Introduction
Childhood obesity prevalence has progressively increased worldwide. Over 340 million children and adolescents aged 5–19 years were overweight or obese in 2016 (World Health Organization, 2020; Singh et al., 2023). According to the World Obesity Federation’s 2019 estimates, 206 million children and adolescents aged 5 to 19 will be obese by 2025, which is expected to increase to 254 million by 2030. Excessive fat accumulation in childhood is associated with an increased risk of the early development of cardiovascular diseases (Stefan, 2020), insulin resistance, diabetes mellitus (Hameed et al., 2020), and cancers (Di Angelantonio et al., 2016). Discovering biomarkers of predisposition to obesity and related comorbidities enables early diagnosis, identifying individuals who better benefit from non-medical interventions such as physical activity and healthy diet, and who need early treatment, thus contributing to disease prevention or better prognostic (Dessì et al., 2014).
Metabolomics is an emerging field recognized as a powerful tool for discovering the metabolic fingerprint of a given phenotypic change (Su et al., 2014). The metabolome represents the total metabolites in a given organism, such as sugars, organic acids, amino acids, lipids, vitamins, minerals, and other compounds. Genetic and epigenetic factors, age, environment, nutrition, drugs, and lifestyle reflect individual metabolome alterations. Moreover, the metabolites are not inert molecules, they can directly or indirectly regulate gene and protein expression and participate in the maintenance of homeostasis (Gerszten and Wang, 2008; Scalabre et al., 2017). Several metabolites have been correlated to disease phenotypes, as their presence or concentrations in biological fluids are directly related to pathogenic mechanisms (Su et al., 2014; Scalabre et al., 2017). Therefore, metabolomics enables the identification of biomarkers of exposure or susceptibility to several diseases, providing remarkable opportunities for better understanding exposure and predicting potential adverse health outcomes (Dessì et al., 2014; Handakas et al., 2022).
Obesity is a complex phenotype depending on several genetic and lifestyle factors. Despite being extensively studied, there is still no complete elucidation of the mechanisms underlying its development, particularly in childhood. Recent metabolomic studies have identified that the concentration of several metabolites including amino acids, lipids, monosaccharides, organic acids, and serotonin are altered in patients with obesity, thus being potential candidates as biomarkers (Hameed et al., 2020; Zhang et al., 2013; Rangel-Huerta et al., 2019; Szczerbinski et al., 2022). Nonetheless, there are fewer metabolomic studies targeting childhood obesity, partly due to methodological and ethical challenges in obtaining blood samples. These studies also need to be expanded, considering the bases of obesity, growth, and pubertal hormones.
Human urine is a rich biofluid for metabolomic studies, since it concentrates a wide range of metabolites that change with age, diet, nutritional status, and environmental exposure, thus being able to characterize a given phenotype (Bouatra et al., 2013; Slupsky et al., 2007; Wahl et al., 2012). Urine production is a constant process and its collection is easy and non-invasive (Slupsky et al., 2007). Urine metabolome has been well explored in studies with adults (Slupsky et al., 2007; Yu et al., 2012; Holmes et al., 2008), but only a few studies have been conducted with children (Zhang et al., 2013; Chiu et al., 2016; McCormack et al., 2013). Considering this, analyzing urine from prepubertal children is very attractive for metabolomic studies, as it targets the molecular mechanisms behind childhood obesity and cardiometabolic alterations.
This study aimed to investigate if there is an amino acid fingerprint in the urine of obese prepubertal children and explore possible mechanisms through which the altered metabolites can contribute to obesity and cardiometabolic diseases.
Discussion
Research groups and health organizations worldwide are unanimous in warning about the rapid increase in overweight and obesity prevalence and its consequences for individuals’ and communities’ health (Vieira et al., 2017; GBD, 2015 Obesity Collaborators et al., 2017; NCD Risk Factor Collaboration NCD-RisC, 2017). These consequences are directly related to higher health service costs, a high risk of disease development, and related mortality (Di Angelantonio et al., 2016; GBD, 2015 Obesity Collaborators et al., 2017; NCD Risk Factor Collaboration NCD-RisC, 2017).
Weight and BMI are anthropometric measures that reflect the quality of nutrition and healthiness of the living environment during childhood and adolescence. These parameters are good predictors of health and developmental outcomes throughout life (Park et al., 2012; Tanner, 1987). As expected, all anthropometric measurements were progressively increased in the groups OW and OB, as shown in Table 1.
