INTRODUCTION

In recent decades Alzheimer’s disease (AD) has recorded rapid growth due to the increase in life expectancy of the population on the planet. The highest rates of AD are observed in developed countries with progressive healthcare infrastructure. It is age-associated progressive neurodegeneration, characterized by progressive loss of cognitive abilities. Alzheimer’s disease is the most common cause of dementia [1]. Despite the complex research with the application of various pharmacological targets, still there is no effective cure for AD. Several mechanisms are involved in the pathogenesis of Alzheimer’s disease, common markers are considered extracellular amyloid-beta plaques and intracellular hyperphosphorylated tau-composed neurofibrillary tangles. There are two types of Alzheimer’s disease: late-onset AD (LOAD) or sporadic AD constituting the majority of cases and early-onset AD (EOAD) or familial AD (FAD). Late-onset AD affects people after the age of 65, while early-onset AD (EOAD) happens earlier and constitutes less than 5% of all cases [2].

Genetic predisposition also has significant weight in the disease pathogenesis. Apolipoprotein E (APOE) genotype is considered to be the main genetic risk factor [3]. Mutations in genes APP, PSEN1, and PSEN2 are known to be responsible for 10–15% of cases of EOAD [4].

Alzheimer’s disease is characterized by the accumulation of insoluble and protease-resistant amyloid plaques, which serve as a primary hallmark of the disease. Amyloid beta, derived from the integral membrane protein known as an amyloid precursor protein, undergoes enzymatic cleavage by beta- and gamma-secretases [5, 6]. Alternatively, physiological cleavage involves the participation of alpha- and gamma-secretases [7]. However, in Alzheimer’s disease, the incorrect cleavage and adhesive properties of amyloid beta subunits result in their aggregation and formation of aggregates. The accumulation of these protease-resistant and insoluble plaques disrupts neuronal metabolism and contributes to neurodegeneration [8, 9].

Another characteristic feature of AD is the presence of neurofibrillary tangles, which are formed from abnormally hyperphosphorylated tau protein, responsible for stabilizing axonal microtubules [1]. The pathological manifestation of tau pathology involves the intraneuronal buildup of neurofibrillary tangles, which disrupt normal metabolism and compromise the integrity of axons. This disruption leads to microtubule dissociation and loss of connectivity between neurons [10]. Furthermore, the disturbed metabolism creates an imbalance between the production of free radicals and the neutralization systems, resulting in oxidative stress. The excessive production of free radicals damages lipids, proteins, and DNA, ultimately leading to cellular death [11, 12]. Additionally, oxidative damage to mitochondria and impaired glucose metabolism negatively impact energy production, resulting in a state of decreased energy levels [13].

AD is characterized by chronic neuroinflammation, which involves persistent inflammation in the nervous tissue. Activated microglia release various pro-inflammatory molecules and generate free radicals [14]. Initiated in the hippocampus, neuroinflammation spreads to other brain regions, including the striatum, amygdala, and hypothalamus, causing cellular damage [15]. These combined factors contribute to synaptic dysfunction, resulting from disrupted intracellular metabolism, oxidative stress, axonal damage, and impaired neurotransmitter synthesis. Symptomatically, synaptic dysfunction is manifested as worsening memory and overall cognitive decline [16–18].

In addition to the mentioned metabolic impairments, AD is associated with the accumulation of metal ions and disturbances in their homeostasis. Imbalances in iron, calcium, zinc, and copper play a significant role in the pathogenesis of Alzheimer’s disease [19, 20].

Generally, AD is a complex pathology encompassing the aforementioned hallmarks, all of which contribute to cognitive impairment.

FERROPTOSIS AS AN ALTERNATIVE CELL DEATH MECHANISM

During evolution, many types of cell death have been developed in the organism. Physiological cell death is essential and does not contain any risk if it goes without an impediment [21, 22].

Parallel to their cell death as a physiological process exist other types of cell death apoptosis, necrosis, autophagy, and ferroptosis, and these processes are involved in the different and important processes in the organism. All these types of cell death have differences in the molecular mechanisms, signaling pathway, morphological features, genetics, and role. Ferroptosis is a form of programmed lytic cell death, which is described as iron accumulation as a result of dyshomeostasis of lipid peroxidation. Ferroptosis has similar features with necrosis, and apoptosis but not completely. During ferroptosis cells have small mitochondria with increased membrane density, reduced or absent crista, as well as rupture of the outer membrane, have normal nuclear size, and do not have changes in the chromatin [23, 24].

