kinase Hemipterous (JNKK/Hep) is overexpressed in neuronal tissue, are resistant to the oxidative stress-inducing compound Paraquat (Wang et al., 2003). Similarly, JNKK/Hep mutant flies exhibit increased stress sensitivity and are deficient in the induction of a transcriptional stress response (Wang et al., 2003). Supporting the protective role for JNK signaling in flies, puc heterozygotes or Hep-overexpressing animals are long-lived under normal conditions (Wang et al., 2003; Libert et al., 2008). Similar consequences of JNK activation have been described in C. elegans. Worms with increased JNK activity due to inhibition of the JNK phosphatase VHP-1 (VH1 dual-specificity phosphatase) are protected against heavy metal toxicity, and overexpression of JNK increases lifespan under normal conditions (Mizuno et al., 2004; Oh et al., 2005).

Genetic interaction studies reveal that all these protective effects of JNK signaling are mediated by DAF-16/Foxo (Oh et al., 2005; Wang et al., 2005). In response to JNK signaling, Foxo promotes the expression of genes that inhibit growth and promote repair, but also of genes that induce apoptosis (Wang et al., 2005; Luo et al., 2007). Accordingly, activation of JNK dominantly suppresses IIS-induced cell growth, and Foxo acts downstream of JNK signaling to extend lifespan in flies and worms, but also to mediate apoptotic responses to excessive DNA damage (Matsumoto and Accili, 2005; Oh et al., 2005; Wang et al., 2005; Luo et al., 2007; Wolf et al., 2008). Activation of Foxo proteins by JNK has further been described in mammalian cells, suggesting that the interaction between IIS and JNK signaling is conserved (Essers et al., 2004; Jaeschke and Davis, 2007).

Interestingly, activation of Foxo by JNK signaling is achieved through multiple cell-autonomous and non-autonomous mechanisms and has been implicated not only in lifespan extension, but also in pathological disruptions of metabolic homeostasis. In fact, JNK-mediated repression of insulin signaling activity has been proposed as a major cause for insulin resistance in obese patients (Waeber et al., 2000; Hirosumi et al., 2002; Solinas et al., 2007; Yang et al., 2007; Emanuelli et al., 2008). JNK can suppress insulin signaling activity and activate Foxo by phosphorylating the insulin receptor substrate (thus preventing signal transduction through this molecule), by phosphorylating 14-3-3 proteins and disrupting their inhibition of Foxo, as well as by phosphorylating Foxo proteins directly (Aguirre et al., 2000; Mamay et al., 2003; Essers et al., 2004; Tsuruta et al., 2004; Sunayama et al., 2005; Yoshida et al., 2005). While C. elegans JNK was shown to phosphorylate DAF-16 (Oh et al., 2005), sequence alignment shows that the JNK phosphorylation sites in mammalian IRS or Foxo4 are not conserved in the Drosophila proteins, while the phosphorylation sites on 14-3-3 are conserved in both flies and C. elegans (Essers et al., 2004; Tsuruta et al., 2004; Sunayama et al., 2005; Yoshida et al., 2005; Nielsen et al., 2008; Karpac and Jasper, 2009). The antagonism between JNK and IIS thus appears to be evolutionarily ancient and critical for survival.

In addition to the cell-autonomous mechanisms of IIS inhibition described above, JNK can promote the expression of cytokines and adipokines that suppress insulin signal transduction at a distance (such as IL-6, which is secreted in response to JNK activity from the adipose tissue and macrophages and prevents insulin signaling in the liver) (Sabio et al., 2008; Sabio and Davis, 2010; Han et al., 2013). JNK thus mediates important cell-autonomous and systemic metabolic adaptations to changing environmental conditions. While chronic engagement of this signaling system can cause metabolic diseases, the acute JNK-mediated repression of insulin signaling activity is an adaptive process that decreases metabolic activity under stress conditions (Karpac and Jasper, 2009).

