All five Wnts are expressed in adulthood
We wanted to examine whether any of the C. elegans Wnt ligands are expressed during aging. To do that, we took advantage of previously generated transcriptional GFP reporter constructs for all five Wnts (Gleason et al., 2006). Using these transgenic lines, we were able to observe temporal and spatial expression from these five Wnt reporters throughout the worm's lifespan, from day 2 (young) to day 12 (old) of adulthood.
We found that expression of lin-44 and egl-20 is high throughout the entire lifespan of the worm and increases during aging (Fig. 1B,C). As in development, lin-44 and egl-20 expression was observed only in the posterior of the animal. Expression of lin-44 is present in tail hypodermis, hyp8, hyp9, hyp10, and hyp11 cells (Fig. S1A). These are exactly the same expression patterns as previously described during larval development (Herman et al., 1995; Gleason et al., 2006). The egl-20/Wnt is expressed in the anal depressor muscle and in the postembryonic rectal epithelial cells (B, F, K, and U) during aging (Fig. S1B). These cells also express egl-20/Wnt during larval development (Gleason et al., 2006). However, we did not observe any expression in hypodermal or muscle cells, as previously reported (Gleason et al., 2006).
The expression dynamics of mom-2/Wnt are quite different throughout the worm lifespan. Expression of mom-2/Wnt increases 3-fold during the first 5 days of adulthood and then decreases 4-fold by day 8 of adulthood, eventually showing little or no expression in old worms (Fig. 1D). In contrast to lin-44 and egl-20, expression localization of mom-2 differs slightly between aging and development. During development, mom-2 is expressed throughout the whole body of the worm, in muscles, hypodermal and intestinal cells, vulva precursor cells, as well as in ventral cord motor neurons (Gleason et al., 2006). In young (day1 and 2) and middle-aged (day 5) adults, mom-2/Wnt expression was observed only in posterior intestinal and intestinal–rectal valve cells (Fig. S1C). We were not able to detect any mom-2 expression in any other tissues.
Expression of both cwn-1 and cwn-2 Wnt ligands decreases during aging (Fig. 1E,F). In very young adult worms, day 1 of adulthood, cwn-1/Wnt is expressed in body-wall muscles and ventral cord motor neurons in the posterior part of the worm (Fig. S2A). However, cwn-1/Wnt expression in body-wall muscles almost entirely disappears in the following 24 h. By day 2 of adulthood, we were able to observe only cwn-1 expression in ventral cord motor neurons. This expression is low in middle-age (day 5) worms and barely detectable in old (day 12) animals (Fig. 1E). Overall, the cwn-1/Wnt expression pattern is similar to that previously observed during development. In addition to body-wall muscles and ventral cord motor neurons, cwn-1/Wnt was detected in intestinal cells, vulval muscle, and fused seam cells during larvae development (Gleason et al., 2006), although this expression pattern is absent in adulthood.
The cwn-2/Wnt expression was observed all along the anterior–posterior body axis of the animal in body-wall muscles (Fig. S2B). We did not detect any expression in ventral cord motor neurons or intestinal cells as was previously observed during worm development (Gleason et al., 2006). This observation was particularly troubling as it has been previously reported, using similar genetic constructs, that cwn-2/Wnt is expressed at high levels in pharynx and SMD neurons in adult worms (Song et al., 2010). These discrepancies could potentially affect our conclusion that cwn-2/Wnt expression decreases with age. To resolve this issue, we collected total RNA from day 2 (young adult) and day 5 (mid-age adult) worms and performed qPCR for the cwn-2/Wnt and mom-2/Wnt (as it also shows a small discrepancy in tissue specific gene expression compared with previous observations). We found that total expression levels for both Wnt ligands significantly increased between day 2 and 5 of adulthood (Fig. 1G,H). It is possible that expression of cwn-2/Wnt decreases in muscle cells, as indicated by analysis using transcriptional fusion, but increases in the pharynx and SMD neurons, where this reporter is silenced. In addition, expression of mom-2/Wnt is highly increased in posterior intestinal cells, but remains unchanged or may be even decreased in other tissues, where we failed to detect any expression using our transcriptional reporters.
