Human adiposity resolves complex interactions among genetic, developmental, behavioral, and environmental influences. Obesity is primarily the result of a net imbalance of caloric intake over energy expenditure. Even small differences resulting in positive energy balance – when integrated over long periods of time – can produce increased adiposity. In the United States, 65% of adults are overweight (BMI 25.0–29.9) and more than 30% of the adult population is obese (BMI ≥30) [1
]. The problem also affects children in whom the percentage with BMI >95 percentile between the ages of 6–19 defined by the CDC growth charts is now 16% [2
]. Increasing in parallel with these trends in obesity are the frequently associated comorbidities of diabetes, hypertension, and cardiovascular disease [3
Evidence for potent genetic contributions to human obesity is provided by familial clustering of increased adiposity, including a 3–7 fold increased relative risk (λs
) among siblings [4
], estimates of heritability for fat mass between 40 and 70% in twin studies [5
], and rare monogenic causes of syndromic and nonsyndromic obesity [7
]. Severe loss-of-function mutations in genes causing monogenic forms of obesity do not account for the heritable component of adiposity in the majority of individuals, although mutations in the Melanocortin 4 Receptor (MC4R)
may account for obesity in 1–5% of the extremely obese [8
]. The phenotypic differences among individuals at the extremes of adiposity presumably reflects – to a substantial degree – allelic variation at genes that affect energy intake, expenditure and the chemical form in which excess calories are stored (‘partitioning’). The nature of these genes and their interactions is of obvious interest.
Both linkage and association studies have been undertaken to identify obesity candidate genes. The 2005 Human Obesity Gene Map update reports over 1,100 studies including 500 genes, markers, and chromosomal regions apparently linked to, or associated with, human obesity phenotypes in specific cohorts [9
]. Yet only 22 genes have been reproducibly implicated in ≥5 studies as statistically significant contributors to common obesity. For monogenic obesity, 176 mutations have been identified in 11 genes, and 50 loci have been mapped corresponding to Mendelian syndromes related to obesity [9
]. Genome-wide association studies have and will likely continue to identify additional genes associated with obesity, but often identify genetic variants that are not themselves functionally significant and genes which were not previously known to be implicated in energy homeostasis. In most individuals, the genetic basis for obesity is complex and likely to involve the interaction of multiple genes as well as gene-by-environment interactions.
In most candidate-based genetic analyses of complex disease phenotypes, fewer than five candidate genes are interrogated in a population. Moreover, the coverage of coding and non-coding polymorphisms genotyped within a given candidate gene may be sparse, and pathways of functionally-related genes are rarely simultaneously investigated. In the present study, we evaluated seven groups of functionally-related obesity candidate genes (30 total genes) in two racially divergent population groups (European Americans and Alaskan Yup'ik Eskimos). Many of the candidate genes were selected from hypothalamic arcuate pathways since the hypothalamus integrates complex neuroendocrine, autonomic, and behavioral signaling pathways that determine food intake, energy expenditure, and nutrient partitioning (fig. ).
Fig. 1 The molecular pathways affecting energy homeostasis. Depicted are candidate genes focused on hypothalamic control of energy homeostasis. The location of neurons in the hypothalamus and brainstem, and cells in the stomach and adipose tissue expressing (more ...)
The seven groups of genes were included amongst the candidate genes:
- (1) The leptin pathway included the hormone leptin (LEP), its receptor (LEPR), and downstream signaling molecules (JAK2, SOCS3, and STAT3).
- (2) The melanocortin pathway included agouti related peptide (AGRP), melanocortin 4 receptor (MC4R), proopiomelanocortin (POMC), and carboxypeptidase E (CPE) that processes preprohormones.
- (3) The ghrelin pathway included ghrelin (GHRL) and its receptor, GHSR.
- (4) The glucagon-like peptide 1 pathway included GCG and its receptor GLP1R.
- (5) The neuropeptide Y pathway included the receptors NPY1R and NPY5R.
- (6) The serotonin pathway included the serotonin receptors HTR2A and HTR2C.
- (7) The Bardet-Biedl group included the genes identified for BBS1, BBS2, ARL6, BBS4, BBS5, MKKS, BBS7, and TTC8.
Five genes that encode proteins involved in the leptin signaling pathway (LEP
were selected for investigation since leptin is a key hormone secreted in proportion to peripheral fat mass as a signal to the hypothalamus regarding the state of long-term energy stores [14
]. Central regulation of energy homeostasis begins as leptin binds receptors on (at least) two sets of neurons in the arcuate nucleus to decrease transcription of Agouti Related Peptide (AgRP)
and Neuropeptide Y (NPY)
or to increase transcription of Pro-Opiomelanocortin (POMC)
and Cocaine and Amphetamine Related Transcript (CARTPT)
that act reciprocally to increase and decrease food intake, respectively [14
]. AgRP is a naturally occurring inverse agonist of MC3R and MC4R that stimulates food intake. Agouti Related Peptide (AGRP)
, Melanocortin 4 Receptor (MC4R)
and Carboxypeptidase E (CPE)
represent orexigenic (AGRP)
and anorexigenic (POMC)
signaling pathways and downstream effector molecules (MC4R)
. CPE processes prohormones in this pathway including POMC. Neuropeptide Y receptors (NPY1R and NPY5R) are expressed on downstream effector neurons responding to the orexigenic NPY signaling events originating in the arcuate nucleus. The orexigenic peptide ghrelin (GHRL) is secreted from the stomach and duodenum and binds to the ghrelin receptor (GHSR) in the arcuate nucleus activating AGRP/NPY neurons to stimulate food intake. Glucagon-like peptide 1 (GCG) is produced by the small intestine and neurons in the nucleus of tractus solitarius post-prandially and decreases food intake and slows gastric emptying upon binding to the glucagon-like peptide 1 receptor (GLP1R) in the brain. Serotonin is involved in central signaling of satiety in hypothalamic downstream effector neurons and acts through the serotonin receptor genes (HTR2A
. Lastly, included in the list of candidate genes are eight genes for Bardet-Biedl syndrome (BBS1
), a syndromic, oligogenic form of obesity resulting from mutations in at least twelve genes, some of which are apparent components of a ciliary motor that can individually, or in some cases in combination, produce a dysmorphic phenotype that includes obesity, polydactyly, and retinopathy [15
In summary, our candidate gene list (table ) includes genes encoding signaling peptides, receptors and downstream signaling molecules expressed in central nervous system and/or peripheral tissues, including adipocytes and gastrointestinal tract, involved in food intake, energy expenditure or energy storage. Bardet-Biedl syndrome is also caused by mutations within a common molecular pathway (ciliary structure), and hence also provides a paradigm for genetic variants in multiple candidate genes within the same pathway interacting to affect adiposity.
This study was designed to prioritize the relative contributions of individual sequence variants, as well as to examine interactions among functionally related candidate genes and pathways. In an effort to identify more universally relevant genes/alleles, two racially divergent populations were genotyped in this study: a population of European American adults living in the New York City tri-state region, and a population of Yup'ik Eskimos from southwest Alaska. GHRL was associated with BMI in the NY European Americans, even after correction for multiple testing, suggesting that genetic variation in GHRL may contribute to relative adiposity in this population.