Ferroptosis: regulated cell death

GPX4 activity may be lowered by direct enzyme inhibition (loss of activity or stimulation of enzyme protein degradation) or inhibition of the antiporter system Xc^-. In the first case, inhibition is most often mediated by RAS-synthetic lethal 3 (RSL3), an alkylating molecule that irreversibly binds to selenocysteine in the active site of GPX4 (30), but there are other inducers of ferroptosis that can deplete or degrade the enzyme protein, such as ferroptosis inducer 56 (FIN56) and caspase independent lethal 56 (CIL56), a molecule that causes non-apoptotic cell death through an acetyl-CoA-carboxylase-1-dependent process (20, 34). Ferroptotic cell death can also be induced through genetic inhibition of GPX4 by siRNA (35).

As for the other pathway – the inhibition of system Xc^- – here is how this membrane-based, sodium-independent and chloride-dependent cystine/glutamate antiporter system works: it imports extracellular cystine in exchange for intracellular glutamate (36) and through the catalytic action of cystine reductase transforms cystine into cysteine, which is required for the synthesis of GSH and subsequently GPX4. Under normal conditions GPX4 catalyses the reduction of PL-PUFA(PE)-OOHs into alcohols (37). When this system is inhibited by small molecules such as erastin and its analogues (piperazine erastin and imidazole ketone erastin) and/or sorafenib, GSH levels drop, GPX4 is inactivated, and L-ROS starts to accumulate (18, 30).

Aerobic organisms are continuously exposed to various reactive oxygen species (ROS) such as superoxide radicals (*O₂^-), hydrogen peroxide (H₂O₂), hydroxyl radical (*OH), and lipid peroxides (L-ROS) such as L-OOH, peroxyl (L-OO*), and L-O* radicals (38). While low, controlled L-ROS levels are acceptable for normal cell and organism functions, higher levels are associated with numerous chronic degenerative processes and acute organ injuries. These conditions are the result of an oxidant-antioxidant imbalance that results in oxidative stress.

The formation of cellular L-ROS involves iron-catalysed spontaneous chain reaction generating toxic radicals (non-enzymatic process) (20, 39) and enzyme-mediated oxidation of PUFAs (40, 41). Lipid compounds that are the most sensitive to lipid peroxidation and are thus involved in triggering ferroptosis are PL-PUFAs, arachidonic acid (AA) in particular, and the elongation product of adrenic acid (AdA) (29).

The formation of PL-PUFAs is mediated by the acyl-CoA long chain synthetase 4 (ACSL4), which catalyses the formation of acyl CoA derivatives (AA-CoA and AdA-CoA). Another enzyme, lysophosphatidylcholine acyltransferase 3 (LPCAT3), esterifies these derivatives into phosphatidylethanolamine (PE) forms AA-PE and AdA-PE that are inserted into the PL membrane. The resulting PL-PUFA(PE) produces L-ROS, which in turn executes ferroptosis (7).

PL-PUFA(PE) oxidation takes place in stages. In the first stage, *OH radicals attack PUFA on the bis-allylic position to create carbon-centred radicals that can react with molecular oxygen and produce L-OO*. In the second stage, L-OO* propagates by abstracting hydrogen from another PL molecule, which leads to the formation of PL-PUFA(PE)-OOH. L-OO* may also be added to the bis-allylic position of another PL molecule and produce a PL-OO-PL dimer. In the third stage, the reaction is continued until two radicals come together and form a non-radical molecule or the chain reaction is interrupted by some lipophilic antioxidant. If weakly bound or if free iron (Fe²⁺, Fe³⁺) is present (42, 43), PL-PUFA(PE)-OOHs can undergo reductive cleavage producing a toxic lipid LO* radical. These lipid radicals can abstract protons from neighbouring PUFAs and start a new cycle of lipid oxidation and damage (44). In addition, secondary lipid peroxidation products can be generated by intramolecular rearrangement and cleavage of PL-PUFA(PE)-OOH such as malondialdehyde (MDA), 4-hydroxynonenal (4HNE), and oxygenated PLs (33).

Enzyme-mediated formation of PL-PUFA(PE)-OOH (35) involves lipid-peroxidising enzymes that contain mononuclear iron centres and can easily receive iron from poly-rC binding chaperone proteins (PCBPs). Their major substrates are AA and linoleic acid. The enzymes are classified according to their positional specificity for AA oxygenation to 5-, 8-, 12-, 15-LOX. They catalyse the introduction of molecular oxygen into PUFAs to produce metabolites of 15-hydroxyeicosatetraenoic acid and 13-hydroxyoctadecadienoic acid (45, 46). The contribution of enzymatically and non-enzymatically mediated lipid peroxidation to ferroptosis differs (41, 47).

