Gas-sensing neurons prime mitochondrial fitness to offset metabolic stress

Abstract

Animals integrate environmental and internal cues to maintain homeostasis and health. The mitochondrial stress response is an essential cytoprotective mechanism, and priming its activation provides a survival advantage. Here, we show that the Caenorhabditis elegans receptor guanylyl cyclase GCY-9 regulates neuropeptide signalling from carbon dioxide sensing neurons to govern a non-canonical mitochondrial stress response in the intestine. This stress response induces atypical mitochondrial chaperone transcription, confers mitochondrial stress resistance, and increases mitochondrial membrane potential and respiration. GCY-9 loss disrupts pathogen avoidance, leading to indiscriminate feeding. We show that starvation decreases GCY-9 expression and propose that the resultant cytoprotective program is launched to offset risks associated with this behaviour. Thus, environmental sensing by peripheral neurons can pre-emptively enhance systemic mitochondrial function in response to metabolic uncertainty.

One-Sentence Summary:

Protecting mitochondria by integrating environmental signals

Coordinating stress responses across tissues is essential for cellular homeostasis and organismal health. Cells employ sophisticated mechanisms to detect and respond to stressors, including the organelle specific mitochondrial and endoplasmic reticulum unfolded protein responses (UPR^mt and UPR^ER) and the cytosolic heat shock response. The ability to anticipate stressors and prime stress responses would offer a powerful cellular and organismal survival advantage. The nervous system is critical for sensing environmental challenges and coordinating stress responses in distal tissues (1–5). Anticipatory signalling offers clear adaptive benefits, especially under fluctuating environmental conditions, however, we understand little of the neuronal and molecular mechanisms that orchestrate this cross-talk.

The Caenorhabditis elegans gas sensing BAG neurons are optimally positioned at the intersection of environmental sensing and physiological adaptability. The BAG neurons are activated by elevated carbon dioxide (CO₂) levels (6), which can influence acid-base homeostasis, cellular respiration, and oxidative stress – all of which are associated with mitochondrial health. Furthermore, our previous studies showed that the BAG neurons influence metabolism and animal behaviour through the ETS-5 transcription factor, an orthologue of mammalian FEV/Pet1 (7, 8). Due to their ability to integrate external and internal environmental cues, we posited the BAG neurons as candidates for anticipating and orchestrating anticipatory mitochondrial stress responses.

Loss of the neuronal transcription factor ETS-5 from the BAG neurons induces a non-canonical UPR^mt

The terminal fate of the BAG neurons – particularly relating to their roles in environmental sensing and systemic metabolism – is mediated by the ETS-5 transcription factor (9, 10). We analysed the hsp-6p::gfp mitochondrial chaperone reporter to determine whether ETS-5 influences the systemic UPR^mt. HSP-6/mtHsp70 is required for importing nuclear-encoded proteins into the mitochondrial matrix and is a key target of the UPR^mt. We found that two independent ets-5 deletion mutants exhibited increased intestinal hsp-6p::gfp expression (Fig. 1A and B). Typically, UPR^mt proteins are upregulated in concert (11–13). However, we found that a second UPR^mt reporter for the mitochondrial matrix chaperonin HSP-60 (hsp-60p::gfp), showed decreased intestinal expression in the ets-5 deletion mutants (Fig. 1D and E). Aberrant hsp-6p::gfp and hsp-60p::gfp expression in ets-5 mutant animals was rescued by driving ets-5 cDNA with the BAG neuron specific gcy-9 promoter (Fig. 1C and F). Canonically, hsp-6 and hsp-60 transcription is upregulated in the UPR^mt by the master transcription factor ATFS-1/ATF5, supplemented by the homeodomain transcription factor DVE-1/SATB and its co-activator UBL-5/Hub1 (11, 14, 15). To examine the requirement of these factors for hsp-6p::gfp upregulation in ets-5 mutant animals, we performed RNA-mediated interference (RNAi) knockdown of ATFS-1, DVE-1 and UBL-5. We used an unc-25 mutant as a positive control for canonical hsp-6p::gfp induction (5). We found that atfs-1 and ubl-5, but not dve-1, are required for the hsp-6p::gfp induction observed in ets-5 mutant animals (Fig. 1G). To confirm DVE-1 independence, we analysed a DVE-1 fusion protein (DVE-1::GFP) which localises to nuclei upon activation (Fig. S1) (15). We found no difference in the number of intestinal nuclear puncta in unstressed or stressed conditions (Fig 1H and Fig S1A) or in the DVE-1 nuclear:cytosolic fluorescence ratio (Fig. S1B–C) between wild-type and ets-5 mutant animals. These data imply that DVE-1 is not required for the atypical mitochondrial stress response observed in ets-5 mutant animals.

