Rapamycin reduces Drosophila longevity under low nutrition

Abstract

Rapamycin treatment is considered a pharmacological intervention with the potential to mimic the longevity benefits of dietary manipulations. However, how rapamycin interacts with nutrition is not fully understood. Here we studied the effect of rapamycin on the longevity of Drosophila under a range of dietary conditions. In diets low in nutrients, rapamycin reduced longevity in a dosage-dependent manner. This dosage effect requires some nutrients as rapamycin has no impact on survival under starvation conditions. Under a balanced diet of yeast and sugar, rapamycin had no repeatable beneficial effect on organismal longevity. These results show that the effect of rapamycin on longevity is sensitive to the nutritional environment and it can reduce lifespan when nutrients are limited.

I. INTRODUCTION

Studies in mammals and invertebrates have identified rapamycin as a drug that may mimic the benefits of diet restriction and have anti-aging effects (1–8)_. Rapamycin inhibits the TORC1 and TORC2 complexes of the mammalian target of rapamycin (mTOR) pathway that regulates nutrient sensing, placing rapamycin at the interface between nutrition and longevity control (1–7). The promise of rapamycin as a pharmacological agent that can extend longevity calls for a thorough understanding of the molecular targets and environmental mediators of rapamycin’s effect on longevity. In genetically heterogeneous mice, rapamycin treatment can extend the life span of both males and females (9,10). In flies, two independent studies reported longevity effects of rapamycin (11,12). The most comprehensive of those studies showed increased lifespan of flies maintained on food supplemented with concentrations of rapamycin ranging from 50μM to 400μM. At a fixed concentration of 200 μM, a concentration at which flies are expected to incorporate a similar amount of rapamycin as that reported in studies of mice, rapamycin extended longevity in several diets (12). Independently, the second study reported that flies maintained on a nutrient rich food supplemented with 1, 10 or 100μM of rapamycin showed no significant differences in life span, but that rapamycin could have detrimental effects at 500μM or higher concentrations (11). Here we undertook an independent study of the effect of rapamycin on Drosophila longevity under a range of dietary conditions.

Since amino acid imbalance is a key determinant of Drosophila lifespan (13), the restriction of food in Drosophila is frequently performed via the dilution of yeast as a protein source. Here we measured the effect of rapamycin on longevity of Drosophila maintained on four different diets with a range of yeast levels (1, 2, 3, 4, 8 and 12% weight/volume) spanning the conditions that extend longevity in previous experiments (14). We further examined the effects of rapamycin on survival on a starvation diet (0.8%w/v agar in water with no additional nutrients). Rapamycin was added to the diets at different concentrations to explore both dose-dependent and presence-absence effects.

II. MATERIALS AND METHODS

All stocks were maintained and conducted under standard conditions (25 °C, 12 h light:12 h dark) on normal media (2% yeast, 11% sugar, 5.2% cornmeal, agar 0.8% w/v in water and 0.2% tegosept -methyl 4-hydroxybenzoate, from Sigma- St. Louis, MO, USA). All stocks were density controlled in replicate bottles using 48-hour egg lays by 20 pairs of parents for two generations prior to collection of flies for experimental assays. Adults were mated for 24 hours, anesthetized with moist CO₂, sorted by sex and allowed to recover for 24 hours before initiating longevity cages. 100 females were transferred to 1 liter demography cages. Dead flies were recorded and removed every 2 or 3 days, at which time fresh food was provided in a vial with 5ml of medium. We used three replicate cages for each treatment/genotype. For starvation studies flies were maintained in a 0.8% Difco bacto-agar in water. 0.1X SY food was prepared as described in (12). Rapamycin was purchased from LC Laboratories (Woburn, MA, USA.). Rapamycin was dissolved in ethanol and added to food, as described in (12).

III. RESULTS

3.1. Longevity effects of rapamycin under different diets

As expected, dietary restriction of yeast at 2, 3, or 4% diet causes a repeatable extension of longevity relative to 1% yeast, in two independent experimental blocks separated by 6 months (Figure 1, A and B). However, the addition of 200μM rapamycin to these diets does not show a consistent effect on longevity. On a diet of 2% yeast rapamycin extended longevity in one experiment (log rank test: χ² = 43.9; P<0.0001) but had no effect when repeated, resulting in a highly significant block effect in a proportional hazards model testing for main effects of rapamycin and block (χ² = 43.3; P<0.0001; see supplementary Table 1). On diets of 3% and 4% yeast the effects of rapamycin were reversed across repeat experiments, with one block showing non significant effects of rapamycin on these two diets and the second block showing significant effects in the opposite directions (see Figure 1 C and D; log rank test on 3% yeast: χ² = 5.91, P<0.015; on 4% yeast χ² = 11.15, P<0.0008). In a proportional hazards model, these data showed no main effect of rapamycin or block, but a highly significant rapamycin X block interaction (supplementary Table 1).

Figure 1. Longevity effects of rapamycin under different diets.

