PPARα and PPARβ/δ are negatively correlated with proinflammatory markers in leukocytes of an obese pediatric population

Background Obesity configures a pathophysiological profile that predisposes the development of metabolic and cardiovascular diseases, critically impacting public health. The chronic dysregulation of immuno-metabolic components triggered by pediatric obesity is a common but scarcely understood aspect of the disease. Peroxisome proliferator-activated receptors (PPARs) are a group of transcription factors essential for energy and immune homeostasis of different tissues. Besides, the glucagon-like peptide-1 receptor (GLP-1R) activation influences insulin secretion, but also regulates the cytokine profile possibly mediated through a PPAR isotype. However, the role of PPARs and GLP-1R in leukocytes from obese pediatric patients remains unclear. Therefore, we examined the expression of PPARs isotypes and GLP-1R in leukocytes, and its correlation with metabolic, hormonal, inflammatory, and anthropometric markers in an obese pediatric population. Results Obese children and adolescents presented a significant increase in anthropometric and body composition parameters, TG, VLDL, TG/HDL, android fat (%)/gynoid fat (%) (A/G%) index, and HOMA score when compared with the control group. Obese participants exhibited a pro-inflammatory profile with an augment of IL-8 (p = 0,0081), IL-6 (p = 0,0005), TNF-α (p = 0,0004), IFN-γ (p = 0,0110), MCP-1 (p = 0,0452), and adipsin (p = 0,0397), whereas displayed a reduction of adiponectin (p = 0,0452). The expression of PPARα and GLP-1R was lower in the leukocytes from obese participants than in lean subjects. Furthermore, PPARα correlates negatively with TNF-α (p = 0,0383), while GLP-1R did not show correlation with any inflammatory variable. However, both receptors correlate negatively with the abdominal skinfold. Although PPARβ/δ expression was similar between groups, it was negatively associated with IL-8 levels (p = 0,0085). Conclusions PPARα and PPARβ/δ expression are negatively correlated with the proinflammatory markers TNF-α and IL-8, respectively, suggesting participation in the regulation of inflammation which was observed to be altered in pediatric obesity. Furthermore, PPARα and GLP-1R are downregulated in leukocytes from obese participants. The low expression of both receptors is correlated with an increase in abdominal skinfold, suggesting a role in fat distribution that could indirectly affect cytokine secretion from different immune and adipose cells, likely triggering an inflammatory profile as a consequence of obesity. Altogether, these findings may impact the understanding and implementation of PPARα or GLP-1R agonists in the clinic. Supplementary Information The online version contains supplementary material available at 10.1186/s12950-020-00264-2.

observed to be altered in pediatric obesity. Furthermore, PPARα and GLP-1R are downregulated in leukocytes from obese participants. The low expression of both receptors is correlated with an increase in abdominal skinfold, suggesting a role in fat distribution that could indirectly affect cytokine secretion from different immune and adipose cells, likely triggering an in ammatory pro le as a consequence of obesity. Altogether, these ndings may impact the understanding and implementation of PPARα or GLP-1R agonists in the clinic.

