Isoflurane aggravates pre-existing proteotoxicity in adult nematodes by suppressing mitochondrial fitness

Abstract

Post operative delirium (POD) is an acute complication, characterized by fluctuating attention and confusion, which may develop and persist into cognitive decline. POD is most commonly observed in elderly patients, particularly those with preexisting cognitive impairments. Isoflurane, a widely used volatile anesthetic, is associated with POD. However. how exposure to isoflurane affects the integrity of the proteome is largely obscure. Utilizing the nematode C. elegans, we found that isoflurane leads to a long-lasting decline in protein homeostasis (proteostasis) of adult animals that express neurodegeneration-causing, abnormally long poly-glutamine stretches. Isoflurane-induced proteostasis impairments are dependent on the aging-regulating transcription factors DAF-16/FOXO and SKN-1/NRF, and can be alleviated by the knockdown of certain components of the mitophagy mechanism. Accordingly, induction of mitochondrial biogenesis protects worms that are challenged by protein aggregation from isoflurane-induced proteotoxicity. Our observations provide novel insights into the mechanism that links isoflurane, proteotoxicity and POD, and highlight the potential of mitophagy modulators as alleviators of POD.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-38591-8.

Introduction

Post-operative delirium (POD) is a common complication that manifests as inattention, lethargy, and confusion that can last for several months¹. The prevalence of POD is tightly and positively correlated with old age² and is most common among elder individuals with preexisting cognitive impairments³. While the exact cause of POD remains debated, some evidence suggest that volatile anesthetics commonly used in clinical settings, enable its manifestation⁴. The observation that POD is most common among elder patients raised the prospect that the aging process plays mechanistic roles in exposing the elderly to this complication. It is possible that aging actively decreases the activities of homeostasis-promoting mechanisms, which reduce the vulnerability of younger individuals to POD. However, how aging influences POD prevalence, and whether homeostasis-promoting mechanisms mitigate its onset in early stages of life, remain largely unknown.

One of the hallmarks of aging is the deterioration of protein homeostasis (proteostasis)⁵. The production of properly folded, functional proteins and the maintenance of their integrity throughout their lifecycle, are challenges faced by organisms across phyla. To maintain a pristine proteome, a nexus of mechanisms act in concert to assist nascent polypeptides in attaining their correct folding, ensure their proper interactions with other molecules, and chaperone them to their accurate cellular locations⁶. Nonetheless, despite the orchestrated activities of the “proteostasis network”, subsets of aggregation-prone proteins fail to fold properly and are directed for degradation by the ubiquitin proteasome system (UPS), or by autophagy⁷. Early in life the proteostasis network efficiently maintains the integrity of the proteome, alas, as the organism ages, the competence of this network declines and subsets of aggregation-prone proteins form insoluble aggregates that accrue within cells⁶. This process jeopardizes cellular and organismal functionality and underlies the development of various maladies that are collectively known as “proteinopathies”⁸. Neurodegenerative illnesses consist of a large and prevalent group of proteinopathies. Abnormally long stretches of polyglutamine (polyQ) in the sequences of several proteins, render them prone-to-aggregate and underlie the development of a subset of late-onset neurodegenerative disorders⁹. These include Huntington’s disease (HD) and Machado Joseph Disease¹⁰. The fact that polyQ-associated disorders are characterized by delirium and cognitive decline, together with the observation that preexisting cognitive impairment is a strong predictor of POD¹¹, suggest that anesthetics-induced POD emanates, at least partially, from the induction or exacerbation of abnormal protein aggregation. Indeed, several studies unveiled that exposure of nematodes¹² and mice¹³ to volatile anesthetics results in the aggregation of human neurodegeneration-causing proteins. Specifically, exposure to isoflurane has led to the aggregation of long polyQ stretches in mammalian cells¹⁴. Moreover, the alteration of aging by the knockdown of daf-2, a prominent aging¹⁵ and proteostasis¹⁶ regulating receptor, protects worms from anesthetic-induced neurotoxicity¹⁷. These observations culminate to link anesthetic-induced toxic protein aggregation (proteotoxicity) with aging and preexisting proteome instability. However, whether isoflurane-induced proteotoxicity is associated with aging and what mechanisms underlie this manifestation, are largely unanswered questions.

Here we employed the nematode Caenorhabditis elegans (C. elegans), to investigate how isoflurane impairs proteostasis in adult worms. While nematodes are distant from humans, fundamental proteostasis mechanisms have been shown to be highly conserved from worms to mammals. These include chaperones that assist and supervise the correct folding of nascent polypeptides, as well as autophagy and the UPS¹⁸. In addition, the nematode is amenable to aging alterations and to the modulation of gene expression by RNA interference (RNAi), nominating this organism as a preferred model for the study of proteostasis¹⁹. Our discoveries indicate that a single exposure, of 3 h to isoflurane is sufficient to modulate gene expression and promote proteotoxicity in adult worms that are challenged by preexisting polyQ-mediated, chronic proteotoxic stress. This deleterious effect of isoflurane has led to the onset of a motility impairment which is associated with mitophagy modulations. Accordingly, reducing PINK-1-dependent mitophagy or administrating a mitophagy/mitochondrial biogenesis regulator, protect worms from isoflurane-induced, polyQ35-YFP-mediated proteotoxicity.

These results suggest that isoflurane exacerbates pre-existing proteotoxicity by suppressing mitochondrial fitness, a phenomenon that may underlie the manifestation of POD. Therefore, a preoperative treatment with mitochondrial fitness regulators may offer therapeutic strategies to mitigate POD in patients with preexisting cognitive impairments.

Results

Isoflurane exacerbates polyQ35-YFP toxicity in adult nematodes

To study the effects of isoflurane on proteotoxicity in the context of aging, we utilized wild-type C. elegans nematodes (strain N2) and examined whether this anesthetic drug influences motility, a phenotype which declines in an age-dependent manner²⁰. The worms were exposed to 8% isoflurane for 3 h on day 3 of adulthood, and subjected to the “thrashing assay”. This assay serves as a measure of locomotor activity, whose decline is indicative of reduced proteostasis, and is suitable for polyQ35-YFP-expressing worms as a measure of proteotoxicity¹⁹. The concentration of 8% isoflurane is equivalent to ~2.66 Monitored Anesthesia Care (MAC) units, and known to induce deep anesthesia and complete immobilization of worms without acute lethality²¹. Thrashing rates were recorded 24 and 72 h after removal from isoflurane. No difference in thrashing rates was observed among isoflurane-treated animals and the control group, in both time points (Fig. 1A). These results indicate that isoflurane does not impair the motility of wild-type animals that are not challenged by a chronic proteotoxic insult.

Fig. 1.

Exposure to isoflurane leads to motility impairments when proteostasis is consistently hampered. (A) No motility impairments were seen in populations of wild-type worms (N2) 24 and 72 hours after exposure to isoflurane. (B,C) Day 1-old AM140 (Q35^muscle) worms that were exposed to isoflurane exhibit significantly reduced thrashing rates 72 hours after exposure (B) however, this proteotoxic effect could not be observed when worms of the same strain were treated with yfp RNAi (C). (D) Isoflurane promotes proteotoxicity in AM1126 animals that express polyQ35-YFP in their neurons, 24 and 48 hours after exposure.

Since previous studies have shown that anesthetics promotes the aggregation of neurodegenerative-associated proteins in different models^13,14, we asked whether isoflurane hampers proteostasis in worms that chronically express aggregation-prone, neurodegeneration-causing proteins. To examine this, we utilized animals that express stretches of 35 glutamine repeats fused to the yellow fluorescent protein (YFP), in their body wall muscles (strain AM140, hereafter also labelled as “Q35^muscle”). The expression of polyQ35-YFP in these worms results in the formation of visible foci that are associated with aging-dependent motility impairments²⁰. Using the thrashing assay, we tested whether exposure to isoflurane aggravates polyQ35-YFP-mediated proteotoxicity. Day 1-old AM140 (Q35^muscle) worms were either exposed to isoflurane for 3 hours or left untreated and subjected to thrashing assays 24 or 72 h after exposure. Exposure to isoflurane did not significantly affect the rate of motility at day 2-old worms (young adults, 24 h after exposure). In contrast, at day 4 of adulthood (post-reproductive adults, 72 h after exposure), isoflurane-treated animals showed a significant reduction in their thrashing rates compared to untreated worms (Fig. 1B). Proteotoxic effect was also observed in AM140 (Q35^muscle) worms that were exposed to isoflurane at day 3 of adulthood in both time points, 24 and 72 h after exposure (days 4 and 6 of adulthood, respectively, Fig. S1A).

