While considerable research has been focused on the biological and environmental mechanisms of food intake, the main focus has been the understanding of the basis for total calorie consumption. In contrast, little emphasis has been directed to the mechanisms underlying food preference. Given the abundance and low-cost of high-calorie foods in Western societies, it has been proposed that a strategy to address obesity would be to favor the choice of foods with lower fat content and calories.1 As food choice becomes central to addressing obesity, it is particularly relevant to study the fundamental biological mechanisms underlying human food preferences. This has traditionally been ascribed to cultural factors, but we hypothesize here that there may be a genetic basis for food choice.
Human diet has changed remarkably during evolution. Early Hominidae are believed to have been herbivorous, and it has been proposed that the adaptation of early hominids to eat diverse foodstuffs made them particularly well suited for life in different environments.2 Moreover, the extended life span of humans may have evolved, at least in part, to allow for the prolonged training to acquire meat.3 Finch and Stanford suggested that this dietary shift to increased consumption of fatty animal tissue in the course of hominid evolution was mediated by selective evolutionary pressure towards ‘meat-adaptive’ genes which could have offered a protective effect by favoring resistance to the harmful effects of fat, toxins and pathogens.4 We present here an alternative hypothesis, which is that there might be a genetic basis for food preferences, affecting the choice of meat in human diet. To test this hypothesis, we conducted a study to determine whether a functional variant of the serotonin receptor 5-HT2A gene is associated with choice of food and micro and macronutrient intake in an outbred, elderly population consuming a stable diet.
There are extensive data from animal and human studies indicating a pivotal role for central serotoni-nergic function in the regulation of energy balance and food intake. Among the various components of the serotoninergic system, we studied a specific functional polymorphism of the of 5-HT receptor subtype, 5-HT2A–T102C, which has been reported to be associated with eating disorders such as anorexia nervosa5 and bulimia nervosa.6 The T102C polymorphism is in linkage disequilibrium with the −1438G/A variant in the promoter region of the 5-HT2A receptor gene.7
We studied a sample of 240 subjects who were genotyped and had their diet patterns recorded for macro and micronutrients. This population is part of an ongoing epidemiological study of aging and non-transmissible disease in the city of Gravataí, state of Rio Grande do Sul, in the south of Brazil. Research subjects live in a large metropolitan area, and are members of a diverse and outbred population that is predominantly of mixed European ancestry. An elderly sample was selected because the diet in this age group is much more consistent and stable than in younger populations, in whom school and work contingencies may impose nutritional biases. To avoid genetic frequency bias, we studied non-related subjects. Two independent sample collections and analysis were conducted at different seasons to ensure reproducibility.
An official institutional ethics committee approved the study and informed consent was obtained from all participants. Spontaneous intake of energy and nutrients were assessed, with a specific goal of recording intake without modifying the diet. Dietary assessment for macro and micronutrient included a 24-h recall asking if the dietary report was a habitual eating pattern or not. A second dietary assessment was made to verify whether the diet record was consistent with the habitual eating pattern in another season, at least 1 year and half after the first interview. Additionally, food frequency questionnaires (FFQ) were applied.8,9
Macro and micronutrient content were analyzed using the Brazilian computer software DietWin Clinic, which is structured from Brazilian food nutritional values. Based on systematically collected, self-report data, the level of physical activity was not different among genotype subgroups.
Clinical and laboratory staffs were blind to genotype and eating behavior respectively, during all experimental procedures. Polymorphism of the 5-HT2A receptor gene was determined from DNA isolated from lymphocytes using previously described methods.10
Subjects were categorized according to 5-HT2A genotypes and the ingestion of macro and micro-nutrients was compared among them using a uni-variate one-way Anova followed by a Bonferroni post hoc test. The distributions of the T102C genotype were in Hardy–Weinberg equilibrium.
Intake of essential amino acids was also compared among subjects with different 5-HT2A genotypes. The data are shown in Table 1a. All essentials amino acids were consumed at significantly higher levels by TT subjects.
Table 1a.
Essential amino acids consumption according to 5-HT2A genotype
| Genotypes
|
P* | |||
|---|---|---|---|---|
| TT | TC | CC | ||
| Tryptophan | 1053.74 ± 464.56a | 810.75 ± 378.98b | 851.21 ± 433.10b | 0.001 |
| Methionine | 2799.77 ± 1043.64 | 2304.69 ± 964.18 | 2235.59 ± 688.42 | 0.001 |
| Valine | 4011.37 ± 2000.54 | 3289.17 ± 1595.26 | 3500.44 ± 1702.79 | 0.031 |
| Threonine | 3892.62 ± 1813.90a | 3103.92 ± 1373.48b | 3018.30 ± 1049.69b | 0.001 |
| Phenyllalanine | 3284.25 ± 1559.29 | 2588.16 ± 1171.30 | 2755.87 ± 1396.00 | 0.004 |
| Leucine | 5764.64 ± 2963.10 | 4623.76 ± 2339.90 | 4931.17 ± 2459.43 | 0.015 |
| Isoleucine | 2036.75 ± 1133.78 | 1515.63 ± 774.40 | 1638.43 ± 912.55 | 0.001 |
| Lysine | 5528.00 ± 2357.82a | 4352.32 ± 2034.79b | 4183.33 ± 1404.32b | 0.000 |
Values are expressed as mean ± s.d.
