Introduction

Mitochondria are essential organelles involved in various cellular functions, including energy production, calcium regulation, cell signaling, and apoptosis. Their reciprocal relationship with aging is widely acknowledged, wherein mitochondrial dysfunction impacts organismal longevity and health, and aging affects mitochondrial homeostasis across variable model animals ^{1, 2}. The degree of mitochondrial dysfunction determines the effect on an organismal lifespan. Mild mitochondrial disruption has been shown to increase lifespan in model animals such as C. elegans, flies, and mice ^3–10. For instance, the lifespan of C. elegans is extended by inhibiting mitochondria-related genes through genetic and RNA interference (RNAi) knockdown, including isp-1 (an iron-sulfur subunit of complex III of the mitochondrial electron transport chain (ETC)) ⁶, spg-7 (a mitochondrial quality control m-AAA protease) ^11–13, and clk-1 (a hydroxylase involved in the biosynthesis of ubiquinone) ^{5, 14}. Like C. elegans clk-1 mutant, mice with the mclk1 mutation, which exhibits normal growth and fertility, also exhibits an increase in lifespan ¹⁵.

The mechanisms underlying longevity enhancements through mitochondrial perturbations bring forth the mitohormesis theory that enhanced stress response pathways contribute to longer lifespans ^{4, 16–21}. It has been demonstrated that some mitochondrial stress requires the mitochondrial unfolded protein response (mitoUPR) pathway to trigger lifespan extension ^{4, 16–19}. The forkhead transcription factor DAF-16/FoxO has also been reported as an additional mediator responsible for extending the lifespan of C. elegans ETC mutants ^{20, 21}. However, certain mitochondrial perturbations and mutations in ETC genes can independently extend lifespan without involving atfs-1, a key regulator of mitoUPR ^{17, 22}. Similarly, it has been documented that DAF-16/FoxO is not indispensable in mediating the lifespan extension in some ETC mutants ⁶, suggesting a complex regulation of organismal aging in response to mitochondrial dysfunction through multiple pathways.

Aging is also accompanied by various changes in mitochondria, including the decreased activity of mitochondrial enzymes, reduced mitochondrial oxygen consumption, increased ROS production, decreased mitochondrial biogenesis, and increased mutations in mitochondrial DNA ^23–27. Interestingly, different types and levels of mitochondrial DNA mutations accumulate in various tissues during aging in humans and model animals ^{28, 29}. Even within a single tissue, such as the brain and muscle, variations in mitochondrial DNA mutations have been reported ^30–32. As a result, there is growing interest in understanding how organisms respond to mitochondrial stress in specific tissues and cells.

Recent studies have elucidated the phenomenon wherein perturbations in mitochondrial function within a specific tissue can elicit lifespan extension and health improvement ^{16, 33, 34}. Remarkably, the nervous system exhibits a particular role in detecting its intrinsic mitochondrial stress and extending organismal lifespan ^{16, 33–36}. However, it remains unclear which neuronal subtype is responsible for extending organismal longevity in response to their mitochondrial stress and what mechanisms underlie the non-cell-autonomous changes in this regard.

γ-aminobutyric acid (GABA) is a widely utilized neurotransmitter found in both vertebrate and invertebrate nervous systems. In vertebrates, a substantial proportion, estimated to be between 30% and 40%, of central nervous system synapses rely on GABA transmission ³⁷. GABA levels tend to gradually decrease as organisms age, and disruptions in GABA neurotransmission have been associated with various neurological disorders and age-related cognitive decline ^{38, 39}. In C. elegans, 26 neurons have been anatomically and functionally characterized as GABAergic neurons ^{40, 41}. Interestingly, recent studies in C. elegans indicate that the GABA signaling pathway plays an important role in regulating the lifespan and overall health of the organism, mitochondrial unfolded protein response, and proteostasis in the post-synaptic muscle tissue ^42–45.

In this study, we utilized C. elegans as a model organism to explore the role of GABAergic neurons in regulating organismal health and aging in response to mitochondrial perturbations, as well as the mechanisms underlying these effects. Our findings indicate that mitochondrial stress in GABAergic neurons can influence the activity of DAF-16/FoxO in peripheral tissues through GABA signaling. This, in turn, leads to non-cell-autonomous alterations in the mitochondria activity and organismal healthspan, reproductive capacity, and lifespan.

Results

Mitochondrial perturbations in GABAergic neurons are sufficient to prolong the organismal lifespan

To explore the influence of mitochondrial dysfunction in GABAergic neurons on lifespan and healthspan, we inhibited the functions of isp-1 and spg-7, well-documented genes that extend organismal lifespan when their normal functions are disrupted ^{11–13, 33, 46, 47}. In line with previous studies, systemic RNA interference (hereinafter referred to as sRNAi) targeting isp-1 and spg-7 significantly extended the lifespan of wild-type N2 animals when compared to control animals subjected to empty vector (EV) control RNAi (Figures 1A and 1B; Table S1). Tissue-specific RNAi in GABAergic neurons (hereinafter referred to as gRNAi) was accomplished by employing a previously well-established transgenic animal ^48–50. In this model, the function of rde-1, a member of the PIWI/STING/Argonaute protein family, was specifically restored in GABAergic neurons by introducing a transgene that expresses RDE-1 under the control of the GABAergic neuron-specific promoter derived from the unc-47 gene (Pgaba) (Figure 1C) ^{48, 51–53}. Additionally, they expressed sid-1, a gene responsible for dsRNA transport. When gRNAi against isp-1 was induced from the L1 stage, it resulted in a median lifespan extension of 55.5% compared to the control (Figures 1D and Table S1). spg-7(gRNAi) animals also showed an extended median lifespan of 33.3% (Figure 1E and Table S1).

Figure 1.