We further tested the relationship between anesthetic dose and proteotoxicity, by exposing day 3 old AM140 worms to 4% isoflurane (~ 1.33 MAC) for 3 hours. Thrashing rates that were measured 24 hours after exposure, showed no significant differences between worms that exposed to 4% isoflurane and untreated controls (Fig. S1B). These findings indicate that unlike exposure to 8%, moderate levels of isoflurane do not exacerbate polyQ35-YFP–mediated proteotoxicity. To scrutinize the notion that this phenotype of reduced thrashing emanates from the chronic proteotoxic challenge that is imposed by polyQ35-YFP, we created a highly efficient yfp RNAi (Fig. S1C). The knockdown of polyQ35-YFP by yfp RNAi rescued the low motility rates typical to untreated AM140 worms (Fig. S1D).

We also assessed whether the isoflurane-mediated motility impairment stems from polyQ35-YFP. AM140 (Q35^muscle) animals were treated from hatching with the yfp RNAi or left untreated (EV) and exposed to isoflurane at day 3 of adulthood. Thrashing assays that were conducted at day 4 (24 h after exposure), indicated that the knockdown of polyQ35-YFP by RNAi abolishes the motility impairment that was observed in isoflurane-treated AM140 (Q35^muscle) animals that were grown on control bacteria (Fig. 1C). The proteotoxic effect of isoflurane is not exclusive to muscles as animals that express the polyQ35-YFP chimeric protein in their neurons (strain AM1126, hereafter also labelled as “Q35^neuron”), exhibited reduced thrashing rates upon exposure to isoflurane at day 3 of adulthood, compared to the control group (Fig. 1D). In addition, no aggravation of proteotoxicity was seen in yfp RNAi-treated AM1126 (Q35^neuron) worms that were exposed to isoflurane, and in isoflurane-treated animals that express polyQ0-YFP in their neurons (strain AM52) (Fig. S1E). It is important to note that in AM1126-based experiments, thrashing assays were conducted at days 4 and 5 of adulthood (24 and 48 hours post exposure, respectively). This temporal difference among AM140- and AM1126-based experiments emanates from to the different features of these worm strains, which drive the expression of polyQ35-YFP by distinct promoters, and exhibit different expression levels of the proteotoxic protein¹⁹.

In sum, these results indicate that exposure to isoflurane promotes proteotoxicity in adult AM140 (Q35^muscle) and AM1126 (Q35^neuron) worms solely when the animals are challenged by chronic polyQ35-YFP-mediated proteotoxic stress. They also show that the proteotoxic effects of isoflurane in muscles can be reproduced in worms that express polyQ35-YFP in their neurons and thus, are of physiological relevance.

To test whether the proteotoxic-enhancing effect of isoflurane is unique to polyQ stretches, or a general phenomenon, we used worms that express the Alzheimer’s-causing Aβ peptide in either muscles (strain CL2006)²² or in neurons (AGD124)²³. The expression of Aβ results in a progressive paralysis within the population, a phenotype that serves as a measure of proteotoxicity¹⁹. The experimental worms were exposed to isoflurane at days 5 and 9 of adulthood, and rates of paralysis were recorded daily and compared to the rates that were observed in the control groups which were not exposed to the anesthetic compound. Our experiments showed that exposure to isoflurane slightly enhances Aβ-mediated proteotoxicity, however, this effect was significant exclusively at day 12 of adulthood in CL2006 worms (Fig. S1F), but not in AGD1246 animals (Fig. S1G). These results may suggest that isoflurane could promote proteotoxicity when other aggregation-prone proteins are expressed, however, this issue requires further clarification.

Since polyQ35-YFP-containing foci are correlated with toxicity²⁰, we next asked whether exposure to isoflurane affects the number of these structures in AM140 (Q35^muscle) animals. The worms were either exposed to isoflurane at day 1 or 3 of adulthood or left untreated, and polyQ35-YFP-containing foci were visualized and counted. A comparison of the numbers of foci in untreated worms and in their counterparts that were exposed to isoflurane, right after exposure (0h), 24 and 72 h thereafter, unveiled that isoflurane significantly increases the number of foci in all time points. This phenomenon was seen in worms that were exposed at day 1 of adulthood (Fig. 2A and B) as well as in those which were treated with isoflurane at day 3 (Fig. S2A).

Fig. 2.

Isoflurane elevates the amounts of polyQ35-YFP-containing foci but reduces the rates of SDS-resistant aggregates. (A,B) Day 1 old AM140 worms were exposed to isoflurane, visualized by fluorescent microscopy right after exposure, as well as 24 or 72 h thereafter (A), and foci were counted (B). Exposure to isoflurane has led to an increase in the number of foci at all time points. (C,D) Day 5 old AM140 (Q35^muscle) worms were exposed to isoflurane and subjected to the filter trap assay. Isoflurane reduced the amounts of SDS-resistant polyQ35-YFP aggregates (C) in a significant manner (D) (the original full-length uncropped scans are provided in Fig. S2B). (E) Exposure to 8% isoflurane does not reduce the thrashing rates of day 3-old AM140 worms that were treated from hatching with RNAi toward either daf-16 or skn-1. In contrast, the knockdown of daf-2 or hsf-1 did not prevent isoflurane from enhancing polyQ35-YFP-mediated proteotoxicity.

Previous studies have shown that oligomers of neurodegeneration-causing proteins, rather than high molecular weight aggregates, are the most toxic species^24,25. The toxicity of these conformers can be mitigated either by their disaggregation or by hyper-aggregation, which assembles oligomers to create high-molecular-weight aggregates of lower toxicity²⁶. Therefore, we used the filter trap assay¹⁹ to test whether exposure to isoflurane enhances or decreases the aggregation of polyQ35-YFP in day 5-old AM140 (Q35^muscle) worms (post-reproductive) and found that an exposure of 3 hours to isoflurane significantly reduces the amounts of SDS-resistant aggregates in three independent experiments (Fig. 2C and D, Fig. S2B).

Given that the rates of protein aggregation have been shown to be controlled by aging-regulating pathways^16,27, we wondered whether the modulations in polyQ35-YFP aggregation and the aggravation of proteotoxicity by isoflurane, involve these pathways. To address this, we examined whether modifying the activity of the insulin/IGF signaling cascade (IIS), a well-established proteostasis²⁸ and aging-regulating¹⁵ pathway which is highly conserved among worms and mammals²⁹, modifies the effect of isoflurane on polyQ35-YFP-mediated proteotoxicity. AM140 (Q35^muscle) worms were treated from hatching with RNAi toward either daf-2, which encodes the nematode’s sole insulin/IGF receptor³⁰, or toward one of the genes that encode its downstream transcription factors; daf-16, skn-1 and hsf-1¸all are well-established proteostasis regulators^16,31. All worm groups were divided into two subgroups, one was exposed to isoflurane and the other was left untreated. All groups were subjected to a thrashing assay on day 4 of adulthood, 24 hours after exposure to isoflurane, side by side with their untreated counterparts. Since DAF-2 negatively regulates the activity of the aforementioned transcription factors, the knockdown of daf-2 hyper-activates them, thereby promoting proteostasis²⁸. Surprisingly, although the knockdown of daf-2 elevated the rate of thrashing compared to untreated animals, it did not prevent isoflurane from enhancing proteotoxicity (Fig. 2E). Treatment with either daf-16 or hsf-1 RNAi, but not with skn-1 RNAi, significantly reduced thrashing rates, however, exposure to isoflurane did not enhance proteotoxicity in daf-16 and skn-1 RNAi-treated animals. In contrast, the knockdown of hsf-1 did not prevent isoflurane from aggravating proteotoxicity (Fig. 2E). Similar results were obtained when the experiment was repeated using day 1-old AM140 (Q35^muscle) worms (Fig. S2C). These observations suggest that isoflurane may inhibit the activity of DAF-16 and thus, when the expression of this factor is knocked down, isoflurane shows no additional proteotoxic effect. Interestingly, the knockdown of skn-1 did not reduce the rate of thrashing compared to that of control animals, however, it prevented isoflurane from aggravating proteotoxicity. This proposes that SKN-1 may play a deleterious role in worms that are challenged by polyQ35-YFP and exposed to isoflurane.

