A small-molecule Psora-4 acts as a caloric restriction mimetic to promote longevity in C. elegans

Abstract

In populations around the world, the fraction of humans aged 65 and above is increasing at an unprecedented rate. Aging is the main risk factor for the most important degenerative diseases and this demographic shift poses significant social, economic, and medical challenges. Pharmacological interventions directly targeting mechanisms of aging are an emerging strategy to delay or prevent age-dependent diseases. Successful application of this approach has the potential to yield dramatic health, social, and economic benefits. Psora-4 is an inhibitor of the voltage-gated potassium channel, Kv1.3, that has previously been shown to increase longevity and health span in the nematode Caenorhabditis elegans (C. elegans). Our recent discovery that Psora-4 lifespan benefits in C. elegans are synergistic with those of several other lifespan-extending drugs has motivated us to investigate further the mechanism by which Psora-4 extends lifespan. Here, we report that Psora-4 increases the production of free radicals and modulates genes related to stress response and that its effect intersects closely with the target set of caloric restriction (CR) genes, suggesting that it, in part, acts as CR mimetic. This effect may be related to the role of potassium channels in energy metabolism. Our discovery of a potassium channel blocker as a CR mimetic suggests a novel avenue for mimicking CR and extending a healthy lifespan.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11357-021-00374-6.

Introduction

5-Phenylalkoxypsoralen (Psora-4) is a potent, small-molecule inhibitor of voltage-gated potassium channels [1, 2]. In a large, unbiased screen of 1280 compounds, carried out by the Petrascheck group at the Scripps Research Institute, Psora-4 was identified as one of only 60 compounds that significantly extended the lifespan of Caenorhabditis elegans. This initial lifespan screen, as well as later confirmation studies, was conducted in a liquid medium [3]. We recently showed that Psora-4 lifespan benefits can be readily reproduced on solid NGM plates and that Psora-4 can extend lifespan on NGM plates, even if exposure only begins in adulthood [4]. Psora-4 is a selective inhibitor of the Kv1.3 potassium channel, has a relatively simple molecular structure and molecular weight of 334.37 g/mol. Studies in mammalian cells show that Psora-4 is both potent and highly selective in terms of the potassium channels that it inhibits [1, 5]. Psora-4 inhibits Kv1.3 by preferentially binding to the C-type inactivated state of the channel [1, 2]. Inhibition of Kv1.3 by Psora-4 has a low-nanomolar half-maximum inhibitory concentration (IC50 = 3nM) and shows up to 70-fold selectivity over most other potassium channels, including members of the closely related Kv1-family of potassium channels, except for Kv1.5 [1].

Psora-4 inhibition of the mitochondrial voltage-gated potassium channel results in increased mitochondrial reactive oxygen species (ROS) production, as shown by a significant increase in fluorescence emitted by MitoSOX, a mitochondrially targeted, ROS-sensitive dye [6]. In in vivo tests of toxicity, however, Psora-4 did not display any acute toxicity in rats, even after five daily subcutaneous injections at 33 mg/kg body weight which is equivalent to approximately 100 μM, well above the concentration where inhibition of Kv1.3 is expected to occur [1].

In addition to being a source of damage, ROS are important signaling molecules, involved in homeostasis, and can trigger adaptive mechanisms [7–9]. Small-molecule modulators affecting ROS-mediated signaling have therefore been explored as pharmacological activators of such defensive responses for therapeutic purposes, e.g., resulting in induction of antioxidant enzymes, increased repair of ROS-mediated damage, or downregulation of endogenous ROS production [9, 10]. The mechanism by which intrinsically harmful stimuli can, at sufficiently low levels, result in improved health is known as “hormesis” [9, 11–13]. Treatment of animals with carefully optimized (hormetic) levels of transient ROS generators activates endogenous antioxidant production and enables animals to better resist later ROS challenges [14]. Such hormetic responses can be translated into lifespan extension and health span improvement [9, 10]. Lithium, arsenate, and d-glucosamine are examples of chemical compounds for which such hormetic benefits have been reported [15–17].

In addition to its effects on ROS production, Psora-4 may increase glucose transport in mammalian cells [18], and loss of Kv1.3 activity has been reported to cause translocation of the glucose transporter GLUT4 to the cell membrane [19, 20]. GLUT4 is the most abundant insulin-sensitive glucose transporter in muscle and adipose tissue and functional GLUT4 is required for sustained growth, glucose and fat metabolism, and optimal longevity [21]. Importantly, caloric restriction (CR) increases the expression of GLUT4 in adipose tissue [22] and mouse muscle [22, 23]. This raises the question of whether Psora-4 mimics caloric restriction through modulation of glucose transport and to what extent, if any, the elevation in ROS resulting from exposure to Psora-4 contributes to its beneficial effects. Our first question, therefore, was whether the rise in ROS production following Psora-4 treatment was required for the lifespan benefits in C. elegans.

