The Role of Ferroptosis in Blood–Brain Barrier Injury

Iron Metabolism

Iron plays an important role in a variety of organisms because it is a transition metal that can easily provide and receive electrons in REDOX reactions (Martin and Russell 2003). For humans and animals, iron is incorporated into proteins as a component of heme, iron–sulfur clusters or other functional groups, which are necessary for biological functions, including mitochondrial respiration, intermediate metabolism, host defense and cell signaling, etc. (Evstatiev and Gasche 2012). The Iron has two oxidation states: Fe²⁺or Fe³⁺. Fe³⁺ binds to transferrin (TF) in serum forms a complex, which is recognized by transferrin receptor 1 (TFR1) on the cell membrane and is transported via endosomes into the cell through endocytosis. Fe³⁺ in the endosome is converted to Fe²⁺ by six-transmembrane epithelial antigen of the prostate (STEAP3), and the free Fe²⁺ can be transported from the endosomes to cytoplasm by the divalent metal transporter 1 (DMT1). In the cytoplasm, Fe²⁺ enters a metabolically active pool called the labile iron pool (LIP). Ferroportin1 (FPN1) is currently the only known iron export protein in mammalian cells. While transporting Fe²⁺ out of the cell, the iron oxidase on FPN1 (ceruloplasmin, ferrous oxidation, enzymes and membrane iron transport auxiliary proteins) will oxidize Fe²⁺ to Fe³⁺, and Fe³⁺ transported out of the cell will immediately combine with TF to participate in the next round of circulation.

During physiological metabolism, excessive iron is stored in cells in the form of ferritin, which is composed of ferritin heavy chain (FTH1) and ferritin light chain (FTL), the former has iron oxidase activity and can catalyze the oxidation of Fe²⁺ to Fe³⁺, providing stable incorporation of iron into the ferritin shell, thus maintaining the equilibrium of the LIP (Arosio et al. 2015). Ferritin allows iron to be stored in an inert form that limits the production of damaging REDOX-active species. Dysregulation of iron content or distribution can lead to the accumulation of iron in the intracellular region, which can lead to damage to cells, tissues and organs. TF and TFR1 are now known to be regulators of ferroptosis, and in RAS-mutant cells, increased TFR1 expression leads to excessive iron uptake, and when TFR1 is knocked out, cells are resistant to Erastin-induced ferroptosis (Dixon et al. 2012). This suggests that excessive iron intake from outside makes cells more sensitive to ferroptosis. The expression of iron regulatory protein 2 (IRP2), a major regulator of iron metabolism, significantly increase the expression of FTL and FTH1, thereby inhibiting ferroptosis induced by Erastin (Yang and Stockwell 2008). FPN1 inhibits tumor growth by limiting iron transport (Kagan et al. 2017), and upregulation of hepcidin antimicrobial peptide (HAMP) may accelerate tumor growth by promoting intracellular iron overload caused by degradation of FPN1 (Bogdan et al. 2016). Depletion of Ceruloplasmin (Cp) leads to intracellular accumulation of Fe²⁺ and ROS, which promotes ferroptosis induced by Erastin and RSL3 (Shang et al. 2020). Iron can bind to heme and non-heme and many enzymes (such as lipoxygenase). Iron can promote lipid peroxidation through Fenton reaction and some enzymatic reactions (Dixon and Stockwell 2014). But either way, iron chelation in the cellular environment is an effective way to prevent lipid peroxidation and associated ferroptosis.

Lipid Metabolism

Lipids can be classified into many types, such as fatty acids, glycerides, or glycerophospholipids. Lipid peroxidation is the main characteristic of ferroptosis that is primarily induced by PUFAs under the action of lipoxygenases and ROS and eventually activates ferroptosis. Fatty acids are divided into saturated fatty acids and unsaturated fatty acids, of which PUFAs have more than one carbon double bond, which makes them vulnerable to free radical attacks, thereby inducing lipid peroxidation. One of the characteristics of ferroptosis is the oxidation of PUFAs. First, Acyl-CoA synthetase long-chain family member 4 (ACSL4), which is the catalytic enzyme for the first step, firstly catalyzes free arachidonic acid (AA)/adrenic acid (AdA) to bind CoA to form AA/AdA–CoA derivatives, facilitating their esterification into phospholipids. Next, lysophosphatidylcholine acyltransferase 3 (LPCAT3) esterifies these derivatives to phosphatidylethanolamine (PE) forms AA-PE and AdA-PE that are inserted into the PL membrane. Subsequently, 15-lipoxygenase (15-LOX) directly oxidizes AA-PE and AdA-PE into lipid hydroperoxides, which subsequently activates ferroptotic signals (Fig. 1) (Doll et al. 2019). Adenosine 5′-monophosphate (AMP)-activated protein kinase (AMPK) can reduce the synthesis of PUFAs, thereby inhibiting ferroptosis (Lee et al. 2020). Conversely, when the activity of PUFAs synthase is enhanced, mesenchymal cells are sensitive to ferroptosis (Viswanathan et al. 2017). These results suggest that regulating PUFAs levels may be an effective mechanism to regulate ferroptosis. In addition, other studies have shown that inhibition of ACSL4 and LPCAT3 significantly inhibits ferroptosis (Shindou and Shimizu 2009). A small scaffold protein that promotes 15-LOX catalysis of AA-PE has been reported to regulate ferroptosis in epithelial and neuronal cells (Wenzel et al. 2017).

