Cell cycle controls stress response and longevity in C. elegans

Inhibition of cell cycle regulates adult lifespan in C. elegans

To assess the function of cell cycle factors in regulation of longevity we first knocked down the cyclin-dependent kinase cdk-2 by RNA interference (RNAi). We used RNAi feeding starting from L4/early adulthood to inhibit cdk-2 and obtain results that were not confounded by its developmental functions. We found that post-developmental inactivation of cdk-2 had a strong effect on lifespan with an increase of 28% compared to wild type control (24.1 d in cdk-2(RNAi) vs 18.8 d control(RNAi)) (Figure 1A and Table 1). Thus, inhibition of cdk-2 function during adulthood is sufficient to confer longevity. The activity of cyclin-dependent kinases (CDKs) and their substrate specificity is regulated by cyclins. CDK-2 acts in concert with the cyclin E homolog CYE-1 to govern mitotic cell cycle [reviewed in 30]. Therefore we next tested knockdown of cye-1. cye-1(RNAi) resulted in lifespan extension similar to cdk-2(RNAi) (24.9 d in cye-1(RNAi) vs 19.5 d control(RNAi)) (Figure 1B and Table 1).

To determine whether the timing of cell cycle gene knockdown might influence lifespan, we initiated cye-1 RNAi feeding either during development (starting from L1 throughout life), in L4, or in adulthood (day 3) when the worms produced eggs. Interestingly, we found that knockdown of cye-1 during development as well as in adulthood resulted in longevity (Supplementary Figure 1A and Supplementary Table S2) suggesting that cell cycle events in the larvae may also modulate the aging process. Furthermore, inhibition of progeny production by supplementation of fluorodeoxyuridine (FUDR) did not affect the longevity of cye-1(RNAi). We obtained comparable results with or without inclusion of FUDR (Supplementary Figure 1B and Supplementary Table S2).

Many C. elegans mutations and manipulations that increase lifespan also confer resistance towards diverse forms of stress [31]. We analyzed heat and oxidative stress response of cye-1 and cdk-2 deficient worms. Knockdown of cye-1 and cdk-2 by RNAi enhanced tolerance to thermal stress (Figure 1C) and oxidative stress (Figure 1D) compared to wild type control.

C. elegans displays age-related degenerative changes including reduction of pharyngeal pumping and body movement during its lifespan [32]. There is a positive correlation between the decline of neuromuscular behavior and survival probability. We therefore assessed whether inhibition of cell cycle genes could also attenuate the decline of locomotory functions. We observed that cye-1(RNAi) and cdk-2(RNAi) delayed the age-related decline of pharyngeal pumping (Supplementary Figure 1C) and body movements (Supplementary Figure 1D). We also quantified the number of progeny after cye-1 and cdk-2 knockdown by RNAi. Consistent with essential roles of cye-1 and cdk-2 for cell cycle progression and embryonic development, cye-1(RNAi) and cdk-2(RNAi) significantly reduced the number of progeny compared to wild type (Supplementary Figure 1E), although the effect was only partially penetrant. We further assessed the morphology of the gonad after 2 generations of RNAi feeding against cye-1 and cdk-2, and observed that a portion of the animals were sterile or even arrested during larval development (Supplementary Figure 1E and 1F).

Given the profound impact that inhibition of cye-1 and cdk-2 had on longevity, we wondered if other components of the cell cycle also control aging. We therefore tested activators of the cell cycle including the cyclin-dependent kinases cdk-1 and cdk-4, their cyclin partners cya-1, cyb-1 and cyd-1, and the CDK-activating phosphatase cdc-25.1. We observed that RNAi-knockdown of these core components of the cell cycle machinery during adulthood significantly extended lifespan (Figure 1E and Table 1). We focused our analyses on cdk-2 and the corresponding cyclin cye-1 because they are well described key regulators of the cell cycle and robustly extended lifespan.

Cell cycle genes regulate longevity through the germline

To determine if an intact germline is necessary for lifespan regulation by cell cycle genes, we examined the effect of cye-1 and cdk-2 knockdown in glp-1 mutant worms. glp-1 encodes a Notch family receptor that is essential for mitotic proliferation of germline stem cells [33,34]. We used the glp-1 temperature-sensitive mutant allele e2141. When grown at the non-permissive temperature of 25°C, these mutants lack the germline and are long lived [7,34]. We observed that knockdown of cye-1 and cdk-2 by RNAi did not further extend the longevity of glp-1(e2141) mutant worms (Figures 2A and S2A, Table 1 and S2) suggesting that cell cycle genes require the germline to regulate longevity. We found that this effect could be reproduced by another germline-defective mutant, mes-1(bn7). The mes-1 gene controls initial development of the embryonic germline [35]. At 20°C, about 50% of mes-1(bn7) mutants lack the germline, and these sterile animals live up to 50% longer than their fertile siblings [7]. Knockdown of cye-1(RNAi) was able to extend the lifespan of fertile mes-1 worms (21%), but not sterile mes-1 animals (Figure 2B and Table 1). We further assessed the function of cell cycle regulators specifically in the germline using rrf-1 mutants. RRF-1 is mainly required for RNAi in somatic tissues. Hence, in animals lacking rrf-1 activity RNAi is efficient in the germline but it functions poorly in somatic cells [36]. We found the cye-1(RNAi) strongly extended lifespan in rrf-1(pk1417) mutants (18% compared to rrf-1 mutants fed with control(RNAi)). The effect was comparable to wild type (Supplementary Figure 2B and Supplementary Table S2). Together, these findings indicate that cell cycle genes function in the germline to influence lifespan.