Background
Obesity is a chronic disease with a high impact on public health due to its increased prevalence. As a result of obesity, adult subjects can develop type 2 diabetes (T2D) and cardiometabolic diseases. Nowadays those events are not limited to adults, but rather they are remarkably increased in children and adolescents with obesity [1]. T2D at early life stages carries on into adulthood, which is related to a more severe damage of β-cells [2], and to increased appearance of additional comorbidities such as cardiovascular disease, kidney and visual failure [3,4].
The metabolic phenotype in obesity is determined by alterations in metabolic regulation of triglycerides and glucose, which can lead to a chronic in ammatory status characterized by the enriched circulation of proin ammatory cytokines [5][6][7][8]. The in ammatory pro le displayed in obese children and adolescents is frequently associated with insulin resistance (IR), hyperinsulinemia, hyperglycemia, and metabolic syndrome, gradually leading to the development of T2D and cardiometabolic diseases [9]. Noticeably, the study of metabolic sensors that regulates the immune proin ammatory response in obesity remains poorly assessed.
Peroxisome proliferator-activated receptors (PPARs) are recognized transcription factors, comprising a group of three isotypes: PPARα, PPARβ/δ, and PPARγ. They play an essential role in regulating the expression of genes responsible for lipid and carbohydrate metabolism, as well as maintenance mitochondrial metabolic rate [10]. PPARs participate in the regulation of the expression of the cytokine pro le from adipose tissue and immune cells with a proven anti-in ammatory role [8,11].
An additional factor that can display a dual role in metabolism and immunity is the glucagon-like peptide-1 (GLP-1)/GLP-1R axis. GLP-1 is a hormonal factor that canonically induced an increase in the pancreatic release of insulin [12,13]. This axis showed a cytoprotective role in pancreatic β-cells, delaying the establishment of obese T2D [12,14]. A non-canonical effect of GLP1 is associated with the activation of GLP-1R to modulate cytokine production. GLP-1R signaling on invariant natural killer T cells (iNKT) can enhance the expression of anti-in ammatory cytokines such as IL-10 [12]. Besides, GLP-1R also downregulates pro-in ammatory cytokines expression, mediated through PPARγ activation and followed by NF-kB inhibition [15].
The regulatory genes characterization of metabolic and in ammatory processes can be accomplished in leukocytes. Leukocytes continuous metabolic sensing signi cantly contributes to the coordination of organism-wide metabolic and immunological responses [16]. Venous blood obtention from humans is a minimally invasive procedure and represents a widely accessible source which can be easily assessed to identify gene expression in leukocytes. Thus, blood analyses are valuable for early diagnosis and opportune intervention for both immune and metabolic disorders. Remarkably, few studies in obese population focus on PPARs isotypes and GLP-1R in leukocytes. Therefore, the aim of this study is to evaluate the expression of PPARs and GLP-1R in leukocytes, and its association with in ammatory, metabolic, body composition, and anthropometric factors in an obese pediatric population.

Results
Obese children and adolescents have an impaired metabolic and anthropometric pro le The anthropometric, body composition and biochemical variables of the studied individuals strati ed by body mass index (BMI) percentile status are shown in Table 1. Age and gender of the participants were not signi cantly different among the control and obese group (Additional le 3 and 4). The median age was 11 years old, and 48% were girls. All mean values of skinfold thickness, waist circumference (WC), arm circumference (AC), waist-to-height ratio (WHR), arm muscle circumference, and neck circumference were signi cantly increased in participants with obesity in comparison with healthy weight children and adolescents (Table 1). Regarding body composition, the percentage of arm, leg, trunk, android, gynoid, and body fat were signi cantly elevated in obese patients compared with control subjects (Table 1). In obese children, fat distribution was similar between different body areas, and there were no signi cant differences in mean values of lean mass between groups.
The mean of fasting glucose and 2-hours glucose fell within the normal ranges, as recommended by the American Academy of Pediatrics (Table 2) [17]. However, insulin was signi cantly higher in children with obesity than in the standard weight group. Accordingly, the HOMA score was increased in obese children; this score overcomes the 3,4 cut-off indicating IR (Table 2) [18]. TG and VLDL levels were higher in obese children than those of lean subjects, whereas HDL-c levels were signi cantly reduced in obese subjects in contrast with the control group ( Table 2). The levels of total cholesterol were similar among groups ( Table   2). TG/HDL and A/G ratios are frequently used as indicators of cardiometabolic risk, and here, those ratios were signi cantly elevated in the obese group in comparison with the lean patients. Based on this data, obese children and adolescents have an increased risk of developing cardiovascular disease and T2D. Incretins gastric inhibitory polypeptide (GIP) and GLP-1 did not show signi cant differences between the study groups ( Table 2).
PPARα and GLP-1R gene expression is reduced in leukocytes from obese subjects To determine the expression of PPAR isotypes in the leukocytes, the expression of each isotype was analyzed by qPCR. PPAR-α expression showed a signi cant reduction (about 50%) in the leukocytes from obese patients in comparison with the control group (p=0,0484) ( Figure 1a). PPAR-β showed no differential expression between the study groups ( Figure 1b). In contrast, PPAR-γ expression was undetectable in leukocytes from all the samples. To corroborate this nding, we used adipose tissue cDNA as a positive control, since this isotype is mainly present in adipose cells. This sample presented a positive ampli cation and a single peak in the dissociation curve, indicating that the reaction was speci c (data not shown). Likely, PPAR-γ transcripts were not expressed on leukocytes or were expressed at low levels but just in some cell lineages of these samples.
Besides, analysis from mRNA expression levels of incretin receptors was also evaluated in the studied individuals. GLP-1R expression was signi cantly reduced in the obese group compared with the healthy weight participants (p=0,1358) (Figure 1c), whereas no difference was observed in GIPR expression between the two groups ( Figure 1d).

