Introduction

The aging process has been extensively studied over the years. One of the best-known definitions for aging is that of Alex Comfort (1956), who described this process as “a progressive increase throughout life, or after a given stadium, in the likelihood that a given individual will die, during the next succeeding unit of time, from randomly distributed causes”. In 1991 Rose (1991), in the book Evolutionary Biology of Aging, suggests a more general definition, characterizing aging as "a persistent decline in the age-specific fitness components of an organism due to internal physiological deterioration resulting in a loss of adaptive response to stress, and an increasing probability of death". According to this definition, as a consequence of aging, some visible changes occur in the individual, such as loss of mobility, and agility, muscle weakness, loss of reproductive capacity, and immune functions (known as immunosenescence), which results in a greater susceptibility to diseases such as Type 2 Diabetes Mellitus (DM2), infections, cancer, and autoimmune, cardiovascular, and neurodegenerative diseases, all known as age-related diseases (Chen and Liao 2007; López-Otín et al. 2013; Graça et al. 2017). The deterioration of the body's functions associated with aging is caused either by internal factors, such as oxidative stress (Pierce et al. 2008; Richardson and Schadt 2014), and DNA replication errors (Hoeijmakers 2009), which cause damage to DNA, proteins, and lipids (López-Otín et al. 2013) or by environmental factors, such as diet, sedentary lifestyle, and exposure to pollutants, microorganisms, and ionizing radiation (Nigam et al. 2012).

We live in a world with an increasingly aging population, so the study of aging is crucial (Carvalho 2012; He et al. 2016; Department of Economic and Social Affairs 2019). The birth rate worldwide has decreased in recent years (Sander et al. 2015), and the average life expectancy, and longevity have increased, especially in more developed countries, so it is expected that in the coming years people older than 65 will be the biggest percentage of world’s population (He et al. 2016). As a consequence of the physiological decline inherent of aging, age-related diseases can appear, which reduce the healthy life span (Rose 1991; Chen and Liao 2007; López-Otín et al. 2013; Graça et al. 2017), and increase the burden of healthcare systems, and families (World Health Organization 2015). To improve the quality of life of the elderly, we can study models of healthy aging, such as centenarians or groups of elderly people without associated diseases (Woo 2017). Studying the mechanisms, and possible biomarkers of healthy (physiological) aging may allow to intervene in the future in a preventive way through anti-aging therapies, which will delay the onset of age-related diseases, and increase the healthy lifespan (Seals et al. 2016; Magalhaes et al. 2018).

The investigation of biomarkers of aging, whether physiological or pathological, has been particularly important over the past years. According to the definition by George Baker and Richard Sprott (1988), a biomarker of physiological aging is “a biological parameter of an organism that either alone or in some multivariate composite will, in the absence of disease, better predict functional capability at some late age than will chronological age”. The American Federation for the Investigation of Aging has defined the criteria to categorize a molecule or panel of molecules as biomarkers of physiological aging: they must predict the rate of aging, be able to monitor the basic mechanisms that influence the aging processes, and not just the effects of associated pathologies, must be able to be tested repeatedly, and easily, without adverse effects, and be able to be measured, for example, in blood tests, and it should apply to both animals, and humans so that it can be tested on animals before being validated on humans (Johnson 2006).

To date, some parameters have been described that can be considered biomarkers of biological age. These include markers of physiological function, such as lung function, cardiovascular function, bone health, body composition, glucose metabolism, and levels of lipoproteins. Some molecular biomarkers have also been described, including biomarkers of the endocrine system, immune system, and nervous system, and some molecular mechanisms, such as oxidative stress, and genomic instability (Graça et al. 2017). Despite the extensive list of potential biomarkers of aging, none of those fits the definition of biomarkers mentioned before, since these biomarkers are mainly used to distinguish healthy, and unhealthy elderly people, as well as to differentiate people at higher risk of developing age-related diseases (Vasto et al. 2010). As such, they can be considered biomarkers of pathological aging, however none of them is specific enough to be considered a biomarker of physiological aging (Lara et al. 2015). Among all biomolecules, lipids appear to be an ideal target in the search for biomarkers. Some lipid alterations have already been associated with age-related diseases, and it is also known that these biomolecules are involved in many physiological processes that suffer alterations with aging. It has also been described that the lipid profile changes with age (Gonzalez-Covarrubias 2013). Thus, the study of the lipidome of the elderly without age-related diseases may prove to be a useful tool for understanding physiological aging.

This work aims to describe the importance of lipids as biomarkers of aging and summarize the changes in the lipidome that have been associated with aging, and longevity. Hopefully, it will help the reader to perceive how lipidome can help to understand age by identifying altered lipid pathways associated with physiological aging.

Metabolomics, and lipidomics applied to the study of aging

Metabolomics is one of the -omics disciplines of systems biology that, together with genomics, transcriptomics, and proteomics, studies the different organizational levels of a biological system (Boccard et al. 2009). More specifically, metabolomics studies the metabolome, which is the set of all biomolecules (metabolites) that result from the genomic, and proteomic activity, and their concentrations, and interactions in complex matrices, such as cells, tissues, biofluids, or organisms (Nicholson et al. 1999; Boccard et al. 2009; Soltow et al. 2010; Vivanco et al. 2010; Johnson et al. 2018). These metabolites are key factors in biological processes, and cellular metabolism, and therefore are representative of the general physiological state, in normal or pathological conditions (Soltow et al. 2010; Barallobre-Barreiro et al. 2013; Johnson et al. 2018). Metabolites can have an endogenous origin if they are compounds resulting from processes of biosynthesis or catabolic reactions or of exogenous origin if they derive from the food (nutrients), exposure to pollutants, or products of the metabolization of drugs (Boccard et al. 2009). The human metabolome consists of thousands of low molecular weight compounds (< 1000 Da), which can be divided into more than 50 chemical classes, including carbohydrates, vitamins, amino acids, fatty acids (FA) or lipids (Boccard et al. 2009; Wishart et al. 2009).