This adaptive process is exemplified by dauer formation in C. elegans, an extreme form of adaptation in which development and metabolic activity are curtailed under unfavorable conditions. Repression of insulin signaling activity is central to dauer formation and can be achieved by stress-induced activation of Foxo, as well as by stress-induced repression of insulin-like peptides. Accordingly, the regulation of insulin-like peptide expression in specific chemosensory neurons in the head (ASI neurons) is crucial for dauer formation. These neurons express the insulin-like peptide DAF-28, which is repressed in response to starvation and dauer pheromone, activating DAF-16 in the periphery (Bargmann and Horvitz, 1991; Li et al., 2003). This peripheral activation of DAF-16, in turn, is sufficient to increase lifespan (Libina et al., 2003). The intestine, the major fat-storing tissue in the worm, has an important role in this lifespan extension, but Foxo activity in other peripheral cells can also extend lifespan (Libina et al., 2003; Zhang et al., 2013).

In Drosophila, similar endocrine regulation of IIS activity by stress signaling has been described (Fig. 1). JNK activation in insulin-producing cells (IPCs) is crucial for metabolic and growth adaptation under stress conditions (Wang et al., 2005; Karpac et al., 2009). This is achieved by repression of Drosophila insulin-like peptide 2 (Dilp2) transcription, and involves Foxo activation in IPCs. As in worms, JNK signaling thus systemically represses IIS activity, allowing coordinated metabolic adaptation of insulin target tissues to environmental changes, and extending the lifespan of flies (Wang et al., 2005; Karpac et al., 2009). This mechanism is part of a complex system of endocrine interactions between insulin-producing tissues, fat-storing tissues and peripheral insulin target tissues that regulate lifespan and metabolic homeostasis (Fig. 1) (Leopold and Perrimon, 2007; Russell and Kahn, 2007; Karpac and Jasper, 2009; Rajan and Perrimon, 2012). Thus, ablation of IPCs is sufficient to extend lifespan in flies (Wessells et al., 2004; Broughton et al., 2005), and overexpression of Foxo in the peripheral fat body also leads to lifespan extension. Interestingly, Foxo activation in this tissue is accompanied by feedback repression of insulin-like peptide gene expression in IPCs (Giannakou et al., 2004; Hwangbo et al., 2004). A recent study has identified a fat-body-derived insulin-like peptide, Dilp6, as a candidate for this feedback signal (Okamoto et al., 2009; Slaidina et al., 2009; Bai et al., 2012). Dilp6 is a peptide that is secreted in non-feeding and fasting states from the fat body, and sustains insulin-dependent growth under such conditions during development (Okamoto et al., 2009; Slaidina et al., 2009). The induction of Dilp6 under starvation conditions in the adult, the repression of brain insulin-like peptides, and the lifespan extension observed when Dilp6 is overexpressed from this tissue highlight the complexity of the regulation of insulin signaling activity in peripheral tissues by a diversity of locally and systemically acting insulin-like peptides. Other examples include insulin-like peptides acting locally in the brain (Chell and Brand, 2010; Sousa-Nunes et al., 2011) and in the intestine (O'Brien et al., 2011), and systemically regulating growth in response to tissue overgrowth and tumor formation (Fig. 1) (Colombani et al., 2012; Garelli et al., 2012).

These findings support the notion that complex, evolutionarily conserved endocrine interactions between insulin-producing and insulin-target tissues influence metabolic homeostasis and lifespan. Stress signaling through JNK appears to be an integral part of this regulatory network. Recent studies have explored this regulation in more detail, and have established the spatiotemporal coordination of stress, innate immune and inflammatory signals that govern this regulatory process in flies (Karpac et al., 2011): local tissue damage (induced, for example, by UV irradiation) initiates a transient systemic repression of insulin signaling activity by inducing inflammatory cytokines (including Upd3, a functional analogue of IL-6) that activate the JAK/Stat signaling pathway in adipose tissue and hemocytes. This activation is required to repress insulin transcription in IPCs. Subsequent activation of Foxo in adipose tissue leads to the activation of NFkB/Relish signaling in these cells, which in turn is required to initiate a recovery phase of the adaptive response. This recovery restores insulin transcription in IPCs and is crucial to resume growth and anabolic functions after damage has been repaired (Karpac et al., 2011).

