Introduction

Maintaining a delicate balance between excitatory and inhibitory (E/I) neurotransmission is critical for optimal brain function¹. 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 ASD¹³. The molecular events underlying the neurodevelopmental deficits in PTEN mutants remain poorly understood.

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^14–20. This system has yielded valuable insights into fundamental synaptic transmission mechanisms^21,22. 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^14,19,20,23. Furthermore, the substantial conservation of the main components of the PI3K/Akt pathway in C. elegans^24,25, enhances the applicability of this model system for investigating the role of this pathway in neurodevelopment.

We here found that mutations in daf-18 (the ortholog for PTEN in C. elegans) result in specific impairments in GABAergic inhibitory signaling due to decreased activity of the transcription factor DAF-16 (the ortholog for FOXO in C. elegans) during neurodevelopment. Interestingly, cholinergic excitatory motor neurons remain unaffected. This targeted impairment of inhibitory signals causes an imbalance between excitatory and inhibitory (E/I) neurotransmission in the animal’s neuromuscular system.

In humans, ketogenic diets (KGDs), which force fatty acids beta-oxidation into ketone bodies, have been utilized for decades to treat pathologies associated with E/I imbalances, such as refractory epilepsies^26–28. More recently, KGDs have also demonstrated effectiveness in alleviating autistic symptoms in humans²⁹ and rodent models of ASD^30,31. The mechanisms underlying these beneficial effects remain largely unknown.

We demonstrated that exposing daf-18/PTEN mutants to a diet enriched with the ketone body β-hydroxybutyrate (βHB) early in development enhances DAF-16/FOXO activity, mitigates morphological and functional defects in GABAergic neurons, and improves behavioral phenotypes. This study not only provides a straightforward system for studying the role of the conserved PI3K/Akt/FOXO pathway in neurodevelopment but also contributes to our understanding of the mechanisms underlying the effects of ketone bodies in neurodevelopment.

Results

Mutants in daf-18/PTEN and daf-16/FOXO are Hypersensitive to Cholinergic Drugs

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^14,15. We analyzed the sensitivity of daf-18/PTEN deficient animals to the acetylcholinesterase inhibitor aldicarb and 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 paralysis³² (Figure 1A). Levamisole also induces paralysis by directly activating muscular cholinergic receptors³² (Figure 1A). We found that daf-18/PTEN mutants are hypersensitive to the paralyzing effects of both drugs (Figures 1B and 1C). Hypersensitivity to cholinergic drugs is typical of animals with an increased E/I ratio in the neuromuscular system, such as mutants in unc-25 (the C. elegans orthologue for glutamic acid decarboxylase, an essential enzyme for synthesizing GABA)^14,15. While daf-18/PTEN mutants become paralyzed earlier than wild-type animals, their hypersensitivity to cholinergic drugs is not as severe as that observed in animals completely deficient in GABA synthesis, such unc-25 null mutants (Figures 1B and 1C) indicating a less pronounced imbalance between excitatory and inhibitory signals. Reduced activity of DAF-18/PTEN has been largely shown to exacerbate the PI3K pathway, precluding the activation of DAF-16/FOXO, the C. elegans ortholog of the FOXO transcription factors family²⁵ (Figure S1A). We analyzed aldicarb and levamisole sensitivity of mutants in this transcription factor. Similar to daf-18/PTEN mutants, we found that daf-16/FOXO null mutants are hypersensitive to the paralyzing effects of aldicarb and levamisole (Figure 1C). 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 (Figure 1C). In addition to daf-16/FOXO and daf-18/PTEN, we assessed the sensitivity to the paralyzing effects of aldicarb and levamisole in loss of function mutants of other components of the PI3K pathway, such as age-1/PI3K, pdk-1, akt-1, and akt-2 (Figure 1D). Unlike mutations in daf-18/PTEN, in these mutants the PI3K pathway is downregulated (Figure S1A). We did not observe significant differences compared to the wild-type. Due to the complete dauer arrest observed in double mutants of akt-1 and akt-2^33,34, we were unable to explore the potential redundancy of these two genes. Interestingly, a gain-of-function mutant in pdk-1, pdk-1(mg142)²⁴, is hypersensitive to aldicarb and levamisole, similar to daf-16/FOXO and daf-18/PTEN mutants (Figure 1D). Given that increased pdk-1 activity is linked to hyperphosphorylation and inactivation of DAF-16/FOXO (Figure S1A), these results support the hypothesis that low activity of DAF-16/FOXO leads to hypersensitivity to these drugs.