The longevity of animals lacking the germline depends on steroid hormone signaling pathways. Mutations in the nuclear hormone receptor daf-12 prevent the lifespan-extending effect when germline stem cells are removed [1,10,37]. To test whether cye-1 genetically interacts with daf-12 to regulate lifespan, we examined the effect of cye-1 knockdown in the daf-12(rh61rh411) null mutant. We found that daf-12(rh61rh411);cye-1(RNAi) animals have significantly shortened lifespans compared with cye-1(RNAi) (Figure 2C and Table 1). Enzymes that produce DAF-12 ligands, such as the cytochrome P450 DAF-9 also contribute to steroid hormone signaling and mutation of daf-9 abrogates the longevity of germline-less animals [13,37]. The loss of function mutant daf-9(rh50) displays a slightly reduced lifespan, presumably because DAF-12 is not fully activated [37]. We observed that the lifespan extension of cye-1(RNAi) was strongly suppressed in the daf-9 mutant background (Figure 2D and Table 1). Together, these observations suggest that cye-1 exerts its effect on lifespan through the steroid hormone signaling pathway including DAF-9 and the nuclear hormone receptor DAF-12.

Longevity associated with inhibition of cell cycle requires DAF-16

It is well established that the activity of the transcription factor DAF-16/FOXO is required for germline removal to extend lifespan, as daf-16 mutants completely abolish longevity of germline-less animals [1]. Upon germline elimination DAF-16 accumulates in the nucleus of intestinal cells and stimulates expression of target genes [12,38,39]. We first analyzed the role of cell cycle genes in the DAF-16-dependent germline longevity pathway. We found that the daf-16(mgDf47) null mutation [40] significantly suppressed the longevity of cye-1(RNAi) (Figure 3A and Table 1) and cdk-2(RNAi) treated worms (Supplementary Figure 2C and Supplementary Table S2). Interestingly, the daf-16(mgDf47) mutation did not fully block the pro-longevity effect of cell cycle inhibition. RNAi against cye-1 and cdk-2 produced a modest but consistent increase in the mean lifespan of the daf-16 mutant (11% and 10%, respectively), but to a much lesser extent than in wild type (30% and 28%) (Table 1 and S2). We tested the genetic interaction with a different daf-16 null allele: again knockdown of cye-1 slightly extended the mean lifespan of daf-16(mu86) mutants (14% compared to 23% in wild type) (Supplementary Figure 2D and Supplementary Table S2). These findings indicate that the longevity benefits of cell cycle inhibition largely depend upon DAF-16 but also raise the possibility that a different mechanism might be involved.

The ankyrin-repeat protein KRI-1/KRIT1 is needed for loss of the germline to target DAF-16 to the nucleus and extend lifespan [10]. Next, we investigated if kri-1 is also required for the longevity of cell cycle deficient worms. We observed that the longevity associated with cye-1(RNAi) and cdk-2(RNAi) was strongly suppressed by lack of kri-1, shortening mean lifespan by up to 18% (Figure 3B and S2E, Tables 1 and S2).

The reproductive pathway functions in a synergistic manner with the daf-2/insulin-like signaling pathway to regulate DAF-16, as loss of the germline greatly extends the long lifespan of daf-2 mutants [1]. Accordingly, we observed that knockdown of cye-1(RNAi) in daf-2(e1370) mutants further increased their longevity (Figure 3C and Table 1) indicating that cell cycle genes act in the DAF-16-dependent germline longevity pathway but not in the insulin signaling pathway.

SKN-1/Nrf is required for longevity from reduced cell cycle activity

It was intriguing that daf-16 null mutants did not completely suppress longevity associated with inactivation of cye-1 and cdk-2. Our findings suggested that inhibition of the cell cycle might affect longevity through a pathway in parallel to DAF-16. The transcription factor SKN-1/Nrf has conserved functions in stress defense, protein homeostasis, and metabolism and promotes longevity [41,42]. Recently, SKN-1 has been shown to promote longevity in the absence of germline stem cells [15,16]. We wondered whether SKN-1 might act in parallel to DAF-16 for longevity under conditions in which cell cycle processes are inactive. We therefore tested the effects of cye-1 and cdk-2 RNAi knockdown in the skn-1(zu67) mutant and wild type, and found that the pro-longevity effect of cye-1 and cdk-2(RNAi) was essentially prevented by mutation of skn-1 (Figure 3D and Supplementary Figure 2F, Table 1 and Supplementary Table S2). These observations indicate that SKN-1 is fully required for the effects of cell cycle factors on longevity.