About 80% of the human metabolome is represented by lipids, and the complete set of all biological lipids is called lipidome (Ejsing et al. 2009; Psychogios et al. 2011). The plasma lipidome consists of thousands of lipids, which perform different functions, and have different structures (Quehenberger et al. 2010). Lipidomics is one of the approaches of metabolomics, and studies the lipidome of a cell, tissue, or organism, by measuring lipids (Ejsing et al. 2009; Gonzalez-Covarrubias 2013). Lipidomics allows the search of biomarkers to assist in the diagnosis, therapy, and prognosis of age-related diseases, and longevity (Gonzalez-Covarrubias 2013).

The role of lipids in the human body

Lipids are a large group of biomolecules that, despite being chemically different, are all insoluble in water. LIPID MAPS initiative created a lipid classification system that divides lipids into eight main categories: FA, glycerolipids, glycerophospholipids, sphingolipids, sterols, prenol lipids, saccharolipids, and polyketides (Fahy et al. 2005; LIPID MAPS 2017). Lipids are essential to the human body, and perform several functions, ranging from energy storage to regulating fluidity of cell membranes (Nelson and Cox 2005), not to mention its role in regulation and signaling, insulation, and protection (Berg et al. 2004). Glycerolipids, in particular triacylglycerols (TG), are the organism's main source of energy storage (Nelson and Cox 2005). Most TG energy reserves are found in adipocytes, cells specialized in their storage, and synthesis, and that are related to insulation, and protection function of this biologic molecule. The energy yield of TG, that results from the oxidation of the FA, is much higher than of other biomolecules (Berg et al. 2004).

Lipids are key components in the signal transduction pathway with products of lipid metabolism serving as intracellular messengers, such as phosphatidylinositol (PI) which, when hydrolyzed, gives rise to diacylglycerol (DAG), the best studied lipid with this function (Hannun 1997; Nelson and Cox 2005). They can be used to synthesize hormones, such as prostaglandins, thromboxanes, leukotrienes, and steroid hormones (progesterone, testosterone, estrogen, and cortisol), which are derived from arachidonic acid (AA), and cholesterol, respectively (Nelson and Cox 2005). So, lipids may be involved in cellular signaling processes such as inflammation, survival, and responses to cellular stress (Hannun and Obeid 2008). Phospholipids are the main constituents of cell membranes (Shevchenko and Simons 2010; Muro et al. 2014). Membrane lipids constitute a large group of structurally different lipids (Klose et al. 2013). It is this enormous diversity that justifies the properties of cell membranes, and the various biological processes in which lipids are involved, such as trafficking across the membrane, the regulation of membrane protein composition, and cell architecture, and the formation of lipid rafts—specific sub-compartments in membranes that contribute to cell function (Lee 2004; Shevchenko and Simons 2010; Bigay and Antonny 2012). According to the KEGG database (last accessed September 12th 2020), there are several processes involved in lipid metabolism in humans, such as FA biosynthesis, degradation, and elongation; synthesis, and degradation of ketone bodies; steroids, bile acids, and steroid hormones biosynthesis; glycerolipids, glycerophospholipids, ether lipids, sphingolipids, AA, and linoleic acid metabolism, and unsaturated FA biosynthesis.

Lipidomics in the study of aging biomarkers

Metabolites may be possible biomarkers because they represent several intermediate metabolic pathways (Nicholson et al. 1999; Soltow et al. 2010), and through the identification of changes in metabolites, it is possible to understand the interactions between gene expression, enzyme activity, and metabolic reactions, and relate them to the phenotype (Nicholson et al. 2005; Soltow et al. 2010). In clinical practice, lipidomics are used to establish lipidic profiles that allow to predict the progression of several diseases. To improve the prognostic, and diagnostic power of these tests, the lipidic profiles can be combined with risk factors, such as demographic risk factors (ethnicity, geography, age, and sex); lifestyle (smoking, physical inactivity, and inadequate diet), or physiological risk factors (body mass index (BMI), DM2, dyslipidemia) (Rasmiena et al. 2013).

Lipidomics appears to be an even more important field in the search for new biomarkers of physiological, and pathological aging (Ishikawa et al. 2014; Meikle et al. 2014) because they are involved in biological processes that change with age, and because they have already been linked to longevity in several long-lived animal species (Gonzalez-Covarrubias 2013; Naudí et al. 2013), and since several diseases have deregulation of lipids in its onset, studying the lipidome, can allow to predict the risk of developing a disease in the future, understand the pathological mechanisms, and monitor the response to therapies, and identify possible biomarkers for the diagnosis, and prognosis of these diseases (Laaksonen et al. 2006; Ng et al. 2012; Rasmiena et al. 2013; Ishikawa et al. 2014; Meikle et al. 2014; Mielke et al. 2014).

Several studies have shown that the composition of lipid cell membranes between vertebrates, invertebrates, and long-lived animals is different, and that the susceptibility of membrane lipids to peroxidation is inversely related to the longevity of the respective animal species (PAMPLONA et al. 2002; Hulbert et al. 2007; Pamplona 2008; Naudí et al. 2013). It is thought that the lipid profile of cell membranes has been optimized during evolution to be more resistant to oxidative stress. The evolution of aerobic life has resulted in the production of reactive oxygen species (ROS), leading to oxidative stress. For this reason, antioxidant defense systems emerged, and cell membranes underwent structural, and functional adaptations, which in turn determine the longevity of different organisms (Pamplona and Barja 2007; Pamplona 2008; Naudí et al. 2013). Jobson et al. (2010) tried to identify genetic targets of natural selection to increase longevity in mammals through a phylogenomic approach. They found that the genes involved in the lipid composition suffered an increased selective pressure in long-lived species, specifically genes involved in cellular lipid metabolism, like mitochondrial fatty acid synthase (FAS II) pathway genes Mecr, and Oxsm, Scd5 gene, and Elovl5 gene, all of them involved in FA elongation, and biosynthesis (Jobson et al. 2010). The plasma lipidome can serve as an improved resource to be associated with longevity, and is specific to each animal species (Jové et al. 2013; Bozek et al. 2017). Besides, some studies have associated mutations in genes related to lipid metabolism with increased lifespan in humans (Barzilai et al. 2001; Vaarhorst et al. 2011; Jové et al. 2017).