Fig. 1. Loss of the neuronal transcription factor ETS-5 from the BAG neurons induces a non-canonical UPR^mt.

A, B Quantification (A) and DIC/fluorescent micrographs (B) of UPR^mt reporter (hsp-6p::gfp) expression in L4 larvae of wild-type, ets-5(tm1734) and ets-5(tm1755) animals. (C) Quantification of UPR^mt reporter (hsp-6p::gfp) expression in L4 larvae of wild-type, ets-5(tm1734), and ets-5(tm1734); gcy-9p::ets-5 cDNA animals. D, E Quantification (D) and DIC/fluorescent micrographs (E) of UPR^mt reporter (hsp-60p::gfp) expression in L4 larvae of wild-type, ets-5(tm1734) and ets-5(tm1755) animals. (F) Quantification of UPR^mt reporter (hsp-60p::gfp) expression in L4 larvae of wild-type, ets-5(tm1734), and ets-5(tm1734); gcy-9p::ets-5 cDNA animals. (G) Quantification of UPR^mt reporter (hsp-6p::gfp) expression in L4 larvae of wild-type and ets-5(tm1734) animals grown on empty vector (EV), ubl-5, dve-1 or atfs-1 RNAi from hatch. (H) Quantification of DVE-1::GFP nuclear puncta in wild-type and ets-5(tm1734) animals. n = 30 animals (A, C, D, F and G) or n = 60 animals (H). P values assessed by one-way analysis of variance (ANOVA) with Tukey’s post hoc test (A, C, D, F and G), or unpaired t test with Welch’s correction (H). Error bars indicate SEM. Scale bar, 250 μm.

Cellular stress responses can be independently activated or integrated across multiple organelles and cellular compartments. We therefore analysed the ER stress response marker hsp-4p::gfp (HSPA6/BiP), the cytosolic stress marker hsp-16.2p::gfp (HSBP1), and two oxidative stress markers sod-3p::gfp (SOD2) and gst-4p::gfp (GSTA4) in ets-5 mutant animals. We detected no change in ER stress (Fig. S2A and B), cytosolic stress (Fig. S2C and D) or oxidative stress (Fig. S2E–H) responses in ets-5 mutants. Together, these data reveal that the BAG neurons control a highly specific and non-canonical systemic UPR^mt (HSP-6 upregulation and HSP-60 downregulation) through the ETS-5 transcription factor.

Loss of ETS-5 improves mitochondrial health readouts

As HSP-6/HSP-60 expression in ets-5 mutant animals is atypical, we aimed to determine how mitochondrial fitness is impacted. First, we performed an acute paraquat survival assay to assess mitochondrial stress resistance in ets-5 mutant animals. Paraquat is a herbicide that disrupts complex I of the electron transport chain, resulting in increased superoxide levels and organismal death (16). We found that ets-5 mutant animals were resistant to paraquat exposure compared to wild-type animals (Fig. 2A). Wild-type levels of paraquat resistance were restored to ets-5 mutant animals by resupplying ets-5 cDNA specifically to the BAG neurons (Fig. 2A). Second, we assessed mitochondrial membrane potential in ets-5 mutant animals using tetramethylrhodamine ethyl ester (TMRE) staining. Higher membrane potential, which is beneficial for mitochondrial function, leads to higher TMRE absorption into the mitochondrial matrix (17). We found that ets-5 mutant animals exhibit increased TMRE fluorescence intensity (Fig. 2B and C), suggesting increased mitochondrial membrane potential. Wild-type levels of TMRE fluorescence intensity were restored to ets-5 mutant animals by resupplying ets-5 cDNA specifically to the BAG neurons (Fig. 2B and C). Finally, we found that loss of ets-5 increases both basal and maximal oxygen (O₂) consumption rates (Fig. 2D–F), suggesting increased capacity for energy production. Together, these data indicate that the atypical stress response induced by ETS-5 loss from the BAG neurons is beneficial for systemic mitochondrial fitness, both at rest and under stress.

Fig. 2. Loss of ETS-5 improves mitochondrial health readouts.