Taken together, these data show that that all five Wnt ligands continue to be expressed after development is completed, and that the Wnt signaling pathway might be active not only in development, but also in aging. However, the expression pattern of most Wnt ligands changes over time, which could imply that Wnt signaling function during aging could be somewhat different than in development.
Effect of Wnt signaling on elt-5/elt-3 GATA transcriptional circuit
The standard readout for the activity of the Wnt signaling pathway is the activation of reporter genes. However, very few targets of the Wnt signaling pathway in C. elegans development have been identified. We know even less about potential targets of Wnt signaling during worm aging. We consulted Wormbase (www.wormabase.org) for a list of potential targets that are regulated by the Wnt ligands and POP-1, the only TCF/LEF factor in C. elegans and a common component in both canonical and noncanonical Wnt signaling pathways. Out of 15 potential targets of the Wnt signaling pathway in C. elegans, we were able to detect expression for 11 genes starting at day 2 of adulthood (Table 1). We examined expression of these genes at different ages (day 2, 5, 7, and 11 of adulthood). We found that expression of end-3 (GATA transcription factor) and sdz-26 (SKN-1-dependent Zygotic transcript) decreases, and expression of elt-5 (GATA transcription factor) is increased with age but that expression of the other 8 genes is steady between young and old worms. Hence, expression changes in end-3, sdz-26, and elt-5 during aging can give us the first insight into Wnt signaling function in adult worms.
Table 1. Relative quantification by RT-qPCR of mRNA expression levels of 11 known targets of Wnt signaling pathway during aging (between days 2 and 11 of adulthood)
|Gene||Function||Age regulation||Fold change||P-value|
| elt-5/F55A8.1 a ||GATA transcription factor||Up-regulated||0.33||0.03|
| end-3/F58A3.5 ||GATA transcription factor||Down-regulated||0.6||0.0001|
|sdz-26/R06A4.6||SKN-1-dependent zygotic transcript||Down-regulated||3||0.0001|
| lin-39/C07H6.7 ||Homeodomain protein||No change||N/A||N/A|
| glr-1/C06E1.4 ||Glutamate receptor subunit||No change||N/A||N/A|
| end-1/F58A3.2 ||GATA transcription factor||No change||N/A||N/A|
|sdz-23/F58G4.4||Putatively secreted protein with a single EGF domain||No change||N/A||N/A|
|psa-3/F39D8.2||Axin family||No change||N/A||N/A|
|ceh-10/C09G12.8||GTPase orthologous to human RAC1||No change||N/A||N/A|
| ceh-22/F29F11.5 ||Homeodomain protein||No change||N/A||N/A|
It has previously been shown that expression of both end-3 and sdz-26 are negatively regulated by POP-1/TCF during endoderm specification (Maduro et al., 2005a; Shetty et al., 2005). Our results suggest that Wnt signaling continues to function and represses expression of end-3 and sdz-26 in adulthood. Unfortunately, expression of both end-3 and sdz-26 is quite low even at day 2 of adulthood, which makes them very poor markers to study the function of Wnt signaling activity all throughout aging process. Therefore, we used elt-5 GATA as a biomarker to access how Wnt signaling activity changes as the worms grow old.
Wnt signaling pathways are known to activate the expression of elt-5 GATA transcription factor during embryonic and larval development (Koh et al., 2002; Cassata et al., 2005). Previously, we have shown that the elt-3/elt-5/elt-6 GATA transcriptional circuit is, at least in part, responsible for driving the aging process in C. elegans (Budovskaya et al., 2008). We have shown that expression of repressor elt-5 GATA increases during aging. This leads to a decreased expression of the elt-3 GATA transcription factor in old age that causes down-regulation of many downstream age-regulated genes. In addition, close examination of the promoter region of the elt-5 gene have revealed that it contains seven potential TCF/LEF (CTTTGWW) binding sites (Maduro et al., 2005b). Therefore, we hypothesized that elt-5 GATA is one of the targets of the Wnt signaling pathway during aging. The Wnt signaling pathway continues to activate expression of elt-5 GATA, which leads to down-regulation of the elt-3 GATA transcription factor, which has detrimental effects on longevity.
