|Home | About | Journals | Submit | Contact Us | Français|
Many ectotherms, including C. elegans, have shorter lifespans at high temperature than at low temperature. High temperature is generally thought to decrease the lifespan of ectotherms simply through its effects on chemical reaction rates. In this study, we questioned this view and asked whether the temperature-dependence of lifespan is subject to active regulation.
We show that thermosensory neurons play a regulatory role in the temperature dependence of lifespan. Inactivation of genes required for thermosensation, or laser ablation of thermosensory neurons, causes animals to have even shorter lifespans at warm temperature. We find that thermosensory mutations decrease expression of daf-9, a gene required for the synthesis of ligands that inhibit the DAF-12/nuclear hormone receptor. In addition, we show that the short lifespan of thermosensory mutants at warm temperature is completely suppressed by a daf-12(−) mutation.
Our data suggest that thermosensory neurons affect lifespan at warm temperature by changing the activity of a steroid signaling pathway which in turn affects longevity. We propose that this thermosensory system allows C. elegans to reduce the effect that warm temperature would otherwise have on processes that affect aging, something that warm-blooded animals do by controlling temperature itself.
Many ectotherms, including C. elegans, have shorter lifespans at high temperature than at low temperature . This change in lifespan is generally attributed simply to the effect of temperature on metabolic rates . Yet ectotherms are clearly capable of influencing the effect that temperature has on biological processes, since, for example, their biological clocks exhibit temperature compensation [3-8].
C. elegans perceives many environmental stimuli through sensory neurons [9, 10], and chemosensory neurons, which detect volatile and soluble compounds in the environment, are known to influence the lifespans of worms and flies [11-13]. C. elegans also contains thermosensory neurons, which allow the animal to migrate towards temperatures previously associated with food . The two bilateral AFD neurons are the main thermosensory neurons required for normal thermotaxis . Animals in which the AFD neurons have been ablated with a laser microbeam, as well as animals carrying a ttx-1 mutation, which specifically disrupts AFD structure , fail to migrate towards previously cultivated temperatures in thermal gradients [14, 15]. Because thermosensory neurons perceive and respond to temperature, they could, in principle, mediate a response to temperature that affects lifespan.
In this study, we examined whether the thermosensory neurons of C. elegans influence the temperature dependence of lifespan. Surprisingly, we found that genetic or laser ablation of the AFD neurons led to even shorter lifespan at high temperature, suggesting that thermosensory neurons counteract the life-shortening effect that temperature would otherwise have on lifespan. We then tested whether any known signaling pathways that affect aging are involved in the effects of the thermosensory neurons on lifespan. One of the pathways we tested was the steroid signaling pathway comprising DAF-9 and DAF-12. daf-9 encodes a cytochrome P450 enzyme (CYP) that is required for the synthesis of sterol hormones, called dafachronic acids, which act as ligands for the nuclear hormone receptor (NHR) DAF-12 [17-23]. We found that the AFD thermosensory neurons influenced lifespan at high temperature through the DAF-9/DAF-12 steroid signaling pathway. Together, our findings suggest that C. elegans actively transmits signals from thermosensory neurons to this steroid signaling pathway to regulate lifespan at high temperature.
To test the idea that temperature sensation may be involved in the longevity response to temperature, we analyzed the lifespans of thermosensory mutants. We found that animals in which the thermosensory AFD neurons had been laser-ablated, as well as animals carrying ttx-1 mutations , lived up to 25% shorter than normal at 25°C, a warm temperature (Figure 1, A and B). We found that neither AFD ablation nor ttx-1 mutation influenced lifespan at 15°C (Figure 1, C and D). This result is consistent with the finding that defects in AFD neurons do not abolish thermosensation towards cool temperature [15, 24]. We also examined the lifespan of ttx-1 mutants at 20°C, an intermediate temperature, and found that it, too, was normal (Figure S1A). Thus the influence of these thermosensory neurons on lifespan was temperature dependent. Together these findings suggest that the effects of warm temperature on lifespan are not produced purely as an unregulated response of chemical reaction kinetics to increased thermal energy. Instead, lifespan at warm temperature appears to be influenced by the sensory perception of temperature, which, surprisingly, acts to counter the effects that increased temperature would be predicted to have on chemical reaction rates.
The process of thermosensation requires a cyclic-nucleotide gated calcium channel whose subunits are encoded by the tax-2 and tax-4 genes, and animals carrying mutations in these genes are defective in thermotaxis [14, 15, 25, 26]. We found that all but one of five tax-2 and tax-4 single mutants that we tested were short-lived (−12 to −43%) at 25°C, as were tax-2; tax-4 double mutants (Figure 1, E to G, and Figure S2). Significantly, the short lifespan of ttx-1 mutants was not further decreased by tax-2 mutation (Figure 1H), suggesting that tax-2 and the AFD neurons act in the same pathway.
