|Home | About | Journals | Submit | Contact Us | Français|
Loss of metabolic homeostasis is a hallmark of aging, and is commonly characterized by the deregulation of adaptive signaling interactions that coordinate energy metabolism with dietary changes. The mechanisms driving age-related changes in these adaptive responses remain unclear. Here we characterize the deregulation of an adaptive metabolic response and the development of metabolic dysfunction in the aging intestine of Drosophila. We find that activation of the insulin-responsive transcription factor Foxo in intestinal enterocytes is required to inhibit the expression of evolutionarily conserved lipases as part of a metabolic response to dietary changes. This adaptive mechanism becomes chronically activated in the aging intestine, mediated by changes in Jun-N-terminal Kinase (JNK) signaling. Age-related chronic JNK/Foxo activation in enterocytes is deleterious, leading to sustained repression of intestinal lipase expression and the disruption of lipid homeostasis. Changes in the regulation of Foxo-mediated adaptive responses thus contribute to the age-associated breakdown of metabolic homeostasis.
In complex organisms, adaptation to changing dietary conditions is critical to maintain metabolic homeostasis (Luca et al., 2010; Roberts and Rosenberg, 2006). Adaptation has to occur at multiple levels, and involves adjusting processes that control nutrient uptake, storage and usage to dietary changes in order to properly maintain energy homeostasis (Roberts and Rosenberg, 2006). It has been proposed that modern (high sugar / high fat) diets can lead to misregulation of evolutionarily conserved adaptive dietary responses, resulting in metabolic dysfunction (Odegaard and Chawla, 2013). Furthermore, the age-associated decline of the ability to metabolically adapt to dietary changes is a likely cause for the increased incidence of metabolic diseases in the elderly (Roberts and Rosenberg, 2006). Our understanding of the mechanisms causing diet-induced or age-related changes in metabolic adaptive responses remains limited, and studies in model organisms are likely to provide critical insight into such mechanisms and into potential strategies for therapeutic or preventive interventions.
Insulin/IGF signaling (IIS) is one of the best understood signaling systems involved in metabolic adaptation and the control of metabolic homeostasis. In multi-cellular organisms, IIS governs energy homeostasis by regulating lipid and carbohydrate metabolism (Saltiel and Kahn, 2001), promotes cell growth and proliferation (Accili and Arden, 2004; Oldham and Hafen, 2003), and influences systemic stress responses (Baumeister et al., 2006; Karpac et al., 2009; Karpac et al., 2011; Niedernhofer et al., 2006; van der Pluijm et al., 2007). Its central role in these processes is highlighted by the fact that attenuating IIS function in invertebrates and vertebrates can extend lifespan (Barzilai et al., 2012; Karpac and Jasper, 2009; Taguchi and White, 2008; Tatar et al., 2003). The transcription factor Foxo, activated in response to decreased IIS activity, is an essential mediator of this lifespan extension (Lin et al., 1997; Ogg et al., 1997; Slack et al., 2011). Foxo regulates a battery of stress response and metabolic control genes (Calnan and Brunet, 2008; Matsumoto and Accili, 2005), and its activation has selective tissue-specific consequences. Thus, activating Foxo in the intestine of worms or in the adipose tissue and muscle of flies can extend lifespan (Demontis and Perrimon, 2010; Giannakou et al., 2004; Hwangbo et al., 2004; Libina et al., 2003; Zhang et al., 2013), while activation in other tissues and cell-types can reduce lifespan or promote cell death (Biteau et al., 2010; Luo et al., 2007).
In mammals, loss of the Insulin receptor in adipose tissue can have positive effects on lifespan (Bluher et al., 2003), while IIS inhibition and Foxo activation in other tissues (such as liver, brain, and muscle) contributes to metabolic and degenerative diseases (Kido et al., 2000; Michael et al., 2000; Mihaylova et al., 2011; Suzuki et al., 2010). Importantly, many age-related metabolic pathologies are associated with chronic insulin resistance in humans (Barzilai et al., 2012). In these conditions, Foxo contributes to the breakdown of lipid and glucose homeostasis, promoting diabetes (Eijkelenboom and Burgering, 2013; Samuel and Shulman, 2012).
