Autophagy activator AA-20 improves proteostasis and extends Caenorhabditis elegans lifespan

Significance

Autophagy, the process by which cells break down and recycle cellular components in the autolysosome, is vital for health and longevity, but declines with aging, contributing to aging-related diseases. In this study, we introduce AA-20, a promising small-molecule autophagy enhancer. AA-20 promotes cellular cleanup by reducing neutral lipid and protein aggregate accumulation in human cells and the model organism Caenorhabditis elegans, improving fitness and extending lifespan in nematodes. Notably, AA-20 enhances autophagy without inhibiting the mammalian target of rapamycin complex 1 (mTORC1) pathway, positioning it as a unique candidate for intervention in aging-related disorders. Our findings suggest AA-20 could hold therapeutic potential for age-related conditions involving protein aggregation and lipid storage issues, offering a promising avenue for mediating healthy aging.

Abstract

The degradation of cellular components through autophagy is essential for longevity and healthy aging. However, autophagy function decreases with aging, contributing to age-related diseases. In this study, we characterized a small-molecule activator of autophagy called AA-20 that enhances autophagy and lipid droplet clearance in human cells and in the nematode Caenorhabditis elegans. AA-20 reduces polyglutamine aggregation in an autophagy-dependent manner in both human cells and C. elegans, where it also promotes fitness. Consistently, we found that AA-20 extends lifespan in WT C. elegans, but not in autophagy-deficient mutants. Interestingly, our findings suggest that AA-20 acts, at least in part, through a mechanism involving the transcription factor EB, but without inhibiting the protein kinase mammalian target of rapamycin complex 1. Collectively, our results identify an autophagy activator AA-20, which may have potential therapeutic implications for aging-related proteinopathies and lipid storage disorders.

Autophagy is a crucial process to maintain organismal homeostasis by which intracellular and extracellular components are delivered to lysosomes for degradation and recycling of molecular components (1). While accumulation of protein aggregates has been recognized to be a common characteristic of neurodegenerative diseases such as Parkinson’s and Huntington’s (2), excessive lipid accumulation is emerging as another prevalent feature of these maladies (3). Autophagy, which degrades these cellular substrates, is a key degradation mechanism to alleviate such pathological conditions (3). In addition to the role that dysregulated lipid homeostasis plays in neurodegeneration, lipid overload has far-reaching consequences for human health, contributing to conditions like obesity, fatty liver disease, and cancer (3). Thus, interventions that enhance lipid droplet clearance and normalize lipid abnormalities represent an attractive therapeutic strategy.

Multiple lines of evidence indicate that impaired autophagy is a hallmark of unhealthy aging across species (4). In model organisms like the nematode Caenorhabditis elegans, conserved longevity pathways and paradigms, including dietary restriction, reduced insulin/insulin-like growth factor 1 signaling, and reduced mammalian target of rapamycin (mTOR) complex 1 (mTORC1) enzymatic signaling, extend lifespan and healthspan, at least in part via autophagy induction (5). Consequently, enhancing autophagy pharmacologically may improve organismal fitness and help ameliorate age-related disorders.

Numerous small molecule autophagy inducers have been reported (2, 6). However, only a handful of these pharmacological agents have undergone direct testing in model organisms to assess their effects on longevity. Among the autophagy-promoting agents investigated for mediating longevity effects, the most extensively studied to date are rapamycin (7), metformin (8), resveratrol (9), spermidine (10), and urolithin A (11). These autophagy activators, either directly or indirectly, stimulate autophagy by inhibiting the mTORC1 complex, which contains the mTOR kinase (12–16), a key autophagy regulator and nutrient sensor in the cell (17). mTORC1 negatively regulates autophagy, and its inhibition (e.g., during nutrient deprivation or rapamycin treatment) activates UNC-51-like autophagy activating kinase (ULK1), leading to the autophagy activation (18). Additionally, mTORC1 inhibition mediates lysosome biogenesis by directing transcription factor EB (TFEB), a major transcription factor regulating autophagy-related and lysosomal genes, to the nucleus (19). While mTORC1 inhibition activates the TFEB transcriptional program (20), TFEB function can also be modulated by other mechanisms, such as calcium signaling via lysosomal calcium channels or by the serine/threonine Protein Kinase B (21).

Notably, the mTOR kinase also governs crucial cellular processes such as protein synthesis and metabolic pathway transcription (22), as well as complex processes on the organismal level, including regulation of the immune system (23). This renders mTOR inhibitor-based autophagy activators poorly selective (24, 25). Consequently, there is an interest in identifying pharmacological interventions that operate independently of mTORC1 inhibition and that can exclusively activate autophagy. Investigating such alternatives offers the possibility to avoid on-pathway immunosuppression by mTORC1 inhibition, or at minimum, complement mTOR-dependent autophagy-promoting agents for treating age-related degenerative diseases.

Recently, we provided a detailed account (26) of how we phenotypically screened a library of one million small molecule autophagy activator candidates to identify compounds that hastened lipid droplet clearance in a noncytotoxic fashion. A subset of these small molecules, including AA-20, enhanced autophagy without affecting phosphorylation of S6 kinase, a known substrate of mTORC1. In this study, we demonstrate that AA-20 activated autophagy in both human cells and C. elegans. In C. elegans, AA-20 reduced neutral lipid levels and protein aggregation in an autophagy gene (atg-3)–dependent manner, consistent with clearance of multiple autophagy substrates. We also found that AA-20 extended both healthspan and lifespan in C. elegans. Since most AA-20–induced phenotypes were dependent on autophagy, we further investigated AA-20’s mechanism of action as an autophagy activator. Interestingly, we found that HLH-30/TFEB, a known transcriptional regulator of autophagy, was required for all AA-20-induced phenotypes we tested, i.e., increased lysosomal acidification, improved lipid clearance, enhanced protein aggregate clearance, and lifespan extension. Interestingly, AA-20 does not seem to function through a discernible TFEB transcriptional program, nor does AA-20 alter the phosphorylation of several known mTORC1 substrates, suggesting AA-20 operates through a mechanism independent of mTORC1 inhibition. Collectively, our results highlight AA-20 as a potential therapeutic candidate for combating age-related diseases featuring dysregulated autophagy.

Results

AA-20 Promotes Autophagy in Mammalian Cells.

The compound AA-20 (Fig. 1A) comprises a 1,3,5-trisubstituted triazine core (colored black) with two 4-methoxyaniline substructures (colored blue) and one benzothiazole substituent (colored red) attached to the triazine sp²-hybridized carbons. To validate that AA-20, discovered in our recently reported high-throughput screen (HTS) (26) (SI Appendix, Fig. S1A summarizes the screening strategy) promoted autophagy, we resynthesized and fully characterized AA-20 (see SI Appendix for details), and used this material for all of the experiments in this study. To determine the stability of AA-20 in our experiments, which were conducted at temperatures ranging from 20 °C to 37 °C, we tested its stability in a PBS buffer solution (pH 7.4) at 25 °C for up to 12 d using liquid chromatography-mass spectrometry. The results showed that AA-20 did not degrade within the test period (SI Appendix, Fig. S1B), demonstrating that it is a relatively stable compound.

Fig. 1.

