Metabolomic signature associated with reproduction-regulated aging in Caenorhabditis elegans

Introduction

Aging is an inevitable and complex part of life and has been a fascinating phenomenon for several thousands of years. Reproductive capacity is closely related to aging, and previous studies have demonstrated the apparent trade-off between reproduction and aging. Consistent with this idea, longevity can be achieved by sacrificing fertile potential in many species, including Caenorhabditis elegans (C. elegans) [1,2], drosophila melanogaster [3], and humans [4], suggesting that the relationship between reproduction and aging is conserved.

Reproduction is an energetically costly process, and it has been reported that reduced reproduction is associated with elevated fat storage and prolonged lifespans in multiple organisms [5,6]. These findings seem to indicate that the longevity of a species is a direct result how it distributes its resources between reproduction and survival. In the nematode C. elegans, laser ablation of germline stem cells (GSCs) precursors or genetic ablation of GLP-1/Notch signaling causes GSC proliferation to be inhibited, which can activate signals in somatic tissues that significantly lengthen lifespan [1] and alter lipid metabolism [7], known as the Glp (germ-line proliferation defective) phenotype. Moreover, researchers have begun to reveal the molecular mechanisms by which signals from the reproductive system influence lipid metabolism and lifespan [8,9].

In C. elegans, during germline quiescence, steroidal signaling (DA/DAF-12), microRNA mir-7, and ankyrin repeat-containing protein KRI-1 regulate and prompt the nuclear localization and activation of DAF-16 [1,10–12]. In addition, DAF-16 nuclear activity is regulated by assuming complexes comprised of FTT-2, PHI-62, and TCER-1. These proteins collaborate in transcriptional complexes to regulate germline signals to extend lifespan [13–16]. In addition, in germline-less worms, SKN-1 was also regulated in parallel with DAF-12 and DAF-16 [17]. Additionally, NHR-80 transcriptional complexes are also regulated by inputs from the germline perhaps in a manner independent of DAF-16 but partly dependent on DAF-12 [18]. Germline loss also stimulated TOR downregulation, which in turn, upregulating PHA-4 and autophagy processes [19]. Altogether, during germline defect, the different transcription factors and nuclear receptors function in a complex and sophisticated network, which initiates a cascade of dramatic events, including autophagy, fatty acid lipolysis, stress resistance and other processes, to enhance homeostasis and increase lifespan.

Compared to the detailed studies on germline signals regulating lipid metabolism and longevity in C. elegans, it remains unknown whether or how other small molecules, including amino acids and sugar, play roles in germline-mediated longevity.

Recent studies have demonstrated that the use of high-throughput ‘-omic’ approaches can increase our understanding of the global variation that accompanies aging and anti-aging [20]. Metabolomics, the untargeted profiling of metabolites, is the real endpoint of physiological regulatory processes, which aims to identify perturbations in biochemical networks and to further extend our understanding of the molecular mechanisms underlying specific biological function in complex organisms [21–23]. By monitoring many related and unrelated small molecules, the metabolic profile can provide a snapshot of highly complex metabolic exchanges, which integrate information from multiple levels of organization, including the genome, the transcriptome, the proteome, the environment and their interactions [24]. Furthermore, metabolomics have been used in aging studies of C. elegans. For example, Fuchs and co-workers found the existence of a common metabolic signature for extending lifespan by investigating metabolic changes of the long-lived worm mutants using high-resolution ¹H NMR spectroscopy [25]. Another study that utilized a high-resolution ¹H NMR metabolomics approach found that, compared with wild type, higher concentrations of branched-chain amino acids (BCAA) were detected in long-lived mutants [26]. Additionally, a previous study also reported metabolic variation associated with aging and specially found that low phosphocholine (PCho) was correlated with high life expectancy based on a ¹H high-resolution magic-angle spinning (HR-MAS) nuclear magnetic resonance (NMR) analysis of intact worms [27]. Other researchers have used a GC-MS metabolomics approach to suggest that a-ketoacids and a-hydroxyacids were related to extend lifespans of long-lived mitochondrial mutants but not of other long-lived mutants or short-lived mutants [28]. Furthermore, questions have been raised as to whether we can investigate the relationship of reproduction and longevity by using a comprehensive metabolomics approach.

