Background
Some persons are genetically resistant to obesity, but only a few studies have evaluated thinness genes for preventing obesity. We aimed to investigate the association of polygenic variants with being underweight and their interaction with the lifestyles of middle‐aged and elderly persons and identify potential new genetic approaches for managing body weight.
Methods
In total, 58,701 participants aged 40–77 years were recruited from urban hospitals in Korea. Underweight (case) was defined as body mass index (BMI) < 18.5 kg m2 (n = 991) and normal weight (control, n = 21,921) was defined as 18.5 ≤ BMI < 23 kg m2. A genome‐wide association study was run to identify thinness‐related single nucleotide polymorphisms (SNPs) after adjustment for compound factors using Gplink. The generalised multifactor dimensionality reduction program was used to identify the genetic variants with SNP–SNP interactions. The polygenic risk score (PRS) was calculated by summing up the number of risk alleles in each SNP and classifying them into low‐, medium‐ and high‐PRS.
Results
The best model included the ANK2_rs7656666, CAST_rs28042, SLC1A3_rs928431867, CHST12_rs2906173, ALOX5_rs1051713, RGS6_rs17180754, ST8SIA5_rs79491311 and DCC_rs35721894 alleles. The participants with high‐PRS had a lower BMI (p < 0.0001) than those with low‐PRS and were 3.834 (2.58–5.70) times more likely to be underweight after multivariate adjustment (p < 0.001). The selected SNPs were correlated with each other and highly expressed in brain‐related genes. The genes with minor alleles of CAST_rs28042 and CHST12_rs2906173 exhibited a higher expression frequency in brain‐related tissues. PRS had significant interactions with protein, sodium, indigestible carbohydrates, calcium intake and exercise (p < 0.05), influencing the underweight state. People with a high‐PRS were more underweight than those with low‐PRS under high protein, sodium, high calcium, low indigestible carbohydrate intake and low exercise by 3.75, 3.88, 7.05, 3.18 and 3.80 times, respectively (p < 0.0001).
Conclusions
In conclusion, adults having a high‐PRS were significantly correlated with being underweight, especially in combination with a particular nutritional status. These results show the potential for thinness genes to be applied to personalised nutrition for preventing obesity through targeted gene therapy.