Distinct mechanisms of non-autonomous UPRER mediated by GABAergic, glutamatergic, and octopaminergic neurons

The capacity to deal with stress declines during the aging process, and preservation of cellular stress responses is critical to healthy aging. The unfolded protein response of the endoplasmic reticulum (UPRER) is one such conserved mechanism, which is critical for the maintenance of several major functions of the ER during stress, including protein folding and lipid metabolism. Hyperactivation of the UPRER by overexpression of the major transcription factor, xbp-1s, solely in neurons drives lifespan extension as neurons send a neurotransmitter-based signal to other tissue to activate UPRER in a non-autonomous fashion. Previous work identified serotonergic and dopaminergic neurons in this signaling paradigm. To further expand our understanding of the neural circuitry that underlies the non-autonomous signaling of ER stress, we activated UPRER solely in glutamatergic, octopaminergic, and GABAergic neurons in C. elegans and paired whole-body transcriptomic analysis with functional assays. We found that UPRER-induced signals from glutamatergic neurons increased expression of canonical protein homeostasis pathways and octopaminergic neurons promoted pathogen response pathways, while minor, but statistically significant changes were observed in lipid metabolism-related genes with GABAergic UPRER activation. These findings provide further evidence for the distinct role neuronal subtypes play in driving the diverse response to ER stress.


Introduction
Regulation of organelle homeostasis is essential for maintenance of cellular health, which has direct implications in organismal health and longevity.The endoplasmic reticulum (ER) is one such organelle, which processes about a third of the proteins and lipids in cells and has dedicated quality control machineries to preserve these and numerous other functions.One primary quality control machinery is the ER unfolded protein response (UPR ER ), a transcriptional response to ER damage or stress, which activates genes essential for maintenance of proper ER function 1 .Activation of the UPR ER involves the ER-membrane protein, inositol-requiring enzyme 1 (IRE-1), which dimerizes upon sensing ER stress to splice X-box binding protein (xbp-1) mRNA into xbp-1s.xbp-1s mRNA encodes a functional transcription factor, XBP-1s, which activates genes essential for restoring ER homeostasis, including protein chaperones, autophagy, ubiquitin proteosome system, among others.This transcriptional response to stress is essential for maintaining proper function of the ER and has direct implications in longevity.Specifically, UPR ER function has been shown to decline with age, and heightened activation of UPR ER maintained stress resilience at old age in C. elegans 2 .Overexpression of xbp-1s solely in neurons is sufficient to enhance C. elegans lifespan due to the whole-body activation UPR ER through neuron-to-body communication medicated by neurotransmitter signaling 2 .Upon neuronal UPR ER activation, a complex signaling event mediated by a combination of dopamine, serotonin 3 , and tyramine 4 , results in dramatic remodeling of peripheral cells.Specifically, intestinal cells can activate proteostasis 4 , lipid metabolism 5,6 , and lysosomal function 7 to drive longevity.These studies revealed numerous neuronal subtypes and distinct mechanistic pathways, including chaperone induction downstream of serotonergic signaling, lipid remodeling through lipophagy downstream of dopaminergic signaling, and proteostasis machinery through tyramine signaling from RIM and RIC neurons, all of which are essential to promote longevity.Finally, this UPR ER signaling is not limited to neurons, as a recent study revealed several glial subtypes were also capable of eliciting a glia-to-body UPR ER signaling event to promote longevity 8 .
Similar homeostatic benefits of UPR ER in neurons are observed in mammals, wherein Xbp1s expression in pro-opiomelanocortin (POMC) neurons has been shown to protect against dietinduced obesity by improving leptin and insulin sensitivity under ER stress 9 .While all these studies utilized an artificial transgenic expression system, two recent studies have shown that neuron-to-body UPR ER signaling is an essential signaling pathway for endogenous pathways.In mice, olfactory perception of food is sufficient to promote POMC Xbp1s expression and activation of post-prandial liver ER adaption 10 .In C. elegans, chemosensation of pathogenic bacteria was found to promote neuronal xbp1-s expression, leading to UPR ER activation in peripheral tissues and extension of lifespan 11 .These studies revealed that endogenous neuronto-body signaling utilized similar mechanistic pathways to xbp-1s overexpression paradigms, which highlight the translatability of using transgenic approaches to dissect the neuronal circuitry of UPR ER signaling.
Building on previous research, we were interested in understanding whether other neuronal subtypes are involved in neuron-to-body UPR ER activity.We sought to determine whether glutamatergic, GABAergic, and octopaminergic neurons are necessary and/or sufficient to drive neuron-to-body communication of the UPR ER in C. elegans.We accomplished this by overexpressing xbp-1s overexpression in these neuronal subtypes and assessing measurements of general health, such as lifespan, healthspan, ER function, and stress resilience.Further, we performed a comprehensive transcriptomic analysis to identify potential mechanistic pathways that drive phenotypic outcomes in these neuronal subtype UPR ER paradigms.

