Nutrigenomic approaches for obesity research
Dr RM Elliott and Professor IT Johnson, Institute of Food Research, Norwich Research Park, Colney, Norwich, NR4 7UA, UK. E-mail: email@example.com; firstname.lastname@example.org
At the individual level, weight gain is essentially the result of energy intake exceeding expenditure for significant periods of time, but this obvious truth provides no insight into the strategies needed to deal with the ever-increasing problem of obesity in Western populations. It is equally obvious, however, that certain individuals are more prone to developing obesity than others. This phenomenon invites the nutrition research community to explore the physiological basis for such differences and ultimately to design more targeted and personalized approaches to the control of body weight (1).
Novel research strategies are required to understand the molecular mechanisms controlling energy balance. In parallel with such studies, there is still much to be learned about the metabolic consequences that follow when an appropriate energy balance is not maintained, and how this relates to risks of diseases such as hypertension, heart disease, stroke, diabetes and certain cancers. The developing fields of nutrigenetics and nutrigenomics, with their accompanying battery of high-throughput technologies, provide an unprecedented opportunity to cope with the complexity of this condition and to develop the knowledge base required.
Terminology: nutrigenetics and nutrigenomics
The term ‘nutrigenetics’ is generally used to refer to the impact of genetic variation on optimal dietary requirements for an individual (i.e. in the simplest terms: gene → diet interactions). Although the term ‘nutrigenomics’, in its broadest sense, encompasses nutrigenetics, more commonly the main focus of nutrigenomics is considered to be on how diet regulates gene function (transcription and translation) and metabolism (i.e. diet → gene interactions) (2).
Nutrigenetics and obesity
Genetic differences play an important role in the development of obesity, although it is clear that these are by no means the only contributing factors. Environmental and social factors are also very important. The relative contributions of genetic and socioeconomic factors to the development of obesity, and the ways in which these interact in human societies, are largely unknown.
The genetic code (DNA sequence) carried by any two unrelated people is approximately 99.9% identical. It is the variation in the sequence of the remaining 0.1% that determines the genetic component of inter-individual differences in disease risk, and presumably also their differing responses to the nutritional environment. Sites in the DNA where the sequences of individuals differ commonly (e.g. in at least 1% of the population) are called polymorphisms; the most common form being a single letter change in the code termed a ‘single nucleotide polymorphism’ (SNP). As each cell contains two copies of every gene (except those present on the sex chromosomes), one individual may carry various combinations of a polymorphism. The term ‘genotype’ refers to the combination of sequences in the two copies of a gene for a particular polymorphism.
The most recent update of the human obesity gene map emphasizes just how complex the genetic component of obesity alone is. There are currently more than 600 genes, markers and chromosomal regions that have been associated or linked with human obesity, and more are added to this list with each update (http://obesitygene.pbrc.edu/) (3). To date, syndromes of obesity because of single-gene mutations have been described for at least 10 different genes. These cases provide immensely valuable insights into the roles of these genes, and to their contributions to key processes, most notably appetite, that influence the development of obesity. However, such syndromes are extremely rare and therefore of limited relevance to the majority of obese individuals.
The effects of the common genetic polymorphisms associated with ‘sporadic’ obesity at the population level are much harder to study for two main reasons. First, the effects of each polymorphism are more subtle, generally modulating the risk of developing obesity by perhaps a few percent, rather than inevitably leading to severe and intractable weight gain. Their effects are more difficult to detect reliably in a diverse population with varied lifestyles. Second, interactions between genotypes for obesity-linked genes may be important. For example, particular combinations of genotypes may cancel each other out. Alternatively, some combinations of genotypes may interact to enhance or reduce risk to a greater extent than the sum of the effect of each genotype considered in isolation.
Given the number of genes implicated so far, and the fact that many more may yet be identified, characterizing all the genes involved in obesity, let alone examining their possible interactions, appears a truly daunting task. However, some recent developments help to make this work more feasible. These include the development of technologies capable of parallel genotyping analysis for hundreds of thousands of SNPs from a single small blood or tissue sample (4). It is estimated that there are about 10 million SNPs in human populations. This scale currently still exceeds the capacity of the new platform technologies, but SNPs that are located close together in the DNA sequence on the same chromosome tend to be inherited together. A set of such associated SNPs is termed a ‘haplotype’ and it turns out that most chromosome regions have only a few common haplotypes. So, while a chromosome region may contain many SNPs, it is possible that analysing only a few ‘tag’ SNPs can provide most of the information on the pattern of genetic variation in that region. Defining these haplotype blocks and the most reliable tag SNPs are the goals of the International HapMap Project (http://www.hapmap.org/) (5). Realizing these goals will help to bring the complexity of genetic studies down towards a level that may be manageable.
