Mendonça and colleagues asked what a non-killing iron load does to neuroblastoma mitochondria. SH-SY5Y cells saw 100 µM iron sulfate for a day. Intracellular iron rose about four-fold. The cells divided more and moved less. Mitochondria broke into fragments and lost membrane potential, but they did not lose mass or OXPHOS coupling. Only proton-leak respiration went up. Cellular oxidants and hydrogen peroxide output went down. Ferritin went up. Ribosomal proteins went down.
Metal sites and redox
The input metal is iron from FeSO4, so the medium salt is ferrous. ICP-MS counts elemental iron, not Fe2+ versus Fe3+ inside the cell. Ferritin is the upregulated store. No [2Fe-2S], [4Fe-4S], heme, or copper site is reported. The named redox product is H2O2, and it fell. That sits with the known steep dependence of superoxide formation on Δψm at Complex I and at the Complex III Qo quinol site. Extra leak can drop potential into a quieter range without uncoupling ATP synthesis if ETS capacity remains. The authors say that. They did not measure the Q/QH2 pool or a specific Fe-S core oxidation level. Do not overwrite the result with a Fenton story the H2O2 data contradict.
What was measured
Viability and cytotoxicity (to define non-toxic), proliferation, migration, ICP-MS iron, confocal morphology and Δψm, respirometry including leak and ETS, oxidant and H2O2 probes, proteomics.
Caveats
One cell line. Leak is a flux, not a molecular identity. The suggestion that blocking this antioxidant adaptation would change tumor progression is not tested here.
What to do with it
When an agent sees iron plus neuroblastoma, do not assume more ROS. This pulse stored iron, leaked protons, dropped potential, and lowered H2O2 while growth sped up. The experiment to demand next is which leak channel, and whether clamping Δψm brings oxidants back.