To test this hypothesis, we compared the expression of elt-3 and elt-5 GATA in control and Wnt (RNAi) mutant backgrounds (Fig. 2). To avoid potentially detrimental effects of reduced Wnt ligands expression on development, we used RNAi to reduce egl-20, cwn-1, cwn-2, mom-2, and lin-44 expression starting on day 2 of adulthood and then observed the effects of down-regulation of the Wnt signaling pathway on the expression of the elt-3pro::GFP and elt-5pro::Cherry reporters on day 5 of adulthood. We found that egl-20(RNAi), cwn-2(RNAi), and lin-44(RNAi) treatments resulted in decreased expression of elt-5pro::Cherry and increased expression of elt-3pro::GFP in the head hypodermis (Fig. 2A,C), whereas RNAi against each of the five Wnt ligands resulted in decreased expression of elt-5pro::Cherry and subsequently increased expression of elt-3pro::GFP in the trunk hypodermis (Fig. 2B,D). Similar results were obtained from analyzing the effect of RNAi treatment against each of the five Wnt ligands on ELT-3 protein expression (Fig. S3). This suggests that even post-translational stability of ELT-3 GATA is not affected in Wnt signaling mutants. Overall, these results suggest that the elt-3 and elt-5 GATA transcriptional factors are still under control of the Wnt signaling pathway during aging and all five Wnts play a role in their regulation in a tissue specific manner. Whether this control of elt-5 expression is direct or indirect will require further investigation.
Figure 2. Effect of Wnt ligands down-regulation (RNAi) on elt-3 and elt-5 expression. (A) elt-3pro::GFP expression in the head hypodermis is increased in egl- 20(RNAi), cwn-2(RNAi), or lin-44(RNAi) animals. (B) elt-3pro::GFP expression in the trunk hypodermis is increased in egl-20(RNAi), cwn-1(RNAi), cwn- 2(RNAi), mom-2(RNAi), or lin-44(RNAi) animals. (C) elt-5pro:: mCherry expression in the head hypodermis is decreased in egl-20(RNAi), cwn- 2(RNAi), mom-2(RNAi), or lin-44(RNAi) animals. (D) elt-5pro:: mCherry expression in the trunk hypodermis is increased in egl-20(RNAi), cwn- 1(RNAi), cwn-2(RNAi), mom-2(RNAi), or lin-44(RNAi) animals. RNAi was induced starting at day two of adulthood by feeding worms bacteria expressing dsRNA. elt-3pro::GFP and elt-5pro::mCherry expression was measured starting at day 5. The y-axis denotes reporters expression (arbitrary units). Average expression and SE from 20 animals are shown.
Download figure to PowerPoint
Wnt signaling activity in aging
Measuring the expression of a few specific target genes could create a bias in our analysis. Therefore, instead of concentrating on one target or group of targets at a time, we wanted to analyze the global role of Wnt signaling activity during aging. In mammals, general Wnt signaling activity is also measured in vitro with the TOPFLASH reporter that consists of multiple TCF binding sites driving expression of luciferase (van de Wetering et al., 1997). For our analysis, we used a previously described homologs reporter for assessing C. elegans Wnt signaling activity (Green et al., 2008). The POPTOP (POP-1 and TCF Optimal Promoter) reporter consists of seven copies of POP-1/TCF binding sites, together with the pes-10 minimal promoter, driving expression of the red fluorescent protein, mCherry (Fig. 3A). This construct exhibits some background expression probably due to low activity of the minimal pes-10 promoter. Therefore, expression levels from the POPTOP reporter must be compared to expression from the control reporter, POPFOP (POP-1 Far from Optimal Promoter), which contains a mutated POP-1/TCF binding sites and cannot be activated by the Wnt signaling pathway.
Figure 3. Wnt signaling activity during aging. (A) Schematic representation of the constructs used to measure Wnt signaling activity in vivo. On top, POPTOP (POP-1 and TCF Optimal Promoter) construct which contains seven copies of POP-1/TCF binding sites upstream of a minimal promoter driving the expression of the fluorescent protein mCherry. Below, POPFOP (POP-1 and Far from Optimal Promoter) control construct is represented, which contains mutated TCF/POP-1 binding sites, used to measure background expression. (B) Number of copies of mCrerry transcript derived from the two constructs was measured by quantitative RT-PCR, using pD4H1-mCherry vector to build the standard curve. (C) POPTOP/POPFOP mCherry expression ratio.