Overexpression of xbp-1s in glutamatergic, octopaminergic, and GABAergic neurons.
In previous studies, serotonergic, dopaminergic, and RIM/RIC neurons have been identified to be involved in neuron-to-body communication of UPR ER 3,4 .However, these four neuron subtypes make up only ~18 of the 302 neurons in C. elegans, raising the question of what other neuronal subtypes may be involved in neuron-to-body UPR ER communication.Previously, we performed a screen of several neurotransmitter signaling pathways involved in neuronal communication of UPR ER , which revealed glutamatergic, octopaminergic, and GABAergic neurons as candidates involved in this signaling event 3 .Glutamate is a widely utilized, excitatory neurotransmitter in both invertebrate and vertebrate systems 12 ; octopamine is a C. elegans-specific neurotransmitter similar to the mammalian norepinephrine, and is involved in immune response 13 ; and gamma-aminobutyric acid (GABA) is a widely utilized neurotransmitter that has been found to function as both an excitatory and an inhibitory signal in C. elegans 14 .C. elegans possess 79 glutamatergic neurons, 2 octopaminergic neurons, and 32 GABAergic neurons in hermaphrodites (Fig. S1A) (Loer CM, Worm Atlas).
To determine the potential involvement of glutamatergic, octopaminergic, and GABAergic neurons in neuron-to-body communication of UPR ER , we overexpressed xbp-1s in each neuronal subtype using the eat-4 promoter for xbp-1s overexpression in glutamatergic neurons 15 (hereafter referred to as glutamatergic xbp-1s); tbh-1 promoter for xbp-1s overexpression in octopaminergic neurons 16 (hereafter referred to as octopaminergic xbp-1s); and the unc-25 promoter for xbp-1s overexpression in GABAergic neurons 17 (hereafter referred to as GABAergic xbp-1s).We confirmed by quantitative PCR (qPCR) that all three subtypes display an increase in xbp-1s mRNA.Although our data did not reach statistical significance, all three neuronal xbp-1s subtypes did display an increase in xbp-1s expression (Fig. S1B).

Glutamatergic, octopaminergic, and GABAergic xbp-1s alter distinct transcriptional pathways.
To more thoroughly investigate the impact of neuronal subtype UPR ER on the periphery, we performed whole-worm RNA sequencing on animals overexpressing xbp-1s in glutamatergic, octopaminergic, and GABAergic neurons.Glutamatergic and octopaminergic xbp-1s resulted in sizable changes to gene expression, while more mild changes occurred with GABAergic xbp-1s (Fig. 1A-C, Table S1).Interestingly, the majority of differentially expressed genes were unique to each condition, suggesting distinct responses were induced by each neuronal subtype (Fig. 1D, Table S2).This adds more insight into a previous study that identified distinct pathways activated downstream of serotonergic and dopaminergic xbp-1s 3 .
To further characterize the similarities and differences between peripheral response to neuronal subtype UPR ER , we directly compared our glutamatergic, octopaminergic, and GABAergic xbp-1s animals to previously published RNA-seq datasets 3,8 .First, we sought to determine the overlap between neuronal subtype xbp-1s overexpression with pan-neuronal xbp-1s overexpression (hereafter referred to as neuronal xbp-1s), as we would expect that neuronal xbp-1s includes each neuronal subtype.We compared neuronal xbp-1s using two different promoters, rab-3p and rgef-1p, and were surprised to find that while there was significant overlap between these two neuronal xbp-1s strains, as a majority of differentially expressed genes were not shared (Fig. S2A, Table S2).Since this could potentially be due to leakiness of the rab-3p compared to the rgef-1p 18,19 , in our subsequent studies, we focused on making comparisons to results from the rgef-1p::xbp-1s strain (which we will continue to refer to as neuronal xbp-1s).
As expected, neuronal xbp-1s animals display altered expression of a large number of direct XBP-1s targets 20 .Interestingly, we see that glutamatergic xbp-1s similarly induces many of these XBP-1s targets and to an even greater extent than neuronal xbp-1s (Fig. 1E, Table S3).These data suggest that glutamatergic xbp-1s activates a more canonical UPR ER signature involved in conventional protein processing pathways.Gene ontology (GO) enrichment analysis supported this idea, as the most enriched biological processes included pathways related to ER function and protein homoeostasis, including ER to Golgi vesicle-mediated transport, protein N-linked glycosylation, endoplasmic-reticulum-associated protein degradation (ERAD) pathway, and proteolysis (Fig. 1F, Table S4).However, a majority of differentially expressed genes in glutamatergic xbp-1s are still distinct from neuronal xbp-1s (Fig. S2B, Table S2), suggesting that these protein homeostatic pathways are being regulated in different ways in each condition.Interestingly, glutamatergic xbp-1s transcriptionally regulates an entirely different set of genes than serotonergic xbp-1s (Fig. S2C, Table S2), although these animals were also shown to induce canonical protein homeostasis pathways 3 .Altogether, these data show that even amongst neuronal subtypes that share a response (e.g., protein homeostasis), the specific genes targeted in these similar pathways are distinct, highlighting the fact that non-autonomous UPR ER is dramatically different based on which neuronal subtype is involved.
Octopaminergic xbp-1s showed smaller gene expression changes to XBP-1s targets in comparison to glutamatergic xbp-1s, being more reminiscent of the levels found in neuronal xbp-1s (Fig. 1E, Table S3).However, similar to glutamatergic xbp-1s, when all differentially expressed genes were compared, the majority of differentially expressed genes were distinct (Fig. S2B).The differentially expressed genes identified were entirely different from those found in serotonergic and dopaminergic xbp-1s (Fig. S2C-D, Table S2).GO analysis identified that the most dramatic changes in gene expression in octopaminergic xbp-1s were defense response pathways, particularly those involved in immune response (Fig. 1G, S2E, Table S3).These data are consistent with previous findings that showed pathogen response in C. elegans is associated with UPR ER induction 21 and a role for non-autonomous signaling in this response 11 , potentially through octopaminergic signaling 13 .These data add an additional downstream function of non-autonomous UPRER in regulation of immune response, potentially downstream of octopaminergic neurons.
Finally, GABAergic xbp-1s activation caused minimal changes in gene expression overall, with very little overlap with other neuronal subtype xbp-1s (Fig 1E, S2B-D).Although the gene expression changes were minor, GO analysis did reveal some pathways previously associated with UPR ER induction, including lipid remodeling 3,5 (Fig. 1H, S2F, Table S3-4).Altogether, our data adds more evidence to the previously proposed model 22 that specific neuronal subtypes participate in activation of unique downstream pathways in response to stress.

