Increased longevity mediated by yeast NDI1 expression in Drosophila intestinal stem and progenitor cells

Expression of ndi1 in intestinal stem and progenitor cells extends lifespan

The intestine is a critical target organ with respect to genetic manipulations that can extend longevity [17], as has been shown previously with dPGC-1 upregulation [11]. To better understand the relationships among mitochondrial respiratory chain activity, intestinal homeostasis, and lifespan determination, we expressed a previously described UAS-ndi1 construct [14, 15] in the Drosophila intestine using the intestine-specific RU486-inducible Gene-Switch driver line TIGS-2 [18]. Unlike the endogenous Drosophila ETC complex I which is sensitive to rotenone inhibition but insensitive to flavone, NDI1 is insensitive to rotenone but inhibited by flavone [19]. Induced expression of ndi1 in the adult intestine produced a robust rotenone-insensitive, flavone-sensitive NADH dehydrogenase activity in mitochondria isolated from intestines (Figure 1A). Control flies from the same background strain that were not provided RU486 did not show detectable levels of rotenone-insensitive, flavone-sensitive NADH dehydrogenase activity, supporting the fidelity of the Gene-Switch system [20, 21] and functionality of the ndi1 transgene and NDI1 protein in the adult fly intestine.

We used this system to examine the impact of intestine-specific expression of ndi1 on Drosophila lifespan. Induced expression of ndi1 using the TIGS-2 driver throughout the life of the fly resulted in a significant increase in lifespan in female flies (Figures 1B and S1A) and no major effect in male flies. RU486 produced no major effects on longevity in control flies (Figure S1B). To examine the impact of targeted expression of ndi1 in intestinal stem cell lineages (ISCs and EBs), we first used the constitutive esgGAL4 driver line and observed a significant extension of lifespan in both female (Figures 1C and S1C-D) and male flies (Figures S1E-F) compared to controls. esgGal4 expression is restricted to ISCs and EBs in the intestine, however, it is also expressed in stem cells within malpighian tubules, germline and somatic stem cells in the testis, and in salivary glands [22]. Therefore, to validate and extend this finding we took advantage of the RU486-inducible 5961GS driver which recapitulates the esgGal4 expression pattern in the digestive tract (ISCs/EBs and malpighian tubule stem cells) [22, 23] but is not expressed in salivary glands [22] or testis (C.L.K. and D.L.J., unpublished data). Induced expression of ndi1 during adulthood, via 5961GS, resulted in a significant lifespan increase in females (Figures 1D and S1G-I) but not in males (Figures S1J-S1K). Expression of ndi1 during adulthood using a Gene-Switch driver that is expressed in EBs and post mitotic enterocytes (ECs) (5966GS, [23]) failed to increase lifespan (Figures S1L-M), implicating expression in ISCs as the major contributor to longevity. The largest and most consistent lifespan extension phenotypes using ndi1 expression were observed with female flies. Therefore, unless noted otherwise, we focused our studies to female flies for the remainder of this study.

ndi1 expression in ISCs/EBs improves markers of intestinal homeostasis during aging

Homeostasis of the digestive tract has been shown to play a central role in lifespan determination in Drosophila [10, 11, 17, 22]. Therefore, we examined markers of intestinal homeostasis in flies that express ndi1 in ISCs/EBs. First, we set out to determine whether ndi1 could delay the onset of markers of ISC proliferation and the accumulation of misdifferentiated ISC daughter cells reported to occur in the aged midgut [8, 9]. Consistent with improved intestinal tissue homeostasis, examination of aged flies that express ndi1 in ISCs/EBs along with an esg reporter (UAS-gfp) revealed a significant decrease in the number of esg positive cells in the midgut relative to controls (Figure 2A-B). In addition, we also observed a delay in the precocious activation of ISC proliferation, as measured by phosphorylation of histone H3 (pHH3), a marker of cell cycle progression through mitosis. Female flies, 50 days post eclosion, expressing ndi1 under the control of esgGAL4 driver contained significantly fewer pHH3⁺ cells, when compared to controls (Figure 2C). No difference in the number of pHH3⁺ cells was observed in 10 day old flies, indicating that ndi1 expression specifically delays the age-related increase in ISC proliferation.