The analytical techniques most used in metabolomics studies are mass spectrometry (MS), and nuclear magnetic resonance (NMR) (Markley et al. 2017). MS-based technologies are the gold standard for metabolomics, and lipidomics studies (Boccard et al. 2009), and therefore have also been applied in the investigation of biomarkers of aging (Mishur and Rea 2012). MS allows the study metabolites using targeted or untargeted methodologies (Fiehn 2002; Boccard et al. 2009; Jové et al. 2016). Untargeted methods have been used to study changes in the lipidome associated with aging, and age-related diseases (Ishikawa et al. 2014; Meikle et al. 2014). Although not so widely used, NMR alone or coupled to MS can also be applied to metabolomics studies (Markley et al. 2017). Its main disadvantage is its lower sensitivity, when comparing to MS (Fan and Lane 2016). However, it has several advantages. With NMR, reproducible data, and accurate quantification of compounds are obtained, and it is possible to study different types of samples (cells, tissues). In vivo images can also be obtained, which allow studying the dynamics of metabolic pathways in tissues, and living organisms. NMR also allows to identify compounds with identical masses, and is more effective in determining structures of unknown compounds. This technique is also preferable for studying compounds that are difficult to ionize by MS (Fan and Lane 2016; Markley et al. 2017).

How the lipidome can help to understand disease: pathological aging

Due to the central role that lipids play in the body, changes in their metabolic pathways can lead to the development of diseases. Some of these diseases are hereditary, and occur due to deficiencies in enzymes that are involved in the metabolism of lipids, which leads to the deposition of these molecules, and their derivatives in cells. Some examples are Niemann-Pick, Fabry, and Gaucher diseases. Niemann-Pick disease develops due to the accumulation of cholesterol, and sphingomyelin in the tissues, due to a deficiency in the sphingomyelinase enzyme (Schuchman and Desnick 2017). Fabry's disease results from a deficiency in the enzyme α-galactosidase A, which leads to the accumulation of globotriaosylceramide in various tissues, and organs, including the skin, eyes, heart, kidneys, and vascular, and nervous systems (Yuasa et al. 2017). Gaucher's disease is due to a deficiency in the enzyme β-glucocerebrosidase, which leads to the accumulation of glucocerebroside in macrophages (Adar et al. 2016).

Several diseases can also arise with aging, such as atherosclerosis, DM2, arterial hypertension, dyslipidemia, cardiovascular, and neurodegenerative diseases (Belikov 2019). These pathologies are also associated with lipid changes, and disturbances in their metabolic pathways (Naudí et al. 2015; Hannun and Obeid 2018). Using metabolomics techniques, several authors have studied the plasma lipid profile characteristic of these pathologies (Kivipelto et al. 2002; Nelson et al. 2006; Bismuth et al. 2008; Li et al. 2009; Solomon et al. 2009; Graessler et al. 2009; Gall et al. 2010; Suhre et al. 2010; Hu et al. 2011; Spijkers et al. 2011; Han et al. 2011; Kulkarni et al. 2013; Whiley et al. 2014; González-Domínguez et al. 2014; Klavins et al. 2015; Marais 2015; Welty 2015; Hatano et al. 2015; Ramasamy 2016; Chan et al. 2017; LeWitt et al. 2017; Müller-Wieland et al. 2017; Taylor et al. 2017; Valaiyapathi et al. 2017; Zhang et al. 2017; Guedes et al. 2017; Staniszewska-slezak et al. 2018; Stoessel et al. 2018; Glaab et al. 2019; Hamid et al. 2019).

In hypertension, studies show differences in plasma levels of different lipid classes, between normotensive, and hypertensive individuals. In men, decreased plasma levels of phosphatidylethanolamine (PE) ethers, phosphatidylcholine (PC) ethers, in which AA is the most abundant FA moiety, are associated with pathogenesis of hypertension (Graessler et al. 2009). The increased plasma levels of ceramides in hypertensive individuals are observed, and this molecules could be associated to endothelial dysfunction, which contributes to hypertensive pathophysiology (Spijkers et al. 2011). Increased plasma levels of PC, and TG were observed in plasma samples from hypertensive individuals, and the lipid metabolism in this patients is different from that normotensive individuals (Hu et al. 2011). Some DAG species have been associated with high systolic, diastolic, and mean blood pressure values, and with a greater risk of normotensive individuals developing hypertension, what suggests an important role for the upregulation of the DAG axis in the pathogenesis of this disease (Kulkarni et al. 2013).

In DM2, some studies indicate that PC, lysophosphatidylcholines (LPC), and PC with polyunsaturated carbon chains are reduced, while PE were found at higher levels in plasma samples of diabetic individuals. This levels are associated with lower HDL, and total cholesterol levels, and higher TG levels in diabetic individuals (Suhre et al. 2010).Some free fatty acids (FFA) were found to be at increased levels in plasma of diabetic individuals (Li et al. 2009). The elevated levels of FFA, and lipids could be a factor for the development of DM2, and is associated with hyperinsulinemia, and insulin resistance (Shulman 20AD; Roden et al. 1996). It is described that there is a decrease in the concentration of lysophospholipids, and long-chain acylcarnitines with increased insulin resistance (Gall et al. 2010). Also, in diabetic mice there was an increase in plasma TG, and longer fatty acyl chains content, and this could be associated with dysregulated lipids metabolism in DM2 (Staniszewska-slezak et al. 2018).

Dyslipidemias consists of quantitative or qualitative changes in blood lipids (Ramasamy 2016). They can be divided into four different types: hypercholesterolemia, which corresponds to increased cholesterol levels (Taylor et al. 2017); hypertriglyceridemia, which corresponds to an increase in TG levels (Valaiyapathi et al. 2017); mixed dyslipidemia, which is a combination of the two previous types (Müller-Wieland et al. 2017), and the decrease in high-density lipoprotein (HDL) levels (Ramasamy 2016). There are other variants, such as hypobetalipoproteinemia, which is due to an increase in the number of low-density lipoprotein (LDL) receptors, and leads to a decrease in plasma cholesterol, abetalipoproteinemia, in which there is a deficiency in the secretion of lipoproteins containing apolipoprotein B (ApoB) (Welty 2015; Ramasamy 2016), and, finally, to dysbetalipoproteinemia or apolipoprotein E deficiency, in which an accumulation of TG-rich lipoproteins occurs (Marais 2015). Dyslipidemias are one of the main risk factors for atherosclerosis, and cardiovascular diseases (Ramasamy 2016). In atherosclerosis, Nelson et al. (2006) found that increased levels of plasma sphingomyelin (SM) are associated with subclinical disease, and severity, since plasma SM are involved in intermediate pathways between traditional risk factors, and subclinical atherosclerosis (Nelson et al. 2006). Also, the accumulation of ceramides seems to influence the development of atherosclerosis, since they are associated with the formation of foam cells, the aggregation of LDL, and the increase in ROS levels (Bismuth et al. 2008).

