H3K27 modifiers regulate lifespan in C. elegans in a context-dependent manner

Abstract

Background: Evidence of global heterochromatin decay and aberrant gene expression in models of physiological and premature ageing have long supported the "heterochromatin loss theory of ageing", which proposes that ageing is aetiologically linked to, and accompanied by, a progressive, generalised loss of repressive epigenetic signatures. However, the remarkable plasticity of chromatin conformation suggests that the re-establishment of such marks could potentially revert the transcriptomic architecture of animal cells to a "younger" state, promoting longevity and healthspan. To expand our understanding of the ageing process and its connection to chromatin biology, we screened an RNAi library of chromatin-associated factors for increased longevity phenotypes.

Results: We identified the lysine demethylases jmjd-3.2 and utx-1, as well as the lysine methyltransferase mes-2 as regulators of both lifespan and healthspan in C. elegans. Strikingly, we found that both overexpression and loss of function of jmjd-3.2 and utx-1 are all associated with enhanced longevity. Furthermore, we showed that the catalytic activity of UTX-1, but not JMJD-3.2, is critical for lifespan extension in the context of overexpression. In attempting to reconcile the improved longevity associated with both loss and gain of function of utx-1, we investigated the alternative lifespan pathways and tissue specificity of longevity outcomes. We demonstrated that lifespan extension caused by loss of utx-1 function is daf-16 dependent, while overexpression effects are partially independent of daf-16. In addition, lifespan extension was observed when utx-1 was knocked down or overexpressed in neurons and intestine, whereas in the epidermis, only knockdown of utx-1 conferred improved longevity.

Conclusions: We show that the regulation of longevity by chromatin modifiers can be the result of the interaction between distinct factors, such as the level and tissue of expression. Overall, we suggest that the heterochromatin loss model of ageing may be too simplistic an explanation of organismal ageing when molecular and tissue-specific effects are taken into account.

Keywords: Ageing; C. elegans; Chromatin; H3K27; Healthspan; Histone demethylase; Histone methyltransferase; Lifespan.

Fig. 1

Lifespan and healthspan regulation by H3K27 methylases and demethylases. a The KDM6 family of H3K27 demethylases comprised utx-1, jmjd-3.1, jmjd-3.2 and jmjd-3.3 in C. elegans (TPR, tetratricopeptide repeat; JmjC, Jumonji C domain). Animals containing mutant alleles of mes-2, jmjd-3.2 or utx-1 were analysed for both lifespan (b, c, e, f) and healthspan (d, g). b Both mes-2(tm5007) and mes-2(ok2480) mutants live significantly longer than the N2 control strain (***p = 0.0003 and ****p < 0.0001, respectively). c Worms that are either homozygous or heterozygous for the mes-2(bn11) allele have extended lifespan when compared to either N2 (****p < 0.0001 and ***p = 0.0005, respectively) or the balancer-containing (SP127) control (****p < 0.0001). d The three mes-2 mutants display an increased thrashing rate, indicative of better health, at day 12 and day 15 of adulthood compared to the N2 control strain (p < 0.005 in all cases). e, f: jmjd-3.2(tm3121) mutant worms (e) and utx-1(tm3118) heterozygotes (f) display extended lifespan compared to the N2 control (**p = 0.005 and ****p < 0.0001, respectively). g jmjd-3.2(tm3121) mutants and utx-1(tm3118)/+ heterozygotes display increased thrashing rates, indicating better health, at day 12 and day 15 of adulthood when compared to N2 controls (p < 0.001 in each case) (see Additional file 3: Table S2 for the full statistical analysis of lifespan data, including repeats). For thrashing assays, n ≥ 10 worms, measured at least 3 times for each strain. Box and whisker plot represents the first and third quartile and minima and maxima of the data points

Fig. 2

Pathway analysis of mes-2, jmjd-3.2 and utx-1 mediated lifespan extension. mes-2, jmjd-3.2 and utx-1 mutant alleles were subjected to daf-16 RNAi and lifespan assays performed in order to investigate possible lifespan-extending pathways. a–c RNAi depletion of daf-16 in mes-2(tm5007) (a), jmjd-3.2(tm3121) (b) and utx-1(tm3118)/+ (c) backgrounds completely suppresses lifespan extension. d Combinatorial knockdown of jmjd-3.2 (using mutant allele tm3121) and utx-1 (using post-embryonic RNAi) results in increased life extension compared to either the jmjd-3.2 allele (**p = 0.007) or utx-1 RNAi alone (*p = 0.03). Controls are shared between experiments a, b and c (as the experiments were performed as a large set) although the graphs are separated for clarity. EV, empty vector control (i.e. worms fed HT115 bacteria transformed with L4440 RNAi vector lacking a genomic insert) (see Additional file 7: Table S4 for the full statistical analysis of data, including repeats)