Importantly, Foxo can be activated independently by stress signaling pathways, in particular Jun-N-terminal Kinase (JNK) signaling, which responds to a variety of inflammatory and stress signals and is a major contributor to insulin resistance, but can also increase lifespan in worms and flies (Biteau et al., 2011b; Hirosumi et al., 2002; Karpac and Jasper, 2009; Samuel and Shulman, 2012). The integration of nutritional responses through IIS with JNK-mediated stress signals by Foxo thus likely determines the beneficial or pathological consequences of attenuated insulin signaling.
Many of the metabolic consequences of altered IIS and Foxo signaling are caused by changes in glucose and lipid metabolism. The role of IIS signaling in regulating systemic glucose homeostasis and the function of Foxo in promoting glycogenolysis and gluconeogenesis in various tissues are well understood (Accili and Arden, 2004; Eijkelenboom and Burgering, 2013; Saltiel and Kahn, 2001; Samuel and Shulman, 2012). At the same time, altered IIS activity is also associated with abnormal lipid metabolism. Mice and flies in which IIS has been genetically manipulated have disrupted 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 (Chakrabarti and Kandror, 2009; Vihervaara and Puig, 2008; 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). Regulation of Foxo thus provides a mechanism by which insulin can regulate lipid metabolism during metabolic adaptation. Changes in IIS/Foxo regulated lipases (such as Adipose triglyceride lipase (ATGL)) and lipogenic transcription factors (such as SREBP-1c) have been linked to the dyslipidemia associated with type 2 diabetes and other metabolic syndromes (Badin et al., 2011; Schoenborn et al., 2006; Shimomura et al., 2000).
The Drosophila intestine is rapidly becoming a productive model system in which to study complex questions relating to the maintenance of tissue function, immune responses, metabolic homeostasis and stress signaling. Importantly, the intestine is key to the control of lipid and cholesterol homeostasis in the fly (Sieber and Thummel, 2012). Disruption of gut enterocyte function (cells required for nutrient uptake and innate immune function within the intestinal epithelium) can lead to organism-wide changes in lipid metabolism (Sieber and Thummel, 2009). Furthermore, as a critical barrier epithelium, the age-related dysfunction of this tissue significantly limits health- and lifespan of flies (Biteau et al., 2010; Rera et al., 2013; Rera et al., 2012). The adult intestine is regenerated by intestinal stem cells (ISCs), which divide to replace enterocytes (ECs) and enteroendocrine cells when needed (Biteau et al., 2011a; Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006). Previous studies have shown that ISCs hyper-proliferate in the aging gut, resulting in loss of tissue homeostasis, and the development of a general inflammatory condition characterized by excessive oxidative stress and chronic JNK activation (Biteau et al., 2008; Biteau et al., 2010; Buchon et al., 2009; Choi et al., 2008; Rera et al., 2012). 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).
Here, we identify a mechanism by which Foxo activity in the intestinal epithelium regulates metabolic adaptation to dietary changes in Drosophila. We show that Foxo activity in the intestine is required to inhibit the expression of evolutionarily conserved lipases as part of an adaptive metabolic response to changes in dietary protein/lipids. On a standard diet, Foxo-mediated repression of these lipases is required to prevent excessive lipid uptake, thus promoting systemic metabolic homeostasis. Strikingly, we find that this regulatory mechanism is constitutively activated in aging animals, contributing to metabolic imbalances and the age-associated breakdown of lipid homeostasis. This activation is caused by chronic activation of JNK, and reflects the development of a general inflammatory condition in the aging gut. Our findings define a distinct case in which an adaptive mechanism promoting metabolic homeostasis is deregulated by an age-associated increase in stress signaling, thus contributing to the breakdown of metabolic homeostasis.