The neurodegenerative diseases most associated with the aging process are Alzheimer's disease (AD), and Parkinson's disease (PD) (Belikov 2019). In AD, higher levels of serum cholesterol during midlife have been associated with an increased risk of developing the disease, having a role in pathogenesis of the AD (Kivipelto et al. 2002), and greater cognitive impairment at older ages. This happens because the higher cholesterol levels in the midlife is associated with executive dysfunction, through increasing vascular pathology (Price et al. 2006). Other studies have shown that plasma levels of specific PC species are decreased, and the levels of two specific LPC species are increased in AD patients (Whiley et al. 2014; Klavins et al. 2015). Also, the ratio PC/LPC can differentiate between healthy individuals, individuals with mild cognitive impairment, and AD patients (Klavins et al. 2015). These results can be explained by aberrant activity of the enzyme phospholipase A2, what catalyze the cleavage of phospholipids in FA, and LPC. The overactivation of this enzyme is responsible for the abnormal metabolism of brain phospholipids, and changes in cell membranes (Farooqui et al. 2004). Moreover, β-amyloid42 peptides can increase the phospholipase A2 activity, and promotes alterations in physical properties of cell membrane (Hicks et al. 2008). González-Domínguez et al. (2014) found that the serum levels of PC with polyunsaturated fatty acids (PUFA), PE, LPC, lysophosphatidylethanolamines (LPE), and plasmalogens, plasmenylethanolamine, and plasmenylcholine, are decreased in these patients, whereas levels of certain PC species with saturated fatty acids (SFA), are increased. Once again, these observed lipid changes may be associated with overactivation of the phospholipase A2 enzyme (Farooqui et al. 2004). In the case of PCs, the overactivation of this enzyme may be related to the dysregulation of biosynthesis, and remodeling of the acyl chains, which will change the profile of FA (Farooqui et al. 1997). The imbalance between saturated FA, and unsaturated FA can also contribute to the alterations in cell membranes (González-Domínguez et al. 2014). The PE alterations are could be related with oxidative stress associated with AD, since they are rich in oxidable FA, and the brain is very susceptible to reactive oxygen species (Migliore et al. 2005; González-Domínguez et al. 2014). The decrease in plasmalogens may be associated with the dysfunction of peroxisomes (Kou et al. 2011), and with the increased activity of the enzyme plasmalogen-selective phospholipase A2 in the brain of AD patients (Farooqui 2010). In addition, plasmalogens are mainly constituted by polyunsaturated carbon chains, which makes them more susceptible to reactive oxygen species, and oxidative stress (Guan et al. 1999; González-Domínguez et al. 2014). In the brain of AD patients it has been described an impairment in metabolism of lysophospholipids, which results from the increased activity of the enzymes lysophospholipid acyltransferase, which recycles lysophospholipids into phospholipids, and lysophospholipase, which converts lysophospholipids in lysophosphatidic acid (Ross et al. 1998; Umemura et al. 2006). Lysophospholipids so can be considered intermediates in the neuronal metabolic pathways involved in the pathophysiology of AD (Frisardi et al. 2011). The levels of SM are decreased, especially those with long aliphatic chains in their composition, and the levels of ceramides are increased in these patients. The changes observed in the two lipid species seem to be involved in the inflammatory stress associated with neuronal apoptosis, and in the AD pathogenesis (Han et al. 2011).

In PD, some authors have found elevated levels of ganglioside GM3 (Chan et al. 2017), and ganglioside-NANA-3 (Zhang et al. 2017) in the plasma of patients. This two members of gangliosides family are precursors for the complex gangliosides of a- and b- series in the brain (Proia 2004; Blanz and Saftig 2016). Those gangliosides are metabolized by enzyme β-glucosidase (GCase), and the reduced GCase activity in the nervous system (Murphy et al. 2014) has been associated with PD (Papagiannakis et al. 2015; Kim et al. 2016), maybe because of the alteration in lipid metabolism by GCase (Chan et al. 2017). In addition, GM3 can regulate α-synuclein-induced channel formation through its binding to the protein's ganglioside-binding domain (GBD), thus being able to play a protective role in the pathogenesis of PD (Di Pasquale et al. 2010). The GM3 also can accelerate α-synuclein aggregation (Grey et al. 2015). The serum levels of some PUFA (eicosapentaenoate) appear to be decreased, what could be associated with α-synuclein patological aggregation (Assayag et al. 2007; Hatano et al. 2015). Plasma levels of some SFA (dodecanoic acid, and hexadecanoic acid) appear to be increased in these patients, which indicates that there is a greater breakdown of lipids, which may be associated with diet or the disease itself (Glaab et al. 2019). In addition, the accumulation of fatty acids may represent changes in their oxidation, which may be involved in the pathophysiology of PD (Ruipérez et al. 2010). Plasma levels of some N-acyl-phosphatidylethanolamines (NAPE) were decreased in patients with PD, and this panel of NAPEs allows to distinguish between women with, and without the disease. These decreased levels may be associated with the transfer of NAPEs from the plasma to the CNS (Hamid et al. 2019). Other authors have observed an increase in the levels of NAPEs in the brain of mice with PD (Basit et al. 2016). The levels of various FA, glycerophospholipids, and sphingolipids are altered in these patients: it has been observed an increase in plasma levels of some PC species, and platelet-activating factors, and a decrease of some PC, PE, and SM species in PD patients (Stoessel et al. 2018). These results indicate that PD is associated with changes in the glycerophospholipids, and sphingolipids metabolism (Kori et al. 2016). In another study, an increase in serum levels of LPE, lysophosphatidylserine (LPS), monohexosylceramide, and some plasmalogens, and SM species, and a decrease in serum levels of dihydrosphingomyelin, phosphatidic acid (PA), and some PE, PC, plasmalogens, and NAPE species were observed in patients’ carriers of the GBA mutation. The accumulation of lipids like monohexosylceramide in carriers of GBA mutation could be associated with disease development (Guedes et al. 2017). In addition, decreased levels of PE can lead to endoplasmic reticulum stress, and the accumulation of α-synuclein, and consequently, cell death (Wang et al. 2014). Some medium- and long-chain FA in plasma have a positive correlation with the disease, being predictive of its progression (LeWitt et al. 2017). In a study of rats treated with rotenone, an mitochondrial toxin, the authors observed that there were major changes in brain, and systemic fatty acid metabolism (Tyurina et al. 2015).

