Understanding how epigenetics influences weight

ResearchBlogging.org In human weight loss studies, response to a given restriction of calories does not produce the same level of weight loss in every subject.  Though genetic factors clearly have some role per monozygotic twin and gene manipulation studies, even then there are differences.  Epigenetics is a likely candidate to explain some of these observations.

So, Bouchard et al. (1) studied the effects of restricting calories on epigenomic and transcriptomic responses.

14 overweight/obese women were subjected to a 500-800 kcal reduction from baseline resting metabolic rate for 6 months as part of another study.  They were classified by how they responded to the reduction; high responders (HR) lost more than 3% of their body fat (average 7.34 +/- 3.05 kg), and low responders (LR) lost less than 3% body fat (average 1.81 +/-1.84 kg).  The range of fat loss was 0.77 kg to 12.98 kg, a substantial range.  Subjects were matched for baseline age, BMI, % body fat, blood pressure, lipids, glucose, insulin, and changes in fat-free mass.

No significant differences in changes of blood pressure, lipids, glucose, or insulin existed.  (This is notable and corroborates the existing evidence on the lack of effect of insulin/glucose differences on weight loss, reviewed in a future post)


Prior to the calorie reduction, 35 genes were identified that were methylated differently (relatively) in adipose tissue between the 2 groups; 3 were hypomethylated and 32 hypermethylated in the HR group.  After reduction, only 3 genes were hypermethylated in the HR group.  The baseline gene differences included those that affect weight and and insulin secretion.


Gene expressions from adipose tissue before the calorie reduction began showed no differences in tested gene expressions (of 47,000) between the 2 groups, but after the 6 months, 334 expressions were upregulated and 342 downregulated in the HR group compared to the LR.  These results suggest that gene expressions are a poor predictor of fat loss, corroborating a recent study by Mutch et al. (2007).  Though, as discussed below, one study in mice suggests otherwise (2).  Among the most notable differences (most studied), cholesterol ester transfer protein (CETP), associated with lipid metabolism in adipose, was expressed greater in the HR group.  Also, genes associated with angiogenesis were down regulated in adipose of HR, suggesting a reduction of blood flow to the fat cells may be a reason for the improved outcome.  Leptin gene expression was downregulated by 1.24 fold in the HR group compared to the LR after the caloric reduction.

The finding that promoter methylation patterns and gene expressions only correlate for a small number of genes suggests an added complexity when interpretating these analyses.  The authors note the additional roles of methylation for protection of the genome against virus and in DNA replication.

The authors noted that limitations exist with the method used to analyze the methylation patterns include that only a fraction of the genome, thus further research is needed to study the relationships.

Animal models

The observation of these dramatic differences in weight loss is not limited to humans.  The C57BL/6J strain of mouse used frequently in studies that require a model of obesity, is genetically indifferent.  Compared to other strains that are more resistant to obesity, it shows unequivocally that certain genetic patterns contribute to obesity, but gauging the amount of contribution is difficult.

Koza et al. (2) noted (along with others) a large variability in fat gain and diabetes incidence in the C57BL/6J strain, even with the same genome and under the same environmental conditions.  Thus, they hypothesized that “unstable stochastic mechanisms, or stable epigenetic modifications” may be reason for the differences.

In a creative experiment, they again found low and high gainers to diets of the same (initial) energy value and macronutrient proportions.  Gene expressions prior to diet manipulations also didn’t differ significantly between the groups with the exception of a few including: leptin, genes of the Wnt family (inhibit preadipocyte differentiation into mature), and those related to angiogenesis (also similar to human results).  Leptin was strongly correlated to adiposity (note leptin gene expression in the human study).  These were correlated with adiposity after the diet manipulations.

Interestingly, they found that the obese phenotypes were maintained even after restricting calories, similar to human observations.  Epigenetic mechanisms may therefore better predict how much weight (if any) one will lose upon reducing energy intake, and perhaps lead us to interventions that can modify these mechanisms.

These findings emphasize how important it is to have large subject numbers in weight loss studies because of the variability of weight loss among subjects.

Since this is an area still without much research, it is difficult to draw conclusions yet.  But the possibilities are intriguing.


1.  Bouchard L, Rabasa-Lhoret R, Faraj M, Lavoie M, Mill J, Pe´russe L, & Vohl M (2009). Differential epigenomic and transcriptomic responses in subcutaneous adipose tissue between low and high responders to caloric restriction American Journal of Clinical Nutrition : 10.3945/ajcn.2009.28085

2.  Koza RA, Nikonova L, Hogan J, Rim JS, Mendoza T, Faulk C, Skaf J, & Kozak LP (2006). Changes in gene expression foreshadow diet-induced obesity in genetically identical mice. PLoS genetics, 2 (5) PMID: 16733553