In this work, we used C. elegans hermaphrodites as a model system to investigate how metabolic pathways changed during aging and how germline signals regulate metabolism to attenuate aging. Here, we assessed the metabolic phenotype of whole C. elegans animals by combining NMR and UPLC-MS. The effect of germline signals on aging-related metabolism was analyzed using multivariate statistics, including unsupervised principal component analysis (PCA), hierarchical and supervised orthogonal projection to latent structure with discriminant analysis (OPLS-DA). By characterizing the metabolic profile of wild-type, glp-1 mutants, and daf-16;glp-1 double mutants, our results demonstrate that glp-1 mutants regulate some age-related metabolic variations to achieve a long-lived phenotype, and some metabolic pathways influenced by germline-less signals are mediated by FOXO/DAF-16.

Results

Discussion

Researchers have recently dedicated considerable effort to the study of the mechanisms of longevity by utilizing OMIC approaches. In this work, we performed a metabolomics analysis by combining NMR and UPLC-MS. Our results indicated that the metabolome revealed promising metabolite markers involved in aging in C. elegans. Moreover, related metabolic pathway variations during aging were observed in glutathione metabolism, glutamate metabolism, taurine and hypotaurine metabolism, the TCA cycle, purine metabolism and pyrimidine metabolism and aminoacyl-tRNA biosynthesis. In addition, variations in trehalose and taurocholate were observed during aging. Furthermore, analyzing the metabotype of long-lived glp-1 mutants suggested that germline-mediated longevity is inseparable from the regulation of age-related metabolic variations.

In C. elegans, previous studies have also analyzed metabolic changes during aging by NMR. For example, Clement and coworkers found that the concentrations of saturated and unsaturated lipids, glycerophosphocholine (GPC), phosphorycholine (PCho), glutamine, and glycine increased with age, while another 14 metabolites (alanine, arginine, isoleucine, leucine, lysine, phenylalanine, tyrosine, valine, formate, cystathionine, glutamate, acetate, lactate and glycerol) decreased [27]. Other researchers compared 10-day-old adult (10A) worms with young adult (YA) worms and found that the concentrations of acetate, arginine, asparagine, betaine, choline, cystathionine, glutamine, glycine, histidine, isoleucine, leucine, malate, phosphorycholine, putresine, serine, threonine, trehalose, valine and glycerophosphocholine were increased in the 10A worms, while those of alanine, glutamate, lactate, proline and succinate were decreased [38]. Furthermore, Neil Copes and co-works also uncovered metabolic changes during aging through an analysis metabolome of glp-4 mutants cultured in liquid medium by GC-MS [39].

Most of these metabolomics studies with C. elegans have focused on finding biomarkers linked to senescence. However, the variation of metabolism during aging is a complex and sophisticated process. Recent work has suggested that, as individuals age, changes occur not only in the levels of specific molecules but also in the way that these molecules interact with one another within networks [40]. By contrast, our work discussed age-specific changes based on the network structure and illustrated that aging is characterized by metabolome remodeling. The metabolic pathways that we identified using our integrative approach also provide new insight into how the metabolic network is regulated by external and internal signals. In addition, our work was performed by combing NMR with UPLC-MS, which enhanced the coverage of the analysis.

In our analysis, we found that, compared with YA N2, the concentrations of glutathione (GSH), taurine and hypotaurine decreased, while oxidized glutathione (GSSG) increased in aging worms (Table S2). Glutathione and taurine are antioxidants, while oxidized glutathione is regarded as an indicator of increased oxidative stress. These results of decreased levels of GSH, taurine and hypotaurine and increased levels of GSSG with aging are consistent with previous studies in other model organisms [41–43], suggesting a vicious cycle of decreased antioxidant ability and increased oxidative stress that accompanies aging.

