Fragile lifespan expansion by dietary mitohormesis in C. elegans

The UPR^mt declines during aging

The induction of hsp-6p::GFP by high dietary glucose was observed in young adult animals (Fig. 1a). It has been hypothesized that aging may occur partly from an age-related decline of protein homeostasis mechanisms [9], thus we wondered if the induction of hsp-6p::GFP by glucose was maintained in aging animals. We observed that exposure to high dietary glucose at three concentrations induced the expression of hsp-6p::GFP in animals at days 1 and 3 (Fig. 2a, b) of adulthood, but this induction was lost in animals at days 5 or 9 of adulthood (Fig. 2c, d). To inspect the morphology of the mitochondrial network we used transgenic worms expressing a mitochondrial targeted signal in muscles cells [3]. Normally the mitochondrial network appears as tubular, parallel structures in muscle tissue [1,10].

We observed that these structures became increasingly disordered when the worms were exposed to high dietary glucose as well as during aging (Supp. Fig. 2). These data suggest that the ability of the UPR^mt to respond to cellular toxicity decreases during aging and mitochondrial health is damaged by high dietary glucose.

Figure 2. Glucose enrichment maintains HSP-6 expression over time. (a-b) Photographs and quantification of increased hsp-6p::GFP expression in strains exposed to 1, 2 or 4% GE at days 1 and 3 of adulthood (****P<0.0001, mean ± SD). (c-d) GE failed induced expression of the hsp-6::GFP reporter in worms at days 5 or 9 of adulthood (mean ± SD). (e) RNAi against atfs-1, dve-1, or ubl-5 reduced the lifespan of N2 worms compared to controls (P<0.0001 compared to empty vector (EV) controls). (f) Worms grown on EV had reduced lifespan only at 4% GE (P<0.0001 compared to controls). Animals grown on (g) atfs-1, (h) dve-1 or (i) ubl-5 RNAi had reduced lifespan at 2% and 4% GE (P<0.0001 compared to EV control).

The UPR^mt protects against glucose-mediated lifespan reduction

We next investigated whether the UPR^mt protects against glucose-induced reduction of lifespan in adult animals. Wild type N2 animals exposed to RNAi against atfs-1, dve-1, or ubl-5 all showed decreased lifespan suggesting that a fully-functional UPR^mt is necessary for promoting longevity (Fig 2e). Wild type N2 worms exposed to 2% glucose had lifespan similar to untreated controls, but exposure to 4% glucose reduced lifespan (Fig 2f). Interestingly, lifespan was reduced in worms exposed to either 2% or 4% glucose and treated with atfs-1, dve-1 or ubl-5 (RNAi) (Fig 2g-i). These data demonstrate that 2% glucose becomes toxic and reduces lifespan in the absence of a functional UPR^mt, thus the UPR^mt may protect organisms against cytotoxicity from dietary sources.

Early developmental exposure to glucose extends lifespan

The induction of hsp-6p::GFP by glucose does not indicate if the UPR^mt is activated to mitigate the toxic effects of glucose, or instead mediates the protective effects of high dietary glucose. Activation of the UPR^mt has been observed after inhibition of mitochondrial DNA (mtDNA) translation, or mtDNA depletion and is involved in longevity phenotypes [1]. Furthermore, the loss of UPR^mt results in increased susceptibility to stress, while prolonged induction reduces lifespan (12-14). To delineate the role of the UPR^mt in glucose-mediated aging phenotypes, we investigated whether activation of the UPR^mt by high dietary glucose could be the cause of glucose-dependent lifespan reduction. Wild-type animals exposed to high dietary glucose for their entire life had reduced lifespan compared to untreated animals, in agreement with previous findings by us [5] and others [6] (Fig 3a). Surprisingly, animals exposed to glucose (2% and 4%) only until the pre-adult L4 larval stage, showed the opposite effect with significant lifespan extension (Fig. 3b). Furthermore, exposing worms to high dietary glucose only during adulthood reduced lifespan compared to control animals (Fig. 3c), suggesting a lifespan-promoting role of glucose during development that becomes deleterious at later adult stages. To determine if the augmented lifespan effect of early glucose exposure was dependent on the UPR^mt, we investigated the contribution of the transcription factor ATFS-1, which is a master regulator of the UPR^mt [7]. We exposed afts-1(gk3094) and atfs-1(et15) mutants to high dietary glucose from hatching until the L4 larval stage and observed that the enhanced longevity phenotype was lost (Fig. 3d,e). These data suggest that endogenous UPR^mt activity via ATFS-1 is required for regulating glucose-dependent lifespan phenotypes.