To explore potential mechanisms within the gut that could mediate IIS-dependent adaptive responses to changes in nutrient availability and therefore promote metabolic homeostasis, we focused on the response of the Drosophila intestine to a dietary shift from a standard yeast/sugar based (St. SY) diet to an isocaloric high sugar/low yeast (HSLY) diet. The HSLY diet greatly reduces the amount dietary protein and lipids, while maintaining calories through excessive carbohydrates (Skorupa et al., 2008). HSLY diet feeding leads to the transcriptional down-regulation of insulin ligands dilp2 and dilp5 (Fig. 1A). These insulin-like peptides are secreted by neurosecretory Insulin Producing Cells (IPCs) in the brain, and have previously been described as responsive to nutrient imbalance (Garofalo, 2002; Geminard et al., 2009; Rulifson et al., 2002). This repression of dilp transcription is sufficient to systemically attenuate IIS activity in the gut, as HSLY diet feeding results in decreases in peripheral phospho-Akt (p-Akt) levels (Fig. 1B). Total Akt levels are also reduced in the intestine during HSLY diet feeding (Fig. 1B.). This suggests that other insulin-independent feedback mechanisms might also contribute to changes in Akt activity under these dietary conditions. The decrease in Akt/p-Akt levels was accompanied by the activation of Foxo target genes (evidenced by increased expression of lacZ from a thor (4E-BP) promoter (thor-lacZ, Fig. 1C) in the intestine, confirming that IIS-dependent signaling is suppressed in this tissue.
To assess whether the decrease in dietary lipids and proteins in the HSLY diet influences the expression of genes involved in metabolic adaptation to dietary changes, we measured the expression of lipA/magro, a gene that encodes an intestine-specific lipase with a described role in fat and cholesterol absorption in the gut (Sieber and Thummel, 2009, 2012). Magro (LipA in mammals) is expressed in distinct set of enterocytes in the adult fly intestine (Fig. S1 A), and is secreted into the intestinal lumen to digest dietary triacylglycerol (TAG) (Sieber and Thummel, 2012). Strikingly, the HSLY diet results in strong repression of lipA/magro in the intestine (Fig. 1A), while re-feeding flies a St. SY diet after 5 days on a HSLY diet could normalize their expression (Fig. 1A), suggesting that the transcriptional regulation of this lipase is part of an adaptive response to dietary changes.
To determine if changes in IIS/Foxo activity are responsible for this response, we further monitored the expression of lipA/magro in various Foxo loss-of-function genetic backgrounds. Reducing Foxo activity in intestinal enterocytes (using the NP1Gal4 driver, Fig. S1B and UAS-Foxo RNAi, Fig. S2A-C) prevented the down-regulation of lipA/magro expression during HSLY diet feeding (Fig 1D), suggesting that Foxo activity is required for the transcriptional repression of this lipase. This notion was confirmed in flies homozygous for the foxo null allele foxoW24, which also showed a significant increase in lipA/magro expression under fed conditions (Fig. 1E; additional allele (foxoΔ94) shown in Fig. S2D). However, it remains unclear how changes in Foxo activation and lipA/magro expression affect intestinal lipid metabolism during HSLY diet feeding.
The regulation of lipA/magro by Foxo suggests that Foxo-mediated responses to dietary changes are critical for metabolic adaptation and the control of lipid homeostasis. To test this hypothesis, we temporally manipulated Foxo activity in intestinal enterocytes using an adult-onset driver. To prevent developmental effects of the expression of UAS-driven transgenes, we used a heat-inducible system in which NP1Gal4 is combined with a temperature-sensitive Gal80 (TARGET system (Osterwalder et al., 2001)). Reducing Foxo activity (using UAS-Foxo RNAi) in enterocytes for 5 days leads to up-regulation of intestinal lipA/magro expression (Fig. 2A). This correlates with a significant increase (28%) in stored TAG levels in the whole animal (Fig. 2A and Fig. S3D). Elevated TAG levels are not due to changes in feeding behavior, as inhibiting Foxo function in enterocytes does not lead to significant changes in activity / rhythmicity or food intake (Fig. S3A-C). These data suggest that Foxo is required in enterocytes to limit intestinal lipase expression and maintain lipid homeostasis in the organism. Conversely, chronic over-expression of wild-type Foxo in intestinal enterocytes strongly represses lipA/magro transcription and disrupts intestinal lipid storage (Fig. 2B-C′). Importantly, intestinal over-expression of Foxo induces thor expression but does not lead to the disruption of epithelial morphology (Fig. S4A-B), showing that these effects are truly a response to Foxo activity in enterocytes and not due to global tissue degeneration. This down-regulation of intestinal lipase transcription correlates with a reduction in intestinal lipolytic activity (Fig. 2D) and a strong decrease (nearly 50%) in stored lipid levels in the whole organism (Fig. 2E and Fig. S3D). Furthermore, flies with chronic intestinal Foxo activation are extremely sensitive to starvation, confirming a breakdown of lipid homeostasis in these animals (Fig. 2F and Fig. S3E). Over-expression of Foxo in enterocytes does lead to a small, but significant, decrease in food intake (Fig. S3C). Further studies are needed to establish whether this change in feeding behavior is a cause or a consequence of the observed decrease in intestinal lipase expression and lipid storage. However, since loss of Foxo in this tissue did not influence feeding behavior, it is likely that the small change observed in Foxo gain of function conditions is a consequence, rather than a cause of tissue-specific metabolic changes elicited by Foxo.