AA-20 promotes autophagy in mammalian cells. (A) Chemical structure of compound AA-20. Black: 1,3,5-trisubstituted triazine core, blue: 4-methoxyaniline substituents, red: benzothiazole substituent. (B) Representative GFP:RFP ratio plot depicting the impact of AA-20 on general autophagy in RPE-1 cells expressing GFP-LC3-RFP-LC3ΔG. Cells were treated with AA-20 (400 nM to 10 µM) and imaged every 2 h for 72 h. The green/red ratio is normalized at each time point to DMSO (AA-20, n = 3; DMSO, n = 6; refer to Dataset S1 for all data). Error bars indicate SEM. Statistical significance denoted as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, by two-way ANOVA with Tukey’s multiple comparison test. (C) Confocal micrographs of HeLa cells immunofluorescence stained with anti-LC3B (green) and DAPI (blue). (Scale bar, 10 µm.) White arrowheads indicate LC3B-positive autophagosomes. (D) Bar graph of LC3B-positive puncta per cell quantified using Imaris software’s Spot Detection function (total ~800 cells analyzed, N = 25 images per condition, three biological replicates). HeLa cells were treated with 10 µM AA-20 or DMSO for 20 h, with bafilomycin A1 (BafA) added at the 18th h to block autophagy at the lysosomal acidification step. (E) Confocal micrographs of HeLa cells incubated with Lysotracker Red and costained with DAPI. Cells were treated with or without 10 µM AA-20 and BafA (250 nM; Scale bar, 20 µm). White arrowheads indicate LysoTracker Red-positive lysosomes. (F) Quantification of average LysoTracker intensity per cell using Imaris software. (G) Further quantification of the average number of LysoTracker Red-positive puncta per cell performed using the Spot Detection function in Imaris software. Analysis included ~250 cells across 25 images per condition with three biological replicates. Error bars indicate SEM. Statistical significance denoted as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, by two-way ANOVA with Tukey’s multiple comparison test. (H) Cathepsin D activity assay in HeLa cells treated with 0.1% DMSO or 10 µM AA-20 for 20 h. Three biological experiments were performed, each with two replicates (error bars represent SEM). *P < 0.05, ***P < 0.001, by one-way ANOVA with Tukey’s multiple comparison test. (I) Relative activity of lysosomal acid lipase (LAL) measured in human fibroblast PCS201-012 primary cells treated with DMSO or 10 µM AA-20 for 72 h (n = 11 to 12, from three independent repeats, error bars represent SEM). ****P < 0.0001, by the unpaired t-test. (J) Relative activity of glucocerebrosidase (GCase) measured in human fibroblast PCS201-012 primary cells treated with AA-20 at 20 µM for 72 h (n = 6, from two independent repeats, error bars represent SEM). n.s. > 0.05, ****P < 0.0001, by ordinary one-way ANOVA.

To monitor changes in autophagy, we utilized a fluorescently labeled LC3/ATG8 autophagy effector protein. LC3/ATG8 is cleaved by the protease ATG4 at its C-terminus, thus facilitating LC3/Atg8 lipidation mediating its incorporation into the autophagosomal membranes, a process that can be visualized by the appearance of fluorescent puncta. We first used retinal pigment epithelial-1 (RPE-1) cells stably expressing GFP-LC3B-RFP-LC3BΔG, a LC3 effector based autophagy reporter commonly used in the autophagy field (27) to measure autophagic activation by ratiometric fluorescent changes. Following cleavage between GFP-LC3B and RFP-LC3BΔG by the protease ATG4, GFP-LC3B is lipidated and incorporated into autophagosomes to mediate autophagy–GFP fluorescence in this ATG4-generated fragment, which is quenched upon fusion with acidic lysosomes. In contrast, RFP-LC3BΔG lacks the LC3B C-terminal glycine residue that facilitates phosphatidylethanolamine attachment (lipidation) that allows LC3B incorporation into the autophagosome membrane. RFP-LC3BΔG remains fluorescent in the cytoplasm longer as an internal control. Thus, a decrease in GFP/RFP ratio indicates an increase in autophagy. When we exposed these reporter cells to AA-20 at doses ranging from 400 nM to 10 µM (26), AA-20 reduced the GFP/RFP ratio in comparison to the dimethylsulfoxide (DMSO) or vehicle control over the 72-h time course, particularly at the 3 µM and 10 µM concentrations (Fig. 1B and Dataset S1). A time-dependent increase in the GFP/RFP ratio was observed under all conditions (Fig. 1B), indicating that basal autophagy decreases as cells approach confluency. This aligns with prior observations in nontumorigenic cells, where autophagosome formation diminishes at high cell density (28). There was no evidence of AA-20 cytotoxicity at concentrations ≤10 µM under these conditions, as indicated by normal growth rates in RPE-1 cells (SI Appendix, Fig. S1C). This suggests that AA-20 at the doses used was nontoxic, and the induced autophagy phenotype was not due to cytotoxic effects (29).

To further explore how AA-20 influences autophagy, we conducted a series of alternative assays in HeLa cells, a human cervical cancer cell line, focusing on endogenous LC3B levels. We employed immunofluorescence to count LC3B-positive puncta (Fig. 1 C and D) and immunoblotting to assess LC3B-II levels (SI Appendix, Fig. S1 D and E). These experiments were performed either with or without the lysosomal inhibitor bafilomycin A1 (BafA), which prevents lysosomal acidification (30) and thus blocks autophagic flux. In the presence of BafA, compounds that activate autophagy lead to a further increase in LC3B levels within deacidified lysosomes relative to the same experiment lacking BafA treatment. Autophagy inhibitors on the other hand can increase LC3B levels, but autophagy inhibitors do not further increase LC3B levels when combined with BafA, as both block autophagic flux (30). AA-20 (10 µM) significantly elevated the steady-state number of LC3B-positive puncta (Fig. 1 C and D) and LC3B-II levels (SI Appendix, Fig. S1 D and E) at the 20 h mark. Adding BafA to AA-20-treated cells during the final 2 h of AA-20 incubation further increased LC3B-positive puncta and LC3B-II levels ~10% compared to AA-20, although it did not reach statistical significance (Fig. 1 C and D and SI Appendix, Fig. S1 D and E). Compared to BafA alone, cotreatment with BafA and AA-20 significantly increased both LC3B-positive puncta counts and LC3B-II levels (Fig. 1 C and D and SI Appendix, Fig. S1 D and E), indicating autophagy activation and not inhibition. Notably, LC3B transcript levels remained unchanged (SI Appendix, Fig. S1F), indicating that AA-20’s effects are not driven by discernible changes in gene expression. The CellTiter-Glo viability assay showed cell viability exceeding 90% in HeLa cells across treatment concentrations (SI Appendix, Fig. S1G), indicating that the observed effects were not due to cytotoxicity.

To further characterize the effects of AA-20 on autophagy, we employed LysoTracker Red to measure the acidification of lysosomes. We observed that AA-20 (10 µM) increased both the intensity of LysoTracker fluorescence (Fig. 1 E and F) and number of LysoTracker-positive puncta in HeLa cells (Fig. 1 E and G). This effect was notably attenuated when cells were cotreated with BafA (Fig. 1 E and G), which suggests that AA-20 enhances both the acidification and the number of lysosomes. To confirm that AA-20 can expand the lysosomal pool, we stained cells for LAMP2 and found a significant increase in LAMP2-positive puncta following AA-20 treatment (SI Appendix, Fig. S1 H and I). Consistently, AA-20 treatment resulted in a significant increase in the activity of several lysosomal enzymes, as determined by a cathepsin D activity assay (Fig. 1H), a lysosomal acid lipase (LAL) activity assay (31) (Fig. 1I), and a glucocerebrosidase (GCase) activity assay (32) (Fig. 1J). Collectively, these results suggest that AA-20 is an autophagy activator that increases acidified lysosome formation and enhances lysosomal hydrolase activities through more acidified lysosomes being produced upon AA-20 treatment.