The Cyst(e)ine/GSH/GPX4 Axis

For many years, the Cyst(e)ine/GSH/GPX4 system was thought to be the only pathway to inhibit ferroptosis. SLC7A11 and solute carrier family 3, member 2 (SLC3A2) on the cell membrane constitute system XC-, which imports the cystine into cells with a 1:1 counter-transport of glutamate. Cystine is rapidly converted intracellularly to cysteine, the rate-limiting amino acid for the synthesis of GSH. The imbalance of the antioxidant defense system dependent on system XC- leads to oxidative damage and ultimately cell death. GPX4 is a unique intracellular antioxidant enzyme that directly reduces peroxidized phospholipids that have been produced in cell membrane (Imai and Nakagawa 2003). GPX4 is a selenoprotein whose biosynthesis is dependent on the co-translation and incorporation of selenocysteine (Sec), and the expression or activity of GPX4 is regulated by selenium and GSH (Friedmann Angeli and Conrad 2018). As GPX4 is synthesized, the nascent polypeptide chain incorporates selenium as selenocysteine (Sec), where selenium replaces the sulfur of cysteine, when a UGA stop codon is “recoded” by a Sec-tRNA and a selenocysteine insertion sequence (SECIS) within the GPX4 mRNA (Tang et al. 2021). Selenium can increase the anti-ferroptosis activity of GPX4 through a selenocysteine residue at 46 (U46) (Ingold et al. 2018). At the transcriptional level, selenium-induced transcription factor AP-2 gamma (TFAP2C) and specific protein 1(SP1) upregulated GPX4 expression to prevent ferroptosis-related intracerebral hemorrhage (Alim et al. 2019). GSH homeostasis is linked to a large network of enzymes; GSH peroxidases (GPXs) decompose both H₂O₂ and lipid peroxides in GSH-requiring reactions that oxidize GSH to GSSG, which in turn is reduced by GSH reductase (Yang and Stockwell 2008; Lewerenz et al. 2013; Rahman and Kinnula 2012). Pharmacological inhibition or gene knockout of GPX4 remains the classic method of inducing ferroptosis (Seibt et al. 2019). The direct effect of GSH levels on ferroptosis emerged from the use of Erastin (one of the ferroptotic activators), which by lowering GSH levels activates cell death identically to that caused by a lack of GPX4 (Pierzynowska et al. 2021).