Taken together, these data show that Foxo can inhibit the expression of intestinal lipases, and therefore control organismal lipid homeostasis through the regulation of intestinal lipid metabolism.
In a previous study, we found that an age-related decline in gut function is associated with changes in lipid metabolism in Drosophila (Biteau et al., 2010). In genome wide expression profiles generated from young and old guts of wild-type flies (using Serial Analysis of Gene Expression (SAGE) in combination with next-gen sequencing) we further found significant transcriptional down-regulation of trypsins (digestive serine proteases, see also (Biteau et al., 2008)) and of lipA/magro and CG6295. (confirmed by qRT-PCR, Fig. 3A-B; also see Table S1). CG6295 is another putative lipase that has over 30% amino acid homology to mammalian pancreatic TAG lipase related proteins, which are gastric lipases important for fat absorption in the intestine (Lowe, 2000). Importantly, the repression of these lipases is not dependent on the disruption of intestinal epithelium structure (i.e. age-related over-proliferation and mis-differentiation of ISC's leading to decreases in functional enterocytes, Fig. S5B), as aged flies in which ISC's have been ablated (preventing the disruption of epithelial structure) also show significant lipA/magro down-regulation in the intestine (Fig. S5A-B).
These results indicated that the disruption of systemic metabolic homeostasis observed in aging animals might be associated with a dysfunction of the adaptive response to dietary changes described above. Supporting this notion, we found that the age-dependent repression of lipA/magro and CG62952 expression correlates with a reduction in intestinal lipid storage, as evidenced by oil red O (neutral lipid) staining (Fig. 3C-C′). Furthermore, over-expression of LipA/Magro specifically in intestinal enterocytes can significantly rescue the age-related reduction of TAG levels in whole animals (Fig. 3D and Fig. S3C). These data support a role for the de-regulation of intestinal lipase expression in the age-related disruption of lipid homeostasis.
Since we found that LipA/Magro can be regulated by Foxo activity in the intestine, we tested whether aging modulates Foxo activity in the gut, and found that Foxo activity indeed significantly increases in the intestine as the animal ages. Foxo activation was measured using both the Foxo reporter thor-lacZ, as well as by measuring the expression of the Foxo target genes thor, InR (insulin receptor), and dlip4 (lipase 4) (Fig. 4A-B). Intestinal activation of Foxo is first observed in a significant number of flies at around 20 days of age, and progressively spreads through the population to reach around 85% at age 40 (Fig. 4C). These data do not rule out that thor-lacZ induction could be invariant with age, but individuals without induction die at younger ages and are eliminated from the population. However, we did not observe significant changes in mortality at 20 days of age (compared to 5 day old flies, data not shown), suggesting that the incidence of intestinal thor-lacZ induction within a population of flies does increase with age. Immunohistochemistry for β-Gal revealed that this age-related increase in Foxo activity occurs specifically in intestinal enterocytes, and not in intestinal stem cells (marked by esgGal4>UAS-GFP, Fig. 4A). Furthermore, reducing Foxo activity in enterocytes was sufficient to prevent the age-associated repression of lipA/magro and CG6295 (Fig. 5D-E). Elevated Foxo activity in the aging gut thus contributes to the inhibition of intestinal lipase expression.