Significance buy one-way Anova. Means followed by different letters differ significantly by Bonferroni test; P < 0.05.
In order to establish the source of differences in protein intake, we analyzed food frequency as presented in Table 1b. The main animal food source whose intake was significantly associated with genotype (TT alleles) was beef. No other food type was consumed differently according to differences in 5-HT2A genotype. The intake of each vitamin and mineral type was analyzed in relation to genotype, and no differences were found (data not shown).
Table 1b.
Frequency of animal food consumption according to 5-HT2A receptor gene T102C polymorphism in a stable elderly Brazilian population
| Animal protein food and products related | Genotypes (%)
|
P* | ||
|---|---|---|---|---|
| TT | TC | CC | ||
| Beef | 65.6 | 31.7 | 37.7 | 0.000 |
| Chicken | 46.9 | 30.5 | 45.5 | 0.073 |
| Eggs | 93.9 | 96.2 | 93.2 | 0.690 |
| Milk | 91.8 | 83.8 | 90.9 | 0.274 |
| Fish | 2.0 | 1.9 | 0 | 0.646 |
Significant by χ2 non-parametric test.
Our results show a relationship between the T102C polymorphism of the 5-HT2A receptor gene and food preference: subjects with TT genotype have higher protein intake than either CC or TC subjects. Additionally, amino-acids analyses showed that such differential intake of beef was the explanation for the differential ingestion of all essential amino acids.
All subjects had the same amount of weekly energy intake (1627 ± 524, 1474 ± 551, and 1586±529 kgCal/week for TT, TC and CC genotypes, respectively, P < 0.2–non-significant). According to systematically collected self-report data there were also no differences in physical activity among these groups. However, subjects with TT genotype had lower body mass index (BMI) than those with TC or CC alelles (BMI = 27.4 kg/m2 for the TT genotype and 29.0 and 29.1 kg/m2 for TC and CC, respectively, P < 0.04). This may be a consequence of their increased protein intake, in the context of similar overall energy intake among the three genotype groups.
Our results suggest that a polymorphism of a gene related to the serotonergic system affects eating behavior, influencing the food choice in a population with stable diet. We present here evidence for a novel concept that food preference, such as consumption of beef, can be determined at least in part by genetic factors, with a significant effect on the final outcome of body weight.
References
- 1.French SA. J Nutr. 2003;133:841S–843S. doi: 10.1093/jn/133.3.841S. [DOI] [PubMed] [Google Scholar]
- 2.Teaford MF, Ungar PS. Proc Natl Acad Sci USA. 2000;97:13506–13511. doi: 10.1073/pnas.260368897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Larsen CS. J Nutr. 2003;133:3893S–3897S. doi: 10.1093/jn/133.11.3893S. [DOI] [PubMed] [Google Scholar]
- 4.Finch CE, Stanford CB. Q Rev Biol. 2004;79:3–50. doi: 10.1086/381662. [DOI] [PubMed] [Google Scholar]
- 5.Enoch M. Lancet. 1998;351:1785–1786. doi: 10.1016/S0140-6736(05)78746-8. [DOI] [PubMed] [Google Scholar]
- 6.Nishiguchi N, Matsushita S, Suzuki K, Murayama M, Shirakawa O, Higuchi S. Biol Psychiatry. 2001;50:123–128. doi: 10.1016/s0006-3223(00)01107-0. [DOI] [PubMed] [Google Scholar]
- 7.Veenstra-VanderWeele J, et al. Am J Med Genet. 2002;114:277–283. doi: 10.1002/ajmg.10192. [DOI] [PubMed] [Google Scholar]
- 8.Kabagambe EK, Baylin A, Allan DA, Siles X, Spiegelman D, Campos H. Am J Epidemiol. 2001;154:1126–1135. doi: 10.1093/aje/154.12.1126. [DOI] [PubMed] [Google Scholar]
- 9.Hankin JH, Wilkens LR. Am J Clin Nutr. 1994;59:198S–200S. doi: 10.1093/ajcn/59.1.198S. [DOI] [PubMed] [Google Scholar]
- 10.Tsai SJ, et al. Psychiatry Res. 1999;87:233–237. doi: 10.1016/s0165-1781(99)00059-1. [DOI] [PubMed] [Google Scholar]