Figure 3. Early or late glucose enrichment has opposite effects on lifespan. (a) GE reduced lifespan of N2 animals at 4% concentration compared to untreated animals (P<0.0001 compared to control). (b) Larval exposure to glucose extends lifespan of N2 animals at concentrations of 2% or 4% compared to untreated animals (P<0.0001 compared to controls). (c) Adult exposure to glucose reduces lifespan of wild-type animals (P<0.0001 for all conditions compared to control). (d-e) GE failed to extend lifespan in atfs-1(et15) and atfs-1(gk3094) animals when treatment ceased at the L4 larval stage. (f) GE failed to extends lifespan in N2 animals when treated with 10 mM N-acetyl cysteine.

Dietary induced mitohormesis extends lifespan

Mitochondria are the major ROS producers at the cellular level and it has long been proposed that ROS molecules and by-products are responsible for organismal aging (15). Recent work suggested that ROS rather acts as a signalling molecule with important roles in regulating longevity [11,12]. This phenomenon is known as mitohormesis [13] and the protective effects are lost in the presence of antioxidant compounds. Thus we wondered if the lifespan enhancing effect from early glucose exposure utilized a mitohormetic mechanism requiring ROS. We observed that animals treated with 10 mM N-acetyl cysteine did not have increased lifespan when exposed to high dietary glucose until the L4 stage GE (Fig. 3f). These data suggest a role for oxidative stress in glucose mediated lifespan extension when applied at early, pre-adulthood stages.

Lifespan extension by glucose requires a functional electron transport chain

Mutations that disrupt mitochondrial function can result in severe pathologies, and mitochondrial dysfunction is linked to multiple neurodegenerative disorders in humans (19-21). Surprisingly, electron transport chain (ETC) dysfunction can extend lifespan not only in C. elegans but also in Drosophila and mice [14]. Thus, we wondered if ETC function was required for the temporal effects on lifespan from high dietary glucose. We observed that the ETC complex II mutant mev-1(kn1) and the complex III mutant isp-1(qm150) strains showed no increase in lifespan when exposed to high dietary glucose until the L4 larval stage (Supp. Fig. 3). These data suggest that functional mitochondria are required for the increased lifespan from early exposure to high dietary glucose. This is consistent with the previous studies demonstrating that lifespan extending interventions can only be produced at late larval stages (L3/L4) when mitochondrial biogenesis is at it's peak [15].

Activation of the UPR^mt during adulthood abolishes lifespan extension

Glucose treatment over the life of the worm is known to reduce lifespan and in this study we have shown that it is linked to increased hsp-6 expression. This suggests that activation of the UPR^mt during adulthood may reduce lifespan. We wondered if early exposure only might have a similar pattern of hsp-6 expression.

Worms were treated with glucose during larval development and then grown on control NGM. At different aging stages, the animals were transferred to cytochrome c-oxydase cco-1(RNAi) plates for a 12h period. cco-1(RNAi) is known to induce the UPR^mt by disrupting ETC function, and knockdown of this gene during development has been reported to mimic ETC dysfunction with similar longevity phenotypes to knockdown of other ETC genes [16]. Worms grown on NGM and placed on cco-1 RNAi as adults showed no hsp-6::GFP increased expression, consistent with the temporal activation of the UPR^mt [17] (Fig. 4). Similar to untreated animals, nematodes exposed to dietary glucose early in life showed no increase in hsp-6::GFP expression. However, animals treated with glucose during development plus adult specific cco-1 RNAi had increased hsp-6::GFP expression (Fig. 4a). This effect was sustained until adult day 5 for all glucose concentrations tested (Fig. 4b, c) and was lost at adult day 9 for the lowest concentration (1%) (Fig. 4d). We wondered if there were functional consequences of additional mitochondrial stress during adulthood on lifespan. We observed that worms grown on NGM plates and exposed to cco-1(RNAi) at day 1 of adulthood had lifespans similar to untreated controls (Fig 4e). We then tested animals that had been exposed to either 2% or 4% GE during development, and then exposed them to cco-1(RNAi) at day 1 of adulthood, and observed that lifespan was reduced compared to control animals (Fig. 4f, g). Thus, early exposure to mitochondrial stress can increase lifespan, but this is a fragile phenotype since further induction of the UPR^mt during adulthood results in lifespan reduction.