Cell cycle inhibition induced a transcriptional stress response that involves SKN-1 and DAF-16

Our results raise the question if the activity of SKN-1 is actually controlled by inhibition of the cell cycle. SKN-1 accumulates in intestinal nuclei and upregulates expression of detoxification genes in response to certain stresses or when mechanisms that limit SKN-1 activation are inhibited [41-44]. We further investigated how RNAi knockdown of cell cycle factors activates SKN-1 dependent transcriptional response and analyzed the expression of the SKN-1 target gene glutathione S-transferase, gst-4 using a transcriptional GFP reporter (Pgst-4::GFP) [45]. In the intestine the Pgst-4::GFP reporter was expressed at very low levels under normal conditions (Figure 4A). Transcription from the transgenic gst-4 promoter was robustly induced by RNAi knockdown of cye-1 and cdk-2 (Figure 4A). This gst-4 induction was greatly abolished by mutation of skn-1 (Figure 4B). Similarly, inhibition of the cyclin cya-1, cyb-1, and cyd-1, the cyclin-dependent kinases cdk-1 and cdk-4, and the CDK activator cdc-25.1 resulted in robust gst-4 activation through skn-1-dependent mechanisms (Supplementary Figure 3 A-C).

We next analyzed effects on endogenous SKN-1–regulated gene expression in adult worms after inhibition of cye-1 through RNAi. Here, we focused on cye-1 representative for different cell cycle factors. We used quantitative PCR to analyze the mRNA levels of a set of well characterized SKN-1 targets involved in various stress processes like glutathione S-transferase gst-4, γ-glutamyl cysteine synthetase gcs-1, oxidoreductase F20D6.11, nitrilase nit-1, and ABC transporter hmt-1 [42,43]. mRNA production of gst-4 and nit-1 has also been shown to depend on SKN-1 under normal (unstressed) conditions, while gcs-1, hmt-1, and F20D6.11 are upregulated by SKN-1 in response to stress [42]. We observed that RNAi knockdown of cye-1 increased the expression of several endogenous SKN-1 target genes in adult worms (Figure 4C). We conclude that inhibition of cell cycle factors promotes the transcriptional activity of SKN-1 for detoxification and stress defense.

Given the importance of DAF-16 for longevity associated with cell cycle inhibition we investigated whether cye-1 influences the expression of DAF-16 target genes. First, we analyzed the expression of the conserved DAF-16/FOXO target superoxide dismutase (sod-3). In adult worms, a sod-3::GFP transgenic reporter was expressed at low levels under normal conditions, but was elevated by cye-1(RNAi) (Figure 4D). The up-regulation of sod-3 was strongly attenuated in daf-16 mutants (Figure 4E). Next, we used qPCR to analyze how knockdown of cye-1 affected the transcriptional profile of well-identified DAF-16 target genes [46,47]. Several stress resistance genes like sod-3, mtl-1 encoding a metallothionein protein, and hsp-12.6 encoding a small heat shock protein were upregulated under cye-1(RNAi) conditions (Figure 4F). We also analyzed the expression of genes known to be upregulated by germline removal, including gpd-2, stdh-1 and nnt-1 [11]. We found that the expression of these genes was strongly induced by cye-1(RNAi) knockdown (Figure 4F). Together, our data indicate that when cell cycle factors are inhibited, SKN-1 and DAF-16 induce a protective transcriptional response.

C. elegans strains

C. elegans strains were grown on NGM agar plates with E. coli OP50 lawns [65]. Mutant strains are described in Supplementary Table S5.

RNA interference

RNAi plasmids were picked from the Ahringer [66] or ORF-RNAi [67] libraries and confirmed by sequencing (see Supplementary Table S6). For RNAi feeding experiments cultures were grown overnight in LB-medium with 12.5µg/ml tetracycline and 50µg/ml ampicillin. The following day, cultures were diluted 1:10 in 50µg/ml ampicillin LB-medium and grown to an OD595 of 0.8-1.1. Bacteria were spun down and resuspended in 1/10 of the original volume containing 50µg/ml ampicillin and induced with 0.7mM IPTG. This culture was seeded onto NGM agar plates containing 50µg/ml ampicillin, 12.5µg/ml tetracycline and 1mM IPTG. Empty vector plasmid pPD129.36 (L4440) was used as control.

