**4. Discussion**

In the present study, we examined the allele and genotype distributions of the *FTO* (rs9939609), *FABP2* (rs1799883), *LEP* (rs2167270), *LEPR* (rs1137101), and *MC4R* (rs17782313) polymorphisms in a group of Caucasian men, who were divided into groups depending on their BMI and FMI. When tested individually, our statistical analyses showed that harboring the specific *FTO* genotype might be associated with FMI, which measures relative fat content. This association was found for the codominant (A/A vs. T/T), dominant (A/T-A/A vs. T/T), and also for the recessive genetic model (A/A vs. T/T-A/T). The chance of having an increased FMI was over two times higher for the *FTO* A allele carriers. This observation constitutes the first important finding of the present study, implying that harboring this specific allele is unfavorable for some individuals. The carriers of the A allele might have an increased accumulation of body fat and a higher risk of many obesity-related disorders.

In 2007, three independent groups demonstrated that a cluster of polymorphisms in the *FTO* first intron was strongly related to body mass and composition parameters and predisposes to overweight and obesity in children, teenagers, and adults [25–27]. Since then, many studies have proved that the *FTO* variants, especially the common A/T polymorphism (rs9939609), are significantly linked to obesity-related traits, e.g., BMI, FMI, body fat percentage, hip circumference, cardiometabolic traits, and many obesity-related medical problems. These associations are found across various ethnic populations and

different age groups [18,25–29]. The A allele, identified as the risk allele, is linked to increased appetite and reduced satiety, a higher intake of dietary protein and fat, poor eating habits, and loss of control overeating, among others [18]. Consequently, it has been linked to the development of overweight and obesity, increasing the risk by 20–30%. About 16% of examined individuals are homozygous for the A alleles and these people weigh ~3 kg more than those without the risk allele [24]. In a previous study including 201 young women from Poland, the A allele was also associated with higher BMI [29]. Additionally, Zmijewski and Leo ´nska-Duniec showed that the SNP within the *FTO* gene can influence athlete status in a study involving 196 elite swimmers and 379 control participants, who were all Caucasians. They found that harboring the T allele might be favorable for achieving success in sports such as swimming [30]. These results are in accordance with our study, which confirms that harboring the A allele is unfavorable for Polish men, who might have an increased accumulation of body fat. These results sugges<sup>t</sup> that the *FTO* (rs9939609) polymorphism is a candidate marker for affecting body mass and body composition parameters in the Caucasian population.

A few potential biological mechanisms underlying the relationship between the *FTO* polymorphism and body mass and composition parameters have been revealed. The research has indicated that these associations are mediated through their functional interactions with distal surrounding genes. The first intron of the *FTO* gene contains a binding site for the transcriptional factor—cut-like homeobox 1 (CUX1). Through controlling retinitis pigmentosa GTPase regulator interacting protein 1 (RPGRIP1L) expression, it interacts with the leptin receptor [27,31]. The leptin signaling is mediated by this specific receptor, which in turn regulates food intake and energy expenditure [16]. Additionally, the *FTO* intron also includes an enhancer sequence, which interacts with the iroquois homeobox 3 (IRX3) promoter region, and thus the *FTO* SNPs regulate IRX3 expression in the human brain. The IRX3 relationship with obesity and the process of browning in adipose cells has been described [27,31,32]. Currently, some studies have indicated that the *FTO* gene is closely related to the regulation of levels of growth hormone and insulin-like growth factor I (IGF-1). IGF-1 is a crucial hormone in the development of metabolic syndrome, due to its influence on lipid and carbohydrate metabolism [33].