The central role of the fat body in coordinating these stress responses is reminiscent of the importance of adipocyte-derived signals in the maintenance of metabolic homeostasis in vertebrates. A recent study has identified Upd2, a cytokine with significant similarity to Upd3, as a factor that is secreted from adipose tissue in the fed state and activates JAK/Stat signaling in a population of GABAergic neurons in the brain that have direct connections to IPCs (Rajan and Perrimon, 2012). This interaction controls Dilp secretion in IPCs, and thus regulates systemic growth and anabolic activity. Strikingly, human leptin can functionally replace Upd2 in this system, highlighting the evolutionary conservation of mechanisms that maintain metabolic and growth homeostasis in metazoans. Furthermore, it has also been shown in flies that amino acids, sensed by the fat body, can relay humoral signals to control insulin release from the brain (Géminard et al., 2009). Slimfast (an amino acid transporter) is required upstream of the amino-acid-sensing and IIS-responsive TOR signaling pathway in the fat body to link dietary amino acid levels with circulating Dilps during metabolic adaptation. It is interesting to speculate that TOR signaling may couple nutrient sensing in the fat body to insulin levels in flies through the regulation of Upd2.

Taken together, these findings show that Drosophila is emerging as model for the study of organ–organ communication (Fig. 1). Studies of physiological cross-talk between fly tissues are revealing mechanistic insight into the regulation of IIS and the control of systemic growth, aging and energy homeostasis, as well as the regulation of tissue homeostasis and innate immunity (reviewed below). The similarity of fly organs to vertebrate organ systems, the evolutionarily conserved signaling pathways encoded in the Drosophila genome, and the highly developed genetic tool kit available for the fly make Drosophila a crucial model organism in which to study the signaling mechanisms and tissue interactions that regulate IIS.

An interesting example of the similarity between the described processes in flies and vertebrates is the repression of insulin/IGF signaling activity in response to DNA damage, which has also been observed in mice. Mutations in genes involved in DNA repair cause repression of IGF signaling, resulting in growth attenuation (Niedernhofer et al., 2006; van der Pluijm et al., 2007; Garinis et al., 2008; Schumacher et al., 2008). The repression of the growth hormone/IGF axis is believed to promote tissue repair in organs afflicted by increased DNA damage, but paradoxically, is also associated with progeroid (or accelerated aging) phenotypes. The complexity of tissue-specific responses to the attenuation of growth hormone/IGF signaling remains to be explored in detail and is likely to provide new insight into the acute as well as chronic mechanisms promoting metabolic adaptation and tissue dysfunction in response to DNA damage and injury.

The antagonistic interaction between stress signaling and insulin signaling is thus an evolutionarily conserved adaptive mechanism to promote cell survival and tissue repair when the organism experiences stressful and damaging conditions. Why does this mechanism then contribute to age-related metabolic diseases? It seems likely that the duration of JNK-mediated IIS repression is crucial for the ultimate outcome of this signaling event. While acute suppression of IIS activity by JNK is likely beneficial, the system can cause significant damage when the organism is stressed continuously, or when stress signaling is otherwise chronically activated (Fig. 2). Such conditions include inflammation that is associated with obesity and aging. In fact, JNK signaling has been identified as a central mediator of insulin resistance in genetic and dietary-induced models of obesity and insulin resistance (Waeber et al., 2000; Hirosumi et al., 2002; Solinas et al., 2007; Yang et al., 2007; Emanuelli et al., 2008). This effect involves complex interactions between JNK activity in the liver, muscle, adipocytes and macrophages (Waeber et al., 2000; Hirosumi et al., 2002; Solinas et al., 2007; Yang et al., 2007; Emanuelli et al., 2008; Han et al., 2013). In obese models, JNK is activated by endoplasmic reticulum stress in adipocytes, as well as by adipose-derived inflammatory signals such as free fatty acids and TNF-α and the concurrent activation of Toll-like receptors (Tsukumo et al., 2007), and results in phosphorylation and inhibition of IRS-1 (Hotamisligil et al., 1996; Yin et al., 1998; Hotamisligil, 1999; Aguirre et al., 2000; Hirosumi et al., 2002; Ozcan et al., 2004; Hotamisligil, 2005; Hotamisligil, 2006; Bouzakri and Zierath, 2007). Furthermore, JNK activation in adipose tissue has been shown to induce inflammatory cytokines such as IL-6, which in turn promotes insulin resistance in the liver by activating JAK/Stat signaling (Sabio et al., 2008; Sabio and Davis, 2010). New therapeutic approaches that specifically target these signaling interactions are expected to be required to restore metabolic and growth homeostasis in these conditions.