Fig. 3

Overexpression of jmjd-3.2 or utx-1 causes lifespan extension. a, b Levels of jmjd-3.2 (a) and utx-1 mRNA (b) were assessed by qRT-PCR, showing significant upregulation in the jmjd-3.2 and utx-1 transgenic lines, respectively, when either wild-type or demethylase dead (DD) constructs were used. Error bars represent the SEM for each data point. Lifespan assays were performed on transgenic lines overexpressing jmjd-3.2 or utx-1 driven by their respective endogenous promoter. c, d overexpression of either jmjd-3.2 (c) or utx-1 (d) in a wild-type background results in lifespan extension (****p < 0.0001). Overexpression of demethylase dead jmjd-3.2 (jmjd-3.2DD) also results in lifespan extension (****p < 0.0001), whereas overexpression of demethylase dead utx-1 (utx-1DD) has no effect. e Overexpression of either wild-type jmjd-3.2 or demethylase dead jmjd-3.2 (jmjd-3.2DD) does not further extend the lifespan of jmjd-3.2(tm3121) mutants. f Overexpression of wild-type utx-1 in a utx-1(tm3118) mutant background promotes lifespan extension beyond that of the long-lived utx-1(tm3118/+) mutant strain (****p < 0.0001), whereas overexpressing the demethylase dead version has no effect (see Additional file 9: Table S5 for full statistical analysis of lifespan data, including repeats). (OE, overexpression; DD, demethylase dead). N2 control strains contained the same coinjection marker (rol-6⁺) as the jmjd-3.2 and utx-1 transgenic lines

Fig. 4

Overexpression of utx-1 enhances resistance to oxidative, UV and heat stressors. Animals were either exposed to 10 mM paraquat from L4 onwards to induce oxidative stress (a), given a 1000-Jm⁻² dose of UVC 3 days post-L4 to induce UV stress (b), or exposed to 35 °C heat for 6 h from 5 days post-L4 to induce heat stress (c), and scored for survival. In all cases, animals overexpressing utx-1 (red survival curve) were more resistant to stress compared to control worms containing the same co-injection marker (rol-6⁺) (blue survival curve) (all ****p < 0.0001) (see Additional file 10: Table S6 for the full statistical analysis of lifespan data, including repeats)

Fig. 5

Lifespan extension due to utx-1 overexpression is independent of insulin signalling but at least partially dependent on daf-16. a Lifespan assays were performed on transgenic animals overexpressing utx-1 in a utx-1(tm3118) mutant background combined with daf-16 RNAi. daf-16 depletion significantly shortened the lifespan of wild-type and transgenic animals compared to EV controls (****p < 0.0001), although utx-1 overexpressing animals subjected to daf-16 RNAi did display a moderate lifespan extension compared to wild-type daf-16(RNAi) animals (****p < 0.0001). b Lifespan assays performed on daf-2(e1370) mutants subjected to utx-1 RNAi resulted in no significant change in survival compared with daf-2(e1370) EV controls, consistent with findings from [14]. The lifespan increase upon utx-1 RNAi in a wild-type background confirms that the utx-1 RNAi was working effectively. c Lifespan assays performed on daf-2(e1370) mutants compared to daf-2(e1370) mutants overexpressing utx-1 showed a significant lifespan extension compared with daf-2(e1370) mutants alone (**p = 0.0023) or utx-1 overexpression alone. EV, empty vector control (i.e. worms fed HT115 bacteria transformed with L4440 RNAi vector lacking a genomic insert); OE, overexpression (see Additional file 11: Table S7 for the full statistical analysis of lifespan data, including repeats)

Fig. 6

Gain or loss of utx-1 expression in specific tissues induces lifespan extension. a Expression of GFP in transgenic lines was monitored to confirm tissue-specific expression. Top left panel: muscle cell expression of utx-1::gfp driven by the myo-3 promoter (white arrows); top right panel: epidermal expression driven by the dpy-7 promoter (white arrows); bottom left panel: neuronal expression driven by the rab-3 promoter (white arrows); bottom right panel: intestinal expression driven by the vha-6 promoter (white arrows). White dashed line is the outline of the worm in each case. b Lifespan assays were performed on transgenic animals overexpressing utx-1 in specific tissues. Lifespan extension was observed when utx-1 was overexpressed in neuronal and intestinal cells (****p < 0.0001 in both cases), but not in epidermal or muscle cells. c–f Lifespan assays were performed in animals subjected to tissue-specific knockdown of utx-1 by RNAi. In the case of muscle, epidermal and intestinal knockdown, the RNAi insensitive rde-1(ne219) mutant was used, rescued by muscle- (c), epidermal- (d) or intestinal-driven (f) rde-1, and animals subjected to utx-1 RNAi post-embryonically. For neuronal knockdown (e), sid-1 was expressed pan-neuronally in the RNAi-insensitive sid-1 mutant background, and utx-1 RNAi performed post-embryonically. Marked lifespan extension was seen in epidermal and neuronal knockdown (****p < 0.0001 in each case), slight lifespan extension was seen in intestinal knockdown (**p = 0.009) and no lifespan extension was associated with muscle-specific knockdown. EV, empty vector control (i.e. worms fed HT115 bacteria transformed with L4440 RNAi vector lacking a genomic insert) (see Additional file 13: Table S8 for the full statistical analysis of lifespan data, including repeats)