Introduction

The most recent National Health and Nutrition Examination Surveys (NHANES) indicate that childhood obesity in the United States has approximately doubled during the past three decades, while adolescent obesity has more than tripled during the same period of time [1, 2]. These studies also report that although the prevalence of childhood and adolescent obesity has not changed since 2007-2008, the prevalence of obesity among 2–5-year-olds (12.1 %), 6–11-year-olds (18 %), and 12–19-year-olds (18.4 %) remains at the highest reported levels since the establishment of NHANES. Similarly, the most recent NHANES reported that although the prevalence of adult obesity has not changed since 2003-2008, the prevalence of obesity among males (35.5 %) and females (35.8 %) remains at the highest reported levels [3].

It is generally accepted that an increased availability of inexpensive calorie-dense foods, a sedentary lifestyle, the built urban environment, and rural food deserts have a primary role in producing what has become known as an “obesogenic environment,” which is primarily responsible for the current obesity epidemic [4–6]. However, until recently our knowledge concerning the role of obesity susceptibility genes in contributing to this major health problem has remained undefined. It has now become increasingly clear that individual heritability of obesity susceptibility genes and interaction with components in the obesogenic environment, particularly dietary macronutrients, such as saturated fats and refined carbohydrates, promote weight gain responsible for obesity [7, 8].

The objective of this review article is to provide an update on i) the identification of obesity susceptibility genes, ii) the interaction of obesity susceptibility genes with dietary macronutrients, and iii) the metabolic pathways that have a central role in regulating energy balance. It is anticipated that increased knowledge of obesity susceptibility genes and interaction with dietary macronutrients that regulate metabolic pathways will result in more effective individual, family, and community-based preventative lifestyle intervention, in addition to targeted nutritional and medicinal therapies.

Energy Balance

The weight gain responsible for obesity has traditionally been considered in terms of energy balance, the basic components of which include energy consumption, energy expenditure, and the storage of excess energy in the form of triacylglycerol within tissues, such as adipose [9•]. When energy in the form of calories obtained from food and drink exceeds i) energy expenditure through energy required to maintain resting metabolic rate (RMR), ii) energy necessary for absorption and metabolism, and iii) energy expended during physical activity, a state of positive energy balance results with the storage of this energy in the form of triacylglycerol within adipose tissue. In contrast, if energy consumption is less than energy expenditure, a state of negative energy balance results characterized by the mobilization of energy from adipose tissue. Therefore, any genetic or environmental component that alters body weight must involve alterations in energy balance over a period of time [10].

Heritability of Obesity and Environmental Factors

It has been recognized for several decades that obesity is a heritable disorder. A landmark study using 250 monozygotic and 264 dizygotic twins demonstrated that familial genetics serves as a strong influence in predisposing individuals to overweight and obesity [11]. However, it also has become apparent that susceptibility genes interact with undefined environmental factors to increase adiposity and body weight, in what has formally become known as a “gene-environment interaction” and defined as “a response or adaptation to an environmental agent, a behavior, or a change in behavior conditional on the genotype of the individual” [12]. The first study clearly demonstrating a gene-environment interaction in relation to weight gain was performed with 12 monozygotic twins who consumed a 1,000 kcal/day surplus of calories for a period of 100 days while maintaining a sedentary lifestyle [13]. The results from this study showed a significant within twin-pair resemblance in adaptation to the excess calories (≥3 times more variance in response between twin-pairs than within twin-pairs in relation to increased body weight), suggesting that genetic susceptibility influenced excess fat storage. A more recent study has quantified the genetic and environmental influences of body mass index (BMI) and waist circumference among children living in an obesogenic environment and reported a substantial heritability for both measures of adiposity (77 %) along with providing a recommendation for more intensive lifestyle intervention for individuals who are genetically predisposed [14]. Together, it has been estimated from several studies that obesity susceptibility genes contribute an estimated 40-70 % to variation in BMI within populations [15•].

