| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,1
* Nestlé Research Center, Vers-chez-les-Blanc, Lausanne, Switzerland; and
Center for Integrative Genomics, University of Lausanne, Lausanne, Switzerland
1 Correspondence: The Scripps Research Institute, 10550 North Torrey Pines Rd., MEM 275, La Jolla, CA 92037. E-mail: dmmutch_sci{at}hotmail.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONA COMPLEX PARTNERSHIP: DISEASE...NUTRIGENOMICSNUTRIGENETICSOUTLOOKREFERENCES |
|---|
Key Words: coronary heart disease • diabetes • integrative metabolism • metabolomics • nutrient-gene interaction • polymorphism • systems biology
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONA COMPLEX PARTNERSHIP: DISEASE...NUTRIGENOMICSNUTRIGENETICSOUTLOOKREFERENCES |
|---|
. This issue can now be eloquently resolved with modern technologies that permit both factors to be considered concurrently, as demonstrated by profound and recent advances in our understanding of health and disease.
These modern technologies have capitalized on the completion of the sequenced human (2
, 3)
, mouse (4
, 5)
, and rat (6)
genomes. The vast quantity of information stemming from these achievements has not only prompted the active development of comprehensive analytical platforms and their appropriate data mining software, but have also revolutionized our approach to understanding health and disease. Gone are the days when one is constrained to study only a single gene or gene product in a biologically out-of-context situation; rather, emerging technologies for investigating all genes, proteins, and metabolites provide the means to scrutinize this information in relation with all other biological elements present in a given sample. As a result, this has led to the creation of popular terms, ending with the Greek suffix ‘ome, meaning "complete" or "all," describing the global analysis of genes (genomics), mRNA (transcriptomics), proteins (proteomics), and metabolites (metabolomics). Although >50 ‘omic terms have been described in the literature, it is the 4 aforementioned terms that are widely established and accepted (7)
. Genomics can be simply defined as the approach describing the mapping, sequencing, and analysis of all genes present in the genome of a given species. The awesome quantity of information comprised within a given genome has led to genomic offspring, each with a specific goal. The predominant genomic derivative has been coined functional genomics, which aims to uncover both the functional roles of different genes and how these genes interact with, and/or influence, each other in the functional network underlying health and disease. However, other extensions such as structural genomics and comparative genomics are rapidly becoming popular methods to further annotate the genome. Transcriptomics, using either cDNA or oligonucleotide microarray technology, describes the approach in which gene expression (mRNA) is analyzed in a biological sample at a given time under specific conditions. This field is certainly the most widely used of the ‘omic technologies, in part due to the intense activity of such initiatives as the "minimum information about microarray experiments" (MIAME)2
(8
, 9)
and the "gene expression omnibus" (GEO)3
(10
, 11)
to standardize experimental procedures and the storage of data, respectively.
Studying the other facets of biological complexity, such as proteins and metabolites, is especially important with the realization that the principle of one gene leads to one protein leads to one metabolite is a simplistic and often incorrect notion, as experimentally demonstrated (12)
. Therefore, proteomics aims to characterize all proteins in a biological sample, including their relative abundance, distribution, post-translational modifications, functions, and interactions with other biological molecules (7)
. Metabolomics, or metabonomics, can be simply defined as the quantitative analysis of all metabolites in an isolated cell system, tissue, or biological fluid. Although these two terms tend to be used interchangeably, metabonomics is generally associated with drug discovery and refers to a systems approach to metabolite profiling, whereas the more widely used term metabolomics refers more specifically to cell-based metabolite profiling. For the purpose of this review, metabolite profiling will be referred to as metabolomics from here on. In contrast with transcriptomics, proteomics and metabolomics are not yet routine and standardized procedures, and continue to face challenges such as sample preparation, technological sensitivity, lack of standardized statistical methods and public databases (13
14
15
16
17
18)
. Nevertheless, their potential benefits for health management are undisputed and have fuelled current efforts to assess, utilize, interpret, and ultimately integrate these global technologies in order to define a phenotype characterizing health status.
These analytical platforms are now widely used by both the pharmaceutical and nutritional communities alike; however, whereas pharmaceuticals have a targeted approach aimed at restoring health, diet is a multi-parametric approach to preserve and/or optimize health. Indeed, the diet is comprised of a multitude of nutritional and chemical molecules each capable of regulating disparate biological processes, and thus cannot use an approach similar to the pharmaceutical industry, i.e., the "one drug one target" paradigm (19)
. Hence, nutrition is a true integrative science that is well positioned to benefit from the exploitation of novel technologies capable of assessing biological networks rather than single endpoints (20)
. Despite the powerful analytical platforms available for the analysis of genes, proteins, and metabolites, few examples have used a comprehensive and integrative approach to understand the influences of nutritional factors on metabolism (21
22
23
24)
. Rather, the great majority of intervention studies performed to date are platform-specific. Inasmuch as each of these analytical platforms, from genomics to proteomics to metabolomics, provides increasingly accurate information describing a given phenotype, it is the integration of these technologies that provides the optimal means to unravel the effects of a biological challenge on an organism; thus, the concept of systems biology (or integrated metabolism) (20
, 25
, 26)
. Indeed, an integrated metabolism approach yields an attractive and exciting future for both pharmaceutical and nutritional communities, and their quest to ameliorate health and prevent disease. Considering multiple facets of biological complexity within the confines of a single analysis provides a means to consider how specific genetic profiles respond to environmental challenges.