We also detected decreasing levels of the intermediates of purine metabolism (glutamine, ADP, AMP and allantoin) and pyrimidines intermediate metabolites (CDP, dUMP, glutamine, 3-aminoisobutanoic acid and CTP) in aged worms, which suggested the downregulation of the purine and pyrimidine metabolism during aging (Figure 2 and Table S2). Previous studies have reported that many diseases such as Alzheimer’s disease, immunodeficiency and growth retardation are related to disorders of purine and pyrimidine metabolism [44,45]. Consistent with our results, the downregulation of purine and pyrimidine metabolism in aging mice was previously indicated in a transcriptional analysis [46].

We also observed that some citric acid cycle intermediate metabolites, such as oxaloacetate, succinate, fumarate, malate and citrate, were upregulated in aged animals (Figure 2, Figure 5A and Table S2). Other researchers found that several TCA cycle genes were upregulated in long-lived Ames dwarf mice and Little mice compared with wild-type controls [47]. In yeast, several TCA cycle gene knockouts exhibited markedly extended lifespans compared with wild-type under DR conditions [48]. In long-lived C. elegans dauer larvae, downregulation of the TCA cycle was observed [49]. A recent study showed that higher concentrations of TCA cycle intermediates were linked to shorten lifespans, possibly due to increased cardiovascular risk [23].

In addition, decreased cystathionine was observed in our aged worms (Figure 2B and Table S2). Polymorphisms in cystathionine beta synthase (CBS), which catalyzes the conversion of homocysteine to cystathionine, is well-known risk factor for Alzheimer’s disease [50]. Our study also revealed that taurocholate, which is a bile acid conjugate of cholic acid and taurine, is a significant marker of longevity (Figure 2B and Table S2). Another recent study reported that increased taurocholate was related to reduce lifespan [23]. Taken together, these results demonstrate that aging is the result of the accumulation of molecular damage and reprogramming metabolism.

Furthermore, extending upon prior works, by analyzing the metabolome of glp-1 mutants, we found that the longevity phenotype of glp-1 mutants was inextricably linked to the metabotype. In C. elegans, the germline integrates nutrient signaling and communicates with other tissues to modulate aging. Previous studies have indicated that reproduction signals that affect aging were connected to the regulation of lipid metabolism, at least in part [51]. Another study reported the metabolic variations in glp-1 mutants by using gas chromatography (GC) and LC-MS [52,53].

We found that germline-less mutants regulated some age-related metabolic pathways, including glycerophospholipid metabolism, the TCA cycle, starch and sucrose metabolism, purine metabolism and pyrimidine metabolism, to attenuate aging. It should be noted that, in our results, for pyrimidine metabolism, the level of some metabolites (e.g., such CDP, UMP, CTP) were not influenced by glp-1 mutants in achieving the longevity phenotype at the young adult stage. This could be interpreted that these nucleotide metabolites are required for cell proliferation [54,55], such that at the young adult stage, fertile N2 can upregulate the level of nucleotides to carry out DNA replication and egg-laying compared with infertile glp-1. Thus, our results did not show increased levels of these nucleotide metabolites in glp-1 mutants at this stage. However, in day 10 of adult worms, the effect of egg-laying had been eliminated such that increased levels of CDP, UMP and CTP were observed in glp-1 mutants compared with N2.

In addition, trehalose was also significantly accumulated in the glp-1 mutants (Figure 4 and Tables S3 and S4). Trehalose accumulation was identified as part of a longevity signature in long-lived mutants [25], and exogenous trehalose treatment could also extend lifespan in C. elegans [56]. Trehalose is a significant storage molecule, which is considered to be an important stress-responsive metabolite and longevity assurance sugar in C. elegans [49].

