The intraflagellar transport in sensory cilia deteriorates with ageing
To explore the effect of ageing on sensory perception, we first examined the response to food in young (day 1 of adulthood, D1) and aged worms (day 10 of adulthood, D10)15. Consistent with the age-related decline of chemotaxis6,16, the food of bacteria no longer affected the movement of aged worms whereas young adults exhibited a clear enhanced slowing response (ESR) to bacteria (Fig. 1a and Supplementary Fig. 1a), indicating a defect of sensory perception with ageing. Cilia defects suppress dye-filling in sensory neurons17. Subsequent dye-filling assay indicated that the staining of DiI in the soma of sensory neurons became remarkably weaker at D10 (Fig. 1b), showing that ageing causes defects in the sensory cilia.
Despite the remarkable deterioration in cilia function, it has been shown that the microtubule organization in sensory cilia does not suffer any obvious changes in D10 worms5, implying that the ciliary structure is unaffected by ageing at D10. We then examined IFT, which is crucial to cilia function, during ageing. The anterograde IFT is driven by kinesin and IFT-B complex, whereas the retrograde IFT by dynein and IFT-A complex (Fig. 1c). The motors of kinesin (OSM-3/KIF17), dynein (CHE-3/DHC2), and the core components of IFT-B (OSM-6/IFT52) and IFT-A (CHE-11/IFT140) complexes were endogenously GFP-tagged using genomic editing to visualize IFT in vivo. The length of GFP signal along cilia from CHE-3::GFP and CHE-11::GFP exhibited a slight increase in aged worms (Supplementary Fig. 1b and 1c). We next examined IFT using live imaging. Indeed, both the frequencies and velocities of these IFT components were remarkably decreased in the sensory cilia of aged worms (Fig. 1d to 1f, Supplementary Movie 1). As the rate of ageing is highly variable among individuals18, the decrease of IFT also exhibited a wide variation in aged worms (Supplementary Fig. 1d). daf-2(-) is a well-established longevity mutant with defective IIS 7. Whereas IFT was impaired in D10 WT worms (Fig. 1), it was well-maintained in daf-2(-) mutants at D10 and D20 (Supplementary Fig. 2a to 2c, Supplementary Movie 2). Therefore, IFT is impaired with ageing but protected in the longevity mutant of daf-2(-).
DAF-19 is critical in the ageing-induced decline of IFT and sensory perception
The proper assembly of the multi-component IFT complexes is required for IFT. Inhibiting the expression of IFT complex components is known to suppress IFT10,19. Therefore, to explore the mechanism underlying the decreasing IFT with ageing, we examined the expression of IFT components in WT worms at D1 and D10. In the soma of amphid and phasmid sensory neurons, endogenously GFP-tagged OSM-3 and OSM-6, but not CHE-11 or CHE-3, exhibited a remarkable decrease with ageing in the WT worms. In phasmid sensory cilia, all four examined IFT components were reduced during ageing (Fig. 2a). DAF-19, an RFX transcription factor, is a master regulator of IFT genes in C. elegans20. Consistent with the decrease of IFT components with ageing, endogenously GFP-tagged DAF-19 was downregulated in the neurons of aged WT worms (Fig. 2b), implying that daf-19 is critical in maintaining IFT during ageing. Indeed, a mild daf-19 RNAi did not disrupt ciliogenesis but abolished the elevated motilities of OSM-3::GFP and OSM-6::GFP in daf-2(-) mutants at D10 to WT levels and substantially downregulated the velocity of CHE-3::GFP and CHE-11::GFP in daf-2(-) mutants (Supplementary Fig. 2d-2f). Mild overexpression of daf-19c, a daf-19 isoform which specifically regulates ciliary genes, with its native promoter effectively ameliorated the decrease of DAF-19 in aged WT worms 20,21 (Fig. 2c). As a result, overexpressing daf-19c increased OSM-6::GFP, CHE-11::GFP, and CHE-3::GFP in young adult worms and upregulated all the examined IFT proteins in aged worms (Fig. 2d). The decrease of OSM-3::GFP and OSM-6::GFP with ageing was also suppressed by daf-19c overexpression (Fig. 2d). Following the increase of IFT proteins, both the velocity and frequency of the four examined IFT components were improved at D10 when daf-19c was upregulated. In young worms at D1, overexpressing daf-19c also increased the frequencies of CHE-3::GFP and CHE-11::GFP and the velocities of OSM-6::GFP and CHE-11::GFP (Fig. 3a to 3c, Supplementary Movie 3). Consistently, overexpressing daf-19c suppressed the diminishing DiI staining in sensory neurons with ageing (Fig. 3d). Therefore, a DAF-19-regulated decrease of IFT components underlies the age-related decline of IFT and in turn impairs the function of sensory cilia in aged worms.