AA-20 Enhances Autophagy in C. elegans.

The process of autophagy is highly conserved between C. elegans, a short-lived nematode, and humans (33). Therefore, we also investigated the effect of AA-20 treatment on autophagy activation in C. elegans. To achieve this, we quantified autophagosome numbers using previously published neuronal or endogenously expressed GFP::LGG-1 (ortholog of GABARAP, which like LC3B is an Atg8 superfamily member) reporters (34–36) in nerve-ring neurons (here referred to as neurons), in the intestine, and in the body-wall muscle of adult animals. To enhance AA-20 absorption in C. elegans, which naturally feed on bacteria, we utilized a 96-well microtiter plate liquid protocol (37) and substituted live bacteria with dead bacteria as the animals’ diet (38). We first quantified GFP::LGG-1/Atg8 puncta in animals fed with 0.4% DMSO, the highest dose of solvent (vehicle) used for administering AA-20 in our assays, to rule out any possible physiological effects of the vehicle (37, 39, 40) on autophagy function. No significant changes were observed (SI Appendix, Fig. S2A), indicating that this amount and duration of DMSO exposure in C. elegans is unlikely to affect autophagy.

To determine the concentration range of AA-20 useful for autophagy induction in C. elegans, we quantified GFP::LGG-1/Atg8 puncta in age-synchronous 4-d-old adults. The onset of adulthood is defined as the first day an animal is reproductive (“day 1”); the reproductive period lasts for ~5 d, and the mean lifespan is ~20 d at 20 °C. Animals were fed with AA-20 at concentrations ranging from 1 to 40 µM from day 1 of adulthood (single treatment with no media change over 4 d). We observed that AA-20 increased the number of autophagic puncta in nerve-ring neurons at concentrations ≥5 µM (SI Appendix, Fig. S2B). Since 5 µM of AA-20 was the minimum effective dose to elicit a significant autophagy response in neurons (SI Appendix, Fig. S2B), unless otherwise noted, we used this concentration in all subsequent assays in day 1 adult C. elegans. Autophagy flux assays in which animals were soaked in 100 µM BafA for the last 2 h of AA-20 incubation before imaging showed increased GFP::LGG-1/Atg8 puncta numbers over AA-20 treatment alone in neurons (Fig. 2 A and B). We also investigated the effects of AA-20 on autophagy in the intestine (Fig. 2 C and D) and in body-wall muscle (Fig. 2 E and F) and found that it consistently increased autophagic puncta numbers in these tissues, and that these numbers were further increased by BafA cotreatment. These data strongly support the hypothesis that AA-20 induced autophagy. Finally, we stained day 4 adult wild-type (WT) animals with LysoTracker Red fed with AA-20 from day 1 of adulthood and found that AA-20 treatment increased intestinal LysoTracker fluorescence levels in the animals compared to control at concentrations ≥5 µM (5 µM data shown in Fig. 2 G and H, concentration-dependent data depicted in SI Appendix, Fig. S2C). Similar to the effect of AA-20 in human cells (Fig. 1H), AA-20 increased the activity of a cathepsin D employing a fluorometric assay in C. elegans (Fig. 2I). Taken together, these findings suggest that AA-20 promotes autophagy in C. elegans, like in human cells.

Fig. 2.

AA-20 enhances autophagy in C. elegans. Autophagy flux was measured in day 4 adult animals expressing rgef-1p::gfp::lgg-1 (A and B) or lgg-1p::gfp::lgg-1 (C−F), treated with vehicle (DMSO) or 5 µM AA-20 from day 1 of adulthood in liquid at 20 °C. Animals were soaked with DMSO or BafA for 2 h before imaging. GFP::LGG-1/Atg8-positive puncta were quantified from three independent experiments in nerve-ring neurons (A and B, N = 18 to 21 animals), proximal intestinal cells (C and D, N = 18 to 28 cells), and body-wall muscle (E and F, N = 15 to 17 animals). (Scale bar, 10 µm.) Error bars represent SEM. **P < 0.01, ***P < 0.001, ****P < 0.0001, by one-way ANOVA with Tukey’s multiple comparison test. White arrowheads indicate GFP::LGG-1-positive autophagosomes. (G and H) Representative images of LysoTracker red positive animals and intensity analysis in day 4 adult WT animals fed with DMSO or 5 µM AA-20 on day 1 of adulthood from three independent repeats (N = 30 to 31 animals). (Scale bar, 500 µm.) (I) Relative activity of cathepsin D measured in day 4 adult WT animals treated with DMSO or 5 µM AA-20 (n = 4, error bars indicate SEM).

AA-20 Reduces Lipid Levels in an Autophagy-Dependent Manner.

In our previous HTS (26), we focused on lipid droplet clearance, as lipid droplets can be degraded through autophagy activation (41, 42). In that HTS, we pretreated cells with a 10 µM concentration of oleic acid to increase cellular lipid droplet levels (24 h). We found that the total lipid droplet area in SW1990 human adenocarcinoma cells preincubated with oleic acid and then treated with AA-20 as a function of concentration (0 to 25 µM) was significantly and dose-dependently reduced with an EC₅₀ of ≈8 µM (SI Appendix, Fig. S3A). To address whether this lipid level reduction depended on lysosomal function, we analyzed neutral lipid levels in SW1990 cells cotreated with BafA. We employed an alternative LipidTOX neutral lipid staining method combined with flow cytometry to measure the mean LipidTOX fluorescence intensity, which indicates lipid levels in the cells. Our results indicated that AA-20 (10 µM) significantly decreased neutral lipid levels, while BafA treatment alone increased lipid levels compared to the DMSO control (SI Appendix, Fig. S3B). When AA-20 was cotreated with BafA, the lipid reduction effect of AA-20 was completely blocked (SI Appendix, Fig. S3B). To assess the neutral lipid substrate autophagic function, we evaluated the ratio of lipid levels (BafA-treated/BafA untreated conditions) for both DMSO and AA-20, as well as the difference in lipid levels between these conditions, respectively. Our analysis revealed that AA-20 significantly increased neutral lipid substrate turnover by autophagy (SI Appendix, Fig. S3 C and D). Collectively, these findings indicate that AA-20 reduces lipid levels by enhancing lipid droplet degradation through autophagy.

Given our observation of conserved autophagy induction by AA-20 in C. elegans, as well as the previously demonstrated degradation of lipid through autophagy in C. elegans (43), we assessed whether the lipid droplet clearance phenotype observed in human cells was conserved in C. elegans. We used Oil Red O (ORO) neutral lipid staining to fluorescently label lipid droplets (44), as well as a C. elegans reporter expressing DHS-3::GFP (a conserved short-chain dehydrogenase lipid droplet binding protein) (45) for imaging analyses. AA-20 treatment significantly reduced both total ORO and DHS-3::GFP intensity levels in day 5 adult WT animals (Fig. 3 A–D), indicating a reduction in lipid levels, as observed in human cells (26) (SI Appendix, Fig. S3A).

Fig. 3.