(A) Survival analysis of wild-type, ets-5(tm1734), and ets-5(tm1734); gcy-9p::ets-5 cDNA animals exposed to 200 mM paraquat from the L4 larval stage. n = 66, 74, 72 animals (top to bottom). B, C Quantification (B) and DIC/fluorescent micrographs (C) of mitochondrial membrane potential reporter TMRE staining intensity in L4 larvae of wild-type, ets-5(tm1734), and ets-5(tm1734); gcy-9p::ets-5 cDNA animals. n = 30. Scale bars, 250μm. D-F Seahorse oxygen consumption rate (OCR) assay. (D) Kinetic OCR analysis of wild-type and ets-5(tm1734) animals, measuring (E) basal, and (F) maximal oxygen consumption. n = 3 replicates (~100 animals per replicate). P values assessed by two-way analysis of variance (ANOVA) with Tukey’s post hoc test (A) one-way ANOVA with Tukey’s post hoc test (B), or unpaired t test with Welch’s correction (E and F). Error bars indicate SEM.

BAG-derived neuropeptides mediate intestinal mitochondrial stress responses

How does ETS-5, acting in the BAG neurons, impact mitochondrial health in the non-innervated C. elegans intestine? We hypothesized that specific neuropeptides may act as BAG-intestinal signals in this context. To investigate this, we used the auxin inducible degron system to degrade UNC-31/CAPS – a phosphoinositide-binding protein required for dense core vesicle release – specifically in the BAG neurons by expressing TIR1 under the BAG specific gcy-9 promoter, as we performed previously (18). When these animals are fed auxin, neuropeptide release is inhibited specifically from the BAG neurons. We found that auxin exposure increased hsp-6p::gfp and decreased hsp-60p::gfp expression – phenocopying ets-5 mutant animals (Fig. 3A and B). This implies that the BAG neurons utilise neuropeptides to control systemic mitochondrial health.

Fig. 3. BAG-derived neuropeptides mediate intestinal mitochondrial stress responses.

(A, B) Quantification of UPR^mt reporter (hsp-6p::gfp) (A) and (hsp-60p::gfp) (B) expression in L4 larvae of wild-type and unc-31 AID; gcy-9p::TIR1 animals, with and without auxin treatment. C-E Quantification (C and D) and DIC/fluorescent micrographs (E) of UPR^mt reporter (hsp-6p::gfp) expression in L4 larvae of wild-type, ins-1(nj32) and gcy-9p::ins-1; ins-1(nj32) (C) and wild-type, flp-19(ok2461) and gcy-9p::flp-19; flp-19(ok2461) (D) animals. F-H Quantification (F and G) and DIC/fluorescent micrographs (H) of UPR^mt reporter (hsp-60p::gfp) expression in L4 larvae of wild-type, ins-1(nj-32) and gcy-9p::ins-1; ins-1(nj32) (F) and wild-type, flp-19(ok2461) and gcy-9p::flp-19; flp-19(ok2461) (G) animals. n = 30. P values assessed by one-way analysis of variance (ANOVA) with Tukey’s post hoc test. Error bars indicate SEM. Scale bar, 250 μm.

To determine which BAG-expressed neuropeptides control the distal mitochondrial stress response, we screened neuropeptide mutants for changes to hsp-6p::gfp and hsp-60p::gfp expression (Fig. S3). We screened insulin-like peptides (INS-1, INS-14 and INS-29) and FMRF-like peptides (FLP-4, FLP-6, FLP-10, FLP-12, FLP-13, FLP-17 and FLP-19), all of which (except FLP-17) are expressed in multiple neurons in addition to the BAG neurons. Of these genes, loss of flp-4 was the only neuropeptide mutant that showed a significant increase in both hsp-6p::gfp and hsp-60p::gfp levels (Fig. S3C and D). Interestingly, ins-1 and flp-19 mutants both showed increased hsp-6p::gfp and decreased hsp-60p::gfp expression. As both ins-1 and flp-19 are transcriptionally regulated by ETS-5 in the BAG neurons (7, 8), and their combined loss causes UPR^mt dysregulation that is equal to loss of ets-5 (Fig. S4), we proposed that the BAG neurons utilise INS-1 and FLP-19 to mediate systemic mitochondrial function. To confirm this, we resupplied a single copy of ins-1 and flp-19 genomic DNA specifically to the BAG neurons in the respective mutants, and found that this rescued the increased hsp-6p::gfp (Fig. 3C–E) and decreased hsp-60p::gfp (Fig. 3F–H) expression.

