Salivary Neurotrophins Brain-Derived Neurotrophic Factor and Nerve Growth Factor Associated with Childhood Obesity: A Multiplex Magnetic Luminescence Analysis

Obesity is linked with higher inflammatory markers and is characterized by chronic low-grade inflammation. Neurotrophins brain-derived neurotrophic factor (BDNF) and β-nerve growth factor (β-NGF), in addition to their neuronal functions, act on several immune cells and have been recently designated as metabokines due to their regulatory role in energy homeostasis and food intake. The current study evaluates the salivary BDNF and β-NGF and their association with anthropometric measurement, blood pressure, and salivary insulin in children. Anthropometric measurements and saliva samples were obtained from 76 children, aged 6–10 years. Multiplex analysis was carried out for the salivary analysis of BDNF, NGF, and insulin by human magnetic Luminex performance assay. Statistical analysis was performed to analyze the best fit diagnostic value for biomarkers and the relationship of the neurotrophic levels of BDNF and NGF with obesity measures and blood pressure. Salivary BDNF and β-NGF showed a significantly higher concentration in obese children than normal-weight children. Both neurotrophins are positively associated with obesity anthropometric measures, blood pressure, and salivary insulin. Multinominal regression analysis reported a significant association between salivary BDNF, β-NGF, insulin, and systolic pressure adjusted for age, gender, income, and maternal education. The salivary concentration of BDNF and NGF was higher in obese children, and it is positively associated with anthropometric measures, suggesting that neurotrophins can be used as a non-invasive predictor of obesity-related complications in children.


Introduction
The global epidemiologic of metabolic disorders is associated with obesity, which is a primary threat to human health in the current century, reducing human lifespan. If the current secular trend continues with the worldwide adult population, an expected overweight of 38% and an additional 20% of the obese population might be reached by 2030 [1,2]. The United States of America's prevalence of obesity has risen by 18.5% since the early 1990s. Over the past few decades, obesity has been a major problem among children and has affected approximately 13.7 million children and teenagers during 2015 and 2016 [3]. Overweight and obesity prevalence in early life elevates the risk of developing obesityassociated health issues in adulthood. Early life detection of the metabolic risk factor could help protect children from the widespread consequences of obesity-related diseases.
Biomarkers for diagnosing disease conditions are analyzed in many body fluids, such as serum, plasma, urine, and cerebrospinal fluid [2,4]. Recent investigations have revealed that disease biomarkers for diagnosing inflammation, obesity, insulin resistance, and other metabolic disorder can be identified in saliva [5][6][7][8]. The significant benefit of saliva is z-score calculation using the growth reference 2017 SPSS macro [30]. Waist circumference (WC) z-score, and waist-to-height ratio (WHtR) z-score were calculated using the R macro package [31].

Measurement of Salivary BDNF, NGF, and Insulin by Multiplex Analysis
The participants were requested to avoid any food or drink 30 min before obtaining the sample. The samples were collected using a saliva collection kit and stored at −80 • C until further analysis. The samples for multiplex analysis were centrifuged at 2800 rpm, 4 • C for 20 min to remove unwanted particulates [28]. Human BDNF and NGF were measured with a human premixed magnetic Luminex assay kit (catalog # LXSAHM, R&D systems, Minneapolis, MN, USA). The assay was carried out as per the manufacturer's protocol. Standards and saliva samples (1:2 dilution) were briefly prepared using calibrator diluent RD6-52. The assay was performed with duplicate standards and samples (50 µL), along with premixed Luminex beads added to the plate and incubated for 2 h at room temperature (RT) with shaking (800 ± 50 rpm). The plate was washed three times with a wash buffer using a Bio-Plex handheld magnetic plate holder and then incubated with 50 µL diluted biotin-antibody cocktail for 1 h at RT on a shaker. Plates were washed with wash buffer and incubated with streptavidin-PE for 30 min at RT with shaking. The plate was washed three times and read using a Bioplex-3D analyzer. The details of the Multiplex measurement of insulin have been described in our previous study [7].

Statistical Analysis
The results are expressed as the mean and standard error of the mean (SEM) in the bar graph. A one-way ANOVA test was performed by GraphPad prism (5.0, GraphPad Software, San Diego, CA, USA) for a three-group comparison. Tukey's post-hoc test was carried out to show the difference between individual groups. Statistical significance between groups was provided on the graph, and the bar graphs without p-value are not significant at p < 0.05 levels ( Figure 1). Data transformation was carried out before processing for further analysis for skewed data with a natural logarithm by SPSS (version 24, IBM, Armonk, NY, USA). The Receiver Operating Characteristic (ROC) curve analysis was performed to determine the diagnostic accuracy of BDNF and NGF. The area under the curve (AUC) was calculated to distinguish the high and low levels of diagnostic biomarkers. Multinomial regression analysis was carried out between parameters to determine the significantly affected biomarkers. The power analyses for the sample size have been described in Selvaraju et al. [7].

