Activated SKN-1 alters the aging trajectories of long-lived Caenorhabditis elegans mutants

Abstract

In the presence of stressful environments, the SKN-1 cytoprotective transcription factor is activated to induce the expression of gene targets that can restore homeostasis. However, chronic activation of SKN-1 results in diminished health and a reduction of lifespan. Here, we demonstrate the necessity of modulating SKN-1 activity to maintain the longevity-promoting effects associated with genetic mutations that impair daf-2/insulin receptor signaling, the eat-2 model of dietary restriction, and glp-1-dependent loss of germ cell proliferation. A hallmark of animals with constitutive SKN-1 activation is the age-dependent loss of somatic lipids, and this phenotype is linked to a general reduction in survival in animals harboring the skn-1gf allele. Surprisingly, daf-2lf; skn-1gf double mutant animals do not redistribute somatic lipids, which suggests the insulin signaling pathway functions downstream of SKN-1 in the maintenance of lipid distribution. As expected, the eat-2lf allele, which independently activates SKN-1, continues to display somatic lipid depletion in older ages with and without the skn-1gf activating mutation. In contrast, the presence of the skn-1gf allele does not lead to somatic lipid redistribution in glp-1lf animals that lack a proliferating germline. Taken together, these studies support a genetic model where SKN-1 activity is an important regulator of lipid mobilization in response to nutrient availability that fuels the developing germline by engaging the daf-2/insulin receptor pathway.

Introduction

Since the discovery of insulin signaling mutants that nearly double the normal lifespan of worms, Caenorhabditis elegans has become an attractive model to understand longevity and aging (Friedman and Johnson 1988, Dorman et al. 1995, Tissenbaum and Ruvkun 1998, Berman and Kenyon 2006, Lee et al. 2006). The longevity of C. elegans can be manipulated via several pathways, including germline ablation, dietary restriction, and deficiencies in insulin growth factor signaling (Lakowski and Hekimi 1998, Arantes-Oliveira et al. 2002). Laser ablation of germline progenitor cells disrupts germline to soma signaling and promotes an increase in lifespan (Hsin and Kenyon 1999). Null mutants for the C. elegans NOTCH ortholog glp-1, when raised at a restrictive temperature, will increase longevity via germ to soma signaling through pathways that modulate DAF-16 activity (Mukhopadhyay and Tissenbaum 2007). Broadly, dietary restriction is a driver of increased longevity (Lakowski and Hekimi 1998, Kaeberlein et al. 2006, Greer and Brunet 2009). In C. elegans, the nicotinic acetyl-choline receptor EAT-2 regulates pharyngeal muscle contractions regulating feeding rate (Lakowski and Hekimi 1998). Reduced feeding rate via eat-2lf mutation results in dietary restriction, several factors such as pha-4 and hlh-30 feed into eat-2lf mediated dietary restriction generally resulting in increased autophagy and proteasomal turnover, which promotes longevity (Panowski et al. 2007, O'Rourke and Ruvkun 2013). Lastly, under nominal conditions, insulin-IGF1-like signaling opposes SKN-1 activation by signaling through to kinases such as SGK-1 which in turn phosphorylate SKN-1 and inhibit its nuclear accumulation (Turner et al. 2024). A reduction in insulin-IGF1-like signaling (IIS) causes an increase in SKN-1 activity that promotes stress resistance and longevity (Tullet et al. 2008).