The interaction between stress and insulin signaling described above not only influences lifespan by promoting metabolic homeostasis, but also has a major impact on proliferative homeostasis in regenerating tissues. Regeneration is crucial for renewal and maintenance of many tissues in vertebrates, and aging, as well as deregulation of insulin/IGF signaling, has important deleterious consequences for regenerative capacity (Jasper and Jones, 2010; Jang et al., 2011; Jones and Rando, 2011). Of particular interest are processes that maintain proliferative homeostasis in high-turnover tissues, as deregulation of such processes is likely to promote cancer or degenerative diseases. In fact, the universal impact of the interaction between stress and metabolic signaling on tissue health is highlighted by recent insights into the development and progression of age-related tissue dysfunction in the intestinal epithelium of Drosophila. This tissue has emerged as a very productive and accessible model system in which to explore the cell-autonomous and systemic regulation of stem cell function, tissue regeneration and the loss of proliferative homeostasis. In the following, we will discuss the implications of this research for our understanding of human diseases and the establishment of therapeutic approaches for age-related proliferative and degenerative diseases.

Drosophila has only recently been recognized as a model in which to study proliferative homeostasis. While flies were long considered post-mitotic animals, two elegant studies in 2006 demonstrated that the intestinal epithelium regenerates in the adult animal (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006). Using lineage tracing techniques, Ohlstein, Micchelli and colleagues identified and characterized a population of pluripotent tissue stem cells in the fly midgut that continuously regenerate this tissue in a process that mechanistically and morphologically resembles cell turnover in the intestinal epithelium of mammals (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006; Radtke et al., 2006; Ohlstein and Spradling, 2007) (Fig. 3). This work thus introduced the fly intestine as an ideal model system in which to characterize the regulation of intestinal stem cell function as well as age-related changes in regenerative capacity of high-turnover tissues.

Interestingly, this tissue also performs metabolic and innate immune functions, and serves as a crucial barrier epithelium, allowing detailed analysis of complex interactions between tissue regeneration, metabolic regulation, innate immunity and inflammation (Biteau et al., 2008; Choi et al., 2008; Buchon et al., 2009a; Biteau et al., 2010; Biteau et al., 2011a; Rera et al., 2012; Sieber and Thummel, 2012; Rera et al., 2013). We have recently shown that the regenerative capacity of stem cells strongly influences lifespan in Drosophila, and that improving intestinal tissue homeostasis also rescues the age-dependent systemic breakdown of lipid homeostasis (Biteau et al., 2010).

Intestinal stem cells (ISCs) are the only dividing cells in the intestinal epithelium, and can give rise to at least two differentiated intestinal cell types: enteroendocrine cells (EEs) and enterocytes (ECs) (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006; Ohlstein and Spradling, 2007) (Fig. 3). When ISC division is triggered, these cells divide asymmetrically, generating a new stem cell that remains basally, while a newly formed progenitor cell, the enteroblast (EB), emerges apically and differentiates into either ECs or EEs (Ohlstein and Spradling, 2007). The asymmetry of the ISC/EB pair is likely to be determined by integrin-mediated adhesion of ISCs to the basement membrane and the underlying visceral muscle (Goulas et al., 2012). The muscle plays an important role not only in maintaining asymmetry in this cell pair, but also in supplying growth factors that promote ISC maintenance and proliferative competence, and has thus been described as the ‘niche’ of the ISC population (Lin et al., 2008; Biteau et al., 2011a; Biteau and Jasper, 2011).

ISC self-renewal and differentiation is controlled by the Notch and Tor signaling pathways. ISCs express the Notch ligand Delta, which in turn activates Notch signaling in the EB, driving their differentiation. The lineage decision between EE and EC fates of the EB is determined by the level of Dl expression in the ISC, and accordingly by the level of Notch signaling activity in the EB, with high Notch signaling promoting EC fates, and low Notch activity promoting EE fates. Differentiation of ECs requires activation of Tor signaling, which is differentially activated in ISCs and EBs. High expression of tuberous sclerosis complex 2 (TSC2), the Rheb GTPase activating protein and inhibitor of the Tor pathway, is required in ISCs to prevent their differentiation, while Notch-mediated suppression of TSC2 expression in EBs is required for differentiation into ECs. Interestingly, TSC2 buffers ISCs from dietary effects, promoting long-term maintenance of ISCs (Amcheslavsky et al., 2011; Kapuria et al., 2012).