Mechanisms underlying ferroptosis stimulation are dysregulation of the xc system, which is normally limited lipid peroxidation, glutathione depletion, GPX4 inactivation, and AA depletion in the presence of serum and glucose. In this process we have positive regulators such as Vdac2/3, Ras, TFR1, NOX, p53, CARS, and negative regulators GPX4, System Xc–, HSPB1, NRF2 and all of these are potential targets for ferroptosis regulation [25, 26].

Despite being a recently discovered mechanism of cell death, they are involved in many dangerous and life-treating conditions: cancer, neurodegenerative disease (AD, PD), stroke, and TBI. Clinical trials for AD used iron chelator DFE, Vitamin E which shows amelioration of AD symptoms [27].

ALTERED IRON METABOLISM IN AD

Iron is one of the most abundant metals in the organism. It is involved in several metabolic processes, such as microsomal hydroxylation, energy production, enzyme activation (cofactor), oxygen transport, oxidation, antioxidant protection, production of free radicals, etc. Iron absorption, storage, and excretion are carefully regulated. Both iron deficiency and excess are dangerous. Iron is involved in several processes taking place in the brain. This includes but is not limited to the production of ATP, hydroxylation to produce neurotransmitters (dopamine, serotonin, noradrenaline, etc.), and antioxidant protection (catalase) [28]. Primarily, the iron deficit is dangerous for the brain because of insufficient oxygen transport, a hypo energetic state, and neuronal death. The excess of iron is also dangerous because of its neurotoxic effect and disturbances in physiological metabolism. Age is among the factors, contributing to increased levels of iron in the brain; this pattern has also been demonstrated in AD patients. High levels of iron were found in the putamen, globus pallidus, substantial nigra, and cortical regions [29, 30]. The reason for the regional specificity of iron accumulation in the brain remains unknown as there is currently no available data on this matter. Aging is considered one of the major contributing factors to iron buildup in the brain. Additionally, aging is associated with neuroinflammation, characterized by a significant increase in astrocyte and microglia activity [31]. During the aging process, ferritin levels were observed to increase in the hippocampus, basal ganglia, cerebellum, cortex, and amygdala. Physiologically, oligodendrocytes contain the highest amount of iron, and this does not change with age. However, with age, microglia that express ferritin display altered morphology and actively phagocytose iron. This iron uptake by microglia leads to oxidative damage and degeneration, which are underlying factors in neurodegeneration [32].

Ferroptosis is considered one of the mechanisms of cell death in AD patients. The underlying cause of ferroptosis is an accumulation of extra amounts of iron. The cause has to be watched in the central iron homeostasis systems. Brain iron intake includes transferrin-mediated transport into endothelial cells of the blood-brain barrier. Astrocytes directly take iron from BBB through DMT1 and release it through ferroportin and ceruloplasmin. Further transport into extracellular space is not clear but is supposed to happen via ferroportin or DMT1. Oligodendrocytes take large amounts of iron, which is mostly used for axon myelination. Microglia activation during neuroinflammation in AD patients disturbs iron metabolism, causing iron accumulation in neurons and microglia. This is due to changes in two iron transporters: an increase in DMT1 and reduced ferroportin in neurons and increased DMT1 and unchanged ferroportin levels in microglia. Astrocytes did not show significant changes in both transporter levels. Microglia and astrocytes have shown increased levels of hepcidin, a hormone responsible for the suppression of ferroportin. Thus, iron accumulation was observed only in neurons and microglia [33]. It causes oxidative stress and the production of aggressive forms of oxygen as a result of the Fenton reaction and the product of the reaction, ⚫OH radical damages intracellular structures, including DNA, membranes, and proteins. The disturbed metabolism creates fertile ground for the release of iron from iron-sulfur proteins and the production of free radicals. This damages mitochondria, contributing to the development of hypo energetic state and further neurodegeneration. Lipid peroxidation of the membranes produces reactive aldehydes, which irreversibly bind to the proteins, modifying native conformation.