Longevity analysis of D. melanogaster strain OreR under 2% (circle) and 1% yeast food (triangle) exposed to rapamycin (open symbols) and vehicle control (filled symbols). Panels A and B show results from two independent experiments, block 1 and 2, respectively. (C and D) Mean longevity of flies under rapamycin treatment (triangle) and control (circle) in 1, 2, 3 and 4% yeast food. Experimental block 1 (C) and block 2 (D) (E) Survival analysis of D. melanogaster OreR strain under 8% yeast food with 200 μM rapamycin added. (F) Mean longevity values for the data in (E). * P< 0.05 log rank test.

Interestingly, in a diet low in nutrients (1% yeast), rapamycin consistently decreased the mean lifespan in both experimental blocks by 28% and 16.5% respectively (see Figure 1 A–D; log rank test for 1% yeast: χ² = 54.09, P<0.0001; χ² = 13.49, P<0.0002 for block 1 and 2 (Figure 1C and D), respectively). A proportional hazards model showed a highly significant effect of rapamycin, no effect of block and a weak rapamycin X block interaction (P< 0.0151) that does not withstand correction for multiple tests for these four diets (Figure 1, and supplemental Table 1). Supplementary Table 1 presents a summary of the rapamycin and repeat-experiment (block) effects for each diet condition (1, 2, 3, an 4% yeast), as revealed by analysis of variance. The 1% yeast diet is unique in that it is the one condition that shows a highly significant main effect of rapamycin that reduces longevity coupled with no significant main effect of block (repeat experiments are consistent). The other diet conditions, ranging from 2 to 4% yeast, either show no significant main effect of rapamycin, or significant main effects of block, implying that the rapamycin effects are not repeatable.

In an effort to determine if the longevity-extending effect of rapamycin were restricted to high-nutrient diets, we conducted a third independent longevity assay using 2% and 8% yeast. The 2% treatment was used to provide a reference to the other experimental blocks, since block effects are common in longevity assays. Rapamycin significantly extended the longevity of flies fed both 2% and 8% yeast in this third experiment, (log rank test for 2% yeast: χ² = 40.663, p<0.001], for 8% yeast: [χ² = 24.475, p<0.001]). Analysis of variance for the effects of rapamycin and yeast showed a significant effect of rapamycin treatment in both diets and a modest effect of the yeast diet (ANOVA: treatment [DF: 1, F: 42.332, p<0.001], yeast [DF: 1, F: 4.116, p=0.043] and yeast*treatment interaction [DF: 1, F: 0.165, p=0.685]) (Figure 1E–F). It should be noted that the mean longevities in this third experiment were higher than in other experimental blocks, but the expected effect of somewhat reduced longevity under richer diet is observed (8% shows shorter longevity than 2% yeast; see Figure 1E–F). The longevity of the OreR strain of D. melanogaster is known to be rather unresponsive to diet in the 5–15% range on SY media, relative to other genotypes (15).

We next conducted a fourth independent study using increasing concentrations of rapamycin (0, 5, 50, 100, 200 and 400 μM) in 2% and 12% yeast diet (Figure 2). Concentrations of rapamycin ranging from 50 to 200 μM extended the mean longevity of flies fed 2% yeast diet, reaching maximum significant extension at 200 μM rapamycin (log rank test for 2% yeast at 200 μM [χ² = 6.435, p<0.01]) and decreasing at 400 μM (Figure 4A and C). However, increasing concentrations of rapamycin were not able to extend longevity of flies fed a 12% yeast diet (Figure 2B and C) suggesting that higher concentrations of rapamycin are needed to extend the longevity of flies fed a very high yeast diet. Although there was not a significant difference in the longevity of the flies fed the yeast diets at 0 μM rapamycin, analysis of variance for the effects of rapamycin and yeast showed a significant effect of the yeast diet (ANOVA: diet [DF: 1, F: 8.911, p<0.003]) and yeast*treatment interaction (ANOVA [DF: 1, F: 3.31, p=0.006]).

Figure 2. The longevity effect of different rapamycin concentrations.

Longevity analysis of D. melanogaster strain w^Dha under 2% (A) and 12%(B) yeast food exposed to 0, 5, 50, 100, 200 and 400 μM. (C) Mean longevity for the data in A and B. *P< 0.05 log rank test for the comparison between 0 and 200 μM.

Figure 4. The effect of rapamycin on starvation resistance.

Survival analysis of D. melanogaster OreR (A) and w^Dha (B) strains under starvation (0.8% agar in water) with 0.5 5, 10, 50 and 200 μM rapamycin added. (C) Mean longevity values for the data in (A) and (B).

One other thing to note is the more linear shape of the survivorship curves in Figure 2 compared to Figure 1. Since the longevity measurement represented in figure 2 was performed at a different laboratory, we attribute this difference to a genotype X environment interaction (GxE) (16).