Figure 1.

In vertebrates, alterations in PTEN activity have been largely shown to impact neuronal development and function by affecting the mTOR pathway³⁵. Consequently, we analyzed 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 (Figure 1E). This suggests that the mTOR pathway is not involved in daf-18/PTEN pharmacological phenotypes.

daf-18/PTEN and daf-16/FOXO Mutants Show Phenotypes Indicative of GABAergic Deficiency

In C. elegans, the body wall muscles receive cholinergic innervation, which induces contraction, and GABAergic innervation, which leads to relaxation. The contralateral activity of cholinergic and GABAergic neurons facilitates the characteristic undulatory movement of the animal (Figure 2A). Hypersensitivity to cholinergic drugs has long been observed in worms where GABAergic signaling is deficient^14,15 (Figure 1). In mutants with severe deficits in GABA transmission, prodding induces a bilateral contraction of the body wall muscles that shortens the body (shrinker phenotype)³⁶. When daf-18/PTEN mutants are touched, there is a slight but significant shortening in body length (Figure 2B). As expected, this shortening is not as noticeable as in animals with a complete deficit in GABAergic signaling, such as mutants in unc-25. Similar to daf-18/PTEN mutants, daf-16/FOXO 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 2B), further supporting the notion that both genes act in the same pathway to impact neuromuscular signaling.

Figure 2.

We also analyzed other behaviors that require a concerted activity of GABAergic and cholinergic systems, such as the omega turns during the escape response³⁷. 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 ³⁷ (Movie 1). The execution of the omega turn involves a hypercontraction of the ventral muscles and relaxation of the dorsal muscles, 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 resumes locomotion in the opposite direction (Movie 1). In response to anterior touch, the vast majority of wild-type worms make a closed omega turn^37,38(Figure 2C). 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 2A)^37,38. 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)³⁷. In agreement with previous reports, we found that 91% of wild-type animals exert a closed omega turn within the escape response (Figure 2C). 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 53% of daf-18 mutants exhibit the typical head-to-tail contact during the omega turn (Figure 2C and Movie 2). Akin to daf-18 mutants, daf-16/FOXO mutants exhibited a decrease in the proportion of closed omega turns (Figure 2C). No additive effects were observed in the daf-18; daf-16 double mutant, suggesting that the increased inactivation of daf-16/FOXO 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^39–41(Figure 2D 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/PTEN and daf-16/FOXO mutants compared to wild-type worms (Figure 2D). 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, a gene encodes the vesicular acetylcholine transporter (VAChT), which leads to muscle contraction and shortened body length^39,41(Figure 2E and Movie 4). Rather than observing reduced shortening in daf-16/FOXO and daf-18/PTEN mutants, we found that cholinergic activation caused hypercontraction of these mutant animals (Figure 2E). 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 2A), 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/PTEN and daf-16/FOXO mutants, there is a specific functional defect in GABAergic neurons, while excitatory neurons do not appear to be affected.

Disruption of daf-18/PTEN Alters Commissural Trajectories in GABAergic Motor Neurons

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) ⁴² (Figures 3A and S2). The commissures have proved useful for studying defects in motor neuron development or maintenance^43,44. We analyzed the morphology of GABAergic motor neurons in L4 animals expressing mCherry under the control of the unc-47 promoter⁴⁵. We found that daf-18/PTEN mutants exhibit a higher frequency of commissure flaws, including guidance defects, ectopic branching, and commissures that fail to reach the dorsal cord (Figures 3B and 3C). 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/PTEN animals (Figure S2).

Figure 3.