AIY interneurons form a neural circuit with the AFD neurons for normal thermosensation and they are also required for normal thermotaxis . To test whether this neural circuit might be involved in the temperature-dependence of the lifespan, we examined ttx-3 mutants, which have defects in the structure and function of AIY neurons. Similar to ttx-1 mutants, the ttx-3 mutants lived shorter than wild type at 25°C but not at 15°C (Figure 1, I and J). Taken together, these findings suggest that temperature cues are processed by a mechanism involving the AFD/AIY neural circuit not only to mediate thermotaxis, but also to influence lifespan.
Many mutations that disrupt chemotaxis increase the lifespan of C. elegans at 20°C [11, 12]. To test the possibility that the chemosensory system influences lifespan in a temperature-dependent fashion, we measured the lifespan of several chemosensory mutants at 25°C. Mutants defective in osm-3, which encodes an anterograde motor kinesin , and osm-5, which encodes an intraflagellar-transport protein , have broad chemosensory defects [16, 29, 30] and both of these mutants are long-lived at 20°C . We found that osm-3 and osm-5 mutations increased lifespan by a similar percentage, relative to wild type, at warm as well as cool temperature (Figure 2, A and B, and Figure S1H). Thus the chemosensory system affects lifespan independently of temperature. The AFD neurons and ttx-1 mutations appear to have little or no effect on chemosensation [14, 16], whereas osm-3 and osm-5 mutations do not appear to affect thermotaxis . Thus, these two sensory systems are likely to act independently of one another to influence longevity.
Some mutations that perturb neural function affect both thermotaxis and chemotaxis [14, 24-26, 31, 32]. Among these are the tax-2 and tax-4 cation-channel mutations. Like other chemosensory mutations, some of the tax-2 and tax-4 mutations that we tested extend lifespan at lower temperature (reference  and Figure S3), so it was interesting that tax-2 and tax-4 mutations produced short rather than long lifespans at 25°C (Figure 1, E to G, and Figure S2). This finding suggested that at warm temperature, the processes triggered by loss of thermosensation at least partially override those triggered by loss of chemosensation. To test this idea directly, we measured the lifespan of double mutants carrying chemosensory osm-3 or osm-5 mutations as well as the thermosensory ttx-1 mutation. As predicted, the ttx-1 mutation decreased the lifespans of osm-3 and osm-5 mutants at 25°C (Figure 2, C and D).
Next, we asked whether thermosensory mutations produce similar changes in the rates of all temperature-dependent processes at 25°C. We found that the rates of pharyngeal pumping (feeding) and reproductive timing of tax-2 mutants were normal at 25°C (Figure S4, A to D). The time for growth to adulthood was slightly slower (not faster) in these animals, and this was the case at all temperatures (Figure S4E). Together these findings suggest that the targets of the thermosensory system will be genes that affect the rate of aging but not the rates of all temperature-dependent processes.
How do the AFD neurons influence lifespan at warm temperature? To address this question, we asked whether any genes already known to influence the rate of aging in C. elegans might be involved. We reasoned that if the thermosensory system affected lifespan by changing the activity of a known longevity gene, then it would be unable to influence lifespan in that gene’s absence. Many proteins that influence lifespan in C. elegans, including the heat-shock factor HSF-1, the deacetylase SIR-2.1, the c-Jun N-terminal kinase JNK-1, the energy sensor AMP kinase and the temporal regulator LIN-14, require the lifespan-extending transcription factor DAF-16/FOXO [33-38] to exert their effects [39-43]. We found that thermosensory mutations were able to further shorten the lifespans of animals carrying null mutations in daf-16/FOXO (Figure 3, A and B), suggesting that thermosensory genes and cells do not act through DAF-16/FOXO. In addition, thermosensory mutations shortened the long lifespan of dietary-restricted eat-2 mutants  and respiration-defective isp-1 mutants , which extend lifespan independently of daf-16, at 25°C (Figure 3, C and D), indicating that thermosensory regulation of lifespan is intact in these mutants.