AA-20 reduces neutral lipid levels in an autophagy-dependent manner in C. elegans. (A) C. elegans stained with ORO and (B) bar graph showing neutral lipid levels in day 5 adult WT (N2) animals, and atg-3(bp412) mutant treated with DMSO or 5 µM AA-20 on day 1 of adulthood (WT-DMSO, N = 52; WT-AA-20, N = 54; atg-3(bp412)-DMSO, N = 34; atg-3(bp412)-AA-20, N = 34, from three biological replicates). Error bars indicate SEM. n.s. > 0.05, *P < 0.05, ***P < 0.001, ****P < 0.0001, by two-way ANOVA with Tukey’s multiple comparison test. (Scale bar, 500 µm.) (C) Animals expressing dhs-3p::dhs-3::gfp. (D) DHS-3 intensity was quantified in day 5 adult animals (WT-DMSO, N = 35; WT-5 µM AA-20, N = 40; atg-3(bp412)-DMSO, N = 32; atg-3(bp412)- 5 µM AA-20, N = 32, from three biological replicates). Error bars indicate SEM. n.s. > 0.05, ****P < 0.0001, by two-way ANOVA with Tukey’s multiple comparison test. (Scale bar, 500 µm.)

To determine whether AA-20-mediated lipid reduction was linked to autophagy, we analyzed ORO staining and the DHS-3::GFP reporter in hypomorphic atg-3(bp412) mutants. ATG-3 is a nonredundant enzyme important for LGG-1/Atg8 lipidation (46). We examined the autophagy function in these mutants by quantifying GFP::LGG-1/Atg8 puncta in neurons, intestine, and body-wall muscle of adult WT and atg-3(bp412) animals. In WT animals, BafA further increased autophagic puncta more than animals fed no BafA, whereas in atg-3(bp412) mutants, GFP::LGG-1/Atg8 puncta levels remained unchanged across all tissues with or without BafA treatment (SI Appendix, Fig. S3 E–G), showing autophagy is impaired in these mutants. Notably, despite this autophagy defect, adult atg-3(bp412) mutants displayed tissue-specific variations in basal GFP::LGG-1/Atg8 puncta levels despite the autophagy defect (SI Appendix, Fig. S3 E–G). By using this hypomorphic mutant, rather than a lethal complete knockout, we could study the consequences of autophagy impairment while avoiding severe developmental defects. Our imaging analyses revealed no significant increase in ORO or DHS-3::GFP intensity in this autophagy-deficient mutant under DMSO conditions compared to WT (Fig. 3 A–D), likely due to the weak bp412 allele rather than a severe knockout. However, we observed that AA-20 did not reduce ORO or DHS-3::GFP intensity in this mutant (Fig. 3 A–D), demonstrating that AA-20 promotes neutral lipid clearance in an autophagy-dependent fashion.

To investigate whether the reduction in neutral lipid levels in C. elegans was associated with an impact on feeding behavior, we analyzed pharyngeal pumping rates, which closely correlate with the ability of C. elegans’ to consume food. Both DMSO- and AA-20-treated animals showed a gradual decline in pharyngeal pumping rates over time (SI Appendix, Fig. S3H), a common characteristic of aging in C. elegans (47). Interestingly, despite this overall gradual decline, we noticed preservation of pharyngeal pumping at day 7 of adulthood in animals fed with AA-20 compared with controls, suggesting that AA-20 may provide mild protection against age-related decline in pharyngeal function in C. elegans. Additionally, we observed no significant difference in bacterial food intake after three days between day 4 adult WT C. elegans treated with AA-20 and those treated with DMSO (SI Appendix, Fig. S3I). This timeframe overlaps with the period assessed for lipid level measurements in day 5 adult animals (Fig. 3 A–D), indicating that the AA-20 mediated reduction in lipid levels is likely not attributable to changes in the animals’ feeding behavior.

AA-20 Reduces PolyQ Aggregation Load in Both Human Cells and in C. elegans.

Recognizing that protein aggregates represent another autophagic substrate associated with neurodegeneration (48), we next explored whether AA-20 treatment could reduce protein aggregate levels. We tested this hypothesis in the context of polyglutamine (PolyQ) aggregates, which are associated with several neurodegenerative disorders including Huntington’s disease; wherein the PolyQ sequence is encoded by cytosine-adenine-guanine triplet repeats (49). We transiently transfected HeLa cells, the same cells used to demonstrate AA-20-induced autophagy flux (Fig. 1 C–F), with an EGFP-Poly-Q74 construct. This construct is known to form large aggregates that accumulate in rat PC12 cells under normal conditions (50). Subsequently, we conducted imaging analysis of GFP-Q74 in HeLa cells. Treatment with AA-20 (10 µM) for 20 h, starting 24 h after transfection, significantly reduced the average number of GFP-Q74 aggregates per cell compared to DMSO treatment (Fig. 4 A and B).

Fig. 4.

AA-20 ameliorates protein aggregate overload in both mammalian cells and in C. elegans. PolyQ aggregate loads were quantified in HeLa cells (A and B) and in C. elegans (C–H) after DMSO or AA-20 treatment. Representative confocal fluorescence micrographs of HeLa cells PolyQ-74 (green) (A). (Scale bar, 10 µm.) Cells were transfected with GFP-tagged PolyQ-74 constructs and treated with DMSO or 10 µM AA-20 and further cotreated with either vehicle (DMSO) or BafA for 20 h. GFP-positive Q74 puncta were visualized by confocal microscopy and quantified using Imaris software’s Spot Detection function (B) (total ~45 to 109 cells analyzed, N = 4 to 7 images per condition, from three independent experiments). Error bars represent SEM. ***P < 0.001, by one-way ANOVA with Tukey’s multiple comparison test. White arrowheads indicate GFP-Q74-positive aggregates. (C−H) PolyQ aggregate loads were quantified in C. elegans expressing PolyQ stretches in different tissues after DMSO or 5 µM AA-20 treatment. Representative images of PolyQ in neurons (Scale bar, 10 µm.) (C), intestine (E), and body-wall muscle (G) tissues (Scale bar, 500 µm.) White arrowheads indicate Q40::YFP- or Q35::YFP-positive protein aggregates, respectively. Neuronal PolyQ aggregate numbers were quantified on day 7 of adulthood in WT or atg-3(bp412) animals expressing rgef-1p::Q40::YFP (D) (WT-DMSO, N = 32; WT-5 µM AA-20, N = 38; atg-3(bp412)-DMSO, N = 39; atg-3(bp412)-AA-20, N = 38, from four biological repeats). Intestinal PolyQ numbers were quantified in day 4 adult animals expressing vha-6p::Q44::yfp after treatment with either DMSO or AA-20 on day 1 of adulthood (F) (WT-DMSO, N = 40; WT-5 µM AA-20, N = 40; atg-3(bp412)-DMSO, N = 35; atg-3(bp412)-AA-20, N = 40, from four biological repeats), and muscle PolyQ numbers were quantified in day 4 adult animals expressing unc-54p::Q35::YFP (H) (WT-DMSO, N = 40; WT-5 µM AA-20, N = 39; atg-3(bp412)-DMSO, N = 38; atg-3(bp412)-AA-20, N = 35, from four biological repeats). Error bars represent SEM. n.s. > 0.05, *P < 0.05, ****P < 0.0001, by two-way ANOVA with Tukey’s multiple comparison test.

