MEETING BRIEF

Advancing the Genetics of Obesity
Strategies and Methods

by David B. Allison and
Moonseong Heo

(Posted August 15, 1997 · Issue 14; archived September 5, 1997)

Abstract

The field of obesity genetics has exploded in the past several years, with recent discoveries injecting new vitality into the field. As in other areas of genetics, these advances draw on many new technologies - in statistics, animal modeling, molecular techniques, and phenotyping. Such advances provide new challenges, including moving from discovering obesity genes in mice to finding obesity genes in humans and the challenges of establishing and coordinating the large collaborative efforts needed for successful population genetics studies. These topics were the focus of an intensive research workshop.

Readers should note this issue's Op-Ed, which focuses on obesity, weight loss, and mortality.

The idyllic Mohonk Mountain House, nestled in the upstate New York hills known as "the Gunks," was the site of a recent meeting sponsored by the North American Association for the Study of Obesity and the National Institute of Diabetes, Digestive, and Kidney Disease (NIDDK). One hundred scientists convened there July 10-13, 1997, to discuss strategies and methods for advancing the study of the genetics of obesity. A sense of vibrancy that permeated the meeting was due not just to the fine mountain air. With a high caliber of scientific participants, and with generous corporate sponsorship funding the participation of nearly forty young investigators, the meeting was a forum for both novice and expert researchers to converse and to strike up new collaborations.

Background

Farmers have known for centuries that livestock can be selectively bred for their degree of adiposity. This indicates a genetic influence on fatness. But until recently, human obesity was generally thought to be a failure of will. Although some data from human twin, family, and adoption studies suggested a genetic component to human adiposity, such studies were typically small and methodologically weak, having little impact. All this changed in the mid-1980s with the publication of several seminal papers by Dr. Albert J. Stunkard of the University of Pennsylvania. Stunkard showed, through several large population-based twin and adoption studies, that there was a substantial genetic contribution to variations in human adiposity. Today, this notion enjoys almost universal acceptance among scientists, and most agree that the genetic component is quite strong.

Subsequently, we have seen an exponential rise in obesity genetics research. In 1992, the first major obesity gene in mice (the agouti gene) was cloned. Subsequently, five other major mouse genes for obesity were cloned (ob, ob receptor, fat, tubby, and mc4r). Of note is the 1994 cloning of the ob gene and the discovery of its protein product, leptin. Leptin appears to be a key hormone for long-term regulation of body energy stores. The resultant research boom on this topic has been jokingly called "leptomania." Finally, several quantitative trait loci (QTLs) for variations in adiposity in mice as well as humans have been identified, setting the stage for indentifying more obesity-related genes.

Dr. Stunkard himself chaired the opening session, titled "Review of Current Knowledge with an Emphasis on Methods." The first presentation, by Dr. Arlen Price of the University of Pennsylvania, reviewed the current knowledge derived from twin, family, and adoption studies about genetic influences on human obesity. He noted that approximately 70% of the variation in human adiposity within modern populations is due to genetic variation among individuals in these populations. Price then discussed the implications of the fact that the quantities of heritability estimated from twin studies are frequently greater than those from family or adoption studies. He indicated that the most probable underlying reason is substantial non-additivity among the genetic influences on human obesity. While this indicates a high degree of genetic influence, the audience was reminded that genes acting in a non-additive fashion can be exceptionally hard to identify. Next came a thorough discussion by Dr. Claude Bouchard of Laval University, who reviewed the methods for locating specific obesity genes in humans. Dr. Streamson Chua of Rockefeller University then discussed information obtained from the study of monogenic obesity syndromes in rodents. He described how the leptin system's ob and ob receptor influences obesity-related phenotypes including those involved in both energy intake and expenditure. To conclude the session, Dr. Janis Fisler of the University of California at Davis discussed methods for isolating QTLs in mice. While these methods have successfully revealed many QTLs in various mouse strains, the field has yet to move from isolating chromosomal regions containing QTLs to identifying specific genes.

Next on the agenda was advances in animal models. Dr. Daniel Pomp of the University of Nebraska discussed the difficulties in identifying QTLs in mice and other animals for obesity-related traits. Although we have isolated genomic regions containing QTLs, Pomp suggested that because of statistical power issues and artifactual defects deriving from some current procedures, we have may have underestimated the number of obesity QTLs. He further believes that the effects of obesity QTLs are likely to have been overestimated. Both conclusions stem in part from his belief that obesity QTLs may be closely linked to one another and therefore may either attenuate or enhance linkage signals in selected regions. Dr. Rudolph Leibel of Rockefeller University further elaborated on the physiology of monogenic obesity syndromes. He pointed out the complexity of these syndromes as demonstrated by their interaction with other variables (e.g., sex). This theme of the complexity of gene expression was continued in the talk by Dr. Gary Truett of the Pennington Biomedical Research Center. Truett argued that it is difficult to disentangle the effects of certain genes because at times they have been set up to co-occur almost exclusively with specific alleles at other loci (e.g., the db mutation, which produces massive obesity and diabetes in mice, and the misty mutation, a coat color gene that may also affect body weight in mice). Truett also pointed out that the expression of these genes in mice may vary dramatically with age. This is consistent with work that has been conducted in humans.

The next session, on advances in molecular diagnostics, began with Dr. James Knowles of the Columbia University Genome Center discussing automated high-throughput fluorescence-based genotyping. His work exemplified the state of the art for accurate and high-speed genotyping. This was followed by a futuristic talk by Dr. David Burke of the University of Michigan. Burke discussed the use of microchip technology for genotyping. Although this technology is not yet widely available, there was excitement in the audience at being on this "star trek" with Burke; we eagerly await further developments in this area. Next came an expert discussion by Dr. Benjamin Taylor of the Jackson Laboratory on the use of DNA pooling for markedly enhancing the efficiency of QTL mapping studies. To date, the use of DNA pooling has not been applied extensively to obesity research, but enthusiasm was high for what may become an ideal method for examining associations within large populations, given many of the forthcoming data sets in the obesity arena.