Glutamatergic xbp-1s promotes protein homeostasis and ER stress resilience.
To determine the physiological impact of the transcriptional changes, we next tested the ability of animals with neuronal subtype xbp-1s to resist specific stressors.Tunicamycin is a wellcharacterized ER stressor which blocks N-linked glycosylation in the ER 23 , and animals with neuronal xbp-1s exhibit an increased resistance to tunicamycin. 2 .Consistent with our transcriptomics data, glutamatergic xbp-1s resulted in a small but significant increase in tunicamycin resistance (Fig. 2A), while octopaminergic (Fig. 2B) and GABAergic activation (Fig. 2C) had no effect on resistance to ER stress.Next, to determine whether UPR ER activation in these neurons can affect proteostasis specifically in distal tissue, we crossed glutamatergic, octopaminergic, and GABAergic overexpressing xbp-1s strains into animals expressing fluorescently-tagged aggregation-prone polyglutamine repeats in the intestine 24 and assessed the extent of aggregation as these animals aged.Strikingly, a significant decrease in fluorescence intensity was observed at days 1 and 5 of adulthood for all three strains as compared to controls, with octopaminergic xbp-1s expression demonstrating the greatest reduction in fluorescence (Fig 2D-E).This suggests that while only glutamatergic xbp-1s improved ER proteotoxic stress resistance, all neuronal subtype xbp-1s animals have improved peripheral protein homeostasis.

Glutamatergic, octopaminergic, and GABAergic xbp-1s improve immune function.
In addition to changes in protein homeostasis pathways, our transcriptomics analysis revealed that glutamatergic, octopaminergic, and GABAergic xbp-1s displayed significant changes in immune response related genes, with octopaminergic xbp-1s animals having defense response against bacteria as one of the most significantly enriched GO terms.Therefore, we measured the impact of xbp-1s overexpression on innate immune response using multiple methods.First, we used a standard pathogen resistance assay using exposure to Pseudomonas aeruginosa 25 .Using a canonical P. aeruginosa fast kill assay, we found that glutamatergic, octopaminergic, and GABAergic xbp-1s animals all displayed a significant increase in survival against PA14, with the octopaminergic xbp-1s displaying the most significant increase in survival even after 8 hours (Fig. 3A).This is consistent with the octopaminergic xbp-1s animals having the highest change in expression of genes associated with immune response and previous reports that indicate a functional role for octopamine signaling in innate immunity in C. elegans 13 .In addition, our data add further evidence that animals with improved protein homeostasis have improved immune response, as C. elegans innate immunity is a protein-synthesis dependent process 21 .
C. elegans also utilize their nervous system for aversive learning behavior to avoid pathogenic bacteria 26 .This avoidance behavior is mediated by several neurotransmitters, including serotonin 27 and octopamine 28 , and certain strains with heightened stress responses have been shown to lack this typical avoidance behavior 29 .Here, we used a previously validated forced exposure method 29 to determine the impact of neuronal xbp-1s overexpression on pathogen apathy.Similar to pathogen resistance, glutamatergic, octopaminergic, and GABAergic xbp-1s animals all displayed increased apathy to pathogens, with glutamatergic animals having the mildest phenotype (Fig. 3B).Thus, it is likely that the heightened resistance to pathogens is directly correlated with a lack of urgency to escape these pathogens.While we did observe an increase in expression of innate immune response genes, it is also possible that the increase in pathogen resistance is due to an increase in gut barrier integrity, as age-associated loss of gut barrier integrity results in infiltration of bacteria and bacterial colonization in the gut 30,31 .Interestingly, glutamatergic, octopaminergic, and GABAergic xbp-1s animals all showed similar breakdown of gut barrier integrity and age-associated bacterial colonization in the gut compared to wildtype controls, with octopaminergic xbp-1s animals having a trend for worse gut barrier integrity (Fig. 3C-D), despite having shown the highest level of resistance to pathogens (Fig. 3A).These data suggest that the pathogen resistance and apathy of glutamatergic, octopaminergic, and GABAergic xbp-1s animals is likely due to a heightened immune response, rather than a gut-barrier-related phenotype.

