Ferroptosis and Immunogenic Cell Death
Ferroptosis is defined by the loss of a single activity. GPX4 is the only enzyme that reduces lipid hydroperoxides inside a membrane, and when it stops, peroxidation propagates until the membrane fails. That makes this death mode unlike apoptosis, which has redundant executioners, and unlike pyroptosis, which has a dedicated pore. It also makes it unusually easy to get wrong: FSP1 is a second, entirely GPX4-independent brake, so a cell can shrug off GPX4 inhibition completely, and ACSL4 decides whether a cell has peroxidisable lipids in its membranes at all. The second half of this map is the question that makes ferroptosis interesting to immunologists rather than only to cell biologists: does a cell dying this way get noticed? Immunogenic death needs calreticulin on the surface, and it needs the phagocyte to weigh that against CD47. Click any protein for the matching Assay Genie In Vivo antibody or ELISA kit.
One enzyme, one backup, and one prerequisite. GPX4 is the only activity that reduces lipid hydroperoxides in a membrane, and every ferroptosis inducer works by removing it directly or by starving it of glutathione \u2014 which is what SLC7A11 blockade does, and why high extracellular glutamate has the same effect. FSP1 is the second brake, entirely independent of GPX4, regenerating ubiquinol at the membrane. Its discovery explained a long list of cells that ignored GPX4 inhibition, and it is the reason a GPX4-only experiment can be confidently wrong. ACSL4 is the prerequisite rather than a brake: it determines which fatty acids enter membrane phospholipids, so an ACSL4-low cell has nothing to peroxidise and is resistant whatever its GPX4 status. In practice ACSL4 expression is the most useful single predictor of whether a cell line will ferroptose at all, and it belongs in the methods rather than the discussion.
Iron is in the name for a reason, and ferritin is two-faced. Labile iron drives the Fenton chemistry that propagates peroxidation alongside the enzymatic ALOX15 route. Ferritin stores iron in a redox-inactive form, so it is protective when intact and dangerous when degraded by ferritinophagy \u2014 which is why total ferritin on its own is a poor read-out. HO-1 is the same problem in sharper form: it is an antioxidant enzyme driven by NRF2, and it liberates iron from heme. Whether it protects or kills is a question of degree, so directional claims about HO-1 are worth checking rather than accepting. NRF2 itself is the master resistance programme, raising GPX4, SLC7A11, ferritin and HO-1 together, and its constitutive activation in many tumours is a substantial part of why they resist this death mode.
Immunogenic death is a balance, not a property of the corpse. Calreticulin translocating to the surface is the eat-me signal, read by LRP1/CD91 on the phagocyte; HMGB1 and ATP are the danger signals that follow. But the dying cell may still carry CD47, and the phagocyte does the arithmetic between them \u2014 the same subtraction that governs the myeloid checkpoint map in this library, arriving here from the opposite direction. HMGB1 adds a complication specific to this death mode: its redox state decides its function \u2014 reduced is chemotactic, oxidised is inflammatory, fully oxidised is inert \u2014 and in a cell that is dying of peroxidation, that state is not incidental. Whether ferroptosis is reliably immunogenic is genuinely unsettled, and a map that asserted otherwise would be overselling it.
The In Vivo tie-in, and where this map is thin. 15 of the 35 nodes carry a functional-grade antibody, and they sit almost entirely downstream of the death itself. The immunogenicity half is well covered: SIRP\u03b1 (CD172a), CD11c, CD68, CSF-1R, class I (H-2Kb), CD8 (Ly-2), CD4 (GK1.5), IFN-\u03b2, IFN-\u03b3 (XMG1.2), TNF, IL-1\u03b2, PD-L1 (10F.9G2), IL-10 and TGF-\u03b2 are all blockable in vivo. That set supports the experiment this field turns on \u2014 the vaccination assay, in which dying cells injected into an immunocompetent animal protect it against later live challenge \u2014 with CD8 and class I blockade to show the protection was antigen-specific rather than bystander inflammation, and CSF-1R to show phagocytes were required. CD71 is the one engine node that is functional grade, which means the iron requirement can be tested rather than assumed. Stated plainly, the rest of the engine is read-out only: GPX4, FSP1, SOD1, HMGB1, LRP1 and the MDA-adduct antibody are research grade; SLC7A11, ACSL4, ALOX15, ferritin H, NRF2, HO-1, GSR, 4-HNE, MDA, calreticulin, annexin A1, P2X7 and CXCL10 are ELISA only; and CD47 is biosimilar. Two measurement cautions worth carrying: MDA is produced by several oxidative routes, so on its own it evidences oxidative stress rather than ferroptosis \u2014 4-HNE is the more specific marker, and the MDA-adduct antibody answers a different question again, since adducts accumulate while free aldehyde turns over. And SOD1 is here as a control rather than a participant: it handles superoxide, not lipid peroxides, so general antioxidant capacity and ferroptosis resistance are not the same measurement. For research use only; not for use in diagnostic or therapeutic procedures.
Every protein node links to a product — ELISA kit, In Vivo antibody or research antibody.