If the resulting L-ROS forms are not successfully detoxified by GPX4, they accumulate in cell and organelle membranes, which results in ferroptosis and membrane disruption. Although the exact mechanism is not clear, it is assumed that “hydrophilic pores” formed on membranes change membrane permeability and lay grounds for the “osmotic catastrophe” (48). Another assumption is that the resulting secondary, lipophilic electrophiles (e.g. MDA) may act as downstream signalling molecules and undetected protein effectors (49, 50).

Current knowledge indicates that LOX enzymes and the non-enzymatic Fenton reaction contribute the most to lipid peroxidation and L-ROS accumulation in ferroptosis (51, 52). The Fenton reaction involves oxidation of the ferrous to ferric form of iron and electron transfer to H₂O₂ (Fe²⁺+H₂O₂→Fe³⁺+*OH+HO^-) producing a very reactive *OH radical (ferric iron from the reaction can be reduced to ferrous iron in the presence of O*₂^- in the Haber-Weiss reaction).

Iron is an essential transition metal for all life forms on Earth. It is necessary for erythrocytopoiesis, for numerous enzymes involved in DNA replication, translation, and repair, for antimicrobial oxidative burst, and for many other biological processes, most often in the form of Fe-S-clusters (53). Thanks to its property to reversibly lose or receive electrons and transit from one valence state to another, it catalyses various biochemical reactions. The presence of free intracellular iron can therefore strongly influence cellular redox status and contribute to oxidative stress in cells. At the cellular level, iron homeostasis is regulated by a post-transcriptional mechanism mediated by the iron responsive protein/iron responsive element (IRP-IRE) system, while at the systemic level (absorption, utilisation, storage, and recycling) it is regulated by the hepcidin-ferroportin feedback loop and numerous other proteins such as divalent metal transporter (DMT1), ferrireductase, haem-oxygenase-1 (HMOX-1), ferroportin, transferrin (Tf), transferrin receptor-1 (TfR1), mitoferrin 1, and frataxin (54, 55, 56).

Under normal conditions, iron is delivered to the cell by Tf and TfR1.There ferric iron is reduced to its ferrous form, which is then introduced into the cytosol by DMT1. If there is excess free, non-transferrin-bound iron (NTBI), it is imported by transmembrane transporter proteins ZIP8 and ZIP 14 (57). Within the cells, iron binds to multifunctional PCBPs, which prevent iron from becoming part of the redox-active intracellular “labile iron pool” (LIP) and eventually its cytotoxic effects (57). In erythroid cells, most of the iron from LIP is transported into mitochondria by mitoferrin 1 and 2 and is stored as mitoferritin or incorporated into haem and Fe-S clusters, required for all electron transport chain complexes (57). In non-erythroid cells, iron from the LIP or the one bound to PCBPs is stored in the form of ferritin. When the stored iron is mobilised, ferritin binds to the autophagic nuclear receptor coactivator 4 (NCOA4). NCOA4 transports it to lysosomes, where it is degraded and released in a process called ferritinophagy (25). Ferritinophagy plays a major role in the recycling of intracellular redox-active iron, which also emphasizes the importance of lysosomes for ferroptosis (25). The resulting radicals and lipid peroxidation end products are highly reactive and cause massive oxidative damage, which is not exactly consistent with the regulated nature of ferroptosis.

Iron-dependent oxidative metabolism has been an indispensable part of life for billions of years. In view of this well-known fact, it was interesting to see that one research from 2010 (57), that is, before ferroptosis was recognised as a process, associated it with cell death. Degenerative changes in many diseases have been found to coincide with dysregulation of iron metabolism (57). Further research will show how consistent these findings are with the demonstrated mechanisms of cell death by ferroptosis in vivo and whether ferroptosis is the oldest and in many contexts the most important form of RCD.

For now, it can be concluded that lipid peroxidation mediated by free redox active iron plays a central role in ferroptosis (18). However, the precise role of iron in this process is yet to be determined. What we know is that iron catalyses the formation of L-ROS through the Fenton reaction and/or by iron-dependent LOX (58). The second assumption involves iron-independent redox activity, which needs further investigation (33, 59).

Little is known about the role of ferroptosis in the development of tissues and organs. Studies have shown that L-ROS levels are increased in embryonic tissues undergoing cell death and that the process can be controlled with GPX4 and lipophilic antioxidants (19, 61). This suggests that ferroptosis is important for maintaining tissue integrity and general homeostasis, although its precise developmental role is still unclear.

