Cardiometabolic Risk Markers in Children With Obesity and Variants in MC4R Pathway-related Genes

Abstract

Obesity is a serious health condition affecting a large number of children and adolescents in the United States and worldwide. The prevalence of obesity in children and adolescents has tripled in the last 4 decades, reaching 19.3% [1]. Obesity is a well-known risk factor for cardiometabolic risk factors including diabetes, metabolic dysfunction associated steatotic liver disease, dyslipidemia, and hypertension [2, 3]. The prevalence of these cardiometabolic comorbidities tends to be higher with increasing severity of obesity [3].

Leptin is a key hormone released by adipocytes [4]. In the fed state, leptin and insulin levels increase and subsequently bind to their receptors on the cellular surface of pro-opiomelanocortin neurons (POMC) that are located in the arcuate nucleus in the hypothalamus [4, 5]. This leads to release of α-melanocyte stimulating hormone, which decreases food intake, increases sympathetic activity, and increases resting energy expenditure through activation of MC4R receptors expressed in the paraventricular nucleus.

Variants in genes that control energy balance via the leptin-melanocortin 4 receptor (MC4R) signaling pathway or cell-signaling genes that influence this pathway peripherally or through their impact on the development of neurons that mediate energy homeostasis have been associated with an increased risk of developing obesity [6, 7].

Variants in LEPR, MC4R, MC3R, PCSK1, POMC, SIM1, and SH2B1 [6] are the most common cause of monogenic obesity contributing to 2% to 6% of individuals with severe obesity [8]. Guidelines from the Endocrine Society recommend genetic testing for children with early-onset severe obesity (before 5 years of age), individuals with syndromic obesity, and/or those with a family history of severe obesity [9].

The MC4R pathway- has been implicated in glucose homeostasis in animal models as well as in humans. In a study of mouse model, restoring leptin activity in POMC neurons normalized blood glucose levels and improved hepatic insulin resistance, dyslipidemia, and hyperglucagonemia [10]. Decreased MC4R signaling in the central nervous system has been associated with decreased glucose utilization in muscle and brown adipose tissue [11]. However, in another study, obese and weight-matched MC4R-deficient mice exhibited improved glucose tolerance as a result of elevated glucosuria [12]. MC4R deficiency was associated with suppression of renal sympathetic nerve activity, decreased circulating adrenaline, and lower renal GLUT2 levels [12]. Variants in the MC4R gene have been associated with an increased risk for type 2 diabetes via obesity-dependent and -independent effects [13, 14]. Increased triglycerides synthesis, lipid uptake, and fat accumulation in white adipose tissue have been noted in mice with genetic disruption of the MC4R. These effects occurred prior to changes in adiposity and were independent of food intake [11]. Studies in children have found an association between a specific variant in the PCSK1 gene (rs155971) and higher total cholesterol and higher low-density lipoprotein cholesterol (LDL-C) in children with obesity [15]. In another study, lipid levels were not any different between children with and without missense variants in the MC4R gene [16]. The MC4R pathway affects blood pressure via activation of the sympathetic nervous system [17].

Therefore, data from animal models and from human studies suggest the potential association between these genetic variants and development of obesity as well as an increased risk of cardiovascular risk factors including type 2 diabetes and dyslipidemia. MC4R deficiency has been noted to have a greater impact on the severity of adiposity, the rate of increase in body mass index (BMI), and the risk of type 2 diabetes during childhood compared with adulthood [18]. This may be partly related to a decrease in hyperphagia in adults [19]. Therefore, studies in children are particularly important as the associations between the genetic variants and obesity as well as other cardiovascular risk factors may be more likely to be detected during childhood.

There is a scarcity of data on the association between the presence of variants in genes related to the MC4R pathway or other obesity-related genes and the severity of obesity and cardiometabolic risk profile in children with obesity. Understanding this association is crucial in children to mitigate future cardiovascular risk. The objective of this study was to compare the severity of obesity and cardiometabolic risk markers in children MC4R-related clinically reported genetic variants relative to children with non-MC4R-related genetic variants and those with no clinically reported variants in monogenic obesity-related genes.