To investigate whether this reduction depends on lysosomal degradation, we cotreated HeLa cells with BafA alongside either DMSO or AA-20 for 20 h. In DMSO-treated cells, BafA did not alter PolyQ74 aggregate levels, consistent with prior studies suggesting that basal PolyQ degradation via autophagy varies (51). However, in cells treated with both AA-20 and BafA, aggregate levels were significantly higher than with AA-20 alone, demonstrating that lysosomal degradation is critical for AA-20’s ability to clear aggregates. We also costained these cells with an anti-LC3B antibody (SI Appendix, Fig. S4A). AA-20 caused a ~15%, nonsignificant increase in basal LC3B levels compared to DMSO, while BafA treatment elevated LC3B levels more than either DMSO or AA-20 alone (SI Appendix, Fig. S4B). The slight LC3B increase with AA-20, coinciding with reduced aggregate levels, suggests increased LC3B turnover rather than a lack of AA-20 efficacy. Additionally, the LC3B increase in DMSO-treated cells with BafA confirmed BafA’s effectiveness as a control, supporting the idea that PolyQ is differentially degraded by autophagy under normal conditions (51). AA-20, however, appears to amplify this process, as its effect was lost when lysosomal function was blocked by BafA. Finally, we evaluated autophagic activity and net degradation flux on PolyQ74 aggregate substrates. Similarly to its effects on lipid substrates (SI Appendix, Fig. S3 C and D), AA-20 significantly enhanced both autophagic fold-change and net degradation (SI Appendix, Fig. S4 C and D). Together, these results suggest that AA-20 reduces aggregate levels by boosting autophagy’s degradative capacity.

Notably, PolyQ::YFP proteins can be expressed in different tissues of C. elegans and accumulate with aging (52). Autophagy can counteract the accumulation of PolyQ proteins and mitigate their potential toxic effects in C. elegans (53). Therefore, we analyzed the ability of AA-20 to affect PolyQ aggregate levels in C. elegans. We quantified the number of PolyQ aggregates per animal that appeared in individual tissues expressing tissue-specific, YFP-tagged PolyQ proteins, which are known to aggregate at different rates (52). Specifically, we scored aggregates in the nerve-ring (PolyQ40, Fig. 4 C and D), or in the intestine (PolyQ40, Fig. 4 E and F), and body-wall muscle (PolyQ35, Fig. 4 G and H) of both WT and atg-3(bp412) animals. AA-20 significantly reduced the Q35 or Q40 aggregate loads per WT animal by 20 to 40% in neurons (Fig. 4 C and D), intestine (Fig. 4 E and F), and body-wall muscle (Fig. 4 G and H). In contrast, AA-20 did not decrease PolyQ aggregates in any examined tissues of autophagy-deficient atg-3(bp412) mutants (Fig. 4 C and H), which aligned with our results in HeLa cells (Fig. 4 A and B), indicating that AA-20 promotes PolyQ aggregate clearance in an autophagy-dependent manner in C. elegans. Collectively, these observations show that AA-20 reduces PolyQ aggregation in both human cells and C. elegans tissues through an autophagy-dependent mechanism.

The proteasome is an alternative, nonautophagic degradative pathway for PolyQ aggregates (54). Notably, PolyQ overexpression increases the protein-degradation load on the proteasome, and disruptions in the proteasomal degradation process can exacerbate the formation of cellular protein aggregates, leading to proteotoxic stress (55, 56). Therefore, building on our understanding that AA-20 enhances autophagy, we investigated its potential to counteract and protect against PolyQ aggregation-associated proteotoxic stress under condition of proteasomal inhibition. We selectively inhibited the 20S subunit of the proteasome with bortezomib (57) treatment in C. elegans expressing PolyQ in the body-wall muscle (SI Appendix, Fig. S4E). As previously demonstrated, bortezomib treatment led to an uncoordinated (Unc) phenotype (58) and shorter body length, whereas AA-20 treatment alone did not produce these effects (SI Appendix, Fig. S4 F and G). Our imaging analysis showed that pretreatment with AA-20 for 12 h at a higher concentration of 33 µM before adding bortezomib on day 1 of adulthood partially reversed the body-length shortening that was observed in day 6 adult animals (SI Appendix, Fig. S4 F and G). These findings suggest AA-20 mitigates the deleterious effects of bortezomib’s proteasome inhibition, potentially supporting animal survival under proteasomal stress. Further direct assessments are required to investigate the relationship between AA-20 and proteasome function.

AA-20 Promotes Healthspan in C. elegans.

Given the established role of autophagy activation in extending lifespan and healthspan (59), we next asked how AA-20 might affect the overall fitness of aging C. elegans. As a start, we examined whether AA-20 had any impact on the reproductive fitness of C. elegans by measuring the reproductive output of C. elegans exposed to AA-20. Our results showed that adding 5 µM AA-20 to eggs or to day 1 adult WT animals on solid nematode growth medium (NGM) media plates coated with AA-20 did not lead to any significant changes in total brood size (SI Appendix, Fig. S5 A and B). These findings suggest that AA-20 does not adversely affect the reproductive fitness of C. elegans. We also investigated the impact of AA-20 on swimming proficiency, which is known to decline over time (47). To measure swimming proficiency, we counted the number of body bends or thrashes per minute on days 5, 7, 12, and 14. Our findings demonstrated that AA-20 significantly increased the thrashing per minute over time (Fig. 5A), suggesting AA-20 enhances baseline mobility in aged C. elegans, an important aspect of fitness. However, the comparable slope of decline in swimming ability between treated and untreated groups suggests that while AA-20 enhanced initial performance, it might only delay the onset of decline in muscle mobility rather than preventing decline over time.

Fig. 5.

AA-20 promotes healthspan and lifespan in C. elegans. (A) The number of thrashes per minute were measured on day 5, day 7, day 12, and day 14 in WT animals that were fed with DMSO or 5 µM AA-20 on day 1 of adulthood (DMSO, N = 42; 5 µM AA-20, N = 42, from four biological replicates). Error bars indicate SEM. ****P < 0.0001, by two-way ANOVA with Šídák’s multiple comparisons test. (B) Representative DIC images of animals subjected to Smurf assay. (Scale bar, 500 µm.) (C) Quantification of percent animals with blue dye in body cavity during day 10 of WT animals. Day 10 animals fed with DMSO or 5 µM AA-20 on day 1 were soaked in blue food dye for 3 h. Animals containing blue dye leakages from the intestinal lumen into the body cavity were considered Smurf phenotype. Animals that presented no blue dye leakages from intestine to body cavity but contain dye in gonads were censored from the analysis. Data are the mean ± SEM of four biological repeats, each with 34 to 70 animals per condition. *P < 0.05, by the unpaired t-test. (D) Survival percentage of WT (N2, WT) animals subjected to heat shock. Animals were fed with DMSO or 5 µM AA-20 on day 1 of adulthood. On day 5 of adulthood, animals were subjected to a lethal dosage of heat shock at 36 °C for 7 h and survival was calculated for each plate (N = 24 to 25 plates per condition, from five independent experiments). Error bars represent SEM. ****P < 0.0001, by the unpaired t-test. (E) WT animals were bleached and cultured in liquid from L1 onward and DMSO or 5 µM AA-20 were added once during day 1 of adulthood for subsequent lifespan analysis (WT-DMSO, N = 139; WT-AA-20, N = 141 animals). A total of 13 independent repeats were conducted (see Dataset S1 for all data). (F) Lifespan was analyzed for liquid cultured WT animals added once with DMSO or 5 µM AA-20 during day 5 of adulthood (WT-DMSO, N = 158; WT-AA-20, N = 110 animals). Three independent lifespan analyses were conducted (refer to Dataset S2 for data). (G) Lifespan analysis of WT compared with atg-3(bp412) animals at 20 °C subjected to DMSO or 5 µM of AA-20 treatment. Liquid lifespan assay was performed in WT and atg-3(bp412) animals fed once with DMSO or 5 µM AA-20 on day 1 of adulthood (WT-DMSO, N = 35; WT-5 µM AA-20, N = 34; atg-3(bp412)-DMSO, N = 106; atg-3(bp412)- 5 µM AA-20, N = 103 animals). See Dataset S3 for lifespan results of three independent experimental repeats.