When the results obtained in our study were incorporated into the complex gene– gene interaction analysis, the novel finding was that all five studied polymorphisms are involved in the formation of obesity-related traits in the Caucasian population. These results imply that some individuals might benefit from carrying some combinations of genotypes. It was shown that for the two-locus model—*FTO* × *LEPR* interaction—both the BMI and FMI division genotypes, TT × AA and TT × AG, were associated with the absence of overweight. The same result showed the genotypes AT × AA with BMI division and AT × GG for FMI division. For three-locus model genotypes, GG × AA × TT, GG × AG × TC, GG × AG × CC, GG × GG × TT, AA × AA × TT, AA × AA × TC, AA × AG × TT, AA × AG × CC, AG × AA × TT, AG × AG × TT, and AG × GG × CC (*LEP* × *LEPR* × *MC4R*) showed a link to the lack of overweight in BMI division. The same association was shown for genotypes AT × GG × AA, AT × AG × GG, TT × GG × AG, TT × AG × AA, TT × AG × AG, and TT × AA × AG (*FTO* × *LEPR* × *MC4R*) in FMI division. For four-locus and five-locus models, only the genotypes TT and AT in *FTO* present when there was an association with the absence of overweight shown (both for BMI and FMI). In all models, when *FTO* was included, the genotypes TT and AT were linked with a lack of overweight, confirming that harboring the *FTO* T allele might be favorable for some individuals.

Although the analysis of individual SNPs showed only one association between harboring the specific *FTO* genotype and FMI, the gene–gene interaction analysis revealed numerous links between the genotypes of studied genes and obesity-related traits. This observation confirmed that the genetic risk of obesity is connected with the accumulation of numerous variants; thus, methods based on numerous SNPs are more informative than methods based on a single polymorphism. Cole et al. showed that the analysis of

gene–gene interactions is a potential source of unexplained heritability, a significant focus of research into complex traits, including obesity, which involves a complex interaction between several genetic variants. Such polygenic traits frequently require etiologies in which complicated biological relations within different tissues, pathways, and networks underlie the trait development [15]. Studying gene–gene interactions has been especially important in the context of obesity [34], which is in accordance with our results.

Although numerous studies refer to the analysis of individual SNPs, which were selected for the present study, the literature on gene–gene interaction analyses is scarce. Therefore, the obtained results cannot be discussed with direct comparisons to other studies. In the study including 2386 individuals, De et al. analyzed interactions between twelve genetic variants robustly associated with obesity (*BDNF*, *ETV5*, *FAIM2*, *FTO*, *GNPDA2*, *KCTD15*, *MC4R*, *MTCH2*, *NEGR1*, *SEC16B*, *SH2B1*, and *TMEM18*). The authors underlined that the used methodology made it possible to reveal the background of interactions between genes known to influence BMI. They characterized the complicated interactions, emphasized new roles of the genes and highlight the involvement of regulatory frameworks in the development of obesity; e.g., rs17066891 in *MC4R* was identified as having the strongest main effect within this network, rs9940128 in *FTO* was identified as having the second strongest main effect in the network, and rs4358154 in *TMEM18* had the highest score for all measures which highlights the potentially significant role of this variant in the context of obesity [34]. In a study including 290 overweight/obese participants and 288 normal-weight controls, Rana et al. examined the effects of gene–gene and gene– environment interactions on the obesity risk in the Pakistani population. They analyzed the five obesity-associated genetic variants (*MC4R* rs17782313, *BDNF* rs6265, *FTO* rs1421085, *TMEM18* rs7561317, and *NEGR1* rs2815752). Surprisingly, the gene–gene interaction was not found to significantly influence any obesity-related anthropometric phenotype, such as BMI or body fat percentage [35].

The study group was very homogeneous, which is a strong point of our study. Participants had the same living conditions, physical activity levels, and meals, and ye<sup>t</sup> the differences in the parameters related to body weight were statistically significant. This allows most of the environmental factors to be ruled out and might indicate a genetic background for the increased body weight. A potential limitation of our experiment is rather the small group size of the participants, which might not show the statistical power necessary to yield a meaningful analysis and the interpretation of the results. Previously, differences between sexes contribute to variation in the obesity-related traits such as levels of fasting glucose and insulin were described. Lagou et al. (2021) indicated that fasting insulin in women shows stronger genetic correlations than in men with waist-to-hip ratio and anorexia nervosa [36]. Unfortunately, this study only included adult men, and thus we did not have the chance to compare the results between different age groups and genders. The participants were also relatively young, healthy, and physically active, which could have influenced the results, because systematic exercise reduces body weight [37]. Additionally, it should be emphasized that this is an observational study and that no causal mechanisms can be inferred.