Methods

This was a retrospective cohort study conducted by chart review of children (age ≤18 years) with obesity who underwent a multigene panel testing for monogenic obesity in the outpatient setting in a multidisciplinary weight management clinic in the division of Pediatric Endocrinology at Mayo Clinic (Rochester, MN) between February 2020 and January 2023. Signed informed consent/assent was obtained from all patients/guardians prior to genetic testing. The study was approved by the Mayo Clinic Institutional Review Board.

Study Participants

Genetic testing was obtained at the discretion of the medical provider based on their clinical judgment including history of early-onset obesity. No formal guidelines were provided to the providers on specific criteria for obtaining the gene panel. Genetic testing was part of clinical practice and was not done for the purpose of research. Inclusion criteria were (1) age ≤18 years of age and (2) BMI ≥95th percentile for age and sex. Exclusion criteria were the presence of secondary causes of weight gain: (1) medication-induced weight gain; (2) the presence of an endocrine disorder that can cause weight gain such as Cushing syndrome, untreated hypothyroidism, or untreated GH deficiency; and (3) acquired hypothalamic dysfunction.

Data Collection

The demographic, anthropometric and biochemical parameters were obtained from a medical record review. Data include age, sex, race/ethnicity, maximum BMI percent of the 95th percentile (between age 2 and current age) and systolic and diastolic blood pressure (Table 1). The BMI percent of the 95th percentile was chosen as the metric for severity of obesity as it is the recommended metric for children with severe obesity. The BMI Z-scores have been shown to have almost no association with adiposity measures such as waist circumference, arm circumference, triceps skin full-thickness, and fat mass in children with severe obesity [20]. The maximum BMI percentile was extracted from electronic medical records and was derived from height and weight measurements that had been recorded in the pediatric endocrinology and primary care outpatient clinic visits. These BMI measurements were reported using the sex-specific Centers for Disease Control and Prevention extended BMI charts.

The laboratory data included maximum value of fasting glucose, hemoglobin A1c (HbA1c), alanine transaminase (ALT), total cholesterol, high-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol, calculated LDL-C, and triglycerides. The maximum BMI percentile and maximum values for laboratory tests were extracted from the medical records and did not have to be within a specific timeframe of each other.

Genetic Testing

Multigene panel done at a CLIA-certified laboratory was obtained in patients after informed consent. The gene panel identified clinically reported variants in 40 genes between February 2020 and June 2021 and 79 genes from July 2021 onward. The multigene panel comprises genes contributing to controlling energy balance via the leptin-melanocortin 4 receptor signaling pathway cell-signaling, genes that interact with the MC4R pathway peripherally or influence the development of neurons responsible for energy homeostasis as well as genes associated with syndromic-monogenic obesity such as Bardet–Biedl syndrome and ciliopathies.

Based on genetic test results, patients were divided into 2 groups. The first group included children with identifiable MC4R-related genetic variants (group A), and the second group included patients without MC4R-related genetic variants (group B). Children in group A were divided into 2 subgroups: those with PCSK1 risk alleles (group A1) and those with MC4R-related variant of uncertain significance (VUS) (group A2). Group A1 was created as the PCSK1 variant was the most frequently reported and the only risk variant identified in our cohort of MC4R-related variants. Group B was subdivided into 2 groups: no clinically reported variants or negative testing (group B1) and those with non-MC4R-related variants (pathogenic, likely pathogenic, and VUS) (group B2). Children with more than 1 genetic variant related to the MC4R pathway and non-MC4R- related genes were included in group A. Maximum BMI percent of the 95th percentile; laboratory parameters; and family history of obesity, type 2 diabetes, hypertension, and dyslipidemia were compared between groups (and subgroups). Prediabetes was defined as fasting blood glucose between 100 and 125 mg/dL or HbA1C between 5.7 and 6.4% [21]. Diabetes was defined as fasting blood glucose ≥126 mg/dL or HbA1C ≥6.5% [21]. Dyslipidemia was defined as total cholesterol ≥200 mg/dL, LDL-C ≥130 mg/dL, non-high-density lipoprotein cholesterol ≥145 mg/dL, or triglycerides ≥100 mg/dL for children between 0 to 9 years of age and ≥130 mg/dL for children and adolescents between 10 to 18 years of age [22]. Clinical steatotic liver disease was diagnosed as ALT ≥2 times the upper range of normal for sex (44 units/L for girls and 52 units/L for boys) [23, 24] and/or if there was evidence of fatty infiltration on imaging (ultrasound, FibroScan, or magnetic resonance elastography).