To exclude the possibility of any remaining systemic RNAi effects in the rde-1 mutant background, we carried out another tissue-specific RNAi strategy ⁵⁴. We expressed isp-1 double-stranded RNAs (dsRNAs) in GABAergic neurons using the Pgaba promoter (hereinafter referred to as Pgaba::isp-1 dsRNA) in the sid-1(qt9) null mutant background, which lacks intercellular dsRNA transport, preventing sRNAi (Figure 1F) ^{55, 56}. The expression of isp-1 dsRNAs in GABAergic neurons within the wild-type N2 animal background proved effective in upregulating the expression of Phsp-6::GFP, which serves as a fluorescent reporter for mitoUPR activity, specifically in the intestine, suggesting that this isp-1 dsRNA expression effectively induced mitochondrial defects (Figure S1A and S1B) ^{4, 34, 57}. Notably, we observed that Pgaba::isp-1 dsRNA expression in sid-1 null mutants also significantly extended organismal lifespan (Figures 1G and Table S1) ^58–60. Additionally, in 9-day-old adult isp-1(gRNAi) and spg-7(gRNAi) animals, we observed reduced lipofuscin fluorescence in the intestine, an aging hallmark that typically increases progressively over time (Figures 1H and 1I) ⁶¹. These findings collectively suggest that mitochondrial perturbation within GABAergic neurons is sufficient to prolong the organismal lifespan and attenuate the aging process.

Mitochondrial stress in GABAergic neurons increases the stress tolerance of the organism

Next, we sought to determine whether mitochondrial stress in GABAergic neurons could also alter the parameters of a healthspan. We tested thermal and oxidative stresses during aging, which are healthspan parameters closely linked to longevity across species ^62–66 (Figure 2A). In mid-age adults (3-day-old adult) groups, gRNAi against isp-1 and spg-7 increased survivability against paraquat exposure compared to controls (Figure 2B). This improvement was also observed in older adult groups (Figures 2C and 2D). Moreover, isp-1(gRNAi) and spg-7(gRNAi) animals displayed a significant increase in survival when exposed to thermal stress at 35 °C (Figures 2E-2G). To validate the efficacy of RNAi, we maximized the RNAi effect by culturing animals under the feeding RNAi condition against spg-7 for three generations (Figure S2A) and observed consistently improved survival rates in response to both oxidative stress (Figures S2B and S2C) and thermal stress (Figures S3D and S3E). In addition, sid-1 null mutants expressing Pgaba::isp-1 dsRNA expression also exhibited enhanced tolerance to thermal and paraquat stresses (Figures 2H and 2I). These findings suggest that mitochondrial stress in GABAergic neurons can enhance healthspan parameters.

Figure 2.

GABAergic neuronal mitochondrial stress alters reproduction

While depletion of some mitochondrial ETC in the entire nervous system of C. elegans and flies increases lifespan in a non-cell-autonomous manner, it does not consistently affect fertility ^{8, 16}. Therefore, we sought to determine if mitochondrial dysfunction in GABAergic neurons similarly has no impact on reproduction. C. elegans exists as a hermaphrodite and produces a limited number of sperm in the L4 stage. In the adult stage, it undergoes oocyte development and self-fertilizes to produce embryos (Figure 3A) ^{67, 68}. While isp-1 mutant has a prolonged lifespan, it has been shown to have a reduced brood size ⁶⁹. Interestingly, isp-1 gRNAi was sufficient to reduce the number of fertile animals (Figure 3B). The total number of embryos produced by fertile isp-1(gRNAi) animals was also significantly decreased (P < 0.0001) (Figure 3C). When we assessed reproductive activity daily, fertile isp-1(gRNAi) animals displayed a reproductive period similar to that of the control group. However, brood sizes from the 2-day-old stage to the 4-day-old stage were significantly reduced (P<0.005 for the 2 and 3-day-old stages and P<0.0001 for the 4-day-old stage) (Figure 3D). Intrigued by these results, we evaluated the impact of isp-1(gRNAi) on the three critical stages of germline development including mitotic germ cell proliferation, meiotic germ cell apoptosis, and oogenesis (Figure 3A). Although the length of the mitotic area was not significantly affected by isp-1(gRNAi) (Figures 3E and S3A), the total number of mitotic germ cells was decreased during the early reproductive periods (1 to 3 days after the L4 stage) (Figure 3F). The number of apoptotic germ cells in the meiotic gonadal loop region, undergoing germline apoptosis, was not different from that in control animals at the 2-day-old adult stage, but it was substantially decreased at the 3-day-old stage (Figure 3G) ⁷⁰. Thus, an increased loss of meiotic germ cells was not likely the primary cause of the reduced brood size. Interestingly, 2-day-old adult isp-1(gRNAi) animals displayed large DNA aggregates in the proximal gonad, a characteristic feature of an endomitotic oocyte phenotype (Figures 3H and 3I)⁷¹. Once embryos were produced, most of them were hatched in isp-1(gRNAi) animals (Figure S3B). Our results indicated that disruption of the mitochondrial ETC in GABAergic neurons negatively affected germline development and reproductive ability.

Figure 3.

Mitochondrial stress in GABAergic neurons enhances mitochondrial function in the peripheral tissues

Accumulated evidence indicates that aging is linked to a decline in mitochondrial function and biogenesis ^72–75. Additionally, experimentally increasing mitochondrial membrane potential and biogenesis is associated with lifespan extension ^{75, 76}. Therefore, we hypothesized that the mitochondrial stress specific to GABAergic neurons could impact the function and homeostasis of mitochondria in other peripheral tissues. At the 2-day-old stage, staining animals with the mitochondrial membrane potential-dependent MitoTracker Red CMXRos dye revealed a higher mitochondrial membrane potential in both the whole body (Figures 4A and 4B) and the intestine (Figure 4C) of isp-1(gRNAi) and spg-7(gRNAi) animals compared to that in the control group ^77,78. Additionally, whole-body extracts from isp-1(gRNAi) animals showed a marked increase in ATP levels (Figure 4D). spg-7(gRNAi) animals showed somewhat higher ATP levels compared to the control group, but it was not significant (Figure 4D). Next, we stained isp-1(gRNAi) and spg-7(gRNAi) animals using the MitoTracker FM Green dye that accumulates in mitochondria in a membrane potential-independent manner, indicating the mass of mitochondria ⁷⁹. We observed that both isp-1(gRNAi) and spg-7(gRNAi) animals exhibited higher MitoTracker FM Green fluorescence in the whole body than the control animals, suggesting an increase in mitochondrial mass (Figures 4E and 4F). In agreement with these results, the copy number of mitochondrial DNA was increased in isp-1(gRNAi) and spg-7(gRNAi) animals compared to control animals (Figure 4G). The expression of mitochondrial DNA polymerase gamma polg-1 was also significantly upregulated in isp-1(gRNAi) animals (Figure 4H) ^{35, 80, 81}. Notably, staining isp-1(gRNAi) and spg-7(gRNAi) animals with the ROS indicator DCF-DA showed lower ROS levels compared to the control animals (Figures 4I-4K) ⁸². Altogether, these results suggest that disturbances in mitochondrial homeostasis within GABAergic neurons could systemically increase overall mitochondrial membrane potential, ATP levels, and mitochondrial mass while decreasing ROS levels.