An increase in reactive oxygen species (ROS) has been implicated in the loss of tissue homeostasis in the aged fly intestine [24, 25]. Previously, we reported that pan-neuronal expression of ndi1 can reduce ROS levels in the aged brain [15]. It is unclear, however, whether ndi1 expression only in progenitor cells of a tissue is sufficient to cause such changes throughout the tissue. To test this idea, we examined the endogenous levels of ROS in the intestines of control and esgGAL4>ndi1 flies using dihydroethidium (DHE), a redox-sensitive dye that exhibits increased fluorescence intensity when oxidized [26]. Targeted expression of ndi1 in ISCs/EBs led to a reduction of DHE fluorescence in these cells and throughout the aged intestine (Figures 2D-E, S2A).

Loss of intestinal integrity can be assayed in living flies by monitoring the presence of non-absorbed dyes (e.g., FD&C blue No. 1) outside of the digestive tract after feeding [10, 11]. To determine whether ndi1 can delay the onset of intestinal barrier dysfunction, we examined flies of different ages fed FD&C blue No. 1 for evidence of this dye outside of the digestive tract. The proportion of aged flies with dye outside of the intestine was significantly lower in flies with ISC/EB ndi1 expression (Figure 2F). This was not a result of altered development, as adult onset induction of ndi1 in ISCs/EBs, using the 5961GS driver, was sufficient to decrease the proportion of flies with dye outside of the digestive tract with age (Figure 2G and S2B). Loss of intestinal integrity has been linked with a systemic increase in expression of immunity-related genes [10]. Hence, we assayed systemic expression levels of several anti-microbial peptides (AMPs) genes in esgGAL4>ndi1 and control flies during aging. In line with decreased intestinal barrier dysfunction, flies that express ndi1 in ISCs/EBs show significantly lower expression of multiple AMPs in whole bodies later in life (Figure 2H). Taken together, our findings show that ISC/EB-specific expression of ndi1 leads to improved intestinal homeostasis during aging.

ndi1 expression in ISCs/EBs does not affect fertility or physical activity but changes sensitivity to some stresses

To gain further insight into intestinal ndi1-mediated longevity, we examined a number of physiological and behavioral parameters in long-lived esgGAL4>ndi1 flies and controls. Neither male nor female flies that express ndi1 in ISC/EBs showed consistent alterations to fertility (Figures S2C-G). Resistance to oxidative stress, assayed by survival under hyperoxia (80% O₂), was similarly unaffected (Figure S2H), suggesting that ROS levels in the intestinal epithelium are not limiting for survival under severely hyperoxic conditions. Survival in elevated environmental temperatures (37°C) and water-only starvation showed considerable differences, with ndi1 expressing flies showing significantly greater sensitivity to elevated temperatures (Figure S2I), and greater resistance to starvation (Figure S2J). These changes were not correlated with significant differences in either spontaneous locomotor activity per time of day (Figure S2K) or cumulative activity over 24-hour periods (Figure S2L). Together, these data indicate that intestinal ndi1-mediated longevity is not associated with a general increase in stress resistance or a decline in reproductive output.

