An Association Between FNDC5, PGC-1α Genetic Variants and Obesity in Chinese Children: A Case-Control Study

Introduction

Childhood obesity has become one of the most important risk factors for chronic non infectious diseases, such as hyperlipidemia, hypertension, diabetes and metabolic syndrome, and is an important indicator for predicting obesity and risks of health in adulthood.^1–3 WHO reported that there were approximately 340 million children and adolescents aged 5–19 with overweight or obese worldwide.⁴ Liu et al reported that the prevalence of overweight or obesity among Chinese children aged 6–18 years was high with 17.62% and 29.05% in boys, 17.57% and 18.04% in girls, respectively.⁵ It was well known that overweight or obesity is the result of the interaction between genetic variants and unhealthy environmental factors (such as unhealthy eating behaviors, insufficient physical activity time, etc). Existing studies have shown that serum irisin and its genetic polymorphism are closely related to dyslipidemia and glucose metabolism, and may become a new target for the treatment of obesity and obesity-related metabolic diseases.⁶

Irisin is a new muscle/adipocytokine discovered in recent years to enhance fat burning and weight loss by increasing energy expenditure, improving glucose tolerance and inducing mitochondrial genetic expression.⁷ Irisin is encoded by the fibronectin type III domain containing protein 5 (FNDC5) gene.^8,9 As a precursor of irisin, FNDC5 is highly expressed in the heart, brain, liver and skeletal muscle, and is essential for maintaining metabolic homeostasis.^10,11 It has been reported that the levels of FNDC5 mRNA in adipose tissue and circulating irisin negatively correlated with hyperglycaemia, triglycerides, visceral adiposity and extramyocellular lipid deposition.¹² Studies have shown that the wild GG genotype of FNDC5 (rs16835198) is also significantly associated with increased fasting glucose levels in diabetic subjects.¹³ However, other studies have shown that the TT genotype of FNDC5 (rs16835198) is associated with several metabolic parameters in circulating serum, suggesting that both TT genotype and T allele may increase obesity susceptibility.¹⁴ Remarkably, results of these study showed the SNPs of FNDC5 gene correlated with obesity and glucose-lipid metabolism, which may be because it modulate the levels of serum irisin.^15,16

Peroxide-proliferator-activated receptor γ coactivator 1α (PGC-1α) is a powerful promoter of mitochondrial biosynthesis and oxidative metabolism, and can induce high expression of brown fat cell uncoupling protein 1 (UCP-1), which plays an important role in inducing heat production and energy consumption of brown fat.¹⁷ Two common SNPs of PGC-1α (rs8192678, rs3755863) were associated with obesity,¹⁸ which may lead to the development of metabolic syndrome, but the effects may be small and depend on interactions of gene-gene and gene-environment. PGC-1α genetic variants may be one of the genetic causes of obesity in different periods of human life. The effects of genetic variants and interaction between PGC-1α and FNDC5 genes on obesity in children and adolescents are still poorly understood. The purpose of the current study was to reveal the relationship between FNDC5 (rs16835198), PGC-1α (rs8192678, rs3755863), healthy eating behaviors, moderate-to-vigorous physical activity time and overweight or obesity in Chinese children and adolescents based on the case-control study.

Material and Methods

Participants

A total of 198 overweight or obese participants aged 8–14 years were screened from two nine-year schools in Bengbu, Anhui Province, and 198 participants with normal weight were matched according to gender and age 1:1. Sample size calculation formula: , . p₁: Risk allele frequency of FNDC5 (rs16835198) in normal weight group was 0.435; p₂: Risk allele frequency of FNDC5 (rs16835198) in obesity group was 0.497; OR=1.33;¹⁹ Z_α/2=1.96, Z_β=1.282. The minimum effective sample size for overweight or obese participants was 162. This study was conducted according to the Declaration of Helsinki and approved by the Ethics Committee of Bengbu Medical College ([2015] No. 003). All parents or their guardians signed written informed consent.

Measurement of Morphological Development Indicators

During the measurement of body height and weight, participants had empty stomach and bladder, barefoot, took off their hats, and only wore underwear. Height was measured by a mechanical scale with height metre to the nearest 0.1 cm. Digital body weight scale was used to measure the body weight to the nearest 0.1 kg.

