Introduction

Maintaining a delicate balance between excitatory and inhibitory (E/I) neurotransmission is critical for optimal brain function ¹. Disruptions in this balance are commonly observed in neurodevelopmental disorders ^2–4. In particular, deficits in inhibitory (GABAergic) signaling have been reported in Autism Spectrum Disorders (ASD) and other related physiopathological conditions ^4,5.

PTEN is a classical tumor suppressor gene that antagonizes the highly conserved phosphatidylinositol 3-phosphate kinase (PI3K)/protein kinase B (PKB/Akt) pathway. Several reports using animal models have highlighted the importance of PTEN in neurodevelopment ^6–12. Moreover, mutations in PTEN were frequently found in human patients presenting ASD ¹³. The molecular events underlying the neurodevelopmental deficits in PTEN mutants remain poorly understood.

Ketogenic diets (KGDs), which are very high in fat and low in carbohydrates, upregulate beta-oxidation of fatty acids and the release of ketone bodies (KBs), which serve as an energy source for other cells. These diets have been shown to be effective in alleviating autistic symptoms ^14–16. It is mostly unknown how the KGDs exert these beneficial effects.

The C. elegans neuromuscular system, where both excitatory (cholinergic) and inhibitory (GABAergic) motor neurons regulate muscle contraction and relaxation, provides an excellent platform for studying the function, balance, and coordination between excitatory and inhibitory signals^17–23. Studies conducted within this system have yielded valuable insights into fundamental synaptic transmission mechanisms^24,25. Over the last decade, numerous studies focused on this simple yet highly informative system have significantly contributed to our understanding of the functioning and dysregulation of human genes associated with neurodevelopmental disorders, epilepsy, and familial hemiplegic migraine^17,22,23,26. Furthermore, the substantial conservation of the main components of the PI3K/Akt pathway in C. elegans^27,28, enhances the applicability of this model system for investigating the role of this pathway in neurodevelopment.

Here we find that mutations in daf-18 (ortholog for PTEN in C. elegans) lead to selective impairments in GABAergic inhibitory signaling due to reduced activity of the transcription factor DAF-16/FOXO during neurodevelopment. Surprisingly, cholinergic excitatory motor neurons remain unaffected. Even more significantly, we demonstrate that exposure of these animals to a diet enriched with the KB β-hydroxybutyrate (βHB) early during development increases DAF-16 activity, ameliorates morphological and functional defects in GABAergic neurons, and improves behavioral phenotypes. Besides providing a behavioral simple system to study the role of the conserved PI3K/Akt/FOXO pathway in neurodevelopment, this study constitutes a proof of concept of the ability of reverting a neurodevelopmental defect with a dietary intervention and contributes to understanding the mechanisms of action of KBs and their potential therapeutic use.

Results

Disturbances in C. elegans cholinergic or GABAergic activity can be detected by analyzing the sensitivity to the paralyzing effects of drugs that exacerbate cholinergic transmission ^17,18. We therefore analyzed the sensitivity of daf-18 deficient animals to the acetylcholinesterase inhibitor aldicarb and to the cholinergic agonist levamisole. Exposure to aldicarb leads to an increase in ACh levels at cholinergic motor synapses, resulting in massive activation of muscular cholinergic receptors and subsequent paralysis ²⁹. Levamisole also induces paralysis by directly activating muscular cholinergic receptors ²⁹. We found that daf-18 mutants are hypersensitive to the paralyzing effects of both drugs (Figures 1A, 1B and S1D).

Figure 1.