For the field of nutrition, this would encompass the ongoing efforts to understand the relationships between the genome and diet, currently termed nutrigenomics and nutrigenetics (27)
. Although these two concepts are intimately associated, they take a fundamentally different approach to understanding the relationship between genes and diet (Fig. 1
). Nutrigenomics aims to determine the influence of common dietary ingredients on the genome, and attempts to relate the resulting different phenotypes to differences in the cellular and/or genetic response of the biological system (28)
. More practically, nutrigenomics describes the use of functional genomic tools to probe a biological system following a nutritional stimulus that will permit an increased understanding of how nutritional molecules affect metabolic pathways and homeostatic control (29)
. Nutrigenetics, on the other hand, aims to understand how the genetic makeup of an individual coordinates their response to diet (28)
, and thus considers underlying genetic polymorphisms. In other words, nutrigenetics embodies the science of identifying and characterizing gene variants associated with differential responses to nutrients, and relating this variation to disease states (30)
. Therefore, both disciplines aim to unravel diet / genome interactions; however, their approaches and immediate goals are distinct. Nutrigenomics will unravel the optimal diet from within a series of nutritional alternatives, whereas nutrigenetics will yield critically important information that will assist clinicians in identifying the optimal diet for a given individual, i.e., personalized nutrition (28)
. Despite the immediate goals differing, the long-term goal of improving health and preventing disease with nutrition requires the amalgamation of both disciplines.
|
|
A COMPLEX PARTNERSHIP: DISEASE AND NUTRITION |
|---|
|
TOPABSTRACTINTRODUCTIONA COMPLEX PARTNERSHIP: DISEASE...NUTRIGENOMICSNUTRIGENETICSOUTLOOKREFERENCES |
|---|
. Indeed, of the nearly 1000 genes that have so far been associated with human disease, 
97% result in monogenic diseases, i.e., a single dysfunctional gene is responsible for disease (32)
. Modifying the consumption of certain dietary compounds can prevent some monogenic diseases, such as galactosemia and phenylketonuria. Galactosemia arises from a rare recessive trait in galactose-1-phosphate uridyltransferase (GALT), leading to the accumulation of galactose in the blood and increasing the risk of mental retardation (33)
. Phenylketonuria is characterized by the defective phenylalanine hydroxylase (PAH) enzyme, resulting in the accumulation of phenylalanine in the blood that drastically increases the risk of neurological damage (34)
. Galactose-free and phenylalanine-restricted tyrosine-supplemented diets are a means to nutritionally treat these monogenic diseases, respectively (35)
.
In contrast, those diseases reaching epidemic proportions in the Western world, such as cancer, obesity, diabetes and cardiovascular disease, often arise from dysfunctional biological networks, and not a single mutated gene (i.e., polygenic diseases). For example, the recent meta-analysis by Segal et al. of nearly 2000 microarray studies spanning 22 different tumor types demonstrated that no single common gene mutation is responsible for the onset of these tumors, but that shared functional modules exist between various cancer types (36)
. Thus, dietary intervention to prevent the onset of such diseases is a complex and ambitious goal that requires not only knowledge of how a single nutrient may affect a biological system, but also how a complex mixture (i.e., diet) of nutrients will interact to modulate biological functions (15)
.
|
NUTRIGENOMICS |
|---|
|
TOPABSTRACTINTRODUCTIONA COMPLEX PARTNERSHIP: DISEASE...NUTRIGENOMICSNUTRIGENETICSOUTLOOKREFERENCES |
|---|
. Inasmuch as these studies are constantly improving the validity and accuracy of our knowledge of these disease states, they all suffer from a similar inherent quandary; namely, that these studies have been designed using current dogma. Thus, the ability to identify novel mechanisms of action and/or biomarkers that may more accurately depict health status is greatly diminished, as only accepted and characterized endpoints are examined. As demonstrated in the following sections, nutrigenomic technologies resolve this tunnel vision by providing a means to identify previously unrecognized and unanticipated molecular endpoints (7)
.
Plant sterols and sitosterolemia
The power of ‘omic protocols to compare gene expression profiles to understand spontaneous differences in phenotype due to disease was creatively extended by inducing differences in phenotype through molecular intervention. This was eloquently executed in a study that simultaneously identified a mechanism for the regulation of sterol uptake in the intestine and the molecular basis for a genetic disease: sitosterolemia (38)
. Until this study, no mechanism could explain why plant sterols, which are structurally similar to cholesterol, are normally absorbed in negligible amounts, nor why sitosterolemic individuals absorb phytosterols to high concentrations. In the study reported by Berge et al. (38)
, mice were treated with a lipid metabolism-altering drug and the global mRNA expression profiles of various tissues were compared with that of normal mice. Differentially expressed genes were evaluated and an unknown gene was found by bioinformatic tools to be homologous to the ATP binding cassette (ABC) family of genes, which include genes coding cellular lipid transport proteins. Indeed, defects in another member of this family (ABCA1) are the basis of the poor cholesterol delivery to HDL that underlies Tangiers disease (39)
. Through the use of a variety of in silico techniques, Berge et al. concluded that the proteins produced from these newly discovered genes, ABCG5 and ABCG8, were responsible for the regulated reverse transport of newly absorbed animal and plant sterols out of the apical surface of intestinal cells. Exploring public gene databases, a human homologue of the putative mouse transporter was identified, cloned and used to screen sitosterolemic humans. In all cases, individuals diagnosed with this disease were found to contain mutations in these genes and thus lack the machinery responsible for the selective and controlled transport of cholesterol, and therefore hyper absorb various sterols (including plant sterols). This study illustrated many of the strengths of a nutrigenomic approach, from identifying phenotypically important genes using microarray technology, to database polling, to deducing structure function from sequence comparison, to characterizing individual variation (polymorphism) linked to health.