Glutamatergic, octopaminergic, and GABAergic xbp-1s do not alter general organismal health.
Next, we sought to test the impact of neuronal subtype xbp-1s on general organismal health, as previous studies have shown that neuronal xbp-1s results in a significant improvement in longevity and animal health, with a reduction in reproductive health 2,16 .Interestingly, we did not find any changes in general lifespan in glutamatergic, octopaminergic, or GABAergic xbp-1s animals (Fig. S3).In addition, while we saw a mild decrease in brood size in glutamatergic, octopaminergic, or GABAergic xbp-1s animals, these differences were not statistically significant (Fig. S4A-C).Finally, general organismal health was also unchanged as no change in motility was observed (Fig. S4D-F).
In addition to improvements in general health, neuronal xbp-1s animals were shown to exhibit changes in ER morphology associated with increased lysosomal function and autophagy 6,7 .These changes were associated with a general increase in secretory capacity of the ER and depletion of lipids, potentially through an increase in lipophagy 6 .Therefore, to measure general changes to the ER, we first performed imaging of the ER using an mRuby::HDEL fused to an HSP-4 signal sequence to localize the fluorophore to the ER 6 .Since we could not successfully make homozygous octopaminergic xbp-1s animals with this mRuby::HDEL marker, we used an mCherry::HDEL fused to a SEL-1 signal sequence, which previous studies have shown display similar ER morphology 6 .Using these ER-localized fluorophores, we did not observe major changes to ER morphology in glutamatergic, octopaminergic, or GABAergic xbp-1s animals (Fig. S5).Next, to measure ER secretory capacity, we utilized the yolk protein marker, VIT-2::GFP.This maternal yolk protein is secreted by the intestinal ER in adults and subsequently endocytosed by developing eggs and is a commonly used marker for secretory capacity 32,33 .Although fluorescent levels appear to be higher in intact animals (Fig. S6A), when we quantitatively measured fluorescent levels in isolated eggs, there was no significant change in VIT-2::GFP signal (Fig. S6B-C), suggesting that there are no changes to ER secretory capacity in glutamatergic, octopaminergic, or GABAergic xbp-1s animals.
Finally, we measured lipid levels using DHS-3::GFP an abundant protein on the surface of C. elegans intestinal lipid droplets 34,35 .Interestingly, we saw a significant decrease in lipid droplet abundance in glutamatergic, octopaminergic, or GABAergic xbp-1s animals (Fig. 4A), despite no major changes in lipid droplet size or morphology (Fig. 4B).To further evaluate changes in lipid content, we utilized a more comprehensive dye, Oil Red O (ORO), which stains neutral lipids, cholesteryl esters, and lipoproteins 36 .Consistent with lipid droplet imaging, we observed a significant decrease in lipid content in glutamatergic, octopaminergic, or GABAergic xbp-1s animals using ORO (Fig. 4C-D).These data suggest that similar to other paradigms of neuronal xbp-1s overexpression 3,5,6 , glutamatergic, octopaminergic, or GABAergic xbp-1s animals results in depletion of neutral lipids, likely resulting in improved lipid homeostasis.