We next performed chemotaxis assay to check whether the DAF-19-IFT-sensory cilia axis is also responsible for the deteriorating sensory perception with ageing. Because the worms at D10 suffer a severe decrease of motility and are not suitable for chemotaxis assay, worms at D5, which are at the end of their reproductive period, were examined. As expected, overexpressing daf-19c improved the chemotaxis to butanone at D5 (Fig. 3e, Supplementary Movie 4). Consistently, the response to food at D10 was also improved by upregulating daf-19c (Fig. 3f). Taken together, these results indicate that overexpressing daf-19c effectively suppresses the ageing-induced degeneration of sensory cilia and perception.
Food perception through sensory cilia activates AMPK signalling
Sensory perception, which requires proper IFT in sensory cilia, is tightly related to metabolism1,10. Since daf-19c controlled IFT underlies the age-related decline of sensory perception, we next examined its impact on metabolism. Metabolism is composed of the biosynthetic anabolism and the energy-yielding catabolism. In C. elegans, IFT is known to modulate insulin/IGF-1 signalling (IIS), a critical signalling pathway in anabolism3,14. IFT mutants of osm-3(-) and osm-6(-) exhibited severe cilia defects and IIS target genes were regulated as reported at D113,14 (Supplementary Fig. 3a). In aged worms, IIS exhibited a decreased modulation by sensory perception, as multiple IIS target genes were no longer changed upon mutating osm-3 or osm-6 at D10 (Supplementary Fig. 3a). Interestingly, overexpressing daf-19c failed to change IIS target genes expression in either young or aged worms (Supplementary Fig. 3b), implying that the enhanced sensory perception may not interfere with IIS.
We next examined the effect of sensory perception on AMPK, a pivotal driver of catabolism22. We first used western blot to check the level of activated AMPK (phosphorylated at the conserved Thr172, p-AMPK) in the whole worm. Overexpressing daf-19c in sensory neurons with its native promoter or in pan neurons with a neuron-specific promoter remarkably increased p-AMPK levels of both young and aged worms20,21 (Fig. 4a and 4b). To examine whether this could be due to any side effects in larval development, we prepared another strain to overexpress daf-19c::degron::gfp by its native promoter. The overexpression of daf-19c::degron::gfp was suppressed by auxin treatment during larval stages and induced specifically in adulthood by removing auxin23 (Supplementary Fig. 4a and 4b). Whereas inhibiting daf-19c upregulation in larvae effectively blocked the increase of p-AMPK in this strain at D1, the adulthood specific overexpression of daf-19c still enhanced p-AMPK level at D10 (Supplementary Fig. 4c). Therefore, it is unlikely that the daf-19c-induced p-AMPK is due to secondary effects in larval development. As western blot shows p-AMPK levels from the whole body, these results also imply that daf-19c could non-autonomously induce AMPK signalling in other tissues. Activated AMPK phosphorylates the CREB regulated transcriptional coactivator (CRTC-1) in C. elegans and induces its cytosolic translocation24. For further confirmation, we then examined the nuclear localization of CRTC-1::RFP in the intestine. As expected, overexpressing daf-19c reduced the nuclear localization of CRTC-1::RFP in the intestine (Fig. 4c), confirming that daf-19c in sensory neurons activates AMPK in other tissues.
To further test whether the daf-19c-induced AMPK activity is due to the enhanced sensory perception or other effects by daf-19c upregulation, worms were incubated in the plates with food, with food odour (food on the lid), or without food and examined for p-AMPK levels. The daf-19c-induced AMPK activation occurred only when worms sensed the food or food odour (Fig. 4a), indicating that it does require sensory perception, especially olfactory perception of food. Moreover, disrupting IFT and sensory cilia via mutating osm-3 fully abolished the elevated p-AMPK levels in both young and aged worms overexpressing daf-19c (Fig. 4d and Supplementary Fig. 4d). Therefore, daf-19c activates AMPK by enhancing IFT in sensory cilia and improving food perception. Consistently, when sensory cilia are disrupted by the mutation of osm-3(-) or osm-6(-), p-AMPK levels decreased at D1 and D10 (Fig. 4d and Fig. 4e), confirming that sensory perception promotes AMPK activity. As daf-19 controls innate immunity25, worms were further incubated on UV-killed bacteria to minimize immunity response and examined for p-AMPK levels. daf-19c overexpression still robustly increased p-AMPK in worms fed with UV-killed bacteria (Supplementary Fig. 4f), indicating that this effect is independent of innate immunity.