Moreover, a large amount of iron is used in microsomal hydroxylation for the production of neurotransmitters, such as dopamine, noradrenaline, etc. Availability of reactive oxygen species oxidizes catechol ring into toxic quinones via iron reduction or enzymatically [34]. The death of neurons through ferroptosis is reached because of an imbalance between extra production of aggressive forms of oxygen and antioxidant system deficiency, particularly, glutathione-peroxidase 4 (GPX4) [35].

Alzheimer’s disease is a complex pathology including different level impairments in proteolysis, axonal transport, energy production systems, free radical production, and neutralization systems, neuronal dysfunction, and altered synaptic transmission.

INTERACTION BETWEEN AMYLOID-BETA ACCUMULATION, TAU PATHOLOGY, AND FERROPTOSIS-RELATED MECHANISMS

Iron is a key player in several physiological processes taking place in the nervous tissue and periphery. Increased levels of iron highly contribute to the formation of free radicals through the Fenton reaction causing oxidative stress and damaging DNA, membranes, and other neuronal structures, as well as facilitating the activity of pro-oxidant enzymes. Increased iron levels were found in Alzheimer’s patient’s postmortem brains, hence it is still poorly understood how iron contributes to the amyloid-beta plaque aggregation. Some authors associate the primary stage of iron-induced oxidative stress with the malfunction of ferritin [36], the storage form of iron. It is worth mentioning that ferritin stores a “safe” form of the iron—Fe3+, which lacks an active electron capable of producing free radicals. Another hypothesis suggests that Amyloid-beta can interact with free iron forms. In vitro studies show that Amyloid beta has the potential to reduce Fe3+ to Fe2+, thus activating it. Fe2+ was found in the composition of senile plaques obtained from AD patients [37]. Moreover Amyloid-beta binds Fe3+ forms of iron reducing them into Fe2+.

Elevated levels of iron contribute to the formation of protein aggregates, such as hyperphosphorylated tau. Iron increase-induced may lead to apoptosis [38] or ferroptosis [39].

Accumulation of metals, such as iron, copper, zinc, etc. was found in the insoluble amyloid plaques and neurofibrillary tangles. These contribute not only to the misfolding of proteins but also increase the “gluing” properties of misfolded proteins.

It was found that for the production of tau, the activation of hem oxygenase has a potentially significant role in metabolizing hem released from damaged mitochondria. But during the catabolism of the heme ferrous iron is also released, which may induce the production of free radicals. Hem oxygenase activation contributes to the overproduction of iron and increases the risk of cognitive decline.

Iron is an inductor for furin, which reduces the activation of alpha-secretase, which precedes gamma-secretase in cutting the APP. Reduced activity of alpha-secretase leads to activation of beta-secretase and contributes to the amyloid plaque production [40]. Reversely, iron deficiency inhibits plaque formation [41]. Not just modulation of furin is crucial for the amyloidogenic pathway, but also modulation of the APP synthesis and its membrane allocation has a key role in amyloid production (Fig. 1).

Fig. 1.
figure 1

The role of ferroptosis in development of neuronal damage at AD. Iron induced cell death via activation of oxidative stress is demonstrated in the figure. Iron transportation and storage in neuronal cell; over accumulation of iron; generation of reactive oxygen species (ROS) via Fenton reaction; ROS initiated peroxidation of membrane phospholipids; damage of cellular organic compounds by end products of oxidative process; GSH–GPX4–NADPH ferroptosis controller system; cysteine uptake via system X– and production from methionine; iron induced activation of furin; modulation of alfa- and beta-secretases by furin; beta amyloid fragments aggregation.

