1. Introduction
Aging can be viewed as a multifactorial process stemming from the interaction of genetic and environmental factors, which include lifestyle. It is characterized by the onset of several age-related diseases (ARDs) such as dementia, osteoporosis, arthritis, diabetes, cardiovascular diseases (CVDs), neurodegenerative disorders, and cancer which, though not unique to old age, are nonetheless closely related to it. The physiological decline experienced by organisms over time is a key factor in increasing the risk of developing ARDs [
1,
2]. As human life expectancy expands, the number of patients with ARDs is rapidly increasing and will continue to mount, posing a serious challenge to healthcare systems globally [
3]. ARDs are also becoming a key social and economic problem [
4]. Progress in the understanding of genetic associations, particularly via genome wide association studies, has disclosed a substantial contribution of genes to human aging and ARDs [
4].
Ample evidence from several species indicates that the maximum age attainable is genetically determined and that multiple mitochondrial DNA polymorphisms are associated with longevity [
5]. The several theories of aging that have been devised over the past few decades have failed to provide an answer to the questions: “why do we age?”, “what can we do to live longer?”. However, the notion of aging as a complex multifactorial process has superseded previous constructs based on single factors [
6]. In fact, whereas some of the processes that characterize physiological aging can be explained by individual factors, no single theory can account for aging as a process. Several different molecular mechanisms linking aging and ARDs have been advanced [
3].
Various lines of research have demonstrated that telomerase activity and telomere length shortening play important roles in aging, Alzheimer’s disease (AD) and type 2 diabetes (T2DM) [
7,
8]; a reduced capacity for DNA repair and genomic instability are commonly seen in both aging and cancer [
4,
9,
10]; mitochondrial dysfunction is a hallmark of aging and ARDs, including CVDs and cancer [
11,
12]; and metabolic syndrome, diabetes, CVD, neurodegenerative diseases, and other ARDs are associated with chronic inflammation [
13,
14,
15].
In fact, most of the phenotypic characteristics of aging are the result of an age-related, low-grade, chronic proinflammatory status that has been designated “inflammaging” [
16], which is partly under genetic control. Moreover, up-regulation of inflammatory responses induces senescence, and inflammatory changes are shared by several age-related diseases. Oxidation-inflammation has thus been hypothesized to be the main cause of aging (oxi-inflamm-aging) [
17]. The aging-related chronic oxidative stress affects all cells, especially those of regulatory systems (nervous, endocrine and immune systems) and the communication among them, adversely affecting homeostasis and the maintenance of the health status. Since the redox state and functional capacity of immune cells are related to longevity, the immune system is also likely to be critically involved in aging and to affect its rate. Moreover, the role of the immune system in senescence is likely to be pervasive, as also confirmed by the demonstration that adequate dietary antioxidants improve immune function, reduce oxidative stress, and increase longevity [
18].
Human molecular processes are influenced by physiological pathways as well as exogenous factors, including dietary components. Since nutrients directly affect physiological changes, the diet has substantive effects. For instance, bioactive molecules capable of selective modulation of specific metabolic pathways affect the development/progression of cardiovascular and neoplastic disease. As bioactive nutrients are increasingly identified, their clinical and molecular chemopreventive effects are characterized and systematic analyses encompassing all “omics” technologies (transcriptomics, proteomics and metabolomics) are conducted to investigate their effects. Nutrigenomic knowledge regarding physiological status and disease risk is expected to lead to the development of improved diagnostic procedures and of therapeutic strategies targeting processes related to nutrition.
The state of the art of aging and nutrigenomics research and the molecular mechanisms underlying the beneficial effects exerted by bioactive nutrients on the main ARDs are reviewed herein.
3. Bioactive Nutrients and Nutrigenomics
The interactions among genomic and environmental factors is crucial in the development/progression of several human diseases. The diet is a key environmental factor with a prominent role in disease aetiology [
135].