Collectively, these results imply that exposure to isoflurane reduces the amounts of high-molecular-weight polyQ35-YFP aggregates and elevates the levels of toxic oligomers. The requirement for transcription factors for the enhancement of proteotoxicity by isoflurane and previous indications that this compound modulates gene expression in the mammalian brain³², raise the prospect that isoflurane aggravates proteotoxicity, at least partially, by modulating gene expression. Therefore, we sought to explore how isoflurane affects the nematode’s transcriptomic landscape in the presence and absence of a chronic proteotoxic challenge.

The modulation of gene expression by isoflurane

Two models can explain how isoflurane-induced changes in gene expression may affect polyQ35-YFP-mediated proteotoxicity. One suggests that isoflurane modulates the expression of genes that are needed for the maintenance of proteostasis, regardless of whether the worm is challenged by proteotoxicity or not. Accordingly, worms that are not challenged by proteotoxic stress may exhibit a temporary proteostasis decline which is reversed after the removal of the worms from the anesthetic compound. However, when the worms are consistently challenged by proteotoxicity, the isoflurane-mediated change in gene expression can initiate a persistent proteotoxic process that hampers proteostasis and underlies functional decline. The other model suggests that exposure to isoflurane modulates the expression levels of a specific set of genes exclusively when the worm is challenged by proteotoxicity. Accordingly, this transcriptomic change is accountable for the persistent proteostasis decline that is observed in AM140 (Q35^muscle) and AM1126 (Q35^neuron) worms after exposure to isoflurane. To distinguish between these models, we conducted an RNA sequencing (RNA-seq) experiment to address two key questions: (i) how exposure to isoflurane modulates gene expression in general, and (ii) the expression levels of which genes are specifically modulated by isoflurane when the worms are challenged by high levels of polyQ35-YFP proteotoxicity?

AM140 (Q35^muscle) worms were cultured from hatching on EV bacteria or on bacteria that express YFP RNAi. The worms were harvested at day 5 of adulthood as at this age proteotoxicity is apparent in AM140 animals (Fig. 1B). Therefore, the worms are likely to exhibit gene expression modulations that stems from the proteotoxic challenge. In addition, at day 5 of adulthood the animals are post-reproductive. This allowed us to characterize gene expression modulations in adult tissues, avoiding potential background from developing embryos. Half of the worms of each group were exposed to isoflurane while the other half was not. All four worm groups were harvested 3 h after exposure, RNA was extracted, and gene expression profiles were analyzed in 3 independent sets by RNA seq (Fig. 3A). Fisher’s Exact Test showed significance (Supplementary Table S1). We harvested the worms three hours after exposure to isoflurane as mRNA molecules are relatively short-lived and thus, changes in the transcriptomic landscape are less likely to be detectable after longer time³³.

Fig. 3.

Exposure to isoflurane modifies the transcriptomic landscape of adult worms. (A) An illustration of the experimental workflow of the RNA-seq experiment. (B) A heat map shows similar gene expression modulations by isoflurane in worms that express either high or low polyQ35-YFP levels. (C) 1748 genes exhibit modulated expression levels upon exposure to isoflurane in worms that express low polyQ35-YFP levels, and 1770 in animals that express high levels of the proteotoxic protein. 1374 of these genes showed modulated expression levels in both worm groups showing a highly significant overlap (Fisher’s Exact test, P < 0.0001). (D,E) WormCat-based clustering of genes that show modulated expression levels as a result of exposure to isoflurane, indicates that very similar functions are affected by exposure to isoflurane in untreated worms (D) and in their yfp RNAi-treated counterparts (E). (F) A heatmap shows changes in the expression levels of a selected subset of genes in AM140 (Q35^muscle) worms that were either treated with isoflurane or not, and were fed with either control bacteria (EV) or with bacteria that express yfp RNAi. These genes include chaperones such as hsp-16.2 hsp-16.41 and hsp-70, as well as the neuropeptide-coding genes nlp-11 nlp-29 and nlp-34.

Plotting a heat map (Fig. 3B), we found that the majority of the significantly affected genes (1374 in total) exhibited different expression levels upon exposure to isoflurane regardless of whether the worms expressed high or low polyQ35-YFP levels (Fig. 3C). In addition, 374 genes exhibited modulated expression levels solely when polyQ35-YFP was knocked down, and 396 showed differential expression levels only in worms that expressed high levels of polyQ35-YFP (Fig. 3C). To statistically assess the overlap between the two gene sets, we performed a Fisher’s Exact test, which demonstrated a highly significant concordance (P < 0.0001, two-sided), indicating that the shared transcriptional response is unlikely to occur by chance (Supplementary Table S1). In total, the shared genes constitute ~78% of each gene set.

Using the WormCat computational tool, we clustered all genes that exhibited significantly modulated levels upon exposure to isoflurane and found that genes which are involved in stress responses, metabolism, and proteolysis are prominently affected by this anesthetic compound in both groups (Fig. 3D and E).

Similarly, clustering the affected genes according to molecular functions showed that genes that are involved in monooxygenase and oxidoreductase activities are prominently affected by isoflurane regardless of whether the worms express high or low polyQ35-YFP levels (Fig. S3A). These results support the notion which proposes that the modulation of gene expression by isoflurane, sets the basis for proteostasis decline when the worms are challenged by a proteotoxic insult.

Interestingly, among the affected genes we found that exposure to isoflurane largely increases the levels of a subset of proteostasis-regulating chaperones (Fig. 3F) including hsp-16.2. hsp-16.41, and hsp-70 (C12C8.1)³⁴. An additional group of genes that exhibit elevated expression levels encode for proteases and components of the autophagy mechanism including asp-7 and sqst-2. We also identified increased levels of genes that encode for certain neuropeptides: nlp-11, nlp-29 and nlp-34. Since neuropeptide signaling was reported to regulate proteostasis at the organismal level^35,36, these observations suggest that exposure to isoflurane jeopardizes the integrity of the proteome and may induce a response of the proteostasis network in an attempt to restore protein homeostasis, across tissues.

Next, we asked which genes exhibit modulated expression levels upon exposure to isoflurane solely when the worms express high polyQ35-YFP levels (EV, Fig. S3B) and which show modulation only in worms that express low levels of the proteotoxic protein (yfp RNAi, Fig. S3C). Interestingly, among the 396 genes that were significantly affected only in worms that express high polyQ35-YFP levels, we identified a few proteostasis-associated genes that were down-regulated. These include cyn-17, a gene that encodes for a chaperone member of the cyclophilin family. Mutations in cyclophilin recognition sites play key roles in the onset of a familial prion disorder³⁷ and inherited Alzheimer’s disease³⁸. Genes that encode proteins of the F-box family, which are involved in protein ubiquitination and degradation³⁹, as well as txt-4 which enhances the activity of the proteostasis-promoting, trans chaperone signaling (TCS) pathway,⁴⁰ were also down-regulated.

Using the thrashing assay, we tested whether the knockdown of the three most significantly affected genes by isoflurane solely in worms that express high polyQ35-YFP levels, abolishes the proteotoxic effects of isoflurane. AM140 (Q35^muscle) worms were treated from hatching with RNAi toward the indicated three genes. At day 3 of adulthood the animals were exposed to isoflurane and subjected to thrashing assays 24 hours thereafter. The knockdown of any of these genes did not prevent isoflurane from enhancing polyQ35-YFP-mediated proteotoxicity (Fig. S3D). Utilizing the same approach, we asked whether the effect of isoflurane on polyQ35-YFP proteotoxicity is dependent on genes that code for the chaperones HSP-16.2 and HSP-16.41 or for the neuropeptides NLP-11and NLP-29, all exhibited modulated expression levels in our RNA-seq dataset. To address this, we conducted thrashing assays utilizing AM140 (Q35^muscle) worms that were treated with RNAi toward the aforementioned genes. Our results indicated that the knockdown of hsp-16.41, but not of any of the other tested genes, prevents isoflurane from enhancing polyQ35-YFP-mediated proteotoxicity (Fig. S3E). To further test this observation, we created an additional specific hsp-16.41 RNAi, and conducted a thrashing assay using this new construct and AM140 (Q35^muscle) animals. Our results validated the deleterious role of hsp-16.41 in the context of isoflurane exposure and polyQ35-YFP proteotoxicity (Fig. S3F).