More important than total body fat, WHtR and WHiR predict fat distribution in the upper part of the body around the abdomen, which is associated with metabolic changes. WHiR and WHtR values of OW (0.84 ± 0.01; 0.48 ± 0.01, p < 0.0001) and OB (0.89 ± 0.01; 0.55 ± 0.01, p < 0.0001) groups, respectively, compared to NW (0.80 ± 0.01; 0.41 ± 0.01) suggest a progressive increase in abdominal fat in greater degrees of excess weight. Zeng et al. (Zeng et al., 2010) looking for biomarkers suggested that the waist-hip ratio, together with total triglycerides, total cholesterol, high-density lipoprotein, and low-density lipoprotein are the most critical parameters that correlate with the metabolic disturbances of childhood obesity (Lo et al., 2016). In a systematic review of thirty-nine studies, Park et al. (Park et al., 2012) observed evidence for associations between childhood BMI and type 2 diabetes, hypertension, and coronary heart disease, as in other studies (Lo et al., 2016; Libert et al., 2018; Juonala et al., 2011; Liu et al., 2021). Moreover, uric acid levels are substantially increased in obese subjects and proportionally associated with BMI (Weihrauch-Blüher et al., 2023), as observed in our results.
Metabolomics has been used to study the metabolic signature of obesity. It describes differential responses to dietary interventions, predicts health outcomes, and allows the study of the effects of specific nutritional patterns on obesity-related metabolites (Zhang et al., 2013; Rangel-Huerta et al., 2019; Szczerbinski et al., 2022). These biomarkers can represent disease diagnostic tools for developing new therapeutic protocols (Hameed et al., 2020; Zhang et al., 2013).
Investigation of serum metabolite concentrations in obese children may lead to new insights into biological mechanisms associated with childhood obesity, for example, branch-chained amino acids and various lipid metabolites, including phosphatidylcholines, cholesteryl esters, triglycerides (Zhang et al., 2013; Szczerbinski et al., 2022). Oberbach et al. (Butte et al., 2015) identified 163 serum metabolites, 12 of which were significantly related to obesity. Among those, GLY, GLN, and glycerophosphatidylcholine 42:0 (PCaa 42:0) serum concentrations were higher, whereas PCaa 32:0, PCaa 32:1, and PCaa 40:5 were decreased in the obesity group compared to the normal weight group. Wahl et al. (Wahl et al., 2012) analyzed serum samples from obese and normal-weight children aged six to 15. Fourteen metabolites and 69 metabolite ratios were significantly different in obese children compared to normal-weight children. Butte et al. (Butte et al., 2015) observed that obese Hispanic children had increased plasma concentrations of LEU, ILEU, and VAL but lower ASN, ASP, GLY, and SER concentrations. Plasma amino acid profile has shown a strong correlation with nutritional status (McCormack et al., 2013; Butte et al., 2015; Wu et al., 2024; Morris et al., 2012).
Despite urine being a promising biological fluid for metabolomics research, few studies have been conducted with this fluid compared to the numerous studies conducted with blood samples. However, urinary metabolomics research focuses on characterizing the metabolic profile present in urine, providing invaluable insights into both physiological and pathological processes. This comprehensive analysis facilitates the discovery of biomarkers for disease diagnosis, treatment monitoring, and elucidation of metabolic pathways (Scalabre et al., 2017; Szczerbinski et al., 2022; Bouatra et al., 2013; Cho et al., 2017; Chavira-Suárez et al., 2020). Detecting metabolites in prepubertal urine represents a gap in the existing literature, and your investigation provides additional insights.
An amino acid signature characteristic of OB group was found in our study, with elevated levels of AAA, ASN, CTH, 1-MeHIS, SER, and aromatic amino acids, and reduced concentrations of APA, and SAR. The OW group presented most expressive urinary levels of APA, GLN, and SAR.
Urinary excretion of 3-MeHIS indicates protein catabolism, as it comes from skeletal muscle actin and myosin, and is also directly related to meat consumption (Cross et al., 2011). The higher levels of 1-MeHIS observed in the OB group may reflect dietary differences in meat intake. A controlled feeding study investigated 1-MeHIS and 3-MeHIS as potential biomarkers of meat intake and found a dose-dependent association between meat intake and urinary excretion of 1- and 3-MeHIS (Cross et al., 2011; Cross et al., 2014).
Our results show increased values of CTH and SER for the OB group. One-carbon metabolism is a metabolic network driven by three interrelated metabolic pathways: the folate cycle, the homocysteine-methionine cycle, and the transsulfuration pathway. If there is abundant methionine, the transsulfuration pathway will become active, by which homocysteine reacts with serine to form cystathionine by cystathionine β-synthase (Zhu et al., 2024; Dwight, 2020). Due to the inability to quantify homocysteine and methionine, their contribution to the increased levels of these other amino acids cannot be directly determined. Nonetheless, elevated levels of CTH and SER may indicate enhanced methionine metabolism and, consequently, increased homocysteine production through cystathionine β-synthase action (Dwight, 2020; Zhang et al., 2020). On the other hand, Butte et al. (Wu et al., 2024) found reduced serum levels of SER, which were associated with risk factors for insulin resistance, hypertriglyceridemia, and hyperuricemia.