SKN-1 is a multifaceted transcription factor that regulates the response to xenotoxins, unfolded proteins, variable proteotoxic insults, and metabolic stress by upregulating appropriate stress response pathways to mitigate the stressful condition (Blackwell et al. 2015, Boocholez et al. 2022, Turner et al. 2024). Loss of SKN-1 function results in a diminished stress response and a reduction in lifespan (Blackwell et al. 2015, Turner et al. 2024). Although mild overexpression of SKN-1, albeit without additional genetic or environmental activation, was shown to increase lifespan (Blackwell et al. 2015), investigations utilizing a genetically encoded constitutively activated skn-1 mutation determined that longevity is severely diminished when SKN-1 activity cannot be turned off (Paek et al. 2012, Turner et al. 2024). Constitutively, active skn-1 mutants display a remarkable resistance to acute exposures to stress (Paek et al. 2012), but at the cost of dysregulated lipid metabolism and mobilization (Lynn et al. 2015) that ultimately reduces overall lifespan (Paek et al. 2012, Lynn et al. 2015, Turner et al. 2024). This phenomenon is also observed in wild-type (WT) animals exposed to pathogens (Nhan et al. 2019), but importantly represents a tradeoff between organismal stress responses and cytoprotection that influences survival that is mediated through lipid storage and distribution. Importantly, SKN-1 regulates cellular detoxification (Turner et al. 2024), proteostasis (Turner et al. 2024), and metabolism (Turner et al. 2024) responses and receives input from multiple regulatory pathways, including nutrient sensation (Turner et al. 2024), germline integrity (Steinbaugh et al. 2015), and insulin signaling (Tullet et al. 2008), which when impaired can result in an extension of lifespan (Blackwell et al. 2015, Turner et al. 2024).

Given the opposing longevity outcomes of SKN-1 activation in response to loss of daf-2 signaling, germline proliferation, and pharyngeal pumping vs constitutive activation of SKN-1 from gain-of-function (gf) alleles, we sought to assess how constitutive SKN-1 activity would impact well-established longevity-promoting mutations.

Materials and methods

C. elegans strains and maintenance

C. elegans were raised on 6 cm nematode growth media (NGM) plates supplemented with streptomycin and seeded with OP50. All worm strains were grown at 20°C except for temperature sensitive strains which were grown at 15°C, and worm strains were unstarved for at least 3 generations before being used (Stiernagle 2006).

Strains used in this study are as follows: WT, N2 Bristol; SPC227, skn-1gf(lax188); DA465, eat-2lf(ad465); CB1370, daf-2lf(e1370); CB4037, glp-1lf(e2141); and GR1307, daf-16(mgDf50).

Double mutants used in this study were obtained by standard genetic techniques. Some strains used in this study were provided by the Caenorhabditis Genetics Center, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD0104400).

Lifespan analysis

Synchronized L4 animals were moved onto NGM plates seeded with OP50 without the supplementation of FuDR. All worms were kept at 20°C and were transferred each day during the reproductive period. Worms that died of vulval bursting, bagging, or crawling off the plate were censored. Lifespan analysis and graphing were performed using GraphPad Prism 10. The lifespans for Fig. 1a–c were performed simultaneously and have the same WT and skn-1gf controls.

Dauer entry and exit assay

Gravid adult worms were bleach spotted and progeny were raised at 25°C for 30 h to allow for the development of dauer larvae. Dauers were counted and proportions calculated as percent of total population. For the exit assay, dauers were shifted to 15°C and plates were scored for the presence of non-dauers at 24, 48, and 72 h.

RNA-seq analysis

Analysis was performed as in Turner et al. (2023). In brief, gravid adult worms were egg prepped and eggs were allowed to hatch overnight for a synchronous L1 population. The next day, L1s were dropped onto seeded NGM plates and allowed to grow 48 h, 72 h, 120 h, or 168 h (L4, day 1 adult, day 3 adult, day 5 adult, respectively) before collection. Animals were washed 3 times with M9 buffer and frozen in TRI reagent at −80°C until use. Animals were homogenized, and RNA extraction was performed via the Zymo Direct-zol RNA Miniprep kit (cat. #R2052). Qubit™ RNA BR Assay Kit was used to determine RNA concentration. The RNA samples were sequenced, and read counts were reported by Novogene. Read counts were then used for differential expression (DE) analysis using the R package DESeq2 created using R version 3.5.2. Statistically significant genes were chosen based on the adjust P-values that were calculated with the DESeq2 package. Gene ontology (GO) analysis was performed using WormCat 2.0 (Holdorf et al. 2020). The transcriptomic analysis revealed lipid metabolism gene expression changes which were assessed functionally by lipid staining analysis.

Oil red O staining

Analysis was performed as in Stuhr et al. (2022). In brief, gravid adult worms were egg prepped and allowed to hatch overnight for a synchronous L1 population. The next day, worms were dropped onto plates seeded with bacteria and raised to 120 h (day 3 Adult stage). Worms were washed off plates with PBS + triton and then rocked for 3 min in 40% isopropyl alcohol before being pelleted and treated with oil red O (ORO) in diH2O for 2 h. Worms were pelleted after 2 h and washed in PBS + triton for 30 min before being imaged at 20 × magnification with LAS X software and Leica Thunder Imager flexacam C3 color camera.