GABAergic motor neurons can be classified based on the muscles they innervate: those that innervate the dorsal muscles are called DDs, while those that innervate the ventral muscles are called VDs. Both types of D neurons send commissures to the DNC^46,47. We found defects in various commissures, some of which correspond to VD and others to DD neurons (Figure S3). We also analyzed GABAergic commissures at the beginning (1 hour post-hatching) of the first larval stage (L1), when only the six DD neurons are formed^46,48,49. We found that in this early larval stage, daf-18/PTEN mutants exhibit defects in the GABAergic commissures (Figures 3D, 3E and S4). While the DD neurons are born and mostly develop embryonically, limited postembryonic axonal outgrowth has been observed in these neurons⁴⁹. We did not find an increase in the number of errors in larvae 5-6 hours post-hatching compared to recently hatched larvae (Figure S4B), indicating that deficiencies in DD neurons in daf-18/PTEN mutants mainly occur during their embryonic development. In contrast, we observed that the prevalence of errors increases significantly at the L4 stage (Figure S4B and S4C). The greater number of defects in L4s likely arises from defects in the VDs, which are born post-embryonically between the mid-L1 larval stage and the L2 stage, and add to the defects already present in the DDs. Taken together, these observations suggest that reduced DAF-18/PTEN activity affects the neurodevelopment of the GABAergic motor system. Since the transcription factor DAF-16/FOXO is one of the main targets of DAF-18/PTEN signaling, we analyzed the morphology of GABAergic motor neurons in daf-16/FOXO null mutants. These animals also exhibit an increased number of defects in GABAergic commissures compared to the wild type (Figure 3C).

Given that daf-18 is ubiquitously expressed in all tissues²⁵, we asked whether DAF-18/PTEN acts autonomously in GABAergic neurons to ensure proper development. We found that specific daf-18/PTEN expression in GABAergic neurons increased the proportion of closed omega turns in daf-18/PTEN null mutants (Figure 3F). In addition, the morphological defects in GABAergic commissures were significantly reduced (Figure 3G), suggesting that DAF-18/PTEN acts autonomously in GABAergic motor neurons to regulate their development.

Collectively, our findings demonstrate that mutations in daf-18/PTEN and daf-16/FOXO 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 arise from the hyperactivation of the PI3K pathway, along with subsequent DAF-16/FOXO inhibition, in GABAergic neurons of daf-18/PTEN mutants.

A diet enriched with the ketone body β-hydroxybutyrate (βHB) Ameliorates Defects in daf18/PTEN Mutants but not in daf-16/FOXO 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 ^27–29. It has been shown that the KB βHB induces DAF-16/FOXO activity⁵⁰. Therefore, we asked whether it is possible to improve the observed phenotypes by modulating the activity of DAF-16/FOXO with βHB. We first evaluated the expression of sod-3, which codes for a superoxide dismutase and is a DAF-16/FOXO transcriptional target gene⁵¹. 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/PTEN and daf-16/FOXO mutant strains. Furthermore, we observed that βHB (20 mM) induces the expression of sod-3 in daf-18/PTEN but not in daf-16/FOXO mutants (Figure S5). 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/PTEN mutants through the transcription factor daf-16/FOXO (Figure S5).

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 egg laying until the time of the assay (typically late L4s or young adults). We found that βHB supplementation significantly reduced the hypersensitivity of daf-18/PTEN mutants to the cholinergic drugs aldicarb and levamisole (Figures 4A and 4B). Moreover, βHB supplementation rescued the post-prodding shortening in daf-18/PTEN mutants (Figure 4C). Accordingly, we found that daf-18/PTEN mutants showed a significant increase in the proportion of closed omega turns during their escape response compared to the naïve condition (Figure 4D). In contrast, βHB exposure does not change the number of closed omega turns in daf-16/FOXO null mutants or the double null mutant daf-18; daf-16 (Figure 4D).

Figure 4.

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/PTEN mutants exposed to βHB, but not wild-type or daf-16/FOXO mutant animals, exhibited increased elongation following optogenetic activation of GABAergic neurons (Figures 4E, 4F, and 4G). Furthermore, we observed that the hypercontraction observed in daf-18/PTEN mutants after the activation of cholinergic neurons is significantly reduced in animals exposed to βHB (Figures 4H, 4I and 4J). These findings suggest that this ketone body can rebalance excitatory and inhibitory signals in the neuromuscular system of daf-18/PTEN C. elegans mutants.