Criteria for Overweight or Obesity

Body mass index (BMI) was calculated as weight (kg)/height (m)², and overweight or obesity was determined according to BMI Classification Standards for Overweight/Obesity Screening of Chinese School-age Children and adolescents.²⁰

Survey on Dietary Behaviors and Physical Activity Time

Food frequency questionnaire was used to investigate the frequency of healthy eating behaviors, such as breakfast, eggs, milk, fresh vegetables, fruits and nuts. The frequency of each item was assigned 7 points for 7 times per week, 5 points for 4–6 times per week, 2 points for 1–3 times per week, 0.5 points for 2 times per month, 0.25 points for 1 time per month and 0 points for never. The total score of healthy eating behaviors was divided into ≥ P₇₅ and <P_75.²¹

Physical activity time was investigated by CLASS questionnaire,²² and the time of moderate-to-vigorous physical activity time was divided into <60 min and ≥60 min.²³

DNA Extraction and SNP Typing Detection

Medical professionals collected 3mL of fasting venous blood from the participants. DNA was extracted using the salting-out method. The national institute of environmental health sciences website (https://snpinfo.niehs.nih.gov/snpinfo/snptag.html) to determine the gene to be detected SNP. The SNP of FNDC5 was rs16835198, the SNPs of PGC-1α were rs8192678, rs3755863. Genotype was performed using the SNaPshot method developed by ABI company. Table 1 contains the sequences of each primer that was used in the qPCR.

Table 1 Sequences of Primers Used for the qPCR Reaction

Statistical Analysis

IBM SPSS 23.0 software was used for statistical analysis. The data were described using proportion and mean ± standard deviation. t-test was used to compare the difference in the measurement variables between two groups. Chi-square test was used to compare the difference in proportion (%) between groups. Logistic regression model was used to analyze the associations between FNDC5, PGC-1α genetic variants and overweight or obesity in children and adolescents based on different genetic models (overdominant, dominant, recessive and additive model), after adjusting healthy eating behaviors and moderate-to-vigorous physical activity time. Generalized multifactor dimensionality reduction (GMDR) was used to analyze the interactions between FNDC5 (rs16835198), PGC-1α (rs8192678, rs3755863), healthy eating behaviors, moderate-to-vigorous physical activity time in occurrence of overweight or obesity in children and adolescents. P<0.05 was statistically significant.

Results

Basic Information

A total of 396 primary and secondary school students aged 8–14 years were enrolled in this case-control study, including 234 boys (59.09%) and 162 girls (49.91%). The height, weight among overweight or obese children and adolescents were significantly higher than those among normal weight (P<0.05). However, the healthy eating behaviors and moderate-to-vigorous physical activity time were not significantly related to overweight or obesity in children and adolescents (P>0.05), as show in Table 2.

Table 2 Association Between Healthy Eating Behaviors, Moderate-to-Vigorous Physical Activity Time and Overweight or Obesity Among Children and Adolescents

Association Between FNDC5, PGC-1α Polymorphisms and Overweight/Obesity in Children and Adolescents Based on Different Genetic Models

The distributions of SNPs of FNDC5 and PGC-1α met the Hardy-Weinberg equilibrium. GT, TT genotypes and T allele frequencies of FNDC5 (rs16835198) in overweight or obese children were significantly higher than those in normal weight children (P<0.05). TT genotype and T allele frequencies of FNDC5 (rs16835198) in girls with overweight or obesity were higher than those in girls with normal weight (P<0.05). GT and TT genotype frequencies of FNDC5 (rs16835198) in boys with overweight or obesity were higher than those in boys with normal weight (P<0.05). Based on overdominant model, GT genotype of FNDC5 (rs16835198) increased the risk of overweight or obesity in boys (OR (95% CI):1.68 (1.00, 2.93)). Based on dominant model, GG genotype of FNDC5 (rs16835198) decreased the risk of overweight or obesity in total students, girls and boys (OR (95% CI): 0.45 (0.28,0.73), 0.45 (0.21,0.97) and 0.45 (0.24,0.83), respectively). Based on recessive model, TT genotype of FNDC5 (rs16835198) increased the risk of overweight or obesity in girls (OR (95% CI): 2.46 (1.19, 5.05)). Based on additive model, TT genotype of FNDC5 (rs16835198) increased the risk of overweight or obesity in total students and girls (OR (95% CI): 2.69 (1.48, 4.91) and 3.82 (1.49,9.80), respectively). However, PGC-1α (rs3755863, rs8192678) were not significantly associated with overweight or obesity in children and adolescents based on different genetic models, as shown in Table 3 and Figure 1.