Statistical Methods

Quantitative variables that were normally distributed were summarized with mean and SD. Those that were not normally distributed were summarized with median and interquartile range. Qualitative variables were summarized with frequency and percentage. Summaries were presented for the total sample, each exposure group, and each subgroup, and groups A vs B were compared. Given that the PCSK1 variant was the most frequently reported and the only risk variant identified in our cohort, we chose to separate patients with MC4R variants (group A) into a distinct subgroup (group A1) for comparison with group A2, which consists of MC4R-related variants other than PCSK1. Similarly, within group B, which included patients without any MCR4-related variants, we aimed to investigate whether non-MC4R genetic variants have an association with BMI (group B2) and cardiometabolic risk markers compared to individuals with negative genetic testing results (group B1). Therefore, additional hypothesis tests were conducted to evaluate the differences between subgroups A1 vs A2 and B1 vs B2 for all outcomes. Post hoc analysis was conducted to evaluate differences between subgroups A1 vs A2, A1 vs B1, and A1 vs B2 for LDL-C and high-density lipoprotein.

Quantitative variables that were normally distributed were tested for differences with univariable linear regression and multivariable linear regression conditional on maximum BMI percent of the 95th percentile. Quantitative variables that were right-skewed were log-10 transformed, and variables that were left-skewed were transformed as follows: 2 ^ (x / 10). The transformed variables were then modeled in the same procedure as the normally distributed variables. Quantitative variables that contained extreme values determined to not be erroneous were modeled with robust linear regression using the M method in place of linear regression as in the approach for normally distributed variables. Fisher's exact test was used to test the prevalence of type 2 diabetes. Estimated model coefficients, 95% confidence intervals (CIs), and P-values were presented.

Binary/categorical variables were tested for differences with univariable logistic regression and multivariable logistic regression conditional on maximum BMI percent of the 95th percentile. Estimated odds ratios, 95% CIs, and P-values were presented.

The data analysis for this paper was generated using SAS software, Version 9.4 for Linux (SAS Institute Inc., Cary, NC). Statistical significance was determined at the 95% confidence level using α = .05.

Results

Data on a total of 106 children who underwent genetic testing between February 2020 and January 2023 were reviewed. Two children were excluded due to a diagnosis of other genetic conditions. Therefore, the study reports on data from 104 children (Fig. 1). The mean (SD) age at testing was 11 years (4). The mean maximum (SD) BMI percent of the 95th percentile was 160 (28). One hundred children out of the 104 (96%) had severe obesity defined as BMI ≥120% of the 95th percentile. Ninety-three (89%) were White, and 88 (85%) were not Hispanic or Latino. Table 1 demonstrates the demographic characteristics of study participants.

Figure 1.

Flowchart illustrating the process of grouping study participants and conducting data analysis.

Open in new tabDownload slide

Reported Genetic Variants

Thirty-nine children had clinically reported variants (pathogenic, likely pathogenic, risk allele, or VUS) within the MC4R pathway (group A), and 65 children had no reported variants in the MC4R pathway (group B). Within group A, 14 children had heterozygous PCSK1 risk alleles (A1) and 1 child was homozygous for the PCSK1 risk allele. One child who was heterozygous for a risk allele also had a likely pathogenic PCSK1 variant with unknown significance. This patient was included in group A1. No patient had clinically reported pathogenic variant in the PCSK1 gene or MC4R-related genes. Twenty-four children had a VUS in the MC4R-related genes (A2). Out of the 65 children without any MC4R-related variants, 38 children had no clinically reported genetic variants or negative testing (group B1) and 27 children had single pathogenic (2 individuals), likely pathogenic (3 individuals), or VUS in non-MC4R- related genes.