Age-dependent somatic depletion of fat quantification

Analysis was performed as in Turner et al. (2023). In brief, ORO-stained worms were placed on glass slides and a coverslip was placed over the sample. Worms were scored and images were taken with LAS X software and Leica Thunder Imager flexacam C3 color camera. Fat levels of worms were placed into 2 categories: non-age-dependent somatic depletion of fat (Asdf) and Asdf. Non-Asdf worms display no loss of fat and are stained dark red throughout most of the body (somatic and germ cells). Asdf worms had most, if not all, observable somatic fat deposits depleted (germ cells only) or significant fat loss from the somatic tissues with portions of the intestine being clear (somatic < germ).

Statistical analysis

Statistical analysis was performed using the Prism 10 software. For lifespan curve comparisons, the log-rank (Mantel–Cox) test was performed. To correct for multiple comparisons, we used the Bonferroni correction method. Transcriptomic analysis was performed on 4 different genetic backgrounds each having 3 biological replicates. Statistics for the DE analysis was performed using the DEseq2 algorithm.

Results

Constitutive activation of SKN-1 reduces the lifespan of canonically long-lived mutants

Animals harboring the gf skn-1 allele, skn-1gf(lax188) (Paek et al. 2012), display constitutive activation of the SKN-1 transcription factor that results in accelerated aging that manifests in a significant reduction of mean, median, and maximal lifespan (Lynn et al. 2015, Nhan et al. 2019,Turner et al. 2024). To assess the impact that constitutive SKN-1 activation has on the longevity of canonically long-lived C. elegans mutants, we generated double mutants of the skn-1gf allele and 3 previously described long-lived mutants, daf-2lf (Dorman et al. 1995, Kimura et al. 1997, Tissenbaum and Ruvkun 1998), eat-2lf (Lakowski and Hekimi 1998, McKay et al. 2004, Kaeberlein et al. 2006), or glp-1lf (Kodoyianni et al. 1992, Hsin and Kenyon 1999, Berman and Kenyon 2006), each of these longevity mutants being missense reduction of function mutants (Kodoyianni et al. 1992, Kimura et al. 1997, McKay et al. 2004).

We found that constitutive activation of SKN-1 shortened the lifespan of each of the long-lived mutants, but to varying magnitude (Fig. 1a–c; Supplementary Table 1). Specifically, the eat-2lf;skn-1gf mutant displays a median lifespan of 9 days, most similar to the skn-1gf mutant alone, and down from an extended 25-day median lifespan in the eat-2lf single mutants. Similarly, glp-1lf;skn-1gf double mutant animals display a median lifespan of 12 days that is down from 27 days for the glp-1lf single mutant. Finally, although the daf-2lf;skn-1gf double mutant animals display a significant reduction in median lifespan as compared to the daf-2lf mutants alone, the presence of the daf-2lf mutation did significantly increase lifespan in the context of the skn-1gf mutation; daf-2lf;skn-1gf double mutant animals are longer lived than WT animals (Supplementary Table 1).

Impaired IIS in daf-2lf mutants drives the alternative developmental program that leads to the stress resistant and long-lived dauer diapause state (Fielenbach and Antebi 2008). As such, we examined the influence of constitutive SKN-1 activation on dauer development by assessing the ability of daf-2lf;skn-1gf double mutant animals to enter the dauer diapause state at 25°C and found no remarkable differences, as compared to the daf-2lf single mutant (Supplementary Fig. 1a). Similarly, the daf-2lf;skn-1gf mutants exited dauer diapause and continued development to adulthood when returned to the permissive temperature (Supplementary Fig. 1b). Taken together, these data suggest that the impact of constitutive SKN-1 transcriptional activity is specific to post-dauer decision activities mediated by the insulin signaling pathway.