Just as there are advances in genotyping technologies, so have there been technological advances in phenotyping. Or have there been? That was the focus of the next session. Dr. Steven Heymsfield (of the Obesity Research Center at St. Luke's/Roosevelt Hospital in New York City and Columbia University's College of Physicians and Surgeons) discussed advances in body composition analytic methods. Although enormous strides have been made, a major limitation is that many of these newer techniques are not adaptable to field settings or to the measurement of a large number of human subjects necessary in gene mapping studies. This theme recurred in the next three talks. Dr. Eric Ravussin of NIDDK discussed measurement of energy expenditure, Dr. L. Arthur Campfield of Hoffman-LaRoche covered hormone and metabolite measurement, and Dr. Barbara Rolls of the University of Pennsylvania discussed eating-related behavior, all noting an inherent tension or conflict in this measurement arena. Certain methods yield highly precise and valid data but are often too expensive, time-consuming, or cumbersome for field work and/or for the large sample sizes that such population genetic studies require. In contrast, many cheaper and simpler procedures are typically of questionable validity. This problem is perhaps greatest in the area of measuring eating behavior, and Dr. Rolls called for more methodologic research to develop tools in this area.

The fifth session concerned statistical methods for the design and analysis of human gene-mapping studies. Dr. Robert Elston of Case Western Reserve described a two-stage approach: One first conducts a genome scan looking for linkage to obesity-related traits, with a low density of genetic markers. Then, using a relatively high alpha-level for statistical significance (to minimize false-positive results), one selects chromosomal regions for follow-up with a higher density of markers. Elston explained how the selection of the alpha-levels as well as marker density can be determined based on factors such as the relative cost of genotyping and phenotyping. Then Dr. Norman Kaplan of the National Institute of Environmental Health Sciences discussed novel methods for disequilibrium mapping, focusing on the transmission disequilibrium test and extensions thereof. He pointed out that this test has been extended to quantitative traits and that work is ongoing to allow its use over the broadest possible circumstances. Such methods should allow efficient mapping of genes in populations in general and small isolated founder populations in particular. Dr. Nicholas Schork of Case Western Reserve University then spoke on incorporating multivariate data into gene mapping analysis. Schork pointed out that most current approaches to gene discovery in humans have involved collecting information on more than one phenotype. However, these multiple phenotypes are typically not incorporated into an analysis simultaneously. This fundamentally discards information, which Schork felt is unnecessary. He discussed methods for including multiple measured phenotypes into a single analysis, and he demonstrated that, under some circumstances, this could markedly improve the power of gene-mapping studies. Next, Dr. John Blangero of the Southwestern Biomedical Research Foundation discussed the use of variance components models in linkage analyses. He indicated that methods relying only on phenotypic differences among relative pairs discard much of the important and useful information that the investigator has collected. He pointed out that the use of variance components models incorporates that information into the statistical analysis and therefore can result in substantially greater power. The final speaker in this session, Dr. David B. Allison (New York Obesity Research Center and the Columbia University College of Physicians and Surgeons), spoke on efficient sampling procedures, and discussed ways in which subjects could be selected that maximized the statistical power and minimized the cost of study.

The final session covered efficient use of global resources. All speakers addressed a common underlying theme: To find genes for obesity in humans we are likely to need enormous sample sizes and resources. To do this most efficiently we need, as a field, to cooperate and pool our resources. From all different angles the speakers attacked this problem. Drs. Janice Sowinski (Human Biological Data Interchange) and Jean Beck (Coriell Institute) discussed how investigators could share data effectively by banking their DNA at a publicly accessible facility. Dr. Mary K. Pelias of the American Association for the Advancement of Science discussed certain ethical issues involved in such large collaborative studies and DNA collection and pooling. Finally, Dr. John Rice of the University of Washington at St. Louis discussed the statistical issues involved in such large collaborative studies and pooling data across several investigators.

The overriding theme that emerged from the meeting it is that, as a field, we are not yet maximally using the information we have. We must more effectively employ statistical approaches that do not discard information, and we can benefit from pooling information across studies. We now await the results of studies that have been stimulated by this meeting.

David B. Allison, Ph.D., is Associate Research Scientist at the Obesity Research Center and Assistant Professor of Psychology in Psychiatry at the Columbia University College of Physicians and Surgeons, New York City.

Moonseong Heo is a postdoctoral research fellow at the Obesity Research Center at the St. Luke's/Roosevelt Hospital Center in New York City.

Send us your comments and ideas for future articles.

Endlinks

Drug Treatment of Obesity - a site compiled by the Academic Computing Center at the University of Missouri, Kansas City.

Anthropometry Resource Center - includes a tutorial and desk references.

The Coriell Institute - is a repository for human and animal mutant cell lines and DNA samples. And the Jackson Laboratory is a source for mice of different genetic strains.

The Centre for Genetic Improvement of Livestock at the University of Guelph - has information and links for livestock genetics research.

Genetics and Ethics Home Page - brings together genome research projects, discussion groups, meeting announcements, public action groups, and ethical and legal issues of human genetics.

World of Genetics Societies - a gateway for accessing these societies.

Obesity research information is presented at the following Web sites:

The Obesity Page at the The Institute of Clinical Nutrition and Metabolism, Princess Alexandra Hospital, Brisbane, Australia

The North American Association for Obesity

National Institute of Diabetes, Digestive and Kidney Diseases maintains an obesity Web site for the lay public

For additional links to obesity research, see those that accompany this issue's Op-Ed.