ETS-5 utilises BAG-derived INS-1 and FLP-19 to mediate systemic mitochondrial health

As ins-1 and flp-19 mutant animals exhibit the same atypical chaperone protein expression pattern as ets-5 mutants, we examined whether this translated to the same enhanced mitochondrial fitness. We measured TMRE staining intensity and found that loss of both ins-1 and flp-19 was equivalent, but not additive, to loss of ets-5 in increasing mitochondrial membrane potential (Fig. 4A–B). We also found that loss of ins-1 or flp-19 conferred paraquat resistance (Fig. S5A and B), and the ins-1; flp-19 double mutant was equivalent to loss of ets-5 (Fig. 4C). Conversely, loss of flp-4, which led to increased hsp-6p::gfp and hsp-60p::gfp expression, did not affect paraquat sensitivity (Fig. S5C). Whereas, loss of flp-17, which caused increased hsp-6p::gfp, and no change to hsp-60p::gfp, resulted in a mild resistance to paraquat (Fig. S5D).

Fig. 4. ETS-5 utilises BAG-derived INS-1 and FLP-19 to mediate systemic mitochondrial health.

A, B Quantification (A) and DIC/fluorescent micrographs (B) of mitochondrial membrane potential reporter TMRE staining intensity in L4 larvae of wild-type, ets-5(tm1734), ins-1(nj32); flp-19(ok2461), ins-1(nj32); flp-19(ok2461) and ins-1(nj32); flp-19(ok2461) ets-5(tm1734) animals. n = 30. Scale bars, 250μm. (C) Survival analysis of wild-type, ets-5(tm1734), ins-1(nj32); flp-19(ok2461) and ins-1(nj32); flp-19(ok2461) ets-5(tm1734) animals exposed to 200 mM paraquat from the L4 larval stage. n = 69, 74, 70, 71 animals (top to bottom). (D) Survival analysis of wild-type, frpr-9(sy1294), flp-19(ok2461) and frpr-9(sy1294); flp-19(ok2461) animals exposed to 200 mM paraquat from the L4 larval stage. n = 65, 67, 66, 72 animals (top to bottom). (E) Survival analysis of wild-type, frpr-9(sy1294), ges-1p::frpr-9; frpr-9(sy1294) and rab-3p::frpr-9; frpr-9(sy1294) animals exposed to 200 mM paraquat from the L4 larval stage. n = 65, 68, 67, 68 animals (top to bottom). F, G Quantification (F) and DIC/fluorescent micrographs (G) of mitochondrial membrane potential reporter TMRE staining intensity in L4 larvae of wild-type, frpr-9(sy1294), flp-19(ok2461), frpr-9(sy1294); flp-19(ok2461), ges-1p::frpr-9; frpr-9(sy1294) and rab-3p::frpr-9; frpr-9(sy1294) animals. n = 30. P values assessed by two-way analysis of variance (ANOVA) with Tukey’s post hoc test (A, D and E) or one-way ANOVA with Tukey’s post hoc test (B and F). Error bars indicate SEM. Scale bars, 250μm.

BAG derived FLP-19 meditates systemic mitochondrial health via intestinal receptor FRPR-9

DAF-2 is the only insulin-like growth factor receptor in C. elegans, with well-documented roles relating to mitochondrial function that stem from many pathways. For example, DAF-2 inactivation in the intestine promotes paraquat resistance (19). We therefore focused our research on the recently discovered FLP-19 receptor, FRPR-9 (20). We found that frpr-9 loss induces increased mitochondrial stress resistance (Fig. 4D) and increased mitochondrial membrane potential (Fig. 4F–G). Importantly, these phenotypes were not additive to loss of flp-19 (Fig. 4D, F–G). FRPR-9 is expressed in both neurons and intestinal cells (21). To determine where FRPR-9 acts to mediate intestinal mitochondrial health, we resupplied frpr-9 genomic DNA in the intestine (ges-1 promoter) or pan-neuronally (rab-3 promoter). We found that intestinal frpr-9 expression, but not neuronal frpr-9 expression, restored wild-type paraquat survival (Fig. 4E) and mitochondrial membrane potential to frpr-9 mutant animals (Fig. 4F–G). These data imply that FLP-19 released from the BAG neurons binds to FRPR-9 on intestinal cells to mediate systemic mitochondrial health.