We next examined the effect of AA-20 on intestinal integrity, another age-related phenotype in C. elegans (60). To do so, we used a so-called Smurf assay, which was adapted from previous studies in Drosophila (61, 62). In this assay, a blue dye was fed to day 10 adult animals treated with either AA-20 or DMSO from day 1 onward. The results showed that AA-20 significantly reduced the percentage of animals with dye leakage from the intestine into the body cavity (Fig. 5 B and C). This result is consistent with our previous findings that enhanced autophagy induced by dietary restriction could maintain intestinal barrier function during aging (60).

Heat shock is a widely recognized stressor that can have detrimental effects on the survival of organisms, including C. elegans (63). Thus, we next investigated whether AA-20 could affect the animals’ stress tolerance following a heat shock. We exposed day 4 adult WT animals to a prolonged heat shock, namely to 36 °C for 7 h as previously described (64), then assessed the survival rate of the animals. Our results revealed a significant increase in the survival rate of day 4 animals treated with AA-20 (Fig. 5D). These findings indicate that AA-20 can protect C. elegans against heat shock, highlighting its potential as a resilience enhancer. Collectively, our results provide evidence that AA-20 is an autophagy activator that promotes increased healthspan in C. elegans.

AA-20 Extends C. elegans Lifespan by an Autophagy-Dependent Mechanism That Operates in Early Adulthood.

Since we observed that AA-20 improved C. elegans overall fitness, we also examined its impact on C. elegans lifespan. We used a liquid lifespan assay, wherein age-synchronous day 1 adult WT animals were fed AA-20 in a liquid format with dead bacteria in 96-well plates, along with fluorodeoxyuridine (FUDR) to maintain sterility (37). We tested 5 µM AA-20, which robustly activated autophagy in all C. elegans tissues (Fig. 2 and SI Appendix, Fig. S2 B and C), and observed a significant and reproducible mean lifespan extension of ~20 to 40% (10 out of 13 repeats with P < 0.0001, log-rank) (Fig. 5E and Dataset S2).

Additionally, we conducted lifespan assays on solid NGM media plates (standard C. elegans husbandry) using dead bacteria as the food source, without adding FUDR. Age-synchronous adult WT animals were transferred to fresh plates coated with AA-20 every other day until they were reproductively halted. We found that 5 µM of AA-20 increased the mean lifespan of C. elegans under these conditions by ~10 to 30% (5 out of 6 repeats with P < 0.0001, log-rank) (Dataset S2), suggesting that the lifespan extension observed in the liquid assay was not due to FUDR.

We also performed liquid lifespan assays using live OP50, the standard C. elegans food source, and observed similar mean lifespan extension (3 out of 3 repeats with P < 0.0001, log-rank) (Dataset S2), suggesting AA-20’s effect on longevity is independent of bacteria’s metabolism or nutritional contribution. Collectively, these results demonstrate that the observed lifespan extension is specific to AA-20 and not influenced by live bacteria, solid media, or the presence of FUDR in the lifespan assays.

To assess the optimal start time of administering AA-20 for effects on lifespan, we conducted liquid lifespan experiments starting administration at different stages of the C. elegans life cycle. AA-20 was administered to media at three different time points (with continuous exposure thereafter): L1 (first larval/developmental stage), day 5 (at/around postreproductive stage), and day 10 adulthood (late adult stage). Feeding AA-20 from the L1 stage did not extend mean lifespan of WT animals (2 out of 2 repeats with n.s., log-rank) (Dataset S3). In contrast, when AA-20 was added during day 5, we observed significantly increased mean lifespan extension, similar to that of animals fed with AA-20 from day 1 of adulthood (3 out of 3 repeats with P < 0.0001, log-rank) (Fig. 5F and Dataset S3). However, the mean lifespan extension was not significant when AA-20 was provided at day 10 (2 out of 2 repeats with n.s., log-rank) (Dataset S3). This lack of lifespan extension could potentially be attributed to a decline in pharyngeal pumping rates and, in turn, reduced compound consumption (SI Appendix, Fig. S3C), or AA-20 may engage pathways that are no longer activatable in older animals.

Additionally, we conducted experiments to assess the impact of short-term AA-20 treatment on lifespan using both solid NGM plates and liquid lifespan methods. To do so, we initially exposed WT animals to AA-20 treatment on day 1 of adulthood and then progressively removed it on a daily basis. Our results revealed that 4 d (i.e., day 1 to day 5) of AA-20 treatment were sufficient to significantly extend the lifespan of C. elegans (5 out of 6 repeats with P < 0.0001, log-rank) (Dataset S3). Altogether, these findings suggest that AA-20’s efficacy in promoting longevity is most pronounced during the early to mid-adulthood phase, with the compound’s beneficial pathways likely being most activatable within this timeframe.

To understand how AA-20 promoted longevity, we conducted epistasis studies using atg-3(bp412) (46, 65) mutants (66), as we did in our lipid (Fig. 3) and PolyQ (Fig. 4) analyses. Notably, AA-20 failed to extend the mean lifespan of atg-3(bp412) mutants (3 out of 3 repeats with P > 0.0001, log-rank) (Fig. 5G and Dataset S4), consistent with AA-20 functioning at least in part by inducing autophagy. Moreover, these results align with established longevity paradigms requiring autophagy genes as a common mechanism for extending lifespan (5). Collectively, we conclude that AA-20 extends lifespan in an autophagy-dependent manner.

AA-20 Promotes Autophagy via a TFEB/HLH-30-Dependent Mechnism without Triggering Global Transcriptional Activation.

Given that many established autophagy activators inhibit mTORC1 and can lead to activation of the helix–loop–helix transcription factor TFEB (67–70), we explored whether AA-20 induced autophagy through similar mechanisms in mammalian cells and in C. elegans. We first assessed mTORC1 activity by analyzing the phosphorylation levels of its key substrates (22)—ULK1, TFEB, S6 kinase, and 4EBP1—via immunoblotting. Unlike the mTORC1 inhibitor Torin 1 (22), AA-20 did not alter the phosphorylation levels of these substrates in HeLa cells (SI Appendix, Fig. S6 A–E). To investigate whether AA-20 affected mTORC1 in C. elegans, we measured the phosphorylation levels of C. elegans S6 kinase ortholog RSKS-1 in AA-20-treated C. elegans and found no changes (SI Appendix, Fig. S6 F and G). Additionally, we evaluated the developmental timing and body length of C. elegans exposed to AA-20, comparing it with Rapamycin, a well-known mTORC1 inhibitor, as mTORC1 activity is critical for growth and development (71). AA-20 showed no significant effect on developmental timing or body length, even at higher doses, whereas Rapamycin significantly shortened body length (SI Appendix, Fig. S6 H–J). Collectively, these results are consistent with AA-20 functioning through an mTOR-inhibition independent mechanism.