Impaired insulin signaling results in reduced lipid metabolism transcripts

Similar to the canonical negative regulation of the FoxO transcription factor DAF-16 (Lee et al. 2001), an established role of the IIS pathway is to restrict the activity of SKN-1¹⁵. However, in light of our finding that the daf-2lf mutation was capable of significantly increasing the lifespan of the constitutively activated skn-1gf mutant, we next examined the transcriptional changes that occur when animals harbor either skn-1gf, daf-2lf, or both alleles. Perhaps unsurprisingly, the transcriptomes to daf-2lf and skn-1gf single mutants were significantly different from each other and WT animals (Fig. 2a; Supplementary Fig. 2a and b and Table 2), but strikingly, the daf-2lf; skn-1gf double mutant was remarkably similar to the daf-2lf mutant alone (Fig. 2a).

With the goal of trying to elucidate the molecular basis underlying the enhanced longevity effect of the daf-2lf allele on animals with constitutive SKN-1 activation, we focused our examination on the transcriptional differences between skn-1gf and daf-2lf;skn-1gf mutant animals that display a clear separation within the principal component space since the transcriptional signatures of the daf-2lf and daf-2lf;skn-1gf mutants animals were similar. We found 3487 genes were differentially regulated by at least 1.5-fold and adjusted P-value less than 0.05 (Supplementary Fig. 2b and Table 2). An analysis of GO terms associated with the differentially effected genes revealed enrichment for signaling, specifically lipid signaling, and metabolism classes of genes (Fig. 2b and c; Supplementary Table 2). In addition, the majority (83%) of the top 100 differentially regulated genes in the daf-2lf;skn-1gf mutants animals are previously confirmed targets of DAF-16 (Tepper et al. 2013) (Supplementary Table 3). To determine if the skn-1gf allele may be affecting DAF-16 activation, we introduced a daf-16p::daf-16::GFP reporter strain with skn-1gf, and we determined that constitutive SKN-1 activation does not induce nuclear accumulation of DAF-16 (Supplementary Fig. 2d). We also generated a daf-16lf;skn-1gf double mutant as IIS signaling converges on DAF-16 as the functional transcriptional effector and found that constitutive SKN-1 activation further reduced lifespan in daf-16lf mutants (Fig. 2d).

Constitutive activation of SKN-1 alters age-dependent somatic lipid redistribution in long-lived mutants

As previously reported, skn-1gf mutant animals display an Asdf phenotype where somatic lipid stores are depleted by mobilizing these lipids to the germline by vitellogenins (Lynn et al. 2015, Perez and Lehner 2019). As such, we used ORO staining to assess the distribution of intracellular lipid stores in the somatic and germline tissues of animals harboring the daf-2lf allele, skn-1gf allele, or both. We found that the loss of insulin signaling in daf-2lf mutant animals was capable of fully suppressing the Asdf phenotype observed in skn-1gf mutants (Fig. 3a; Supplementary Fig. 3a). We next examined the impact of the eat-2lf and separately the glp-1lf mutations on the skn-1-dependent somatic lipid depletion phenotype. In support of our previous findings (Lynn et al. 2015), loss of eat-2, which results in reduced food consumption (Lakowski and Hekimi 1998) that can independently activate SKN-1 (Bishop and Guarente 2007), resulted in reduced somatic lipid stores (Fig. 3b; Supplementary Fig. 3a), but surprisingly, the glp-1lf mutation suppressed Asdf (Fig. 3c; Supplementary Fig. 3a), which suggests that the somatic lipid depletion phenotype in animals with constitutive SKN-1 activation requires functional GLP-1 signaling in the germline.

Strikingly, the daf-2lf;skn-1gf double mutant animals display reduced expression of lipid metabolism genes in classes such as fatty acid chain elongation, beta oxidation, and acetyl transferases (Supplementary Fig. 3b and Table 2), which given the well-established connection between lipid homeostasis, SKN-1 activity, and lifespan (Hou and Taubert 2012, Paek et al. 2012, Lynn et al. 2015, Johnson and Stolzing 2019) could explain the maintenance of somatic lipid distribution and increased longevity of skn-1gf mutants when IIS is impaired.

Taken together, our genetic analysis of the impact that constitutive SKN-1 activity exerts on established genetic perturbations that drive longevity suggests an interaction model where skn-1 pathways monitor nutrient uptake and integrate with the insulin signaling pathway in order to induce somatic lipid redistribution, which cannot occur without signals from the proliferative germline (Fig. 3d).