The study of the lipid profile of age-related diseases may, in the future, help to better understand them, and assist health professionals in their diagnosis, and treatment. In the same way, the study of the lipid profile of healthy elderly people can allow us to understand the differences in the metabolic pathways in physiological aging (Gonzalez-Covarrubias 2013).

Lipid changes in physiological and healthy aging

Throughout aging, there are changes in several physiological processes, which translates into changes in the lipidome. Therefore, the scientific community has joined efforts to understand which of these alterations can characterize the aging process, trying to assess which lipid classes suffered alterations over the years, in plasma, and serum samples, mainly using MS, and NMR-based approaches (Lawton et al. 2008; Lee et al. 2009; Yu et al. 2012; Collino et al. 2013; Gonzalez-covarrubias et al. 2013; Ishikawa et al. 2014; Montoliu et al. 2014; Mielke et al. 2015; Jové et al. 2016, 2017; Darst et al. 2019; Pradas et al. 2019; Wong et al. 2019).

Concerning to FA, different studies present different results. Collino et al. (2013) observed that the levels of 11,12- dihydroxy-eicosatrienoic acid (11,12-DiHETrE), 9-hydroxy-octadecadienoic acid (9-HODE), and 9-oxo-octadecadienoic acid (9-oxo-HODE) decreased in centenarians, on the opposite to the levels of 15-hydroxy-eicosatetraenoic acid (15-HETE), and leukotriene E4 (LTE4), which were found to increase, when compared to adults, and elderly groups. Other authors verified that the levels of several FA, and their derivatives, like AA (Lee et al. 2009), resolvin D6, hydroxyl FA, prostaglandin (Jové et al. 2016), and oxidized FA (oxFA) (in women) (Ishikawa et al. 2014) decrease with age, in plasma, and serum samples, and the levels of acylcarnitines (Yu et al. 2012; Darst et al. 2019), oxFA (in men) (Ishikawa et al. 2014), β-hydroxybutyrate (Lawton et al. 2008), long-chain FA, and others FA (Lawton et al. 2008; Darst et al. 2019) increase. Jové et al. (2017) show that increased levels of SFA, and longer carbon chains, and decreased levels of PUFA are associated with extreme longevity. The authors realized that the FA C16:0, and C22:1n-9 appear to be promising biomarkers of extreme longevity. Gonzalez-Covarrubias et al. (2013) found that the lipidome associated with longevity is characterized by a high monounsaturated fatty acids (MUFA)/PUFA ratio.

Regarding glycerolipids, Gonzalez-Covarrubias et al. (2013) found that there are decreased levels of long-chain TG in descendants of nonagenarians, similar to Montoliu et al. (2014), that observed a decrease in long-chain TG, and DAG, but also an increase in very long-chain TG in centenarians. Finally, Jové et al. found that the levels of DAG, and TG were at increased levels in groups of adults, and centenarians (Jové et al. 2017), and that the levels of monoacylglycerol 22:1, and DAG 33:2 were decreased with age (Jové et al. 2016). We can conclude that the differences observed in the levels of TG, and DAG between the studies, may be due to the increase or decrease in the levels of specific glycerolipid species.

Phospholipids are one of the classes that experience different changes during aging, either increased or decreased levels, depending on the species. For example, PC ethers have been associated with longevity, being found at high levels in the female descendants of nonagenarians (Gonzalez-covarrubias et al. 2013). In centenarians, polyunsaturated PC ethers were found at increased levels, while saturated PC ethers were found at decreased levels (Montoliu et al. 2014). Pradas et al. (2019) found that there was a general decrease in the levels of ether lipids in the plasma of centenarians comparatively with adults, and no differences comparatively to elderly. Also, they found that lipids with alkyl bonds were reduced in groups of elderly individuals, and centenarians compared to adults, and lipids with alkenyl bonds (plasmalogens) were reduced in centenarians compared to adults, and elderly subjects. These authors observed that there is a specific signature of ether lipids in the plasma of centenarians, with a significant increase of 15 ether lipids derived from PC with a predominant presence of the alkyl form, consisting of shorter carbon chains, and less content of double bonds; a significant decrease in 6 ether lipids derived from PE with a predominant presence of the alkenyl form, consisting of longer carbon chains, and greater content of double bonds. This specific lipid signature can mean greater resistance to lipid peroxidation in the centenarians, and consequently, to lower levels of oxidative stress, which is known to be related to exceptional longevity in various organisms, including humans (Pamplona and Barja 2007; Pradas et al. 2019). However, Collino et al. (2013) described a decrease in PC ethers in centenarians. In general, the levels of PC species appear to be increased in centenarians (Collino et al. 2013; Montoliu et al. 2014), and elderly men, and women (Ishikawa et al. 2014). Levels of diacyl phosphatidylcholine increased with age in women (Yu et al. 2012), and LPC levels decrease with age (Collino et al. 2013). Concerning the remaining phospholipids, Gonzalez-Covarrubias et al. (2013) observed that the levels of PE 38:6 were found to be decreased in female descendants of nonagenarians. Jové et al. (2017) found that the levels of specific species of LPC, LPE, lysophosphatidylglycerol (LPG), PA, PE, phosphatidylglycerol (PG), and phosphatidylserine (PS) were increased in young adults, and centenarians, while other specific species of LPE, lysophosphatidylinositol (LPI), PA, PG, and PS decreased. The variability in the results presented by the different studies for the levels of phospholipids is notorious. This difference may be related to the specific species of phospholipids that are analyzed, which may vary between studies. In addition, there is a vast amount of structurally different phospholipids that perform different body functions, integrating mechanisms that change with aging. Thus, it is expected that the levels of the various phospholipids will vary depending on the changes in the different pathways in which they are involved.