Collectively, our observations suggest that exposure to isoflurane lowers the proteomic fitness of the worm by hampering the activities of proteostasis-promoting mechanisms, such as stress response. They also imply that isoflurane affects the integrity of the proteome, at least partially, at the post-translational level and that cells respond to this hazardous impairment by elevating the expression levels of chaperones (Fig. S3B) that attempt to restore proteostasis. Nevertheless, the chaperone HSP-16.41 appears to play a deleterious role in the face of isoflurane. To explore the effects of isoflurane at the proteomic level, we next asked which proteins aggregate upon exposure to this compound and whether proteotoxic stress is a prerequisite for this aggregation.

Proteomic destabilization by isoflurane in worms that are challenged by polyQ35-YFP proteotoxicity

To test the hypothesis that isoflurane differently destabilizes subsets of aggregation-prone proteins in worms that are challenged by proteotoxic stress and in their counterparts that express low level of the same proteotoxic protein, we cultured AM140 (Q35^muscle) worms on either control (EV) or on YFP RNAi bacteria. At day 5 of adulthood, the worms were exposed to isoflurane for 3 hours and harvested 3 hours thereafter. Aggregated proteins were sedimented by ultracentrifugation and identified by mass spectrometry using label-free quantification (LFQ) (Fig. 4A). Fisher’s Exact Test showed significance (Supplementary Table S2).

Fig. 4.

Analysis of aggregated proteins in isoflurane-treated and control worms. (A) A description of the experimental workflow of the proteomic analysis of aggregated proteins in control and isoflurane-treated worms that express either high or low polyQ35-YFP levels. (B) 273 proteins exhibited modified levels in the aggregated fraction of worms that express low polyQ35-YFP levels and 147 in animals that express high levels, upon exposure to isoflurane. Only 43 proteins were identified in aggregate fractions of both groups. (C) A list of the 10 most prominent proteins that sediment upon exposure to isoflurane only in worms that express high polyQ35-YFP levels (LFQ ratio > 2 or < 0.5, p-value <0.01). (D) A thrashing assay indicates that the knockdown of tomm-20 by RNAi, and rab-35 albeit, less prominently, abolishes the proteotoxic effect of isoflurane in AM140 (Q35^muscle) worms. No such effect was observed when any of the other tested genes was knocked down. (E) The knockdown of tomm-20 by RNAi prevents isoflurane from enhancing polyQ35-YFP proteotoxicity in neurons as measured in AM1126 (Q35^neuron) worms.

We created two lists of proteins that exhibited significantly modulated levels (ratio of LFQ intensity <3/4 or >4/3, p-value < 0.05), upon exposure to isoflurane. One contained proteins that appeared or disappeared from the pellet of isoflurane-treated animals, compared to the pellets of worms that were not exposed to isoflurane, in worms that expressed low polyQ35-YFP levels (YFP RNAi treatment, 273 proteins). The second list consisted of proteins that sediment or disappeared from the pellet, upon exposure to isoflurane, in worms that expressed high polyQ35-YFP levels (untreated, 147 proteins). Only 43 proteins were found in both lists (Fig. 4B), A Fisher’s Exact test revealed that this overlap is significantly greater than expected by chance (P < 0.0001, two-sided), Nevertheless, the shared proteins constitute only ~16% of the modulated proteins in the low-polyQ35 group and ~29% of those in the high-polyQ35 group, indicating that isoflurane differentially affects the worms’ proteome when the animals are challenged by proteotoxic stress of polyQ35-YFP.

To characterize cellular entities that are most significantly affected by the isoflurane-induced protein sedimentation, we clustered the proteins that were sedimented in the presence and absence of proteotoxic stress (Fig. S4A). Many proteins that reside in the mitochondrial matrix or the mitochondrial ribosome (Fig. S4A, red arrows) were exclusively identified in the pellet of isoflurane-treated worms which expressed high polyQ35-YFP levels. Of note, mitochondrial membrane and envelope proteins were exclusively identified in the pellet of worms which express low polyQ35-YFP levels (Fig. S4A, blue arrows). These results suggest that mitochondria are involved in the proteostasis decline that we observed upon exposure to isoflurane.

To further test which of the sediment proteins are involved in enhancing proteotoxicity upon exposure to isoflurane, we generated a list of the proteins that are most significantly affected by isoflurane in worms that express high polyQ35-YFP levels (Fig. 4C). AM140 (Q35^muscle) worms were treated from hatching until day 3 of adulthood with RNAi toward each of the genes which encode a protein of interest, exposed to isoflurane and thrashing rates were measured 24h thereafter. We discovered that knocking down the expression of tomm-20, a gene which encodes a mitochondrial protein that enables protein import into the mitochondria⁴¹, prevents isoflurane from enhancing polyQ35-YFP-mediated proteotoxicity (Fig. 4D). A similar, but less prominent outcome, was seen when the worms were treated with RNAi toward rab-35, which is expressed in endosomes⁴², but not toward all other tested genes (Fig. 4D). The knockdown of tomm-20 was found to prevent the aggravation of polyQ35-YFP proteotoxicity also in neurons (Fig. 4E).

To further scrutinize the notion that mitochondrial components are involved in proteostasis we asked whether GAS-1, the nematode’s orthologue of the 49-kDa subunit of the mitochondrial NADH: ubiquinone-oxidoreductase (complex I of the respiratory chain), which was reported to be a target of isoflurane⁴³, has a role in the regulation of proteostasis. To address this, AM140 (Q35^muscle) worms were either treated from hatching with gas-1 RNAi or left untreated and subjected to the thrashing assay at day 4 and 6 of adulthood. Our results (Fig. S4B) indicate that the knockdown of gas-1 aggravates the toxicity of polyQ35-YFP in both ages. To examine whether GAS-1 inhibition is inter-connected with isoflurane-induced proteotoxicity, we utilized AM140 worms and treated them from hatching with gas-1 RNAi. At day 3 of adulthood the animals were exposed to isoflurane and their trashing rates were recorded 24 hours thereafter. Exposure to isoflurane does not significantly affect thrashing rates of gas-1 RNAi-treated worms (Fig. S4C), suggesting that isoflurane reduces GAS-1 activity to promote proteotoxicity. Therefore, the knockdown of gas-1 cannot further reduce thrashing rates upon exposure to isoflurane.

Together, these results point at the mitochondria as a key arena of proteostasis modulation by isoflurane.

Isoflurane suppresses proteostasis by modulating mitophagy

The involvement of tomm-20 (Fig. 4, D and E) and gas-1 (Fig. S4C) in isoflurane-induced proteotoxicity, and their possible roles in the induction of mitophagy^44,45 have led us to ask whether mitochondrial quality control mechanisms play active roles in the proteotoxic effect of isoflurane in AM140 (Q35^muscle) animals. Protein degradation by autophagy⁴⁶ in general, and mitochondria recycling by mitophagy in particular⁴⁷, were shown to be crucial proteostasis-promoting activities. Thus, we first asked whether autophagy-related genes exhibit modulated expression levels upon exposure to isoflurane in worms that express high or low polyQ35-YFP levels. Our RNA-seq dataset indicated that four genes which encode for proteins that are known to be involved in autophagy, namely sqst-1, sqst-2, dct-1, and bec-1, exhibit significantly (p<0.05) elevated expression levels in isoflurane-treated animals (Fig. 5A).

Fig. 5.

The roles of autophagy in isoflurane-induced proteotoxicity. (A) An analysis of our RNA-seq data shows that four autophagy-related genes, namely sqst-1, sqst-2, dcr-1 and bec-1, exhibit significantly elevated levels upon exposure to isoflurane regardless if the worms express low or high polyQ35-YFP levels. (B) AM140 worms that were treated with RNAi toward pink-1 or sqst-1 do not exhibit significant changes in thrashing rates upon exposure to isoflurane. (C) The knockdown of sqst-2 and of lgg-1 by RNAi, but neither of dct-1 nor of bec-1 and pdr-1, prevents the aggravation of polyQ35-YFP-mediated proteotoxicity by isoflurane in AM140 worms. (D) Similar results were observed when AM140 worms were crossed with animals that carry mutations in the sequence pink-1, pdr-1 or dct-1. While animals that express polyQ35-YFP in their muscles and carry mutated pdr-1 (strain EHC154), or dct-1 (strain EHC155), exhibited reduced thrashing rates upon exposure to isoflurane, no aggravation of proteotoxicity was observed in worms that express the aggregation-prone chimeric protein as well as mutated pink-1 (strain EHC153). (E) The ratio of nuclear to mitochondrial DNA declines in isoflurane-treated feminized worms that express polyQ35-YFP in their muscles (strain EHC143), indicative of reduced quantities of mitochondria in these animals. This decline is reversed by treatment with VL-004. (F,G) Treatment with VL-004 protects AM140 (F) and AM1126 (G) worms from proteotoxicity enhancement by isoflurane.