The diet primarily meets the metabolic and energy requirements of body composition homeostasis. However, it may also enhance health through regulation of specific processes [
136], positively counteract inflammaging and the epigenetic changes associated with aging, and promote health [
137,
138]. Indeed, nutrients are considered as dietary signals capable of affecting both metabolic programming and cell homeostasis. Bioactive nutrients or chemopreventive molecules exert effects on human health and reduce disease risk through specific molecular mechanisms [
139,
140,
141]; experimental and epidemiological evidence emphasizes the potential of dietary components, both macronutrients (carbohydrates, protein, fat, and fibre) and micronutrients (antioxidant vitamins and minerals) as first-line interventions in the prevention and treatment of cancer and other diseases [
140,
141]. Nutritional research studies span numerous disciplines and are conducted at the molecular and genetic level [
142,
143]. For instance, a recent review of studies into the effect of the Mediterranean diet (MD) on inflammaging, cancer, and most ARDs has found that the MD and its individual bioactive nutrients modulate several interconnected processes involved in tumorigenesis, the inflammatory response (e.g., free radical production, NF-κB activation, and the expression of inflammatory mediators), and the eicosanoid pathway. In particular, the authors highlight the evidence indicating that the MD can affect the balance between pro- and anti-inflammaging and some emerging topics, such as the maintenance of gut microbiota homeostasis and the epigenetic modulation of carcinogenesis through specific microRNAs [
144]. Moreover, a parallel randomized trial has investigated the effect of a healthy diet on inflammaging and its consequences on the prevention of age-related decline in European elderly individuals [
145].
The effects of food at the genetic and epigenetic level are examined by two new approaches, nutrigenetics and nutrigenomics, which assess the influence of dietary components on health and disease onset, progression, and treatment. Nutrigenetics examines how genetic variation affects the response of an organism to a given diet, evaluating the risks and benefits of specific diets and dietary components and formulating “personalized nutrition” recommendations. Nutrigenomics investigates how nutrients affect gene expression and downstream processes [
146].
Nutrigenomics and nutrigenetics clearly straddle multiple research fields that span from nutrition to bioinformatics, molecular biology, genomics, functional genomics, epidemiology, epigenomics, transcriptomics, metabolomics, proteomics, lipidomics, and the microbiome [
147].
To an extent, nutrigenomics approaches pharmacogenomics, which involves the systematic study of the effect of drugs on the genome [
148]. However, whereas drugs are pure compounds, acting with affinity and selectivity on a limited number of biological targets through administration of precise and low doses, nutrigenomics addresses the complexity and variability of the diet, where some nutrients may be consumed in high albeit non-toxic concentrations (from µmM to mM) and may also bind to targets with different affinities and specificities [
149,
150].
Eating patterns influence gene, protein expression and metabolism and may thus be considered as endogenous cellular mediators. Once it is absorbed at the cell level, a nutrient is capable of interacting through specific signalling pathways, and even small changes in its structure may involve differential activation of metabolic steps. Fatty acids and their degree of carbon chain unsaturation are a valuable example, since
n-3 polyunsaturated fatty acids promote anti-inflammatory pathways, whereas
n-6 polyunsaturated fatty acids induce synthesis of proinflammatory molecules; in addition,
trans fatty acids increase plasma LDL-cholesterol [
151,
152,
153] whereas
n-3 polyunsaturated fatty acids do not have this effect.
It is likely that several dietary compounds exert their protective and restorative action through modulation of distinct signal transduction pathways. Nutrients have been found to affect gene expression as a consequence of a direct interaction with transcription factors [
154]; for instance fat-soluble ligands, such as vitamin A/D, activate their cognate nuclear receptors for ligand-dependent transcriptional regulation. Population studies based on dietary questionnaires supply useful data to relate dietary intake to phenotypes and to the risk of developing a number of diseases [
155]. Moreover, actual nutritional intake can be monitored by measuring specific molecules in blood, urine, fat, and tissues, an approach that can also enable identification of the nutritional biomarkers that connect nutrition and health. Altered serum lipids (e.g., cholesterol, triglycerides), increased blood pressure, and reduced insulin sensitivity are common predictors of diet-related diseases. A broad biomarker panel, rather than a search limited to single markers, would be able to provide exhaustive information to characterize the health status of individuals.
The application of “omics” technologies to nutrition and bioinformatic data analysis allows integrating information from the closely interconnected fields of transcriptomics, proteomics, and metabolomics, to identify specific differences related to nutritional habits and to investigate the mechanisms that cause changes [
147,
148,
156,
157,
158,
159,
160,
161].
The risk of developing disease is partly under genetic control. Nutrigenetics investigates the genetic variations induced by individual nutrients, for instance by relating single nucleotide polymorphisms and point mutations in DNA sequences to diet responsiveness [
162,
163]; population differences in single nucleotide polymorphisms can help risk assessment and prediction, enabling formulation of lifestyle recommendations. Phenylketonuria has been the first case of a condition induced by a single gene defect to be successfully treated with a dietary, i.e., nutrigenetic, intervention, namely a low-phenylalanine diet [
164].