In spite of the rapid pace of technical developments, the history of genetic association studies addressing subtle and complex associations indicates that the reliable detection of true associations, and the avoidance of false positives, will continue to be a significant challenge (6,7). Appropriate study design and improved statistical approaches will be vital (8–10). Ultimately, this type of work will require studies involving very large numbers of human subjects. There is therefore an obvious need to promote new international collaborations, bringing together the large and well-defined cohorts of human subjects that have already been established, to achieve the study power necessary (11).
Nutrigenomics and obesity
The potential impact of functional genomic approaches (transcriptomics, proteomics and metabolomics) in nutrition has been reviewed extensively (12–16). This potential is now starting to be realized, with the publication of an increasing flow of nutrigenomic studies each giving new mechanistic insights.
The transcriptome is the complete collection of RNA transcripts produced from the DNA in a genome. Transcriptomics is performed using microarray technology, which enables the transcript levels for many tens of thousands of genes to be studied simultaneously. This technology is ideally suited to the study of the metabolic syndrome and the associated inflammatory signals that underlie many of the comorbidities linked to the obese state. Microarrays have been used to define the changes in patterns of gene expression at the level of RNA in the adipose and other tissues of different strains of lean and obese mice, revealing characteristic and tissue-specific alterations in the expression of genes involved in adipogenesis, inflammation and gluconeogenesis (17,18).
More limited work has been performed with samples from human subjects. Some regional differences in gene expression within different fat depots have been described and a number of studies have examined the effects of weight loss/caloric restriction on patterns of gene expression in adipose tissue from obese subjects (19). Preliminary studies have also been performed on the patterns of gene expression in regions of the human brain that are known to show differential responses to nutritional stimuli in obese vs. lean individuals (20). These types of studies provide a much broader perspective on the effects of obesity than was possible before the development of microarrays and a wealth of new information and research leads. However, a more comprehensive and focused programme will be required to obtain a robust overview for the changes in gene expression related to obesity and their biological significance in relation to health.
The proteome is the full complement of proteins produced from the transcriptome, including all subsequent modifications that the proteins may undergo. To date, proteomics has been used less extensively in nutrigenomic studies than transcriptomics but it has just as much potential (21). The use of protein expression patterns as ‘biomarkers of vulnerability’ to obesity-related diseases such as colorectal cancer is one approach (22). Studies relating directly to the physiology and biochemistry of obesity have examined patterns of protein expression in adipocytes during differentiation (23,24), the effects of a high-fat diet on protein expression in different target tissues in mice (25) and have compared skeletal muscle of lean and obese women (26).
In addition to its use in the analysis of gene expression in tissues, proteomics provides a possible route for the identification and validation of new protein biomarkers that can be detected in plasma. The international HuPO Plasma Proteome Project is providing new resources (protein databases and optimized experimental standards) that will be essential for this kind of work (27).
Metabolomics is the study of the sum total of endogenous and exogenous metabolites in a cell, an organ, or in body fluids, and is the newest of the ‘omic’ technologies. It is ideally suited, both from a technical and scientific standpoint, to the global analysis of metabolite patterns in body fluids (plasma/serum/urine, etc.), which are comparatively easy to access in human volunteers.
The mass spectrometry and nuclear magnetic resonance techniques that are used to analyse the composition of these fluids are capable of very high sample throughput at comparatively low cost (after the initial set-up of the machinery). It is therefore possible to generate large datasets very fast. These approaches have already been demonstrated to be sensitive enough to detect the often subtle effects of dietary modification, and should lend themselves readily to the detection of metabolic differences, both between individuals with differing susceptibilities to chronic weight gain, and within individuals who are undergoing significant changes in body weight and adiposity (28,29).
There is also increasing interest in the use of metabolomic profiles as markers of dietary habits and as a descriptor of nutritional phenotype (30). Before this concept can be developed further, the influence of potential confounding factors (e.g. age, gender, ethnicity, physical activity and gut microflora) has to be defined.
Another challenge is that present metabolomic technologies generate metabolite profiles for which the majority of signals are not immediately identified. While these profiles may be used with multivariate statistical tools for diagnostic/classification analyses, the absence of the full range of identified metabolites limits the biological interpretation of the data.