Results

A low dose of Psora-4 extends healthy lifespan in nematodes

We previously tested the effect of Psora-4 on nematode lifespan at different doses, finding that the optimal dose in our laboratory was 100 μM of Psora-4 [4, 24]. We, therefore, used 100 μM for all follow-up experiments, including those described here. Lifespan studies in C. elegans can be highly variable [25]. To accurately establish the typical effect size of Psora-4 under our laboratory conditions, we combined six independent lifespan studies comprising all previous trials carried out in our laboratory [4] with new data generated for this study (Figure S1). We found the typical benefit in terms of mean lifespan across these six independent experiments was 25±4% (mean ± SE) (Fig. 1a). These data allow us to directly compare the lifespan benefits of Psora-4 with those of the most efficacious lifespan-extending drugs currently known [27]. In our hands, adult-only exposure to Psora-4 extends lifespan to a similar degree as most efficacious lifespan-extending drugs such as rapamycin and metformin [4].

Fig. 1.

Psora-4 increases healthy lifespan of wild-type N2 nematodes. a Mean lifespan in days of several repeats of different doses of Psora-4 (previous and new data combined). 100 μM showed the best lifespan extension (log-rank). b Health span category of worms. Worms under Psora-4 treatment experienced a significantly larger percentage (51%) of their already extended lifespan at optimal health compared to untreated controls (39.2%) (P-value < 0.01). 100 worms per condition. Log-rank test was used to determine statistically significant differences in health span by taking category “A” animals as “healthy” and combining category “B” and “C” animals as “compromised.” c Health span category of worms using the scoring scheme of Herndon et al. Worms under Psora-4 treatment showed better health span across all stages of lifespan. d Multiple sequence alignment of human Kv1.3 potassium channel with C. elegans homolog, using EMBL-EBI Clustal Omega [26]. (Table S4 for percent identity score). Shadowed boxes indicate SF and S6 amino acid sequence regions respectively. “*”means the same letter/amino acid in each; “:”means very similar sequence; “.”means similar sequence; red means strongly aligned; blue means not strongly aligned. One-way ANOVA, ***P < 0.001, *P < 0.05

Another important question with any lifespan-extending intervention is whether lifespan extension is accompanied by health detriments or if the extension of lifespan is associated with an equal or larger benefit in health span. To determine whether Psora-4 treatment improved the health span of nematodes, we categorized the health status of each nematode based on their motility, following the scoring scheme of Herndon et al. [28]. Worms that showed clear, spontaneous sinusoidal movement were categorized as “healthy” (category “A”). Worms exhibited abnormal/non-sinusoidal movement or that only moved when prodded were classed as category “B” whereas those with only head or tail movement (partially paralyzed) were classed as category “C.” Interestingly, cohorts treated with Psora-4 contained a higher percentage of individuals in the healthiest categories (category “A”) compared to age-matched control across their lifespan (Fig. 1b, c). Together, these data show that Psora-4 extended not just lifespan but also health span and that treated animals enjoy better health for longer. Treated animals spent a larger fraction (51% or 13 days) of their already extended lifespan at optimal health compared to controls (which on average enjoy 39.2% or 8 days of good health) (Fig. 1b, c). There was, therefore, a significant increase in healthy days following Psora-4 treatment. By contrast, at the end of their lives, both Psora-4-treated and untreated animals spent on average the same amount of time (12 days) in compromised health (category “B” and “C”). However, this period of decline (ill health) represents only 49% of the lifespan of Psora-4-treated animals (12 days of the 25 days of mean lifespan) but it represents 60.8% of the total lifespan of control animals (12 days of the 20 days mean lifespan). Since both treated and untreated animals spend, on average, 12 days in ill health, there is no change in the time spent in ill health in absolute terms (all of the added lifetime is in good health). However, these 12 days of ill health represent a smaller percentage of the total lifespan of Psora-4-treated animals than of untreated controls. In other words, while there was no change in morbidity in absolute terms, there was a statistically significant compression of morbidity in relative terms (fraction of life spent in ill health is significantly smaller, P-value < 0.01).

Psora-4 as a potassium channel blocker

The amino acid sequence of human Kv1.3 potassium channel is highly similar to its C. elegans homolog (EMBL-EBI Clustal Omega [26], Fig. 1d, Table S4). The binding site for Psora-4 in human Kv1.3 involves the conserved channel S6 amino acids (residues 497 to 516) and SF (residues 479 to 484). The S6 sequence is highly similar, with 16 out of the 20 amino acids (80%) being identical or similar within the S6 helix and all the amino acids of the SF motif (TTVGYG) perfectly conserved (highlighted regions of Fig. 1d). The Psora-4 binding site is therefore highly conserved, including those residues that, when mutated, interfere with inhibition of Kv1.5 by Psora-4 in Xenopus laevis [2]. The channel domains of related nematode potassium channels, including the putative Psora-4 binding site, are highly similar, including that of Kv1.3, suggesting that Psora-4 is likely able to inhibit these channels in similar ways to the mammalian orthologs.