Download figure to PowerPoint
Although POPTOP expression is visible at all larval stages during worm development, we were unfortunately not able to measure POPTOP expression in vivo in adult worms due to the weak signal of mCherry and high levels of gut autofluorescence. Therefore, we decided to examine levels of POPTOP expression in vitro using quantitative RT-PCR. We found that POPTOP expression increased about fivefold in the first 5 days of adulthood (Fig. 3B,C). The POPTOP expression decreased slightly by day 10 of adulthood, but still remained twofold higher then in young animals. These results indicate a general increase in Wnt signaling activity during aging in C. elegans.
Wnt influences Caenorhabditis elegans longevity
We hypothesize that increased levels of Wnt ligands in young adult worms lead to increased levels of Wnt signaling that promotes aging (developmental drift). Therefore, inactivation of either one or all Wnts by either RNAi or mutations should increase the lifespan of the worm. Another possibility is that high levels of Wnt expression are protective for the organism; for example, they activate genes that are responsible for damage repair and protection from pathogenic infections (damage accumulation). If that were the case, then inactivation of these Wnts would decrease the worm's lifespan. Our analysis of elt-3 and elt-5 GATA expression shows that all five Wnts function in the same manner: they increase expression of elt-5 GATA and consequently repress expression of elt-3, which at least in part causes aging in C. elegans (Budovskaya et al., 2008). This would indicate that mutations in any of the Wnt ligands should prolong the lifespan.
To test this hypothesis, we used RNAi treatment against each Wnt ligand to investigate their effect on longevity. We found that only mom-2(RNAi) treatment significantly increases the lifespan of wild-type, N2 animals by about 33%. Unexpectedly, lin-44(RNAi) treatments significantly shorten (by about 15%) the lifespan of wild-type, N2 animals (Table 2). This result implies that in C. elegans, the Wnt signaling pathway plays a different role in longevity than in development.
Table 2. Median lifespan changes in Wnt ligands mutants
| N2 || N2 ||19.5||21||20|
| Control * || N2 ||14.3||15|| |
| mom-2(RNAi)*|| N2 ||19.1d||20.2d|| |
| Control || N2 ||25||24||25|
| egl-20(RNAi)|| N2 ||28a||21||26a|
| cwn-1(RNAi)|| N2 ||24.5||25||24|
| cwn-2(RNAi)|| N2 ||24.1||23.3||26.2|
| lin-44(RNAi)|| N2 ||21.1b||20.5b||21.4b|
It has been previously shown by the smFISH technique that mom-2/Wnt is expressed in germ cell precursors (Z2 and Z3) during embryonic development and continues to be expressed in the germline throughout larval development (Harterink et al., 2011). Mutations in mom-2/Wnt influence germline function have been shown to induce sterility (Ceron et al., 2007). In addition, it has been well established that germline signals act through insulin/IGF-1 signaling cascade to promote aging in C. elegans (Hsin & Kenyon, 1999). We have not observed any mom-2/Wnt expression in the germline. However, we cannot exclude the possibility that mom-2/Wnt continues to be expressed in the germline at low levels and affects germline function and, as consequence, the worm's longevity. To assess this possibility, we analyzed the rate of reproduction and brood size of worms treated with mom-2(RNAi) at day 1 of adulthood compared to animals feed with control RNAi (Fig. S4). We observed no difference in germline function in these animals. mom-2(RNAi) did not cause any change in the general brood size or in the number of eggs that did not hatch. Therefore, we can conclude that mom-2/Wnt acts to promote aging independently of germline function.
The RNAi treatments against egl-20, cwn-1, and cwn-2 Wnts did not have any effect on longevity of N2 animals. It is possible that the RNAi treatment did not reduce levels of Wnt ligands enough to have a significant effect on longevity. To overcome this problem, we used genetic mutants that affect function of various Wnts.