Full size image

LIPID PEROXIDATION AND ITS ROLE IN FERROPTOSIS

As was already described above, the main pathological pathway of ferroptosis can be characterized by the accumulation of iron with consequent generation of ROS and L-ROS (lipid reactive oxygen species). The process of ferroptosis in the brain deserves special attention due to its metabolic features. High metabolic activity of the brain requires a high level of oxygen consumption, in addition, the cell membranes of the brain are rich in polyunsaturated fatty acids (PUFA). A higher number of double bonds in PUFA increases the sensitivity to oxidation. Several in vitro studies showed that monounsaturated fatty acids block membrane lipid ROS accumulation and inhibit ferroptosis [42]. Moreover, in comparison to other tissues in the brain the level of redox ions, such as iron, is higher and the activity of the antioxidant system is lower [43, 44]. All these facts make the brain vulnerable to oxidative stress. Iron triggers ROS formation through the Fenton reaction, in which hydroxyl radicals are generated and enhance the oxidation of membrane lipids. Membrane PUFAs are the primary target of ROS with two dangerous consequences of lipid peroxidation: structural modification of the membrane and generation of primary products. The main primary products of lipid peroxidation are lipid hydroperoxides. Among the secondary products, the most abundant products are malondialdehyde (MDA), acrolein, 4-hydroxy-2-hexenal (HHE), and 4-hydroxy-nonenal (HNE). Several studies reported the toxic effect of MDA and 4-HNE on cell metabolism by interaction with nucleic acids, proteins, and other nucleophilic compounds. Furthermore, end products may initiate ROS signaling and activate the mitochondrial caspase pathway and amplify the mechanisms of cell death. Peroxidation of membrane lipids leads to the structural and chemical transformations of a phospholipid bilayer, formation of the pores, and alteration of the barrier function.

Interestingly, the main hallmarks of ferroptosis, such as iron accumulation and lipid peroxidation, have been observed in Alzheimer’s disease pathology. The involvement of iron in AD pathology has been proven by the discovery of iron and ferritin around the glial cell in senile plaques and neurofibrillary tangles [45–48]. Advanced methods showed that iron accumulation in the brain is associated not only with neurodegeneration but also with the aging process in general. An increasing number of studies showed the accumulation of lipid peroxidation markers in brain-specific regions, cerebrospinal fluid, blood, and urine of AD patients. Post-mortem studies reported increased amounts of thiobarbiturate reactive substances, HNE, and acrolein in diseased regions of the AD brain in comparison to controls [48, 49]. Moreover, proteins modified by end products of lipid peroxidation were detected in diseased regions of AD brain, and not detected in regions uninvolved in AD and in brains of age-matched control individuals [50]. A few studies measured the concentration of HNE, F2-IsoPs, and F4-NeuroPs in cerebrospinal fluid received from the AD patient and received a significantly increased level compared to controls. Additionally, the F2-IsoPs level correlated with global brain mass, and degree of cortical atrophy, but not the density of neuritic plaques or neurofibrillary tangles [51]. Furthermore, urinary and plasma levels of detected compounds were significantly correlated with cerebrospinal fluid levels, whereas only CSF levels were partially correlated with cognitive decline [52]. That indicates that the periphery reflects the central nervous system and these compounds may be considered biomarkers for AD diagnosis.

ANTIOXIDANT SYSTEM FUNCTIONING IN FERROPTOSIS

In line with iron accumulation, the other inducer of ferroptosis is the inhibited antioxidant system. There are two main pathways involved in this process: the first one integrates cystine-glutamate transporter (system \({\text{X}}_{{\text{c}}}^{ - }\)), and the second one directly inhibits glutathione peroxidase activity [35, 53]. System \({\text{X}}_{{\text{c}}}^{ - }\) is an antiporter located in the cell membrane, it transfers glutamate out of the cell and cystine into the cell. Immediately after the transportation cystine converts to cysteine and is used for the biosynthesis of glutathione (GSH) [54]. Glutathione is an essential coenzyme of glutathione peroxidase (GPX), which is an important antioxidant and scavenger of free radicals. Among all three amino acids, cysteine is the most essential one for GSH production and its deficiency may trigger ferroptosis in vitro [23]. Under oxidative stress conditions, an additional source of cysteine is activated, which is the transsulfuration of serine by sulfur of methionine. In GPX-catalyzed reaction, the role of GSH is donating electrons to lipid hydroperoxide and neutralization its toxic effect. During the reaction glutathione itself oxidized with the formation of GSSG. Glutathione reductase is required for regeneration of oxidized GSSG, NADPH acts as an electron donor in this reaction. In vitro, studies showed a linear correlation between decreasing GSH and neurodegeneration in cell culture obtained from AD triple transgenic mice [55]. Furthermore, GPX mutation in neurons causes neurodegeneration in mice with defined ferroptosis features. It was shown that knockout of GPX leads to cognitive decline in mice and neurodegeneration which can be reduced by ferroptosis inhibitors [56]. To compensate for the loss of GPX, another oxidoreductase (ferroptosis–suppressor–protein1, FSP1) reduces ubiquinone and accelerates the scavenging of radicals [57]. Ubiquinone is a strong antioxidant widely distributed in cell membranes, the reduced form of ubiquinone traps lipid radicals and inhibits lipid peroxidation. However, the role of FSP1 in cell death remains unclear. There are few studies, indicating that the anti-ferroptotic function of FSP1 is independent of GPX and even glutathione levels. That may suggest that FSP1 does not interfere with canonical ferroptosis mechanisms, and at the same time, it does not protect cells against other proapoptotic factors. Thus, the FSP1–CoQ–NADPH and GPX–GSH–NADPH are the main parallel systems that inhibit ferroptosis by modulating oxidative stress and the generation of L‑ROS.