Reduced activity of DAF-18 has been largely shown to exacerbate the PI3K pathway, precluding the activation of DAF-16, the C. elegans ortholog of the FOXO transcription factors family ²⁸ (Figure S1A). We analyzed aldicarb and levamisole sensitivity of mutants in this transcription factor. Similar to daf-18 mutants, we found that daf-16 null mutants are hypersensitive to the paralyzing effects of aldicarb and levamisole (Figures 1A and 1B). Furthermore, we did not observe significant differences in aldicarb and levamisole sensitivity between daf-18; daf-16 double null mutants and the respective single mutants, suggesting that both genes affect neuromuscular signaling by acting in the same pathway (Figures 1A and 1B). In addition to daf-16 and daf-18, we assessed the sensitivity to the paralyzing effects of aldicarb and levamisole in mutants of other components of the PI3K pathway, such as age-1/PI3K, pdk-1, akt-1, and akt-2 (Figure S1E). We did not observe significant differences compared to the wild-type in most cases, except for the mutant containing the mg142 allele, which exhibits a gain-of-function mutation in pdk-1²⁷ (all other mutations lead to reduction or complete loss of function mutations in their respective genes). This gain of function mutant displays hypersensitivity to aldicarb and levamisole, similar to daf-16 and daf-18 mutants. Given that increased pdk-1 activity is linked to hyperphosphorylation and inactivation of DAF-16 (Figure S1A), these results support the hypothesis that low activity of DAF-16 leads to hypersensitivity to these drugs. In vertebrates, alterations in PTEN activity have been largely shown to impact neuronal development and function by affecting the mTOR pathway³⁰. Consequently, we conducted an analysis to determine whether mutations in components of the C. elegans TOR complexes (TORC) would lead to significant changes in sensitivity to aldicarb and levamisole. Our findings indicate that neither animals with loss of the essential TORC1 component raga-1/RagA nor animals with a loss of function in the essential TORC2 component rict-1/Rictor exhibited significant alterations in sensitivity to cholinergic drugs compared to wild-type animals (Figures S1F). This suggests that the mTOR pathway is not involved in daf-18 pharmacological phenotypes. Hypersensitivity to cholinergic drugs has long been observed in worms where there is a deficiency in GABAergic signaling ^17,18 (Figures 1A and 1B). In mutants with severe deficits in GABA transmission, prodding induces a bilateral contraction of the body wall muscles that shortens the body (shrinker phenotype) ³¹. When daf-18 mutants are touched, there is a slight but significant shortening in body length (Figure 1C). As expected, this shortening is not as noticeable as in animals with a complete deficit in GABAergic signaling, such as mutants in unc-25 (the C. elegans orthologue for glutamic acid decarboxylase) (Figure 1C). Similar to daf-18 mutants, daf-16 animals also exhibit a mild decrease in body length after prodding. Consistent with our aldicarb and levamisole results, there are no significant differences in body shortening between daf-18; daf-16 double mutants and the corresponding single mutants (Figure 1C), further supporting the notion that both genes act in the same pathway to impact neuromuscular signaling.

We also analyzed other behaviors that require a concerted activity of GABAergic and cholinergic systems, such as the omega turns during the escape response ³². In C. elegans the escape response can be induced by a gentle touch on the head and involves a backward movement that is usually followed by a sharp omega turn and a 180° change in its direction of locomotion ³² (Movie 1). The execution of the omega turn involves a hypercontraction of the ventral muscles and relaxation of the dorsal muscles (see Figure 1D and Movie 1), allowing the animal to make a sharp turn, where the animal’s head slides on the ventral side of the body (closed omega turn), and resume locomotion in the opposite direction. In response to anterior touch the vast majority of the wild-type worms make a closed omega turn^32,33. Ventral muscle contraction is triggered by cholinergic motor neurons (VA and VB neurons) that synapse onto ventral muscles, while dorsal muscle relaxation is induced by GABAergic motor neurons (DD neurons) that synapse onto dorsal muscles (Figure S1C) ^32,33. Ablation of DD GABAergic neurons reduces dorsal muscle relaxation, therefore preventing the head from touching the ventral side of the body during the escape response (open omega turn)³². In agreement with previous reports, we found that 93% of wild-type animals exert a closed omega turn within the escape response (Figure 1D). We observed that, similar to wild-type animals, gentle anterior touch with an eyelash induces daf-18 mutants to move backward and initiate an omega turn (Movie 2). However, only 54% of daf-18 mutants exhibit the typical head to tail contact during the omega turn (Figure 1D). Akin to daf-18 mutants, daf-16 mutants exhibited a decrease in the proportion of closed omega turns (Figure 1D). No additive effects were observed in the daf-18; daf-16 double mutant, suggesting that the increased inactivation of daf-16 is primarily responsible for the defects observed in the escape response of daf-18 mutants.