Discussion

The biology of aging research field has focused on single gene mutations that promote longevity and separately mutations that can accelerate aging, but interactions between the two remain understudied. The complexity of the aging process and the genetic variation that contributes to health outcomes requires an examination of effects of genetic variation in polygenetic models. The constitutively active SKN-1 mutants display accelerated aging phenotypes yet present an interesting paradox, where moderate activation (Blackwell et al. 2015) and even moderate overexpression (Blackwell et al. 2015) of SKN-1 can be beneficial for the worm and even required in some circumstances to positively modulate lifespan; however, having SKN-1 always active results in a severe reduction in lifespan (Bishop and Guarente 2007, Tullet et al. 2008, Onken and Driscoll 2010, Paek et al. 2012, Blackwell et al. 2015, Turner et al. 2024). Our combination of constitutively active skn-1 mutations with established longevity-promoting mutations has allowed us to uncover new ways that SKN-1 cytoprotection can interact with longevity paradigms.

While eat-2lf mutants are dietarily restricted due to the reduction in feeding rate (Avery 1993, Lakowski and Hekimi 1998), skn-1gf mutants replicate a starvation response by the induction of lipid utilization genes (Paek et al. 2012). Our findings confirm that the eat-2lf mutants undergo somatic redistribution of lipids to the germline which is in line with the activation of SKN-1 that occurs in response to the actual starvation state of these mutants (Lynn et al. 2015). Furthermore, our observation that the eat-2lf;skn-1gf are exceptionally short-lived confirms previous observations that the degree of SKN-1 constitutive activation can correlate with the measured reduction in lifespan (Paek et al. 2012, Turner et al. 2023). The previous observation that SKN-1 activity in the pair of ASI sensory neurons is required for longevity of dietarily restricted animals (Bishop and Guarente 2007) and for Asdf (Turner et al. 2023) lends further evidence to the role(s) SKN-1 can play in monitoring nutrient status (Bishop and Guarente 2007). Additionally, our findings may provide evidence of a decoupling between lipid mobility and longevity. The eat-2lf mutants display a lipid mobilization phenotype similar to skn-1gf while displaying increased longevity. The addition of skn-1gf does not further change the lipid phenotype; however, it does ablate the longevity phenotype suggesting that there are likely lipid-independent mechanisms for the change in longevity.

The insulin/IGF-1 signaling pathway is an evolutionarily conserved modulator of healthspan and lifespan (Tissenbaum and Ruvkun 1998). Although DAF-16/FoxO mediates the longevity responses through insulin signaling (Kwon et al. 2010), SKN-1 has also been demonstrated to potentiate the physiological outcomes from reduced insulin signaling (Tullet et al. 2008). Intriguingly, our finding that impaired insulin signaling can increase the lifespan of animals with constitutively activated SKN-1 suggests that insulin-like signaling has roles downstream of constitutive SKN-1 activity or effects parallel lifespan processes. Alternatively, previous work has demonstrated that redirection of SKN-1 activity away from specific promoters can mitigate the negative health outcomes of constitutive SKN-1 activity (Nhan et al. 2019). It is possible that the activation of DAF-16 in the daf-2lf mutants could result in a competition model that redirects the transcriptional focus of SKN-1gf targets (Fig. 3d) that results in the maintenance of somatic lipids and improves lifespan. Another way of analyzing the genetic relationship between skn-1gf and daf-2lf is that rather than impaired insulin signaling being beneficial to skn-1gf, constitutively activated SKN-1 is detrimental to animals with impaired insulin signaling. Given that SKN-1 activation can suppress DAF-16-mediated stress tolerance (Deng et al. 2020), it is possible that constitutive SKN-1 activation abrogates the longevity-promoting effects of DAF-16 in insulin signaling mutants. This idea is supported by our transcriptomic analysis, where we find that the expression of several DAF-16 target genes is altered in the daf-2lf; skn-1gf animals as compared to daf-2lf mutants alone (Supplementary Tables 2 and 3). Furthermore, examination of SKN-1 target expression between skn-1gf and skn-1gf;daf-2lf reveals that SKN-1 targets are not differentially expressed (Supplementary Table 2) indicating that a change in skn-1 gene expression may influence longevity in the context of reduced IIS. Additionally, our observation that constitutive skn-1 activation reduces daf-16lf longevity indicates that activation of SKN-1 under reduced IIS may inhibit longevity in pathways parallel to daf-16lf. Furthermore, our examination of lipid homeostasis has provided insight into the role that insulin signaling plays in the distribution of stored intracellular lipids. daf-2lf mutants have been demonstrated to store increased levels of lipids in their somatic tissues (McElwee et al. 2006), and our findings reveal that even with constitutive activation of SKN-1, daf-2lf mutant animals do not deplete somatic lipids as skn-1gf mutants. This suggests that SKN-1, when activated, requires the insulin signaling pathway to modulate lipid redistribution to the germline. Lastly, it is important to consider the tissue specificity of both the insulin-like signaling pathway and skn-1. Evidence from tissue-specific expression assays suggests that expression of daf-2/IIS in the neurons is capable of rescuing longevity of daf-2 mutants (Wolkow et al. 2000). Additionally, neuronal-specific expression of daf-16 can also moderately rescue longevity (Libina et al. 2003). Taken with the observation that SKN-1 activity is required in the neurons to mediate dietary restriction longevity (Bishop and Guarente 2007), and constitutive SKN-1 activation only in neurons is sufficient to recapitulate phenotypic outcomes of the skn-1gf allele (Turner et al. 2023, Nair et al. 2024), it is probable that signaling between the IIS and the SKN-1 cytoprotection pathway occurs within the nervous system, albeit action in distinct neurons is also important.