Long-chain polyunsaturated fatty acids and carcinogenesis
Currently, no human in vivo data exists in which nutrigenomic technologies were used to elucidate those molecular mechanisms modulated after the consumption of long chain polyunsaturated fatty acids (LC-PUFA); however, several animal and in vitro microarray studies have demonstrated that these dietary compounds clearly regulate the cellular machinery in a variety of healthy and carcinomic tissues. Indeed, there is an active interest to comprehensively unravel the biological functions of this class of dietary compounds, as epidemiological studies have demonstrated that the consumption of LC-PUFA beneficially affect physiological processes including growth, neurological development, lean and fat mass accretion, reproduction, innate and acquired immunity, infectious pathologies of viruses, bacteria and parasites; and the incidence and severity of virtually all chronic and degenerative diseases including cancer, atherosclerosis, stroke, arthritis, diabetes, osteoporosis, and neurodegenerative, inflammatory, and skin diseases (40
41
42
43
44
45
46
47)
. Their disparate roles on health can be associated with their ability to serve as or be processed to ligands for a number of critically important transcription factors, including peroxisome proliferator activated receptors, hepatic nuclear factor-4, sterol regulatory expressed binding proteins, liver-X-receptor 
, and nuclear factor 
ß, which are able to disseminate the biological activity of LC-PUFA throughout the various tissues of the body (48
, 49)
. Additionally, fatty acids have been found to modulate protein activity (50)
and localization (51)
. Furthermore, LC-PUFA are associated with alterations in lipid metabolism and the remodeling of several lipid species, such as phospholipids (PLs), triglycerides (TGs), and cholesterol esters (CEs) (52
53
54
55)
. Indeed, the extensive influence of dietary LC-PUFA on metabolism, in addition to the many epidemiological studies demonstrating their beneficial health effects, renders this class of nutrients ideal for study by a nutrigenomics approach.
Carcinogenesis is a process composed of multiple stages in which gene expression, and protein and metabolite function begin to operate aberrantly (56)
. In the post-genomic era, the cellular events mediating the onset of carcinogenesis, in addition to their modulation by dietary factors, has yielded important information in our understanding of this disease. The most commonly studied dietary lipids, with respect to their effects on cancer, are LC-PUFA (57)
. Previous findings have found that fish oil, rich in omega-3 fatty acids, inhibit the growth of colonic tumors in both in vitro and in vivo systems (58
59
60)
. Modern nutrigenomic technologies, coupled with bioinformatics, have begun to reveal the complexity of LC-PUFA signaling (Fig. 2
); however, current lipid metabolic pathways have proven to be relatively underdeveloped (57)
. Thus, the deluge of information derived from microarray technology provides a potential method to further annotate the biological functions of these common dietary compounds and complete the poorly defined pathways. Preliminary microarray work has found that LC-PUFA mediate the functions of several transcription factors, cell-cycle regulatory genes, RNA transcription processes, prostaglandin synthesis, and inducible nitric oxide synthase and related proinflammatory genes (61
62
63)
; however, it remains unclear to what extent LC-PUFA are directly and indirectly involved in modulating these biological processes. Further work, in which specific inhibitors, small inhibiting RNA technology and alternate analytical (proteins, metabolites) platforms are incorporated, can clarify the biological functions mediated by dietary lipids. Although clearly at the beginning, nutrigenomics offers an attractive means to decipher the molecular networks regulated by LC-PUFA, and promises an exciting future for nutritionists and biochemists alike.
|
Rosiglitazone and type 2 diabetes mellitus
More recently, metabolomic technologies have established their ability to identify novel and potential biomarkers characterizing unseen tissue toxicity. Currently, no concrete examples exist in the field of nutrition; thus the following demonstrates the power of these modern technologies to identify subtle metabolic changes in response to a drug compound. As diabetes is characterized by metabolic changes, the goal is to address these changes while minimizing undesirable side effects. Using an appropriate mouse model, it was shown that lipid metabolism in the liver, adipose and heart tissues was altered after treatment with rosiglitazone, a drug commonly used by type 2 diabetic patients (64)
. The authors revealed that fatty acid synthesis was markedly increased in the aforementioned tissues, leading to lipotoxicity; however, this increased synthesis was not reflected in the plasma lipid profile, where cholesterol esters and triglycerides were decreased. Therefore, examining standard biomarkers would have falsely indicated that rosiglitazone decreases de novo lipid synthesis. In reality, lipid synthesis was increased locally and the processes regulating lipid import/export into tissues were apparently modified, concealing the side effects stemming from chronic rosiglitazone therapy. The acyl composition (namely 16:1n7 and 18:1n7) of plasma lipids reflected the increased hepatic lipid synthesis, suggesting that the use of these fatty acids as potential biomarkers for hepatic lipotoxicity should be further explored.