Discussion
The UPR ER is involved in diverse cellular processes that impact organismal health, including proteostasis 2,37 , autophagy 7 , lipid metabolism 5,6 , and immune response 38 .Many of these functions can occur in a non-autonomous fashion, whereby neural cells with XBP-1s activation can coordinate a signal to the body to coordinate a homeostatic response 22 .Numerous neural circuits have been implicated in this response, including serotonin, dopamine 3 , tyramine 4 , RIM/RIC interneurons 11 , and glial cells 8 .This study adds to this complex neural circuitry, adding additional functional roles for glutamatergic, octopaminergic, and GABAergic neurons in non-autonomous UPR ER signaling.
In our current study, we overexpress xbp-1s in glutamatergic, octopaminergic, and GABAergic neurons, which requires several considerations.First, promoter strength and neuron number can drive different phenotypes across each neuronal xbp-1s overexpression paradigm, which have nothing to do with the biological significance of neuron identity.This is especially important to consider in the face of using two different neuron-specific promoters, rab-3p and rgef-1p, displaying dramatically different downstream transcriptional responses.However, our data showed that xbp-1s expression level alone does not purely drive phenotypic outcome.While previous studies utilizing pan-neuronal, serotonergic, or dopaminergic xbp-1s showed a significant increase in xbp-1s overexpression 3 , in our study, although we can see a trend for an increase in xbp-1s expression in glutamatergic, octopaminergic, and GABAergic xbp-1s, these data did not reach statistical significance.Despite this lack of a significant increase in xbp-1s levels, we still saw dramatic changes in transcriptome, especially in glutamatergic and octopaminergic xbp-1s, with many genes implicated as canonical XBP-1s targets 39 , including protein homeostasis 2 and immune response 21 .In addition, while octopaminergic neurons represent the lowest number of total neurons (2 neurons) compared to glutamatergic (79) and GABAergic (34), octopaminergic xbp-1s displayed some of the strongest phenotypes observed in our study.Taken together, these data argue that neuron number and promoter strength alone do not drive phenotypic outcomes, and neuronal identity is a critical factor in non-autonomous signaling, even when using an artificial system such as xbp-1s overexpression.
Another potential limitation of ectopic gene overexpression is whether our findings correlate with the endogenous roles these neuronal subtypes play in signaling or gene regulation.Previous work in numerous animal models provide sufficient evidence that this is the case: in C. elegans, olfactory sensation of pathogenic bacteria utilize neuron-to-body XBP-1s signaling through TGFβ signaling to improve longevity and healthspan 11 .In mice, Xbp1s overexpression in POMC neurons promotes adipose tissue UPR ER to improve metabolic health 9 , which is very similar to hepatic Xbp1s activation to promote metabolic health downstream of food perception 10 .In D. melanogaster, glutamate signaling can promote lipid mobilization as a systemic metabolite, altering lipid metabolism 40 .These reports suggest that even ectopic genetic models can provide mechanistic insight into important endogenous physiological processes.One interesting finding in our study is that although GABAergic xbp-1s had minimal changes to gene expression, these animals still displayed significant changes to organismal health, including improved proteostasis and immune response.While transcriptional changes are not the only change that could translate to physiology, as altered protein function, organelle dynamics, and metabolism can all occur in the absence of transcriptional change, technical limitations could also be responsible for the lack of difference observed in GABAergic xbp-1s.
Here, we used whole-worm transcriptomics, and it is entirely possible that opposite changes in gene expression in different tissues could result in a net result of no change.Indeed, in terms of xbp-1s overexpression, this is seen where whole-body overexpression of xbp-1s does not result in lifespan extension, likely do to the summation of negative effects in the muscle and positive effects in the intestine and neurons 2 .Thus, it is entirely possible that GABAergic xbp-1s may drive differential effects in different tissue, as it does drive depletion of lipids and increased protein homeostasis in the intestine.Future tissue-specific studies can reveal whether these physiological outputs are dependent on gene expression changes in the intestine in these animals.
One final consideration is that in this study, we are generalizing neuronal signals as separate entities, although neuronal circuits are often intertwined and complex 41 .There are known neuron-neuron interactions even amongst the limited neuronal subtypes in this study.For example, dopamine and octopamine coordinate appetite regulation in crickets 42 ; and excitatory glutamatergic and inhibitory GABAergic synapses strategically converge on specific cell types of the brain 43 .While our previous study has separated the utility of dopamine in serotonergic xbp-1s signaling and vice versa 3 , this does not preclude the convergence of other neuronal subtypes.For example, we find that although glutamatergic xbp-1s animals display mostly changes to protein homeostasis-related pathways, they also show improved immune function and lipid depletion.While the increase in protein homeostasis could be responsible for increased immune function 21 , it is entirely possible that glutamatergic neurons can also recruit octopaminergic or GABAergic signaling to alter immune response and lipid metabolism.Future studies mapping the neural circuitry across subtypes will be necessary to develop a full neural map of non-autonomous XBP-1s signaling.Overall, our study adds to the complex literature of non-autonomous XBP-1s signaling, adding three additional neuronal subtypes to this rapidly expanding map.

C. elegans maintenance
All strains utilized in this investigation are derived from the N2 wild-type worm sourced from the Caenorhabditis Genetics Center (CGC) and are detailed below.The worms are maintained at 15°C, fed with OP50 E. coli B strain.Animals are bleached and L1 arrested as described below for all experimentation and transferred to growth conditions at 20°C, utilizing HT115 E. coli K strain for all experiments.Experiments employed HT115 bacteria carrying an empty pL4440 vector referred to as empty vector (EV).

Bleaching
Experiments were conducted on animals of the same age, synchronized using a standard bleaching protocol.Worms were collected into a 15 mL conical tube using M9 solution (22 mM KH2PO4 monobasic, 42.3 mM NaHPO4, 85.6 mM NaCl, 1 mM MgSO4) and subjected to a bleaching solution (1.8% sodium hypochlorite, 0.375 M NaOH in M9) until complete digestion of carcasses.Intact eggs were then washed four times with M9 solution by centrifugation at 1.1 RCF for 30 seconds.After the final wash, animals were L1 arrested by incubating overnight in M9 at 20°C on a rotator for a maximum of 24 hours.