To determine if this increase in Foxo activity and repression of lipase transcription correlates with attenuation of IIS, similar to the response to HSLY, we assayed p-Akt levels in young and old guts. Surprisingly, both p-Akt and total Akt levels are slightly increased in aging intestines (Fig. 4D). The age-related increase in total Akt protein is accompanied by increases in akt expression (Fig. 4E), an activation that is likely driven by Foxo, as akt has recently been described as a Foxo target gene (Alic et al., 2011). These data suggest that the IIS pathway is activated rather than repressed in older intestines, and that Foxo is activated by IIS independent mechanisms.
Since high steady-state p-Akt levels may not reflect changes in the sensitivity of the IIS pathway to extracellular ligands, we also confirmed that the IIS pathway remains sensitive to systemic insulin ligands in aging intestines. We performed ex vivo insulin stimulation tests on young and old guts by exposing dissected guts in culture media to mammalian insulin, and then measuring p-Akt levels. Insulin stimulation of both young and aged intestines revealed a strong up-regulation of phosho-Akt, and older intestines appear to be even more sensitive to this stimulation than young controls (Fig. 4D). This increased sensitivity is likely caused by the increased Akt and InR expression levels.
Taken together, our data reveal that Foxo becomes chronically activated in intestinal enterocytes with age, while the IIS pathway appears to be active and hyper-sensitive. We propose that this deregulation of IIS pathway control is at the center of the metabolic dysfunction observed in aging flies.
To identify alternative, IIS-independent, mechanisms responsible for the age-related increase in Foxo activity, we turned to JNK (Jun-N-terminal kinase), a stress-activated protein kinase that can activate Foxo in both vertebrates and invertebrates. JNK activity increases in older guts (Biteau et al., 2008), and we thus hypothesized that this elevated JNK activity may be responsible for the observed age-related increase in Foxo activity and the subsequent breakdown of intestinal lipid metabolism. Importantly, we confirmed that increased JNK activity in the aging intestine, like Foxo, occurs primarily in enterocytes (as determined using the JNK reporter puc-lacZ, Fig. 5A). Inhibiting JNK (using a dominant-negative form of the Drosophila JNK Basket, Bsk) in intestinal enterocytes significantly blocked the induction of thor-lacZ in aging guts (Fig. 5B-B′), suggesting that elevated JNK activity drives Foxo activation in the aging intestine. We further found that acute JNK activation in young guts is sufficient to repress intestinal lipase expression in a Foxo-dependent manner: Over-expression of a constitutively active JNKK (Hemipterous (Hep) in Drosophila) in intestinal enterocytes induces thor expression (Fig. S6), and also strongly represses lipA/magro and CG6295 expression (Fig. 5C). This lipase repression is reduced in a Foxo mutant background (Fig. 5D). Importantly, the transcriptional down-regulation of intestinal lipases is not caused by JNK-mediated apoptosis of enterocytes (Fig. S4, as the same suppression was observed in animals expressing the anti-apoptotic molecule p35.
Supporting a role for chronic JNK/Foxo activation in intestinal in the age-related disruption of lipid homeostasis, we found that reducing JNK activity (using UAS-BskDN) in enterocytes was sufficient to prevent the age-associated repression of lipA/magro and CG6295 (Fig. 6A). Importantly, inhibiting intestinal JNK activity also rescued the age-related reduction of lipid levels in the whole organism (Fig. 6B). These data support the notion that JNK activation in enterocytes is a major contributor to the overall breakdown of lipid homeostasis that occurs in the aging fly. To further test this concept, we assayed starvation sensitivity in the young and old flies described above. Aging flies are more sensitive to starvation, which correlates with age-related changes in glucose and lipid metabolism. Inhibiting JNK activation in enterocytes had no effect on starvation resistance in young animals, but led to significantly improved starvation resistance in older flies (as compared to wild-type controls, Fig. 6C-C′).
Together, these results suggest that JNK-mediated Foxo activation (independent of IIS) in the aging intestine promotes the de-regulation of intestinal lipase expression and subsequent disruption of intestinal and organismal lipid homeostasis in Drosophila.