We next performed an RNAi-based, directed screen using AM140 (Q35^muscle) worms and the thrashing assay, to examine whether the knockdown of a subset of autophagy-promoting genes enhances isoflurane-induced proteotoxicity. Surprisingly, we found that knocking down the expression of lgg-1, sqst-1, sqst-2 and pink-1 by RNAi, largely reduces the change in thrashing rate upon exposure to isoflurane, as compared to the control group (Fig 5B and C). Yet, no such effect was observed when bec-1, dct-1, or pdr-1 were knocked down by RNAi (Fig. 5C). To further test the possibility that these genes affect isoflurane-mediated proteotoxicity we used mutants. AM140 animals were crossed with either pink-1 mutant (strain tm1779), pdr-1 mutant (strain VC1024) or dct-1 mutant (strain tm376) worms. The resulting strains (EHC153, EHC154 and EHC155, respectively) were exposed to isoflurane at day 3 of adulthood and thrashing rates were recorded. No aggravation of proteotoxicity was seen in EHC153 mutant worms (Fig. 5D). Although mitophagy has been reported to alleviate neurodegeneration-associated phenotypes⁴⁷, it is possible that while certain types of mitophagy protect the worms from the toxicity of polyQ35-YFP upon exposure to isoflurane, other types are not involved in this protection. In fact, these puzzling results point at several possibilities. First, the observation that knocking down pink-1, but not pdr-1 (coding the worm’s orthologue of Parkin), alleviates isoflurane-mediated proteotoxicity, suggests that a mitophagy pathway that is PINK1-dependent but Parkin-independent is involved in isoflurane-induced proteotoxicity. They also suggest that the DCT-1-dependent mitophagy pathway⁴⁸ has no role in the effect of isoflurane on mitochondrial fitness.

To regulate the quality and quantity of mitochondria within the cell, mitophagy is closely intertwined with mitochondrial biogenesis⁴⁹. These mechanisms work in concert to ensure a proper balance between the creation of new, functional mitochondria and the elimination of damaged mitochondria. This dynamic interplay is crucial for preserving mitochondrial fitness and regulating cellular energy production⁴⁹. Our results (Fig. 5B-D) have led us to hypothesize that exposure to isoflurane disrupts the balance between mitochondrial biogenesis and mitophagy in somatic tissues, particularly under chronic proteotoxic challenge. To test this hypothesis, we utilized AM140 (Q35^muscle) worms that were crossed with CF512 animals (strain EHC143), to measure the relative amounts of mitochondria in untreated adult worms and in their counterparts that were exposed to isoflurane. These animals express polyQ35-YFP in their muscles and are feminized upon exposure to 25 °C. Therefore, they allow the comparison of mitochondria amounts in adult tissues with no background that may stem from developing embryos. We also used in this experiment VL-004, a mitophagy and mitochondria biogenesis inducer, which functions in a PINK-1 and DCT-1 dependent, but PDR-1-independent manner⁵⁰, to test whether the enhancement of mitochondrial fitness protects from isoflurane-promoted proteotoxicity. Four AM140 (Q35^muscle) worm populations were grown from hatching on EV bacteria. Two groups were exposed at day 3 of adulthood to isoflurane for 3 hours while the two others were not. One isoflurane treated, and one untread group were transferred onto plates that were supplemented with VL-004. All worm groups were harvested and subjected to qPCR 24 hours thereafter, at day 4 of adulthood, to measure the relative amounts of mitochondrial DNA (measured by nduo-1) to nuclear DNA (measured by ned-8). Animals that were exposed to isoflurane showed a non-significant trend of approximately 30% reduction in the number of mitochondria compared to their untreated counterparts (Fig. 5E). Interestingly, in 3 days old worms, VL-004 reduced the ratio of nuclear to mitochondrial DNA by approximately 27%, however, when the animals were exposed to isoflurane and then to VL-004, the observed ratio was elevated by about 58% (Fig. 5E). These results suggest that when mitochondrial fitness is well preserved, VL-004 enhances mitophagy and reduces the levels of mitochondria within the worm, however, when the mitochondria quantity is reduced due to exposure to isoflurane, it enhances mitochondria biogenesis.

Given that VL-004 appears to enhance mitochondrial biogenesis (Fig. 5E) and improves mitochondrial fitness⁵⁰, it is expected that this compound will alleviate the physiological damage of isoflurane. To test this hypothesis, we exposed AM140 (Q35^muscle) worms to isoflurane, treated them with VL-004, and compared their thrashing rates with those of their untreated counterparts. VL-004-treated animals that were exposed to isoflurane, showed a small but not significant reduction in thrashing rates compared to the control group (Fig. 5F). Treating AM140 (Q35^muscle) animals with VL-004 before and after exposure to isoflurane showed similar protective effect (Fig. S5A). Comparable results were observed when a similar experiment was conducted using AM1126 (Q35^neuron) worms (Fig. 5G), indicating that VL-004 mitigates isoflurane-induced proteotoxicity in muscles and neurons alike.

Finally, we asked whether the documented effect of isoflurane on mitochondrial membrane potential⁵¹, is long lasting. To address this, we employed AM140 (Q35^muscle) worms which were either treated from hatching with yfp RNAi or left untreated. At day 3 of adulthood, half of the worms of each group were exposed for 3 hours to 8% isoflurane, while the other half was left untreated. Twenty-four hours after exposure the animals were soaked for 1 hour in TMRE and visualized using a fluorescent microscope. Our results showed no effect on mitochondrial membrane potential 24 hours after exposure to isoflurane (Fig. S5, B and C), These results suggest that the effect of isoflurane on mitochondrial membrane potential is temporary.

Discussion

What underlies the onset of POD in elder patients, and why POD is more common in individuals with pre-existing cognitive impairments, are largely unanswered queries. Here we employed the nematode C. elegans, to test the theme that isoflurane-mediated anesthesia compromises proteostasis. This model organism offers several advantages for the study of proteostasis in the context of aging. The nematodes are short-lived and amenable for aging manipulation¹⁵. The IIS²⁸ and TGFβ signaling cascade⁵² were shown to serve as critical proteostasis regulators and to control mechanisms that are highly conserved between worms and humans. Thus, despite the evolutionary distance between worms and mammals, key aspects of the cell biology and organismal coordination of proteostasis, can be studied in these animals.

Our findings indicate that exposure to isoflurane does not impair motility of worms that are not challenged by proteotoxicity (Fig 1A). However, animals that experience constant proteotoxic stress, due to the expression of the prone-to-aggregate polyQ35-YFP protein (Fig. 6), exhibit persistent motility impairments upon exposure to isoflurane. This phenotype was observed regardless of whether the protein aggregates were present in muscles or neurons, supporting the notion that different cells types similarly respond to polyQ35-mediated proteotoxic challenges This proteotoxic effect is associated with reduced levels of high molecular weight aggregates (Fig. 2C and D) and probably with an increased rate of small polyQ35-YFP oligomers. Additionally, the aging-regulating transcription factors, DAF-16 and SKN-1, are involved in the toxicity that stems from exposure to isoflurane (Fig. 6I).

Fig. 6.

A model. Exposure to isoflurane modulates the expression of proteostasis-promoting genes, setting the basis for the onset of a persistent proteotoxic insult. These transcriptomic changes are controlled, at least partially, by the transcription factors DAF-16 and SKN-1 (I). In worms that are challenged by chronic polyQ35-YFP aggregation, these transcriptomic modulations result in elevated protein aggregation (II) and in enhanced PINK-1 dependent mitophagy (III). Together these events lead to impaired mitochondrial fitness (IV), and to the aggravation of proteotoxicity (V), which results in motility impairments. Created in BioRender (https://biorender.com). Elami, T. (2026) https://BioRender.com/9go92qw.