Figure 4.

GABAergic neuronal Mitochondrial Stress Enhances the DAF-16/FoxO Pathway

Despite the increased total mitochondrial mass and activity, the decreased ROS levels in isp-1(gRNAi) and spg-7(gRNAi) animals suggest the possibility that their ability to mitigate oxidative stress has improved. Therefore, we analyzed changes in stress response regulators associated with mitochondria and ROS, including DAF-16/FoxO, mitoUPR, and SKN-1/Nrf ^{21, 57, 83–85}. It has been demonstrated that mutations in isp-1 and RNAi targeting isp-1 (isp-1 sRNAi) result in enhanced expression of reporter genes associated with the mitoUPR pathway, such as hsp-6 or hsp-60 ^{4, 34, 57}. In line with this, when the expression of Pgaba::isp-1 dsRNA was induced in N2 wild-type worms, leading to systemic isp-1 RNAi, a significant increase in Phsp-6::GFP expression was observed (Figures S1A and S1B). Additionally, we observed that sRNAi against isp-1 effectively triggered the expression of Phsp-6::GFP, and this induction was dependent on atfs-1, a crucial mediator of the mitoUPR pathway ⁸⁶ (Figure S4A). These findings indicate the efficiency of our RNAi feeding conditions. However, isp-1 gRNAi did not significantly elevate the mRNA levels of hsp-60 and hsp-6, as evaluated by quantitative reverse transcription PCR (RT-qPCR) analysis (Figure 5A). The mRNA level of gst-4, a downstream target of the SKN-1/Nrf pathway, was also not significantly increased by isp-1(gRNAi) (Figure 5A) ^{87, 88}. In contrast, the mRNA levels of three DAF-16/FoxO downstream genes, including sod-3, hsp-16.2, and dlk-1, were substantially increased by isp-1(gRNAi) (Figure 5A) ^89–91. Additionally, sid-1(eq9) null mutants expressing isp-1 dsRNA in GABAergic neurons also showed increased expression of DAF-16/FoxO target genes (Figure 5B). The intensity of a fluorescent reporter for the DAF-16/FoxO pathway (muIs84 [Psod-3::gfp]) was also enhanced in this condition (Figures 5C and 5D). The elevated expression of sod-3 and dlk-1 induced by isp-1(gRNAi) was suppressed by the loss of DAF-16, suggesting that their expression is mediated by the DAF-16/FoxO pathway (Figure 5E).

Figure 5.

The non-cell-autonomous effects of GABAergic neuronal mitochondrial defects require DAF-16/FoxO

Further investigations were conducted to evaluate the functional involvement of DAF-16/FoxO in the non-cell-autonomous effects of mitochondrial stress in GABAergic neurons. It was found that sRNAi against isp-1 required DAF-16/FoxO function to extend lifespan, as previously reported (Figure S4B and Table S1) ^{20, 21}. DAF-16/FoxO was also necessary for the enhanced tolerance against paraquat in isp-1(sRNAi) worms (Figure S4C). Notably, the lifespan extension by isp-1 gRNAi was suppressed in daf-16(mgDf47) null mutants (Figure 6A). Additionally, DAF-16/FoxO was required for the increased stress tolerance against thermal (Figure 6B) and paraquat (Figure 6C) stresses in isp-1(gRNAi) animals. DAF-16 loss also suppressed the upregulation of mitochondrial membrane potential (Figures 6D and 6E) and mitochondrial mass (Figures 6F and 6G) induced by isp-1 gRNAi. Moreover, gRNAi against isp-1 in the daf-16(mgDf47) null mutant background failed to increase mitochondrial DNA copy number (Figure 6H) and polg-1 mRNA levels (Figure 6I). Finally, the daf-16 null mutation suppressed the abnormalities in daily reproductive ability (Figure 6J) and total brood size (Figure 6K) caused by isp-1 gRNAi. Interestingly, the double RNAi knockdown of isp-1 and daf-16 in GABAergic neurons (isp-1+daf-16 gRNAi) also resulted in a normal lifespan, to a level similar to that displayed in EV(gRNAi) and daf-16(gRNAi) conditions, suggesting that DAF-16 function in GABAergic neurons is required to mediate the lifespan extension (Figure S4D and Table S1). These findings collectively suggest that DAF-16/FoxO plays a critical role in mediating non-cell autonomous changes resulting from GABAergic neuronal mitochondrial stress, influencing organismal lifespan, stress tolerance, mitochondrial homeostasis, and reproductive capacity.

Figure 6.

Mitochondrial stress in GABAergic neurons causes non-cell-autonomous changes by acting on the same mechanisms as GABA signaling

Recent studies in C. elegans have revealed that GABA signaling plays a role not only in the regulation of GABAergic neuronal function but also in governing organismal longevity. Specifically, loss of unc-25, which encodes glutamic acid decarboxylase, has been shown to prolong lifespan and enhance stress tolerance ^{42, 43}. Hence, we conducted epistatic tests to investigate the possible interaction between the loss of GABA signaling and isp-1 knockdown conditions in GABAergic neurons and their impact on organismal health and longevity. We found that targeting isp-1 specifically in GABAergic neurons by expressing Pgaba::isp-1 dsRNAs in sid-1(qt9) mutants did not lead to a further extension of lifespan in unc-25 null mutants (Figures 7A and S5A; Table S1). Additionally, GABAergic neuronal mitochondrial stress did not lead to an additive increase in stress tolerance beyond the levels observed in unc-25(e156); sid-1(qt9) mutants (Figures 7B and 7C). The enhanced GFP expression driven by the sod-3 promoter due to isp-1 dsRNA expression in GABAergic neurons was not further increased in the unc-25 null mutant condition (Figure 7D).