Notably, TFEB/HLH-30 regulates many lysosome and autophagy-related genes (72, 73), and several conventional autophagy activators depend on or influence TFEB/HLH-30 (67–70) activity. We used hlh-30(tm1978) loss-of-function mutants, which lack expression of essential autophagy genes (66), to assess whether TFEB/HLH-30 was required for the increase in lysosomal acidification (measured by LysoTracker Red staining) induced by AA-20 (Fig. 2 G and H and SI Appendix, Fig. S2C). While we did not observe a significant decrease in lysosomal acidification in hlh-30(tm1978) mutants compared to WT, possibly due to compensatory mechanisms (74), AA-20 was unable to increase LysoTracker Red intensity levels in these mutants (Fig. 6A). Additionally, since AA-20 required atg-3 to reduce lipid levels (Fig. 3) and PolyQ aggregate load (Fig. 4), we examined whether TFEB/hlh-30 was similarly necessary for AA-20 to decrease these substrates. Notably, unlike in WT animals, AA-20 failed to reduce DHS-3::GFP levels (Fig. 6B), and did not decrease the number of PolyQ aggregates in the nerve-ring (PolyQ40, Fig. 6C), in the intestine (PolyQ40, Fig. 6D), or in body-wall muscle (PolyQ35, Fig. 6E) in hlh-30(tm1978) mutants. Similarly, AA-20 did not extend mean lifespan in hlh-30(tm1978) mutants (3 out of 3 repeats) (Fig. 6F and Dataset S4). We also tested AA-20 on another transcription factor mutant, Forkhead box, class O protein (FOXO)/DAF-16, which is essential for numerous longevity paradigms. For this purpose, we utilized daf-16(mu86) null mutants. Our results demonstrated that AA-20 extended the mean lifespan of daf-16(mu86) mutants by 20-30% (3 out of 3 repeats with P < 0.0002, log-rank) (Dataset S4), indicating that daf-16 is not essential for the lifespan extension induced by AA-20 like hlh-30 is. Altogether, these findings support that TFEB/HLH-30 is essential for the beneficial effects of AA-20 on C. elegans lifespan, lysosomal acidification, and substrate reduction.

Fig. 6.

AA-20 promotes autophagy in C. elegans via a TFEB/HLH-30-dependent mechanism without triggering its transcriptional activation (A) LysoTracker red staining intensity analysis in day 4 adult WT (N2) and hlh-30(tm1978) animals fed with DMSO or 20 µM AA-20 during day 1 of adulthood from three independent repeats (N = 18 to 22 animals). (B) DHS-3 intensity was quantified in day 5 adult animals (WT-DMSO, N = 30; WT-5 µM AA-20, N = 26; hlh-30(tm1978)-DMSO, N = 27; hlh-30(tm1978)- 5 µM AA-20, N = 28, from three biological replicates). (C–E) PolyQ aggregate loads quantification in WT or hlh-30(tm1978) animals expressing PolyQ stretches in different tissues after DMSO or 5 µM AA-20 treatment. Neuronal PolyQ aggregate numbers were quantified on day 7 of adulthood in WT or hlh-30(tm1978) animals expressing rgef-1p::Q40::YFP (C) (WT-DMSO, N = 26; WT-5 µM AA-20, N = 30; hlh-30(tm1978)-DMSO, N = 25; hlh-30(tm1978)-AA-20, N = 25, from four biological repeats). Intestinal PolyQ numbers were quantified in day 4 adult animals expressing vha-6p::Q44::yfp after treatment with either DMSO or AA-20 on day 1 of adulthood (D) (WT-DMSO, N = 31; WT-5 µM AA-20, N = 36; hlh-30(tm1978)-DMSO, N = 41 hlh-30(tm1978)-AA-20, N = 38, from four biological repeats), and muscle PolyQ numbers were quantified in day 4 adult animals expressing unc-54p::Q35::YFP (E) (WT-DMSO, N = 40; WT-5 µM AA-20, N = 35; hlh-30(tm1978)-DMSO, N = 36; hlh-30(tm1978)-AA-20, N = 40, from four biological repeats). Error bars represent SEM. n.s. > 0.05, *P < 0.05, **P < 0.01, ****P < 0.0001, by two-way ANOVA with Tukey’s multiple comparison test. (F) Lifespan was assayed in WT and hlh-30(tm1978) animals treated with 0.05% DMSO or 5 µM AA-20 coated on NGM plates (WT-DMSO, N = 114; WT-AA-20, N = 114; hlh-30(tm1978)-DMSO, N = 105; hlh-30(tm1978)-AA-20, N = 102 animals). P-values by the log-rank test. See Dataset S4 for experimental repeats. (G) Representative images of day 2 adult C. elegans expressing hlh-30p::hlh-30::gfp + rol-6. (Scale bar, 500 µm.) White arrowheads highlight HLH-30::GFP localized in the nuclei. (H) The percentage TFEB/HLH-30 nuclear localization was quantified for the population within each image of C. elegans after 24 h of treatment (DMSO, N = 28; 5 µM AA-20, N = 30 animals) across three independent experiments. Error bars represent SEM. *P < 0.05, ****P < 0.0001, by the unpaired t-test. (I) Expression of TFEB/hlh-30 and putative autophagy-related and lysosomal target genes was measured using RT-qPCR in day 2 adult C. elegans (5 µM AA-20) WT animals raised at 20 °C, from three independent experiments. Error bars represent SEM. *P < 0.05, ****P < 0.0001, by the multiple t-test.

We next investigated how AA-20 affects TFEB/HLH-30 function. TFEB/HLH-30 translocates to the nucleus in response to several activating stimuli, including dietary restriction, and HLH-30/TFEB is nuclear localized in multiple long-lived C. elegans mutants (66). We therefore examined the subcellular localization of HLH-30/TFEB following AA-20 treatment, using a HLH-30::GFP reporter driven by its endogenous promoter (66). On day 2 of adulthood, we found that animals treated with AA-20 were 50% more likely to display nuclear localization of TFEB/HLH-30 compared to those treated with DMSO (Fig. 6 G and H). This result suggests that AA-20 promotes the activation of TFEB/HLH-30.

To directly test whether AA-20 affected TFEB/HLH-30 transcriptional activity, we performed RT-qPCR analyses using day 2 adult animals which showed increased HLH-30 nuclear localization (Fig. 6 G and H) on a panel of TFEB-regulated autophagy-related genes, including BECN1/bec-1, MAP1LC3/lgg-1, ATG9/atg-9 (for vesicle formation and elongation), P62/SQSTM1/sqst-1, LAMP1/lmp-1, VPS11/vps-11, VPS18, ATP6V1H/vha-15 (for autophagosomes/lysosome fusion), CTSA, CTSB/cpr-1, and CTSD/asp-1 (66, 75). Our RT-qPCR results showed that AA-20 modestly increased the expression of LAMP1/lmp-1 and vha-15 levels, significantly reduced BECN1/bec-1 expression, and had minimal effects on the other genes tested (Fig. 6I). This pattern contrasts with the strong transcriptional activation typically observed with autophagy activators like Torin 1 (76), which broadly affect TFEB transcriptional programs. Collectively, these results suggest that AA-20 functions independently of mTORC1, yet requires TFEB/HLH-30 as a prerequisite for augmenting the number of acidic lysosomes, as demonstrated by intensified LysoTracker Red staining. Despite this, AA-20 does not appear to activate a canonical TFEB/HLH-30-mediated transcriptional response.

To address the counterintuitive results on TFEB/HLH-30’s involvement, we asked whether AA-20 could boost lysosomal function via a transcription-independent pathway. To do this, we conducted experiments in C. elegans by treating them with both AA-20 and 30 µg/mL actinomycin D, which inhibits RNA polymerase and thus prevents new mRNA synthesis (77) in C. elegans (78) (SI Appendix, Fig. S6K). This approach assessed whether AA-20-induced lysosomal acidification, shown by LysoTracker Red staining (Fig. 2 G and H and SI Appendix, Fig. S2C), depended on transcription or occurred via nontranscriptional mechanisms. As expected, AA-20 increased LysoTracker Red staining in day 3 adult WT animals (SI Appendix, Fig. S6L). Although adding actinomycin D reduced the magnitude of this increase, AA-20 still significantly enhanced lysosomal acidification compared to controls treated with either vehicle or actinomycin D alone (Fig. 6L). This indicates AA-20 promotes autophagy at least partly independently of transcriptional upregulation, without requiring a typical TFEB-mediated transcriptional response.

