Jennifer M. De Guzman, Ginger Ku, Ryan Fahey, Yun-Hee Youm, Ignatius Kass, Donald K. Ingram, Vishwa Deep Dixit, Indu Kheterpal
PMID: 22661299 PMCID: PMC3705111 DOI: 10.1007/s11357-012-9430-x

Calorie restriction (CR) remains the most robust metabolic intervention to extend lifespan and improve healthspan in several species. Using global and targeted mass spectrometry-based metabolomics approaches, here we show that chronic CR prevents age-related changes in specific metabolic signatures. Global metabolomic analysis using ultra-performance liquid chromatography–tandem mass spectrometry detected more than 7,000 metabolites in sera from ad-libitum-fed young, aged, and aged C57BL/6 mice maintained on 40 % CR. Multivariate statistical analysis of mass spectrometry data revealed a clear separation among the young, aged, and aged–CR mice demonstrating the potential of this approach for producing reliable metabolic profiles that discriminate based on age and diet. We have identified 168 discriminating features with high statistical significance (p ≤ 0.001) and validated and quantified three of these metabolites using targeted metabolite analysis. Calorie restriction prevented the age-related alteration in specific metabolites, namely lysophosphatidylcholines (16:1 and 18:4), sphingomyelin (d18:1/12:0), tetracosahexaenoic acid, and 7α-dihydroxy-4-cholesten-3-one, in the serum. Pathway analysis revealed that CR impacted the age-related changes in metabolic byproducts of lipid metabolism, fatty acid metabolism, and bile acid biosynthesis. Our data suggest that metabolomics approach has the potential to elucidate the metabolic mechanism of CR’s potential anti-aging effects in larger-scale investigations.

In the present study, we have utilized a global and targeted metabolomics approach to demonstrate its potential to investigate the effects of CR on aging processes. Serum from three groups of mice (3-month-old fed ad libitum (Y–AL), 26-month-old fed ad libitum (A–AL), and 26-month-old on chronic CR diet (A–CR)) were analyzed using UPLC–tandem MS. Mass spectrometry-based metabolomics methodology was used to take advantage of this technique’s sensitivity, wide dynamic range, selectivity, and capabilities for structural elucidation and quantification. Since blood constituents reflect biological processes occurring simultaneously in the tissues, analyzing metabolite profile of serum provides a unique first insight into the changes that occur due to CR. In addition, blood is readily available and can be obtained in a minimally invasive manner.

Global analysis was performed first to detect as many metabolites as possible without any bias towards specific classes of compounds. Over 7,000 metabolic features were detected in this study. Global analysis also allows measurement of changes in levels of all metabolites simultaneously. Multivariate statistical analysis tools were used to facilitate recognition of 168 features that contributed most significantly to discrimination between Y–AL and A–AL, A–AL and A–CR, and Y–AL and A–CR sample groups. Since many metabolites can have identical mass, database inquiries based on accurate m/z are not sufficient for identification. We therefore utilized high energy data, physico-chemical rules in mass spectrometry combined with mass spectral information from the literature and databases for metabolite identification. Targeted analysis using MRM–UPLC–ESI–MS was used to validate and quantify representative identified metabolites, namely oleoyl carnitine, palmitoyl carnitine, and docosapentaenoic acid.

Specifically, the age-related decrease in levels of two of the phospholipids, namely lysoPCs (16:1) and (18:4) were completely restored by CR. Fatty acid composition of membrane phospholipids in rats has been shown to shift from low to high degree of unsaturation with age, which increases the peroxidizability index whereas membranes from rats on CR, while exhibiting higher levels of essential fatty acids, showed lower levels on a peroxidizability index (Laganiere and Yu 1993). Because membranes with a high index of polyunsaturation have increased susceptibility to oxidation, lipid peroxidation is a common occurrence with aging, releasing oxidation products that contribute to chronic inflammation (Walton et al. 2003; Tahara et al. 2001). Evidence from long-lived animals have shown that the degree of unsaturation in membrane phospholipids is low in these animals compared with short-lived animals and the deacylation–reacylation cycle is implicated in this change in lipid profile (Portero-Otin et al. 2001). We have also identified a sterol lipid derivative, 7α-dihydroxy-4-cholesten-3-one, which is involved in bile acid biosynthesis pathway. This metabolite exhibited an age-related decline and was partially restored by CR. In addition, we show that levels of 15 age-associated metabolites are completely or partially restored by CR. Our data also suggest that the effect of CR is more pronounced than the influence of aging alone in the serum metabolic profile of mouse.

In conclusion, we were able to employ a MS-based global and targeted metabolomics approach to investigate serum metabolomic profiles related to CR and aging. This approach allowed the determination of discriminatory metabolites which were used to identify the metabolic pathways implicated in CR. This report does not intend to be a comprehensive metabolomic analysis of aging and CR, but rather our objective is to describe a methodology with potential for conducting larger scale studies. Identification of differentiating features without definitive identification is relatively straightforward and can be used to distinguish between various sample types and to evaluate CR mimetics. Because the serum metabolome is comprised of the sum of changes from all different tissues, our future studies will focus on use of this methodology to identify tissue specific alterations due to CR and aging. These future studies will also include investigation of effect of CR and aging across various mice strains and age groups.