Figure 7.

Next, we assess if GABAergic neuronal mitochondrial stress affects GABA function by testing the sensitivity of animals to aldicarb, an acetylcholinesterase inhibitor that can cause post-synaptic receptor hyperstimulation and paralysis due to reduced acetylcholine breakdown ⁹². As previously reported, depletion of GABA in unc-25 mutants increased the sensitivity to aldicarb ^92–94 (Figures 7E). GABAergic neuronal expression of isp-1 dsRNA also significantly elevated the sensitivity to aldicarb, particularly when animals were exposed to it for a prolonged period (120 min). The expression of Pgaba::isp-1 dsRNA did not further increase the hypersensitivity to aldicarb in unc-25 mutants, indicating that they function in the same pathway. Together, these findings suggest that diminished GABA signaling and mitochondrial stress within GABAergic neurons trigger alterations in organismal lifespan and healthspan through a common pathway.

Neuropeptide signaling in GABAergic neurons regulates organismal aging and health without additive effects with mitochondrial stress in GABAergic neurons

Next, we tested the potential roles of neuropeptide signaling in the non-cell autonomous effects of GABA neuronal mitochondrial stress ³³. We measured the lifespan of mutants with mutations in unc-31, required for dense-core vesicle exocytosis ⁹⁵. In agreement with the previous report, unc-31(e928) mutants exhibited prolonged lifespan (Figure 8A and Table S1). There is no additive lifespan increase by Pgaba::isp-1 dsRNA expression in the unc-31 mutant background, suggesting the potential involvement of neuropeptide signaling in regulating lifespan (Figure 8A and Table S1). GABAergic neurons have been suggested to express several neuropeptides, including flp-10, flp-11, flp-13, and flp-22 ^{96, 97}. In line with previous studies, flp-13 was expressed in a subset of C. elegans GABAergic motor neurons (Figure S6A) ^{97, 98}. We found that flp-13+EV gRNAi extended the lifespan of animals compared to control animals, indicating its non-cell autonomous function in GABAergic neurons in regulating organismal lifespan (Figure 8B); Table S1). Double gRNAi against flp-13 and isp-1 did not further increase lifespan compared to flp-13+EV gRNAi alone, suggesting that depletion of FLP-13 and mitochondrial disruption in GABA neurons could extend lifespan through a common mechanism (Figure 8B and Table S1). While gRNAi against flp-13+EV did not affect the tolerance of animals against heat stress, it improved stress tolerance against paraquat to a level comparable to that observed in animals treated with flp-13+isp-1 double gRNAi (Figures 8C and 8D). Maximal treatment of flp-13 gRNAi, without mixing with EV bacteria, did not significantly increase the heat tolerance compared to the control group, suggesting that the lack of impact of flp-13+EV gRNAi on heat stress is unlikely to be attributed to a diluted RNAi efficiency due to the double RNAi method (Figure S6B). Additionally, maximal treatment of flp-13 gRNAi still showed enhanced paraquat tolerance and lifespan (Figures S6C and S6D; Table S1). Collectively, these findings indicate a novel role for the neuropeptide FLP-13 in regulating organismal aging and health. This regulation appears to be mediated by a common mechanism associated with non-cell autonomous effects resulting from GABA neuronal mitochondrial dysfunction. The selective response of flp-13 mutants to specific stressors implies the existence of additional mechanisms that respond to mitochondrial stress within GABAergic neurons.

Figure 8.

Discussion

Our findings suggest that GABAergic neurons play a critical role in sensing mitochondrial stress and regulating longevity. Additionally, disrupting the mitochondria in GABAergic neurons alone was sufficient to induce alterations in organismal health and reproduction. Enhanced DAF-16/FoxO activity is necessary for mediating the non-cell autonomous changes observed in animals. Furthermore, we identified GABA signaling and one neuropeptide signaling as a factor associated with the non-cell autonomous effects caused by mitochondrial stress in GABAergic neurons.

Previous studies have demonstrated that mitochondrial stress in all neurons is sufficient to extend organismal lifespan. However, it remains incompletely understood which neurons respond to their mitochondrial stress to mediate this non-cell autonomous effects on organismal lifespan. Sha and colleagues have tested 18 neurons out of the total 302 in C. elegans regarding their potential to activate non-cell-autonomous effects and demonstrated that a specific subset, consisting of ASK, AWA, AWC, and AIA neurons, exhibits the capacity to induce non-cell-autonomous activation of the mitoUPR pathway in response to mitochondrial stress their mitochondria perturbation ³³. Interestingly, these conditions do not alter organismal lifespan, suggesting an unlink between mitoUPR and lifespan in certain conditions ³³.

The specific role of GABAergic neurons in non-cell autonomous regulation of aging and health in response to mitochondrial dysfunction has not been reported. This could be because these GABAergic neurons have not been tested or that previous studies primarily focused on neurons that non-cell autonomously affect the mitoUPR pathway in the peripheral tissues ^{16, 33–36}. However, a recent study reveals that induced ROS production in GABAergic neurons can trigger mitoUPR induction in other peripheral tissues through activating UNC-49/GABA_A receptor signaling, suggesting a central role of GABAergic neurons in regulating organismal stress response ⁴⁵. In our study, we prioritized monitoring changes in the lifespan rather than detecting mitoUPR reporter activity and found that inducing mitochondrial stress, specifically in GABAergic neurons, was sufficient to extend the lifespan of the animal. We employed two independent RNAi strategies to knock down two genes related to mitochondria. The tissue-specific RNAi approach, utilizing the rde-1(ne219) null mutant background and exclusively restoring rde-1 in target tissues, has been successfully utilized in multiple studies ^{48, 50, 51, 53}. To eliminate the possibility of incomplete gRNAi, we also conducted isp-1 knockdown in GABAergic neurons by expressing isp-1 dsRNAs in sid-1 null mutant backgrounds, and we consistently observed an extension in lifespan in both RNAi strategies. isp-1 mutants and Pgaba::isp-1 dsRNA expression in wildtype animals robustly increase mitoUPR. In contrast, our qPCR results showed that there was no increase in mitoUPR reporter expression in isp-1(gRNAi) animals ⁴. Additionally, the isp-1(gRNAi) animals did not fully recapitulate the phenotypes of isp-1 mutants, including normal ATP level, increased ROS level, and decreased mitochondrial membrane potential ^99–102, supporting that our two isp-1 knockdown conditions did not occur systemically other than in GABAergic neurons.