Our work identifies Foxo-mediated repression of intestinal lipases as a critical component of an adaptive response to dietary changes in Drosophila. Interestingly, misregulation of this metabolic response also contributes to the age-associated breakdown of lipid homeostasis, as elevated JNK signaling leads to chronic Foxo activation and subsequent disruption of lipid metabolism due to chronic repression of lipases. This age-related deregulation of an adaptive metabolic response is reminiscent of insulin resistance-like phenotypes in vertebrates, which can also be triggered by chronic activation of JNK (Hirosumi et al., 2002; Samuel and Shulman, 2012), and thus highlights the antagonistic pleiotropy inherent in metabolic regulation. The adaptive nature of signaling interactions that drive pathology (such as JNK-mediated insulin resistance) has remained elusive in many instances, and our work provides a model for age-related changes in an adaptive regulatory process that ultimately lead to a pathological outcome. We believe that this system can serve as a productive model to address a number of interesting questions with relevance to the loss of metabolic homeostasis in aging organisms:
In mammals, JNK has been shown to promote insulin resistance both cell-autonomously and systemically (through inflammation), subsequently affecting lipid homeostasis in various tissues (Samuel and Shulman, 2012). Our results further introduce a mechanism by which JNK can alter cellular and systemic lipid metabolism through the regulation of lipases, independent of changes in IIS. Thus, JNK-mediated Foxo activation in select tissues may be able to alter intracellular lipid metabolism, changing metabolic fuel substrates and disrupting metabolic homeostasis in other tissues without altering insulin responsiveness. While our data show that Foxo activation leads to the transcriptional repression of intestinal lipases, especially LipA/Magro, it remains unclear if this control is direct or indirect. Foxo is classically described as an activator of transcription, but recent reports have shown that Foxo can transcriptionally repress genes through direct association with promoters (Alic et al., 2011; Deng et al., 2012). The promoter regions of LipA/Magro and CG6295 do not contain conserved Foxo transcription factor binding sites, suggesting that the regulation of these genes may be indirect, potentially through Foxo-regulated expression of secondary effectors. Thus, tissue specific control of lipid homeostasis by IIS/Foxo might be achieved through the regulation of lipogenic or lypolitic transcription factors that can elicit global and direct changes in cellular lipid metabolism. Previous reports have shown that the nuclear receptor dHR96, a critical regulator of lipid and cholesterol homeostasis, promotes lipA/magro expression (Sieber and Thummel, 2009). However, dhr96 expression is up-regulated in aging intestines (data not shown), suggesting that the age-related repression of intestinal lipases is not merely due to decreases in dHR96 levels. dhr96 transcript levels are strongly induced in genetic conditions where Foxo is activated and intestinal lipases are repressed (data not shown), again suggesting that Foxo does not mediate its effects on lipase transcription by antagonizing dhr96 expression. Furthermore, age-related changes that are independent of JNK/Foxo activation may also contribute to the repression of intestinal lipase expression and disruption of lipid metabolism, such as an age-associated decline in feeding / food intake (Wong et al., 2009).
The reasons for the increase in JNK and Foxo activity in aging ECs remain to be explored. Buchon et al. have also shown that age-related activation of JNK in the intestinal epithelium is dependent on the presence of commensal bacteria, as maintaining animals axenically reduces activation of JNK in the first 30 days of life (Buchon et al., 2009). Thus, bacteria induced inflammation and subsequent JNK activation appears to be a likely cause, in part, for age-related increases in Foxo activity. In a separate study, however, we find that Foxo activation still occurs in intestines of old (40 day old), axenically reared flies, suggesting that age-related activation of Foxo may also occur through JNK-independent processes (L.Guo, J.K., S.T., and H.J., unpublished data). Supporting this idea, our results show that inhibiting JNK function in enterocytes can attenuate, although not completely inhibit, this Foxo activation. Additional factors, such as sirtuins or histone deacetylases (HDACs), recently shown to deacetylate and activate Foxo in response to endocrine signals (Mihaylova et al., 2011; Wang et al., 2011), may also lead to age-related increases in intestinal Foxo activity.