Psora-4 increases ROS production and activates stress resistance

Inhibition of Kv1.3 by Psora-4 in human lymphocytes was associated with a significant increase in ROS, as measured by an increase in fluorescence using the MitoSOX dye [6]. We, therefore, hypothesized that the same might be true in C. elegans and that lifespan benefits might be mediated by a mitohormetic response to an increase in ROS. This potential hormetic mechanism is consistent with the relatively narrow dose-response range in terms of lifespan that we observed for Psora-4 (Figure S1).

To test this hypothesis, we first determined whether treatment with Psora-4 induced ROS production at the same concentration at which it was most effective in extending lifespan. Given the mitochondrial origin of most ROS and the previous in vitro evidence for increased mitochondrial ROS production, we decided to use the same fluorescence ROS detector dye (MitoSOX) to test for in vivo changes in ROS levels in C. elegans [6]. After exposure to Psora-4 for 2 days, mitochondrial ROS production was evaluated by incubating intact nematodes with MitoSOX (see methods for details). We found that treatment with 100 μM Psora-4 indeed resulted in a consistent and significant increase in ROS (Fig. 2a).

Fig. 2.

Psora-4 increases ROS production and improves stress resistance. a Psora-4 resulted in a significant increase in mitochondrial ROS production (mean ± SD). NAC completely blocks the increase in ROS in worms treated with Psora-4 but does not affect endogenous ROS levels in untreated animals. 150 worms per well, 8 replicates. b Psora-4-treated worms exhibit improvement in stress resistance as measured by survival time on solid NGM plates containing 20 mM paraquat. 50 worms per condition. Log-rank test, P < 0.05. c The exogenous antioxidant/redox modulator NAC alone does not extend or decrease the lifespan of nematodes. Lifespan extension by Psora-4 is not affected by co-treatment with NAC. Lifespan extension in the presence and absence of NAC is unchanged (statistically insignificant, log-rank test). d Mean lifespan of nematode under psora-4 and NAC treatment. 100 worms per condition. One-way ANOVA, ***P < 0.0001

According to the mitohormesis model of lifespan extension, this increase in ROS production should result in increased stress resistance and be required for lifespan extension [29, 30]. To test if Psora-4-treated animals were more resistant to further stress when compared to untreated controls, we challenged animals that had been pre-treated with Psora-4 or untreated controls using N,N′-dimethyl-4,4′-bipyridinium dichloride (paraquat), a redox-cycling herbicide, by adding it both to the solid NGM agar and to the bacterial food source at a final concentration of 20mM. Treatment with paraquat results in increased ROS production and causes oxidative stress [29]. In untreated, wild-type C. elegans, exposure to paraquat at this level increases ROS production and oxidative damage and is lethal within 3 days [4]. Consistent with the mitohormesis hypothesis, animals pre-treated with Psora-4 were able to survive the subsequent oxidative stress caused by the paraquat challenge for a significantly longer time than untreated controls (Fig. 2b).

Psora-4 transcriptionally activates stress responses

To further investigate the mechanism underlying the increased stress resistance and lifespan extension, we next re-analyzed transcriptional changes following Psora-4 treatment of WT animals using a transcriptomics dataset that we had collected previously [9]. To identify genes and pathways that may be involved in lifespan extension of Psora-4, we determined differential expression of genes (DEG), comparing Psora-4-treated nematodes to untreated controls. Psora-4 treatment resulted in 681 DEGs with 71 genes being significantly upregulated and 610 genes downregulated following Psora-4 treatment (|LFC| > 2, P-value < 0.05) (Table S1). Based on these DEGs, we next identified the top pathways affected by Psora-4 treatment using Metascape [31]. Consistent with the mitohormesis hypothesis, defensive responses, which include targets of the transcription factor daf-16 and heat shock proteins (HSP16), were the top gene ontology (GO) category that was enriched (Fig. 3a, Table S2). Furthermore, “antioxidant activities” was the top molecular function category enriched based on Protein ANalysis THrough Evolutionary Relationships (PANTHER) [34] (Fig. 3e). Further splitting these DEGs into up- and downregulated genes followed by analysis separately for each category (up/down), we found that “oxidation-reduction process” and “biological oxidations” were the top GO terms associated with upregulated genes (Fig. 3c), while the top downregulated pathways were “metabolic processes including single-organism catabolic process,” “metabolism of amino acids and derivatives,” and “sulfur compound metabolic process” (Fig. 3b). We then repeated this GO term analysis using DAVID [32] pathway analysis and then summarized the GO terms using REVIGO [33] to remove redundant terms, identifying metabolic processes among the top enriched GO terms (Fig. 3f). In particular, we identified “oxidation-reduction processes,” “protein metabolism,” “cellular metabolism,” “transmembrane transport,” “carbohydrate metabolism,” “carbohydrate transport,” and “defensive response” as significantly affected (Fig. 3f). The specific upregulated genes include abu-8 (activated in blocked UPR), gst-8 (probable glutathione s-transferase 8), cbs-2 (cystathionine beta synthase), and sdz-1 (skn-1 dependent zygotic transcription 1) (Fig. 3f and Table S1). Taken together, these changes in the transcriptome suggest that exposure to Psora-4 indeed activates a wide range of defensive responses, including increases antioxidant activity and that this makes worms resistant to further stresses (Figs. 1 and 3). However, we also noted that most canonical antioxidant genes, such as superoxide dismutase genes, catalase, and peroxiredoxin genes, were not transcriptionally induced.