Transgenic strain synthesis
The sequence for xbp-1s expression was defined as per 2 and is provided below.Coding sequences were cloned from cDNA synthesized via reverse transcriptase using RNA isolated from N2 worms, the endogenous eat-4p, tbh-1p, and unc-25p promoter was cloned from gDNA isolated from N2 worms, and an unc-54 3′UTR was cloned from gDNA isolated from N2 worms.Plasmids were injected into N2 worms using a standard microinjection protocol as described 44 with 10 ng/μl of overexpression plasmid, 2.5 ng/μl of myo-2p::mCherry or myo-2p::GFP as a coinjection marker.Both injections and integration of constructs were performed by SUNY Biotech.All integrated animals were then backcrossed to our N2 lines to eliminate mutations and create an isogenic line.All sequences used in this manuscript are as follows:

C. elegans microscopy Stereoscope
For whole-worm imaging of vha-6p::Q40::YFP, DHS-3::GFP, and VIT-2::GFP strains, synchronized animals were grown on RNAi plates plated with HT115 empty pL4440 vector from L1 stage.Animals were imaged at day 1 of adulthood and for aging experiments also imaged at day 5 and day 9 of adulthood.10+ animals were placed in a pool of 100 mM sodium azide in M9 on standard NGM plates without bacteria to induce paralysis.Paralyzed animals were then lined alongside each other and imaged on a Leica M205FCA automated fluorescent stereomicroscope running LAS X software and equipped with a standard GFP filter, Leica LED3 light source, and Leica K5 camera.For all imaging experiments, 3 biological replicates were performed with 2 technical replicates each, and 1 representative image was chosen for use in figures.For vha-6p::Q40::YFP quantification, Fiji 45 was used to draw a region of interest along the posterior half of each group of worms, and integrated density was measured.Graphing and statistical analysis was performed with GraphPad Prism 10 software using a Mann-Whitney test.

Widefield and confocal imaging
Widefield imaging utilized a Leica THUNDER Imager equipped with a 63x/1.4Plan AproChromat objective, standard dsRed filter (11525309), Leica DFC9000 GT camera, a Leica LED5 light source, and run on LAS X software.For high resolution imaging of DHS-3::GFP and ER morphology, imaging was performed using a Leica Stellaris 5 confocal microscope equipped with a white light laser source and spectral filters, HyD detectors, 63x/1.4Plan ApoChromat objective, and run on LAS X software.Animals were placed in 100 mM sodium azide solution on a glass slide to induce paralysis and imaged within 5 minutes of slide preparation to prevent artifacts from prolonged exposure to sodium azide.For all imaging experiments, 3 biological replicates were performed with 2 technical replicates each, and 1 representative image was chosen for use in figures.

Intestinal bacteria invasion assay
Assessing intestinal bacteria invasion was performed as previously described 46 .Animals were L1 synchronized via bleaching and plated on RNAi plates containing a bacterial lawn derived from a mixture of 80% HT115 bacteria containing empty pL4440 vector and 20% HT115 bacteria expressing mCherry.Once at the desired age, animals were manually transferred onto a standard OP50 plate and allowed to feed on OP50 for 2 hours at 20°C to facilitate clearance of mCherry bacteria.For imaging, worms were paralyzed by exposure to M9 solution containing 100 mM sodium azide and arranged on a standard NGM plate without bacteria.Images were captured using a Leica M205FCA fluorescent stereomicroscope equipped with a standard dsRed filter as described above.For each of the 3 biological replicates, 2 technical replicates with 13 animals per replicate were performed and 1 representative image was used for figures.The percentage of animals exhibiting bacterial invasion was quantified and plotted with GraphPad Prism 10 software for each technical and biological replicate.Statistical analysis was conducted across all replicates using a Mann-Whitney test with GraphPad Prism 10 software.

ER Secretion Assay
Assaying of ER secretory function was performed as described previously 47 .Transgenic control animals and eat-4p::xbp-1s, tbh-1p::xbp-1s, and unc-25p::xbp-1s animals expressing VIT-2::GFP were bleached to obtain eggs.Eggs were then placed on glass slides and imaged using a Leica THUNDER Imager equipped with a 63x/1.4Plan AproChromat objective, standard dsRed filter (11525309), Leica DFC9000 GT camera, a Leica LED5 light source, and run on LAS X software.Images were quantified using Fiji and drawing a region of interest around each individual egg to obtain an integrated density value.4 independent biological replicates were performed.SuperPlots 48 were created using GraphPad Prism 10 software where large dots represent the median value of each biological replicate and small dots represent single eggs with different intensities of colors representing eggs from the same biological replicate; lines indicate median and interquartile range.All statistical analyses were conducted using Mann-Whitney testing with GraphPad Prism 10 software.For whole worm imaging, animals were raised to day 3 of adulthood and imaged using the Leica M205FCA fluorescent stereomicroscope equipped with a standard GFP filter as described above.