We also evaluated the morphology of GABAergic motor neurons in daf-18/PTEN animals exposed to βHB. We found that βHB supplementation reduced the frequency of defects in GABAergic processes (Figure 4K). Consistently, βHB exposure did not significantly reduce the defects on GABAergic neurons of either daf-16/FOXO 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/PTEN mutants.

It is noteworthy that we did not observe any improvement in either neuronal outgrowth defects in the AIY interneuron or the migration of the HSN motor neurons (Figure S6) in daf-18/PTEN mutants exposed to βHB, even though these defects were shown to depend on the reduction of DAF-16/FOXO activity ^52,53. AIY neurite and HSN soma migration take place during embryogenesis ^52,54. It is therefore possible that βHB may not go through the impermeable chitin eggshell of the embryo, as has been reported with other drugs (see below) ⁵⁵.

β-Hydroxybutyrate Exposure During Early L1 Development Sufficiently Mitigates GABAergic System Defects

In the above experiments, animals were exposed to βHB throughout development. We next asked whether there is a critical period during development where the action of βHB is required. We exposed daf-18/PTEN mutant animals to βHB-supplemented diets for 18-hour periods at different developmental stages (Figure 5A). The earliest exposure occurred during the 18 hours following egg laying, covering ex-utero embryonic development and the first 8-9 hours of the L1 stage. The second exposure period encompassed the latter part of the L1 stage, the entire L2 stage, and most of the L3 stage. The third exposure spanned the latter part of the L3 stage, the entire L4 stage, and the first 6-7 hours of the adult stage (Figure 5A). Interestingly, we found that the earliest exposure to βHB was sufficient to increase the proportion of daf-18/PTEN mutant animals executing a closed omega turn during the escape response. However, when the animals were exposed to βHB at later juvenile stages, their ability to enhance the escape response of daf-18/PTEN mutants declined (Figure 5B). Moreover, exposing animals to βHB for 18 hours starting from egg laying was enough to reduce morphological defects in the GABAergic motor neurons of these mutants (Figure 5C and Figure S7). Interestingly, βHB has no effect on GABAergic commissures in either recently hatched L1s or L4 daf-18/PTEN mutants when exposure is limited to the first 9 hours after egg laying (where ex utero embryonic development occurs), possibly due to the impermeability of the chitinous eggshell (Figure S7). Thus, it is likely that βHB acts at early L1 stages to mitigate neurological GABAergic defects in daf-18/PTEN mutants.

Figure 5.

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/FOXO by this ketone body.

Discussion

Mutations in daf-18/PTEN are linked to neurodevelopmental defects from worms to mammals^11,52,53. 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/PTEN deficient mutants have a shorter lifespan⁵⁶. 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 in DD neurons are already evident at the early L1 stage. In C. elegans, DD GABAergic motor neurons undergo rearrangements during the L1-L2 stages^48,49. In newly hatched L1 larvae, each DD motor neuron innervates ventral muscles and extends a commissure to the dorsal nerve cord to receive synaptic inputs from cholinergic DA and DB neurons^48,49,57. In adults, the DD commissure morphology is maintained, but the synaptic output shifts to dorsal muscles, and input is provided by VA and VB cholinergic motor neurons in the ventral nerve cord⁵⁷. A plausible possibility is that this remodeling process makes DD neurons specifically sensitive to PI3K pathway activity. However, since the DD commissures are defective already at the very early L1 stage (prior to rewiring) and similar defects are observed in VD neurons, which are born post-embryonically between late L1 and L2 stages and do not undergo this remodeling^46,49,57, we can infer that the deficits caused by daf-18/PTEN deficiency affect the development of the entire GABAergic system, independently of the synaptic rearrangement of DD neurons.