ndi1 expression in ISCs/EBs stimulates feeding behavior

A moderate reduction in food intake, dietary restriction (DR), can extend lifespan in diverse organisms, possibly by reducing the intake of specific nutrients [4]. To determine if a gross difference in food intake could play a role in ndi1-mediated longevity, we assayed feeding behavior in esgGAL4>ndi1 flies and controls. Surprisingly, total food consumption, measured using a capillary feeding (CAFE) assay [27], revealed an overall increase in feeding in flies that express ndi1 in ISCs/EBs at both young and aged time points (Figure 3A). An independent assay of feeding using a modified dye-tracking assay [28] was used to parse the feeding behavior into the proportion of flies that feed within the assay period and the meal size of flies that feed. Expression of ndi1 in ISCs/EBs resulted in significant increases in both the proportion of flies that feed (Figure 3B) and their meal sizes (Figure 3C) in both young and aged flies. 24 hour activity profiles of ndi1 expressing flies are similar to controls, suggesting that an altered activity at different times of day is not responsible for the increased feeding during the assay period (Figure S2K). Moreover, adult-onset expression of ndi1 for 10 days in ISCs/EBs, using the 5961GS driver, was sufficient to confer an increase in total food consumption (Figure 3D) and meal size (Figure 3E). The presence of the inducing drug itself had no significant effect on total feeding (Figure S3B) or meal size (Figure S3C).

To determine whether increased feeding was associated with alterations in defecation, we examined the material excreted by esgGAL4>ndi1 flies and controls. Although ndi1 expressing flies ate significantly more than controls, excreta number were not significantly different than controls, at both young and aged time points (Figure S3D). Recent work has shown that qualitative analysis of excreta can provide insight into intestinal transit and fluid homeostasis [29]. Specifically, flies that are starved for nutrients and fluids, as during times of high fecundity in females, were shown to have increased frequency of “reproductive oblong deposits” (RODs) and lower fecal pH. Closer examination of excreta shape of young and aged esgGAL4>ndi1 flies showed a significant reduction in the frequency of RODs (Figures 3F and S3E) indicating improved fluid availability, less concentrated intestinal contents, and quicker intestinal transit [29]. Similarly, fecal pH analysis of flies maintained on bromophenol blue (BPB) containing diets showed less acidic fecal deposits in ndi1 expressing flies at the young time point (Figure 3G), consistent with a quicker transit through the intestinal tract. Together, these findings indicate that expression of ndi1 in ISCs/EBs in addition to improving tissue homeostasis, improves intestinal function.

Next, we set out to determine whether esgGAL4>ndi1 flies show systemic physiological changes that are consistent with increased nutrient uptake. Whole body weight measurements indicated that esgGAL4>ndi1 flies are heavier than controls, and maintain their weight during aging (Figure 3H). As with feeding behavior, adult-onset expression of ndi1 for 10 days in ISCs/EBs, using the 5961GS driver was sufficient to confer an increase in body weight (Figure 3I), and this was not a result of the inducing drug itself (Figure S3F). Moreover, aged esgGAL4>ndi1 flies display increased protein levels and triglyceride stores relative to controls at aged timepoints (Figures 3J-K, S3G). Unlike triglycerides, levels of glycogen declined similarly in both ndi1 expressing flies and controls with age (Figure S3H).

ndi1 expression in ISCs/EBs alters systemic metabolic signaling pathways

We set out to further characterize the physiology of long-lived ndi1 flies by examining steady state effects of ndi1 expression on systemic nutrient sensitive pathways. AMP-activated protein kinase (AMPK) is a crucial metabolic gauge that is activated by low cellular energy status [16]. Since expression of ndi1 in ISCs/EBs stimulates feeding, we reasoned that these flies may show reduced systemic AMPK activity. Indeed, Western blots specific for phosphorylated AMPKrevealed significantly decreased phosphorylation at Thr184 in whole bodies of esgGAL4>ndi1 flies relative to controls (Figures 4A-B and S4A). AMPK activation has been shown to stimulate sirtuin1 (SIRT1) activity, which deacetylates FOXO and increases its transcriptional activity [30]. Consistent with this, systemic dFOXO transcriptional activity, assayed by measuring transcript levels of multiple direct downstream targets of dFOXO in whole bodies, was significantly decreased in esgGAL4>ndi1 flies (Figure 4C).