Longevity is associated with healthspan, but not in all cases ¹⁰³. Our results show that mitochondrial stress in GABAergic neurons not only prolonged lifespan but also improved stress tolerance, a typical healthspan parameter. Notably, we also found enhancements in mitochondria activity and ATP levels. These improvements in mitochondria could be, in part, a result of an increase in mitochondria population evidenced by increases in mitochondrial DNA copy number and mass. Consistently this notion, we also found increased mRNA levels of polg-1, encoding the mitochondrial DNA polymerase that is responsible for the replication of the mitochondrial genome, suggesting a potential increase in mitochondria biogenesis by GABAergic neuronal mitochondrial stress ^{35, 104}. The concept of aging in model organisms has long been linked to a decrease in mitochondrial function and biogenesis ¹. Similarly, in humans, mitochondrial function declines with age ^{2, 105}. Enhancing mitochondrial function has been proposed as an intervention strategy against aging. In mice, the brain-specific overexpression of Sirt1, a pivotal regulator of mitochondrial biogenesis through PGC-1a, not only extended lifespan but also induced non-cell-autonomous effects in skeletal muscle including improved mitochondria homeostasis ⁷⁵. A recent study indicates that experimentally rejuvenated mitochondrial membrane potential is sufficient to extend C. elegans lifespan ⁷⁶. Additionally, while isp-1 mutants have been shown to have elevated levels of ROS, a central factor in aging, our findings suggest that knockdown of isp-1 in GABAergic neurons leads to a reduction in ROS levels, despite an increased mitochondrial population and heightened activity ^{6, 102, 106}. Therefore, both enhanced mitochondrial homeostasis and reduced ROS levels could contribute to the non-cell-autonomous enhancement in aging and health.

The reduced ROS also proposed the potential involvement of the stress response pathway against ROS. Strikingly, our qPCR results showed no evidence supporting the activation of mitoUPR by GABAergic-neuronal mitochondrial stress ^{16, 33–36}. It has been shown that afts-1 acts cell-autonomously in certain neurons to mediate non-cell autonomous effects of mitochondrial stress in specific neurons ³³. Thus, it is still possible that mitoUPR activation in GABAergic neurons plays a role in mediating lifespan extension and improving stress resistance. Recent studies have demonstrated that experimentally enhanced ROS production in GABAergic neurons activates the mitoUPR and this requires UNC-49 GABA_A receptor ⁴⁵. However, unc-49 mutants exhibit a normal lifespan ^{42, 43}. Therefore, the extended lifespan resulting from mitochondrial defects in GABAergic neurons is not primarily mediated by increased ROS production and non-cell-autonomous activation of mitoUPR.

Our results suggested that GABAergic neuronal mitochondrial stress requires DAF-16/FoxO. We found that GABA-neuronal mitochondrial stress promoted the expression of sod-3, hsp-16.2, and dlk-1, which are regulated by DAF-16 ^89–91. A recent study also reveals that ROS induction in GABAergic neurons results in non-cell autonomously increased expression of sod-3 along with hsp-6 ⁴⁵. DAF-16 is another pathway suggested to be activated by mitochondrial dysfunction, contributing to the extended lifespan observed in C. elegans with whole-body mitochondrial dysfunction ²¹. However, another study reports that while DAF-16 is required for the longevity of mitochondrial mutants in certain conditions, it was not sufficient to fully account for the observed lifespan extension ⁶. We found that depletion of DAF-16 suppressed a series of non-cell-autonomous changes in lifespan, stress resistance, mitochondrial DNA (mtDNA) copy number, and polg-1 expression. These results indicate the functional importance of DAF-16 in mediating the non-cell autonomous effects of mitochondrial disruption in GABAergic neurons.

Recent studies on C. elegans have reported that GABA signaling regulates lifespan and health ^{42, 43}. Notably, GABA loss increases DAF-16/FoxO activity in the intestine, thereby increasing organismal longevity and health span ^{42, 43}. Additionally, GABA signaling modulates protein homeostasis in C. elegans post-synaptic muscle cells ⁴⁴. In contrast to the role of GABA signaling in C. elegans lifespan, knockdown of the Drosophila GABA_B receptor shorts lifespan ¹⁰⁷. Nevertheless, these results indicate that GABA signals could have a conserved role in regulating organismal health and aging. Our aldicarb assay results suggest that GABAergic neuronal mitochondrial stress could interfere with GABAergic neuronal activity. It has been revealed that mitochondria abundantly accumulate at the presynapse and affect synaptic activity by regulating the supply of ATP and calcium homeostasis, which are required for proper neurotransmitter release and recycling ^108–113. Our previous studies have demonstrated that over 75% of mitochondria localize at the presynapse in GABAergic motor neurons in C. elegans ¹¹⁴. A screen for essential genes required for GABAergic neuronal function, utilizing a GABAergic neuron-specific feeding RNAi in C. elegans, has identified several mitochondrial-related genes. Knockdown of these genes in GABAergic neurons triggers hypersensitivity against aldicarb ⁴⁸. Therefore, GABAergic neuronal mitochondrial perturbation could mimic effects induced by a reduction in GABA signaling. This hypothesis is supported by our epistatic analysis results, suggesting that GABAergic neuronal mitochondrial stress and the loss of GABA signaling in unc-25 mutants may share a common mechanism influencing longevity, stress tolerance, and sod-3 expression. Understanding how mitochondrial stress can modulate downstream mediators such as DAF-16 through altering GABA singling and ultimately affecting longevity and health can be complex. Mitochondrial stress in GABAergic neurons may partially reduce GABA signaling to varying degrees rather than completely turning it off ^{112, 115}. Moreover, GABA signaling differentially regulates lifespan and each healthspan parameter through three receptors and a combination of four downstream pathways ⁴³. Thus, reduced GABA signaling caused by GABA mitochondrial stress could affect each downstream receptor pathway to varying degrees, depending on the ability of each receptor to respond to GABA ligands.