Obese children and adolescents showed a proin ammatory pro le of adipokines and cytokines
To identify the in ammatory serum pro le in obese children and adolescents, the levels of several adipokines and cytokines were measured by ow cytometry using a multiplex assay. Obese subjects showed a signi cant increase in the levels of IL-8 (p=0,0081), IL-6 (p=0,0005), TNF-α (p=0,0004), IFN-γ (p=0,0110), and MCP-1 levels (p=0,0452), compared with the normal weight group ( Figure 2). Serum levels of IL-10 and IP-10 were detected but did not differ signi cantly between the groups.
The serum adiponectin concentration presented a signi cant reduction in the obese group (p=0,0452) compared with the control group ( Figure 3). Conversely, adipsin displayed a signi cantly higher concentration (p=0,0397) in the obese group in comparison with the control group, which had a concentration of 7,688 ± 6,116 pg/mL ( Figure 3). Resistin showed similar levels between the study groups, with concentrations of 13,102 ± 8,342 pg/mL and 16,792 ± 8,302 pg/mL for the obese and control participants, respectively ( Figure 3). Finally, Retinol-binding protein 4 (RBP4) did not present a representative concentration in any of the study groups, indicating that it is not secreted in serum at detectable levels under the evaluated conditions. PPARα and PPARβ/δ correlate negatively with proin ammatory markers Analysis of the relationships between PPAR-α, PPAR-β, and GLP-1R expression, and the levels of cytokines, chemokines, and anthropometric parameters of the obese subjects, showed that PPAR-α transcript levels had a signi cant negative correlation with TNF-α levels (r = -0,583, p=0,03883) (Table 4), as well as with abdominal skinfold (r = -0,712, p=0,0016). Besides, PPAR-β showed a signi cant negative correlation with IL-8 levels (r = -0,667, p=0,0085) and with arm fat percentage (r = -0,651, p= 0,0132) ( Table 4