Dietary Macronutrients and Weight Gain

The increased consumption of a high-fat diet, particularly a high-fat diet enriched with saturated fatty acids, has been found to be strongly associated with increased adiposity in overweight and obese children [16, 17]. Moreover, a recent study indicates that increased consumption of saturated fatty acids (but not plant protein, carbohydrates, or other types of fat) are highly associated with increased body weight among children at 6 and 18 months of age [18]. Consistent with these results, studies indicate that obesity susceptibility genes tend to preferentially interact with saturated fatty acids, but not monounsaturated fatty acids or polyunsaturated fatty acids, to promote weight gain [19, 20]. More recent studies have demonstrated that a greater intake (≥2 servings per day) of sugar-sweetened beverages (fruit punches, lemonades, and colas) interact with obesity susceptibility genes and associated with adiposity [21•]. Together, these results are consistent with saturated fatty acids and refined carbohydrates interacting with obesity susceptibility genes to promote weight gain, while diets with high protein content and low glycemic index are protective against obesity among children and adults as reported in the Diet, Obesity, and Genes (DiOGenes) randomized family-based studies [22, 23].

Identification of Obesity Susceptibility Genes

A decade (1996-2005) of intense investigation performed using candidate gene approaches and gene-linkage analysis culminated in the identification of 127 obesity susceptibility genes [24]. However, due to a general lack of power (low number of mice or participants used in these studies) and an inability to reproduce results obtained from these studies, only a small percentage of candidate genes now comprise the current list of approximately 50 validated obesity susceptibility genes identified using genome-wide association studies (GWAS). The more powerful GWAS performed during the past several years have revealed the robust association of gene variants with different measures of adiposity (body weight, BMI or class of obesity, and percent body fat) using tens of thousands of individuals [15•, 25, 26].

In brief, the first wave of GWAS performed using BMI as a measure of adiposity identified the FTO gene, which to date possesses the largest effect size (~1,100 g/risk allele) [27]. A second wave of GWAS performed using BMI as a measure of adiposity confirmed the FTO gene and identified the MC4R gene which possesses a more moderate effect size (~650 g/risk allele) [28]. A third wave of GWAS performed using BMI as a measure of adiposity identified several novel obesity susceptibility genes (TMEM18, GNPDA2, SH2BI, MTCH2, KCTD15, NEGR1, ETV5, BDNF, SEC16B, FAIM1, NPC1, MAF, and PTER), which also possess moderate effects sizes (~200-900 g/risk allele) [29–31]. Finally, a fourth wave of GWAS performed using BMI as a measure of adiposity identified an additional 20 novel obesity susceptibility genes (SLC39A8, PRKD1, GPRC5B, QPCTL, RBJ, TFAP2B, MAP2K5, NRXN3, LRRN6C, CADM2, FANCL, FLF35779, LRP1B, MTIF3, TMEM160, TNNI3K, ZNF608, NUDT3, PTBP2, and RPL27A) which also possess moderate effect sizes (~200-600 g/risk allele) [32].

These initial GWAS performed using BMI served as a good estimate of adiposity but did not accurately reflect the overall distribution or percent body fat among individuals. Therefore, a number of GWAS have since been performed using measures of waist circumference, waist/hip ratio, and percent body fat. The first wave of GWAS performed using waist circumference and waist/hip ratio identified two novel obesity susceptibility genes (MSRA and LYPLAL1) along with four previously identified obesity susceptibility genes (FTO, MC4R, TFAP2B, and NRXN3) identified using BMI [33, 34]. A second wave of GWAS performed using waist/hip ratio identified 13 novel obesity susceptibility genes (RSPO3, NFE2L3, VEGFA, TBX15-WARS2, GRB14, DNM3-PIGC, ITPR2-SSPN, HOXC13, LY86, ADAMTS9, ZNFR3-KREMEN1, NISCH-STAB1, and CPEB4) after adjustment for BMI [35]. A GWAS and population-based study performed using percent body fat as a measure of adiposity identified two novel obesity susceptibility genes (IRS1 and SPRY2) along with two other obesity susceptibility genes (FTO and NPC1) previously found to be associated with measures of adiposity using BMI [36, 37]. Finally, the latest GWAS meta-analysis of a population-based discovery cohort with 33,533 individuals has reported that two novel genes (FGF21 and TANK) along with the FTO gene are associated with the total consumption of protein and carbohydrate [38].

To date, probably the best characterized human obesity susceptibility genes that interact with dietary macronutrients to promote positive energy balance represent a gene identified using a candidate gene approach followed by extensive validation with epidemiological studies and two genes identified using GWAS. These obesity susceptibility genes are the apolipoprotein A-2 (APOA2) gene, the fat mass and obesity (FTO) gene, and the Niemann-Pick C1 (NPC1) gene. The following sections will provide a brief update describing what is known concerning the APOA2, FTO, NPC1 genes and interaction with a high-fat diet to promote positive energy balance responsible for obesity.