THE FERROPTOSIS THERAPEUTIC TARGET FOR ALZHEIMER’S DISEASE

All the above-mentioned support that iron metabolism alteration and related ferroptosis can be considered as a therapeutic target for AD treatment. Anti-ferroptosis and antioxidant factors are the potential candidates and are discussed below. All of them have a set of benefits and drawbacks.

Iron chelation therapy is widely used in medicine to balance the amount of iron accumulated in different tissues. There are two iron chelators currently available and approved by the FDA: deferoxamine and deferasirox. The main features of these drugs are bioavailability and blood–brain barrier permeability, as well as affinity to intracellular iron. Both of them show therapeutic effects in different pre-clinical and clinical studies. For instance, deferoxamine showed an inhibitory effect on APP processing and amyloid fibril aggregation in animal models. Deferasirox demonstrates a neuroprotective effect by reducing ferritin and its receptor expression, reversing Ab metabolism in the rodent model. Deferiprone was tested in different clinical trials directed to different neurodegenerative diseases with brain iron accumulation. Results showed a slowdown of cognitive decline [58]. Some randomized controlled studies showed that deferiprone ameliorates iron-related neurological symptoms [59]. The neuroprotective effect of deferiprone was also described on cell lines and animals. It was shown that deferiprone decreases the level of hydrogen peroxide and Ab1-40 [60].

Enhancing the antioxidant system is the second therapeutic strategy in the context of ferroptosis-induced neurodegeneration. Different antioxidant agents were tested in animal and human studies, for instance, tocopherol, N-acetylcysteine (NAC), selenium, polyphenols, and ubiquinone.

Different preclinical studies showed that deficiency of vitamin E may initiate neurodegeneration, while treatment may contribute to the inhibition of ferroptosis. It was shown that the level of tocopherol is low in serum and CSF of AD patients [61]. Recent clinical studies found that high levels of vitamin E may have an anti-inflammatory effect by reducing the number of activated microglia [62]. Selenium is the other potential antioxidant agent which is essential for GPX activity. Several studies showed decreased activity of GPH after the replacement of selenocysteine with cysteine [63]. That indicates that selenium is directly involved in the catalytic activity of GPX and its deficiency may lead to oxidative stress and attenuate protection against ferroptosis. Animal studies showed the neuroprotective effect of selenium injection into the brain, which also had GPX independent mechanism of protection against cell death [64]. Whereas clinical trials with the supplementation of selenium did not describe functional improvement [65]. N-Acetylcysteine is used as a precursor of cysteine with a high BBB permeability level. Increasing cysteine level is a main mechanism of the NAC effect, which leads to the increase of GSH level and protects cells against oxidative stress. The anti-ferroptosis activity of NAC was demonstrated in mouse AD models, with improved learning and memory deficit, as well as neuron function along with the reduced MDA level [66]. Some clinical trials showed improvement in cognitive function in AD patients.

However, a larger sample size is required to determine the therapeutic effectiveness of all mentioned compounds in AD and other cases of neurodegeneration.

OVERVIEW OF IN VITRO AND IN VIVO MODELS USED TO STUDY FERROPTOSIS IN AD

Since 2012 when ferroptosis first was described, several in vivo and in vitro experimental models of ferroptosis were developed. In vitro models focus on neuronal cell lines, astrocytes, and stem cell cultures and are implicated to investigate ferroptosis-related intracellular changes in Alzheimer’s disease.