The relationship between reproduction and organismal longevity across organisms is well-established (Le Bourg 2007, Antebi 2013, Brooks and Garratt 2017). In C. elegans, the specific loss of germline progenitor cells leads to a significant increase in lifespan (Nhan et al. 2019), and similarly, the loss of the germ cell proliferation mediated by GLP-1/NOTCH signaling can extend lifespan (Berman and Kenyon 2006). The finding that glp-1lf;skn-1gf double mutant lifespan was unremarkable as compared to skn-1gf single mutants suggests that the GLP-1/NOTCH signaling effect on longevity requires the ability to turn off SKN-1 activation. Interestingly, glp-1lf;skn-1gf double mutants do not undergo redistribution of somatic lipids which uncouples reduced lifespan and Asdf phenotypes associated with constitutive SKN-1 activity. This finding also suggests that a germ cell proliferation is required for lipid redistribution and not simply the somatic gonad, which glp-1lf mutants still possess. This could indicate that during stressful events that activate SKN-1, the proliferative germline signals to somatic tissues, like the intestine, where lipids are stored to mobilize these resources to promote reproduction and ensure fitness (Lynn et al. 2015).

Each of these pathways likely activates SKN-1 post-transcriptionally as an examination of our RNA-seq with daf-2 and prior transcriptional profiles assembled from each of the longevity mutants used in this study (Steinbaugh et al. 2015, Ng et al. 2020, Zhang et al. 2022, Ham et al. 2024) did not reveal a significant change in skn-1 expression. Collectively, our study demonstrates the need to tightly regulate SKN-1 activity in the context of long-lived genetic backgrounds to ensure maximal longevity. Although other genetic pathways are likely influenced by constitutive SKN-1 activity, our work provides a model to approach the examination of the polygenic nature of organismal longevity and health across the lifespan.

Data availability

Plasmids and strains will be available upon request. All data are available in the main text or the Supplementary material. RNA-seq data are available at the NIH (GEO) Gene Expression Omnibus GSE270810.

Supplemental material available at GENETICS online.

Acknowledgments

We thank S. Ledgerwood for the technical assistance and C.M. Ramos for the critical reading of the manuscript.

Funding

This work was funded by the NIH R01AG058610 and Hevolution Foundation award HF AGE-004 to SPC and T32AG052374 to CDT. We also thank the USC School of Gerontology Imaging Core that is funded in part by the Nathan Shock Center of Excellence (P30AG068345). Some strains were provided by the CGC, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). We thank WormBase for database curation and data access.

Author contributions

Conceptualization: S.P.C. Methodology: S.P.C. Investigation: C.D.T. and S.P.C. Visualization: C.D.T. and S.P.C. Supervision: S.P.C. Writing: C.D.T. and S.P.C.

Literature cited

Author notes