Discussion
Childhood and adolescence are critical periods where fundamental metabolic and hormonal changes start and have an impact on adulthood health. Excess weight at these life stages is a critical factor for the onset of cardiometabolic disease and T2D. Thus, those early stages are vital periods for timely intervention. We evaluated the expression of essential metabolic and immune sensor genes, such as isotypes of PPAR and GLP-1R in leukocytes, and their correlation with metabolic, immune, and anthropometric factors in a Colombian's pediatric population with obesity.
PPARs participate in the control of the in ammatory process produced by obesity, modulating the expression of proin ammatory cytokines in adipose cells [10]. However, other tissues can contribute to the in ammatory process. Studies of PPARs, particularly of PPARβ/δ and PPARα, on leukocytes from obese children and adolescents, are scarce. However, the assessment of PPARs expression and relation with variables associated with obesity represents an accessible tool for a better comprehension of obesity pathophysiology, and for the development of improved therapeutic strategies [19].
In this work, we identi ed the expression of PPARα and PPARβ/δ, but not PPARγ in leukocytes from obese children and adolescents. The activation of PPARβ/δ has been mainly studied in "endurance-exercise mimetics" to reduce in ammation as an obesity treatment, particularly focused on the liver, skeletal muscle, and adipose tissue [20,21]. Moreover, PPARβ/δ is involved in lipid metabolism, regulation of fatty acid oxidation (FAO), in ammation, and modulation of glucose and cholesterol levels [22]. Direct activation of this isotype improves insulin sensitivity and disorders associated with metabolic syndrome in humans [23]. In this study, PPARβ/δ did not show signi cant differences between the standard weight and obese groups. However, this gene displayed a signi cant negative correlation with IL-8 levels and arm fat percentage. IL-8 (CXCL8) is a proin ammatory chemokine which mediates the crosstalk between obesity and cardiovascular disease [24,25]. IL-8 is mainly secreted by adipose tissue, broblasts, endothelial cells, monocytes, and macrophages when exposed to IL-1β, TNF-α, or LPS [26]. Indeed, we detected a signi cant increase in TNF-α levels from the serum of obese subjects, which could contribute to the IL-8 secretion in those individuals.
Although the direct link of PPARβ/δ to IL-8 is suggestive, some studies have already related the activation of PPARβ/δ with the modulation of expression of IL-8 in endothelial cells and in endometrium participating in in ammation [27][28][29]. In line with this evidence, our results suggest that a decrease in the PPARβ/δ expression in leukocytes might consequently augment IL-8 levels and, thereby, the in ammatory response during obesity. Indeed, experiments in mice have associated the deletion in the expression of PPARβ/δ, with a reduction in the number of hepatic M2 macrophages, which results in alterations of lipid metabolism, IR, and decreased expression of anti-in ammatory cytokines [30]. Those ndings indicate a protective role of PPARβ/δ in the control of lipid metabolism and the development of in ammatory processes.
On the other hand, PPARα acts as a lipid sensor actively expressed in tissues with high metabolic rates, such as the liver, heart, muscle, kidney, intestinal mucosa, and brown adipose tissues [31,32]. The use of natural ligands or agonists for this isotype has revealed its participation in glucose homeostasis, IR development, in ammation, atherogenesis, and in the expression of genes related to lipid metabolism [33,34]. In leukocytes, PPARα expression has barely been studied. However, the expression of PPARα has been reported in immune cells such as monocytes/macrophages, peripheral blood mononuclear cells (PBMCs), and CD4 + T cells [10,35,36]. Also, subjects with metabolic syndrome signi cantly decreased PPARα expression in eosinophils by approximately 21% [37]. Our pediatric obese population showed reduced expression of this isotype in leukocytes when compared to the lean group. Likewise, studies on mice with induced obesity, exhibited reduced expression of PPARα and carnitine palmitoyltransferase I (CPT-1) in the liver [38]. PPARγ expression is also reduced in PBMC from obese children and adolescents [39]. Indeed, obese mice treated with beza brate (BZ), an agonist that preferentially activates PPARα and PPARγ, improves biochemical parameters and reduce the WAT proin ammatory state in obese mice [38]. These ndings suggest that obesity in the early stages of life in uences the expression of PPARα in leukocytes, and consequently, its reduction can be related to the observed altered lipid pro le and reduced insulin sensitivity.
We observed a signi cant increase in TG and a reduction in HDL levels in obese children and adolescents in comparison with the control group. Both parameters overcome the reference values for pediatric individuals. Noticeably, insulin sensitivity was also affected in the obese group. Thus, we used the HOMA index to evaluate the IR status, which is the most used method for its diagnosis in the pediatric population. The obese subjects displayed a mean value of 4,28, signi cantly higher than lean subjects, which showed a 1,50 score, predicting IR for the obese group.
As a complementary measure to con rm the prevalence of IR, we calculated the TG/HDL ratio, which closely correlated with this parameter in adults at a value ≥ 3 [40]. This ratio is also considered as a predictor of cardiometabolic risk and re ects an atherogenic lipid pro le [41]. In the pediatric population, a high TG/HDL ratio is associated with lower insulin sensitivity and cardiovascular risk [42]. We found a signi cantly increased TG/HDL ratio for the obese group (3,02 p <0,0001) in comparison with the lean group (1,60). These results con rm a prevalent condition of IR and cardiovascular risk in obese participants that could be associated with the downregulation of PPARα. Regarding this, PPARα has been reported to improve insulin sensitivity and provide a protective property against the development of IR [37,43].
Besides, PPARα is recognized by its anti-in ammatory role. In this study, we observed a proin ammatory pro le in obese pediatric subjects with an increase in the serum levels of adipsin, MCP-1, IL-8, IL-6, IFN-γ, and TNF-α, and decreased levels of the anti-in ammatory adipokine adiponectin. Hence, PPARα downregulation in leukocytes from obese children suggests an association with their pro-in ammatory cytokine pro le. Indeed, PPARα displayed a signi cant negative correlation with TNF-α levels and the abdominal skinfold. TNF-α promotes IR and decreases the uptake of fatty acids by adipocytes, increasing them in circulation while stimulating lipolysis [44,45].
Furthermore, TNF-α increases the production of in ammatory cytokines such as IL-6 and suppresses the production of adiponectin [45]. TNF-α altogether with IFN-γ and IL-6 leads to an in ltration of immune cells in adipose tissue and dysfunction of in ammatory immune cells [46]. In this work, we found that the concentration of TNF-α was signi cantly higher in obese patients than in the control group. These ndings are consistent with the high HOMA index, IL-8 and IL-6 levels, and the decreased adiponectin concentration in obese children and adolescents.
The consequences of obesity on the development of IR and activation of a chronic in ammatory response allowed to establish a relationship between GLP-1R and PPARs. GLP-1 is a hormonal factor that canonically improves insulin secretion on pancreatic β-cells [12,14]. Otherwise, non-canonical functions of GLP-1R are reported in some lineages of immune cells, such as an immunomodulatory role in the migration and differentiation of T and B lymphocytes [12,47,48], and a reduction in monocyte/macrophage migration and in ammatory cytokine production [49,50]. In this study, we observed a reduced expression of GLP-1R in obese children and adolescents. Although its expression showed no signi cant association with in ammatory parameters, GLP-1R and PPARα expression exhibited a negative correlation with abdominal skinfold (Figure 4). These ndings suggest a role of both receptors in the adipose tissue distribution that could indirectly affect the role of different immune cells and the secreted cytokines. Besides, GLP-1R and PPARα effects have been associated in pancreatic islets favoring GLP-1R expression and sensitivity to GLP-1 [51], and, in endothelium, leading to downregulation of proin ammatory cytokines [15]. Therefore, both receptors may present a crosstalk in leukocytes from an obese population. This association is highly needed to be explored in the future.
In summary, adipose tissue accumulation in obesity is associated with disturbances in the homeostasis of several tissues, disruption of the mechanisms controlling lipid and glucose metabolism, and dysregulation in serum proin ammatory cytokines and adipokines [46]. It was possible to establish a relationship between anthropometric alterations and body fat distribution observed in obesity with associated comorbidities, such as IR, chronic in ammatory pro le, and risk for the development of cardiovascular diseases. In addition, PPARα and GLP-1R represent potential gene targets to be explored in pediatric obesity, based on their roles in abdominal fat distribution and in ammatory pro le (Figure 4).