Fig. 3.

Transcriptome profile of Psora-4 confirmed that nematodes treated with psora-4 have enhanced stress response. a GO term analysis, top GO terms targeted by Psora-4 were defensive responses, transmembrane transport, and oxidation-reduction process, based on DEGs with |LFC| > 1 and P-value < 0.05. b As in a only for downregulated genes. Top hits are metabolic processes. c As in a for upregulated genes. Top hits are oxidation-reduction processes. d Top signaling pathways targeted by Psora-4 differentially expressed genes based on panther pathway analysis. e Panther GO-slim molecular function showed the top classes of differentially expressed genes by Psora-4 are antioxidant activities. f GO terms enriched by DEGs of Psora-4 were analyzed by DAVID [32] and summarized by REVIGO [33]. The top GO terms are metabolic processes. X- and Y-axes represent GO terms’ semantic similarity measures in multidirectional scaling; the closeness of the groups on the plot reflects their relative similarity based on this metric

Psora-4 extends lifespan in the presence of exogenous antioxidant

While the above evidence supports a ROS-mediated adaptive mechanism as underlying the increased stress resistance, it was unclear if this mechanism was causally linked to the observed lifespan extension. We, therefore, asked whether the increased ROS production was required for Psora-4 lifespan extension. It has previously been shown that phenotypes based on mitohormetic mechanisms can sometimes be blocked by co-exposure to N-acetyl-cysteine (NAC), presumably due to redox-signaling and thiol-driven blunting of the mitohormetic response by NAC [30]. To explore whether NAC was able to modulate Psora-4-mediated ROS and whether elevated mitochondrial ROS levels were required for Psora-4 lifespan benefits, we co-treated nematodes with 100μM Psora-4 and 5mM NAC [30, 35]. In untreated wild-type animals, in the absence of Psora-4, NAC did not significantly affect either basal ROS production or lifespan (Fig. 2a, d), but co-treatment with NAC completely abrogated the elevation of ROS production (as evaluated by MitoSOX) associated with Psora-4 treatment (Fig. 2a). Next, we tested if this loss of the ROS signal also prevented lifespan extension by Psora-4. While NAC successfully blocked the Psora-4-induced ROS signal, we found that this did not diminish its lifespan benefits. Psora-4 was able to extend the lifespan of nematodes in the presence of NAC to the same extent as in its absence (Fig. 2c, d). If Psora-4 lifespan extension was mediated by a hormetic or mitohormetic mechanism related to the observed increase in ROS, we would expect that lifespan benefits would be blunted or prevented by NAC treatment, as seen with other mitohormetic stimuli [36].

Psora-4 and eat-2 extends lifespan via an overlapping mechanism

During our analysis of transcriptional changes, we noticed the large number of genes differentially expressed by Psora-4 treatment were genes related to metabolism (Fig. 3f). This is consistent with the previous reports, implicating Psora-4 in the regulation of glucose metabolism through effects on GLUT4 [18]. Furthermore, cbs-2, an enzyme involved in H₂S synthesis and previously implicated in mediating benefits of CR was also differentially expressed by Psora-4 treatment [37].