Strikingly, cholinergic neurons have no noticeable morphological or functional defects in daf-18/PTEN 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 ⁴⁴. Similar to PTEN, mutations in neuronal integrins have been linked to neurodevelopmental defects ⁵⁸. Interestingly, the PI3K/Akt/FOXO pathway and integrin signaling are interrelated in mammals ⁵⁹. This observation opens the possibility that one of the mechanisms by which daf-18/PTEN 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^19,60, 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^8,61,62. However, our results suggest that, in the C. elegans neuromuscular system, decreased activity of DAF-18/PTEN 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 kingdom⁶³. There is increasing evidence demonstrating the key role of this transcription factors family in neurodevelopment^64–66. Downregulation of FOXO activity early in development reproduces neuropathological features found in ASD patients, i.e., increased brain size and cortical thickness^67,68. The autonomic activity of DAF-18/PTEN and DAF-16/FOXO coordinates axonal outgrowth in C. elegans AIY interneurons and rat cerebellar granule neurons ⁵². On the other hand, DAF-18/PTEN and DAF-16/FOXO in the hypodermis control neuronal migration of the HSN neuron during development⁵³. Our rescue experiments strongly suggest that the PI3K/Akt/DAF-16 pathway modulates the development of GABAergic motor neurons by acting autonomously in these cells. Noteworthy, autonomic DAF-16/FOXO activity in GABAergic motor neurons is also key for axonal growth during regeneration⁶⁹. 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^27,28. Furthermore, they can reduce behavioral abnormalities in individuals with ASD²⁹. While the mechanisms underlying the clinical effects of ketogenic diets remain unclear, it has been shown that these diets correlate with increased GABA signaling^70–72. We demonstrate here that dietary supplementation of the ketone body βHB ameliorates morphological and functional defects in GABAergic motor neurons of daf-18/PTEN 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 functions⁷³. 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 gene⁷⁴. HDAC1 and HDAC2 play an important role as redundant regulators of neuronal development⁷⁵. Interestingly, in C. elegans βHB inhibits the class I HDACs to extend worm lifespan in a DAF-16/FOXO-dependent manner⁵⁰. Therefore, βHB-mediated HDAC inhibition may upregulate transcription of DAF-16/FOXO 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 muscle⁷⁶. Moreover, several reports have shown that βHB administration reduces phosphorylation and activity of Akt/protein kinase downstream of the insulin receptor ^77,78. In C. elegans, inhibition of AKT-1 activates DAF-16/FOXO ⁷⁹. 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 defects⁸⁰. However, recent evidence shows that phenotypes associated with certain neurodevelopmental defects can be ameliorated by interventions during adulthood⁸¹. Our results show that βHB can ameliorate the phenotypic defects of daf-18/PTEN 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.

Our experiments do not allow us to distinguish whether the effect of βHB is preventive, reversive, or both. Our results suggest that the improvement is not due to prevention in DDs because the defects are present in newly hatched larvae regardless of the presence or absence of βHB, and DD post-embryonic growth does not add new errors. Unlike in early L1 stages, the protective effect of βHB becomes evident when analyzing the commissures of L4 animals. In this late larval stage, not only the DDs but also the VD neurons are present. This leads us to speculate that βHB may have a preventive action on the neurodevelopment of VD neurons. We also cannot rule out that this improvement may be due, at least partially, to a reversal of defects in DD neurons. It is intriguing how exposure to βHB during early L1 could ameliorate defects in neurons that mainly emerge in late L1s (VDs). We can hypothesize that residual βHB or a metabolite from the previous exposure may prevent these defects in VD neurons. βHB, in particular, has been shown to generate long-lasting effects through epigenetic modifications⁸². Further investigations are needed to elucidate the underlying fundamental mechanisms regarding the ameliorating effects of βHB supplementation on deficits in GABAergic neurodevelopment associated with mutations in daf-18/PTEN.

Across the animal kingdom, food signals increase insulin levels leading to the activation of Akt/PI3K pathway^83–85. In C. elegans, the L1 larval stage is particularly sensitive to nutritional status. C. elegans adjusts its development based on food availability, potentially arresting in L1 in the absence of food⁸⁶. Strong loss-of-function alleles in the insulin signaling pathway exhibit constitutive L1 arrest⁸⁷, highlighting the critical importance of this pathway during this larval stage. Hence, it is not surprising that dietary interventions targeting the PI3K pathway at these critical early L1 stages can modulate developmental processes. Our pharmacological experiments showed that mutants associated with an exacerbation of the PI3K pathway, which typically inhibits the nuclear translocation and activity of the transcription factor DAF-16/FOXO, lead to E/I imbalances that manifest as hypersensitivity to cholinergic drugs. We demonstrated that these imbalances arise from defects that occur specifically in the neurodevelopment of GABAergic motor neurons. Interestingly, mutants inhibiting the PI3K pathway do not show differences in their sensitivity to cholinergic drugs compared to wild-type animals. This observation can be explained by a critical period during neurodevelopment when the Insulin/Akt/PI3K pathway must be maintained at very low activity (or even deactivated). These low activity levels of the Insulin/Akt/PI3K pathway would allow for a high level of DAF-16/FOXO activity, which, according to our results, appears to be key for the proper development of GABAergic neurons.