Nutriproteomics, a recent branch of proteomics, studies protein structure and function and protein-protein interactions to identify the molecular targets of dietary components [
165,
166,
167,
168,
169,
170,
171,
172,
173,
174,
175,
176,
177,
178,
179,
180,
181,
182,
183,
184,
185,
186,
187,
188,
189,
190,
191,
192,
193,
194,
195]. Here, too, the goal is to detect differences in protein patterns induced by given interventions, for instance a dietary treatment. In turn, proteomics analyzes the effect of dietary components at various levels, investigating peptides as bioactive markers and seeking information on nutritionally relevant biological pathways in view of the development of dietary interventions to be applied in the clinic. Proteomics, also combined with gene expression analysis, has been used in cancer prevention studies to identify novel biomarkers [
196,
197,
198,
199,
200,
201,
202,
203,
204,
205,
206].
A quantifiable change connecting a normal or pathological condition to modulation of mRNA, a protein, or the concentration of a metabolite can be used as a molecular biomarker. In particular, a protein concentration is a practical biomarker and a valuable diagnostic tool, due to its reproducible and accurate determination [
207]. The dietary levels of most nutrients are only weakly biologically active and probably have several targets. When a biomarker is tested, the timing of its response(s) should be considered according to the nutrient’s bioavailability and bioefficacy. Notably, although biomarkers may correlate with nutrient intake, their modulation may in fact be the result of a more complex process including intake, absorption, metabolism, and excretion, as well as environmental factors and genetic predisposition. Based on these considerations, successful investigation may well need a combination of proteomic biomarkers and information from other “omics” technologies.
The levels of individual metabolites can be considered as the final step in a biological process that is influenced by genetic and environmental factors, including, critically, nutritional intake.
The metabolome (from genome) is the complete set of metabolites in an organism, and metabolomics studies classify and quantify them individually in a biological fluid, cell culture, or tissue sample. Their levels, determined by analytical methods, supply information on how enzymes and the other functional proteins affect cellular homeostatic mechanisms [
180]. Nutrients can interact directly with our body at the level of organs, cells, and molecules. They usually come in complex mixtures, where the amount of each compound and its interaction with multiple components are crucial, since they influence bioavailability and bioefficacy. Metabolomics allows systematic investigation of small organic molecules, and in conjunction with nutrigenomics establishes how those molecules can reflect the effects of different diets [
182]. A key goal of nutritional metabolomics is to detect and identify all endogenous human metabolites and exogenous components consumed through food that are found at least transiently in human body fluids. The development of metabolite panels related to specific nutrition states would be able to characterize physiological and pathological conditions more exhaustively than dosing of a single molecule, and such information could be integrated with the data obtained with the other “omics” technologies [
147,
183,
184,
185].
A further approach that deserves to be mentioned in relation to the role of nutrients in health and disease is nutritional epigenetics, which endeavours to explore gene-diet interactions and can provide information on the role of nutrition in aging and ARD development [
137,
208]. Epigenetic traits are defined as heritable DNA modifications that regulate chromosome architecture and modulate gene expression, without changes in the underlying bp sequence, ultimately determining phenotype. DNA methylation and post-translational histone modifications are well-established levels of epigenetic regulation. Epigenetic phenomena are critical from embryonic life to old age, and epigenetic pattern aberrations are recognized as aetiological mechanisms in several ARDs including cancer, CVD and neurodegenerative disorders. Nutrients can act as sources of epigenetic modifications and can regulate the site where they take place. Nutrients involved in one carbon metabolism, i.e., folate, vitamin B12, vitamin B6, riboflavin, methionine, choline, and betaine, are involved in DNA methylation through modulation of the levels of the universal methyl donor
S-adenosylmethionine and of the methyltransferase inhibitor S-adenosylhomocysteine. Other nutrients and bioactive components of food—e.g., retinoic acid, resveratrol, curcumin, sulphoraphane and tea polyphenols—affect epigenetic patterns by modulating the levels of
S-adenosylmethionine and
S-adenosylhomocysteine or the enzymes that catalyze DNA methylation and histone modifications. Aging and ARDs are associated with profound epigenetic changes, even though it is unclear whether such changes are programmatic or stochastic [
209]. Future work in this field will need to characterize the epigenetic pattern of healthy aging, to learn which nutritional measures can contribute to maintain or achieve it.