We have previously compared Psora-4 phenotypes with those of eat-2 mutants [4, 24], a mutant strain often considered a genetic model of CR in C. elegans [38]. Interestingly, eat-2 and Psora-4 treatment both increase respiration in nematodes and our previous data also revealed that Psora-4 is unable to further extend the lifespan of eat-2 mutants [4]. These data are consistent with the notion that Psora-4 might be mimicking the same aspect of CR seen in eat-2 mutants [4]. To further evaluate the degree to which Psora-4-treated animals resembled eat-2 (eat-2 (ad1116)), we directly compared the transcriptome of Psora-4-treated wild-type nematodes with that of eat-2 mutants. Comparing these two sets of DEGs, we identified what appeared to be a significant overlap between Psora-4-treated N2 and eat-2 mutants (Fig. 4a, b). The purple link in Fig. 4b indicates genes commonly affected by both eat-2 and Psora-4 whereas the blue link indicates genes affected only by either Psora-4 or eat-2 but that share the same GO term. More blue and purple lines indicate a higher similarity of DEGs and GO terms affected by Psora-4 and eat-2. We next applied a hypergeometric probability test to determine if the similarity between eat-2 and Psora-4-treated N2 was statistically significant (http://nemates.org/MA/progs/overlap_stats.html). We found that among the downregulated genes, 175 of 610 genes affected by Psora-4 were also downregulated in eat-2 mutants and this intersection was statistically significant (representation factor = 2.5, P-value < 0.56e⁻³³, see “Methods”). Among the upregulated genes, we observed a similar pattern (representation factor = 3.4, P-value < 3.8e⁻⁴) (Fig. 4a). As a negative control, we also evaluated the intersection between genes that were affected by Psora-4 treatment and eat-2 but for which changes between Psora-4 and eat-2 had opposite directions. We found that there was no significant overlap within this set (representation factor = 0.9, P-value = 0.44). These data suggest that transcriptional changes associated with Psora-4 were significantly more similar to those of eat-2, both in terms of the genes affected and in terms of the direction of these effects, than expected by random chance. To determine if this holds true specifically for the genes that are known to regulate aging and lifespan, we repeated the analysis but restricted it to genes found in the GenAge database [39]. We first identified all GenAge genes differentially expressed in eat-2 and Psora-4-treated N2 nematodes (Table S3). We then, again, applied a hypergeometric probability test to determine if the overlap of Psora-4 and eat-2 is statistically significant. Interestingly, the lifespan-extending GenAge genes affected by Psora-4 overlapped with those affected by eat-2 to an even higher degree (representation factor = 91.7, P-value < 9.0e⁻⁵⁴). Psora-4 affected 43 lifespan-extending GenAge genes, of which 26 were also affected by eat-2. Most importantly, this overlap was due to the overlap of genes affected consistently (changes having the same direction), meaning genes upregulated by Psora-4 were also upregulated in eat-2 mutants and vice versa. From 26 GenAge genes affected by both Psora-4 and eat-2, 23 changed in the same direction.

Fig. 4.

Comparison of Psora-4 with eat-2 and effect on the lifespan of worms fed on different concentrations of bacteria. a A Venn diagram of differentially expressed genes. 193 genes were affected by both Psora-4 and in eat-2 mutants. Among these genes impacted by both Psora-4 and eat-2, 26 were known aging genes (as determined by having entries in the GenAge database). Overall, Psora-4 impacted 43 GenAge genes; therefore, 59% of GenAge genes (26 out of 43) affected by Psora-4 were also affected in eat-2 mutants. b Circos plot (generated by Metascape) shows gene expression overlaps for Psora-4 and eat-2. The outside arc represents genes that are up and downregulated by eat-2 or Psora-4. On the inside arc, dark orange represents genes affected by both eat-2 and Psora-4 and light orange represents genes unique to either conditions. Purple lines link genes commonly affected by eat-2 and Psora-4. c Lifespan of untreated worms fed on different concentrations of OP50-1. d Psora-4 extends the lifespan of worms fed on a normal diet (10¹⁰ CFU/ml). e Psora-4 results in only a small lifespan extension effect in worms fed on 10⁹ CFU/ml OP50-1 (P-value < 0.05). f Psora-4 was unable to further extend the lifespan of worms fed on 10⁸ CFU/ml (CR). g Direct comparison of Psora-4 lifespan in worms fed on different concentrations of OP50-1; lifespan of Psora-4-treated animals is similar on all food concentrations and indistinguishable from lifespan at highest level of CR (lowest food concentration)

In summary, the transcriptional changes between Psora-4-treated wild-type N2 and eat-2 mutants were similar in both magnitude and direction, and these transcriptional similarities were more pronounced for the subset of genes known to be involved in lifespan extension. This insight supported the hypothesis that the mechanism of Psora-4 mimics key responses to caloric restriction, in particular those involved in lifespan extension.

Lower concentrations of OP50-1 (e.g., 10⁸ CFU/ml) have been used previously to induce CR more directly and in a way that is independent of mutations such as eat-2 [40]. We compared the effect of Psora-4 on normal-fed controls (10¹⁰ CFU/ml of OP50-1) with worms under CR conditions, fed on either 10⁹ or 10⁸ CFU/ml OP50-1. As expected, worms fed with OP50-1 at a concentration of 10e⁸ CFU/ml showed significantly longer lifespans compared to those fed at higher CFU/ml (either 10⁹ CFU/ml or our standard “normal” food concentration of 10¹⁰ CFU/ml) (Fig. 4c). Importantly, exposure to Psora-4 was unable to further extend the lifespan of worms at the lowest food concentration (10⁸ CFU/ml) and had only a small beneficial effect on worms feed with 10⁹ CFU/ml (P-value < 0.05), whereas it consistently resulted in significant lifespan extension effects on the standard diet (Fig. 4d–g). In agreement with our eat-2 result, Psora-4 was not able to further extend lifespan under CR conditions. Our data suggest that Psora-4 treatment reproduces lifespan benefits of CR (Fig. 4g).