In contrast, an exposure to isoflurane only slightly, and not significantly, affected proteotoxicity in worms that express the Alzheimer’s-causing peptide Aβ in muscle (strain CL2006 (Aβ^muscle) or neurons (Strain AGD1246 Aβ^neuron, Fig. S1F and G, respectively). While this observation is counter intuitive, it supports the notion that the proteostasis network differentially responds to distinct proteotoxic challenges. For instance, while the knockdown of the torsin-coding genes, tor-1 and tor-2, mitigates the toxicity of Aβ it aggravates polyQ35-YFP-mediated proteotoxicity³⁵. Similarly, the expression of Aβ and of polyQ35-YFP elicits modulations in the expression levels of distinct chaperone networks that only partially overlap⁵³. These reports and our observations suggest that exposure to isoflurane enhances proteotoxicity that stems from certain proteotoxic proteins, but much less prominently when other proteotoxic proteins are expressed.

Characterizing the changes in the transcriptomic and proteomic landscapes of worms that express high and low polyQ35-YFP levels, we obtained two key insights. First, isoflurane enhances the expression of proteostasis-promoting genes, such as hsp-16.41 and hsp-16.2 which encode for molecular chaperones (Fig. 3F), thereby, it probably sets the basis for the onset of a persistent proteotoxic insult (Fig. 3). Secondly, prominent proteotoxicity-associated changes occur at the proteomic level upon exposure to isoflurane (Fig. 4B and 6II). Our results point at the mitochondria as the cellular organelle that exhibits the most prominent proteomic modulations in isoflurane-treated worms that express high polyQ35-YFP levels (Fig. S4A). Specifically, the mitochondrial protein TOMM-20, as well as proteins which participate in mitophagy, were found to play key roles in the isoflurane-induced proteotoxic effect. These results propose that while mitophagy is critical for proteostasis maintenance under certain circumstances, an over activation of this mechanism (Fig. 6III) can be detrimental when worms that are challenged by a proteotoxic insult, exposed to isoflurane. They also suggest that under these conditions, the biogenesis of new mitochondria is incapable of replacing the damaged ones. Accordingly, mitochondrial fitness is impaired (Fig. 6IV) and proteotoxicity is enhanced (Fig. 6V).

These observations raise the intriguing question of what proteostasis-related roles are played by autophagy and mitophagy, in isoflurane-treated worms. On one hand, it is well-documented that autophagy is a key protein quality control mechanism that promotes proteostasis and alleviates proteotoxicity⁴⁶. On the other hand, we found that the knockdown of pink-1, a gene that codes for a kinase that is known to be involved in the promotion of mitophagy⁵⁴, and whose knockdown is associated with POD⁵⁵, prevented isoflurane-induced motility impairment (Fig. 5B and D). Similarly, knocking down the expression of the autophagy-related genes; sqst-1, sqst-2 and lgg-1 protected the nematodes from isoflurane-induced polyQ35-YFP proteotoxicity (Fig. 5B and C). The idea that under pathophysiological conditions, autophagy can be a double-edged sword⁵⁶ is supported by our results which suggest that autophagy in general, and mitophagy in particular, may be deleterious when the organism is exposed to isoflurane. It is plausible that upon exposure to isoflurane, mitophagy declines proteostasis by impairing the delicate balance between the removal and biogenesis of mitochondria. The interplay between these processes is pivotal for mitochondrial fitness. Thus, under these specific circumstances, the inhibition of mitophagy slows the pace of mitochondria removal and assists cells restoring their reservoir of mitochondria. This notion is supported by the observation that the compound VL-004, which prevents the decline in the quantity of mitochondria upon exposure to isoflurane (Fig. 5E), alleviates the proteotoxic effect of this anesthetic agent (Fig. 5F, G and S5A).

Exposure to isoflurane did not affect mitochondrial membrane potential, 24 h after exposure, regardless of whether the worms were challenged by high or low levels of polyQ35-YFP proteotoxicity (Fig. S5B and C). This observation suggests that the documentaed effect of isoflurane on mitochondrial membrane potential⁵¹ is temporary. Therefore, the long-lasting effect of isoflurane on mitochondrial fitness may be associated with reduction in ATP production, or with modulated activation of the mitochondrial unfolded protein response, which was shown to exhibit elevated activity in worms that express polyQ stretches²³.

Our results also show that knocking down gas-1 reduced thrashing rates regardless of whether the worms were exposed to isoflurane or not (Fig. S4C). This observation together with the well-documented involvement of GAS-1 in sensitivity to general anesthesia⁴³, and with the possible roles of this protein as a regulator of mitophagy⁴⁵, suggest that inhibition of GAS-1 by isoflurane can be the first step in the excessive mitophagy induction that may lead to proteotoxicity.

An additional interesting aspect of the interplay between mitochondria removal and biogenesis is the possible role of SKN-1 in the regulation of this balance. Knocking down the expression of this proteotoxicity-mitigating transcription factor³¹ appears to be protective, as it maintains the thrashing rate of isoflurane-exposed worms to a similar level to this of control, untreated animals (Fig. 2E). SKN-1 has been shown to coordinate induction of both mitochondrial biogenesis and mitophagy⁴⁹. Therefore, it is possible that exposing worms that are challenged by chronic proteotoxic stress to isoflurane, dysregulates these opposing processes. This dysregulation may leads to high activation of mitophagy by SKN-1 and jeoperdizes mitochondrial fitness. Accordingly, the knockdown of skn-1 reduces this detrimental effect. The observation that skn-1 can play detrimental roles in the face of proteotoxicity is consistent with our previous results³⁵.

Knocking down another known proteotoxicity-mitigating transcription factor, DAF-16, resulted in a reduced thrashing rate regardless if the worms were exposed to isoflurane or not (Fig. 2E). This phenomenon can be explained by the key importance of this transcription factor for proteostasis. Accordingly, reduced DAF-16 level efficiently impairs proteostasis and thus, exposure to isoflurane cannot further reduce thrashing. Further research is needed to elucidate the roles of SKN-1 and DAF-16 as proteotoxicity modulators upon exposure to isoflurane.

The novel insights that were obtained in this work may have clinical relevance as they propose that the induction of mitochondrial biogenesis using compounds such as VL-004, prior or after anesthesia, could reduce the risk of POD.

Materials and methods

Caenorhabditis elegans maintenance

All C. elegans strains were grown at 20 °C unless stated otherwise (a complete strain list appears as Supplementary Table S3). CF512 and EHC143 strains were maintained at 15 °C. For experiments, animals of these two strains were let hatch and developed at 25 °C (to feminize the worms) and transferred into a 20 °C incubator for the rest of the experiment. All worms were grown on Nematode Growth Medium (NGM) plates that were supplemented with 100 µg/ml ampicillin and seeded with Escherichia coli HT115 bacteria. Worm populations were synchronized using sodium hypochlorite (bleach) and potassium hydroxide following standard methods⁵⁷.

Gene knockdown by RNA interference (RNAi)

All RNAi experiments were carried out on NGM-ampicillin plates seeded with Escherichia coli bacteria that harbor the appropriate RNAi clone, as previously described⁵⁸. The bacteria were grown overnight at 37 °C in LB medium and supplemented with 100 mM isopropyl b-d-1-thiogalactopyranoside (IPTG; final concentration of 4mM) to induce the expression of the dsRNA. RNAi constructs targeting daf-2 (pAD48), daf-16 (pAD43), and the empty vector control (pAD12) were kindly provided by Prof. Andrew Dillin. The hsf-1 RNAi clone was obtained from the Ahringer library⁵⁹. The RNAi against yfp, pink-1, sqst-1 and gas-1 were created by the Cohen lab by amplifying the relevant sequences (see Supplementary Table S4) using PCR and cloning the amplicons into the L4440 plasmid using the indicated restriction enzymes. All other RNAi bacterial strains were obtained from the Vidal ORFeome library⁶⁰.

Exposure to isoflurane

Worms were placed on foodless NGM plates and exposed to 8% isoflurane for 3 hours using a custom-made glass chamber at 20°C. Isoflurane concentration was monitored by gas chromatography. The control group remained in the same conditions but was not exposed to the anesthetic agent.