The NAD(P)H/FSP1/CoQ10 System

Induction of ferroptosis by inhibition of GPX4 has emerged as a therapeutic strategy to trigger cancer cell death. GPX4 is thought to be a major regulator of ferroptosis inhibition, but inhibition of GPX4 does not inhibit ferroptosis in some cell lines, suggesting that there are other mechanisms to inhibit ferroptosis. To uncover a potential pathway that inhibits ferroptosis independent of the GPX4 pathway, Bersuker et al., employing a synthetic lethal CRISPR/Cas9 screen in osteosarcoma cells, identified ferroptosis suppressor protein 1 (FSP1) (previously known as apoptosis-inducing factor mitochondrial 2 (AIFM2)) as a novel ferroptosis resistance factor (Bersuker et al. 2019). FSP1, a traditional inducer of apoptosis in mitochondria, has recently been identified as an antioxidant regulator of ferroptosis. FSP1 is translocated from mitochondria to the cell membrane, where NADPH is used to catalyze regeneration of non-mitochondrial reductive coenzyme Q10 (CoQ10), thereby inhibiting lipid peroxidation in a GPX4-independent manner (Bersuker et al. 2019; Doll et al. 2019). CoQ10, also known as ubiquinone, is an endogenous isoprenoyl benzoquinone compound ubiquitous in nature and an effective lipophilic antioxidant. FSP1 is recruited to the plasma membrane by myristoylation, where it functions as an oxidoreductase that catalyzes the CoQ10 using NAD(P)H. Ubiquinol, the reduced form of CoQ10, is a lipophilic radical-trapping antioxidant (RTA), which regulates ferroptosis by halting the propagation of lipid peroxides (Bersuker et al. 2019, 2018; Doll et al. 2019; Elguindy and Nakamaru-Ogiso 2015; Yan et al. 2020). In mechanism, CoQ10 is reduced by FSP1 as a lipophile RTA, which can directly reduce lipid free radicals or indirectly be recycled by alpha-tocopherol (α-TOH), which is the most important lipophile antioxidant in nature. In addition to mediating the mechanism of CoQ10, FSP1 inhibits ferroptosis by activating the endosomal sorting complex required for transport (ESCRT-III) in the plasma membrane, supporting the anti-injury effect of membrane FSP1 (Dai et al. 2020). However, the underlying mechanisms involved are unclear, with key downstream targets of FSP1 remaining to be identified in future. Although FSP1 was originally identified as a p53 response gene, recent studies have shown that FSP1 expression is insensitive to p53 activators Nutlin-3 and doxorubicin (Bersuker et al. 2019; Doll et al. 2019), suggesting that p53 activation does not necessarily induce FSP1 upregulation, and that some other response may be required to fully initiate the process. FSP1 is negatively regulated by the long non-coding RNA maternal expression gene 3 (MEG3)-miR-214 axis, which has been shown to promote ferroptosis by targeting and decreasing activation of transcription factor 4 (ATF4) levels (Cho et al. 2019; Bai et al. 2020).

The GCH1/BH4/DHFR System

Tetrahydrobiopterin (BH4) is a redox-active cofactor involved in the production of nitric oxide, neurotransmitters, and aromatic amino acids (Kraft et al. 2020; Cronin et al. 2018). The first step in the conversion of GTP to BH4 is the conversion of GTP to 7,8-Dihydroneopterin triphosphate (NH2TP) by GTP cyclohydrolase-1(GCH1), followed by the conversion of NH2TP to BH4 by two other enzymes, GCH1-6-pyruvoyltetrahydropterin synthase (PTS) and sepiapterin reductase (SPR), with GCH1 being the rate-limiting enzyme for BH4 synthesis (Levine et al. 1990; Kojima et al. 1995). A recent functional screening has shown that activation of GCH1 has a protective effect during GPX4 inhibition, and that GCH1 expression is a determinant of the availability of BH4, a potent cofactor of REDOX activity that protects lipids from peroxidation and regeneration by dihydrofolate reductase (DHFR). Studies have shown that the genetic or pharmacological blockade of DHFR and GPX4 inhibition had a synergistic effect on the promotion of ferroptosis (Lei et al. 2020; Wang et al. 2019). BH4 converts phenylalanine to tyrosine, which in turn is converted to 4-hydroxybenzoate (a precursor of CoQ10) to promote the synthesis of CoQ10, thereby inhibiting ferroptosis. Loss of BH4 induces the accumulation of CoA, GSSG, and NADP (Soula et al. 2020), while an increase in reduced CoQ10 was detected in GCH1-overexpressed cells. These results suggest that the GCH1-BH4-phospholipid axis acts as a master regulator of ferroptosis resistance, controlling endogenous production of the antioxidant BH4, abundance of CoQ10, and peroxidation of unusual phospholipids with two polyunsaturated fatty acyl tails. This demonstrates a unique mechanism of ferroptosis protection that is independent of the GPX4/GSH system.

DHODH Inhibits Ferroptosis

The rate-limiting enzyme in the pyrimidine nucleotide synthesis pathway, Dihydroorotate dehydrogenase (DHODH), is located on the inner mitochondrial membrane and catalyzes the dehydrogenation of dihydroorotate (DHO) to orotic acid, ultimately producing uridine and cytidine triphosphate (UTP and CTP). Recent studies have shown that treatment of cancer cells with GPX4 inhibitors results in acute depletion of N-carbamoyl-l-aspartate acid, an intermediate in pyrimidine biosynthesis, accompanied by the accumulation of uridine. Supplementation of DHO or DHODH substrates and products attenuated or enhanced ferroptosis induced by inhibition of GPX4, respectively, especially in cancer cells with low expression of GPX4 (Fig. 1). However, the role of DHODH in ferroptosis in the BBB is not clear.