The challenges and potential of nutrigenomics and nutrigenetics
The potential of nutrigenomic and nutrigenetic approaches is starting to be realized. A great deal of progress has already been made and, by applying new analytical tools to the data already generated, it has proven possible not only to obtain lists of gene products and metabolites that change in response under defined conditions but also to gain insights into the overall biological processes involved.
Another emerging challenge that may well carry implications for the development of obesity research is that of ‘epigenomics’. This can be defined as the study of heritable epigenetic signals, encoded in patterns of DNA methylation and histone acetylation within the chromatin, that modulate the expression of genes (31). Epigenetic marks have recently been shown to change in response to environmental factors over an individual’s lifetime (32), so even identical twins may ultimately develop differing susceptibilities to adverse environmental factors. As the massive task of mapping the human epigenome progresses, it will become possible to explore the role of epigenetic effects, both as causes and possible consequences of obesity.
Continuing development of improved statistical and bioinformatic tools means that the conclusions of the data analyses are becoming more robust and sensitive. New text-mining tools are starting to make it easier to interrogate the full body of scientific literature and thus to place new findings within the context of current scientific knowledge. The next challenge is to develop tools to integrate the different types of data and start to realize the vision of nutritional systems biology. In all these areas, the European Nutrigenomics Organization (NuGO, http://www.nugo.org) is working to identify bottlenecks and emerging technical requirements and seeking solutions to them.
Not least among the many challenges are the needs for quality control, standardization, data capture and storage of nutrigenomic and nutrigenetic data. The ‘omic’ tools produce vast quantities of data rapidly. If we are to make use of this information, rather than drown in the flood, it is essential that the data collected are of high quality and are captured in a manner that enables them to be stored and exchanged readily. Standardization of data capture for microarray studies has already been addressed (33) and equivalent procedures are in development for proteomic and metabolomic studies (34,35). Refinements to these data capture systems are likely to include appropriate data-quality metrics and specialty-specific metadata. For example, NuGO is working with international organizations from other specialties on the development of the Reporting Structure for Biological Investigations Tiered Checklist (RSBI-TC, http://www.mged.org/Workgroups/rsbi/rsbi.html), both through contributions to the design of core modules and through the development of the nutrition-specific component.
Finally, beyond the ‘omic’ technologies, there is clearly an emerging need for the development of non-invasive techniques that will allow biological processes to be visualized in remote tissues in vivo. These will be essential for future studies with human volunteers to confirm that the dietary effects characterized in model systems also occur in target tissues in humans in the manner predicted.
Thus, to date, nutrigenomic/nutrigenetic research in obesity has provided insights in three major areas. First, the identity of many genes in which polymorphisms can affect the propensity to develop obesity and related conditions such as diabetes and cardiovascular disease and, second, characteristic changes in patterns of gene expression in adipose and other tissues associated with obesity. In turn, these have provided indications as to the processes involved and their biological consequences.
Cutting-edge work in this area at the moment involves further validation, optimization and standardization of the ‘omic’ technology platforms and how they are used for nutritional studies as well as the development of improved statistical and bioinformatic methods to enable the full biological meaning of the vast datasets produced to be extracted. New biomarkers of health, and the more rigorous characterization of lean and obese phenotypes at the molecular level are being developed. The different ‘omic’ and bioinformatic approaches to realize the vision of nutritional systems biology are in the process of integration, and non-invasive techniques to facilitate future studies in humans are in development.
What could the full exploitation of nutrigenomic/nutrigenetic approaches provide during the next 25 years?
In 5–10 years, a detailed overview of the molecular mechanisms that control energy balance will be available. In 5–15 years, a mechanistic definition of the metabolic consequences of failure to maintain appropriate energy balance and how this relates to the development of associated diseases is likely. A longer time frame (up to 20 years) will see detailed analyses of the basis of individual variation in obese individuals, both in the propensity to develop obesity and the propensity to develop associated diseases. In parallel with this, and perhaps 25 years away, is the creation of a comprehensive knowledge base to be used for the development of targeted strategies to reduce obesity incidence and severity and the burden of chronic disease at the population level.
However, it is important to note what the main barriers are in this area of research, as these could prevent such a full exploitation of the potential of nutrigenetics and nutrigenomics. The human system is immensely complex and individual variation very diverse. Coping with this may be difficult, as will designing and executing studies of sufficient power to define the effects of, and interaction between, genetic, epigenetic and dietary factors, which may be subtle in the short term but profound over many years or a lifetime.
Conflict of Interest Statement
No conflict of interest was declared.