It has been well-documented that cell-autonomous responses to mitochondrial defects can vary depending on the stressor conditions, including disruptions in each ETC component and disruption modes, pathological conditions, and environmental stressors ^{101, 116}. These variations highlight the complex responses to mitochondrial disturbances. Interestingly, the non-cell autonomous longevity changes also differ depending on the type of mitochondrial perturbation. It has been reported that inhibition of spg-7 and cco-1, encoding a cytochrome c oxidase-1, in the entire nervous system prolongs lifespan ^{16, 33}, but it was not suppressed by mitoUPR loss. Mitochondrial stress induced by pan-neuronal expression of polyglutamine repeats (polyQ40) and mutation in ucp-4, encoding a mitochondrial uncoupling protein-4, do not further prolong the lifespan than wild-type animals, but mitoUPR function is required to preserve normal lifespan ³³. Also, the deletion of spg-7 in AIY neurons is sufficient to induce mitoUPR in distal tissues but does not prolong lifespan ^7,32,35. Additionally, pan-neuronal expression of KillerRed, which induces oxidative stress, resulted in a reduced lifespan, despite an enhanced mitoUPR pathway in remote tissues ³³. In this study, we targeted to inhibition of isp-1 and spg-7 in GABAergic neurons and observed consistent changes in lifespan and stress resistance. Notably, isp-1 gRNAi resulted in decreased reproductive ability. Previous studies report that depletion of CCO-1, a component of mitochondria ETC, in the nervous system results in mitoUPR activation in the peripheral tissues and prolongs lifespan without affecting brood size ^{16, 33}. Similarly, in Drosophila, ETC reduction in the nervous system increases longevity but maintains normal fertility ⁸. However, it remains unclear whether other mitochondrial stressors in GABAergic neurons can produce similar non-cell-autonomous effects as seen with isp-1 and spg-7 knockdown, which could be one reason the role of GABAergic neurons in regulating the non-cell-autonomous effects of mitochondrial disturbance has not been elucidated previously.

Finally, studies have demonstrated that neuropeptide signaling mediates the non-cell autonomous regulation of mitoUPR and organismal aging in response to mitochondrial stress in specific neurons ^{42, 117–122}. For instance, mitochondrial stress in ASK, AWA, AWC, and AIA neurons utilizes the FLP-2 neuropeptide to induce mitoUPR in peripheral tissues ³³. These results suggest that there could be additional signaling molecules mediating the non-cell autonomous effects of GABAergic neuronal mitochondrial stress. GABAergic neurons have been found to express several neuropeptides, including flp-10, flp-11, flp-13, and flp-22 ^{96, 97}. Our finding indicates that gRNAi against unc-31 increased lifespan resulting from depletion of dense core vesicle exocytosis was not further increased by GABA-neuronal mitochondrial stress, suggesting that neuropeptide signaling could be involved in non-autonomous changes caused by GABA-neuronal mitochondrial damage ⁹⁵. FLP-13 loss increased lifespan and stress resistance, which is similar to the phenotype of animals with mitochondrial perturbations in GABAergic neurons. Animals with both conditions did show significant changes in stress resistance and lifespan compared to animals with either single condition, indicating that they work through the same mechanism to mediate non-cell autonomous effects. Therefore, it is possible that GABAergic neuronal mitochondrial stress reduces FLP-13 function, which inhibits lifespan and healthspan. Future studies are needed to elucidate the specific role of FLP-13 and other neuropeptides in the non-cell-autonomous effects of GABA-neuronal mitochondrial stress.

Acknowledgements

We thank Drs. Shinichi Someya, Christiaan Leeuwenburgh, Stephanie Wohlgemuth, and Sejin Lee for their valuable discussions. Some strains were provided by the CGC, funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). This research was conducted while SMH was a Hevolution/AFAR New Investigator Awardee in Aging Biology and Geroscience Research (AWD16056). SMH was also supported by the National Institute on Aging (NIA) of the National Institutes of Health (NIH) under award numbers R56AG066654 and R01AG081270, as well as the Glenn Foundation for Medical Research and the American Federation for Aging Research under award numbers AWD06577 and AWD11009 to SMH. Additionally, support was provided by NIA/NIH under award number AG060373-01 and the National Science Foundation under award number IOS (2132286) to MHL.

Author Contributions

S.M.H. supervised the study. M.H.L. and S.M.H. conceived the study. L.R. and S.M.H. wrote the manuscript with input from all authors. L.R. and S.C. designed and performed experiments and analyzed and interpreted data. Y. P. and M.H.L designed germline experiments and interpreted data. B.R. and T.M. assisted with planning and performing the stress response and lifespan assays. Y.S. and R.X generated reagents and assisted with planning and interpreting data. K.M., W.C., and R.X. provided expertise and resources and reviewed the manuscript.

Declaration of Interests

The authors declare no competing interests.

Inclusion and diversity

We support inclusive, diverse, and equitable conduct of research.

Data availability

All numerical data supporting the findings of this study are available in the supplemental information and from the corresponding author upon reasonable request.