A large number of metabolically active metabolites are produced by the microbiome, particularly the gut microbiome, which may be viewed as a complex organ capable of influencing host health. Recent studies suggest that it should actually be regarded as an “immune system” that can promote health but sometimes initiates disorders such as inflammatory bowel disease, metabolic syndrome, obesity-related disease, diabetes, liver disease and colorectal cancer [
210]. Changes in the diet may exert profound effects on the microbiome and are capable of altering the overall bacterial composition. Interactions between the microbiome and the metabolism of dietary components such as phosphatidylcholine and carnitine have been reported to modulate the CVD risk [
147]. Moreover, the microbiome could constitute a novel therapeutic opportunity, because in some cases it may be used to detect gut-related diseases earlier than conventional diagnostic workups. In the future, this information could be harnessed to stratify patients more accurately and for more effective treatment [
210]. These research fields are still in their infancy. It is therefore critical to clarify the relationship among genetics, diet, microbiome, and health risk.
Finally, useful data are expected to come from systems biology, a highly cross-disciplinary approach to biology research; in turn, biochemical systems biology includes and combines genomics, biochemistry, and molecular biology integrating them with mathematical and computational analysis, engineering practices, and “omics” technologies such as transcriptomics, proteomics and metabolomics [
211]. A number of experimental strategies are combining quantitative measurements of cell components (mRNA, proteins, and metabolites) using mathematical and computational models [
147,
212,
213,
214,
215,
216,
217,
218,
219,
220,
221,
222,
223,
224,
225,
226,
227,
228,
229,
230,
231,
232]. Such high-throughput technologies are providing a huge amount of functional genomic data that are expected to deliver breakthrough in aging research. According to Özdemir and colleagues, now that the goals and tasks of nutrition science and nutrigenomics have become clearly established, it would be desirable to achieve the integration of four key domains that are naturally connected—agrigenomics, nutrigenomics, nutriproteomics, and nutrimetabolomics—which address complementary issues in relation to individual differences in response to food-related environmental exposure. Although the knowledge and findings of these four omics have still failed to be integrated, they have a very high innovation potential. In the future, personalized nutrition interventions are expected to benefit from the integration of life sciences funding, research, and practice from “farm to clinic to supermarket to society,” and from “genome to proteome to metabolome” [
233].
4. Molecular Pathological Epidemiology
Epidemiological research typically investigates the factors that are associated with the overall risk of developing certain diseases, including the relationship between exposure and a disease entity in population-based cohorts, whereas pathology research traditionally explores aetiology, development, and histopathological and molecular characteristics to predict prognosis and response to treatment. The merging of the two approaches through the incorporation of molecular pathology into epidemiological research has created a new population health science field that straddles several research areas, which has been designated Molecular Pathological Epidemiology (MPE) [
234,
235]. Its close relationship to both aetiology and prognosis involves that MPE pursues to gain a greater understanding of how particular exposures influence disease risk through the search and evaluation for molecular pathological markers. Disease processes are influenced by a wide range of exogenous (e.g., acquired genetic and epigenetic alterations, diet, lifestyle, smoking, medications, microorganisms) as well as inherent factors (e.g., germline genetic variations, sex, ethnicity) that induce significant interindividual variability in all phases of the disease process [
236]. Compared with conventional approaches, where patients diagnosed with similar symptoms or disease manifestations are assumed to make up a homogeneous group and to share similar causative factors, MPE employs molecular pathological signatures to refine patient categorization and identify subgroups that share more homogenous, to gain insight into disease heterogeneity respect to both aetiologies and pathogenic process. By this approach, the MPE multidisciplinary method explores whether exogenous and endogenous factors are associated with differential molecular signatures and disease subgroups [
235,
236].
The notion of MPE was first introduced by Ogino and Stampfer in a comment on a case–control study of body mass index (BMI) and the risk of colorectal cancer (CRC) in relation to tumour microsatellite instability (MSI) (MSI-high vs. microsatellite stability [MSS] status) [
234,
237]. The authors demonstrated that pre-diagnosis BMI was associated with an increased risk of developing CRC, and that the excess risk associated with BMI was limited to MSS tumours [
237].