Given that our results suggest a deficit in GABAergic functionality in daf-18/PTEN mutants, we used optogenetics to specifically activate these neurons in mutant worms. The expression of Channelrhodopsin (ChR2) in GABAergic motor neurons (using the unc-47 promoter, orthologue for the vesicular GABA transporter SLC32A1) elicits a flaccid paralysis of the worms upon exposure to blue light. This obvious and robust response results in an increase in body length that can be used as a clear readout ^34–36(Figure 1E and Movie 3). Interestingly, we found that the elongation of the animal after the specific activation of GABAergic neurons is significantly decreased in daf-18 and daf-16 mutants compared to wild-type worms (Figure 1E). While these results suggest a defect in GABAergic transmission, it could also be possible that general neuronal transmission is affected. Consequently, we reciprocally activated the cholinergic motor neurons in animals expressing ChR2 under the unc-17 promoter, which leads to muscle contraction and shortened body length ^34,36(Figure 1F and Movie 4). Rather than observing reduced shortening in daf-16 and daf-18 mutants, we found that cholinergic activation caused hypercontraction of these mutant animals (Figure 1F). Since the activation of cholinergic motor neurons not only activates muscles but also stimulates GABAergic neurons to produce counteractive muscle relaxation in the other side of the animal (Figure S1C), it is expected that a GABAergic deficit would lead to increased muscle contraction and body shortening upon cholinergic activation. In summary, these results strongly suggest that in daf-18 and daf-16 mutants, there is a specific functional defect in GABAergic neurons, while excitatory neurons do not appear to be affected.

Since our previous results imply perturbations of neuromuscular transmission, we explored the morphology of C. elegans motor neurons. The cell bodies of both cholinergic (A and B-type) and GABAergic (D-type) motor neurons that innervate body wall muscles are located in the ventral nerve cord (VNC), and a subset extends single process commissures to the dorsal nerve cord (DNC) ³⁷ (Figures 2A, S1C and S2). These commissures allowed us to study defects in the development or maintenance of GABAergic and cholinergic neuronal processes ^38,39. We analyzed the morphology of GABAergic motor neurons in L4 animals expressing mCherry under the control of the unc-47 promoter ⁴⁰. We found that daf-18 mutants exhibit a higher frequency of commissure flaws, including guidance defects, ectopic branching, and commissures that fail to reach the dorsal cord (Figures 2B and 2C). GABAergic neuron anomalies were also evident at the first larval stage (L1), suggesting that reduced DAF-18/PTEN activity affects the neurodevelopment of the GABAergic system (Figures 2D and 2E). Since the transcription factor DAF-16/FOXO is one of the main targets of DAF-18/PTEN signaling, we analyzed the morphology of GABAergic D-type neurons in daf-16 null mutants. These animals also exhibit an increased number of defects in commissures compared to the wild-type (Figure 2C).

Figure 2.

Given that daf-18 is ubiquitously expressed in all tissues ²⁸, we asked whether DAF-18 acts autonomously in GABAergic neurons to ensure proper development. We found that specific daf-18 expression in GABAergic neurons increased the proportion of closed omega turns in daf-18 null mutants (Figure 2F). In addition, the morphological defects in GABAergic commissures were significantly reduced (Figure 2G), suggesting that DAF-18 acts autonomously in GABAergic motor neurons to regulate their development.

In contrast to our findings in GABAergic neurons, we observed no obvious differences in the frequency of commissure defects when we compared cholinergic motor neurons in control and daf-18 animals (Figure S2).

Collectively, our findings demonstrate that mutations in daf-18 and daf-16 result in developmental defects in GABAergic neurons, leading to both altered morphology and function, while leaving cholinergic motor neurons unaffected. Our experiments strongly suggest that these specific defects in the inhibitory transmission arises for the hyperactivation of the PI3K pathway, along with subsequent DAF-16 inhibition, in GABAergic neurons of daf-18/PTEN mutants.