Material and Methods

C. elegans strains

All C. elegans strains were maintained at 20 °C on nematode growth medium (NGM) plates seeded with the OP50 strain of Escherichia coli (E. coli) as described before ¹²³. A list of strains used in this study is the following: N2 (Wild type), XE1375 (wpIs36 I; wpSi1 II; eri-1(mg366) IV; rde-1(ne219) V; lin-15B(n744) X), SJ4100 (zcIs13 [hsp-6p::gfp]), CB156 (unc-25(e156) III), HC196 (sid-1(qt9) V), GR1352 (daf-16(mgDf47)I), TJ356 (zIs356 IV [daf-16p::daf-16a/b::gfp; rol-6(su1006)]), CF1553 (muIs84 [(pAD76) sod-3p::gfp + rol-6(su1006)]), CL2166 (dvIs19 III [(pAF15)gst-4p::gfp::NLS]). HAN260 (daf-16(mgDf47) I; wpIs36 I; wpSi1 II; eri-1(mg366) IV; rde-1(ne219) V; lin-15B(n744) X), HAN186 (unc-25(e156); sid-1(qt9)), HAN356 (unc-25(e156); sid-1(qt9)+ sbsEx27 [unc-47p::isp-1 sense+ unc-47p::isp-1 antisense]), HAN357 (sid-1(qt9)+sbsEx27 [unc-47p::isp-1 sense+ unc-47p::isp-1 antisense]).

RNAi assay

RNAi was performed by the feeding method ^{124, 125}. Briefly, freshly streaked single colonies of HT115(DE3) bacteria containing either empty L4440 vector (control) or isp-1 and spg-7 RNAi plasmid were grown overnight at 37°C in Luria broth (LB) medium supplemented with carbenicillin (25 μg/ml). HT115(DE3) bacterial feeding strains were obtained from the genome-wide library ¹²⁵. PCR and sequencing were used to confirm that strains contained the correct clones. RNAi bacteria were seeded on NGM plates containing IPTG (1 mM) and cultured overnight before transferring animals. To prevent undesired non-specific effects, we did not use 5-fluoro-2′-deoxyuridine (FUdR) in seeded RNAi plates. Duplex RNAi was performed by mixing two HT115(DE3) bacterial strains, each containing the desired RNAi plasmid or empty vector in a 1:1 volume ratio. The efficiency of double RNAi, which involves feeding a 1:1 volume ratio mixture of bacteria strains targeting two different genes, was compared to that of single gene RNAi, which involves feeding a 1:1 volume ratio mixture of bacteria strains targeting one gene and an empty vector.

Lifespan assay

Age-synchronized animals were prepared by egg prep from adult animals cultured on each RNAi or regular NGM plate at 20°C ¹²⁶. The isolated embryos were transferred and allowed to hatch on RNAi NGM plates for the desired gene or regular NGM plates. Then, approximately 50 animals at the L4 stage were transferred to fresh RNAi plates. Every day, animals that failed to respond to gentle prodding with platinum wire were scored as dead. Lifespan data were statistically analyzed for significance by the log-rank test, comparing survival curves using GraphPad Prism software. Lifespan assays were performed at least in triplicate. To prevent undesired non-specific effects, we did not use 5-fluoro-2′-deoxyuridine (FUdR) in the seeded RNAi plates.

Aldicarb sensitivity assay

Aldicarb sensitivity assay was performed as described previously ⁹². Briefly, approximately 40-50 mutant animals were grown on OP50-seeded NGM plates at 20 °C for 72 hours. Approximately 40 L4 stage mutant animals were then picked and placed on 35 mm OP50-seeded NGM plates containing 0.5 mM aldicarb (Sigma Aldrich, 33386: prepared aldicarb NGM plates 1 day before the assay) and scored for paralysis every 30 minutes over a 120-minute period. Animals were considered paralyzed when they failed to show any movements in response to touching at a fixed time. This assay was carried out in triplicate for each experimental condition.

Reproductive assay

After pre-exposure to spg-7 and isp-1 RNAi from the L1 stage, age-synchronized XE1375 animals at the L4 stage (n=∼20-30) were individually transferred to fresh spg-7 and isp-1 RNAi NGM plates. Every day, we moved mother animals to new spg-7 and isp-1 RNAi NGM plates until egg-laying stopped. The embryos produced daily were counted, and the total number of produced embryos was used to calculate the brood size. The egg production period was used to calculate the reproductive span, and the number of embryos produced each day was used to analyze reproductive trends. The hatching rate was scored 24 hours after egg-laying. The brood size, reproductive span, hatching rate, and reproductive trend data were generated from independent experiments. At least three replicate experiments were performed for each assay.

Stress assays

For the paraquat assay, 30-50 age-synchronized animals were pretreated to RNAi for each gene from the L1 stage. Then, 2-day-old adult animals were transferred to 96-well plates containing M9 with paraquat, a reactive oxygen species generator (Sigma-Aldrich, 36541-100MG), at a concentration of 0 mM, 50 mM, 100 mM, and 150 mM in a total volume of 100 µl per well. The survival of animals was evaluated after 24 hours, and animals that failed to respond to platinum wire touch were scored as dead. In the heat stress resistance assay, age-synchronized adult animals pretreated with isp-1 or spg-7 RNAi were exposed to thermal stress at 37°C for 5 hours. The survival after heat shock was recorded every hour for 5 hours by gently prodding with a platinum loop. Animals that failed to respond were scored as dead. All experiment was independently repeated at least three times.