Increased levels of AAA were observed in the OB group, as reported by Libert, Nowacki, and Natowicz (Libert et al., 2018), who also found elevated levels in adult subjects with obesity and diabetes. AAA can be generated by LYS metabolism. L-lysine is first converted to saccharopine by condensation with α-ketoglutarate, which is then reduced to 2-aminoadipic semialdehyde, releasing GLU. Subsequently, 2-aminoadipic semialdehyde is interconverted to AAA (Dwight, 2020). Catabolism of AAA forms 2-ketoadipic acid, and in TRP catabolism, it also occurs through L-kynurenine degradation (Dwight, 2020). High values of AAA and TRP, but not LYS, may suggest changes in these metabolic pathways in the OB group. Wang et al. (Wang et al., 2013) observed that AAA predicted the development of diabetes in normoglycemic individuals and hypothesized that AAA levels were increased in response to hyperglycemia, increasing insulin secretion and contributing to a compensatory mechanism to maintain glucose homeostasis in early insulin resistance. It also suggests that AAA is a marker of diabetes risk and a potential modulator of glucose homeostasis in humans. Additional investigations should be done to link AAA to weight gain, insulin resistance, and T2DM.
In addition to AAA, GLN has also been linked to insulin resistance (Cheng et al., 2012; Newgard et al., 2009). Hanzu et al. (Hanzu et al., 2014) found high levels of GLN and ALA in the visceral adipose tissue of individuals with obesity. Due to the high gluconeogenic effect, the increased amount of these amino acids released by visceral adipose tissue contributes to hyperinsulinemia and the development of insulin resistance (Hanzu et al., 2014). Higher GLN levels in the OW group suggest that being overweight may modify glucose metabolism due to increased body fat, especially visceral fat (Payab et al., 2021; Yan et al., 2023), evidenced by the augmented waist and hips circumferences and relative ratios. Despite the evidence, other groups have observed reduced levels of GLN in serum samples from children and urine samples from obese adolescents (Wahl et al., 2012; Cho et al., 2017).
Furthermore, it is well established that branched-chain and aromatic amino acids are indicators of the development of insulin resistance in normoglycemic young adults. This fact indicates a strong association between amino acids (Singh et al., 2023), particularly aromatics (TRP, PHE, and TYR), and the risk of future development of diabetes mediated partially by insulin resistance (Rangel-Huerta et al., 2019; Newgard et al., 2009; Martos-Moreno et al., 2005).
Under conditions of obesity, most studies demonstrate significant changes in blood and urinary values of aromatic (Butte et al., 2015; Cho et al., 2017; Haufe et al., 2016) and branched-chain amino acids (Szczerbinski et al., 2022; Butte et al., 2015; Bagheri et al., 2018). TRP, PHE, and TYR levels were elevated in the OB group, as other studies have also shown an association between increased concentrations of aromatic amino acids and obesity in both younger and older children, with different metabolic disturbances involved in the progression from overweight to obesity between the two age groups (Butte et al., 2015; Wu et al., 2024; Newgard et al., 2009; Payab et al., 2021; Bagheri et al., 2018; Kim et al., 2010; Yu et al., 2018).
Exploring amino acids through Volcano Plot analysis comparing children with normal weight (NW) and those with excess weight (OW + OB), using the NW group as a reference, the amino acids that exhibited a log2 fold change of less than −2 and -log10(p) of less than 1.3 — phenylalanine, tyrosine, α-aminoadipic acid, cystathionine, and lysine — showed greater significance and were negatively regulated in the NW group. Notably, tyrosine emerged as the amino acid with the most significant changes and the most considerable magnitude of change.
Tyrosine levels were related to increased hepatic fat content, suggesting hepatic dysfunction associated with a metabolic disorder (Libert et al., 2018; Haufe et al., 2016). In addition, tyrosine contributes to the profile described in obese children (Payab et al., 2021). It may be a possible predictor of insulin resistance in these children and the most sensitive metabolite for the classification of obesity (Handakas et al., 2022; Butte et al., 2015; Martos-Moreno et al., 2005). Increased plasma levels of phenylalanine and tyrosine have been observed in most analyses of amino acid biomarkers in obesity and T2DM (Payab et al., 2021; Haufe et al., 2016; Zhao et al., 2016; Park et al., 2015). In this regard, elevated levels of PHE and its hydroxylation product, TYR, provided strong relevance as biomarker metabolites predictive of the development of cardiovascular disease and diabetes type 2. Suzuki et al. (2019) described that a state of unbalanced or increased amino acids associated with obesity may exacerbate obesity and insulin sensitivity (Suzuki et al., 2019).
Several limitations of the present study need consideration. The relatively small size of groups may have interfered with statistical significance. Moreover, our study used BMI-for-age to classify obesity, rather than body fat content and distribution, which would be a more accurate parameter of adiposity and metabolic changes. However, determining criteria for cutting off body fat in children remains under discussion in the pediatric area. Additionally, other children’s data should be compared with these results to trace the relationship between nutritional status and the metabolic profile of amino acids. Our working group is already outlining new data crossings.