There are many mutant alleles for different Wnt ligands. Unfortunately, many mutants cause severe developmental defects and are not suitable for longevity analysis as they produce sick individuals. For our purposes, we picked only mutations that do not compromise worm development (no delay in developmental timing) and produce normal, properly developed, individuals.
The mom-2/Wnt plays a very important role in the noncanonical, or ‘Wnt/β-catenin asymmetry’, signaling pathway that is required during early embryogenesis to insure proper endoderm induction (Eisenmann, 2005). Most of the mom-2 mutants are temperature sensitive and exhibit a strong embryonic lethal phenotype. For our analysis, we used the mom-2(or77) mutation. This mutation is a weak allele of mom-2/Wnt that affects a splice acceptor site at the last exon. The mom-2(or77) mutants develop normally (only 8% of embryos lack a gut at 15 °C) (Thorpe et al., 1997). After development, a small percentage (~10%) of the adult hermaphrodites retains eggs (bag of worms phenotype). We performed our lifespan analysis on the mom-2(or77) mutants and observed, as in the case of RNAi treatment, a significant increase in longevity of the mom-2(or77) mutant compared to wild-type controls (Fig. 4A).
Figure 4. Effect of mutations in various Wnt ligands on longevity. Weak mutation in mom-2/Wnt, mom-2(or77) (A), or deletion of cwn-2/Wnt, cwn- 2(ok895) (D), leads to significantly increased life span of the worms compared to N2 controls (P-value <0.0001). The mutation in lin-44/Wnt, lin-44(n1792) (B), or egl- 20/Wnt, elt-20(n585) (C) shows significantly decreased life span compared to N2 worms (P-value <0.0001). Deletion of cwn-1/Wnt, cwn-1(ok546) (E), does not affect longevity of mutant worms compared to wild type control.
Download figure to PowerPoint
Next, we tested the effect of the lin-44(n1792) mutant on longevity. In this mutant, lin-44 gene contains a nonsense mutation in the second exon that results in premature termination of transcription. These worms display reverse polarity and loss of asymmetry during B- and T-cell divisions (Rocheleau et al., 1999; Sawa et al., 2000). Despite the reversed polarity, lin-44(n1792) mutants are generally healthy and do not display severe morphological or physiological defects. The only side effect of lin-44(n1792) is an egg-laying defect. At 20 °C, about 50% of the adult hermaphrodites retain eggs. We performed lifespan analysis of the lin-44(n1792) mutation at 15 °C and observed that only 20% of the population was egg-laying defective. These individuals were censored and not counted in our longevity analysis. As in the case of RNAi treatment, the lin-44(n1792) mutation significantly decreases longevity of the worms compared to wild-type controls (Fig. 4B).
For egl-20/Wnt, we chose the egl-20(n585) mutation to analyze its effect on longevity. The egl-20(n585) mutation has a single-nucleotide substitution that causes replacement of the conserved cysteine with a serine (Harris et al., 1996; Maloof et al., 1999). This mutation was picked for our analysis because it has been shown that it significantly lowers gene and protein expression levels (Pan et al., 2008). However, egl-20(n585) produces several severe developmental defects at a low frequency, such as a worm carcass that gets filled with retained eggs that hatch inside (bag of worms phenotype) (Trent et al., 1983), a delay in egg laying (Desai et al., 1988), reversion of cell polarity, defects in cell migration (particularly in HSN motor neurons and Q neuroblast migration), and very mild defects in hypodermal cell development (Harris et al., 1996; Pan et al., 2008). It has been reported that these egl-20 mutants are slightly temperature sensitive; therefore, we performed our longevity assays at 15 °C to minimize these effects. Individuals that displayed severe developmental defects were censored and not counted. The elg-20(n585) mutation significantly decreased lifespan of the worms compared to wild-type controls (Fig. 4C).
The fourth Wnt ligand, cwn-2/Wnt, plays an important role in regulating nerve ring placement. Mutations that affect cwn-2/Wnt function are not embryonically lethal and do not cause any morphological or physiological defects. We used the cwn-2(ok895) mutant for our analysis. We found that the cwn-2(ok895) mutant lives significantly longer than wild-type worms, implying that cwn-2 function is detrimental for longevity (Fig. 4D).