APOA2 Gene

The APOA2 gene is a member of the apolipoprotein multigene (APOA1, APOA2, and APOA4) family for which the encoded APOA2 protein primarily associated with high-density lipoprotein (HDL) modulates activity of lipoprotein lipase to influence liver lipogenesis and adipose lipolysis [39–42]. An initial study indicated that individuals homozygous affected for the loss-of-function APOA2 variant (rs5082) had increased measures of adiposity (body weight, BMI, and waist circumference) and characterized with increased consumption of food composed of fat and protein compared with individuals homozygous normal or heterozygous for this gene variant [43]. A second study identified an association between individuals homozygous affected for the same gene variant and consumption (≥22 g/day) of saturated fat (but not unsaturated fat) compared with individuals homozygous normal or heterozygous for this gene variant [44••]. Finally, a third study indicated an interaction between this gene variant and consumption (≥22 g/day) of saturated fat (but not unsaturated fat) resulting in a 6.8 % greater BMI compared with individuals homozygous normal or heterozygous from his gene variant in the same population [45].

FTO Gene

The fat mass and obesity (FTO) gene is a member of the non-heme dioxygenase family that encodes the FTO protein localized to the nucleus where it functions to catalyze demethylation of nucleic acid bases (1-methyladenine, 3-methylthymine, and 3-methylcytosine) associated with single-strand DNA or RNA. An early study indicated that children (4-10 years of age) with the loss-of-function FTO gene variant (rs9939609) have an increased body weight and BMI, which also were associated with increased energy intake (but not weight of food that was consumed or resting energy expenditure), thereby suggesting a preference for energy-dense foods responsible for the weight gain [46••]. A subsequent study demonstrated that individuals homozygous affected for the FTO gene variant have increased BMI resulting from consuming increased saturated fats after adjustment for total energy consumption compared to individuals homozygous normal or heterozygous for this gene variant [47]. Similarly, another study indicated that children homozygous affected for the FTO gene variant tended to consume a greater percentage of calories from fat and possessed loss-of-control eating compared with individuals homozygous normal for this variant [48]. However, in contrast to these studies, a recent study comprised of five European countries (European Prospective Investigation into Cancer and Nutrition or EPIC) demonstrated that the FTO gene variant (9939609) does not interact with dietary factors (fat, protein, carbohydrates, or glycemic index) yet still associated with BMI and waist circumference [49]. Therefore, additional studies are now required to resolve whether the FTO gene variant actually interacts with dietary components, particularly saturated fatty acids, to promote weight gain responsible for obesity.

NPC1 Gene

The Niemann-Pick C1 (NPC1) gene is a member of the sterol-sensing resistance-nodulation-division family which encodes the NPC1 protein primarily associated with a unique late endosome-like compartment [50, 51]. Studies have demonstrated that the NPC1 protein has a central role in regulating the transport of lipoprotein-derived lipids (cholesterol and fatty acids) from late endosomes/lysosomes to other cellular compartments [52, 53]. To date, studies have found that the NPC1 gene variant (rs1805081) is associated with measures of extreme obesity (early-onset and morbid-adult) and percent body fat [29, 36, 54]. Consistent with these results, Npc1 heterozygous (Npc1 +/-) mice characterized with decreased gene dosage are predisposed to weight gain when fed a high-fat diet (but not when fed a low-fat diet) compared with wild-type mice, consistent with a gene-diet interaction responsible for weight gain [55]. Moreover, a recent study indicates that the Npc1 gene is down-regulated by dietary fatty acids (but not dietary cholesterol) through feedback inhibition of the sterol regulatory element-binding protein (SREBP) pathways [56•]. Therefore, additional studies must be performed to further characterize the Npc1 gene in relation to positive energy balance responsible for obesity.