Primary cilia in cultured cells modulate AMPK via LKB1, a kinase phosphorylating AMPK26. By tagging the worm ortholog of LKB1 (PAR-4) with GFP using genomic editing, we found that it was expressed in sensory neurons (Fig. 5a). RNAi against par-4 specifically in sensory neurons or in all neurons blocked the increase of p-AMPK in worms overexpressing daf-19c (Fig. 5b and 5c), indicating that sensory cilia control AMPK activity via par-4. We further pursued the molecule transducing AMPK signalling non-autonomously from sensory neuron. Activated AMPK in neurons affects other tissues through octopamine27. To examine whether octopamine is required in the sensory perception-induced AMPK signalling, two octopamine biosynthetic enzymes (tbh-1 and tdc-1) were mutated 27. Indeed, mutating either of them blocked the increase of p-AMPK in worms overexpressing daf-19c (Fig. 5d and 5e). Supplementing the mutants of tbh-1(-) and tdc-1(-) with octopamine rescued the elevated p-AMPK induced by daf-19c overexpression (Fig. 5e). Therefore, octopamine is involved in the sensory perception-induced AMPK signalling.
Sensory perception upregulates AMPK activity for metabolic homeostasis and longevity
Activated AMPK is a positive regulator of longevity, driving critical catabolic processes including autophagy7. Since enhanced sensory perception activates AMPK (Fig. 4 and 5), we next explored whether it also promotes autophagy in the intestine using a mCherry-GFP-tagged LGG-1 reporter. GFP in this reporter is specifically quenched in autolysosomes (ALs), thus labelling autophagosomes (APs) with both mCherry and GFP and ALs with mCherry12. As expected, upregulating daf-19c in sensory neurons remarkably increased the number of ALs whereas mildly reduced APs in the intestine (Fig. 6a). Either blocking or enhancing autophagy flux could change the numbers of APs and ALs. Chloroquine blocks autophagy flux and should not regulate APs and ALs when autophagy is already blocked12. Chloroquine suppressed the change of APs and ALs upon daf-19c overexpression (Fig. 6a), indicating that autophagy is active in the worms overexpressing daf-19c and enhanced sensory perception promotes autophagy in the intestine.
Metabolism is closely related to ageing. We next examined a series of hallmarks of healthy ageing in worms overexpressing daf-19c. Overexpressing daf-19c improved chemotaxis, the enhanced slowing response (ESR) to food, and motility in aged worms (Fig. 3e, 3f and 6b), which are key health metrics15. Interfering AMPK activity by neuron-specific RNAi against either par-4/LKB1 or aak-2/AMPK abrogated the enhanced motility in the worms overexpressing daf-19c at D10 (Supplementary Fig. 5a), indicating that sensory perception promotes healthy ageing through AMPK signalling. The motility of worms is closely related to the integrity of myofibers in the body wall muscle (BWM), which is prone to ageing18,28. Sensory neuron specific overexpression of daf-19c reduced myofilaments abnormalities in aged worms, whereas a neuron-specific RNAi against daf-19 had an adverse effect (Supplementary Fig. 5b and Supplementary Fig. 5c). AMPK and autophagy are critical in maintaining the balance of protein metabolism and in turn promote health and longevity7. The accumulation of polyglutamine (polyQ) aggregates is a marker of the deteriorating protein homeostasis12. Consistently, overexpressing daf-19c reduced polyQ aggregates (Fig. 6c), indicating improved protein homeostasis with the enhanced sensory perception. Taken together, sensory perception promotes healthy ageing through AMPK.
We next checked the potential effect of sensory perception on lifespan. With ageing, the motilities of IFT components decrease at variable rates among worms (Supplementary Fig. 1d). We then examined the lifespan of worms with remarkably different velocities of IFT components (i.e., OSM-3, OSM-6, and CHE-11) at D10. The worms with faster IFT lived longer than the worms with slower IFT (Fig. 6d and Supplementary Fig. 5d; Supplementary Table 3), showing a positive correlation between IFT function and longevity. We further examined the lifespan of the worms with improved sensory perception. Indeed, two strains overexpressing daf-19c with its native promoter, and another strain overexpressing daf-19c with a neuron-specific promoter all exhibited extended lifespans (Fig. 6e and Supplementary Fig. 5e; Supplementary Table 3). Disrupting sensory cilia by an adult-specific RNAi against osm-3 abrogated the extended lifespan of the worms overexpressing daf-19c (Fig. 6f; Supplementary Table 3), indicating that daf-19c promotes longevity by enhancing sensory perception. Since sensory perception activates AMPK to improve worm motility (Supplementary Fig. 5a), neuron-specific RNAi against par-4/LKB1 or aak-2/AMPK was performed in adult worms to test whether the sensory perception-induced longevity also requires AMPK. Indeed, the lifespan of worms overexpressing daf-19c is reduced to the same level as the WT worms upon either of the two RNAi treatments (Fig. 6g; Supplementary Table 3). Therefore, sensory perception promotes longevity via upregulating AMPK signalling.