To test the effect of cwn-1/Wnt on longevity, we used a deletion allele, cwn-1(ok546). This mutation has been created by the International C. elegans Gene Knockout Consortium. During embryogenesis, cwn-1/Wnt is required for cell fate specification in the C-lineage that gives rise to ectodermal and muscles tissue (Baugh et al., 2005). It has been shown that cwn-1(RNAi) treatment results in abnormal tail morphogenesis of C. elegans larvae. Although the cwn-1(ok546) mutant is a null allele of cwn-1, it displays very mild vulva-less phenotypes. About 40% of the animals lack a vulva at 20 °C and have the bag of worm phenotype. This defect is significantly reduced at 15 °C. Therefore, we performed our lifespan analysis of the cwn-1(ok546) mutation at 15 °C and observed that only 15–20% of the population was egg-laying defective. These individuals were censored and not counted in our longevity analysis. The cwn-1(ok546) mutant did not affect median lifespan of the worm, but caused significant extension of the maximal lifespan compared to wild-type controls (Fig. 4E).
These results indicate that even through all five Wnt ligands act in the same manner to activate expression of the elt-5 GATA transcription factor during aging, their general activity in adulthood is quite different. Activities of mom-2/Wnt and cwn-2/Wnt are detrimental for longevity, whereas activities of lin-44/Wnt and egl-20/Wnt are beneficial for long life.
Wnt signaling does not affect stress resistance during aging
There are two possibilities about how Wnt signaling regulates longevity. First, Wnt signaling continues to function after development and regulates the same set of genes during aging, which have detrimental effects on longevity (developmental drift theory). Indeed, we found evidence that in the case of elt-5 and elt-3 GATA transcription factors, Wnt signaling regulates their expression in both development and aging. Our results suggest that all five Wnts function to induce expression of elt-5 GATA, to repress expression of elt-3 and to drive aging.
The second possibility is that Wnt signaling changes function during aging, for example, from development to maintenance and stress response (damage accumulation theory). We hypothesized that Wnt signaling activity is increased as a result of stress and damage accumulation.
In test of this theory concerning Wnt function, we first analyzed whether the increased expression of Wnt ligands is influenced by oxidative damage, stress, impaired mitochondrial function, or pathogenic infections. To do this, we used a previously published microarray analysis of heat shock (McCarroll et al., 2004), hypoxia response (Shen et al., 2005), mitochondria dysfunction (Falk et al., 2008), and response to Pseudomonas aeruginosa infection (Shapira et al., 2006) to determine how these stresses affect Wnt ligands expression. Interestingly, we found that none of these stresses changes expression of Wnt ligands during aging suggesting an intrinsic mechanism of regulation.
Next, we tested the hypothesis that Wnt ligands switched their function from development to maintenance. If Wnt ligands function to promote aging (cwn-2 and mom-2) by repressing expression of stress response genes, then cwn-2(ok895) and mom-2(or77) mutants would have higher resistance to stress and cellular damage. On the other hand, if Wnt ligands have an anti-aging role (egl-20 and lin-44) by activating stress response genes and protection form oxidative stress and cellular damage, then egl-20(n585) and lin-44(n1792) mutants would be even more susceptible to stress and cellular damage than wild-type animals. We measured the survival of 1-day-old worms exposed to two stress conditions that most likely linked to aging, heat shock, and oxidative stress. Interestingly, none of the mutants showed significant resistance or sensitivity to heat shock (Fig. 5A–B) or oxidative stress (Fig. 5C–D) (see also Table S1). These data suggest that the mechanism by which Wnt signaling regulates aging might be different than just regulation of its ability to respond to stress and damage. Thus, a comprehensive analysis of targets of Wnt signaling during aging is crucial to this endeavor.
Figure 5. Wnt signaling does not affect worm's thermo-tolerance and resistance to oxidative stress during aging. None of the mutants show resistance or sensitivity to heat shock stress (A, B) or to oxidative stress (C, D). As a positive control daf-2(e1370) was included in both assays. The graphs depict the averages from three independent experiments.
Download figure to PowerPoint