Dietary conditions and metabolic signals not only influence differentiation of ISCs through the Tor signaling pathway, but also determine proliferative activity of ISCs through the insulin signaling pathway, thus influencing overall tissue growth. Interestingly, the insulin receptor has both cell-autonomous and non-autonomous roles in the regulation of ISC function. ISCs mutant for the insulin receptor do not divide, while increased insulin signaling activity can induce proliferation in widely quiescent ISCs (Amcheslavsky et al., 2009; Biteau et al., 2010; McLeod et al., 2010; Choi et al., 2011; Kapuria et al., 2012). Loss of the insulin receptor in EBs, however, impairs EB differentiation, most likely because of reduced Tor signaling in these cells. Unresolved adhesion of undifferentiated EBs to ISCs, in turn, inhibits ISC proliferation (Choi et al., 2011).

The activity of the insulin signaling pathway in ISCs is determined by both local and systemic Dilps. When insulin-producing cells in the brain are ablated, ISC proliferation is significantly reduced, highlighting the importance of systemic insulin-like signals (Amcheslavsky et al., 2009; Biteau et al., 2010; Kapuria et al., 2012). Dilp3 secreted from the visceral muscle, in contrast, influences ISC proliferation in the young maturing gut and in response to fasting/refeeding cycles, regulating the overall size of the intestinal epithelium (O'Brien et al., 2011). Strikingly, this response can induce ISCs to divide symmetrically, thus managing the overall numbers of available stem cells.

Dietary conditions, and the associated nutrient-responsive signals, thus influence ISC activity and maintenance, and can also significantly influence age-related changes in proliferative homeostasis of the intestinal epithelium. This regulation is integrated into a complex network of regulatory processes that determine compensatory proliferation of ISCs in response to cellular damage, stress and infection. These signaling interactions are complex, yet their impact on tissue homeostasis, health and lifespan suggest that further insight into these processes is crucial to understanding age-related inflammatory and proliferative diseases in vertebrates (Biteau et al., 2011a).

ISC proliferation rates are regulated by JNK signaling, but also by the Jak/Stat signaling pathway and by p38 MAPK, EGFR and insulin signaling (Biteau et al., 2008; Amcheslavsky et al., 2009; Apidianakis et al., 2009; Buchon et al., 2009a; Buchon et al., 2009b; Chatterjee and Ip, 2009; Jiang et al., 2009; Biteau et al., 2010; Buchon et al., 2010; Biteau and Jasper, 2011; Jiang et al., 2011). These pathways regulate ISC proliferation through a number of local and paracrine interactions involving both ECs and the muscle. Stress, infection, injury or tissue damage leads to JNK activation in ECs, which in turn induces production of the Jak/Stat ligands Upd1–3. These ligands are also induced by activation of Yorkie, which occurs in response to a loss of tissue integrity and repression of the Hippo pathway. Upds activate the Jak/Stat pathway in ISCs, triggering proliferation. At the same time, cell-autonomous activation of JNK signaling in ISCs can promote proliferation of these cells in response to oxidative stress. All these stress-induced pathways determine the activity of the ISC pool by changing the ratio of quiescent to active stem cells. Signaling through receptor tyrosine kinase pathways (including InR and EGFR signaling), however, is crucial to maintaining the proliferative potential of ISCs. The Fos transcription factor integrates JNK signaling with EGFR signaling (which is activated in ISCs by ligands derived from both the muscle and ECs), and thus serves as a crucial node in a signaling network that governs ISC activity (Buchon et al., 2010; Biteau and Jasper, 2011).