Likewise, metabolomic profiling after the consumption of food compounds promises to identify biomarkers capable of predicting nutrition-related diseases; yet, the multidimensionality of diet-induced metabolic changes suggests that identifying these biomarkers may prove challenging. In contrast to medicinal compounds, such as rosiglitazone, nutrients will have subtle effects on the metabolome; however, well-controlled experimental designs coupled with the appropriate statistical models may prove sufficiently robust for the identification of biomarkers indicative of nutrition-related diseases.
|
NUTRIGENETICS |
|---|
|
TOPABSTRACTINTRODUCTIONA COMPLEX PARTNERSHIP: DISEASE...NUTRIGENOMICSNUTRIGENETICSOUTLOOKREFERENCES |
|---|
7.7 million SNPs for which 4.8 million can be linked to discrete locations on the whole human genome map (65)
. The importance of genetic variation in the need for, and physiological response to, particular nutrients is already well described in principle, and examples continue to emerge. Furthermore, it is already apparent that there are many polymorphisms that influence risk in cardiovascular disease and diabetes. SNP analysis provides a powerful molecular tool for investigating the role of nutrition in human health and disease, and their consideration in clinical, metabolic and epidemiological studies can contribute enormously to the definition of optimal diets (66)
.
Clearly, the potential influence of nutrients on the genome cannot be underestimated; however, there will be certain genes (often referred to as "candidate" or "susceptibility" genes) that will have a heightened sensitivity to diet (67
68
69)
. The identification of these candidate genes often meets one or more of the following conditions:
Using such criteria, several studies have correlated a small number of important SNPs to particular phenotypes, and many new trials are being planned to examine the correlation between subsets of the millions of newly discovered sites of genetic variation (66)
. It is important to note that uniform standards for running nutritional epidemiological studies do not exist, thus the biological correlations made between diet and genes in the following sections stem from a multitude of studies differing in both cohort size and the severity of background genotyping. Nevertheless, these important studies linking genetic variation and dietary intervention to insulin resistance, type 2 diabetes mellitus (T2DM) and CVD demonstrate how individual genotyping will assist with managing diets for disease prevention.
Coronary heart disease
Coronary heart disease (CHD) is one of the most prevalent diseases in Western society, and has been attributed the cause of one in every five deaths in America each year (70)
. Epidemiological studies and long-term outcome trials have established a clear link between lipids and the development of CHD (71)
, leading The National Cholesterol Education Program to specify two lipoproteins as primary targets in combating CHD: low density lipoproteins (LDL) and high-density lipoproteins (HDL). Lipoproteins are macromolecular complexes comprised of a lipid-rich core surrounded by a surface monolayer containing phospholipids, unesterified cholesterol, and specific proteins. Various complexes exist and are defined based on their composition, which ultimately affects their density: chylomicron (high lipid to protein ratio, highest in TG as % of weight), VLDL (very low density lipoprotein, 2nd highest in TG as % of weight), IDL (intermediate density lipoprotein), LDL (highest in CE as % of weight), and HDL (high protein to lipid ratio). Clinical data have demonstrated the efficacy of lipid-lowering modes of therapy for the prevention of CHD. The first line of therapy, termed therapeutic lifestyle changes, is composed of alterations in the individual’s diet and physical activity, coupled with reductions in smoking and weight if applicable. The second line of therapy involves pharmaceutical compounds, such as statins, which have been found to efficaciously inhibit the activity of hepatic 3-hydroxy-3-methylglutaryl CoA reductase (HMG-CoA reductase) and ultimately decrease the levels of circulating cholesterol (72
, 73)
. However, it is well known that individuals respond differently to the aforementioned therapeutic interventions and that subsets of the population, in response to a given therapy, are unable to achieve the recommended levels of lipids (71)
. These varying degrees of response have been accredited to genetic differences in the population, thereby emphasizing the importance of identifying those genes that play a critical role in the onset or protection from CHD, and unravel their interactions with common dietary compounds (74)
. To date, several candidate genes, and their common SNPs, have been identified and preliminary evidence supports the notion that genetic variations in such genes as cholesterol ester transfer protein (CETP), lipoprotein lipase (LPL), hepatic triglyceride lipase (HL), LDL-receptor, apolipoprotein E (APOE), apolipoprotein A1 (APOA1), ATP binding cassette transporter A1 (ABCA1), and lecithin-cholesterol acyltransferase (LCAT) will alter individual sensitivity to developing CHD (75
76
77
78
79
80
81
82)
. Additionally, a nutrigenetic approach has begun to reveal that several of the aforementioned genes and their polymorphisms, such as APOA1 and LPL, are susceptible to dietary intervention and may modulate the onset of CHD.
Apolipoprotein A1 (APOA1)
HDL plays a critically important role in the prevention of CHD, predominantly through the transport of cholesterol from peripheral tissues to the liver for excretion in bile. APOA1 is a major component of plasma HDL (83)
and is thought to contribute to the transport of cholesterol from ABCA1 to HDL particles (84)
. Recently, it has been found that HDL levels are sensitive to dietary factors, such as PUFA. In the context of the Framingham Heart Study (85)
, individuals with a polymorphism in the APOA1 gene promoter region (–75 G/A) were found to differentially respond to dietary PUFA (86)
. In brief, the authors found that individuals with the A allele showed an increase in HDL levels following an increased consumption of PUFA. In contrast, those with the more common G allele showed an inverse relationship between HDL levels and PUFA consumption. Furthermore, the authors revealed that differences in sex also mediate the response. Indeed, men did not show a relationship between HDL and PUFA consumption, irrespective of their APOA1 polymorphism. However, women with the A allele and a high intake of PUFA (>8% of energy derived from PUFA) were found to have high levels of HDL. In comparison, women with the common G allele were found to have high HDL levels when fed low quantities of PUFA (<4% of energy derived from PUFA). This example highlights the complex interactions that arise between dietary molecules and genotype, and the additional influence that gender may have.