In summary, our study has unveiled AA-20 as a conserved autophagy activator that promotes elements of fitness in C. elegans. Furthermore, AA-20 extends lifespan via autophagy genes, consistent with it being an inducer of autophagy. Although the exact mechanism of action by AA-20 requires further elucidation, our findings suggest that AA-20 operates independently of mTORC1-inhibition but involves TFEB/HLH-30. Future research should focus on uncovering the precise molecular mechanisms involved and exploring AA-20’s potential therapeutic applications in age-related diseases by targeting autophagy pathways.

Discussion

Lysosomal dysfunction is strongly associated with aging and neurodegeneration (79), yet pharmacological interventions targeting the enhancement of lysosomal efficiency are limited. While pharmacological autophagy inducers like rapamycin have shown therapeutic benefits in neurodegenerative disease mouse models by inhibiting mTOR (80–82), the potential autophagy-independent effects of mTORC1 inhibition in these studies have not been fully investigated. On the other hand, trehalose, an autophagy activator independent of mTOR, has demonstrated therapeutic benefits (83–86), yet its effectiveness in enhancing autophagy has been debated (27, 87). Considering that the majority of identified autophagy-inducing compounds operate by inhibiting mTORC1 kinase activity, the scarcity of mTORC1-independent autophagy enhancers (86, 88) indicates a need for more extensive exploration in this area. Herein, we introduce a validated hit compound, AA-20, that enhances autophagy, apparently without inhibiting mTORC1 kinase activity.

In this study, our findings characterize AA-20 as an unconventional autophagy inducer, setting it apart from typical triggers such as starvation or mTORC1 inhibitors. Unlike the strong LC3 responses commonly seen with these conventional inducers, AA-20 produces more subtle effects in LC3/LGG-1-based assays. Its impact on net LC3B flux (the difference in LC3 levels between BafA-treated and untreated conditions) and the LC3-GFP/RFP ratio is relatively modest. This muted LC3B-specific response suggests that AA-20’s influence on autophagy may extend beyond LC3B, possibly engaging other ATG8/LC3 homologs, an avenue that remains to be further investigated. Initially, we posited that AA-20 primarily promotes autophagosome biogenesis. However, its capacity to expand the lysosomal pool and enhance substrate clearance flux points to a broader mechanism. Alternatively, AA-20 may also enhance autophagy by improving substrate recruitment efficiency, thereby increasing degradation capacity. Notably, while traditional autophagy enhancers typically suppress mTORC1, our analysis of mTORC1 activity markers and C. elegans developmental parameters reveals that AA-20 operates through a distinct mechanism, independent of mTORC1 inhibition. Overall, these findings highlight AA-20’s unique role as an autophagy inducer, distinct from mTORC1-dependent pathways, and emphasize the need to explore its interactions with the broader autophagic machinery.

TFEB/HLH-30 is crucial for AA-20’s effects on autophagy and lifespan extension, yet our findings within a 24-h window suggest AA-20 might engage TFEB in nontranscriptional functions. This could include TFEB’s phosphorylation for nuclear localization, although our examination in HeLa cells showed that phosphorylation at site 211, critical for nuclear localization (20), remained unaffected. Alternatively, AA-20 might enhance TFEB’s interaction with key autophagy proteins, improving efficiency without significant transcriptional changes. It could also regulate TFEB’s role in autophagy dynamics, like autophagosome-lysosome fusion, or boost the activity of existing autophagic proteins (72), enhancing lysosomal enzyme efficiency. Our RT-qPCR results, indicating minimal transcriptional changes in TFEB/HLH-30-regulated autophagy genes, support the notion that AA-20’s action is predominantly nontranscriptional. However, we noted a significant decrease in the expression of the TFEB/HLH-30 regulated gene BECN/bec-1, which might hint at compensatory mechanisms or could suggest involvement of noncanonical autophagy processes (89). However, our experiments with BafA cotreatment, known to also promote noncanonical or secretory autophagy (90), revealed further substrate (e.g., neutral lipids, protein aggregates) accumulation with AA-20, reducing the likelihood of noncanonical autophagy, though this still merits further investigation. Additionally, our findings with actinomycin D, an RNA polymerase inhibitor, showed that AA-20 still increased levels of acidified lysosomes, reinforcing the idea that AA-20 enhances lysosomal function independently of new gene transcription. This suggests AA-20 might operate via posttranscriptional modifications or alternative regulatory pathways, warranting deeper experimental exploration. Collectively, these findings position AA-20 as an autophagy-enhancing compound with a mechanism that may not rely solely on transcriptional activation, presenting a potential avenue for therapeutic intervention.

Growing evidence suggests a close relationship between accumulation of lipids and protein aggregate cargoes with reciprocal impacts (91, 92). Lipid droplets have been identified as temporary storage sites for hydrophobic proteins prone to aggregation, such as α-synuclein and apolipoprotein B, which are often associated with neurodegenerative disorders (93, 94). Studies have shown that reducing oxidized lipids, also prone to aggregation (95, 96), decreases oligomerization of misfolded proteins or aggregates (91). In this study, we found that AA-20 consistently reduced lipid levels across various model systems. Notably, among the three lysosomal enzymes we tested, AA-20 appears to be particularly effective in increasing LAL activity, the primary enzyme for lipophagy (97) (selective degradation of lipid droplets by autophagy), suggesting a preference for lipid droplets as a cargo. Further research is needed to explore the dependence of AA-20 on LAL/lipophagy regulators, its tissue-specific effects on autophagic kinetics and substrates, and its influence on membrane fluidity relevant to organismal health (98, 99). Nonetheless, this study highlights AA-20’s high efficacy in reducing lipid levels. Given its effectiveness in reducing lipid load, a prioritized investigation should also test the efficacy of AA-20 in reducing aggregates associated with lipid droplets, such as Lewy bodies (100). These efforts will provide valuable insights into the interplay between lipid composition and aggregate clearance.

In our study, we demonstrated that AA-20 enhances multiple aspects of fitness and promotes longevity in C. elegans, suggesting its potential as a possible intervention against age-related health deterioration and in promoting healthy aging. Our results indicated that AA-20 extended lifespan when administered between postdevelopment (day 1 of adulthood) and postreproductive age (day 5 of adulthood). No significant mean lifespan extension was observed with AA-20 from the L1-stage or at older ages (day 10 of adulthood). This lack of effect in early life could be due to the treatment altering developmental progression (101), potentially negating later benefits. In contrast, in older adults, the decline in mechanisms like autophagy (5) might reduce AA-20’s effectiveness, highlighting the importance of the timing of administration. Consistent with this, we found that AA-20–induced longevity is dependent on autophagy, as observed for many existing longevity paradigms (5). Moreover, AA-20 protected C. elegans from proteotoxic stress induced by PolyQ aggregate overload when proteasomal function was impaired, underscoring the critical role of its autophagy-activating effects in maintaining organismal health and lifespan. While autophagy appears central to AA-20’s effects, it remains possible that additional intracellular processes are involved in AA-20–mediated lifespan extension. It will also be interesting to investigate other hits from our original million-molecule HTS (26) and their effects on longevity. The identification of AA-20 and the close link between autophagy and longevity (5) raise the intriguing possibility of discovering additional autophagy activators from our screen that could promote fitness, extend lifespan, and ameliorate neurodegeneration in the future.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Dataset S03 (XLSX)

Dataset S04 (XLSX)

Dataset S05 (XLSX)

Dataset S06 (XLSX)

Dataset S07 (XLSX)

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.