Two additional genes that affect lifespan are daf-9 and daf-12, which encode components of a steroid signaling pathway [17-23, 46, 47]. daf-9 encodes a cytochrome P450 that synthesizes sterol ligands for DAF-12, a nuclear hormone receptor . At high temperature, the reduction-of-function daf-9(rh50) mutant is short lived . Unlike any of the other mutants we tested, we found that the lifespan of this mutant was not further shortened by ttx-1 or tax-2 mutations (Figure 4, A and B). This finding, along with the healthy appearance of ttx-1 and tax-2 mutant animals (data not shown), argued against the possibility that ttx-1 or tax-2 mutations shorten lifespan through a non-specific effect on their health. In contrast, this finding suggested that thermosensory neurons affect lifespan by perturbing a regulatory pathway involving steroid hormones.
One way that AFD could influence daf-9-dependent hormone signaling would be to control daf-9 gene expression. To test this idea, we examined the expression of a DAF-9::GFP translational fusion protein, expressed from the endogenous daf-9 promoter, in transgenic animals. DAF-9::GFP is expressed in the XXX neurosecretory cells, the hypodermis and the spermatheca [17, 18, 20, 21, 48]. We found that DAF-9::GFP levels were decreased in tax-2 mutants (Figure 4, C to E). Likewise, we observed a sharp decrease in daf-9 mRNA levels in ttx-1 and tax-2; tax-4 mutants using quantitative RT-PCR (Figure 4, F and G). Thus, signals from thermosensory neurons regulate daf-9 expression in other tissues.
To test the significance of this change in daf-9 expression, we asked whether expression of daf-9 from a heterologous promoter could rescue the reduced-longevity phenotype of ttx-1 or tax-2 mutants (Figure 4, H and I, and Figure S5). Expressing daf-9 from the sdf-9 promoter , which is active only in the XXX cells , suppressed the short lifespan of ttx-1 and tax-2 mutant animals, but had no effect on wild type (Figure 4, H and I). Likewise, expressing daf-9 from a hypodermal promoter in tax-2 mutants could suppress the shortened 25°C lifespan (Figure S5B). These findings argue that loss of thermosensory function shortens lifespan at 25°C by reducing daf-9 expression. We note that expression of daf-9::gfp under the control of daf-9’s own promoter (Pdaf-9::daf-9) failed to rescue the short lifespan of tax-2(p671) animals at 25°C (Figure S5A). However, as described above, this transgene was greatly downregulated in tax-2 mutants (Figure 4, C to E), suggesting that it could not produce rescuing levels of daf-9 mRNA in the absence of functional thermosensory neurons.
DAF-9/CYP carries out a step in the synthesis of sterol ligands for DAF-12/NHR. These ligands, called dafachronic acids , stimulate some aspects of DAF-12/NHR activity and inhibit others [17, 18, 20-22, 47, 49]. We found that the short 25°C lifespan of daf-9 mutants was completely suppressed by the daf-12 null mutation rh61rh411  (Figure 5A), whereas the lifespan of wild type was unaffected. This result implies that daf-9/CYP influences lifespan at 25°C by regulating the activity of DAF-12/NHR. Consistent with this idea, daf-12 mutation completely suppressed the short 25°C-lifespans of ttx-1 and tax-2 mutants (Figure 5, B and C). Together, these findings indicate that thermosensory neurons influence lifespan by controlling steroid signalling. Specifically, in wild-type animals grown at 25°C, a signal produced by thermosensory neurons stimulates daf-9 expression, which in turn affects lifespan by raising the level of a sterol ligand that inhibits DAF-12/NHR.
During thermotaxis, AFD is thought to respond to an increase in temperature, rather than to warm temperature itself. Increases in temperature trigger visible changes in the level of calcium within the AFD neurons, whereas constant incubation at warm temperature does not [51, 52]. Thus it is interesting that AFD can regulate daf-9 expression when the animals are cultured continuously at 25°C. This finding raises the possibility that AFD has the ability to distinguish warm from cool temperature even if the temperature is not changing. To investigate this further, we asked whether loss of AFD function (through ttx-1 mutation) would trigger changes in daf-9 expression at low as well as high temperature. We found that it did not (Figure 6A). Together these findings suggest that AFD has the capacity to distinguish warm from cool temperature, and that it does so in its regulation of lifespan.
In addition to allowing worms to migrate up a thermal gradient, the AFD neuron has recently been shown to mediate the animal’s classical heat-shock response . Heat shock, and overexpression of the heat-shock factor HSF-1, are known to increase lifespan [39, 42, 54, 55], so another possibility we considered was that AFD neurons mediate their effects on lifespan at 25°C, at least in part, by activating a heat-shock response. This idea seemed unlikely because, as with thermotaxis, AFD activates the heat-shock pathway in response to rising temperature, but not in response to constant temperature . We tested this idea directly, and found, using quantitative RT-PCR, that the loss of AFD function (through ttx-1 mutation) did not influence the level of expression of either of two known heat shock protein genes in animals cultured continuously at either warm or cool temperature (Figure 6, B and C). In addition, the lifespan extension caused by overexpressing HSF-1 requires DAF-16/FOXO , yet, as described above, loss of AFD function influences lifespan in a daf-16-independent fashion (Figure 3, A and B). Together these findings argue against the idea that AFD influences lifespan in animals grown continuously at warm temperature by affecting the animal’s heat shock response.