In general, there seems to be an increase in the levels of ceramides, and dihydroceramides with age (Ishikawa et al. 2014; Mielke et al. 2015). However, Jové et al. (2017) found decreased levels of ceramide d39:0, glucosylceramides, and ganglioside GM3 in adults, and centenarians. Regarding SM, they seem to increase in healthy aging (Yu et al. 2012; Collino et al. 2013), and are found at high levels in centenarians (Montoliu et al. 2014), and descendants of nonagenarians (Gonzalez-covarrubias et al. 2013). Darst et al. (2019) reported an increase in 24 sphingolipids with age. To the best of our knowledge, only Collino et al. (2013) described decreased levels of SM in centenarians.

Regarding sterols, during the aging process it appears to be an increase in cholesterol, and cholesteryl esters levels (Lawton et al. 2008; Ishikawa et al. 2014; Jové et al. 2017). Nevertheless, Lawton et al. (2008) found that dehydroepiandrosterone sulfate (DHEA-S) levels decreased with age. Also, Darst et al. (2019) concluded that the 11-ketoethiocololone glucuronide, an androgenic steroid, and cortisol increased with age but the levels of 29 other sterols decreased with age. The levels of 7β-hydroxycholesterol, and 27-hydroxycolesterol were also found to decrease with age, after the age of 50 (Lee et al. 2009). Wong et al. (2019) realized that, in APOE ε3 homozygous individuals, there is a generalized, and significantly decrease in plasma levels of all lipid classes, except for DAG species, especially after 95 years of age. However, all lipid classes were negatively affected by age. This decrease is independent of sex, and BMI. The authors also describe a significant positive interaction of age with sex for species of ceramide, DAG, TG, PC, PE, and cholesteryl esters 20:X, and a more marked negative effect of age on men than on women. Tables 1, and 2 show the summary of the described lipid changes in serum, and plasma, respectively, in healthy aging, and longevity, divided by lipid classes.

Table 1 Brief summary of lipid alterations in serum samples, in healthy aging, and longevity
Full size table
Table 2 Brief summary of lipid alterations in plasma samples, in healthy aging, and longevity
Full size table

Possible effect of dietary habits in lipid profile during aging

Dietary habits change throughout life. Therefore, it is expected that the basic diet of older individuals will be different from the diet of individuals of other ages (Yannakoulia et al. 2018; Kehoe et al. 2019). Changes in dietary habits in the elderly may be related to several factors, such as loss of appetite due to decreased sense of smell and taste, the interference of drugs with food intake, absorption and metabolism, the deterioration of oral health that requires the choice of some foods over others and social and psychological changes that can modify the way people eat (Ship 1999; Amarya et al. 2015; Pilgrim et al. 2015). Dietary habits thus influence the nutritional status of individuals, and in the elderly this can translate into malnutrition states (Zhu et al. 2010; Wysokiński et al. 2015). For example, several micronutrients, such as vitamins D, B2 and B12, folate and calcium are at levels below those recommended for the elderly (Kehoe et al. 2019). There are no significant changes in protein and carbohydrate intake with age (Yannakoulia et al. 2018), however there is an increase in salt (Cappuccio et al. 2015; Moreira et al. 2018) and sugar intake (Sluik et al. 2016; Ruiz et al. 2017).

In the elderly, there is a gradual decrease in the total intake of lipids (Anderson et al. 2011). More specifically, there is an increase in the intake of SFA (which should be as low as possible) and a decrease in the intake of MUFA and PUFA (Ter Borg et al. 2015). With regard to PUFAs, there appears to be a deficit in the intake of ω3-PUFA, such as α-linoleic acid, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) in the elderly (Carrière et al. 2007). This specific group of PUFA is extremely important as it has numerous benefits. ω3-PUFA are associated with the maintenance of bone health and muscle tone, inhibit TG synthesis in the liver, decrease the inflammatory process and decrease the risk of cognitive performance decline associated with aging (Uauy and Valenzuela 2000; Farina et al. 2011; Kesse-Guyot et al. 2011; Smith et al. 2011; Úbeda et al. 2012). EPA and especially DHA further reduce the risk of cardiovascular and neurodegenerative diseases (Freund-Levi et al. 2006; Gebauer et al. 2006). Currently, there are already several dietary supplements in order to increase the levels of ω3-PUFA (mainly EPA and DHA) (Plourde and Cunnane 2007). Often, the ω3-PUFA used in supplements are derived from fish oils (Uauy and Valenzuela 2000; Hulbert et al. 2014).

It would make sense that changes in dietary lipid intake associated with aging would influence the plasma or serum lipidome. However, studies that referred to changes in the lipid profile induced by the diet in the elderly are missing. Abbott et al. (2012) tried to understand the relationship between the FA ingested in the diet and the phospholipids of the cell membranes and the TG of the adipose tissue and plasma. The authors concluded that diet SFA, MUFA and PUFA do not significantly influence membrane lipids. On the other hand, the composition of TG in adipose tissue and plasma appeared to be influenced in accordance with the dietary FA. It should be noted that some of the authors tried to minimize the influence of the diet on the lipidomic results by controlling the timing of sample acquisition and in some of the studies, blood samples were taken in the morning after donors fasted overnight (8h-14h) (Yu et al. 2012; Collino et al. 2013; Ishikawa et al. 2014; Montoliu et al. 2014; Jové et al. 2016; Darst et al. 2019; Pradas et al. 2019).

