Lifelong impacts of puberty timing on human plasma metabolic profiles: A metabolome‐wide Mendelian randomization study

Abstract Aim The aim was to investigate the causal relationship between puberty timing and plasma metabolites, accounting for birth weight, childhood and adulthood adiposity. Materials and Methods The meta‐analysis of genome‐wide association studies (GWAS) for puberty timing was extracted from the ReproGen Consortium, involving 329 345 women of European ancestry. Summary data for 174 plasma metabolites were retrieved from a recently conducted cross‐platform GWAS that involved a meta‐analysis of three cohort studies (i.e. the Fenland, European Prospective Investigation into Cancer‐Norfolk and INTERVAL studies) and three publicly available studies and included up to 86 507 participants. We conducted a two‐sample Mendelian randomization (MR) analysis to infer the causal relationship of puberty timing on 174 plasma metabolites, complemented by a two‐step and multivariable Mendelian randomization (MVMR) analysis to assess direct and indirect effects. Additionally, summary‐level data from the UK Biobank were used for our replication analysis. Results The results of the two‐sample MR provide moderate evidence supporting a causal relationship between puberty timing and 23 of 174 plasma metabolites (i.e. 7 acylcarnitines, 8 amino acids, 2 biogenic amines and 6 lysophosphatidylcholines). Even after single‐nucleotide polymorphisms associated with birth weight and childhood adiposity were excluded, causal effects persisted for 16 metabolites (i.e. 8 acylcarnitines, 4 amino acids, 2 biogenic amines and 2 lysophosphatidylcholines). The two‐step MR analysis provided evidence that the relationship between puberty timing and plasma metabolites was mediated by adulthood adiposity. Additionally, moderate evidence emerged for an independent causal effect of puberty timing on 10 metabolites through an MVMR analysis (i.e. 5 acylcarnitines, 2 amino acids, 1 biogenic amine, 1 lysophosphatidylcholine and 1 phosphatidylcholine). Furthermore, the replication analysis suggested the robustness of our results. Conclusions In summary, our study provides compelling evidence that puberty timing has a causal influence on certain plasma metabolites, although this influence is largely mediated by adulthood adiposity.