Reduction of HSF-1 activity is known to cause an accelerated aging (progeric) tissue phenotype in animals grown continuously at 20°C , and HSF-1 has been shown to act in multiple tissues to influence lifespan . Thus, even without the involvement of AFD, it was still possible that HSF-1 might exert a temperature-dependent effect on lifespan. To test this, we measured the lifespan of hsf-1(sy441, RNAi) animals grown continuously (without heat shock) at different temperatures. We found that these hsf-1 mutant animals lived much shorter than wild type at warm temperature (22.5°C) (Figure 7A). In contrast, the lifespan-shortening effect of this hsf-1 reduction was minor at low temperature (15°C) and intermediate at an intermediate temperature (20°C) (Figure 7, B and C). We observed a similar pattern with hsf-1(RNAi) animals (data not shown). Thus, it seems likely that both HSF-1 and AFD neurons independently extend lifespan at high temperature.
As temperature is decreased, ectotherms live longer. The thermosensory system of C. elegans also senses cool temperature, as it causes animals to migrate towards cool temperatures that were previously associated with food [9, 10]. Does the thermosensory system also influence the longevity response to cool temperature? This question is difficult to address: ablating the AFD neurons does not completely abolish thermotaxis to low temperature  and we observed no effect on lifespan at low temperature (Figure 1B). Mutations and cells that affect thermotaxis at cool temperature do exist [14, 24-26, 31, 32], but all of these affect chemotaxis as well [9, 10]. Until genes that affect thermotaxis but not chemotaxis at low temperature are identified, it will be not possible to dissociate the effects of these two sensory systems from one another.
In summary, the major finding of this study is that C. elegans’ longevity response to warm temperature is not a purely passive consequence of increased temperature, but instead is subject to regulation by thermosensory neurons. Our findings indicate that thermosensory neurons influence lifespan by controlling the activity of a steroid-signalling pathway. Specifically, they suggest that at high temperature, AFD produces a signal that stimulates expression of the daf-9/CYP gene, which in turn affects lifespan by changing the level of a ligand that influences the activity of DAF12/NHR.
Through its effects on steroid signalling, the C. elegans neuroendocrine system limits the effect that warm temperature would otherwise have on the lifespan of C. elegans. Warm-blooded animals protect themselves from the effects of temperature in a different way, by controlling temperature itself. Interestingly, in mice, transgenes that increase the temperature of the hypothalamus (the region of the brain that controls temperature homeostasis) decrease body temperature and extend lifespan . Perhaps aspects of the neuroendocrine response to high temperature in mammals have a common evolutionary origin with the lifespan-extending neuroendocrine response to warm temperature that we observe in this ectotherm.
The following strains were analyzed in this study.
|N2||Wild type (WT)|
|CF2424||ttx-1(p767) V||PR767 outcrossed 4 times to Kenyon-|
|CF2360||tax-2(p671) I||PR671 outcrossed 4 times to Kenyon-|
lab N2 (M. Gaglia)
|CF2363||tax-2(p691) I||PR691 outcrossed 4 times to Kenyon-|
|FK100||tax-2(ks10) I||Outcrossed 3 times to N2 by I. Mori lab|
|FK101||tax-2(ks15) I||Outcrossed 3 times to N2 by I. Mori lab|
|CF2380||tax-4(p678) III||PR678 outcrossed 4 times to Kenyon-|
lab N2 (M. Gaglia)
|CF2426||ttx-3(ks5) X||FK134 outcrossed 4 times to Kenyon-|
|CF2552||osm-3(p802) IV||PR802 outcrossed 4 times to Kenyon-|
|CF2553||osm-5(p813) X||PR813 outcrossed 3 times to Kenyon-|
|CF1042||daf-16(mu86) I||Outcrossed 11 times to Kenyon-lab N2|
|CF2479||daf-12(rh61rh411) X||AA86 outcrossed 3 times to Kenyon-|
lab N2 (S. Korenblit)
|CF2531||daf-9(rh50) X||AA111 outcrossed 4 times to Kenyon-|
lab N2 (T. Yamawaki)
|CF1908||eat-2(ad1116) II||DA1116 outcrossed 3 times to Kenyon-|
lab N2 (L. Mitic)
|CF2172||isp-1(qm150) IV||MQ887 outcrossed 3 times to Kenyon-|
lab N2 (J. Pinkston)
|CF2495||hsf-1(sy441) I||PS3551 outcrossed 3 times to Kenyon-|
|The transgene is termed Pdaf-9::daf-9|
in this manuscript for consistency. The
same abbreviation (eg. Pdpy-7::daf-9)
was applied to all the other daf-9
transgenic animals described below.