Results
Anthropometrics are quantitative body measures applied to evaluate growth, development, and health parameters. In the current study, participants' anthropometric measurements were recorded. The obesity measurements showed a significant increase in the OW and OB groups compared to NW. The present study continues our previous study, and these results are shown in Table S1 of Supplementary Material [7].
The expression of salivary BDNF, the second member of the neurotrophic factors' family, showed a significant increase in the OB (46.55 ± 3.63 pg/mL; p < 0.02) group compared to the NW group. The OW (37.32 ± 4.95 pg/mL) group showed an increased expression of BDNF, but there was no significant difference compared to the NW group (28.33 ± 3.94 pg/mL). Similarly, the best-known neurotrophin NGF, also recently identified as adipokines associated with obesity, exhibited a significant elevation in the OB group (11.08 ± 0.94 pg/mL; p < 0.005) in comparison with NW participants. NGF was higher in the OW group (7.53 ± 1.37 pg/mL) compared to the NW group (5.41 ± 1.03 pg/mL) but did not show significance. The results are expressed in the bar graph in Figure 1. In addition, we assessed the participant's blood pressure and heart rate. OB group participants expressed significantly higher systolic (108.81 ± 2.85; p < 0.008) and diastolic (70.50 ± 2.40; p < 0.04) blood pressure compared to NW participants. The systolic (105.20 ± 3.00) as well as diastolic (64.25 ± 1.53) blood pressure levels of the OW group did not show a significant difference with respective NW participants (systolic: 98.22 ± 1.72, diastolic: 63.90 ± 1.52, respectively). The OW and OB group participants did not show any significant difference in heart rate compared to the NW group ( Figure 1).

Results
Anthropometrics are quantitative body measures applied to evaluate growth, development, and health parameters. In the current study, participants' anthropometric measurements were recorded. The obesity measurements showed a significant increase in the OW and OB groups compared to NW. The present study continues our previous study, and these results are shown in Table S1 of Supplementary Material [7].
The expression of salivary BDNF, the second member of the neurotrophic factors' family, showed a significant increase in the OB (46.55 ± 3.63 pg/mL; p < 0.02) group compared to the NW group. The OW (37.32 ± 4.95 pg/mL) group showed an increased expression of BDNF, but there was no significant difference compared to the NW group (28.33 ± 3.94 pg/mL). Similarly, the best-known neurotrophin NGF, also recently identified as adipokines associated with obesity, exhibited a significant elevation in the OB group (11.08 ± 0.94 pg/mL; p < 0.005) in comparison with NW participants. NGF was higher in the OW group (7.53 ± 1.37 pg/mL) compared to the NW group (5.41 ± 1.03 pg/mL) but did not show significance. The results are expressed in the bar graph in Figure 1. In addition, we assessed the participant's blood pressure and heart rate. OB group participants expressed significantly higher systolic (108.81 ± 2.85; p < 0.008) and diastolic (70.50 ± 2.40; p < 0.04) blood pressure compared to NW participants. The systolic (105.20 ± 3.00) as well as diastolic (64.25 ± 1.53) blood pressure levels of the OW group did not show a significant difference with respective NW participants (systolic: 98.22 ± 1.72, diastolic: 63.90 ± 1.52, respectively). The OW and OB group participants did not show any significant difference in heart rate compared to the NW group ( Figure 1).

Figure 2.
Correlation between anthropometric measurements BMI z-score, WC z-score, and WHtR z-score with neurotrophins BDNF and NGF.