Thrashing assay

Thrashing assays were conducted as previously described (17). Briefly, individual AM140, N2, or AM1126 worms were subjected to the experiment at the indicated age. Each animal was placed in 10 µl M9 buffer on a glass slide, allowed to acclimate for 30 seconds, and then body bends in one lateral direction were counted for 30 s. 16–20 worms were used per treatment per an independent experiment and two or three independent experimental repeats were performed.

Fluorescent microscopy to visualize foci

For polyQ35-YFP foci measurement, worms were collected at the indicated age and immobilized in 20 mM sodium azide on a glass slide containing a 2% agarose pad in the center. The worms were visualized immediately using an AZ100 Nikon microscope. Foci were counted manually in 12–25 worms per treatment across three independent experiments.

Filter trap assay

Filter trap assays were conducted according to the method of Wanker et al.⁶¹. AM140 worms were grown on HT115 bacteria harboring the empty RNAi vector (EV). At day 5 of adulthood the worms were exposed to isoflurane, washed with M9 buffer, collected, and snap-frozen in liquid nitrogen. The worm pellets were thawed on ice and resuspanded into a lysis buffer (50 mM Hepes pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X‐100) supplemented with an EDTA‐free protease inhibitor cocktail (Roche), and disrupted using a bullet blender 24 device (Next Advance, NY, USA). Debris was sediment by low‐speed centrifugation (8000 × g spin for 5 min, 4 °C). 100 μg of each protein extract was supplemented with SDS to the final concentration of 0.5% and loaded onto a cellulose acetate membrane (Cytiva #10404180) assembled in a slot blot apparatus (Bio‐Rad cat#1703938). The membrane was washed with 0.2% SDS, and the retained polyQ35‐YFP was assessed by immunoblotting using a GFP antibody (Cell Signaling, mAb #2956). Band intensities were compared using ImageJ software (https://imagej.net/ij/).

Paralysis assay

Paralysis was assessed in CL2006 and AGD1246 worms as described by Boocholez et al.,³⁵. Synchronized populations were grown to day 1 of adulthood, and 120 animals were transferred onto 6-cm NGM plates seeded with the corresponding E. coli strain (12 worms per plate). Worms were scored daily for paralysis and transferred to fresh plates throughout the reproductive period (days 1–6), and subsequently every other day until day 12 (CL2006) or day 13 (AGD1246), when age-dependent paralysis typically occurs. A worm was classified as paralyzed when exhibiting the “windshield wiper” phenotype, defined as head movement without the ability to crawl. Isoflurane exposure was performed on days 5 and 9 of adulthood.

Measuring mitochondrial membrane potential using TMRE

AM140 worms were collected at day 4 of adulthood (24 h post-isoflurane exposure). Approximately 200 animals were incubated for 60 min at room temperature in 200 μL M9 containing 4μM TMRE, 0.1% Poloxamer 188, and 0.1% Pluronic F-127, with gentle rotation (~9 rpm) and protection from light. Following staining, worms were washed three times with M9 and transferred onto NGM plates seeded with the corresponding RNAi treatment (EV or yfp RNAi) for 1 h. Animals were then immobilized on 2% agarose pads using 20 mM sodium azide and imaged immediately using a fluorescence microscope (Zeiss Z1). Images of TMRE and YFP were acquired using red and green channels respectively at identical imaging parameters. Fluorescence intensities were quantified in ImageJ using Integrated Density (IntDen), and corrected values were calculated as: Corrected Integrated Density (CID) = IntDen − (Area × Mean Background Intensity).

RNA isolation, next generation sequencing

The isolation of total RNA and next generation sequencing were conducted as described by Zhu et al.⁶². In brief, total RNA was extracted using a NucleoSpin® RNA kit (Macherey‐Nagel, #740955) according to the manufacturer instructions. For each group, 10,000 synchronized eggs of AM140 worms were placed on NG‐ampicillin plates seeded with HT115 bacteria that harbored either the control RNAi plasmid (EV) or yfp RNAi. The worms were washed daily with M9 to get rid of progeny until harvesting at day 5 of adulthood. One group of worms that were grown on EV bacteria and one that was treated with yfp RNAi were exposed to isoflurane for 3 h and two identical groups were not. Three hours after exposure to isoflurane all four worm groups were collected in M9 and flash-frozen in liquid N₂. The worms were then thawed and subjected to mechanical disruption using a 2 mL glass Dounce homogenizer (Kimble, cat # 885301‐0002). Homogenates were transferred into microcentrifuge tubes and centrifuged at 14,000 × g for 5 min. The supernatants were transferred to NucleoSpin® Filter (NucleoSpin® RNA kit; Macherey‐Nagel), and total RNA was purified according to the manufacturer’s instructions. The RNA was quantified using a NanoDrop 2000c spectrophotometer. For RNA-sequencing, we used RNA ScreenTape kit (catalog #5067‐5576; Agilent Technologies), D1000 ScreenTape kit (catalog #5067‐5582; Agilent Technologies), Qubit® RNA HS Assay kit (catalog # Q32852; Invitrogen), and Qubit® DNA HS Assay kit (catalog #32854; Invitrogen). mRNA libraries were prepared using KAPA Stranded mRNA kit with mRNA Capture Beads (KAPA Biosystems, KK8421). In brief, 1 μg was used for the library construction; library was eluted in 20 μL of elution buffer. All DNA sample libraries were pooled into 10 nM samples. Multiplex sample pools were loaded on NovaSeq 6000 (Illumina) using NovaSeq 6000 SP Reagent Kit v1.5, 100 cycles (cat# 20028401), with 122 cycles of single‐end sequencing (raw data are available at: GSE272857).

Computational analyses of next generation sequencing data

Raw reads were processed for quality trimming and adaptors removal using fastx_toolkit v0.0.14 and cutadapt v2.10 (Marcel M. et al., EMBnet.journal 2011, 17.1:10-12, https://cutadapt.readthedocs.io/en/v2.10/installation.html). The processed reads were aligned to the Caenorhabditis elegans transcriptome and genome version WBcel235 with annotations from Ensembl release 106 using TopHat v2.1.1 (Kim D et al., Genome Biology 2013, 14: R36, http://ccb.jhu.edu/software/tophat). Counts per gene quantification was done with htseq-count v2.01 (Anders S et al., Bioinformatics 2015, 31 (2):166-169, https://htseq.readthedocs.io/en/master/history.html). Normalization and differential expression analysis were done with the DESeq2 package v 1.36.0 (Love MI et al., Genome Biology 2014, 15:550. https://bioconductor.org/packages/release/bioc/html/DESeq2.html). Normalized counts were further batch-corrected for the effect of the 3 repeats of the biological procedure. Pair-wise comparisons were tested with default parameters (Wald test), without applying the independent filtering algorithm. The significance threshold was taken as padj<0.1. Gene set enrichment analysis was performed using WormCat (http://wormcat.com/) and clusterProfiler R https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html) packages. For enrichment analysis using WormCat, differentially expressed genes were analyzed using ‘worm_cat_fun’ function. For enrichment analysis using clusterProfiiler, differentially expressed genes were enriched against Gene Ontology molecular function using enrichGO’, (https://www.rdocumentation.org/packages/clusterProfiler/versions/3.0.4/topics/enrichGO) function with ‘MF’ argument.

Comparing the amounts of mitochondria by qPCR

Four groups of EHC143 worms were used in this experiment. At day 3 of adulthood, two groups were exposed for 3 hours to 8% isoflurane, and two identical groups were not exposed to the anesthetic agent. Immediately after exposure, the worms were transferred to new plates which were either supplemented with VL-004 or not. Twenty-four hours thereafter 10 individual worms were transferred into an Eppendorf tube containing 50 µl of worm lysis buffer (50 mM KCL, 10 mM Tris-HCl pH8.3, 2.5 mM MgCl₂, 0.45% v/v Nonidet P-40, 0.45% Tween-20, 0.01% w/v gelatin, 0.1 mg/ml proteinase K) and immediately froze them in -20 °C. Upon thawing, samples were vortexed, and incubated in 60 °C for 90 min, followed by 95 °C for 30 min, then kept at 4 °C. qPCR: Crude worm lysate was diluted 1:2 and 4 µl was used for each reaction. Each reaction contained 0.375 µM of Forward and Reverse primers (see primer sequences at Supplementary Table S5), 4 µl of crude worm extract diluted 1:2, 6 µl of SYBR mix (#172-5124, Bio Rad) in a total volume of 12 µl. ubq-2 was used to normalize the expression data. Each reaction was performed in triplicates. Relative quantities of the mitochondrial gene nduo-1 and nuclear gene ned-8 were compared in each sample and the ratio of the relative level of mtDNA per nuclear genome was calculated. At least three independent repeats were performed.