Lipoprotein lipase (LPL)
LPL is responsible for catalyzing the hydrolysis of TGs present in circulating chylomicrons and VLDL particles, providing nonesterified fatty acids and 2-monoacylglycerol for tissue utilization. Dysfunctional LPL has been associated with various disease states, such as atherosclerosis, chylomicronaemia, obesity, Alzheimer’s disease, and the dyslipidemia related with diabetes and insulin resistance (87)
. Furthermore, several polymorphisms have been described that disrupt normal LPL function and contribute to the premature development of CHD, primarily through the increased levels of circulating TGs (88)
. Indeed, several of the common LPL polymorphisms described by Merkel et al. have recently been established to influence circulating lipid levels in pregnant women, whom are often characterized with high levels of circulating TGs and increased total cholesterol. Although TG levels were unaffected, certain LPL SNPs modulated HDL levels and may alter the susceptibility of pregnant women to developing CHD (89)
; however, further studies are required to definitively define a relationship between lipid levels, CHD, and pregnant women. Therefore, the importance of LPL in the whole-body regulation of lipid metabolism has been avidly demonstrated and merits further exploration.
One common LPL polymorphism, known as T495G HindIII, has been extensively examined and demonstrates the complexity of disease prediction associated with a single SNP. Indeed, preliminary indications suggest that this polymorphism may play a role in the onset of several important diseases, such as CHD, diabetes and obesity. This SNP has been associated with the higher plasma TG and lower HDL levels characteristic with the early onset of diabetes (90)
. Preliminary results have also suggested a positive association between the HindIII polymorphism and a predisposition to developing obesity (91)
. Finally, this polymorphism has been associated with variations in lipid levels and heart disease, and that these alterations were attenuated by such environmental factors as physical exercise (92)
and low calorie diets (93)
, reiterating the important interactions arising between lifestyle, nutrition, and disease. Although these associations are not conclusive, they do suggest that LPL variants play a critically important role in the regulation of whole-body lipid metabolism that may predispose an individual to the onset of several metabolic diseases.
A relationship was established between a low calorie diet and the circulating lipid profile in obese individuals with the HindIII polymorphism (93)
. Homozygotes (H2H2) were found to have significantly higher levels of plasma VLDL-TG and APOB than heterozygotes (H1). Caloric restriction reduced lipid levels in both H2H2 and H1 individuals to a point where no difference was observed between the groups. Although H2H2 individuals responded more strongly (larger decreases in plasma lipids) to the low calorie diet, these preliminary results identify an important relationship between LPL polymorphisms, function, and diet.
Insulin resistance and type 2 diabetes mellitus
T2DM is a metabolic disorder, stemming from either an insufficient secretion or impaired action of insulin, characterized by hyperglycemia, dyslipidemia, and associated with impaired carbohydrate, protein and lipid metabolism (94
, 95)
. The importance of lifestyle has been linked to the high incidence of T2DM in Westernized societies through several observational studies, which identified a higher rate of incidence in individuals leading inactive lives and/or with high levels of food consumption (94)
. Furthermore, since the initial study by West and Kalbfleish (96)
, diets high in fat and energy have been repeatedly shown to play a fundamental role in the onset of T2DM. However, evidence exists that both environmental and genetic factors are important in the etiology of T2DM. Recent mechanistic work using mouse models with various genetic modifications affecting lipid metabolism have conclusively demonstrated a relationship between lipids and disease state (97)
. Furthermore, the genetic manipulation of genes involved in fatty acid and TG synthesis have demonstrated that neither T2DM nor insulin resistance are monogenic diseases; rather, a myriad of susceptibility genes involved in regulating lipid metabolism and insulin sensitivity have been demonstrated to modulate the risk of disease onset and novel targets continue to be discovered (98)
. The following examples do not comprise the inclusive list of genes associated with T2DM (e.g., recent findings have identified adiponectin (99)
and resistin (100)
as possible susceptibility genes for T2DM); however, preliminary evidence suggests that those genes discussed below are responsive to diet, and thereby demonstrate important interactions between genotype and nutrition in an individual’s predisposition to T2DM.
Sterol response element binding protein (SREBP)
SREBP-1a and 1c, splice variants arising from alternate start sites, have been repeatedly shown to have an important role in fatty acid synthesis. Using transgenic and knockout mouse models, these membrane-bound transcription factors were found to directly activate the expression of more than 30 genes involved in the synthesis and uptake of cholesterol, fatty acids, TGs, and PLs (101
, 102)
. Two transgenic models leading to the overexpression of SREBP-1c in either the liver or adipose tissue have demonstrated that this transcription factor plays an important role on the development of diabetes (97)
. When hepatic SREBP-1c was overexpressed, mice developed a fatty liver, associated with increases in TGs and CEs (103)
. In comparison, overexpressing SREBP-1c in adipose tissue had far more drastic effects. These mice were characterized with abnormal white adipose tissue development, fatty livers, hypertriglyceridemia, and severe insulin resistance and T2DM (104)
. Hence, perturbations in fatty acid synthesis, coupled with observations that dysfunctional SREBP-1c expression initiates the onset of T2DM, suggest SREBP-1c is a suitable target for the management of disease progression through dietary intervention.