A large number of investigations have subsequently been identified as belonging to MPE. Several of these have identified links between diet, lifestyle, environmental exposure, and alteration of molecular patterns characterized as distinctive features of specific diseases.
Over the past few years, the molecular changes related to the risk of developing CRC have extensively been explored. MPE findings suggest that the effects on outcome of alterations in the WNT signalling pathway and cadherin-associated protein β 1 (CTNNB1 or β-catenin) are modified by BMI and physical activity. In particular, CTNNB1 activation was associated with longer CRC-specific survival and overall survival among obese patients, whereas post-diagnosis physical activity was associated with longer CRC-specific survival only for patients with negative nuclear CTNNB1 status [
238]. According to another study, obesity and physical inactivity are associated with a higher risk of CTNNB1-negative CRC, but not of CTNNB1-positive CRC, suggesting that the energy balance and the metabolic state exert effects on a specific carcinogenesis pathway that is less likely to be dependent on WNT/CTNNB1 activation [
239]. An important role for BMI has also been found in relation to tumour TP53 mutations, which are key factors in CRC development, and on fatty acid synthase (FASN), which is overexpressed in some colon cancers and in involved in the energy metabolism of fatty acids. MPE data support a dual role for TP53 alterations in cell-cycle deregulation and cell autonomy in relation to the energy balance [
240], while FASN-negative and FASN-positive tumours have been reported to be associated with a significantly different CRC risk [
241]. Moreover, an excess energy balance may influence the immune and inflammatory status, suggesting an association of BMI with a heightened CRC risk regardless of the level of the lymphocytic response to the tumour [
242].
Dietary compounds affect specific pathways related to cancer development. High alcohol consumption increases the CRC risk because one-carbon metabolism triggers a DNA methylation reaction, which affects the CpG island methylator phenotype (CIMP), with tumour epigenetic features modulating the cancer risk [
243,
244]. Preclinical and epidemiological studies have provided evidence of a protective effect of vitamin D against CRC [
245], whereas a higher calcium intake has been associated with a lower risk of developing CRC, especially distal colon cancer; the overall inverse association was linear and did not differ in relation to intake source [
246]. Similarly, a high intake of marine ω-3 polyunsaturated fatty acids (ω-3 PUFAs; including eicosapentaenoic acid, docosahexaenoic acid and docosapentaenoic acid) has been documented to have an antineoplastic action. An increased intake of marine ω-3 PUFAs after CRC diagnosis may still confer benefits [
247].
A large role for microRNAs has also been documented by MPE studies. The expression level of miR-21 is associated with a worse CRC clinical outcome [
248], whereas let-7 family microRNAs suppress adaptive immune responses, contributing to immune evasion by the tumour [
249]. Circulating miR-21-5p and miR-126-3p have been shown to play a role as dynamic biomarkers of systemic inflammatory/angiogenic status, and could have an even greater role in managing T2DM [
250]. Furthermore, microRNAs have an established potential in the diagnosis and prognosis of several cancers and of pollution exposure [
251]. For example, a pool of deregulated circulating and tissue microRNAs with biomarker and therapeutic potential has been identified in malignant mesothelioma, a lethal cancer related to asbestos exposure [
252].
MPE findings have also linked leukocyte telomere length (LTL) and genetic variants in the telomerase reverse transcriptase gene region to the risk of pancreatic cancer [
253]. LTL shortening is found in a number of ARDs, including T2DM. Analysis of its possible association with mortality was analyzed in this study. Recently, time-dependent mortality risk stratification has allowed demonstrating that T2DM duration and LTL combined with clinical parameters can provide additive prognostic information on mortality risk in these patients [
8].
Finally, several lines of MPE evidence have confirmed the link between microbiota and disease. For instance, the abundance of
Fusobacterium nucleatum, which increases gradually from the rectum to the caecum, reflects the pathogenic influence exerted by the gut microbiota on neoplastic and immune cells, and may promote CRC growth by inhibiting T-cell-mediated immune responses against the tumour. A greater amount of
F. nucleatum DNA in CRC tissue is associated with shorter survival and may serve as a prognostic biomarker [
254,
255].
It is highly likely that the MPE data reviewed above can be harnessed to develop new disease prevention and early detection approaches, and that the molecules and pathways identified by MPE research can be used to devise new treatments by targeting altered regulatory mechanisms.