Oil Red O Staining
Oil Red O fat staining was performed as previously described 49 .Briefly, worms were bleached, and eggs were plated to obtain a synchronous population.Worms were grown on RNAi plates with a lawn of HT115 bacteria containing an empty pL4440 vector and aged to day 3 of adulthood or day 5 of adulthood.Aging was performed in the absence of FUDR by gravity settling in M9 solution and aspirating to remove progeny.For staining, worms were washed off plates using a PBS + 0.01% Triton solution, rocked for 3 minutes in 40% isopropyl alcohol, pelleted, and then stained with Oil Red O in diH2O for 2 hours while rocking at room temperature.Worms were pelleted and washed in PBS + 0.01% Triton for 30 min before being imaged at 20× magnification with a Leica THUNDER Imager Flexacam C3 color camera and run on LAS X software.To quantify somatic fat depletion, worms were scored as previously described 50 .The level and distribution of fat was placed into categories of non-somatic lipid depletion, displaying no loss of fat and being darkly stained throughout the body, and somatic lipid depletion, being stained largely in the germ cells.At least 100 worms were scored for each condition over 3 biological replicates.

Caenorhabditis elegans RT-qPCR and RNA-seq analysis
For collection of RNA, we used glp-4(bn2) animals to eliminate progeny.After bleaching L1 arresting, all animals were raised at 22°C (the restrictive temperature for our backcrossed glp-4(bn2) strain) for 3 days to collect animals at day 1 of adulthood.Approximately 1000 animals were used per condition.Worms were collected using M9 and transferred to TRIzol solution and underwent 3 freeze/thaw cycles between liquid nitrogen and a 37°C bead bath with a 30-second vortexing step between each cycle to lyse worms.Following the final thaw, chloroform was added at a ratio of 1:5 chloroform/TRIzol, and aqueous separation of RNA was achieved by centrifugation using a heavy gel phase-lock tube (VWR, 10847-802).The aqueous phase was mixed with isopropanol at a 1:1 ratio and applied to a QuantaBio Extracta Plus RNA kit (95214) for RNA purification according to the manufacturer's instructions.
Library preparation and sequencing was conducted at Novogene using their standard pipeline using paired-end, polyA selection, first-strand synthesis, and an Illumina NovaSeq6000.Each condition was measured with 3 biological replicates.Gene expression levels were quantified using kallisto 51 with WBcel235 as the reference genome.Fold changes were determined using DESeq2 52 .Gene targets of XBP-1 were defined based on previous experimental findings 53 .Gene Ontology (GO) enrichment analysis was performed using WormEnrichr 54,55 .rgef-1p, dat-1p, tph-1p, eat-4p, tbh-1p, and unc-25p driven expression of xbp-1s was compared to the N2 wild-type control.
For RT-qPCR, cDNA synthesis was conducted using qScript cDNA SuperMix (QuantaBio, 101414-102) with 500 ng of RNA.RT-qPCR was performed using NEB Q5 DNA polymerase following the manufacturer's guidelines and utilizing the primers listed below.Each condition was assessed using 3 biological replicates.QuantStudio 3 (Thermo Fisher) was used for quantification using a standard curve method.

Lifespan measurements
For lifespan experiments, animals were grown on RNAi plates on either EV or RNAi bacteria from L1 stage.At day 1 of adulthood, animals were washed off plates with M9 solution and then moved to plates containing 100 µL of 100 mg/mL FUDR to eliminate progeny.One replicate of lifespan assays was performed using a lifespan machine 56 with others being done by hand.Tunicamycin and thermotolerance lifespan assays were performed based on established protocols 57 .For tunicamycin assays, animals were moved onto plates supplemented with 25 µg/mL tunicamycin in DMSO directly in the plate.Animals were grown at 20°C and checked every 2 days for viability.For thermotolerance assays, day 1 adult animals were placed at 34°C and scored for viability every 2 hours.Animals were considered dead if they did not exhibit any movement when prodded with a platinum wire at both the head and the tail.Animals that exhibited bagging, intestinal leakage, desiccation on the side of the plate, or other deaths unrelated to aging were scored as censored.All lifespans were performed on 3 biological replicates.Lifespan assay survival curves were plotted using GraphPad Prism 10 software and statistics were performed using a Log-Rank test in GraphPad Prism 10.
Representative data are depicted in figures and a table of all lifespan assays performed is available in Table S5.
Caenorhabditis elegans brood size assay Brood assays were measured as previously described 57 .Bleaching was used to obtain a synchronized population of animals, and 10 L4 stage animals were transferred onto individual plates.Every 12 hours, the animals were moved onto new plates, while plates containing eggs were stored in a 15°C incubator for 2-3 days.All surviving progeny on each egg-laying plate were counted and totaled to determine the brood size.SuperPlots 48 were created using GraphPad Prism 10 software where large dots represent the median value of each biological replicate and small dots represent single animals with different intensities of colors representing animals from the same biological replicate; lines indicate median and interquartile range.All statistical analyses were conducted using Mann-Whitney testing with GraphPad Prism 10 software.