Recently, SREBP-1c was identified as a candidate gene in the regulation of human insulin resistance (105)
. Maudes and colleagues identified two missense mutations (P87L and P416A) in exons coding the aminoterminal transcriptional activating domain and found a correlation with individuals displaying severe insulin resistance; however, their preliminary findings revealed that the DNA binding ability of normal SREBP-1c and the two mutants were equivalent. Furthermore, data mining the Cambridgeshire Case Control Population cohort database identified a significant association between an intronic SNP (C/T) between exons 18c and 19c and the onset of diabetes in men, but not in women. Although only a preliminary study, these findings suggest that mutations in SREBP-1c may render a heightened sensitivity to developing diabetes; however, as no differences were observed in DNA binding between the various mutants, conclusive evidence concerning the role of SREBP-1c polymorphisms on the manifestation of disease remains to be demonstrated (105)
.
Nevertheless, SREBP-1c appears to be susceptible to diet, thus rendering this candidate gene a target for nutritional intervention. Examining two mouse models (CBA/JN and DBA/2N), Nagata and colleagues found that SREBP-1c hepatic gene expression differed between the models after the consumption of high fructose diets (106)
. Whereas hepatic SREBP-1c mRNA was significantly induced in CBA/JN mice, no response was observed in DBA/2N mice. Comparing the nucleotide sequences between the two mouse strains revealed a polymorphism (–468 A/G) that altered the binding potential of an unidentified nuclear protein. These findings demonstrate that a single polymorphism cannot only regulate a molecular event, but also ultimately modulate the sensitivity of a gene to dietary intervention.
Peroxisome proliferator-activated receptors (PPARs)
PPARs are an important group of ligand-activated transcription factors that mediate the cellular response to synthetic and biological compounds, such as fibrates, thiazolidinediones (TZDs), and fatty acids and their derivatives (107
, 108)
. Three PPAR isotypes with tissue-specific expression, termed , ß/
, and 
, play critical roles in mediating fatty acid catabolism, adipocyte differentiation, and wound healing (109
110
111)
. The pharmaceutical industry has focused much attention to PPAR-, as its activation promotes adipogenesis and the storage of TGs by regulating an important subset of adipose genes, including those involved in fatty acid transport, assembly into TGs, and glyceroneogenesis (97)
. Thus, activating PPAR- with TZDs reduces circulating fatty acid levels by promoting their uptake in adipocytes, and ultimately increasing insulin sensitivity. The creation of several PPAR transgenic and knockout animal models have additionally demonstrated their potential as targets for therapeutic intervention (112)
.
Initial work with PPAR- demonstrated that a homozygous null mouse was not viable, whereas the heterozygous animal is viable and has since been used in several studies (113
114
115
116)
. Curiously, studies using PPAR- +/– mice have demonstrated that the tissue-specific molecular functions of this transcription factor are not yet fully understood. For example, it was unexpectedly found PPAR- +/– mice are more sensitive to insulin, in contrast to what the authors expected (115
, 116)
. However, a liver-specific PPAR- deficient mouse with an obese-phenotype stemming from leptin deficiency was found to lack the ability to store lipids in the liver, resulting in the aggravated development of T2DM and insulin resistance (117)
. Although discrepant results exist regarding our understanding for the mechanistic actions of PPAR-, the efficacy of TZDs to alleviate symptoms associated with T2DM remains undisputed.
The functionally important role of PPAR- has led to it being considered as a candidate gene that may have a fundamental role in the onset of T2DM (118)
. Preliminary work has demonstrated that a missense mutation (proline 
alanine) in PPAR-2 is associated with improved insulin sensitivity and a decreased risk of developing T2DM (119)
. Indeed, both the Bogalusa Heart study (119)
and the Go-DARTS study (120)
have found that Pro12Ala polymorphism in the PPAR-2 gene beneficially affects insulin resistance and the risk of myocardial infarction associated with T2DM, respectively. However, a number of earlier studies found the Pro12Ala polymorphism to be associated with an increased body mass index (a predisposing factor for T2DM) (121
, 122)
. Although recent work would suggest that the Pro12Ala has an overall modest protective effect against T2DM, this effect has not considered the influence of diet. For example, preliminary indications suggest that the Pro12Ala polymorphism responds differently to changes in the dietary polyunsaturated/ saturated fatty acid ratio. More specifically, when this ratio is low, individuals with the Ala allele had a greater BMI than individuals with the Pro allele. The opposite was seen when the dietary ratio was increased (123)
. Thus, it is clear that PPAR-2, and its associated polymorphisms, play an important role in mediating the onset of insulin resistance and T2DM. Furthermore, it is imperative that a continued nutrigenetic approach be considered if we are to ultimately unravel the influence of PPAR-2 polymorphisms on the individual’s response to dietary intervention.
Intestinal fatty acid binding protein (IFABP)
The gastrointestinal tract (GIT) is the body’s first tissue boundary to interact with dietary compounds. The importance of unraveling the molecular mechanisms regulating the bioavailability (i.e., the degree or rate at which a substance is absorbed or becomes available at the site of physiological activity) of a given nutritional molecule are fundamentally important if we are to understand their bioactive roles throughout the body; however, this remains a difficult goal to achieve primarily due to the lack of well-characterized in vitro model systems and the in vivo complexity of this organ. Indeed, factors such as gastric motility, pH, disease status and the enzymatic contribution of the diverse microbial community will play a critical role in defining the bioavailability of a substance (124
, 125)
. Associations between the various aforementioned factors of the GIT and physiological states suggest that understanding the interactions between diet and disease states should begin in the GIT. For example, recent work has revealed that the presence of a gut microbiota mediates energy storage by increasing fat reserves and insulin resistance, and thus may have an influence on the development of T2DM and obesity (126)
. In another example, certain fibers increased the viscosity of the GIT luminal contents and resulted in decreased plasma cholesterol levels, demonstrating that GIT motility modulates cholesterol absorption and plays an important role in defining nutrient bioavailability (127)
. To date, little work has explored the role of genetic polymorphisms in the intestine on nutrient bioavailability; however, examining the role of polymorphisms in transporter genes regulating drug bioavailability suggests that this line of research will reveal important information explaining the inter-individual and inter-ethnical variability in drug/nutrient disposition and response (128
129
130)
.