Mutations in PTEN are linked with Autism Spectrum Disorders (ASD) ⁶. Ketogenic diets, which force the endogenous production of KBs, have proved to be effective for the treatment of neurological disorders associated with E/I imbalances, such as epilepsy and, more recently, ASD^14,41,42. It has been shown that the KB βHB induces DAF-16/FOXO activity ⁴³. Therefore, we asked whether it is possible to improve the observed phenotypes by modulating the activity of DAF-16 with βHB. We first evaluated the expression of sod-3, which codes for a superoxide dismutase and is a DAF-16 transcriptional target gene⁴⁴. We used a strain expressing a GFP transcriptional reporter for sod-3 and determined fluorescence intensity upon dietary supplementation of βHB. Consistent with previous reports, the levels of SOD-3::GFP are reduced in daf-18 and daf-16 mutant strains. Furthermore, we observed that βHB (20 mM) induce the expression of sod-3 in daf-18 but not in daf-16 mutants (Figure S3). Importantly, we did not detect increased sod-3 expression in daf-18; daf-16 double deficient animals, strongly suggesting that βHB induces sod-3 expression in daf-18 mutants through the transcription factor daf-16 (Figure S3).

Next, we evaluated behavioral phenotypes and GABAergic neuronal morphology of animals that were raised on an E. coli diet supplemented with 20 mM βHB throughout development (from embryogenesis to L4). We found that βHB supplementation significantly reduced the hypersensitivity of daf-18 mutants to the cholinergic drugs aldicarb and levamisole (Figures 3A and 3B). Moreover, βHB supplementation rescued the post-prodding shortening in daf-18 mutants (Figure 3C). Accordingly, we found that daf-18 mutants showed a significant increase in the proportion of closed omega turns during their escape response compared to the naïve condition (Figure 3D). In contrast, βHB exposure does not change the number of closed omega turns in daf-16 null mutants or the double null mutant daf-18; daf-16 (Figure 3D).

Figure 3.

We subsequently analyzed the changes in body length induced by optogenetic activation of both GABAergic and cholinergic neurons in animals exposed to a diet enriched with βHB. Interestingly, we found that daf-18 mutants exposed to βHB, but not wild-type or daf-16 mutant animals, exhibited increased elongation following optogenetic activation of GABAergic neurons (Figures 3F, 3G and 3H). Furthermore, we observed that the hypercontraction observed in daf-18 mutants after the activation of cholinergic neurons is significantly reduced in animals exposed to βHB (Figures 3I, 3J and 3K). These findings suggest that this ketone body can rebalance excitatory and inhibitory signals in the C. elegans neuromuscular system.

We also evaluated the morphology of GABAergic motor neurons in daf-18 mutants exposed to βHB. We found that βHB supplementation reduced the frequency of defects in GABAergic processes (Figure 3E). Consistently, βHB exposure did not significantly reduce the defects on GABAergic neurons of either daf-16 null mutants or daf-18; daf-16 double mutants. Taken together, these results demonstrate that dietary βHB ameliorates the defects associated with deficient GABAergic signaling in daf-18 mutants.

It is noteworthy that we did not observe any improvement in either neuronal outgrowth defects in the AIY interneuron or in the migration of the HSN motor neurons (Figure S4) in daf-18 mutants, even though these defects were shown to depend on the reduction of DAF-16 activity ^45,46. Unlike the development of D-type GABAergic neurons that mainly occurs at L1 larval stages ⁴⁷, AIY neurite and HSN soma migration take place during embryogenesis ^45,48. It is therefore possible that βHB may not go through the impermeable chitin eggshell of the embryo, as has been reported with other drugs ⁴⁹.

In the above experiments, βHB was supplemented throughout development: from embryonic stages to the L4 larval stage. We next asked whether there is a critical period during development where the action of βHB is required. We exposed daf-18 mutant animals to βHB supplemented diets at different periods of development (Figure 4A). Interestingly, we found that exposure to βHB during early development (L1) was sufficient to increase the proportion of daf-18 mutant animals executing a closed omega turn during the escape response. However, when the animals were exposed to βHB at later juvenile stages (L2-L3 or L3-L4), its ability to enhance the escape response of daf-18 mutants declined (Figure 4B). Moreover, exposing animals to βHB solely at the L1 stage was enough to reduce morphological defects in the GABAergic motor neurons of this mutants (Figure 4C). Thus, it is likely that βHB acts at very early stages of development to mitigate neurological defects in daf-18 mutants.

Figure 4.