Altered lipidic pathways associated with physiological aging

Centenarians are recognized as representing the ideal physiological aging model, and therefore, are widely studied (Franceschi and Bonafè 2003; Franceschi et al. 2007). Although the causes of centenarians' longevity are not yet completely understood, it is believed that it may be due to a lower incidence of serious diseases (Barzilai et al. 2003; Andersen et al. 2012; Kheirbek et al. 2017), and the increased activity of protective mechanisms (Andersen et al. 2012). Although they also show signs of inflammation with aging, these individuals do not appear to suffer from its negative consequences. This is because there seems to be a balance between pro- and anti-inflammatory factors, which causes a slower, limited, and balanced increase in inflammation levels with age, compared to elderly non-centenarians (Franceschi and Bonafè 2003; Montoliu et al. 2014). As oxidative stress, and inflammation regulate a series of metabolic processes, it is possible to infer that the study of changes in the metabolome in these individuals allows to understand which are the processes inherent to physiological aging (Montoliu et al. 2014).

The role of different lipid classes in cell senescence

Sphingolipids

It has been reported that senescent cells have increased levels of glycerophosphocholine, reinforcing the importance of phospholipid metabolism in cellular senescence (Gey and Seeger 2013). Choline is associated to the synthesis of SM, and PC, from which DAG, and ceramides are produced (Tang et al. 2013; Montoliu et al. 2014). Sphingolipids appear to play an important role in healthy aging (Jové et al. 2017). SM can be constituents of cell membranes, and be associated with cholesterol, also forming lipid rafts, essential for the formation of the cell membrane, transport, and metabolism processes (Ito et al. 2000). In neuronal cells, they can even promote signal transduction (Yu et al. 2012). Also, some sphingolipids species, such as ceramides, are involved in cell signaling (Maceyka and Spiegel 2014), mediating processes such as cell senescence, and differentiation, autophagy, and apoptosis (Hannun and Obeid 2008, 2011). Oxidative stress accelerates the degradation of SM in ceramides. SMAses activity increases with age, which leads to a greater degradation of SM in ceramides (Smith et al. 2006; Smith and Schuchman 2008).

Phospholipids

As previously mentioned, changes in membrane's phospholipids will change its fluidity, which will interfere with the function of membrane proteins, and the transport of substances across the membrane (Shevchenko and Simons 2010). One of the compartments of cell membrane that suffer alterations with age is lipid rafts/caveolae, that include G protein-coupled receptor (Barnett-Norris et al. 2005; Naru et al. 2008). Caveolin-1, a component of caveolae, and whose levels increase with aging, induces changes in signal transduction, and in cell morphology associated with senescence (Park et al. 2000; Park 2005). Also, the G protein is sensible to alterations in the lipid composition of the cell membrane (Alemany et al. 2007), and senescence leads to changes in cellular properties, such as cell morphology, and loss of homeostasis (Naru et al. 2008). In senescent cells, there is a greater uptake of specific PC species stearic acid (C18:0), and AA (C20:4), and long-chain FA, so the expenditure of specific phospholipids may be related to membrane changes associated with cell senescence, therefore the importance of depicting the lipidome in chemical detail (Naru et al. 2008; Engelmann and Wiedmann 2010). The increase in uptake of PC species with long-chain FA can also be associated with the development of some age-associated diseases, such as cardiovascular diseases (Engelmann and Wiedmann 2010). These senescent-related changes can be associated with an increase in the surface area during senescence (Naru et al. 2008). Phospholipids, PC, and PE are the main components of cell membranes. It has also been previously described that the PC/PE ratio is an modulator of membrane integrity, since the alterations in PC levels could change the characteristics of lipid bilayer, and consequently their fluidity (Li et al. 2006). Also, PC/PE ratio could be associated with inflammation. The decrease in PC/PE ratio results in a change of membrane potential that allows the flow of ions, and ion radicals what is associated with the beginning of inflammation. The loss of membrane integrity also allows the influx of extracellular components, such as cytokines (Kmiec 2001; Li et al. 2006). It was proven that, in centenarians, and long-lived individuals, the lipid composition of erythrocyte membrane allows a higher membrane fluidity, and a decreased susceptibility to a peroxidation (Caprari et al. 1999; Rabini et al. 2002; Puca et al. 2008).

FA

Some studies report an inverse relationship between the content of double bonds of membrane lipids, and longevity in mammals (Pamplona et al. 1999).The susceptibility to lipid peroxidation increases exponentially with the increase in the number of FA double bonds (Hulbert et al. 2007). Decreased levels of PUFA are associated with less lipid peroxidation, and consequently, less oxidative damage (PAMPLONA et al. 2002).

Ether lipids

Plasmalogens are components of the cell membranes, and exist mainly in cardiac cells, and neurons. They perform several functions: participate in the fusion of membranes, in the signal transduction mediated by G protein, in the transport of cholesterol, and assist in the control of vesicular function, and the function of membrane proteins. Also, they are involved in maintaining some properties of the membranes, such as their fluidity, and the stability of the lipid raft microdomains. They may also be involved in cell signaling, giving rise to second messengers (Wallner and Schmitz 2011; Braverman and Moser 2012; Dean and Lodhi 2018). Plasmalogens could have an antioxidant effect, being able to protect other unsaturated membrane lipids from oxidation, since due to their vinyl ether bond, they are more susceptible to oxidation by reactive species (Goldfine 2010; Wallner and Schmitz 2011; Braverman and Moser 2012; Dean and Lodhi 2018; Pradas et al. 2019). Moreover, studies suggests that plasmalogens could protect unsaturated lipids from membranes from oxygen singlet oxidation (Broniec et al. 2011), and that plasmalogens can scavenging some ROS (Morand et al. 1988; Maeba et al. 2002; Skaff et al. 2008; Wu et al. 2019).

The plasma/serum lipidome, and aging

Sphingolipids

An increase in the concentration of SM may indicate a decrease in the risk of developing age-related diseases (Gonzalez-covarrubias et al. 2013). This is especially relevant in women, where higher levels of SM are associated with a lower risk of developing AD (Mas et al. 2012). Elevated levels of ceramides have been reported to regulate processes such as cell senescence, diabetes, insulin resistance, inflammation, neurodegenerative disorders, and atherosclerosis (Samad et al. 2006; Bismuth et al. 2008; Arana et al. 2010; Han et al. 2011; Spijkers et al. 2011; Guedes et al. 2017). Increased levels of SM, and decreased levels of ceramides can indicate greater protection against oxidative stress, and greater inclusion of SM in cell membranes, mechanisms that can be associated with healthy aging (Ichi et al. 2009; Corre et al. 2010; Gonzalez-covarrubias et al. 2013).