Staining assays for ROS level and mitochondrial homeostasis

We used the fluorescent probe H₂DCF-DA dye (Invitrogen, D399) to detect ROS levels in vivo, as previously described ⁸². 2-day-old adult stage mutants or XE1375 animals pretreated with spg-7, isp-1, or control EV RNAi from the L1 stage were washed with M9 buffer to remove the bacteria. After washing 3 times, animals were collected in 300 µl of PTw buffer (1xPBS with 0.1% tween20). Then, around 30-35 animals were transferred to 96-well plates containing 10 mM H₂DCF-DA. The fluorescence was recorded using Spectra Max M2 multimode microplate reader (Molecular Devices) at 485 nm excitation and 530 nm emission. The change in fluorescence was recorded for 120 min at 20 min intervals at 37 °C. The experiment was performed 3 times independently. To perform the H₂DCF-DA assay using an image, we transferred 30-40 pretreated spg-7 and isp-1 RNAi animals in 500 µl M9 buffer and washed 3 times. Afterward, animals were incubated in 500 µl M9 buffer containing 10 mM H₂DCF-DA for 1 hour. Animals were washed with M9 buffer at least 3 times, transferred to 2% agarose pads on glass slides, and visualized using a GFP filter and imaged using a 40x objective (Andor, DSD2 spinning disk confocal, Andor, Zyra4.2 camera). To evaluate mitochondrial membrane potential and mass, we used MitoTracker CMXRos (Invitrogen, M7512) and FM green dye (Invitrogen, M7514), respectively, as described in previous studies ^{127, 128}. The lyophilized dyes were dissolved in DMSO and made to the final concentration of 100 μM. Animals were pretreated from the L1 stage to spg-7 and isp-1 gRNAi. L4 stage animals (n=30-40) were transferred to spg-7 and isp-1 RNAi plates with MitoTracker CMXRos or MitoTracker FM green dye and incubated for 48 hours at room temperature under dark conditions. Next, stained animals were transferred to fresh NGM plates (without the dye) for 1 hour to remove the bacterial fluorescent background inside the gut. The animals were observed using a 10x magnification to visualize the whole-body staining and a 40x objective to observe the anterior body, including the intestine.

ATP assay

We used previously reported protocols with a minor modification ⁵². Approximately 150 age-synchronized animals treated with spg-7 and isp-1 RNAi from the L1 stage were prepared.1-day-old adult animals were washed 5 times with M9 buffer to remove the intestinal bacteria and washed with TE buffer. The animals were frozen at −80 °C. The sample was sonicated for 15 seconds with a 15-second interval, followed by boiling for 10 minutes to release ATP and block ATPase activity. Debris was removed by centrifuging at 4 °C for 10 minutes. The supernatant was collected, and the ATP levels were measured using the bioluminescence detection kit (Promega ENLITEN ATP Assay System, FF2000) and Spectra Max M2 multimode microplate reader (Molecular Devices). Luminescence was normalized to protein content measured with a Pierce BCA protein determination kit (Thermo Scientific, 23227).

mtDNA quantification

mtDNA quantification was performed using a quantitative PCR (qPCR) based method ¹²⁹. About 30 age-synchronized L4 staged spg-7(gRNAi) and isp-1(gRNAi) animals were collected in 30 µl of lysis buffer (freshly added proteinase K) and frozen immediately at −80 °C for 10 minutes before lysis at 65 °C for 1 hour, followed by 95°C for 15 minutes, and then maintained at 4 °C. Relative quantification was used for determining the fold changes in mtDNA. 2 µl of lysate sample was used as template DNA in each triplicate reaction. qPCR was performed using the SYBR green mixture in a CFX96 Touch Real-Time PCR System (Bio-Rad). Primers (mtDNA target-specific primer) for cox-4 and nd-1 were used to determine the mtDNA copy number. The cox-4 forward primer 5’GCCGACTGGAAGAACTTGTC-3’ and reverse primer 5’-GCGGAGATCACCTTCCAGTA-3’. The nd-1 forward primer 5’-AGCGTCATTTATTGGGAAGAAGAC-3’ and reverse primer 5’-AAGCTTGTGCTAATCCCATAAATGT-3’. All qPCR results were performed in triplicates.

Quantitative reverse transcriptase PCR

RNA isolation and quantitative reverse transcriptase PCR (qRT-PCR) analysis were performed as previously described ¹³⁰. Animals raised on spg-7 or isp-1 gRNAi plates were synchronized by egg prep. Then, total RNA was extracted using the TRIzol (Invitrogen, 15596026) method from age-synchronized animals (approximately 1,500 animals at the 3-day-old adult stage). RNA was purified using a Qiagen Rneasy kit, and 2 µg RNA was used for cDNA synthesis (Thermo Scientific, Verso cDNA synthesis kit, AB1453A). qPCR was performed using SYBR Green master mix (Bio-Rad Laboratories). qPCR primers are listed below. Act-1 Forward 5’-GCTGGACGTGATCTTACTGATTACC-3’, act-1 Reverse 5’-GTAGCAGAGCTTCTCCTTGATGTC-3’, daf-16 Forward 5’-CCAACACATTCATCCCAGTG-3’, daf-16 Reverse 5’-GATGGGATAGAGGTAGCATT-3’, sod-3 Forward 5’-CTGATGGACACTATTAAGCG-3’, sod-3 Reverse 5’-AAGTGGGACCATTCCTTCCAA-3’, gst-4 Forward 5’-GCTGAGCCAATCCGTAT-3’, gst-4 Reverse 5’-GTAAAATGGGAAGCTGGC-3’, dlk-1 Forward 5’-TCGACGCTATCTCCGAACTT-3’, dlk-1 Reverse 5’-TGCTTGATCTCGGTCTCCTT-3’, hsp-6 Forward 5’-CGAAAGCTATTTGGGAACCA-3’, hsp-6 Reverse 5’-GCTCGTTGATGACACGAAGA-3’, hsp-60 Forward 5’-CCGTCTCTGTCACTATGGGC-3’, hsp-60 Reverse 5’-CTCGAATCCCTCTTTGGCGA-3’, polg-1 Forward 5’-GTTACGGCCGACGAGATACG-3’, polg-1 Reverse 5’-CGTAGCTTCCGGACTCCAAA-3’. All qPCR results were performed at least in triplicates.

Statistics

Statistics were performed using GraphPad PRISM (version 9 and 10). No statistical method was used to pre-determine sample size. For Figures 5D and 7D, outliers were identified using the ROUT method at Q = 1% and excluded from further analysis. Within experimental groups, animals were randomized for each experimental replicate. The qPCR experiments were analyzed using the delta-delta Ct method ¹²¹. All experiments were reliably reproduced at least 3 times independently. For the qPCR assays, in cases where the control and subjects were not paired, the average control delta Ct value was utilized to calculate delta-delta Ct values. However, if the control and experimental groups were paired, individual control delta Ct values were used to calculate delta-delta Ct values for each paired experimental group.