Taken together, our findings demonstrate that mutations in daf-18, the C. elegans orthologue of PTEN, lead to specific defects in inhibitory GABAergic neurodevelopment without significantly affecting cholinergic excitatory signals. These GABA-specific deficiencies manifest as altered neuronal morphology, hypersensitivity to cholinergic stimulation, reduced responses to optogenetic GABAergic neuronal activation, mild body shortening following touch stimuli, and deficits in the execution of the omega turn. We have determined that these impairments in GABAergic development result from reduced activity of the FOXO orthologue DAF-16 in daf-18/PTEN mutants. Importantly, our study’s pivotal finding is that a βHB-enriched diet during early development, robustly mitigates the deleterious effects of daf-18/PTEN mutations in GABAergic neurons. This protective effect is critically dependent on the induction of DAF-16 by this ketone body.

Discussion

Mutations in daf-18/PTEN are linked to neurodevelopmental defects from worms to mammals^11,45,46. Moreover, decreased activity of PTEN produces E/I disequilibrium and the development of seizures in mice ⁷. The mechanisms underlying this imbalance are not clear. Our results demonstrate that reduced DAF-18/PTEN activity in C. elegans generates guidance defects, abnormal branching, incomplete commissural outgrowth and deficient function of inhibitory GABAergic neurons, without affecting the excitatory cholinergic neurons.

daf-18 deficient mutants have a shorter lifespan ⁵⁰. One possibility is that the defects in GABAergic processes are due to neurodegeneration associated with premature aging rather than developmental flaws. However, this idea is unlikely given that the neuronal defects are already manifest at the L1 stage. In C. elegans, the GABAergic motor neurons rearrange during L1-L2 stages ^47,51. Since the commissures are defective already at the early L1 stage (prior to rewiring), we can infer that this specific impairment in the neurodevelopment of GABAergic motor neurons is not a consequence of deficient rearrangement during the early larval stages.

Strikingly, cholinergic neurons have no noticeable morphological or functional defects in daf-18 mutants. Loss-of-function mutants in the neuronal integrin ina-1, ortholog of human ITGA6, affect the guidance of GABAergic commissures, without affecting cholinergic neurons ³⁹. Similar to PTEN, mutations in neuronal integrins have been linked to neurodevelopmental defects ⁵². Interestingly, the PI3K/Akt/FOXO pathway and integrin signaling are interrelated in mammals ⁵³. This observation opens the possibility that one of the mechanisms by which daf-18 mutants have defects in GABAergic neurodevelopment involves integrin expression and/or function. Interestingly, mutations in eel-1, the C. elegans ortholog of HUWE1, or in subunits of the Anaphase-Promoting Complex (APC), lead to developmental and functional alterations in GABAergic neurons but not in cholinergic neurons^22,54, despite their expression in both neuronal types. This observation suggests the existence of compensatory or redundant mechanisms in cholinergic neurons that may not be present in GABAergic neurons.

In mammals, defects in PTEN mutants have been typically related to altered function of the mTOR pathway ^8,55,56. However, our results suggest that, in the C. elegans neuromuscular system, decreased activity of DAF-18 affects GABAergic development due to a downregulation of DAF-16/FOXO transcription factor activity. The FOXO family of transcription factors is conserved throughout the animal kingdom ⁵⁷. There is increasing evidence demonstrating the key role of this transcription factors family in neurodevelopment ^58–60. Downregulation of FOXO activity early in development reproduces neuropathological features found in ASD patients, i.e., increased brain size and cortical thickness ^61,62. The autonomic activity of DAF-18 and DAF-16 coordinates axonal outgrowth in C. elegans AIY interneurons and rat cerebellar granule neurons ⁴⁵. On the other hand, DAF-18 and DAF-16 in the hypodermis control neuronal migration of the HSN neuron during development ⁴⁶. Our rescue experiments strongly suggest that the PI3K/AKT-1/DAF-16 pathway modulates the development of GABAergic motor neurons by acting autonomously in these cells. Noteworthy, autonomic DAF-16 activity in GABAergic motor neurons is also key for axonal growth during regeneration ⁶³. These results further emphasize the importance of DAF-16/FOXO in neuronal development and axonal growth.

