A fin-loop-like structure in GPX4 underlies neuroprotection from ferroptosis Original paper

What was studied?

This primary research study dissected how glutathione peroxidase 4 (GPX4) prevents ferroptosis in the nervous system and why a patient-associated missense variant, GPX4 p.R152H, causes progressive neurodegeneration. The authors combined structural biology (NMR and X-ray crystallography), membrane-binding assays, and ferroptosis phenotyping in human patient-derived cells and conditional mouse models to connect a specific GPX4 structural element to neuronal survival and Alzheimer’s-like molecular signatures.

Who was studied?

The study evaluated three individuals from two unrelated families carrying the homozygous GPX4 c.455G>A (p.R152H) variant with a protracted, hypomorphic Sedaghatian-type spondylometaphyseal dysplasia course and progressive CNS atrophy on MRI. The mechanistic work used patient fibroblasts, patient iPSC-derived cortical neurons, and forebrain organoids, alongside mice with inducible neuron-specific Gpx4 loss or neuron-specific expression of the patient-mimicking allele to model cortical and cerebellar degeneration in vivo.

What were the most important findings?

GPX4 protected neurons from ferroptosis through a hydrophobic “fin-loop”-like structure that anchored the enzyme to membranes, positioning it at the site where lethal phospholipid peroxidation occurs; the p.R152H substitution collapsed this fin-loop and weakened membrane engagement while leaving GPX4 expression, stability, and catalytic peroxide-reduction capacity largely intact. In patient-derived fibroblasts and iPSC-derived cortical neurons, the homozygous p.R152H genotype drove rapid, lipid peroxidation–linked cell death that ferroptosis inhibitors (notably liproxstatin-1) prevented, and forebrain organoids showed degeneration and collapse unless ferroptosis was pharmacologically blocked. In vivo, temporally controlled neuronal Gpx4 deletion or neuron-specific expression of the p.R152H allele produced progressive cortical atrophy, Purkinje neuron loss, and escalating neuroinflammation; neuronal injury rose early as reflected by marked plasma neurofilament light chain increases, and ferroptosis inhibition reduced neuronal death signals. Proteomic profiling of affected brains converged on dementia-like pathways, including enrichment of Alzheimer’s disease signatures and upregulation of canonical AD-associated proteins such as APOE and CLU, while immune landscapes shifted toward disease-associated microglial states (e.g., TREM2-linked activation).

What are the greatest implications of this study/ review?

This work makes ferroptosis mechanistically actionable in neurodegeneration by showing that neurons fail when GPX4 cannot properly localize to membranes, even if its enzyme chemistry remains competent. Clinically, the findings elevate ferroptosis inhibition from a correlative idea to a targetable upstream driver of neuronal loss and secondary neuroinflammation, and they justify biomarker strategies (e.g., neurofilament light chain dynamics) to track ferroptosis-linked neuroaxonal damage in therapeutic development.