Role of Cyst(e)ine/GSH/GPX4 Axis in BBB

GSH is a tripeptide composed of glutamate, cysteine and glycine. Cysteine is a rate-limiting precursor to GSH synthesis because of its limited intracellular concentrations. The XC- system delivers cysteine in its oxidized form into the cell, where it is immediately converted to reduced form cysteine. GSH plays a key role in protecting cells from oxidative stress and xenobiotics, as well as maintaining the thiol redox state, especially in the CNS (Dringen 2000). Aging-related decreases in GSH content have been reported in many tissues, while levels in the brain are lower than in other organs. Compared to other organs, the brain has low levels of antioxidant enzymes, such as SOD, catalase, and GPX, which makes the brain particularly susceptible to oxidative stress and eventually leads to cell death (Dringen 2000). The concentration and synthesis of GSH are highly regulated in the CNS and are limited by the availability of the sulfhydryl amino acid l-cysteine (l-cys), which is primarily transported from the blood to neurons via BBB. Basal GSH levels were different in BBB endothelial cells as well as neurons, astrocytes, microglia and oligodendrocytes from different brain structures, especially in astrocytes (AC), where the XC- system exchanger uptake of l-cys2 into the cells with 1: 1 to release the excitatory amino acid l-Glu (Huang et al. 2020b; Valdovinos-Flores and Gonsebatt 2012). Extracellular l-cys is oxidized to l-cystine and transported through the system into AC, which release GSH via the multidrug resistance protein 1 (MRP1) as the first step in supplying cysteine for GSH synthesis in neurons. XC- system is related to oxidative protection, BBB physiology, brain tumor growth, etc. (Augustin et al. 2007).

BBB injury significantly accelerated the progression of brain diseases and increasing the level of endothelial GSH could prevent BBB damage (Fig. 1). Compared to primary endothelial cells, AC contained relatively high levels of GSH-related metabolites, and external damage significantly disrupted REDOX homeostasis in BMECs, but not AC (Huang et al. 2020a). Combining isotope labeling/tracer and time-of-flight mass spectrometry, it was demonstrated that AC continuously transported GSH to BMECs even under resting conditions, and the transport was accelerated in the presence of in vitro damage. siRNA-mediated gamma-glutamyl cysteine ligase (GCL) knockdown affects the production and secretion of GSH in astrocytes and significantly compromises BMECs barrier integrity (Dringen 2000). Therefore, AC-GSH shuttle is critical for maintaining the redox balance of BMECs and the stability of BBB, indicating that the supplement of GSH improve recovery after brain injury (Huang et al. 2020a). Post-transcriptional modifications of tightly junction proteins dynamically regulate complex assembly and localization (Rao 2008). Particularly tyrosine phosphorylation (pTyr) of occludin and claudin-5 C terminus results in the attenuation of their interaction with ZO-1 and cytoskeleton, leading to increased BBB permeability (Kale et al. 2003; Shen et al. 2011). Exogenous GSH supplementation improves BBB function by preventing hypoxia-induced tyrosine phosphorylation (pTyr) of occludin, a modification that determines the localization of the BMECs tight junction protein, complex stability, and regulatory protein integration, maintaining the integrity of the BBB (Wong 1997; Yamamoto et al. 2008; Engelhardt et al. 2014). The protective effect of GSH may be mediated by two different mechanisms: (1) Inhibition of ROS imbalance to prevent phosphorylation; (2) Non-ROS-mediated pathways; Protein tyrosine phosphatase (PTP) may be directly activated by GSH (Yasothornsrikul et al. 1999; Filomeni et al. 2002).

Role of NAD(P)H/FSP1/CoQ10 System in BBB

Mitochondria use oxygen to efficiently produce energy. The excessively stimulated NADPH and electron transport chain in the mitochondria lead to the overproduction of ROS, which is considered to be one of the main sources of oxidative stress (Dröge 2002). Although moderate concentrations of ROS/RNS play important roles in physiological processes (such as signaling pathways, induction of mitotic responses, and defense against infectious pathogens), its overproduction and imbalance of endogenous antioxidant defense systems can lead to oxidative damage (Gan and Johnson 2014; Bhat et al. 2015), such as post-translational modification and oxidation of proteins and lipid oxidation. In general, the brain is a highly metabolically active organ, with less ability to regenerate than other organs. Therefore, the influence of active substances on brain is significant (Bhat et al. 2015).