In many patients suffering from epilepsy, ketogenic diets can control seizures ^41,42. Furthermore, they can reduce behavioral abnormalities in individuals with ASD ¹⁴. While the mechanisms underlying the clinical effects of ketogenic diets remain unclear, it has been shown that these diets correlate with increased GABA signaling ^64–66. We demonstrate here that dietary supplementation of the ketone body βHB ameliorates morphological and functional defects in GABAergic motor neurons of daf-18 mutants. Although ketone bodies were historically viewed as simple carriers of energy to peripheral tissues during prolonged fasting or exercise, our findings confirm more recent reports showing that βHB also possesses a variety of important signaling functions ⁶⁷. We can hypothesize several distinct, non-mutually exclusive models by which βHB can induce DAF-16/FOXO-dependent signaling. βHB directly inhibits mammalian histone deacetylases HDAC1 and HDAC2, increasing histone acetylation at the FOXO3a promoter and inducing the expression of this gene ⁶⁸. HDAC1 and HDAC2 play an important role as redundant regulators of neuronal development ⁶⁹. Interestingly, in C. elegans βHB inhibits the class I HDACs to extend worm lifespan in a DAF-16-dependent manner ⁴³. Therefore, it is possible that βHB-mediated HDAC inhibition upregulates transcription of DAF-16 counterbalancing hyperactivation of the PI3K pathway in daf-18/PTEN mutants. Another potential mechanism for the effect of βHB involves the inhibition of the insulin signaling pathway. In mammals, the administration of βHB downregulates the insulin signaling in muscle ⁷⁰. Moreover, several reports have shown that βHB administration reduces phosphorylation and activity of Akt/protein kinase downstream of the insulin receptor ^71,72. In C. elegans, inhibition of AKT-1 activates DAF-16 ⁷³. Although understanding the mechanism behind βHB’s action will require further studies, our results demonstrate that this ketone body positively modulates DAF-16/FOXO during neuronal development.

Multiple reports, from C. elegans to mammals, suggest that there is a sensitive period, typically early in development, where pharmacological or genetic interventions are more effective in ameliorating the consequences of neurodevelopmental defects ⁷⁴. However, recent evidence shows that phenotypes associated with certain neurodevelopmental defects can be ameliorated by interventions during adulthood ⁷⁵. Our results show that βHB can rescue the phenotypic defects of daf-18 mutants only when exposure occurs during an early critical period. The inefficacy of βHB at later stages suggests that the role of DAF-16/FOXO in the maintenance of GABAergic neurons is not as relevant as its role in development.

Since the PI3K/Akt pathway is highly conserved in animals, this study provides universally relevant information on how alterations in this pathway lead to neurodevelopmental defects and the mechanisms underlying the positive effects of ketogenic diets on neuronal disorders characterized by GABA dysfunction and altered E/I ratios.

Acknowledgements

Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). We thank Mark Alkema, and Alex Byrne for strains. We thank Andrés Garelli, Guillermo Spitzmaul, Gabriela Salvador, Mark Alkema, Inés Carrera and Claire Bénard for helpful discussions. In addition, we would like to acknowledge Ignacio Bergé, Adrian Bizet, Carolina Gomila, Marta Stulhdreher and Carla Chrestía for technical support

Funding

This work was supported by Grants from: 1) Consejo Nacional de Investigaciones Científicas y Técnicas, Argentina to INIBIBB (PUE-N22920170100017CO) and to DR/ MJDR (PIP No. 11220200101606CO). 2) Agencia Nacional de Promoción de la Ciencia y la Tecnología ANPCYT Argentina to DR (PICT 2019-0480 and PICT-2021-I-A-00052) and MJDR (PICT-2017-0566 and PICT-2020-1734) and 3) Universidad Nacional Del Sur to DR (PGI: 24/B291) and MJDR (PGI: 24/B261). The funders had no role in the study design, data collection, and analysis, decision to publish, or preparation of the manuscript.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Materials and methods

C. elegans culture and maintenance

All C. elegans strains were grown at room temperature (20°C) on nematode growth media (NGM) agar plates with OP50 Escherichia coli as a food source. The wild-type reference strain used in this study is N2 Bristol. Some of the strains were obtained through the Caenorhabditis Genetics Center (CGC, University of Minnesota). Worm population density was maintained low throughout their development and during the assays. Transgenic strains were generated by microinjection of plasmid DNA containing the construct Punc-47::daf-18cDNA (kindly provided by Alexandra Byrne, UMASS Chan Medical School) at 20 ng/µL into the germ line of (daf-18 (ok480); lin-15 (n765ts)) double mutants with the co-injection marker lin-15 rescuing plasmid pL15EK (80 ng/µl). At least three independent transgenic lines were obtained. Data are shown from a single representative line.