Lifespan assays

Lifespan assays were performed unless otherwise indicated at 20°C. Animals were age-synchronized by timed egg laying and allowed to develop on control empty vector RNAi plates at 20°C. At L4 larval stage, worms were transferred on NGM RNAi plates containing 50 µM FUDR to prevent hatching of larvae. Animals were examined every day until the end of the assay. Worms that did not respond to gentle prodding with movement were scored as dead. Worms that crawled off the plates, showed extruded organs, or exploded due to internally hatching larvae were censored. All lifespans were plotted with L4 as time-point 0. Data were analyzed using a log-rank test in SigmaPlot 11.0.

To induce sterility in glp-1(e2141) mutants, adult animals were allowed to lay eggs at 15°C and the progeny were shifted to 25°C at L1 until L4/early adulthood. They were then placed on plates containing FUDR and RNAi bacteria for lifespan assays at 20°C. To prevent dauer formation of daf-2(e1370), daf-12(rh61rh411), and daf-9(rh50) mutants, these strains were allowed to develop to early adulthood at 15°C before starting the lifespan at 20°C.

Stress resistance assays

For the analysis of resistance against oxidative stress and heat, wild type worms were fed with RNAi for 4 days starting at L4 stage at 20°C. Worms were then either placed on plates containing 7.5 mM tert-butyl hydroperoxide (TBHP) or at 35°C preheated seeded RNAi plates. The criteria for analysis and censoring were essentially the same as described for lifespan assays.

Healthspan assays

Worms were raised as described for lifespan assays. To quantify pharyngeal pumping rates, worms were examined under a dissecting microscope every other day starting at the first day of adulthood. The number of contractions of the terminal bulb was counted for 30 sec.

For thrashing assays, worms were transferred into phosphate-buffered saline and observed for thrashing movements after 5 to 10 minutes of recovery. Quantification of thrashing movements was performed on alternate days beginning on the first day of adulthood. A thrash was defined as a full change in the lateral bending direction of the whole worm corpus. Pairwise statistical comparison was performed using a Mann-Whitney U-Test.

Brood size

Hermaphrodites were allowed to lay eggs on empty vector control. Their progeny was raised at 20°C until L4. Individual larvae were placed on fresh RNAi plates containing either empty vector control, cye-1(RNAi), or cdk-2(RNAi) and transferred on new plates every day until egg-laying ceased. The total number of progeny was counted. Pairwise statistical comparison was performed using a Mann-Whitney U-Test.

Phenotypic analysis

Worms were fed with empty vector control, cye-1(RNAi), or cdk-2(RNAi) for two generations at 20°C. At day 2 of adulthood worms were mounted on 2% agar pads, immobilized and images were taken with a Zeiss Axioplan-2 microscope.

Microscopy and scoring of transgenic GFP reporters

Microscopy was performed using a Zeiss Axioplan-2 microscope equipped with an AxioCam camera and AxioVision-Software Rel.4.8. Worms were placed on 2% agarose pads on slides and stunned with 2 mM levamisole. GFP was detected using an EGFP-filter set (480/20 nm excitation, 510/20 nm emission). Images were processed with AxioVision Rel 4.8 and Adobe Photoshop CS6.

Intestinal Pgst-4::GFP expression was assessed as described [43]. Worms were raised on RNAi until L4/young adult stage. Scoring was as follows: “high”, strong GFP levels throughout most of the intestine; “medium”, intense GFP levels anteriorly and/or posteriorly; “low”, only barely visible or no GFP expression observed.

For Psod-3::GFP expression L4 larvae were placed on RNAi and four days later GFP intensity was scored in the adult worms. Fluorescence intensity was categorized in “high”, worms displayed robust GFP intensity throughout their bodies or showed strong nuclear signal; “medium”, easily detectable nuclear signal at weak levels; “low”, no nuclear GFP signal observed. p values were determined from a Chi-square test.

RNA isolation and quantitative PCR

For RNA preparation approximately 500 animals were grown at 20°C on empty vector or cye-1(RNAi) plates for 4 days starting from L4 stage. Worms were washed in M9 buffer to remove bacteria. RNA was extracted with TriReagent (Sigma), treated with DNAse (Quiagen), and purified using a RNA purification column (ZYMO Research). RNA concentration and quality was assessed using a NanoPhotometer P-Class. cDNA was synthesized with reverse transcriptase (SuperScript III First-Strand Synthesis System, Invitrogen) and oligo-dT primer. qPCR runs were performed in technical triplicates on a Roche Light Cycler 480 using the Takyon No Rox SYBR MasterMix blue dTTP (Eurogentec).

Samples were analyzed by the 2^-ΔΔCt method with normalization to the geometric mean of the reference genes cdc-42 and Y45F10D.4. At least three biological replicates were examined for each sample. Primer sequences are listed in Supplementary Table S7.