Table 3 Associations Between FNDC5, PGC-1α Genetic Variants and Overweight or Obesity in Children and Adolescents Based on Different Genetic Models

Figure 1 OR values of FNDC5 (rs16835198) and PGC-1α (rs8192678, rs3755863) associated with overweight or obesity based on different genetic models.

Interactions of Gene (FNDC5, PGC-1α) Variants, Healthy Eating Behaviors, Physical Activity Time in Occurrence of Overweight or Obesity in Children and Adolescents

Generalized multifactor dimensionality reduction (GMDR) was used to explore the interaction of gene (FNDC5, PGC-1α) variants, healthy eating behaviors, physical activity time in the occurrence of overweight or obesity in children and adolescents. The results showed that the interaction between FNDC5 (rs16835198) and PGC-1α (rs3755863, rs8192678), healthy eating behaviors, moderate-to-vigorous physical activity time was significant in the occurrence of overweight or obesity in children and adolescents (P=0.01). There was significant interaction between PGC-1α (rs8192678) and moderate-to-vigorous physical activity time in the occurrence of overweight or obesity in children and adolescents (P=0.01). As shown in Table 4 and Figure 2.

Table 4 Interaction of FNDC5 (rs16835198), PGC-1α (rs3755863, rs8192678), Healthy Eating Behaviors, Moderate-to-Vigorous Physical Activity Time in Occurrence of Overweight or Obesity in Children and Adolescents

Figure 2 Combined action model of association between FNDC5 (rs16835198), PGC-1α (rs8192678, rs3755863), healthy eating behaviors and healthy dietary behaviors, moderate-to-vigorous physical activity time and overweight or obesity in children and adolescents. Notes: The left band of each small cell represents the total of positive integral, the right band represents the total of negative integral; dark gray cell represents high risk, light gray cell represents low risk.

Discussion

Obesity has been a common and complex multifactorial disease, which leads to excessive fat accumulation under the combined action of environmental and genetic factors.^24,25 This study revealed the association between FNDC5 (rs16835198), PGC-1α (rs3755863, rs8192678), eating behaviors, physical activity time and overweight or obesity in children and adolescents. There was significant association between FNDC5 (rs16835198) and overweight or obesity, interaction between FNDC5 (rs16835198) and PGC-1α (rs3755863, rs8192678), healthy eating behaviors, moderate-to-vigorous physical activity time, and interaction between PGC-1α (rs8192678) and moderate-to-vigorous physical activity time in the occurrence of overweight or obesity in children and adolescents.

The results of this study showed that healthy eating behaviors, moderate-to-vigorous physical activity time were not associated with overweight or obesity in sample children and adolescents. Traub et al²⁶ reported that skipping breakfast and increasing screen time were risk factors for overweight or obesity in children and adolescents. Raistenskis et al²⁷ found that obese children had less physical activity than those with normal weight, and sedentary lifestyle was not conducive to the health outcomes of children and adolescents.²⁸ Of course, some studies have shown that the relationships between eating behaviors, physical activity and obesity are not significant.²⁹ In addition, it can also be related to the memory bias of the subjective measurement method of eating behavior and physical activity time. This may also be due to the fact that obese children and adolescents have a higher awareness of healthy eating behaviors and physical activity, and intentionally improve their unhealthy eating behaviors or increase physical activity time to loss weight.

This study found that the SNP of FNDC5 (rs16835198) was significantly associated with overweight or obesity in children and adolescents. T allele frequency of FNDC5 (rs16835198) in overweight or obese children was higher than that in normal weight children. Based on different genetic models, the results showed that TT genotype of FNDC5 (rs16835198) increased the risk of overweight or obesity in children and adolescents. Todendi et al³⁰ reported that TT genotype of FNDC5 (rs16835198) increased risk of obesity in South Brazilian children and adolescents. The studies showed that the obese adults with TT genotype of FNDC5 (rs16835198) had higher levels of total cholesterol (TC), triacylglycerol (TG) and low-density lipoprotein (LDL-C), lower levels of high-density lipoprotein (HDL-C).¹⁴ Tang et al³¹ reported that the interaction between FNDC5 (rs16835198) and BMI could affect fasting insulin level in Chinese population. However, the studies also showed that TT genotype of FNDC5 (rs16835198) was associated with decreased risk of T2DM, and G allele of FNDC5 (rs16835198) was associated with elevated insulin resistance in Egyptians with no effect on renal complications.^32,33 We knew that FNDC5 gene encode irisin that regulates adipose tissue metabolism, and plays an important role in energy metabolism and obesity. Studies also demonstrated that irisin is not only a myokine but also an adipokine, with important autocrine and paracrine functions.³⁴ Luo et al³ proposed that lack of irisin was associated with a poor browning response, glucose/lipid derangement, and decreased bone mass in mice. These may be a reasonable explanation for the association between T allele, TT genotype of FNDC5 (rs16835198) and overweight or obesity.