The strains used in this manuscript were:

CB156 unc-25(e156) III

MT6201 unc-47(n2409) III

CB1375 daf-18(e1375) IV

OAR144 daf-18(ok480) IV

GR1307 daf-16(mgdf50) I

OAR115 daf-16(mgDf50); daf-18(ok480)

OAR161 daf-18(ok480); Punc-47::daf-18

LX929 vsIs48 (Punc-17::gfp)

IZ629 ufIs34 (Punc-47::mCherry)

OAR117 ufis34 (Punc-47::mCherry); daf-18(ok480)

OAR118 vsIs48 (Punc-17::gfp); daf-18(ok480)

OAR142 ufis34 (Punc-47::mCherry); daf-16(mgDf50)

OAR143 ufis34 (Punc-47::mCherry); daf-16(mgDf50); daf-18(ok480)

CF1553 muIs84 [(pAD76) Psod-3::gfp + rol-6(su1006)]

OAR140 muIs84 [(pAD76) Psod-3::gfp + rol-6]; daf-18(ok480)

OAR141 muIs84 [(pAD76) Psod-3::gfp + rol-6]; daf-16(mgDf50)

OH99 mglS18[Pttx-3::gfp]

OAR83 daf-18(ok480); mglS18[Pttx-3::gfp]

MT13471 [Ptph-1::gfp]

OAR112 [Ptph-1::gfp]; daf-18(ok480)

IZ805 [Punc-17::ChR2]

ZM3266 [Punc-47::ChR2]

OAR177 [Punc-17::ChR2]; daf-18(ok480)

OAR178 [Punc-17::ChR2]; daf-16(mgDf50)

OAR179 [Punc-47::ChR2]; daf-18(ok480)

OAR180 [Punc-17::ChR2]; daf-16(mgDf50)

TJ1052 age-1(hx546)

GR1310 akt-1(mg144) GR1318 pdk-1(mg142)

JT9609 pdk-1(sa680)

VC204 akt-2(ok393)

VC222 raga-1(ok386)

KQ1366 rict-1(ft7)

Paralysis assays

Paralysis assays were carried out in standard NGM plates with 2 mM aldicarb (Sigma-Aldrich) or 0.5 mM levamisole (Alfa Aesar). 25-30 L4 worms were transferred to each plate and paralyzed animals were counted every 15 or 30 minutes. An animal was considered as paralyzed when it did not respond after prodding three times with a platinum wire on the head and tail ¹. At least four independent trials with 25-30 animals for each condition were performed. The area under the curve (AUC) for each condition in each experiment was used for statistical comparisons

Escape response

Escape response assays were performed on NGM agar plates seeded with a thin bacterial lawn of OP50 E. coli. To maintain tight control of growth and moisture, 120 μL of bacteria were seeded 24 hours before the assay and grown overnight at 37° C. The day of the assay, L4 worms were transferred to the plates and allowed to acclimate for at least 5 min. Omega turns were induced by gentle anterior touch with fine eyebrow hair and were classified as closed when the worm touched the tail with its head as previously described ². Between 4 and 7 independent trials with ∼20 animals for each condition were performed.

Body length assays

Body length measurements were performed in standard NGM agar plates without bacteria. L4 synchronized worms were transferred into the plates and allowed to acclimate for at least 5 min. Worms were recorded with an Amscope Mu300 camera. Animal body length, before and after touching with a platinum pick, was measured using FIJI Image J software. Quantification of body shortening after touching was calculated as the decrease of body length related to the length of the animal before being touched.