Impaired ferroptosis is associated with various pathological conditions such as malignancies, neurodegenerative diseases, ischaemia/reperfusion, acute kidney disease, liver and heart diseases, haemochromatosis, (62, 63), and neuropsychiatric illnesses (such as bipolar disorders, depression, schizophrenia, and Huntington’s disease) (64, 65). Many of these conditions are also accompanied by iron, glutamate, or GSH imbalance and increased lipid peroxidation (66).

In vitro and animal studies investigating the role of ferroptosis in tumourigenesis indicate that cancer cell lines derived from the brain, ovary, kidney, bone tissue, and soft tissue are susceptible to ferroptosis, whereas the lines derived from the pancreas, breast, stomach, and upper respiratory system are not (30, 67). These differences in sensitivity to ferroptosis have been attributed to differences in the basal metabolic state of these particular cell types, especially in lipid metabolism. Ferroptosis in cancer cells seems to be promoted by tumour suppressor gene p53, NADPH-oxidase (NOX), HMOX-1 and inhibited by miR-137 and transcriptional factors Nrf2 and p53 (10, 68).

Either way, they directly or indirectly target iron metabolism or lipid peroxidation and show a potential for genetic or pharmacological interventions utilising ferroptosis to eliminate malignant cells and treat different types of cancer. However, it has yet to be established in which cancer types ferroptotic therapy would be effective.

Malignant cells usually have high concentrations of iron and are consequently in persistent oxidative stress (69). This is why some cancers contain somatic mutations in the Nrf2/Keap 1 pathway to enhance transcription of antioxidant enzymes (19, 70, 71). As a consequence, ferroptosis is not frequent in the development of cancers. To stimulate ferroptosis as a therapeutic strategy against cancer, the most common therapeutic targets are the system Xc^-, GPX4, iron-related genes, GSH, coenzyme Q10 (CoQ10), and LOXs (67). Tumour suppressor gene p53 (more specifically, acetylation-defective mutant p53^3KR) is also known to induce ferroptosis (19) by suppressing the antiporter system Xc^-component SLC7A11. However, the roles of wild and mutant p53 in ferroptosis appear to be extremely complex. Depending on the context, different cells and different test conditions can stimulate or suppress ferroptosis by different mechanisms (72, 73, 74, 75). These studies are ongoing.

The mechanisms controlling ferroptosis have also been studied as therapeutic targets in various pathological conditions. For ferroptosis-based cancer therapy, iron-based nanomaterials have been designed and synthesised in recent times, such as ferumoxytol (otherwise approved by the United States Food and Drug Administration for iron deficiency therapy) and amorphous iron nanoparticles. Due to certain disadvantages of iron-based nanomaterials, other metals with multiple oxidation states (e.g. manganese dioxide-coated mesoporous silica nanoparticles) and GPX4-inhibiting nanomaterials (e.g. metal-organic network coated on the surface of polyethylenimine/p53 plasmid complex) are also being tested (74).

Chemotherapeutics (e.g. doxorubicin/adriamycin, cytarabine/ara-C, cisplatin) in combination with erastin show a significantly synergistic antitumor activity (75). Other drugs with the ability to stimulate ferroptosis, such as sulfasalazine, cisplatin, artesunate, sorafenib, lapatinib, salinomycin, and ironomycin are still being tested on cell lines of various cancers (10).

The role of ferroptosis in the treatment of cancer, however, is not yet sufficiently clear. According to Krysko et al. (76), tumour cell death by ferroptosis is a “double-edged sword”. Namely, ferroptotic cells can activate antitumour immune response through immunogenic cell death and thus enhance the effects of anticancer therapy. However, they can also suppress antitumor immune response and contribute to tumour progression. At this stage, no definitive conclusions can be drawn as to whether ferroptosis is an immunogenic or immunosuppressive form of cell death. This is why some authors call for further basic research on animal models (77).

To confirm the findings of ferroptosis studies, those in vivo in particular, it is necessary to define reliable molecular biomarkers of this process. Studies conducted so far are mainly based on the use of different inducers and on monitoring the effects of ferroptosis inhibitors and increased L-ROS values (19). A reliable and specific biomarker has not yet been defined. Candidates include increased expression of prostaglandin E synthase mRNA 2 (PTGS2), glutathione-specific gamma-glutamylcyclotransferase 1, and HMOX 1, which have been reported following erastin-induced ferroptosis (30, 78). Other potential ferroptosis biomarkers could be higher activity of cyclooxygenase 2 (30, 31), higher MDA, and lower NADPH levels (79, 80, 81). Various methodological approaches have been used in vitro to determine cell viability, most commonly flow cytometry (30, 82), followed by the measurement of intracellular iron using a specific dye and GSH depletion (83).