Figure 4. Early exposure to glucose sensitizes nematodes to UPR^mt activation. (a-d) Early exposure (L4) to glucose at different concentrations (1, 2 and 4%) had no effect on hsp-6 expression during adulthood compared to control NGM conditions. RNAi against cco-1 was able to induce hsp-6 expression in animals treated at all glucose concentrations until day 5 of adulthood, and only at 2% and 4% at day 9 (mean ± SD., *P<0.05, **P<0.01 and ****P<0.0001 compared to control). (e) The lifespan on N2 worms treated with cco-1(RNAi) during adulthood was similar to N2 control worms. The lifespans of N2 worms grown on (f) 2% or g. 4% GE plates was reduced when treated with adult-stage specific cco-1(RNAi) compared to worms treated with EV (P<0.01 for 2% GE + cco-1(RNAi) and P<0.0001 for 4% GE + cco-1(RNAi) versus controls.)

Worm strains and genetics

Worms were maintained on standard NGM plates streaked with OP50 E. coli. In some experiments D-glucose was added to NGM plates. All strains were scored at 20°C unless indicated. Strains used in this study were obtained from the C. elegans Genetics Center (University of Minnesota, Minneapolis) unless indicated and include: atfs-1(gk3094); atfs-1(et15); mev-1(kn1); isp-1(qm150); zcIs13[hsp-6::GFP]; zcIs9[hsp-60::GFP]; zcIs39[dve-1p::dve-1::GFP] ; zcIs19[ubl-5p::ubl-5::GFP]; zcIs14[myo-3::GFP(mit)] ; ccIs4251[myo-3p::GFP(NLS)::LacZ (pSAK2) + myo-3p::GFP(mit) (pSAK4) + dpy-20(+)] and sid-1(pk3321).

Fluorescence microscopy

Animals were immobilized in M9 with 5 mM levamisole and mounted on slides with 2% agarose pads. Animals were visualized with a Leica 6000 and a Leica DFC 480 camera. A minimum of 30 animals was scored per treatment over 3 trials. Mito-GFP animals were visualized at days 1, 5 and 9 of adulthood with a Leica TCS_SP5 confocal microscope and analyzed with LAS AF software. The mean and SEM were calculated for each trial and two-tailed t-tests were used for statistical analysis.

Lifespan assays

Worms were grown on NGM or NGM + 1%, 2% or 4% glucose and transferred on NGM or NGM + glucose. 30-40 animals/plate by triplicates were tested at 20°C from adult day 1 until death. Worms were declared dead if they didn't respond to tactile or heat stimulus.

RNAi experiments

RNAi-treated strains were fed with E. coli (HT115) containing an Empty Vector (EV), atfs-1 (ZC376.7), ubl-5 (F46F11.6), cco-1 (F26E4.9) RNAi clones from the ORFeome RNAi library (2) and dve-1(ZK1193.5) clone from Arhinger RNAi library. RNAi experiments were performed at 20°C. Worms were grown on either NGM or NGM + 2% glucose both enriched with 1 mM Isopropyl-b-D-thiogalactopyranoside (IPTG).

Statistics

For paralysis and stress-resistance tests, survival curves were generated and compared using the Log-rank (Mantel-Cox) test, and a 60-100 animals were tested per genotype and repeated at least three times.