Conclusion
We have shown that PPARα and PPARβ/δ expression are negatively associated with the proin ammatory markers TNF-α and IL-8, respectively, suggesting participation in the regulation of in ammation observed in pediatric obesity. Furthermore, PPARα and GLP-1R are downregulated in leukocytes from obese children and adolescents, and both receptors are associated with the abdominal skinfold, suggesting a role in fat distribution that could indirectly affect the function of different immune cells and the secreted cytokines. Thereby, these ndings may impact the understanding and implementation of PPARα or GLP-1R agonists in the clinic.

Materials And Methods
Participants This is a cross-sectional study with primary data collection performed in children and adolescents from February 2017 to June 201,8 in Bogotá-Colombia. Children and adolescents aged 7 to 17 years old from seven public schools were invited to participate voluntarily but with parental authorization. The study population included participants from low-to middle-income, which represents most of the city population. Finally, 37 participants (female, n=18, and male, n=19), which ful lled the inclusion criteria, agreed to participate in this study. The inclusion criteria included female and male participants aged 7 -17, with a healthy weight (lean control group) and obesity (de ned as BMI, ≥ 85th, and ≥95th percentile) [52]. Exclusion criteria included: children and adolescents with diabetes mellitus (both type I and II), secondary causes of obesity, psychiatric illness, mental disorders, with weight equal to or greater than 100 kilograms, as well as those who were receiving medications which promoted weight loss or gain, or which altered in any way the metabolic pro le, including glycemia.
The Institutional Review Board of the Antonio Nariño University and the Medical Ethics Committee at DEXA DIAB (CE-CC-00721) approved the study protocol. Written informed parental consent and child assent from participants were obtained before any research procedures. This study was carried out following the ethics committee code of the American Medical Association endorsed by the National Institutes of Health (NIH).

Anthropometric assessment
Anthropometric measures were obtained using standard methods [53]. Weight was measured to the nearest 0.1kg on a calibrated digital scale without shoes and wearing light clothing. Height was measured to the nearest 0.1cm on a stadiometer. To evaluate the nutritional status of children and adolescents, the BMI Z-score was used, according to the criteria proposed by the World Health Organization [52]. Waist circumference (WC) was measured via exible measuring tape with an accuracy of 1 mm. The measurement was made with the subject standing, taken at the midpoint between the right lower costal ridge and the iliac crest, without skin compression from the measuring tape, and at the end of a normal expiration. BMI and waist-height were calculated as ratios. Skinfold thicknesses were assessed using a Harpenden Skinfold Calliper, according to a previous protocol [53]. The measurements were performed in triplicate at the non-dominant side on biceps, triceps, subscapular, and suprailiac to the nearest 0.1 mm, and the mean values were calculated [53].