Interactions between JNK and IIS/Foxo are mediated by various mechanisms. In mammals, JNK phosphorylates the insulin receptor substrate (IRS), inhibiting insulin signaling transduction (Aguirre et al., 2000; Mamay et al., 2003). JNK has also been shown to directly phosphorylate and activate Foxo in mammalian cell culture (Essers et al., 2004). While JNK has clearly been shown to antagonize IIS (activate Foxo) in C. elegans and Drosophila (Oh et al., 2005; Wang et al., 2005), that exact mechanism by which Foxo activation is achieved may be divergent in mammals. For example, no IRS homolog has been identified in worms, and the JNK phosphorylation site in mammalian IRS is not conserved in flies. Our data show that JNK-mediated Foxo activation in the aging fly intestine is not achieved through IIS antagonism upstream of Akt, suggesting either a direct interaction between Foxo and JNK or changes in other regulators of Foxo. Recent studies have shown that JNK-mediated phosphorylation of 14-3-3 proteins results in the release of their binding partners, including Foxo (Sunayama et al., 2005; Yoshida et al., 2005). The conservation of 14-3-3 proteins between vertebrates and invertebrates makes 14-3-3 an interesting candidate in promoting Foxo function via JNK in the aging fly intestine (Nielsen et al., 2008). This chronic intestinal Foxo activation, and subsequent metabolic changes, provides a physiological system in Drosophila to genetically dissect the cross-talk between IIS/Foxo and JNK signaling. Detailed analysis of these signaling interactions promises to provide important insight into the pleiotropic effects of IIS/Foxo function and the pathogenesis of age-related metabolic diseases.
Our data further reveal the pleiotropic consequences of Foxo activation in regard to healthspan and longevity in Drosophila. Over-expressing Foxo in the fatbody or muscle of flies leads to lifespan extension (Demontis and Perrimon, 2010; Giannakou et al., 2004; Hwangbo et al., 2004). Conversely, we have recently described that over-expression of Foxo in the intestinal stem cell lineage shortens lifespan by inhibiting stem cell proliferation and thus blocking proper tissue regeneration (Biteau et al., 2010). Over-expression of selected cytoprotective Foxo target genes in stem cells, on the other hand, is sufficient to prevent age-associated dysplasia and extend lifespan (Biteau et al., 2010). The data presented here show that chronic Foxo activation in intestinal enterocytes disrupts lipid metabolism by deregulating intestinal lipases, and thus highlight how cell- and tissue-specific consequences of Foxo function play an important role in determining either the beneficial (i.e. lifespan extension) or pathological (i.e. disruption of lipid metabolism) outcome of Foxo activation.
Recent work in C. elegans has begun to explore the relationship between lipid metabolism and longevity, revealing that increases in intestinal lipase expression can extend lifespan (O'Rourke et al., 2013; Wang et al., 2008). The beneficial effects of elevated lipase expression appear to be mediated by increases in specific types of fatty acids, which can activate autophagy and lead to lifespan extension (O'Rourke et al., 2013). Interventions that promote lipid homeostasis with age, such as JNK/Foxo inhibition in intestinal enterocytes, might thus affect healthspan and/or longevity through means other than primarily maintaining energy homeostasis.
The following strains were obtained from the Bloomington Drosophila Stock Center: OreR, w1118, y1w1, thor-lacZ, UAS-p35, and tub-Gal80ts. UAS-FoxoRNAi was obtained from the Vienna Drosophila RNAi Center (transformant ID 106097). NP1Gal4 was kindly provided by D. Ferrandon; UAS-Magro by C.Thummel; the foxoW24 mutant allele by M.Tatar and the foxoΔ94 allele by L. Partridge; GMRGal4, UAS-HepACT and UAS-BskDN by M. Mlodzik; esg-Gal4 and esg-LacZC4-1 by S. Hayashi; Su(H)-GBE-lacZ by S. Bray; puc-lacZ (pucE69) by E. Martín-Blanco. UAS-Foxo was previously described in (Puig et al., 2003). For aging and starvation sensitivity experiments, UAS-FoxoRNAi, UAS-BskDN, UAS-Magro, and UAS-Foxo transgenenic lines were backcrossed 8-10 generations into the w1118 background. Backcrossed UAS flies and w1118 male siblings were used to set up the crosses in order to reduce genetic background effects. The foxoW24 mutant was backcrossed into the yw background, and yw siblings were used as controls (Fig. 1E).
