Identification of circulating protein biomarkers for childhood obesity using mendelian randomization

Childhood obesity is a major public health problem affecting one in 10 youth and is associated with several short and long term cardiometabolic complications. Characterisation of early biomarkers for pediatric obesity is necessary to develop screening tools, to prioritize children at risk for preventive interventions, while these biomarkers represent potential therapeutic targets. Recent advances in genomics and proteomics have unveiled genetic variants for thousands of circulating proteins in genome-wide association studies (GWAS). Here, our goal was to identify circulating proteins with causal effects on childhood obesity, leveraging proteomic and pediatric body mass index (BMI) GWAS data, in a two-sample Mendelian randomization (MR) study. We identified genome-wide significant (P < 5x10-8) polymorphisms (SNPs) associated with circulating levels of 1,297 proteins in 5 large GWAS (Sun et al, Emilsson et al, Folkersen et al, Yao et al , Suhre et al). We obtained the effects of SNPs on the level of each protein from these GWAS and their effects on pediatric BMI from a GWAS by Vogelezang et al. (N=61,111 children). Subsequently, we estimated the causal effect of circulating proteins on BMI by calculating the Wald ratio for proteins with a single SNP-instrument or by meta-analyzing the effects of multiple SNP-instruments per protein on pediatric BMI using the Inverse Variance Weighted (IVW) approach. Among 210 proteins tested to date (with SNPs available in Suhre et al, Folkersen et al and Yao et al GWAS), we identified two proteins whose MR effect (Wald ratio) had a P-value below the Bonferroni-corrected threshold (< 2.4 x10-4). These are the ENG (endoglin; MR beta: -0.07, 95% CI= -0.10, -0.05, P= 4.4 x10-5) and the FABP4 (fatty acid binding protein 4, MR beta: -0.33 95% CI -0.5, -0.16, P=1.3 x10-4), both having a negative effect on BMI per standard deviation increase in their serum level. ENG is expressed in the vascular wall, while FABP4 has a known role in fatty acid transport and metabolism. Neither ENG nor FABP4 have been previously associated with adult or pediatric BMI in GWAS. MR studies are ongoing to test proteins from additional proteomic GWAS for causal effects on pediatric BMI. Subsequently, our MR results will be validated by genetic colocalization studies, aiming to identify which among the MR-prioritized proteins share causal SNPs with childhood BMI. Finally, we will use the candidate proteins to identify protein-protein interaction networks that will point to predominant pathophysiological pathways in pediatric obesity.