Figure 2.
Correlation between anthropometric measurements BMI z-score, WC z-score, and WHtR z-score with neurotrophins BDNF and NGF.
ROC curves were used to analyze the diagnostic efficiency of BDNF and NGF among the study participants. The area under the curve (AUC) results showed β-NGF had a higher AUC (0.757; 95% CI: 0.650-0.864; p < 0.005) and cut-off values of 2.08 from the transformed data with a sensitivity of 0.938 and specificity of 0.317. Evaluating the diagnostic performance for early detection of biomarker showed AUC for BDNF (0.727; 95% CI: 0.615-0.839; p < 0.002) with sensitivity (0.875), specificity (0.367), and the cut off values (3.551). The diagnostic power of β-NGF and BDNF showed significant AUC values, and this helps for the early detection of obesity and associated risk factors in children ( Figure 4A). ROC curves were used to analyze the diagnostic efficiency of BDNF and NGF among the study participants. The area under the curve (AUC) results showed β-NGF had a higher AUC (0.757; 95% CI: 0.650-0.864; p < 0.005) and cut-off values of 2.08 from the transformed data with a sensitivity of 0.938 and specificity of 0.317. Evaluating the diagnostic performance for early detection of biomarker showed AUC for BDNF (0.727; 95% CI: 0.615-0.839; p < 0.002) with sensitivity (0.875), specificity (0.367), and the cut off values (3.551). The diagnostic power of β-NGF and BDNF showed significant AUC values, and this helps for the early detection of obesity and associated risk factors in children ( Figure  4A). To determine the relationship between BDNF and β-NGF in obese children, linear regression analysis was performed. The results of the BDNF and β-NGF association demonstrate a significant positive correlation, with an r-value of 0.970 (95% CI: 0.607-0.681; p < 0.001) ( Figure 4B). Our previously published data showed a higher expression of insulin in OB group participants compared to the NW group [7]. In this study, we analyzed insulin association with neurotrophic parameters BDNF and NGF to measure insulin dependency in NW and OB participants. The results showed a positive correlation of insulin with BDNF and β-NGF, with a respective r-value of 0.725 (95% CI: 0.671-1.051; p < 0.001) and 0.749 (95% CI: 0.470-0.712; p < 0.001) ( Figure 5).  To determine the relationship between BDNF and β-NGF in obese children, linear regression analysis was performed. The results of the BDNF and β-NGF association demonstrate a significant positive correlation, with an r-value of 0.970 (95% CI: 0.607-0.681; p < 0.001) ( Figure 4B). Our previously published data showed a higher expression of insulin in OB group participants compared to the NW group [7]. In this study, we analyzed insulin association with neurotrophic parameters BDNF and NGF to measure insulin dependency in NW and OB participants. The results showed a positive correlation of insulin with BDNF and β-NGF, with a respective r-value of 0.725 (95% CI: 0.671-1.051; p < 0.001) and 0.749 (95% CI: 0.470-0.712; p < 0.001) ( Figure 5). ROC curves were used to analyze the diagnostic efficiency of BDNF and NGF among the study participants. The area under the curve (AUC) results showed β-NGF had a higher AUC (0.757; 95% CI: 0.650-0.864; p < 0.005) and cut-off values of 2.08 from the transformed data with a sensitivity of 0.938 and specificity of 0.317. Evaluating the diagnostic performance for early detection of biomarker showed AUC for BDNF (0.727; 95% CI: 0.615-0.839; p < 0.002) with sensitivity (0.875), specificity (0.367), and the cut off values (3.551). The diagnostic power of β-NGF and BDNF showed significant AUC values, and this helps for the early detection of obesity and associated risk factors in children ( Figure  4A). To determine the relationship between BDNF and β-NGF in obese children, linear regression analysis was performed. The results of the BDNF and β-NGF association demonstrate a significant positive correlation, with an r-value of 0.970 (95% CI: 0.607-0.681; p < 0.001) ( Figure 4B). Our previously published data showed a higher expression of insulin in OB group participants compared to the NW group [7]. In this study, we analyzed insulin association with neurotrophic parameters BDNF and NGF to measure insulin dependency in NW and OB participants. The results showed a positive correlation of insulin with BDNF and β-NGF, with a respective r-value of 0.725 (95% CI: 0.671-1.051; p < 0.001) and 0.749 (95% CI: 0.470-0.712; p < 0.001) ( Figure 5).  To test the relationship between socioeconomic status (SES) and health outcomes, the odds ratio was calculated by multinomial regression analysis. The OB group showed significant unadjusted β-coefficient values with reference to the NW weight category for salivary BDNF (OR = 1.036; p < 0.011), β-NGF (OR = 1.163; p < 0.004) and insulin (OR = 1.006; p < 0.002), along with both systolic (OR = 1.085; p < 0.005) and diastolic (OR = 1.084; p < 0.021) blood pressure. The measures were adjusted for age, gender, income, and maternal education to verify the SES correlation. After adjustment with cofounders, the β-coefficient was shown to be significant in salivary markers and systolic blood pressure. Diastolic pressure did not show as significant after adjusting with cofounders (Table 1). BDNF-Brain derived neurotrophic factor; β-NGF-β-nerve growth factor; NW-Normal-weight; OW-Overweight; OB-Obese; OR-Odds ratio; SE-Standard error. The p-value in bold represents the significant odds ratio of the OW or OB participants compared to the NW group.