All flies were raised on standard yeast and molasses - based food, at 25°C and 65% humidity, on a 12 h light/dark cycle, unless otherwise indicated. Food for high-sugar/low-yeast (HSLY) diet feeding (and standard sugar/yeast (St. SY) control food) was made with the following protocol (as described in (Skorupa et al., 2008)): HSLY – 1.5 g agar / 40 g sucrose / 2.5 g yeast (yeast flake) / 0.3 mL propionic acid / 100 mL water; St. SY – 1.5 g agar / 10 g sucrose / 10 g yeast / 0.3 mL propionic acid / 100 mL water. Ingredients were combined, heated to at least 120°C, and cooled before pouring.
Total RNA from 8-10 guts or from 5 whole flies (without heads) was extracted using Trizol and cDNA synthesized using Superscript II (Invitrogen). Real time PCR was performed using SyberGreen, a Biorad IQ5 apparatus and the primers pairs described in Supplemental Information (Table S2). Results are average +/− standard error of at least 3 independent samples, and quantification of gene expression levels were measured relative to the expression of Actin5c.
Guts (10-20) were homogenized in protein sample buffer; proteins were separated by SDS-PAGE and transferred to nitrocellulose membrane using standard procedures. The following antibodies (all from Cell Signaling) were used: phsopho-Akt (anti-p-Drosophila Akt (S505); rabbit, 1:1000), total Akt (anti-pan Akt; rabbit, 1:1000), and β-actin (anti-beta-actin; rabbit, 1:5000). Signal was detected using HRP-conjugated anti-rabbit and chemi-luminescence (Pierce), according to manufacturer instructions.
Intact guts were dissected in PBS + 2mM MgCl2 and fixed for 10 minutes in 0.5% glutaraldehyde. Detection of β-galactosidase activity was carried out at room temperature in staining buffer (PBS, 2 mM MgCl2, 5 mM K4(Fe[CN]6), 5 mM K3(Fe[CN]6), 0.1% X-gal).
For TAG assays, 4 to 5 females (without the head) were homogenized in 150μl of PBST (PBS, 0.1% Tween 20), heated at 70°C to inactivate endogenous enzymes (homogenate cleared by centrifugation). 30μl of cleared extract was used to measure triglycerides concentrations according to the manufacturer instructions (StanBio Liquicolor Triglycerides). TAG levels were normalized to weight (measured using MT XS64 scale).
Guts were dissected in PBS and fixed in 4% formaldehyde/PBS for 20 minutes. Guts were then washed twice with PBS, and incubated for 30 minutes in fresh Oil Red O solution (6 mL of 0.1% Oil Red O in isopropanol and 4 mL DEPC water; and passed through a 0.45 –μm syringe), followed by rinsing with distilled water.
Intact guts were fixed at room temperature for 45 minutes in 100 mM glutamic acid, 25 mM KCl, 20 mM MgSO4, 4 mM Sodium Phosphate, 1 mM MgCl2, 4% formaldehyde. All subsequent incubations were done in PBS, 0.5% BSA, 0.1% TritonX−100 at 4°C.
The following primary antibodies were used: mouse anti-β-Gal (1:200), mouse anti-Prospero (1:250), and mouse anti-Armadillo (1:100) from Developmental Studies Hybridoma Bank. Fluorescent secondary antibodies were obtained from Jackson Immunoresearch. Hoechst was used to stain DNA.
Confocal images were collected using a Leica SP5 confocal system and processed using the Leica software and Adobe Photoshop.
Guts (8-10) were dissected in PBS, and then transferred to M3+BPYE (without serum) cell culture media with or without bovine insulin (0.1 units per 40 μL media, from Sigma) for 20 minutes. Samples were then prepared for protein extraction / Western blot analysis as previously described.
This work was supported by the National Institute on Aging (NIH RO1 AG028127 to H.J.), and by an AFAR/Ellison postdoctoral fellowship to J.K.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.