Among the group with clinically reported variants in the MC4R pathway, PCSK1 risk variants were most reported in 15 children (14%) followed by CEP290 variants in 8 children (7.7%) including 1 pathogenic variant and 7 VUS (Fig. 2).

Figure 2.

The frequency of reported genetic variants.

Open in new tabDownload slide

Severity of Obesity

The maximum BMI percent of the 95th percentile was not different between the group with MC4R-related variants (group A) and the group without any MC4R-related variants (group B) [mean (SD) in group A 154% (28) and 163% (28) in group B, P = .116]. Subgroup analysis did not reveal any difference between children with PCSK1 risk alleles (A1) relative to those with MC4R-related VUS (A2) [155% (23) and 153% (31) in group A2, P = .872]. Similarly, there was no difference in BMI percent of the 95th percentile between the children with negative genetic testing (group B1) vs children with non-MC4R-related variants (group B2) [161% (26) and 166% (32) in group B2, P = .458] (Table 2, Fig. 3).

Figure 3.

Comparison of body mass index percentile between groups stratified by genetic test results.

Open in new tabDownload slide

Upon combining children with positive genetic variants in either MC4R- and non-MC4R-related genes into a single group, referred to as “positive results,” there were no significant differences in BMI percentiles between this combined group and children with no reported genetic variants, classified as “negative results” [159% (26) vs 161% (30), P = .748]. Also, the number of variants captured by the testing panel per each patient was not associated with an increase in BMI (P = .173) (Fig. 4). Family history of obesity was not associated with the presence of a variant in monogenic obesity-related genes [odds ratio (95% CI) 1.58 (0.58,4.33), P = .375].

Figure 4.

Analysis using the number of variants reported by the clinical multigene panel to test for association with BMI measurement. A higher number of genetic variants identified per individual was not found to be associated with an increase in BMI severity.

Abbreviations: BMI, body mass index.

Open in new tabDownload slide

Cardiometabolic Risk Markers

Considering the effect of severity of obesity on glycemic status and various cardiometabolic parameters, we adjusted our dataset concerning cardiometabolic risk markers by accounting for maximum BMI. The comparative analysis of cardiovascular markers, unadjusted for BMI, is detailed in Table 3. The overall prevalence of dyslipidemia in those with clinically reported genetic variants was 51%. Two patients were excluded from the analysis because of a diagnosis of familial hypercholesterolemia. The LDL-C was not different between children with MC4R-related variants and children with no MC4R-related variants [mean (SD) 101 mg/dL (31) and 92 mg/dL (31) respectively, P = .132]. The LDL-C was higher in children with the c.661A > G PCSK1 risk allele relative to those with MC4R-related VUS [(mean (SD) 114 (37) and 94 (24) respectively, P = .047]. The mean LDL-C was also higher in children with PCSK1 risk alleles relative to those with negative genetic testing (P = .012) and those with non-MC4R-related variants (P = .048) (Table 2, Fig. 5). There was no significant difference in LDL-C values in those with positive genetic testing relative to those with no clinically reported genetic variants. The total cholesterol, high-density lipoprotein cholesterol, and triglycerides were not different between group A vs group B. Subgroup analysis did not reveal significant differences between groups (Table 2, Fig. 5).

Figure 5.

Comparison of fasting cholesterol levels stratified by genetic test results (A-C) and triglyceride levels (D-F).

Open in new tabDownload slide

The prevalence of type 2 diabetes was 8% in children with MC4R-related variants and 3% in the non-MC4R group (P = .368). One patient was excluded from the analysis because of a diagnosis of type 1 diabetes. The fasting blood glucose in children with MC4R pathway variants was not different from those with non-MC4R-related variants [median (Q1-Q3): 92 mg/dL (86-97) and 92 mg/dL (87-97) respectively, P = .682]. HbA1c level, blood pressure percentiles, and ALT were not different between the groups and subgroups (Table 2, Fig. 6).

Figure 6.

Comparison of fasting blood glucose and hemoglobin A1c between groups stratified by genetic test results.