As previously stated, little work has concentrated on intestinal gene polymorphisms; however, preliminary work aimed at unraveling the genetics behind T2DM in Pima Indians have suggested an association between insulin action, plasmatic nonesterified fatty acids (NEFA), and polymorphisms in the intestinal fatty acid binding protein (IFABP) (131)
, suggesting that SNPs have the ability to modulate intestinal transport function. The FABP family is comprised of nearly 20 soluble proteins capable of binding long chain fatty acids (LCFA), bile acids or retinoids with a high affinity (132)
. One member, IFABP, is exclusively expressed in the small intestine along with the liver-FABP and ileal bile acid binding protein (IBABP). Although the precise physiological functions of these proteins have not been elucidated, IFABP is believed to bind and transport LCFA in the cytoplasm of columnar absorptive epithelial cells of the small intestine (131
, 133)
. A polymorphism at codon 54 of the IFABP gene (Ala54Thr), resulting in a change from alanine to threonine, has been associated with a heightened affinity to bind LCFA (134)
and increase TG secretion (135)
. Although these findings have been reported using in vitro model systems, the Thr54 allele has been associated with impaired insulin action and increased fat oxidation in several populations (134
, 136)
. Pratley and colleagues demonstrated that healthy Pima Indians homozygous for the Thr54 form have higher plasmatic concentrations of NEFA and an increased insulin response after the consumption of a high fat meal; however, such findings have not been observed in all studies (136)
. Nevertheless, these preliminary findings suggest that elucidating the role of IFABP polymorphisms on dietary LCFA transport, and, ultimately on insulin action, merit further investigation. In conclusion, both IFABP and the previously discussed ABCG5/ABCG8 highlight that polymorphisms in genes coding for intestinal transporters may compromise health by modulating the bioavailability of dietary components.
|
OUTLOOK |
|---|
|
TOPABSTRACTINTRODUCTIONA COMPLEX PARTNERSHIP: DISEASE...NUTRIGENOMICSNUTRIGENETICSOUTLOOKREFERENCES |
|---|
.
The observed differences in an individual’s response to diet have been attributed to differences in the underlying genetic makeup, prompting exploration into the role of nutrient-gene interactions in the determination of a healthy phenotype. This avenue of research has been facilitated by our recent entrance into the post-genomic era, which has witnessed the development of analytical platforms and complementary data mining algorithms capable of comprehensively studying and interpreting a biological system. However, unlike the pharmaceutical industry, which aims to target a specific dysfunctional gene or gene product to improve health, the nutritional industry must manage health through a complex mixture of nutritional molecules (i.e., diet), each with their own influences on the biological system. Thus, in comparison with a medicinal compound, consuming a diet drastically increases the number of molecular endpoints that are capable of influencing phenotype, and thereby places the field of nutrition in a prime position to benefit from the technological innovations brought forth by the post-genomic era. Indeed, the one drug–one target paradigm of the pharmaceutical industry is simply not an appropriate or feasible goal for nutrition. Hence, the advent of nutrigenomics and nutrigenetics: two approaches with distinct immediate goals for unraveling the relationship between diet and genes, but a common ultimate objective to optimize health (28
, 29)
. Although both of these approaches individually will begin to unravel nutrient-gene interactions, it is critical that these approaches are amalgamated if we are to achieve the ambitious goal of a personalized nutrition. Furthermore, the diverse genetic-makeup of humans implies that the clear distinction of these two approaches is difficult to achieve. Using genetically identical animal models permits gene-diet interactions to be explored by either approach; however, this is nearly impossible to achieve in humans and suggests that these approaches could fall under the more general label of "nutritional genomics." Irrespective of the terminology used, understanding nutrient-gene interactions will benefit from the concept of systems biology (or integrative metabolism), which aims to understand physiology and disease by integrating and considering molecular pathways, regulatory networks, cells, tissues, organs, and ultimately the whole organism (26
, 137)
; however, such an approach is not routinely used in nutrition and is conceptually challenging due to the large quantities of data generated. Currently, few examples exist in which an integrated approach has been used to examine the influences of exogenous factors on metabolism (22
23
24
, 138
, 139)
; yet, even these eloquent examples have only provided the genomic and metabolomic profiles of a single organ. Nevertheless, an underlying theme begins to emerge with these studies. Indeed, there is an inherent, albeit counterintuitive, advantage with such an approach that is able to minimize the known limitations associated with these analytical platforms (e.g., such as those observed with microarray data analysis (140)
). The integration of various platforms has the ability to reduce the quantity of distracting "noisy" data associated with each individual technology and diminish the number of molecular endpoints to be interpreted by focusing on only those endpoints common between the various experimental platforms (Fig. 3
). Although at first glance this may seem contradictory, as the use of multiple platforms will yield large quantities of data, the appropriate statistical models can filter through this data and highlight only those important changes. Indeed, this reductionist concept has been previously demonstrated, where a combined lipid-metabolomics and transcriptomics approach identified a single hepatic gene target responsible for changes in lipid-metabolites after the consumption of arachidonic acid, despite nearly 300 genes being identified as differentially expressed after the consumption of this diet (24
, 62)
.