BAG carbon dioxide machinery controls systemic mitochondrial health

We previously showed that flp-19 expression in the BAG neurons depends on the CO - sensing receptor guanylate cyclase GCY-9, a transcriptional target of ETS-5 (22). We therefore wondered if ins-1 expression is also influenced by gcy-9 loss. We found that the number of BAG neurons expressing an ins-1 transcriptional reporter is drastically reduced in a gcy-9 deletion mutant (Fig. 5A and B). We therefore hypothesised that GCY-9 – and therefore the ability to sense CO₂ – may be mechanistically involved in the BAG-mediated mitochondrial stress response. Indeed, we found that gcy-9 loss caused increased hsp-6p::gfp, and decreased hsp-60p::gfp expression, which was rescued by expressing gcy-9 cDNA specifically in the BAG neurons (Fig. 5C and D). A gcy-9 ets-5 double mutant was not additive to either single mutant with respect to hsp-6p::gfp and hsp-60p::gfp regulation (Fig. S6), indicating that ets-5 and gcy-9 act in the same pathway in the BAG neurons to mediate the systemic UPR^mt. Furthermore, gcy-9 loss conferred increased mitochondrial stress resistance (Fig. 5E) and increased mitochondrial membrane potential, phenotypes that were not enhanced by simultaneous ets-5 loss (Fig. 5F and G). Intriguingly, and consistent with previous studies, we found that gcy-9 loss did not increase intestinal fat storage (23) (Fig. S7), in contrast to ets-5 mutants, which accumulate excess intestinal fat (7). Furthermore, ets-5 expression is not influenced by loss of gcy-9 (Fig. S8). These results imply that increased fat storage and the associated behavioural phenotypes observed in ets-5 mutant animals exist in a separate, but potentially linked, molecular pathway to the stress response pathway described here.

Fig. 5. BAG carbon dioxide machinery controls systemic mitochondrial health.

A, B Quantification (A) and fluorescent micrographs (B) of ins-1p::gfp reporter expression in the BAG neurons in L4 larvae of wild-type and gcy-9(n4470) animals. n = 30. Scale bar, 50μm. (C) Quantification of UPR^mt reporter (hsp-6p::gfp) expression in L4 larvae of wild-type, gcy-9(n4470) and gcy-9p::gcy-9; gcy-9(n4470) animals. n = 30. (D) Quantification of UPR^mt reporter (hsp-60p::gfp) expression in L4 larvae of wild-type, gcy-9(n4470) and gcy-9p::gcy-9; gcy-9(n4470) animals. n = 30. (E) Survival analysis of wild-type, ets-5(tm1734), gcy-9(n4470) and ets-5(tm1734) gcy-9(n4470) animals exposed to 200 mM paraquat from the L4 larval stage. n = 68, 73, 67, 71 animals (top to bottom). F, G Quantification (F) and DIC/fluorescent micrographs (G) of mitochondrial membrane potential reporter TMRE staining intensity in L4 larvae of wild-type, ets-5(tm1734), gcy-9(n4470) and ets-5(tm1734) gcy-9(n4470) animals. n = 30. Scale bar, 250μm. H, I Quantification (H) and DIC/fluorescent micrographs (I) of gcy-9p::gfp reporter expression in BAG neurons in L4 larvae of wild-type exposed to live, dead or no bacteria for 24 hours. n = 30. Scale bar, 50μm. P values assessed by unpaired t test with Welch’s correction (A) one-way ANOVA with Tukey’s post hoc test (B), or (C, D, F and H) or two-way analysis of variance (ANOVA) with Tukey’s post hoc test (E). Error bars indicate SEM.

CO₂ is a metabolic by-product, and fluctuations in environmental CO₂ can signal key environmental factors such as food quality, overcrowding, and hypoxia. Why would impaired CO₂ sensing enhance stress resistance? We propose that loss of peripheral CO₂ sensing functions as a “blackout alarm”, reflecting a failure to detect environmental status and triggering a primed stress response in high metabolic tissues such as the intestine. The unique stress response described here likely represents an adaptive failsafe triggered to ensure survival in ephemeral environments.

Previous studies showed that gcy-9 mutants fail to avoid pathogenic bacteria (24). Combined with our data, this raises the possibility that gcy-9 expression – under the transcriptional control of ETS-5 – is dynamically regulated in response to environmental cues such as food. We thus wondered if, during starvation, animals might downregulate gcy-9 expression in the BAG neurons to suppress avoidance of potentially pathogenic but nutritionally valuable bacteria. Indeed, we found that gcy-9 transcriptional reporter expression decreased following starvation or exposure to dead, metabolically inactive bacteria (Fig. 5H and I). We propose that this transcriptional mechanism balances feeding with cellular protection and exists to buffer the risk associated with non-discriminatory feeding when food is scarce. By suppressing gcy-9 expression, animals increase their likelihood of consuming available food, while simultaneously activating an atypical systemic mitochondrial stress response that primes tissues to offset the potential harm of metabolic stress. This hypothesis is supported by previous research, revealing that FRPR-9 loss enhanced resistance to both gram negative and gram-positive pathogenic bacteria (25).