Open in new tabDownload slide

Family history of type 2 diabetes was not statistically different between those with negative vs positive genetic testing [odds ratio (95% CI) 1.51 (0.55, 4.12), P = .426)] or with the presence of MC4R-related variants vs those with no MC4R-related variants [odds ratio (95% CI) 1.09 (0.39, 3), P = .873]. Similarly, a family history of dyslipidemia was not associated with the presence of any genetic variant [odds ratio (95% CI) 0.66 (0.21, 2.1), P = .479] or the presence of any MC4R-related variants [odds ratio (95% CI) 0.7 (0.22, 2.23), P = .543] (Table 4).

For most outcome variables, there was no evidence of interaction between sex and study group. For those variables, the estimates resulting from models adjusted for maximum BMI percent of the 95th percentile, testing age, and sex did not meaningfully differ from those resulting from models adjusted for only maximum BMI percent of the 95th percentile.

Discussion

In this study of children and adolescents with obesity who underwent testing for monogenic obesity in a tertiary healthcare clinical practice, 63% of patients were found to have a clinically reported variant in genes associated with monogenic obesity. We did not find an association between the presence of clinically reported genetic variants and severity of obesity, blood pressure percentiles, glycemic parameters, and ALT. To our knowledge, this is the first study to examine the association between the presence of clinically reported genetic variants in any of the monogenic obesity genes and, specifically, MC4R-related genes and cardiometabolic risk factors in a pediatric cohort.

The prevalence of reported variants was 61.5% in a study in which patients had been selected mostly due to self- or caregiver-reported overeating to excess or binge eating (80.3% of study population) or sneaking food or eating in secret (59% of study population) [7]. The prevalence of reported genetic variants was different in another study from the United States (44%—total number of subjects of 117) that had a large age distribution with 30% subjects being 18 years of age and older [25]. The high prevalence of positive results highlights the need to support and educate clinicians with an understanding of the implications of these findings and the importance of appropriate utilization of these results for selecting children who might be candidates for treatment with MC4R agonists such as setmelanotide, as the list of genetic variants that qualify for treatment may expand in view of the ongoing phase 2 and phase 3 clinical trials (NCT04963231, NCT05093634).

Of those who had a positive test (n = 66), 59% (n = 39) had a genetic variant in a gene related to MC4R. This emphasizes the high prevalence of these variants among children with obesity. Our study specifically identified 15 children (14%) with the common PCSK1 risk allele c.661A>G - p. (Asn221Asp), which has been linked to suboptimal genetic expression resulting in a decrease in the secreted enzyme level [26-29]. Of note, this variant has an allele frequency of 6.6% in the South Asian population and is as high as 7.5% in the Indian population. The prevalence of the risk allele in our multiethnic obesity cohort was almost 2 times the highest prevalence in general populations, highlighting its contribution to the development of obesity.

We did not find any association between the presence of clinically reported variants and severity of obesity. These results are similar to those from previous studies [25, 30]. There was no significant difference in BMI status in those who have a PCSK1 risk allele relative to those with and without other clinically reported variants in the monogenic obesity-related genes. One study found that heterozygous children for the PCSK1 risk allele had a 1.25-fold higher risk (95% CI = 1.081-1.444, P = .027) of increased waist circumference than patients with the normal genotype, after adjustment for age, sex, and BMI. However, this impact was not observed on BMI status [15]. Another study in China found a significant association between a polymorphism in PSCK1 (rs155971) and a higher risk of obesity [31].