In this study, the results of generalized multifactor dimensionality reduction (GMDR) showed that there was significant interaction between FNDC5 (rs16835198) and PGC-1α (rs3755863, rs8192678) in the occurrence of overweight or obesity in children and adolescents. Boström et al³⁵ showed that in mouse that PGC1-α expression in muscle stimulated an increase in expression of FNDC5. The level of FNDC5 mRNA is increased in skeletal muscle in some exercise paradigms.³⁶ As a key regulator of energy metabolism, PGC-1α can not only regulate glucose metabolism in liver and muscle, but also regulate lipid oxidation and adipocyte differentiation. Studies suggest that PGC-1α can be a potential therapeutic target for various metabolic diseases.³⁷ Lin et al³⁸ demonstrated that A allele of PGC-1α (rs8192678) was an independent risk factor of nonalcoholic fatty liver disease in obese children and adolescents. These may be a reasonable explanation for the interaction between FNDC5 (rs16835198) and PGC-1α (rs3755863, rs8192678) in the occurrence of overweight or obesity.

The results of this study showed that there was significant interaction between FNDC5 (rs16835198) and healthy eating behaviors, moderate-to-vigorous physical activity time in the occurrence of overweight or obesity in children and adolescents, and the TT genotype of FNDC5 (rs16835198) had the most significant effect. This shows that the children with TT genotype of FNDC5 (rs16835198) are more likely to be obese if they have insufficient healthy eating behavior or moderate to high intensity physical activity. The FNDC5 (rs16835198) variant may effect encoding of irisin, thus affecting role of irisin in energy metabolism. Studies showed that exercise increased levels of irisin in children, which may be more instrumental in mediating lipid metabolism.³⁹

The interaction between PGC-1α (rs8192678) and moderate-to-vigorous physical activity time was significant in the occurrence of overweight or obesity in children and adolescents, and the GA genotype of PGC-1α (rs8192678) had the most significant effect. This means that the children with GA genotype of PGC-1 (rs8192678) and insufficient physical activity are more likely to be obese. In a meta-analysis of the PGC-1α (rs8192678) variant, the A allele and AA genotype were suggested to be beneficial for athletic performance.⁴⁰ The study showed that the adolescents with A allele of PGC-1α (rs8192678) had higher TG than those with G allele of PGC-1α (rs8192678).⁴¹ Some studies showed that PGC-1α gene regulated the induction of muscle adaptation training.⁴² Endurance training can increase expression of PGC-1α mRNA.⁴³ However, the association between PGC-1α (rs8192678) and physical performance was not reproducible in different races and populations.⁴⁴ Eynon et al⁴⁵ found that the GG genotype of PGC-1α (rs8192678) was conducive to the increase in aerobic capacity. Yang et al reported that the G allele of PGC-1α (rs8192678) was associated with endurance performance, while the A allele of PGC-1α (rs8192678) was associated with power performance in Chinese Han Population.⁴⁶

There were also some limitations. First, the subjective methods were used to measure healthy eating behaviors and physical activity time, which might have recall bias. Second, children suffering from obesity may increase their awareness of healthy eating behaviors and physical activities, so that they can change unhealthy eating behaviors or strengthen physical exercise. Third, we only studied Chinese children and adolescents, the generalizability to other ethnic groups was limited. Finally, the exact molecular mechanism of association between FNDC5 (rs16835198) and childhood overweight or obesity remained to be elucidated.

Conclusion

This study showed that FNDC5 (rs16835198) was significantly independently associated with overweight or obesity, and played an interactive role with PGC-1α (rs3755863, rs8192678), healthy dietary behaviors and moderate-to-vigorous physical activity time in the occurrence of overweight/obesity in Chinese children.