Metabolic Pathways and Regulation of Energy Balance

The functional integration of SREBP, peroxisome proliferator-activated receptor (PPAR), and other ligand-binding transcription factors have a central role in regulating the expression of genes responsible for macronutrient oxidation and fatty acid/triacylglycerol synthesis. The metabolic pathways involved have largely been demonstrated with the use of specific ligands for each of these transcription factors. For example, the PPARα agonist class known as fibrates are prescribed clinically to decrease hyperlipidemia [57]. These agonists specifically activate PPARα to increase transcription of genes involved in β-oxidation of fatty acids, thus diminishing liver production of triacylglycerol that would otherwise be packaged into very low density lipoproteins (VLDL) and secreted into the bloodstream [58]. Evidence that PPARα activators increase β-oxidation of fatty acids and reduce body weight is provided by the observation that mice given fenofibrate while fed a high-fat diet for a period of 14 weeks weigh significantly less than mice fed the high-fat diet without fibrates, even though there was no difference in the calories consumed by these two groups of mice [59]. Moreover, PPARα null mice become obese with age while in contrast transgenic mice with continuous up-regulation of PPARα are protected from obesity [60, 61]. It is now well known that PPARα is activated by diverse fatty acids and their metabolites and that omega-3 fatty acids can be prescribed clinically to lower hypertriacylglycerolemia [58, 62]. Adiposity also may be reduced by omega-3 fatty acids; 6 grams of dietary fat replaced by 6 grams of fish oil rich in omega-3 fatty acids for a period of 3 weeks significantly reduced body fat mass and reduced respiratory quotient (providing evidence of enhanced β-oxidation of fats) among healthy adults [63].

The current understanding of metabolic pathways with a central role in regulating oxidation or storage of energy and the regulatory elements that respond to dietary macronutrients, mainly monosaccharides and fatty acids, are provided (Fig. 1). The transcription factors that increase expression of genes involved in fatty acid β-oxidation are part of the nuclear receptor superfamily and include PPARα, PPARβ/δ, and retinoid X receptor (RXR) [64]. In addition, liver X receptors (LXR), carbohydrate response element-binding protein (ChREBP), and SREBP-1c are involved in de novo lipogenesis, while PPARγ has a role in the differentiation of preadipocytes and mesenchymal stem cells required to mature adipocytes for the storage of excess triacylglycerol [65, 66]. Differential tissue expression of these regulatory transcription factors mirrors the predominant metabolism of dietary macronutrients in these tissues. For example, the expression of PPARα and PPARβ/δ in muscle is responsible for fatty acid β-oxidation in this tissue, while the expression of SREBP-1c and PPARγ in liver and adipocytes is responsible for the synthesis and storage of triacylglycerol in these tissues [58, 67].

Fig. 1
figure 1

Transcription factors that have a central role in maintaining energy balance and partitioning. Liver X receptor (LXR), retinoid X receptors (RXR), sterol regulatory element-binding protein-1c (SREBP-1c), peroxisome proliferator receptor element (PPRE), peroxisome proliferator-activated receptor α (PPARα), carbohydrate response-element binding protein (ChREBP), carbohydrate response element (ChoRE), LXR response element (LXRE,) sterol regulatory element (SRE). LXR ligands include oxysterols and glucose. PPAR ligands include fibrate drugs, thiazolidinedione drugs, saturated and unsaturated fatty acids, and phytanic acid. RXR ligands include 9-cis retinoic acid, arachidonic acid, docosahexaenoic acid, phytanic acid.

Full size image

Dietary fatty acids serve as natural ligands for PPARs which function as fatty acid sensors. PPARα is activated by both saturated and unsaturated fatty acids to increase expression of genes encoding enzymes that catalyze β-oxidation of fatty acids in mitochondria and peroxisomes. Fatty acids with carbon lengths of less than 14 or greater than 20 do not serve as ligands for PPARα as effectively as the more common dietary fatty acids [58]. In brief, PPARα forms an obligatory dimer with RXR (which is activated by 9-cis retinoic acid) followed by this complex binding to the peroxisome proliferator receptor element (PPRE) within the promoter region of genes that encode enzymes that catalyze β-oxidation of fatty acids [64]. It has also been found that activation of LXR by agonist ligands (glucose and oxysterols) prompts formation of an LXR-RXR complex and binding to LXR binding elements while at the same time inhibiting formation of the PPARα:RXR complex and binding to the PPREs [68, 69]. Thus, high levels of cellular glucose promote fatty acid and triacyglycerol synthesis.

In contrast to PPARα that responds to increased amounts of fatty acids to increase β-oxidation of fatty acids, the natural ligand for PPARγ remains undefined. The function of PPARγ in differentiation of adipoctyes and in insulin-sensitization has been elucidated through the use of drugs known as thiazolidinediones, which serve as strong synthetic ligands [58]. Arachidonic acid and eicosapentaenoic acid are the natural fatty acid ligands that bind most readily, but elicit only weak activation, and many other small lipophilic molecules, medium chain fatty acids and eicosanoids also elicit weak activation [70–72]. Despite their efficacy in improving insulin sensitivity, thiazolidinediones have many side effects. Therefore, new partial agonists that may improve insulin sensitivity without toxic side effects are being sought [72]. Weight gain is considered one of the undesirable side effects. However, adipocyte hyperplasia induced by these drugs may be one of the important insulin-sensitizing mechanisms of PPARγ activation and weight gain from hypertrophy of adipocytes and insulin resistance may result from lower activity of PPARγ and other adipogenic genes [73]. Consistent with these results, a recent study has demonstrated an age-dependent (1-6 years of age) and gender-dependent (female) interaction between a PPARγ gene variant (Pro12Ala) and saturated fatty acids/total fat that was positively associated with BMI, waist circumference, and skinfold thickness [74••].

