A network pharmacology approach reveals new candidate caloric restriction mimetics in C. elegans

Abstract

Caloric restriction (CR), a reduction in calorie intake without malnutrition, retards aging in several animal models from worms to mammals. Developing CR mimetics, compounds that reproduce the longevity benefits of CR without its side effects, is of widespread interest. Here, we employed the Connectivity Map to identify drugs with overlapping gene expression profiles with CR. Eleven statistically significant compounds were predicted as CR mimetics using this bioinformatics approach. We then tested rapamycin, allantoin, trichostatin A, LY-294002 and geldanamycin in Caenorhabditis elegans. An increase in lifespan and healthspan was observed for all drugs except geldanamycin when fed to wild-type worms, but no lifespan effects were observed in eat-2 mutant worms, a genetic model of CR, suggesting that life-extending effects may be acting via CR-related mechanisms. We also treated daf-16 worms with rapamycin, allantoin or trichostatin A, and a lifespan extension was observed, suggesting that these drugs act via DAF-16-independent mechanisms, as would be expected from CR mimetics. Supporting this idea, an analysis of predictive targets of the drugs extending lifespan indicates various genes within CR and longevity networks. We also assessed the transcriptional profile of worms treated with either rapamycin or allantoin and found that both drugs use several specific pathways that do not overlap, indicating different modes of action for each compound. The current work validates the capabilities of this bioinformatic drug repositioning method in the context of longevity and reveals new putative CR mimetics that warrant further studies.

Keywords: Caenorhabditis elegans; aging; drug repositioning; lifespan; longevity; pharmacogenomics.

Figure 1

Percentage survival of wild‐type (N2) or eat‐2 mutant worms alone or treated with allantoin (A), trichostatin A (B), rapamycin (C) or LY‐294002 (D). Worms that did not respond to touch stimulation were recorded as dead. Both eat‐2 and treated worms of both genotypes showed both a significantly longer lifespan than N2 controls and no significant variation from each other (Table 3). (A) Allantoin‐treated N2 had a 21.9% increase in lifespan. (B) Trichostatin A‐treated N2 had a 22.1% increase in lifespan. (C) Rapamycin‐treated N2 had an 18.9% increase in lifespan. (D) LY‐294002‐treated N2 had a 21.6% increase in lifespan. Results are shown for combined data from two representative trials (full range of results is shown in Table S3).

Figure 2

Percentage survival of daf‐16 worms alone or treated with rapamycin, allantoin or trichostatin A. A significant increase in lifespan was observed in response to treatment with rapamycin (21.7% increase), allantoin (19.7% increase) or trichostatin A (23.8% increase) compared to untreated daf‐16 worms.

Figure 3

The pharyngeal pumping rate of either N2 control worms or those treated with allantoin (A), trichostatin A (B), rapamycin (C) or LY‐294002 (D). The pumping rate was recorded on days 1, 5, 10 and 15 post‐L4 moult when treatment with the drug began. The treated worms showed a slower decline of pharyngeal pumping compared to the untreated controls. The rates of pumping in allantoin‐ (83.5 pumps per minute), trichostatin A‐ (46.7 pumps per minute), rapamycin‐ (45.3 pumps per minute) and LY‐294002 (33.88 pumps per minute)‐treated worms are greater than those of untreated worms (25.3 pumps per minute) to a statistically significant degree at day 10 (two‐tailed t‐test): allantoin (P < 0.001, N = 10), trichostatin A (P < 0.05, N = 10), rapamycin (P < 0.05, N = 10) and LY‐294002 (P < 0.01, N = 10). *Significance of at least 0.05. Error bars indicate ±1 standard error.

Figure 4

The movement rate of either N2 control worms or those treated with geldanamycin. The movement rate was recorded on days 1, 5, 10 and 15 post‐L4 moult when treatment with the drug began. The geldanamycin‐treated worms show a much faster initial decline in movement compared to the untreated controls before levelling out after day 5 with no decrease between days 5 and 10. Movement rate decreases faster in geldanamycin‐treated worms with the rate of movement at days 5, 10 and 15 for treated worms (2.8, 2.7 and 0 body bends per minute, respectively) being lower than the rate in untreated controls (10.9, 7.7 and 3.7 body bends per minute, respectively), to a significant degree (two‐tailed t‐test) at day 5: (P < 0.001); day 10: (P < 0.001); day 15: (P < 0.05). *Significance of at least 0.05; **significance of at least 0.005. Error bars indicate ±1 standard error.

Figure 5

Visual representation of allantoin, LY‐294002, rapamycin and trichostatin A and their gene/protein targets. Included is the mapping of longevity‐associated genes, CR‐mediating genes and genes with expression changes under CR.

Figure 6

The worm longevity network. The network (which includes longevity‐associated genes and their interaction protein partners) contains many potential links to drug targets through which allantoin, LY‐294002, rapamycin and trichostatin A might have an effect on its functionality. For an integrative visualization, all drug targets have been displayed (including the targets that do not belong to the network).