|
Although such results demonstrate the advantage of an integrated approach for the study of exogenous compounds, it is important to note that these studies used genetically identical mice; therefore the role of genetic polymorphisms has not been considered. Furthermore, while the targeted study of single organs produce invaluable information concerning the metabolism of these compounds, the influence of the remaining biological system is often not considered. Not only do exogenous factors have differential responses both within and between various tissues, but underlying differences between the distinct regions of an organ (e.g., the gastrointestinal tract) may also mediate metabolism. Therefore, the presence of genetic polymorphisms that can influence nutrient absorption, metabolism and transport throughout the body must be acknowledged prior to making dietary recommendations to ameliorate health. For example, using margarines enriched with plant sterols to assist in reducing circulating cholesterol levels could be counterproductive in individuals with mutations inhibiting the proper function of ABCG5 or ABCG8 transporters.
If the goal of a personalized nutrition is to be realized, we must begin to think of the biological system as a physiological network that is intricately connected rather than composed of individual and unrelated elements. Comparable to the reverberations felt by a spider when an insect crashes into its web, nutritional molecules have the ability to trigger inter-connected molecular mechanisms in various tissues throughout the body. Although it is experimentally feasible to use a systems biology approach in humans (similar to that recently performed with the ApoE* Leiden transgenic mouse (138)
), it is ethically impossible to even conceptualize that such an analysis will be performed in humans in the foreseeable future. Therefore, alternate methods to assess nutrient-gene interactions must be envisioned.
Profiling metabolites, which best depict an organism’s acute cellular function, offers an attractive platform to identify disease biomarkers (141
142
143)
. Furthermore, defining the metabolite profile associated with disease and health states will offer an ideal starting point to decipher how nutritional molecules can modulate these phenotypes. It is enticing indeed to imagine that a drop of blood could be rapidly analyzed for its comprehensive metabolite profile, uploaded into software capable of comparing this profile with those present in a massive database, and return information enabling a physician to make dietary recommendations to optimize health; however, several important points must be resolved if we are to achieve such an attractive aspiration. First, there is currently no single technology capable of comprehensively assessing all of the metabolites in a biological sample. Nuclear magnetic resonance, chromatography and mass spectrometry are powerful analytical platforms; however, the complete spectrum of metabolites is currently assessable only through the integration of these technologies (144)
. Thus, the creation of a unique platform that is capable of noninvasively assessing the complete metabolome will become a valuable tool in both a clinical and research environment. Second, in order to characterize the metabolic profile underlying a healthy phenotype, one must characterize the variability existing in metabolites across healthy individuals. Indeed, this prompts the question, "What metabolic profile characterizes healthiness?" The influence of such factors as asthma (145)
, smoking (146)
and physical activity (147)
have been reported to affect metabolism; however, are these differences significant when considering all metabolites simultaneously? A comprehensive approach to metabolite profiling, coupled with robust statistical algorithms, may reveal that these perturbations have a minimal effect on the overall metabolite profile. For this to be assessed objectively, it becomes critically important that individuals participating in such epidemiological studies are genotyped prior to their inclusion in a cohort and that cohorts are of a sufficient size to permit statistical algorithms to function optimally. Only then can we determine whether metabolomics can identify a "healthy phenotype." Third, the metabolite profile associated with various disease states must be characterized. A suitable and feasible starting point would define the metabolite profile of monogenic diseases, which stem from a single dysfunctional gene. In contrast, the metabolite profiles of diseases such as CHD, cancers, and T2DM are difficult to correlate with single genes, as dysfunctional networks characterize these conditions. Rather, the use of annotated microarray results to identify common biological functions, such as that used by Segal and colleagues to classify different tumor types (36)
, suggests that these polygenic diseases may have similar metabolite profiles despite varied gene expression profiles. For the moment, the metabolite profiles of polygenic diseases have not been comprehensively explored; however, it insinuates that associating a candidate disease gene with a metabolite profile may not be an appropriate endpoint. However, relating perturbed functional networks with metabolite profiles may have more relevance from a clinical perspective. Finally, it is vital that this information be stored in publicly available databases that have stringent and universally accepted criteria for accepting data (144)
. The vast quantity of metabolite data produced by profiling diseased and healthy individuals will be difficult to interpret by a human; however, such databases will empower supervised biomathematical algorithms to compare a given metabolomic profile with those referenced in the database. Although there is currently no existing database empowering the scrutinization of metabolite profiles and permitting nutritional recommendations to be proposed, the initiatives of groups such as the American Society for Nutritional Sciences Long Range Planning Committee will manifest itself into a database with precisely this objective.
In this regard, the future lies not with the technologies, but with the storage, management, and interpretation of the vast quantity of metabolomic data. No single lab will be able to achieve the concept of a personalized nutrition alone; rather, a collective effort by the scientific community to adhere to guidelines put forth regarding experimental designs, analysis, and data storage will generate a database that is readily available to researchers and clinicians alike. Thus, nutrition in the 21st century is poised to be an exciting and highly relevant field of research, as each new day is accompanied by advances in our understanding of how the interactions between lifestyle and genotype contribute to health and disease, taking us one step closer to achieving the highly desirable goal of personalized nutrition.
|
FOOTNOTES |
|---|
3 http://www.ncbi.nlm.nih.gov/geo/
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONA COMPLEX PARTNERSHIP: DISEASE...NUTRIGENOMICSNUTRIGENETICSOUTLOOKREFERENCES |
|---|