Mass-spectroscopy

Sample preparation for protein identification by mass-spectroscopy

For each group, 10,000 synchronized eggs of AM140 worms were placed on NG‐ampicillin plates seeded with HT115 bacteria that harbored either the control RNAi plasmid (EV) or yfp RNAi. The worms were washed daily with M9 to get rid of progeny until harvesting at day 5 of adulthood. One group of worms that were grown on EV bacteria and one that was treated with yfp RNAi were exposed to isoflurane for 3 h and two identical groups were not. Three hours after exposure to isoflurane all four worm groups were harvested. Upon harvesting the worms were washed twice with M9 buffer (RT) and resuspended in 300 µl PBS, supplemented with a protease inhibitor cocktail (Millipore, Billerica MA #539134). Then the nematodes were homogenized using a Dounce homogenizer and cleared by low-speed centrifugation (5 min, 1000g). Supernatants were transferred onto new tubes, supplemented with TX-100 and deoxycholic acid (final concentration of 1% of each detergent) followed by incubation on ice for 30 minutes. The samples were transferred twice through thin needles (0.4mm) and spun at 10,000g (30 min, 4 ºC). Supernatants were transferred onto new tubes, supplemented with 1% Sodium lauroyl sarcosinate (Sarkosyl), and incubated on ice for 30 min followed by ultra-centrifugation (1 h, 200,000g, 4 ºC). Supernatants (soluble fractions) were transferred onto new tubes and pellets (aggregate proteins) were frozen at –80C and sent to the Stein Family mass spectrometry center of the Silberman Institute of Life Sciences, Hebrew University of Jerusalem.

Sample preparation for LC/MS

Samples were prepared for LC/MS analysis as described previously by Siddiqui et al.⁶³. Protein samples were digested and cleaned using S-Trap microcolumns (Protifi, LLC, Huntington, NY). Proteins were initially denatured by incubation in 8M Urea with 10mM DTT in 25mM TRIS-HCl buffer (pH 8) at room temperature for 30 minutes. Samples were alkylated in 55 mM iodoacetamide and incubated for 30 min at room temperature in the dark. Phosphoric acid was added to a final concentration of 1.2%. Methanol-TRIS buffer (90% MeOH, 10% TRIS 0.5M pH7.1) was added to the samples at a ratio of 6:1 (buffer:sample) and loaded onto S-Trap columns by centrifugation at 1,000g for 1 minute. Columns were subsequently washed twice with 150 µl Methanol-TRIS buffer at 4,000g (second wash). Sequencing grade modified trypsin (Promega Corp., Madison, WS) was loaded onto the column (1 µg per column) and incubated at 47 °C for 90 min. Peptides were eluted from column, acidified and desalted on homemade C18 stage tips (Rappsilber J, Mann M, Ishihama Y. Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips. Nat Protoc. 2007;2(8):1896-906.). From the resulting peptides 0.35 µg (determined by Absorbance at 280 nm) from each sample were injected into the mass spectrometer for analysis. Dataset is available at: https://www.ebi.ac.uk/pride/archive/projects/PXD054547

LC-MS/MS analysis

LC/MS analysis and data analysis were conducted as described previously by Siddiqui et al.⁶³. Briefly, MS analysis was performed using a Q Exactive-HF mass spectrometer (Thermo Fisher Scientific, Waltham, MA USA) coupled on-line to an Ultimate 3000 Dionex (Thermo Fisher Scientific, Waltham, MA USA) UHPLC. Peptides were separated on a 120 min acetonitrile gradient run at a flow rate of 0.15 μl/min on a reverse phase 25-cm-long C18 column (75 μm ID, 2 μm, 100Å, Thermo PepMapRSLC). Survey scans (300–1650 m/z, target value 3E6 charges, maximum ion injection time 20 ms) were acquired and followed by higher energy collisional dissociation (HCD) based fragmentation (normalized collision energy 27). A resolution of 60,000 was used for survey scans and up to 15 dynamically chosen most abundant precursor ions, with “peptide preferable” profile was fragmented (isolation window 1.6 m/z). The MS/MS scans were acquired at a resolution of 15,000 (target value 1E5 charges, maximum ion injection times 25 ms). Dynamic exclusion was set to 20 s. Data were collected with Xcalibur software. (Thermo Scientific, https://www.thermofisher.com/il/en/home/industrial/mass-spectrometry/liquid-chromatography-mass-spectrometry-lc-ms/lc-ms-software/lc-ms-data-acquisition-software/xcalibur-data-acquisition-interpretation-software.html). To avoid a carryover, the column was washed with 80% acetonitrile, 0.1% formic acid for 25 min between samples.

MS data analysis

Mass spectra data were processed using the MaxQuant computational platform, version 2.0.3.0 (https://maxquant.org/). Peak lists were searched against C. elegans proteome (UP00000194) obtained from Uniprot on 12.01.2023. The search included cysteine carbamidomethylation as a fixed modification, N-terminal acetylation and oxidation of methionine as variable modifications and allowed up to two miscleavages. The ‘match-between-runs’ option was used. Peptides with a length of at least seven amino-acids were considered and the required FDR control was applied at 1% for both peptide and protein identifications, as per MaxQuant defaults. Relative protein quantification in MaxQuant was performed using the label-free quantification (LFQ) algorithm (Cox, J. et al. MaxLFQ allows accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction (Mol. Cell. Proteomics 13, 2513–2526 (2014)).

Statistical analysis was performed using the Perseus software (Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, Mann M, Cox J. Nat Methods. 2016 Sep;13(9):731-40. 10.1038/nmeth.3901. Epub 2016 Jun 27, PMID: 27348712., maxquant.net/perseus/) The Perseus computational platform for comprehensive analysis of (prote)omics data.). Analysis was performed using LFQ values of proteins after filtration of those annotated as “reverse”, “potential contaminant” and “only identified by site”. In addition, any protein that was not identified in all three samples in at least one of the two sample groups was excluded from the analysis. The remaining proteins which still had zero LFQ intensity were replaced with values from a normal distribution using the default Perseus parameters.

The different lists of proteins were created by using LFQ and P-values as mentioned.

clusterProfiiler R package was utilized to cluster the identified proteins according to cellular components using ‘enrichGO’ function with ‘CC’ argument.

Data are available via ProteomeXchange with identifier PXD054547. Reviewer account details: Username: reviewer_pxd054547@ebi.ac.uk Password: DC6V5rQY1wWB

Statistical analysis

For thrashing assay and fluorescent microscopy (to visualize foci), the results are presented as the mean with 95% confidence intervals. Statistical significance was calculated using multiple unpaired t-test (*p < 0.05, **p < 0.01, ***p < 0.001, **** p < 0.0001).

For qPCR experiments and filter trap assays, statistical significance was calculated using the unpaired t-test function (*p < 0.05, **p < 0.01, ***p < 0.001, **** p < 0.0001).

The statistical analyses and plotting of the data were performed using GraphPad Prism 10 (GraphPad Prism Software, Inc., La Jolla, CA, USA, https://www.graphpad.com/features).

Author contributions

EC and TE designed, initiated this study and wrote the manuscript. TE performed most of the experimental work including thrashing assays, preparation of samples for RNA-sequencing, mass spectrometry and computational analysis, as well as cloning procedures. VS and EG characterized and assisted with experiments that involve VL-004, HZ crossed worm strains, IC performed qPCR analyses and AZ conducted filter-trap assays and analyzed computational data. RBH assisted with cloning and crossing procedures.

Data availability

RNA sequencing data (presented at Fig. 3) are freely available at the Gene Expression Omnibus (GEO) dataset (from the NCBI), [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi] (https:/www.ncbi.nlm.nih.gov/geo/query/acc.cgi) Access code: GSE272857. MS data shown as Fig. 4, are available via ProteomeXchange with identifier PXD054547.

Declarations

Competing interests

Einav Gross is an inventor of VL-004 and has a share in a patent application (No. PCT/IL2019/050278) that covers the rights of using this compound. The other authors declare no competing interests.