Elevated acute or chronic PIC levels are the primary cause of increased BBB permeability and the neuroinflammatory cascade characteristic of neurodegenerative and neuroprogressive diseases (Capaldo and Nusrat 2009). Some of the mechanisms supporting this effect, such as some cytokines can inhibit ZO-1 transcription and activate nicotinamide adenine dinucleotide phosphate (NADPH) in BMECs, leading to BBB damage (Capaldo and Nusrat 2009). NO synthase and NADPH oxidase activate microglia. Then activated microglia release a series of neurotoxic molecules, such as ROS, NO, PGE, COX-2, quinolinic acid, chemokines monocyte chemoattractant protein 1 (MCP-1), C-X-C motif chemokine ligand-1 (CXCL-1), macrophage inflammatory protein-1α (MIP-1α), PICs, IL-6, TNF and IL-1β), the above factors all adversely affect the integrity of microglia (Morris and Maes 2014). ROS production by microglia increases BBB permeability through several different pathways, including upregulation of PI3K/Akt and JNK-c-Jun signaling pathways, activation of MMP-9, MMP-3 and MMP-2 leading to cytoskeleton remodeling, and downregulation of TJ proteins (Claudin-5, 1 and 11), ZO-1 and occludin at the transcriptional level (Schreibelt et al. 2007). Previous studies have shown that membrane lipid peroxidation of BMECs is another mechanism of increased BBB permeability mediated by ROS produced by activated microglia cells (Chodobski et al. 2011).

CoQ10 plays an important role in oxidative phosphorylation where it acts as an electron carrier in the MRC. Its major role is in accepting electrons derived from complex I and II (NADH ubiquinone reductase; succinate ubiquinone reductase) and transporting them to complex III (ubiquinol cytochrome c reductase) (Ernster and Dallner 1995). In addition, CoQ10 can also act as an antioxidant and an important part of the “respiratory system” (Acín-Pérez et al. 2008), which can prevent electron leakage and ROS production. CoQ10 acts as a superoxide scavenger in neuroglial mitochondria, promoting mitochondrial proliferation, improving lipid peroxidation in the inner mitochondrial membrane, activating uncoupling proteins, and having a positive effect on plasma membrane REDOX system. Several studies have revealed the anti-apoptotic, antioxidant and anti-inflammatory effects of CoQ10. For example, significant reductions in TNF and IL-6 levels following CoQ10 lead to reductions in oxidative stress and inflammation in cardiovascular disease (Salehpour et al. 2019). Intravenous injection of CoQ10 reduces apoptosis by regulating BAX-Bcl-2. CoQ10 reduced the degree of tissue damage and nervous system defects after ischemia in hyperglycemia rats (Lu et al. 2017). Studies have shown that CoQ10 reduces MMP-2, MMP-9 and inflammatory factors, and this effect of CoQ10 is related to the reduction of tissue damage, cerebral vascular permeability, nervous system defects and brain edema (Ghasemloo et al. 2021). In addition to activating mir-149-5p-related pathways and playing a positive role in improving ischemic complications, CoQ10 may also have a stronger protective effect by increasing antioxidant activity in the body. Mir-149-5p and CoQ10 can exert a protective effect on the brain by reducing MMP and inflammatory factors after cerebral ischemia (Ghasemloo et al. 2021).

Role of GCH1/BH4/DHFR System in BBB

BH4 plays a vital antioxidant role because it interacts with NOS to prevent superoxide formation. Oxidative stress oxidizes BH4 to dihydrobiopterin (BH2), leading to decoupling of the two enzymatic domains of NOS and the production of superoxide instead of NO (NOS uncoupling). So increased utilization of BH4 in the brain may result in better protection against oxidative damage. ROS induces the oxidation of BH4 to BH2, which reduces the level of BH4 in endothelial cells (Najjar et al. 2013). The subsequent decrease in the BH4-to-BH2 ratio leads to the inhibition of eNOS while uncoupling arginine as its substrate, thereby allowing it to contact environmental oxygen and increased production of superoxide ions (Najjar et al. 2013; Moens and Kass 2006). As mentioned above, O²⁻ in turn combines with NO to form ONOO–-, which promotes the oxidative conversion of BH4 to BH2, which further reduces the activity of eNOS, forming a positive feedback loop (Chen et al. 2010).

Conflict of interest

The authors report no conflict of interest.

Ethical Approval

Not applicable.

Consent to Participate

Not applicable.

Consent for Publication

Written informed consent for publication was obtained from all participants.