Psora-4 decreases triglycerides (TAG) accumulation

We previously reported that Psora-4 lifespan extension requires the transcription factor daf-16 but is capable of further extending the lifespan of mutants of the IGF receptor, daf-2, upstream of it [4]. Together with the fact that Psora-4 is unable to further extend the lifespan of eat-2 mutants [4] and that it caused transcriptional changes that were highly similar to those of eat-2 (Fig. 4), these data suggest that the key downstream genes responsible for Psora-4 lifespan extension are downstream of both daf-16 and controlled by mechanisms also impacted by the eat-2 mutation but that are independent of daf-2. The subset of daf-16 target genes that are explicitly dependent on daf-2 has previously been reported [41]. Based on these data and our own [4], we identified genes that were differentially expressed both in eat-2 and Psora-4 treatment but that are not controlled through the daf-2/daf-16 axis.

Based on this analysis, we identified only 11 genes (highlighted in purple in Fig. 5e) that showed consistent gene expression changes (changed in the same direction) within the DEGs of both eat-2 and Psora-4 treatment (Fig. 5e) but that were not part of the established gene-set downstream of daf-2/daf-16. Within this set, six genes have no known function (Table S3). The remaining 5 genes comprise one gene involved in xenobiotic metabolism (oat-1), a potassium channel (twk-33) with the remaining three genes being involved in lipid metabolism. The latter set includes the lipase (lips-13), the Δ9 desaturase (fat-5), and an acyl-CoA synthetase (acs-2). These three genes are directly controlled by the SBP-1 transcription factor [4, 42]. SBP-1 is a major transcription factor involved in regulating lipid metabolism [43]. We have previously shown that lifespan extension of Psora-4 is dependent on functional SBP-1 [4] and that the expressions of Δ9 desaturases and lipases are diminished in sbp-1 (ep79) mutant strains [4]. Therefore, since Psora-4 was unable to extend the lifespan of sbp-1 mutants, these changes are likely functionally important. This suggests that modulation of lipid metabolism through sbp-1 may be at the core of Psora-4 lifespan extension and CR mimetic efficacy.

Fig. 5.

Psora-4 decreases accumulation of long-chain fatty acyl–containing TAGs. a Heatmap of TAG species. Based on TAG abundance Psora-4-treated N2 (N2 + Psora-4) and untreated eat-2 (eat-2) were separated from untreated N2 (N2) and clustered together. b Two days of Psora-4-treated N2 worms and eat-2 mutants has lower total TAG accumulation than wild-type N2 nematodes. c Both Psora-4 treatment and eat-2 did not affect the abundance of medium-chain fatty acyl–containing TAGs. d The decrease in total TAGs is due to a decrease in long-chained fatty acyl–containing TAGs. e DEGs of Psora-4 treatment and eat-2 mutant that are downstream of daf-16 but not daf-2. Highlighted in purple are genes downregulated by Psora-4 and eat-2 in a similar direction. *** P-value < 0.001, ** P-value < 0.01

Eat-2 mutants are known to store significantly less fat than wild types [44]. Based on the above observations, we hypothesized that Psora-4 may affect lipid metabolism, in particular triglycerides (TAGs) in a similar fashion to eat-2. We have previously reported the effect of lifespan-extending drugs and drug combinations on the lipid profile of nematodes and made the complete dataset publicly available [24]. We re-analyzed these data to determine the effect of Psora-4 on fat content and to specifically compare lipid content and composition between eat-2 and WT with and without Psora-4 treatment. Based on the abundance of TAG species, untreated eat-2 mutants and WT nematodes treated with Psora-4 clustered together and were separated from untreated WT (Fig. 5a). We found that, as expected, eat-2 mutants indeed had significantly less storage fat (total TAGs) compared to wild-type N2 nematodes (Fig. 5b). An increase in medium-chain fatty acyl–containing TAGs is advantageous for health and lifespan [45]. Consistent with this insight, the decrease in total TAGs of eat-2 mutants was due predominantly to a decrease in long-chain fatty acyl TAGs while the medium-chain fatty acyl TAGs remained unaffected (Fig. 5c, d). Similarly, 2 days of Psora-4 treatment was sufficient to decrease the total TAGs in wild-type N2 nematodes by 20%, P-value < 0.001 (Fig. 5b). Again, as was the case for eat-2, this decrease in total TAGs was due predominantly to a decrease in long-chain fatty acyl–containing TAGs while the medium-chain fatty acyl TAGs remained unaffected (Fig. 5c, d).

Discussion

Our previous results confirmed that Psora-4 extends lifespan in C. elegans under the condition typically used in our laboratory to a similar extent as originally reported for liquid culture by the Petrascheck group [3, 4]. Lifespan-extending drugs are often considered mimetic CR [46, 47]. Indeed, several lifespan-extending drugs have previously been classified as CR mimetic based on similarity in transcriptional changes with those seen in eat-2 mutants [4, 48–50]. Some of the key metabolic effects of Psora-4 also resemble changes seen in eat-2 mutants (Fig. 5). We, therefore, compared the transcriptome profile of Psora-4-treated N2 nematodes with those of eat-2, finding a highly significant similarity between Psora-4 treatment and eat-2, both in the subset of genes affected and the direction of these effects.