Among the age-related metabolic pathways regulated by glp-1, the TCA cycle and pyrimidine metabolism showed patterns of DAF-16 dependence (Figure 6B). The TCA cycle was downregulated in the glp-1 mutants, but was not downregulated or moderately elevated in the daf-16;glp-1 double mutants. A similar trend was found for pyrimidine metabolism. It should be noted that, theoretically, the metabolic pathways were elevated or repressed in the glp-1 mutants but not in the daf-16;glp-1 double mutants. However, in our study, the double mutants showed a slight trend of opposite regulation. This result could suggest that several upstream signals affect DAF-16 to regulate aging and that the signaling pathways and metabolic pathways are complex and engage in crosstalk. Furthermore, our results illustrate that the glp-1 mutants extend lifespan by regulating various downstream targets, including DAF-12, FOXA/pha-4 and NHR-49, in addition to DAF-16 [19,36,37], such that some metabolic regulation in the glp-1 mutants was independent of DAF-16, such as purine metabolism, starch and sucrose metabolism and glycerophospholipid metabolism (Figure 6B).

We also noted the limitations of this study. First, a previous study has reported that the use of 5’-fluorodeoxyuridine (FUdR) may have a significant effect on the levels of many metabolites (including betaine, lactate, NAD, glycerol, tyrosine, glutamate, glutamine and GPC) [57], but the biomass of worms required for metabolomic analysis cannot be collected without using FUdR or a similar intervention. In our study, in order to reduce the influence of FUdR, all strains were maintained on the same plates either with or without FUdR, and before collecting aged samples, the worms maintained on the plate containing FUdR were transferred to a plate without FUdR and maintained for two days.

Another limitation associated with the use of the glp-1 strain is the required temperature shift of the eggs from 20 ^oC to 25 ^oC to induce sterility. This shift could affect stress resistance. Stress resistance is often directly proportional to longevity [58]. However, overall, in our trial, all C. elegans stains underwent the same treatment, which reduces the different effects of stress resistance on metabolism among the strains.

Comprehensive metabolome analyses often observe pairs of metabolites that are correlated across samples but which are apparently unrelated to the metabolic reaction. In the future, a study combining untargeted metabolomics with targeted metabolomics is necessary to exclude the unrelated information. More challenging, it remains to be seen whether age-related changes in metabolite correlations can be altered through genetic or pharmacological approaches or by consuming metabolites whose correlations change with age to further verify the discovery linked to aging by metabolomics.

Overall, aging is a complicated process related to cellular metabolism, and systems approaches such as comprehensive non-targeted metabolomics have the potential to disclose the linkage between metabolic changes with aging and varying stress resistance. This study combined NMR and UPLC-MS technology to display a comprehensive and unbiased footprint of the metabolic variations that accompany aging in different mutants at the whole organism level in C. elegans. Comparing young adults with older adults among different strains, we found that age-varying metabolic pathways included glutathione metabolism, glutamate metabolism, purine metabolism and pyrimidine metabolism, taurine and hypotaurine metabolism, and the TCA cycle, suggesting that aging emerges as a metabolome remodeling process and the accumulation of damage. In addition, in long-lived glp-1 mutants, we observed different metabolic characteristics compared with N2, and the metabolic differences were more marked in old stage worms than in YA worms, indicating that aging is associated with increasing metabolite diversity. Moreover, our results showed that long-lived glp-1 mutants influenced some aging-related metabolic variation, which demonstrates that the longevity phenotype of the glp-1 mutants is associated with its metabotype. Additionally, we found that the TCA cycle and pyrimidine metabolism was particularly significantly associated with the longevity regulated by germline-less signals, which is mediated by downstream DAF-16 transcription factors. Altogether, these results demonstrate that high-sensitivity metabolomic studies have excellent potential not only to reveal mechanisms leading to senescence but also to help reveal pathways associated with overall survival and longevity. Therefore, in subsequent studies, extending the analysis to lipidomics and combining other molecular level analyses, including transcriptomic or proteomic analyses, will provide further insight into the interrelationships between metabolic function with respect to longevity, as well as the association between germline-less signal regulation longevity and metabolism.

Materials and Methods

Supplementary Materials

Supplementary Information