Phospholipids

The observed increase in the levels of PC, and PE in centenarians emphasizes that, in these individuals, there is greater conservation of cell membranes (Montoliu et al. 2014). PE is also a modulator of inflammation, and apoptosis (Gibellini and Smith 2010): PE 38:6, which is highly polyunsaturated, carries pro-inflammatory precursor molecules, such as those involved in the AA lipid pathway, from which eicosanoids are synthesized (Maskrey et al. 2007; Gonzalez-covarrubias et al. 2013). Decreased levels of this specific PE may suggest increased protection against lipid peroxidation, and the synthesis of pro-inflammatory molecules (Gonzalez-covarrubias et al. 2013). Another important lipid class for the inflammatory process is PI, from which AA is derived (Leslie 2015). Centenarians benefit from a perfect balance between pro- and anti-inflammatory eicosanoids, which translates into decreased inflammation due to their unique ability to modulate the AA metabolic pathway. This modulation can be proven with the increase in the levels of PI, and most of the PE species already reported (Montoliu et al. 2014). Also, changes in plasma phospholipid levels have been linked to impaired memory in elderly individuals (Mapstone et al. 2014).

Glycerolipids

Regarding TG, which may be associated with TG-rich lipoproteins, it is known that high levels of these molecules are related to an increased risk of developing some pathologies, such as atherosclerosis (Ramasamy 2016; Valaiyapathi et al. 2017). TG with long carbon chains, and highly unsaturated preferentially suffer peroxidation (Rhee et al. 2011). Since TG with long carbon chains undergo β-oxidation in mitochondria, and peroxisomes, and the metabolic capacity of these organelles decreases with age. The attenuation of the increase of unsaturated TG species may mean that, in physiological aging, there is a more efficient β-oxidation process (Gonzalez-covarrubias et al. 2013). Their levels increase significantly in women after menopause (Sugiyama and Agellon 2012). This increase is believed to be related to the decrease in estrogen secretion (Bjørnerem et al. 2004) since this hormone regulates lipoprotein metabolism (Knopp et al. 1994). Some TG species (with a higher number of carbons, and double bond content) have also been associated with a decreased risk of developing DM2 (Rhee et al. 2011).

FA

It is described in the literature that in adult individuals, and centenarians, there is a lower number of double bonds in FA than in the elderly. Therefore, in centenarians, lipids are not so prone to undergo lipid peroxidation, and this occurs at a similar level to that of young adults. So FAs such as C16:0, and C22:1n-9 appear to be promising biomarkers of extreme longevity (Jové et al. 2017; Pradas et al. 2019). Also, the ratio MUFA/PUFA is considered a biomarker of longevity (Caprari et al. 1999), and the conversion of polyunsaturated lipids to monounsaturated can be a hallmark of physiological aging (Gonzalez-covarrubias et al. 2013).

Ether lipids

Some studies show that plasmalogens are negatively related to several diseases in which there is an increase in oxidative stress, including age-related diseases, such as obesity, pre-diabetes, and DM2, cardiovascular diseases, cancer, and AD (Wallner and Schmitz 2011; Braverman and Moser 2012; Huynh et al. 2017; Meikle and Summers 2017; Dean and Lodhi 2018). Other studies indicate that there is a decrease in the levels of plasmalogens with age (Maeba et al. 2007, 2008). According to Pradas et al. (2019) in the centenarians there is an increase in the levels of PC ethers with alkyl groups, with shorter carbon chains, and lower number of double bonds, and a decrease in the levels of PE ethers with alkenyl groups, with longer carbon chains, and greater number of double bonds. The susceptibility of lipids to peroxidation increases as a function of the content of double bonds in carbon chains. Given this composition pattern, it can be concluded that the lipid ethers of the centenarians have a greater resistance to lipid peroxidation. The profile of ether lipids observed in centenarians indicate a physiological adaptation to an inherently low oxidative stress, which can be considered an optimized mechanism of exceptional longevity. This adaptation may be due to differences in the maintenance, and catabolism of poly and monounsaturated lipid species, and plasmalogens biosynthesis (Meikle and Summers 2017).

Conclusion

In a world in which life expectancy is increasing, it is necessary to understand the biochemical mechanisms that are related to aging. The use of metabolomic techniques has allowed the study of biochemical changes associated with various diseases, and physiological processes, such as aging. The development of these techniques, and the appearance of increasingly sophisticated devices allow for a better, and detailed analyses of the human metabolome in particular lipidome. It is possible to use plasma, serum, urine to search for new biomarkers, with the advantages of they are easily accessible since they are collected almost non-invasively.

During aging, there are disturbances in the metabolic pathways of the main biomolecules in the body. However, in physiological aging, these disorders are not so pronounced as in pathological aging, especially what concerns lipid metabolism. Based on that, one can say that the lipid signature varies with aging, and that in centenarians, descendants of centenarians, and elderly without age-related diseases, there are pronounced changes in the lipid profile. Table 3 summarizes putative biomarkers of longevity as well as putative biomarkers of healthy aging. In fact, healthy aging can be assessed by measuring molecules, in particular lipids, in biological fluids. Increases in PC ether 32:1, SM 24:1, SM 16:0, sterols and glycerolipids and decreased levels of and LPC 18:2, LPC 20:4 and sphingolipids are found in healthy aged individuals. These variations can be considered predators of healthy aging and reflects changes in metabolic lipidic pathways upon age and reflects in increased antioxidant capacity, and lower levels of lipid peroxidation, and inflammation.

Table 3 Brief summary of possible biomarkers of longevity and healthy aging
Full size table

Lipidomic methods of analysis are becoming more accessible, easier to implement and more sensitive. Thus, it is possible to identify well-known lipidic biomarkers in individuals in order to establish their molecular age. Lipidomic era makes possible to further investigate new biomarkers that characterize longevity, and that allow the development of therapies that may increase healthy life span, acting preventively avoid or slow the development of age-related diseases.