Among the best-characterized nutrient-sensing pathways known to mediate lifespan extension by CR are AMPK, mTOR, FOXA3, and IGF. Psora-4 extends the lifespan of long-lived daf-2 mutants but not that of short-lived daf-16 mutants [4]. While daf-16 is one of the key downstream targets of daf-2, several other pathways also contribute to the modulation of daf-16 activity. Since Psora-4 lifespan extension is dependent on daf-16 but not downstream of daf-2, we investigated one such alternative mechanism upstream of daf-16, related to its primary function as a potassium channel blocker. Psora-4 is known to block mitochondrial potassium channels involved in modulating mitochondrial respiration [2]. Consistent increases in mitochondrial respiration have previously been linked to lifespan extension, possibly through hermetic and mitohormetic mechanisms [30]. Treatment with Psora-4 increases oxygen consumption in C. elegans and also increases ROS production as well as inducing activation of genes involved in defensive response, including increasing the expression of genes involved in oxidative stress responses (Figs. 2a and 3a). However, we find that the increase in ROS levels, as evaluated by MitoSOX, can be blocked by the addition of NAC without reducing Psora-4s effect on lifespan, suggesting that the key mechanism is not consistent with a classic mitohormetic mechanism.

Our comparison of gene expression changes following Psora-4 treatment with changes seen in eat-2 mutants revealed a high degree of similarity (Fig. 4). This similarity was also seen when comparing the effects of eat-2 mutants with Psora-4-treated animals in terms of metabolism as well as abundance and composition of storage lipids (Fig. 5), suggesting that inhibiting voltage-gated potassium channels may mimic key aspects of CR. In a previous study, we showed that combining Psora-4 with either metformin or rapamycin, two other drugs considered to function as partial CR mimetics, is non-synergistic. By contrast, combining Psora-4 with a non-CR mimetic compound (rifampicin) resulted in beneficial interactions and further lifespan extension [4]. This further supports our argument that Psora-4 acts by mimicking aspects of CR and potentially places its effects upstream of the CR-like function of rapamycin and metformin. Psora-4 has previously been reported to modulate glucose transport in mammalian cells [18], resulting in increased glucose uptake by facilitating GLUT4 translocation (depending on Ca²⁺ signaling). Because of its role in insulin-sensitive glucose transport in muscle and adipose tissue and because its expression is increased in adipose tissue following CR, modulation of GLUT4 is a possible alternative/additional mechanism by which Psora-4 might mimic CR [22, 23].

Caloric restriction is one of the few interventions that have been shown to extend lifespan and improve health span consistently in evolutionarily widely separated organisms and even in humans [51, 52]. However, practicing CR regiment in humans is difficult and it is associated with eating disorders and malnutrition and can lead to undesirable physiological and psychological effects (review in [53]). Therefore, drugs such as metformin, rapamycin, and Psora-4 that mimic key aspects of CR are promising avenues for aging interventions [4, 54, 55]. Together with evidence for lack of any acute toxicity [1], these data suggest that potassium channel blockers such as Psora-4 might be a promising class of compounds to explore as novel CR mimetic compounds.

Methods

Culturing C. elegans

For all experiments, Bristol wild-type N2 or mutant nematodes were grown and maintained on Nematode Growth Medium (NGM) agar plates at 20 °C using Escherichia coli OP50-1 bacteria as a food source unless otherwise noted. After plates were poured and dried, they were sealed and stored at 4°C. E. coli was spotted on plates the previous evening and allowed to dry. For Psora-4 treatment, all agar plates were prepared from the same batch of NGM agar and treatment plates were supplemented with the respective concentrations of Psora-4 or vehicle as a control. Fresh plates were prepared every week. The following worm strains were used: wild-type N2, eat-2 (DA1116), daf-2 (CB1370), daf-16 (CF1038). All strains were obtained from the Caenorhabditis Genetics Center (CGC).

Bacterial preparation method

A single-colony OP50-1 E. coli was picked and incubated at 37°C for 20 h with shaking at 180 rpm in LB broth supplemented with Streptomycin (final concentration 200μg/ml). Bacterial colony-forming units per ml (CFU/ml) were determined spectrophotometrically at a 600-nm wavelength, and the bacterial stock was diluted to 10¹⁰ CFU/ml, and frozen at −80 °C. NGM plates were loaded with bacteria at this final concentration and used for lifespan assays without further incubation. Drugs or vehicles were added to the bacteria (in addition to being added to NGM) just before transfer to plates.

