Causal landscape of the circulating proteome in systemic iron overload revealed by Mendelian randomization and experimental validation

Abstract Systemic iron overload precipitates progressive and often irreversible organ dysfunction, yet the clinical toolkit for non-invasive, organ-specific monitoring of iron accumulation remains inadequate. This study leverages a proteome-wide causal inference framework to identify circulating biomarkers that drive or reflect iron deposition in the liver, spleen, and pancreas. We integrated large-scale plasma pQTL data (n = 35,559) with GWAS summary statistics of organ-specific iron content using two-sample Mendelian randomization (MR). Candidate proteins were prioritized through Bayesian colocalization to identify shared genetic etiologies. The findings were validated in a systemic iron-overload mouse model, assessing the concordance between genetically predicted effects and experimental protein expression, histopathology, and systemic iron markers. MR analysis identified PRSS1 as a high-confidence causal mediator and prioritized BGLAP, FTMT, and ZHX3 as candidate proteins associated with tissue iron homeostasis. Genetically predicted BGLAP levels exhibited a negative causal effect on liver iron content (β = -0.294, P = 6.35 × 10⁻⁶), while FTMT and ZHX3 were positively associated with splenic iron accumulation. Notably, PRSS1 demonstrated a robust negative association with pancreatic iron (β = -0.217, P = 1.28 × 10⁻⁵) and showed strong evidence of genetic colocalization (PP.H4 = 0.982), suggesting a direct regulatory role. These MR-prioritized protein targets exhibited directionally concordant in vivo associations with organ iron accumulation and pharmacological iron chelation in mouse models of systemic iron overload. Our study establishes a genetically anchored proteomic map of systemic iron variation, prioritizing PRSS1 alongside BGLAP, FTMT, and ZHX3 as candidate proteins associated with organ-specific iron homeostasis. Rather than providing definitive clinical tools, these findings offer hypothesis-generating insights that warrant rigorous prospective validation in clinical iron-overload patient cohorts and dedicated functional models.