Obesity is an inflammatory condition associated with the production of inflammatory mediators and adipokines, leading to subclinical inflammation, insulin resistance, and eventually overt cardiovascular disease [32]. In individuals with obesity, certain variants in MC4R pathway-related genes have been identified as potentially influencing glucose homeostasis and lipid metabolism. A study examining the connection between the likely pathogenic MC4R p. (Ile269Asn) variant and the risk of type 2 diabetes in Mexican adults showed that the variant was linked to an increased risk for type 2 diabetes, even after accounting for BMI adjustments (odds ratio = 1.70, 95% CI 1.13-2.56, P = .011). Other studies revealed the potential influence of some PCSK1 variants on glucose metabolism in children with obesity [29]. Noteworthy, a study in the Thai population reported an association of homozygous PCSK1 rs155971 polymorphism and high cholesterol and LDL-C levels in children [15]. Within our study cohort, individuals carrying the c.661A>G PCSK1 risk allele exhibited higher LDL-C levels compared to those with MC4R-related VUS, individuals with negative genetic testing, and those with non-MC4R-related genetic variants. This finding underscores a potential link between PCSK1 variants and increased LDL-C levels. There were no significant differences observed in other lipid profile parameters between the groups. Moreover, no significant differences were observed in fasting blood glucose, HbA1c, ALT, and blood pressure percentiles between these groups.

Even though most patients in our cohort met the criteria for genetic testing defined by the American Endocrine Society, a history of early-onset obesity or a positive family history were not sufficient to predict the presence of a pathogenic variant in monogenic obesity-related genes. Thus, further studies are warranted to better identify criteria for genetic testing in children with obesity. Such criteria are particularly valuable, especially in light of the approval of setmelanotide for the treatment of individuals aged 6 years and older who carry pathogenic, likely pathogenic, or VUS, leading to leptin, POMC, or PCSK1 deficiency. Additionally, setmelanotide has been approved for individuals with Bardet–Biedl syndrome due to promising results with improvement in body composition [33-35].

Major limitations of our study are the small sample size and lack of racial and ethnic diversity, both of which limit the generalizability of the findings. Other limitations include the lack of a screening questionnaire for hyperphagia, the cross-sectional and retrospective nature of the study, and lack of information on eating habits and physical activity, both of which can influence BMI and cardiometabolic risk markers. There is a possibility of a selection bias as the decision to test was based on the clinical judgment of the medical provider and not done as part of a research protocol. Moreover, the multigene panel employed in this study focused solely on screening for monogenic obesity-related genes, with no report of polymorphisms other than the known risk alleles, potentially overlooking genetic variants that contribute to polygenic predisposition of obesity. Furthermore, since most of the reported genetic variants are rare, the analyzed groups consisted of individuals with various unique variants. These genetic variants within each group could have produced heterogeneous effects, and that could potentially result in no clinically significant distinction between the groups. Additionally, some variables in this study exhibited nonlinearity and displayed outliers, necessitating the use of different analysis methods. Another limitation was the absence of weight and height information from birth in all our patients due to the retrospective nature of the data. Therefore, we were not able to compare the age of attaining maximum BMI percentile in the study. Additionally, the LDL-C and fasting glucose can be influenced by pubertal status, and this information was not available for a large proportion of the study participants.

In conclusion, our findings indicate that almost two-thirds of the screened children with a suspected genetic cause for obesity carry at least 1 genetic variant in an obesity-related gene. The most frequently reported variant in our cohort was the PCSK1 c.661A>G risk allele, accounting for 14% of cases, all of whom were White. However, the presence of a monogenic obesity-related variant did not demonstrate a significant association with higher BMI status or increased cardiometabolic risk, except for the risk allele, which was associated with higher LDL-C levels compared to those with VUS in MC4R-related genes, individuals with negative genetic testing, or those with variants in non-MC4R- related genes. This might highlight that patients harboring this variant may warrant enhanced surveillance pertaining to lipid profiles or potential alterations in management/dietary strategies preemptively, prior to the onset of hyperlipidemia. This proactive approach aims to mitigate the long-term risk of cardiovascular events. Additionally, a family history of obesity did not correlate with the presence of reported monogenic obesity-related genetic variants. These findings underscore that other genetic and nongenetic factors likely play a crucial role in determining the severity of obesity and its associated cardiometabolic risk factors. Larger-scale studies are needed to further evaluate the correlation between genes within the MC4R pathway and cardiometabolic risk.

Disclosures

S.K. is a clinical trial investigator for Rhythm Pharmaceuticals. A.C. has received study support from Medtronic. None of the authors have any conflicts pertinent to the current manuscript.

Data Availability

Original data generated and analyzed during this study are included in this published article or in the data repositories listed in References.

References