As regeneration through the ISC lineage is crucial for normal homeostatic tissue turnover as well as for epithelial recovery after damage or infection (Buchon et al., 2009a; Buchon et al., 2009b; Jiang et al., 2009), the activity of ISCs strongly influences stress and infection tolerance, as well as lifespan of flies. Interestingly, ISCs become deregulated in aging flies, overproliferating and producing a large number of cells that undergo incomplete differentiation (Fig. 3). This triggers intestinal dysplasia, which is characterized by loss of apico-basal organization of the epithelium and accumulation of mis-differentiated cells at the basement membrane (Biteau et al., 2008; Choi et al., 2008). The gut then loses both its resorptive and its barrier functions, resulting in increased mortality (Buchon et al., 2009a; Buchon et al., 2009b; Rera et al., 2011; Rera et al., 2012). This phenotype is caused by an age-related increase in global oxidative stress in the intestine that is associated with the presence of commensal bacteria (Buchon et al., 2009a) and results in the widespread activation of stress-responsive signaling pathways (Biteau et al., 2008). This includes p38 MAPK, JNK and the Jak/Stat signaling pathways, which are influenced by the commensal flora and are activated in the gut after exposure to pathogenic bacteria (Apidianakis et al., 2009; Buchon et al., 2009a; Buchon et al., 2009b; Chatterjee and Ip, 2009; Jiang et al., 2009). In the absence of infection, basal activation of JNK and Jak/Stat activities in ISCs by commensal bacteria is sufficient to maintain low levels of epithelial renewal in young flies (Buchon et al., 2009a). In older flies, however, the commensal population seems to cause excessive activation of JNK and JAK/Stat signaling, inducing the intestinal dysplasia described above. Accordingly, midguts from 30-day-old axenically raised flies exhibit a marked decrease in mitotic ISCs, as well as reduced levels of JNK and Jak/Stat signaling pathways compared with control animals (Buchon et al., 2009a).

It is likely that similar interactions between signals that regulate stem cell activity and inflammatory processes cause age-related deregulation of tissue homeostasis in barrier epithelia of vertebrates. Loss of tissue homeostasis and cancer progression is associated with an inflammatory state (Mantovani, 2005; Colotta et al., 2009; Mantovani, 2009), and individuals genetically predisposed to chronic inflammation display a higher incidence of various cancers, while tissue inflammation is present in most tumor microenvironments (Mantovani, 2005; Grivennikov et al., 2010).

This association between inflammation and cancer is particularly significant in barrier epithelia such as the intestinal epithelium, which mount frequent and required immune responses to pathogenic microorganisms (Xavier and Podolsky, 2007; Garrett et al., 2010). Accordingly, patients with inflammatory bowel diseases (including ulcerative colitis or Crohn's disease) have an increased risk of developing colon cancer (Gonda et al., 2009). Interestingly, studies to determine the genetic basis of colon cancer indicate that changes in microflora composition and load might be a prominent etiological factor for chronic inflammation and cancer development (Gonda et al., 2009; Uronis et al., 2009; Kaser et al., 2010; Niwa et al., 2010; Qin et al., 2010). Germ-free mice, accordingly, exhibit reduced intestinal inflammation and are refractory to carcinogen-induced colon tumor formation (Uronis et al., 2009).

Better mechanistic understanding of (1) the interactions between the microflora, (2) the innate immune responses and (3) the regenerative processes in the intestinal epithelium are thus key to the development of models of disease progression and are likely to significantly impact our understanding of the development of age-related tissue dysplasias. Intestinal dysplasia significantly limits Drosophila lifespan, and reducing the rate of stem cell proliferation in the aging intestine is sufficient to extend lifespan of the organism (Biteau et al., 2010; Rera et al., 2011). Similar management of stem cell proliferation in mammalian barrier epithelia is therefore likely to be beneficial for health and lifespan. Fly studies are leading the way to identify strategies that will promote stem cell quiescence, reduce dysplasia and improve health of the tissue. Reducing IIS and Tor activity, limiting JNK activation, promoting activity of the Nrf2 transcription factor, promoting mitochondrial function, as well as overexpressing general antioxidant genes and chaperones are all strategies that will limit ISC proliferation, delay or prevent dysplasia, and increase lifespan in this model (Biteau et al., 2010; Rera et al., 2011).

The integration of studies characterizing the tissue-specific regulation of metabolic and proliferative homeostasis with studies focusing on systemic control of metabolic homeostasis, fitness and longevity in flies promises to significantly advance our understanding of aging and the development of age-related diseases. It is clear that stress- and nutrient-responsive signaling pathways (including IIS/FOXO, TOR and JNK signaling) play a central role in governing homeostasis and longevity by coordinating metabolic functions, cell growth and proliferation, and stress responses throughout the organism (Figs 1, 2). However, many questions remain: how are these signaling pathways integrated under specific environmental (nutritional and stress) conditions; what are the major determinants that maintain the balance between acute stress responses, tissue repair and metabolic adaptation, and why does this balance degrade with age? It can be expected that further systematic analysis of the role of signaling interactions in the control of stress, regenerative and metabolic responses in the fly will continue to provide new insight and elucidate potential therapeutic avenues against age-associated pathologies.