The fine regulation of insulin and insulin-like signaling during early development is a conserved process in animals^88,89. In mammals, for instance, conditions like Gestational Diabetes Mellitus (GDM), characterized by fetal hyperinsulinemia and high levels of IGF-1⁹⁰, are associated with neurodevelopmental defects⁹¹. Our results lead to the intriguing idea that dietary interventions that increase DAF-16/FOXO activity, such as βHB supplementation, could constitute a potential therapeutic strategy for these pathologies. Future studies using mammalian models are crucial to shed light on the potential of this hypothesis.

Acknowledgements

Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). We thank Mark Alkema, Michael Francis and Alex Byrne for strains. We thank Andrés Garelli, Guillermo Spitzmaul, Gabriela Salvador, Mark Alkema, Inés Carrera, Claire Bénard, Alexandra Byrne and Michael Francis for helpful discussions. In addition, we would like to acknowledge Ignacio Bergé, Adrian Bizet, Carolina Gomila, Marta Stulhdreher, María José Tiecher, Edgardo Buzzi 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 (22°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. All experiments were conducted on age-synchronized animals. This was achieved by placing gravid worms on NGM plates and removing them after two hours. The assays were performed on the animals hatched from the eggs laid in these two hours.

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) I; daf-18(ok480) IV

• OAR161 daf-18(ok480); wpEx173[Punc-47::daf-18 + myo-2::GFP]

• 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 nIs121[Ptph-1::gfp]

• OAR112 nIs121[Ptph-1::gfp]; daf-18(ok480)

• IZ805 ufIs53[Punc-17::ChR2]

• ZM3266 zxIs3[Punc-47::ChR2::YFP]

• OAR177 ufIs53[Punc-17::ChR2;;YFP]; daf-18(ok480)

• OAR178 ufIs53[Punc-17::ChR2::YFP] ; daf-16(mgDf50)

• OAR179 zxIs3[Punc-47::ChR2::YFP];daf-18(ok480)

• OAR180 zxIs3[Punc-47::ChR2::YFP] ; 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 paralyzed when it did not respond after prodding three times with a platinum wire on the head and tail¹. 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, young adult 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. Young adult 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) on NGM agar plates seeded with E. coli OP50. We synchronized the animals similarly to other experiments by placing gravid animals in NGM plates containing 20 mM of βHB and removing them after two hours.

For experiments involving exposure at different developmental stages, animals were transferred between plates with and without βHB as needed. For the earliest exposure, eggs were laid on plates containing 20 mM of βHB and then transferred to drug-free plates to complete their development. Conversely, for later exposures, animals were born on βHB-free plates and subsequently transferred to βHB-containing plates at the specified time (See Figure 5).

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). ∼35-60 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 compare the changes induced by optogenetic activity between different strains, the body area measurements for each animal were averaged from second 6 (1 second after the blue light was turned on) to second 9 (1 second before the light was turned off). These average values were used for the statistical comparisons detailed in each figure legend.

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 (Figure 2F and 2G). 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 ± SD. Typically, one-way ANOVA was employed for analyzing multiple parametric samples, and Tukey’s post hoc test was used for pairwise comparisons among all groups. For comparisons against a control group, Dunnett’s post hoc test was used. For multiple non-parametric samples, the Kruskal-Wallis test was applied followed by Dunn’s post hoc test, which was also utilized for comparisons against the control group. In cases where comparisons were made between two independent conditions, a t-test was utilized for parametric data, while the Mann-Whitney U test was employed for non-parametric data. We used the software GraphPad Prism version 6.01 to perform statistics. The statistical information is indicated in the figure legends. For all assays, the scoring was done blinded. All raw data are available in Open Science Framework (https://osf.io/mdpgc/?view_only=3edb6edf2298421e94982268d9802050).