Commissure analysis

Synchronized L1 or L4 animals carrying the fluorescence reporters vsIs48 (Punc-17::GFP, cholinergic neurons) or ufIs34 (Punc-47::mCherry, GABAergic neurons) were immobilized with sodium azide (0.25 M) on 2% agarose pads. Commissures of GABAergic and cholinergic neurons were scored with a Nikon Eclipse TE 2000 fluorescence microscope. A commissure is generally composed of a single process, and occasionally two neurites that extend together dorsally. Defects on commissures, including guidance defects, abnormal branching, and incomplete commissures were classified similarly to previous reports ³. The percentage of animals with at least one commissure defect was calculated for each neuronal class (e.g., cholinergic or GABAergic). At least three trials (∼20 animals per condition in each trial) were analyzed for each individual experiment. Representative images shown in the figures were collected using laser confocal microscopy (ZEISS LSM 900 with AirScan II) with 20× and 63× objectives.

β-hydroxybutyrate assays

Worms were exposed to 20 mM DL-3-hydroxybutyric acid sodium salt (Acros Organics) in NGM agar plates seeded with E. coli OP50. Exposure time varied in each trial as indicated (See Figure 4).

sod-3 expression

sod-3 expression levels were analyzed in transgenic strains containing the transcriptional reporter muIs84, as described previously ^4,5. Synchronized L4 animals were anesthetized with Sodium Azide (0.25 M) and mounted on 2% agarose pads. Images were collected using a Nikon Eclipse TE 2000 fluorescence microscope. GFP fluorescence intensity was quantified in same-sized ROIs at the head of the animal using Image J FIJI software. Results were normalized to control conditions (wild-type individuals without βHB). ∼40-60 animals for each genotype/condition were analyzed.

Optogenetic assays

We examined young adult animals (6-8 hours post-L4 stage) that express Channelrhodopsin (ChR2) in either GABAergic (Punc-47::ChR2) or cholinergic neurons (Punc-17::ChR2). We transferred these animals to a NGM 6mm agar plate without food, let them acclimate for 5 minutes, and recorded each animal at 15 frames per second using an Allied Vision Alvium 1800 U-500m camera. To stimulate neuronal activity, we exposed the animals to 470 nm light pulses for 5 seconds. These light pulses were delivered using a custom Python script (VIMBA Peron) to an Arduino Uno microcontroller, which operated a Mightex compact universal LED controller (Mightex SLC-MA02-U). The light emission was achieved through a Mightex High-Power LED Collimator Source (LCS-0470–03-11). To precisely track the changes in the worm’s body, we continuously monitored its area from 5 seconds before the light stimulus, during the light stimulus and until 5 seconds afterward. We developed a FIJI-Image J macro capable of automatically tracking the body area in each frame, capitalizing on the clear contrast between the worm’s body and the background. As demonstrated in Movies 3 and 4, changes in body area directly corresponded to alterations in the animal’s length. To validate our measurement system, we manually measured the width of 6-8 animals at the 2.5-second point of light stimulation and compared them to the body area and length. Our observations consistently showed that, regardless of whether the area increased or decreased (depending on the activation of GABAergic or cholinergic neurons), the width remained mostly unchanged (Fig. S1G). Therefore, the observed changes in the animaĺs area measured by our FIJI-Image J macro indeed represent alterations in the animal’s length

AIY and HSN analysis

Synchronized L4 or Young Adult worms carrying the fluorescence reporters Pttx-3::gfp (AIY interneurons expressing GFP) and Ptph-1::gfp (HSN expressing GFP) were immobilized with sodium azide (0.25 M) on 2% agarose pads and analyzed with a Nikon Eclipse TE 2000 fluorescence microscope. AIY neurons morphology were sorted in qualitative categories (see figure legend) while the migration of HSN was classified in quantitative categories using ImageJ software.

Statistical Analysis

The results presented in each figure are the average of at least three independent assays. Bars represent mean ± SEM. Comparison between two groups was carried out using Student’s t-test, while differences among more than two groups were carried out by One-way analysis of variance (ANOVA). ns p > 0.05; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. We used the software GraphPad Prism version 6.01 to perform statistics. Other statistical information is indicated in the figure legends. For all assays, the scoring was done blinded.

Supplementary Figures

Suppl. Figure 1.

Suppl. Figure 2.

Suppl. Figure 3.

Suppl. Figure 4.