3.2. Longevity effects of rapamycin under low nutrition diets

We hypothesized that low-nutrient diets might require lower concentrations of rapamycin to obtain the beneficial effects on longevity. To test this hypothesis we measured lifespan on 1% yeast using two standard laboratory strains (w^Dha and OreR) and rapamycin concentrations ranging from 5 to 200μM. At 5μM, rapamycin decreased median lifespan of w^Dha strain but did not significantly affect the lifespan of OreR, relative to the 0μM control. At higher concentrations, rapamycin decreased the life span of both strains in a dose-dependent manner (Figure 3C). Previous studies reported an extension of longevity by rapamycin in all diets tested, even in those with low nutritional value (12). Because the low nutritional value diet used in that study differs from the diet used in our experiments, we repeated the survival analysis using the same low nutritional diet as published by Bjedov et al. 2010 (0.1X SY,(12)). Rapamycin significantly shortened the mean lifespan of both strains when maintained in a 0.1X SY diet (Figure 3D,E).

Figure 3. Longevity effects of rapamycin under low nutrition diets.

Survival analysis of D. melanogaster strains OreR (A) and w^Dha (B) on 1% yeast food with 0, 5, 50 and 200 μM rapamycin added. (C) Mean longevity for the data in A and B. (D) Survival analysis of OreR and w^Dha strains in 0.1X SY food with 200μM rapamycin added. (E) Mean longevity values for the data in (D). * P< 0.05 log rank test.

3.3. The effect of rapamycin on starvation resistance

We also studied the effect of rapamycin under starvation conditions. w^Dha and OreR strains maintained in a starvation diet were exposed to 5, 50 and 200 μM of rapamycin. Rapamycin did not have any effect on the resistance to starvation of w^Dha and OreR flies (Figure 4A, B). Interestingly, w^Dha strain showed an overall higher resistance to starvation than OreR strain (Figure 4C). Several technical points are worth noting for these starvation assays. Bjedov et al (2010) used 1.5% agar in starvation assays – twice the concentration of our assay – and flies survived 1–2 days longer in their study than what we report here, suggesting that agar may have some nutritional value for Drosophila. Rapamycin has low solubility in water (17), and the amount of rapamycin delivered to flies in the no-nutrient agar-water medium may be lower than in media containing nutrients that could act as co-solvents for rapamycin delivery. However, all rapamycin exposures used ethanol as a co-solvent to enhance solubility(17).

IV. CONCLUSION

In summary, there are three conclusions from these experiments: 1) Under limited nutrition, rapamycin decreases the longevity of two Drosophila strains in a dosage-dependent manner (Figure 3) in two independent low-nutrition diets (1% yeast and 0.1X SY food). 2) Rapamycin does not consistently extend longevity of Drosophila melanogaster under unrestricted diets (Figure 1 and 2). 3) The detrimental effects of rapamycin are not observed in two Drosophila strains when under starvation conditions (Figure 4). The lack of a detrimental effect of rapamycin under starvation suggests that the pharmacological effects of rapamycin depend upon the nutritional state of the fly. Thus, as a dietary restriction mimetic, the addition of rapamycin to a diet that is already low in nutrients may exacerbate malnutrition and reduce longevity. It may be that decreased longevity is not observed under starvation because rapamycin cannot further restrict dietary nutrients. However, rapamycin may need more time than only a few days to modulate longevity and its effect under starvation might be hindered by the short longevity of starved flies.

Although we did find that rapamycin can extend longevity at balanced diets with non-limiting nutritional conditions (Figure 1), these effects were not repeatable in different experimental blocks. Of the seven longevity experiments conducted on yeast diets in the 2–4% range, rapamycin extended longevity in three experiments, had no effect in three experiments, and reduced longevity in one experiment (Figure 1). These data suggest that longevity extension by rapamycin under particular diets is not sufficiently repeatable that it can over-ride variable environmental conditions that are typically associated with replication of longevity experiments at different times, even using identical nutrient conditions.

Environmental interactions in rapamycin treatment might also explain the different results described in the literature (11,12). Rapamycin strongly reduces female fecundity (12), and reproductive output typically shows a negative correlation (trade-off) with longevity. The variable evidence for longevity extension by rapamycin in Drosophila may be partly attributable to indirect effects on longevity due to variable inhibition of reproductive effort when nutrients are sufficient to support egg production. The detrimental effects of rapamycin under low nutrients that we describe here may be heightened if the nutrient level is sufficiently low that egg production is already turned off, preventing any further beneficial effect of rapamycin from its inhibitory impact on reproduction. While this trade-off hypothesis is plausible, it remains possible that the distinct results from different experiments represent subtle variation in nutrient levels that confound the delivery or metabolism of rapamycin upstream or downstream of its site of action.

While rapamycin may indeed be a promising drug for the treatment of age related diseases or the extension of lifespan, the complex interactions reported here suggest that the precise nutritional and environmental conditions under which rapamycin may realize this promise are not fully understood. There is ongoing controversy in the field about the generality of the effects of rapamycin (18,19), and it is incumbent on researchers in this field to resolve this uncertainty through independent, replicated experiments. Given the variable results from model organism studies conducted in different laboratories with standard diets, temperatures and growth conditions, the use of rapamycin to treat human ailments will be challenging given the widely varying dietary and environmental conditions to which the general public are exposed. With additional studies of how diet may modify the effects of rapamycin, these challenges may well be overcome.