Caenorhabditis elegans thrashing assay
Thrashing assays were conducted on animals synchronized via bleaching and aged on RNAi plates containing FUDR from day 1 of adulthood.Upon reaching the desired age, plates containing adult animals were flooded with 100 μL of M9 solution, and 30-second videos were recorded using an M205FCA stereomicroscope equipped with a Leica K5 microscope running LAS X software.Thrashing movements were manually recorded over a 10 second period.A bending of more than 50% of the animal's body in the opposite direction was deemed a single thrash.Representative data from 3 independent biological replicates are presented.SuperPlots 48 were created using GraphPad Prism 10 software where large dots represent the median value of each biological replicate and small dots represent single animals with different intensities of colors representing animals from the same biological replicate; lines indicate median and interquartile range.All statistical analyses were conducted using Mann-Whitney testing with GraphPad Prism 10 software.

Fast kill assay
Fast kill assays were performed as previously described with minor modifications 58,59 .
Pseudomonas aeruginosa (PA14) cultures were grown overnight at 37°C for 14-15 hours.5 μL of overnight culture was spread over 3.5 cm peptone glucose media plates (1% Bacto-Peptone, 1% NaCl, 1% glucose, 1.7% Bacto Agar) containing 0.15 M sorbitol, using a spreader made from an open loop tipped glass pasture pipette.The plates were incubated at 37°C for 24 hours and then at 25°C for 48 hours.Following this, 30 to 40 synchronized L4 animals were placed on each plate.Assays were performed at 25°C.Survival of animals was plotted over a period of 8 hours with intervals of 2 hours.An animal was deemed dead when it no longer responded to touch. 3 biological replicates with 3 technical replicates were performed for a total of 9 replicates.Survival rates were measured wherein 100% survival was indicated as an integer "1" and the fraction of survival populations at every time point was represented as decimal values.

Forced food choice
Forced food choice assays were performed as described previously 29,60 .Plates were made using the same recipe for NGM plates aside from the addition of 0.35% peptone.A single colony of Pseudomonas aeruginosa (PA14) and E. coli OP50 bacteria were inoculated into separate 3 mL of LB for overnight primary culture at 37C.The following day, the OD600 of each of the cultures was diluted to an OD600 of 1.0.A culture of PA14 was transferred as a line along the center of an NGM plate using a glass Pasteur pipette bent at a 90-degree angle.Next, 15 μL of OP50 culture was seeded as a dot onto the plate 2.5 cm away from the center and 0.5 cm away from the edge of the plate.The plates were dried and transferred to 37C for 24 hours, followed by incubation at 25C for 48 hours.On the day of the assay, the plates were removed from the 25C incubator and allowed to reach room temperature.Worms were washed three times in M9 solution before being placed onto the assay plate diametrically opposite to the OP50 dot at 2.5 cm away from the center.The proportion of worms found on or off each food was recorded after 1 hour, 2 hours, 4 hours, 6 hours, and 8 hours.After each time point, the population was scored in which -1 represented 100% of the population on PA14 and +1 represented 100% of the population on OP50.The movement index was then calculated towards the OP50 dot using the formula: ("A" population of worm on OP50 dot -"B" population of worms on PA14 line) ("A" population of worm on OP50 dot + "B" population of worms on PA14 line) Each assay was done in biological triplicate with technical triplicates for a total of 9 replicates.Statistical analyses were conducted using Mann-Whitney testing with GraphPad Prism 10 software.
Foundation for Medical Research and AFAR Grant for Junior Faculty Award.Some strains were provided by the CGC, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440).Some gene analysis was performed using Wormbase, which is funded on a U41 grant HG002223.S1.Heat map of (E) lipid homeostasis (GO:0055088) or (F) immune response (GO:0006955) gene expression under pan-neuronal, glutamatergic, octopaminergic, and GABAergic xbp-1s overexpression.Warmer colors indicate increased expression, and cooler colors indicate decreased expression.See Table S3.Lifespans were scored every 2 days and data is representative of 3 biological replicates.All statistical analysis is available in Table S5.

Fig. S4 .
Fig. S4.Glutamatergic, octopaminergic, and GABAergic xbp-1s does not increase healthspan.Measurements of fecundity of control (blue) and (A) glutamatergic xbp-1s (green, eat-4p), (B) octopaminergic xbp-1s (yellow, tbh-1p), and (C) GABAergic xbp-1s (pink, unc-25p) animals.Total number of eggs that hatched were counted per animal.Measurements of thrashing of control (blue) and (D) glutamatergic xbp-1s (green, eat-4p), (E) octopaminergic xbp-1s (yellow, tbh-1p), and (F) GABAergic xbp-1s (pink, unc-25p) animals.Number of thrasheswas assessed over a 10 second period in animals at day 1 adult (young), day 4-5 adult (middle), and day 9 adult (old) in M9 solution with each thrash being counted as a movement from a concave to a convex formation.For SuperPlots, each small dot represents a single animal with various intensities of colors representing independent biological replicates and each large dot is the median value of each biological replicate.Lines represent the median across all biological replicates and whiskers indicate interquartile range.Statistical analysis was performed using a Mann-Whitney test.