Discussion
The study analyzed the BDNF and NGF levels in the saliva of NW, OW, and OB children. The results show that BDNF and NGF concentrations were significantly increased in OB children compared with NW children, which correlates with obesity measures, indicating a correlation between neurotrophic proteins and fat mass. Similarly, systolic and diastolic blood pressure was found to be high in OB participants compared to NW. We could not detect any difference in the heart rate between NW and OW or OB group participants. The current study was designed to evaluate the neurotrophic proteins BDNF and NGF as the salivary marker of obesity in children. Childhood obesity is considered one process that generates low-grade inflammation [32]. Neurons are protected by neurotrophic factors from inflammatory outcomes and play an important role in regulating dietary intake and changes in body weight in children [33]. The results of BDNF and NGF concentrations correlate with the previously published studies. The cross-sectional and longitudinal data analysis performed on normal-weight and obese children showed significantly increased serum BDNF in obese children than in normal-weight children [34]. Plasma BDNF levels in the pediatric population age range of 5-13 years were shown to be significantly elevated in overweight and obese children compared to normal-weight children [35]. Neurotrophin NGF concentration in adult female participants with different degrees of obesity showed a 1.4-fold increase in overweight and obese participants in comparison to normal-weight participants. Plasma NGF was decreased in morbidly obese participants compared to overweight and obese participants, but its level was more elevated than in normal-weight participants [13]. Ha et al. reported that higher circulating CRP and neurotrophin (BDNF and NGF) were observed in obese adolescents. Our result converged with previous find-ings [36]. The higher amount of BDNF is in accordance with the previous study that the increase could be linked to genetic factors [37].
In the present study, systolic and diastolic blood pressure was higher in obese children compared to overweight and normal-weight groups. Previous findings recorded a 31% elevated blood pressure in overweight and obese children, and in adolescents aged 9-17 years [38]. It has been reported in a study conducted on children and adolescents that there was a significantly higher systolic and diastolic blood pressure in overweight and obese compared to control participants [39]. A higher systolic and diastolic pressure was recorded in obese compared to lean participants in a study including middle-aged and elderly subjects. The significant correlation between diastolic pressure and BDNF in both males and females strongly suggests that plasma BDNF is important for cardiovascular disease [40]. In addition, the study showed a positive association of BDNF with BMI in female participants, but not in males. As shown in previous studies, a significantly positive correlation was observed between obesity measures (BMI, WC z-score, and WHtR z-score) and neurotrophins (BDNF and NGF) in our study. Roth et al. demonstrated a significant positive association of BMI and BDNF and leptin levels. Haploinsufficiency in human BDNF is negatively correlated with obesity, which emphasizes the pivotal role of BDNF in energy homeostasis regulation [41]. A recent report on plasma BDNF levels correlated significantly with WC, BMI, glucose, and systolic pressure among the pediatric population. The results evidence that BDNF levels tend to increase in association with BMI, with an energy homeostasis regulation in obese children [35].
On the other hand, the plasma concentration of NGF showed a significantly positive association with BMI, but the participants did not express a significant association with BDNF [13]. The results describe the relationship between salivary BDNF and NGF concentration, adiposity index, and fat mass in children. Although BMI may be a good indicator for overweight, it is an indirect measure of fat mass. In the current study, salivary BDNF and NGF levels were positively associated with BMI z-score, WC z-score, and WHtR z-score. Concerning blood pressure, a large school-based children's study conducted in China resulted in a positive correlation of BMI and WC with systolic and diastolic blood pressure [42]. A cross-sectional evaluation of school children showed a gradual elevation of blood pressure observed with increasing waist circumference [43]. A community-based cross-sectional study conducted on the Taiwan study population showed a positive correlation between BMI, WC, and systolic blood pressure [44], and these findings are congruent with our results.
The receiver-operating characteristics curve of BDNF and NGF is a good marker of obesity and metabolic complications with high sensitivity and specificity. In recent clinical study data of healthy and type 2 diabetes patients, diagnostic values of BDNF and NGF showed high sensitivity and specificity with very good AUC values [45]. The ROC curves for obesity prediction on the BDNF of a case-control study on school children expressed a prediction value above 0.65 in two different models [46]. In the current study, BDNF and NGF showed a positive association with insulin, and this result is supported by Levinger et al. A study conducted on middle-aged individuals reported a positive correlation between plasma BDNF and insulin [47]. Pedersen et al. reported in the Childhood Health, Activity, and Motor Performance School Study Denmark (CHAMPSstudy DK) that serum BDNF was positively associated with insulin in the healthy adolescent population [48]. We also found a positive correlation between both neurotrophic factors (NGF and BDNF), and this is consistent with the higher expression of NGF and BDNF of overweight and obese participants [13].

Conclusions
BDNF and NGF levels were higher in obese children compared to normal-weight children, indicating the connection between neurotrophic factors and fat mass. The outcome of the ROC curve had a good diagnostic value of BDNF and NGF, which suggests that they can be used as salivary biomarkers. Both neurotrophic factors have a positive association with insulin and obesity measures. These findings suggest that both BDNF and NGF may play pivotal roles in the pathogenesis of childhood obesity and act as a salivary biomarkers for predicting obesity-related complications.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/diagnostics12051130/s1, Table S1: Anthropometric measurements of study subjects.  Informed Consent Statement: Written informed consent was obtained from all the parents and participants.

Data Availability Statement:
The study datasets of the current manuscript are available from the corresponding author upon request.