Determination of lifespan and health span

For determination of C. elegans lifespan, nematodes were age-synchronized by bleaching and allowed to hatch and 100 young adult worms per condition were transferred to three 35-mm culture plates. Worms were transferred to fresh plates every day until progeny production ceased, and every 2 to 3 days afterward, until all worms had died. The number of worms that were alive was determined every other day, and dead worms were removed from the plate. Worms were considered dead when no touch-provoked movement was observed. Animals that crawled off the plate, exhibited extruded internal organs, or displayed an egg-laying defect with the subsequent hatching of larvae inside the mother were censored. All lifespan assays were blinded and repeated at least two times. Health span was determined based on their motility, following the scoring scheme of Herndon et al. [28]. Worms with spontaneous sinusoidal movement are categorized as healthy category “A.” Worms exhibiting abnormal/non-sinusoidal movement or that only move when prodded are categorized as category “B” whereas those with only head or tail movement (partially paralyzed) are classed as category “C.” Log-rank test was used to determine statistical significance in health span by taking category “A” animals as “healthy” and combining category “B” and “C” animals as “compromised.” The GraphPad Prism 5.0 software and OASIS 2 software [56] were used to calculate the mean adult lifespan and to perform statistical analysis of significance using the log-rank test.

Resistance to oxidative stress

Age-synchronized adult worms were transferred to a fresh NGM plate containing drug or vehicle. 5-fluoro-2′-deoxyuridine (FUdR) was added to prevent egg hatching. After 48 h of treatment, worms were transferred to plates with 20mM paraquat (methyl viologen dichloride hydrate, Sigma Aldrich) to induce oxidative stress in the worms. Thereafter, dead worms were counted every day until all of the worms were dead.

ROS production

Mitochondrial ROS production was determined using MitoSOX assay. Briefly, 150 adult nematodes were transferred into each well of a black, opaque 96-well plate containing 100μl M9 buffer. 10μM final concentration of MitoSOX red reagent was added into each well. Using a microplate reader, kinetics was measured every 2 min for 5 h at excitation of 396 nm and emission of 579 nm at room temperature.

RNA extraction, RNA-seq, and pathway analysis

Transcriptomics profiling was done based on our previous method. In brief, after 2 days of treatment with drug or vehicle, adult worms were washed off the plates, and worm pellets were then used for RNA extraction. Total RNA was isolated using QIAGEN RNeasy microkit (QIAGEN, Hilden, Germany) following the standard protocol. Afterward, RNA was quantified photometrically with NanoDrop 2000 and stored at −80 °C until use. The integrity of total RNA was measured by Agilent Bioanalyzer 2100. For library preparation, an amount of 2 μg of total RNA per sample was processed using Illumina’s RNA Sample Prep Kit following the manufacturer’s instruction (Illumina, San Diego, CA, USA). Libraries were sequenced using Illumina HiSeq4000 sequencing platform (Illumina, San Diego, CA, USA) in a paired-end read approach at a read length of 150 nucleotides. Sequence data were extracted in FastQ format. The RNA-seq reads from each sample were mapped to the reference C. elegans transcriptome (WBcel235) with Kallisto (v0.43.0) [57] using 100 bootstrap samples and sequence-based bias correction. The estimated counts were imported from Kallisto to the R environment (v3.3.2) and summarized to gene-level using the tximport package (v1.2.0) [58]. The DESeq2 package (v1.14.1) [59] was used to identify DEGs. Pathway and GO term enrichment analyses were determined by DAVID [32] and Metascape [31]. To identify the top GO terms from Metascape, we use an input gene list of DGE with P-value < 0.05 and P-value for Metascape enrichment analysis < 0.05.

Determination of the representation factor and the associated probability for gene overlap

Two groups of genes are compared and found to have x genes in common. A representation factor and the probability of finding an overlap of x genes are calculated. The representation factor is the number of overlapping genes divided by the expected number of overlapping genes drawn from two independent groups.

A representation factor > 1 indicates more overlap than expected of two independent groups, and a representation factor < 1 indicates less overlap than expected.

x = the number of genes in common between two groups.

n = the number of genes in group 1.

D = the number of genes in group 2.

N = the total genes.

The representation factor = x / expected number of genes.

Expected number of genes = (n * D) / N.

Lipidomics analysis

Lipidomics analysis is fully described in our previous paper [4] and raw data with detail description is available in [24].

Author contribution

T.D.A. and J.G. designed the study and wrote the manuscript. T.D.A. performed most experiments and analyzed results. T.D.A. and D.B. contributed to bioinformatics analysis. J.G., A.C.-G., M.R.W., and T.D.A. contributed to lipidomics experimental design. T.D.A. and K.C.B performed the lipidomics experiment. N.L.F. was involved in experimental design and writing.

Funding

This work was funded by the Ministry of Education Singapore (Grants MOE2014-T2-2-120 and IG17-LR005).

Data availability

All sequencing data that support the findings of this study have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO). The accession number for the sequence reported in this paper is GSE108263. The accession number for the lipidomics data reported in this paper is MetaboLights: MTBLS648.

Declarations

Conflict of interest

The authors declare no competing interests.