Measurement of body composition by Dual X-ray Absorptiometry
Measurements of total and regional body composition were acquired using a Dual-Energy X-ray Absorptiometry (DEXA) scan (GE Lunar Prodigy advance, GE Healthcare) by a trained technician. Before each acquisition, the scanner was calibrated according to the manufacturer's instructions. The total effective radiation dose during each examination was <0.05 micro Sievert for a three-minute scan. For analysis, reference data for standard deviation scores were provided from GE Lunar Body composition software (enCORE 2010; version 11.3; GE Healthcare, Madison, WI, USA)

Biochemical analyses
Blood samples for laboratory analysis were collected by venipuncture from all participants after an overnight fast (10-12 hours). Participants underwent a 2-hour oral glucose tolerance test (OGTT) (1.75 g/kg, maximum 75 g) according to the American Diabetes Association criteria [54].

Leukocyte isolation
An additional tube with an anticoagulant was collected for the isolation of leukocytes by the whole blood lysis method as described previously [55]. The sample was centrifuged at 800 g for 5 minutes and the leukocyte layer was carefully transferred to a conical tube with Red Blood Cell Lysis Buffer (RBC lysis). The tube was mixed by inversion and incubated for 5 minutes at room temperature. The samples were centrifuged, and the supernatant was discarded. The procedure was repeated twice. The cell pellet was washed with PBS and preserved in trizol at -80 °C for RNA extraction.

Gene expression analysis
Leukocyte total RNA was extracted using Trizol reagent (Life Technologies, Inc) according to the manufacturer's instructions. RNA was quanti ed with a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies), and cDNA was synthesized from 1 μg ARN with the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scienti c, Cat. # 4368814). ). Real-time PCR was performed in duplicate using CFX96 equipment (BioRad) and SYBR® Green PCR Master Mix, according to the manufacturer's protocol. PPAR-α, PPAR-β, PPAR-γ, incretins receptor genes GLP-1R and GIPR, and reference genes PGK1 and YWHAZ were quantitated using the gene-speci c primers (Additional le 1). cDNAs were ampli ed for 40 cycles consisting of 10 s of denaturation at 95 °C, 15 s of annealing temperature for each primer (Additional le 1), and 10 s of extension at 72 °C. Standard curves for all genes were generated using serial dilutions of pooled cDNAs from all samples. Relative mRNA expression was calculated with the method ΔΔCt. Data are shown as normalized ratios between target gene expression and geometric media of the two reference genes [56]. All expression assays were performed following the MIQE guidelines [57] [56].
Hormone and cytokine levels Preprandial GIP and GLP1 levels were measured in serum using a speci c enzyme-linked immunosorbent assay (Elabscience, Cat. #E-EL-H2061 and E-EL-H6025) according to the procedures provided by the manufacturer. Levels were expressed in ng/mL and pg/mL. The presence of adiponectin, adipsin, RBP4, MCP-1, IL-1β, IP-10, IL-10, IL-8, IL-6, IFN-γ, resistin, and TNF-α in serum samples was analyzed using a multiplex immunoassay based on uorescence-encoded beads according to the manufacturer's instructions (No. 40196, BioLegend). The acquisition was performed in an Accuri C6 ow cytometer. Offline analysis was performed with LEGENDplex™ data analysis software (version 8.0), and the data were expressed as the mean reporter uorescence intensity PE (MFI) as a function of concentration (pg/mL). Each assay was performed with two technical replicates.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 6. Descriptive statistics were expressed as mean ± standard deviation (SD) whenever applicable, and variables were tested for normality by Kolmogorov-Smirnov test. Non-parametric tests were used whenever data were not normally distributed or Student t-test when data were normally distributed. We used Wilcoxon signed-rank tests to assess the signi cance of inter-group differences. The correlation of target genes and evaluated parameters were determined using the Spearman rank test. For all analyses, statistical signi cance was considered as p<0.05.

Declarations
Ethics approval and consent to participate The study protocol was approved by the Institutional Review Board of the Antonio Nariño University and the Medical Ethics Committee at DEXA DIAB (CE-CC-00721). Written informed parental consent, and child assent from participants were obtained before any research procedures were conducted. This study was carried out in accordance with the ethics committee code of the American Medical Association endorsed by the National Institutes of Health (NIH).

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests  Differences between groups were evaluated using Wilcoxon signed-rank tests. *p < 0.005, ** p < 0.0001.