The daf-9 transgenic animals were gifts
from G. Ruvkun and were crossed to
other strains in our laboratory.
tax-2(p671); daf-9(e1406) dpy-
9 cDNA::gfp; mec-7::gfp]
9 cDNA::gfp; mec-7::gfp]
12p::daf-9 cDNA::gfp; mec-7::gfp]
9 cDNA::gfp; mec-7::gfp]
All lifespans were measured as described previously, starting with Day 1 adults . Strains were cultured at a given temperature (15°C, 20°C, 22.5°C or 25°C) under standard growth conditions for at least two generations before lifespan analysis. The chemical 2′fluoro-5′deoxyuridine (FUDR, Sigma, 75 μM) was added to pre-fertile young-adult animals to prevent their progeny from developing unless stated otherwise. Animals that ruptured, bagged (i.e., exhibited internal progeny hatching), or crawled off the plates were censored but included in the lifespan analysis as censored worms as described previously . STATA version 10.0 software (StataCorp, USA) was used for statistical analysis and P values were calculated using the Log-rank (Mantel-Cox) method.
All pictures of transgenic animals were captured using a Retiga EXi Fast1394 CCD digital camera (QImaging, Burnaby, BC, Canada) attached to a Zeiss Axioplan 2 compound microscope (Zeiss Corporation, Germany). Quantification of the GFP intensity (Fig. 4E) was ‘determined using the “Show Measurements” function in the OpenLab version 4.02 program (Improvision, UK). The mean fluorescence intensity was normalized by subtracting mean fluorescence intensity of pictures of non-transgenic control worms taken with the same exposure. P values were calculated using the unpaired Student’s t-test (two-tailed).
RNA extraction, purification and reverse transcription of RNA were performed as described . Quantitative RT-PCR was carried out in a 7300 Real Time PCR System (Applied Biosystems) and analyzed using the Ct method (Applied Biosystems Prism 7700 Users Bulletin No. 2 http://docs.appliedbiosystems.com/pebiodocs/04303859.pdf). mRNA levels of the actin gene, act-1, was used for normalization. Average of at least 2 technical repeats were used for each biological data point. Primer sequences are available on request.
AFD neurons or the gonad precursor (Z1 and Z4) cells in synchronized, newly-hatched L1 animals were ablated with a VSL337 nitrogen-pumped dye laser (Laser Sciences, Inc.) as described [12, 47]. oyIs18 animals, which express the AFD-specific transgene, gcy-8::gfp , were used to identify AFD neurons.
L1 larvae were synchronized by overnight hatching and starvation of bleached eggs. Approximately 100 L1 larvae were placed on each OP50-seeded NG plate and incubated at 15°C, 20°C or 25°C. After most of the animals had developed to L3 or L4 larvae, they were examined every 2 hours and their developmental stage was recorded. The time from L1 to adult was defined as the time when 50% of the worms had become pre-fertile adults (the ‘M’-shaped vulva stage).
Pumping rates were measured under a dissecting microscope as described .
Single late-L4 stage worms were placed on individual plates for each strain at 15°C, 20°C or 25°C (10 worms per condition). The animals were transferred to new OP50-seeded plates every 12 hours for 8 days while they produced progeny. Worms that bagged or ruptured were censored. All the plates containing progeny were incubated for 2 days after removing the parental worms, and then held at 4°C until the number of worms that developed was counted.
We thank Dr. G. Ruvkun and the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources, for providing strains, and all Kenyon lab members for sharing outcrossed strains and for helpful comments on the experiments, data analysis and manuscript. We also thank Dr. D. Shin for valuable comments on the manuscript. S.J.L. was an Ellison Medical Foundation fellow of the Life Sciences Research Foundation and a postdoctoral fellow of the American Heart Association, Western States Affiliate. This work was supported by NIH grant #AG020932 to C.K., who is the director of the UCSF Hillblom Center for the Biology of Aging, an American Cancer Society Professor, and a co-founder and director of the biotechnology company Elixir Pharmaceuticals.