FOXO activity is independently regulated by a number of different signaling pathways, including the insulin/insulin-like growth factor signaling (IIS) pathway [31]. To determine whether the observed decrease in dFOXO activity in esgGAL4>ndi1 flies was also associated with an increase in systemic IIS, we assayed the activation state of the direct mediator of FOXO activity in the IIS pathway, phosphoinositide-3-OH-kinase-dependent serine/threonine protein kinase (AKT). In whole bodies of esgGAL4>ndi1 flies the phosphorylation of AKT was not significantly affected at either the IIS phosphorylation site (Thr423) or the TORC2 phosphorylation site (Ser505), nor was there a difference in total AKT protein levels (Figures 4D-E and S4B). Similarly, Western blot analysis of S6K, a component of the TOR pathway revealed no significant changes in phosphorylation at the TORC1 target residue (Thr398) or in total S6K levels in whole bodies of ndi1 expressing flies (Figures S4C-D). Thus, the observed decrease in FOXO transcriptional activity in esgGAL4>ndi1 flies is not associated with alterations in systemic IIS/TOR pathway activities.

ndi1 expression affects both feeding behavior and longevity, both of which have been shown to be regulatable by Drosophila insulin-like peptides (dilps) signaling [32-34]. Therefore, we checked the expression levels of several dilps to determine if dilp levels in heads of esgGAL4>ndi1 flies were altered. Transcript levels in heads of dilp2,dilp3 and dilp5, that are expressed in the insulin producing cells (IPCs) upon feeding, were significantly decreased in esgGAL4>ndi1 flies (Figure 4F) while the transcript level of dilp1, which is not expressed in adult heads [32], was not altered significantly (Figure S4E). Direct quantification of DILP2 protein levels in IPCs by immunofluorescence showed a similar decrease in DILP2 fluorescence in IPCs of esgGAL4>ndi1 flies (Figure 4G and 4H). The Drosophila ortholog of mammalian neuropeptide Y, short neuropeptide F (sNPF), is expressed in the nervous system and regulates food intake [35] and along with its cognate receptor, sNPFR1, regulates the expression of dilps in the fly brain [36]. To determine if decreased dilp transcription in heads of esgGAL4>ndi1 flies is linked to altered sNPF/sNPFR1 expression, we measured transcript levels of sNPF and sNPFR1 in heads. In line with dilp transcript levels in heads, both sNPF and sNPFR1 transcript levels were decreased in heads of esgGAL4>ndi1 flies relative to controls (Figure 4I and 4J).

Fly lines, culture, and genotypes

UAS-ndi1 lines [15] were backcrossed 10 times into a w¹¹¹⁸ background. TIGS-2 line was provided by L. Seroude, 5961GS was provided by H. Jasper, and esgGAL4 was provided by A. Christiansen. Culturing of flies and measurements of lifespan were performed as previously described [11]. See Supporting Information Materials and Methods for details.

Genotypes:TIGS>+:+;+;TIGS-2/+,TIGS>ndi1:+;UAS-ndi1/+;TIGS-2/+,esgGAL4>+:+;esgGAL4,UAS-gfp/+;+,esgGAL4>ndi1:+;esgGAL4,UAS-gfp/UAS-ndi1;+,5961GS>+:+;5961GS,UAS-gfp/+;+,5961GS>ndi1:+;5961GS,UAS-gfp/UAS-ndi1;+,5966GS>+:+;5966GS/+;+,5966GS>ndi1:+;5966GS/UAS-ndi1;+

Feeding Assays

Capillary feeding (CAFE) assays and dye-tracking assays were adapted from previously described protocols with slight modifications [27, 28]. See Supporting Information Materials and Methods for details.

Molecular Biology and Physiology

Protocols and reagents used for complex I and NDI1 activity assays, immunofluorescence staining, ROS staining, quantitative real-time polymerase chain reaction (qRT-PCR), intestinal barrier and transit assays, and measurements of fertility, stress resistance, spontaneous activity, weights, and metabolites are provided in Supporting Information Materials and Methods.

Statistical Analyses

Unless otherwise indicated, significance was determined using a two-tailed, unpaired t test from at least three independent experiments and expressed as p values. All error bars reflect standard error of the mean.