Dietary carbohydrates activate the ChREBP and SREBP-1c transcription factors responsible for glycolysis, de novo fatty acid synthesis, and production of triacylglycerol in liver and adipocytes when excess carbohydrates are consumed [65, 75]. As mentioned previously, physiological levels of glucose in the liver activate LXR. In addition, when there are high glucose levels, the glucose-6-phosphate metabolite interacts with the low glucose inhibitory domain to prevent inhibition on the glucose response conserved element (GRACE) so that ChREBP enters the nucleus, dimerizes with max-like protein X (Mlx), and binds to the carbohydrate response element (ChoRE) in the promoter region of genes involved in fatty acid synthesis [75]. Genes encoding pyruvate kinase, acetyl CoA carboxylase, fatty acid synthase, stearoyl-CoA desaturase 1, and glycerol-3-phosphate acyltransferase are all induced by both ChREBP:Mlx heterodimers and SREBP-1c that interact with response elements in their promoter regions whenever carbohydrate intake exceeds glycogen storage capacity [65, 76, 77].

Polymorphisms in these transcription factors underscore the importance of this pathway in healthy accumulation and distribution of body fat. For example, the many mutations in human PPARγ that result in loss or lowering of function are associated with less total body fat, lipodystrophy and/or insulin resistance [70]. Common polymorphisms in PPARγ explain only 1 % of variation in BMI [78]. Several PPARα polymorphisms have been identified [64]. At least one genetic variant (Val227Ala) was associated with lower rates of obesity, but exacerbated the amount of liver fat in patients with nonalcoholic fatty liver disease [79]. Similarly, polymorphisms in SREBP-1c and LXR have been found to be associated with overweight, obesity and/or morbid obesity [80, 81].

Another current area of research focuses on the epigenetic heritability of body weight that may occur through interaction with environmental factors, especially nutrition. Epigenetics is generally defined as the study of heritable changes in gene expression or function without modification of the DNA sequence [82]. For example, hypermethylation or hypomethylation of nucleotide bases, primarily cytosine, and histone modification through acetylation, methylation or phosphorylation is capable of modulating the transcription of genes involved in regulating energy balance [83, 84]. Moreover, nutritional status can change dietary methylation molecules during prenatal or early postnatal life and has been found to change methylation status for several genes involved in metabolic or endocrine function, changes of which are associated with childhood or adulthood adiposity [85, 86]. Another mechanism of epigenetic adaptation occurs through microRNAs, generally derived from noncoding areas of the genome and active in suppressing gene expression. However, recently the surprising ability of microRNAs from food (rice, Chinese cabbage, wheat, and potato) that are stable to cooking and accumulate in human plasma to modify mammalian gene expression in significant ways has been published; thus plant MIR168a can decrease expression of low-density lipoprotein receptor adapter protein 1 in liver and decrease LDL removal from mouse plasma [87•]. This discovery suggests a novel type of interaction between nutrition and gene function. However, epigenetic-environment interactions in relation to obesity remains controversial, because it is difficult to determine whether epigenetic modifications causes obesity or whether obesity causes epigenetic modifications [88, 89]. As a result, whether epigenetics can be applied to clinical prevention or treatment of obesity will require further investigation [90].

Conclusions

A large number of obesity susceptibility genes have now been identified, but only a limited number were shown to interact with environmental components, such as various dietary macronutrients, to alter regulation of metabolic pathways that influence energy balance and partitioning. Although the identification of these obesity susceptibility genes has occurred in a relatively short period of time, it is suspected that defining the molecular basis for how these genes interact with environmental components to promote weight gain will take much longer. Nonetheless, it is anticipated that increased knowledge of obesity susceptibility genes and interaction with dietary macronutrients that regulate metabolic pathways will result in more effective individual, family, and community-based preventative lifestyle intervention, in addition to targeted nutritional and medicinal therapies.