Discussion

Here, we identify a pair of peripheral sensory neurons (the C. elegans gas-sensing BAG neurons) that mediate protective plasticity through non-autonomous priming of the intestinal mitochondrial stress response, enabling organismal resistance to metabolic stress. Within the BAG neurons, the ETS-5–GCY-9 regulatory axis controls expression of the INS-1 and FLP-19 neuropeptides that signal to the intestine. We show that the FLP-19 receptor FRPR-9 acts in the intestine to modulate mitochondrial membrane potential and stress resistance.

The specificity of the ETS-5-controlled UPR^mt may be due to a distinctive combination of transcription regulators used that does not involve the DVE-1 homeodomain transcription factor. DVE-1 and UBL-5 function together to positively regulate both hsp-6 and hsp-60 transcription (15). However, previous studies have shown that DVE-1 can act independently of UBL-5 to induce the UPR^mt and regulate lifespan (26, 27). UBL-5 can also act independently of DVE-1 in yeast (28), recently confirmed in C. elegans (29), where it functions in pre-mRNA alternative splicing. This mechanism may be conserved in humans, as human UBL5 interacts with the cyclin-like kinase CLK4, which plays an important role in spliceosome formation (30). These studies provide evidence to support additional roles for UBL-5 outside of DVE-1 coactivation in C. elegans.

In our model, we propose that loss of CO detection through GCY-9 in the BAG neurons functions as a “blackout alarm,” prompting anticipatory stress responses in metabolically active tissues such as the intestine. Priming stress responses to sensory deficits has been observed in other systems. In mice, ablation of olfactory sensory neurons promotes resistance to diet induced obesity and improves insulin sensitivity, suggesting that sensory input impacts systemic metabolism (31).

Integration of environmental and metabolic cues by the BAG neurons to mediate systemic physiology suggests an intriguing evolutionary parallel to mammalian carotid bodies. The carotid bodies are chemoreceptors located bilaterally at the bifurcation of the common carotid arteries that detect changes in arterial blood CO₂ and O₂. They also respond to metabolic signals such as blood glucose, insulin, and adrenalin (32–34). Overactivation of the carotid bodies during sleep apnoea is associated with increased fasting glucose, insulin resistance, reduced glucose tolerance and impaired pancreatic β cell function (32, 35, 36). Carotid body ablation restores insulin sensitivity and glucose tolerance in rats with diet-induced metabolic syndrome (37), and denervation prevents the development of hypertension and insulin resistance induced by hypercaloric diets (33). However, bilateral carotid body resection confers a significant risk of O₂ desaturation during mild hypoxia (38). Therefore, research into pharmacological modification of key biomolecular pathways in peripheral chemosensors controlling systemic metabolism – such as described here – is required. The BAG neurons also share functional similarities with pancreatic β cells. ETS-5 directly regulates transcription of the insulin-like peptide gene, ins-1, in the BAG neurons (8). Likewise, the mammalian ETS-5 orthologue – Pet1 – positively regulates insulin expression in mouse pancreatic β cells (39). Thus, appropriate regulation of insulin signalling is likely essential for maintaining systemic mitochondrial and metabolic function across evolution.

These functional comparisons to the carotid bodies and pancreatic β cells position the BAG neurons as a powerful evolutionary model for investigating how sensory perception, behaviour and systemic mitochondrial fitness are integrated. Collectively, our findings underscore the nervous system’s capacity to prime protective mitochondrial responses in peripheral tissues by interpreting changes in sensory information.

Supplementary Material

Materials and Methods

Figs. S1 to S8

Table S1

References (18, 40–47)

Data S1

Funding Statement

National Health and Medical Research Council grants GNT1137645 (RP) and GNT2000766 (RP, AH)

Data and materials availability:

All data is available in the main text or supplementary materials. In addition, Source Data are provided with this paper. There are no accession codes, unique identifiers, or weblinks in our study and no restrictions on data availability. Materials will be available upon request from the Pocock laboratory.