Models based on Neuronal Cell Lines

HT-22 cells are mouse hippocampal neurons that are used to evaluate oxidative stress-induced cell death. Chu and others have used HT22 neuronal cell line as a model to reveal the role of ferroptosis in glutamate-induced cell death and the effectiveness of ferrostatin-1 on cell death. HT-22 cells were planted in a 6-well plate and were exposed to different dosages (1.25, 2.5, 5, 10, 20 mM) of glutamate for 12, 24, 36, and 48 h. Cells showed the lowest vitality rate when exposed to 5 mM glutamate for 12 h. These conditions were selected to be used in the model. HT-22 cells were treated with ferrostatin-1, a selective inhibitor of ferroptosis, for 16 h, which was followed by adding 5 mM glutamate. LDH release assay has shown that ferrostatin-1 dose-dependently reduces cell death exposed to glutamate, showing maximal effect at a 12 μM dose. DAPI staining also has shown a reduced apoptosis rate for the glutamate + ferrostatin-1 co-treated cells.

Glutamate exposure affected the mitochondria of the cells, which were found to be smaller, with a higher density of the membrane. To ensure the ferroptosis involvement in glutamate-induced cell death, other types of cell death inhibitors (apoptosis inhibitors, autophagy inhibitors, necrosis inhibitors, and iron chelators) also were co-treated with the glutamate on the HT-22 cell cultures. Among the inhibitors first three were ineffective, hence iron chelator DFO prevented cell death in the established model. The authors stated the involvement of ferroptosis in glutamate-induced cell toxicity. Ferrostatin-1 reduces the formation of free radicals and lipid peroxidation in glutamate-exposed HT-22 cell culture [67].

Another widely used in vitro model is the SH-SY5Y cell line, which is a human neuroblastoma cell line. These cells are characterized by low differentiation levels, which enables their differentiation using various protocols. During differentiation SH-SY5Y cells undergo the formation and extension of dendrites and axon, experience elevated electrical excitability of the membrane, synthesize neurotransmitters and their receptors, and form synapses. All the listed processes lead to the generation of cells, similar to primary neurons. In the context of neurodegenerative diseases, this cell line was used primarily to study impaired dopamine metabolism in Parkinson’s disease [68]. It was found that ferroptosis has a major role in dopaminergic neurodegeneration, which was reversed by ferroptosis inhibitors. The SH-SY5Y cell line was also used to study ferroptosis involvement in AD. It has been shown, that specifically designed plant-based flavonoid precursors—hydroxylated chalcones inhibit aggregation of amyloid-beta subunits, erastin-induced Xc path, and reversed inhibition of GPX4, thus inhibiting ferroptosis [69].

Induced Pluripotent Stem Cell (iPSC)-Derived Models

iPSC-derived neurons: Reprogrammed iPSCs can be differentiated into neurons, allowing the generation of patient-specific or disease-specific neuronal models to investigate AD-related ferroptosis

Many neurodegenerative disorders are associated with late age, and there is no obvious genetical basis, it is hard to model these diseases in vitro. Pluripotent stem cells derived neurons are currently used as in vitro models for neurodegenerative diseases. Among the limitations of in vitro models is lack of the aging markers. The iPSC-derived neurons are “freshly made” cells and do not relate to the real age of the donor. Somatic cell (fibroblast)-derived neurons undergo reverse differentiation via Yamanaka transcription factors and thus are relatable with the age of the donor, so can be considered as better model to study the role of ferroptosis in Alzheimer’s disease [70]. Evidence indicates that the produced cell has to reach a similar state to aged neurons, such as similar epigenetic changes, mitochondrial dysfunction, and other age-associated alterations. Similarly, the somatic-cell-derived neurons have a similar epigenetic level. Thus, this model is a successful one due to relatable age-associated issues and can be used to study iron excess-derived neurodegeneration.

Obtaining iPSCs from Alzheimer’s disease patients and differentiating them into neurons show an accumulation of Amyloid beta oligomers, as well as mutations of Amyloid precursor protein and a high rate of oxidative stress. Hence the patient-derived form of iPSCs has great potential to study ferroptosis in AD.

Glial Cell Culture-Based Models

Glial cells encompass microglia, astroglia, and oligodendrocytes, which have a key role in the maintenance of brain functioning, including immune protection, synthesis of several neurotransmitters, regulation of synaptogenesis and synaptic plasticity, and maintenance of structural integrity [71]. The set of functions makes the glial cells a good target for modeling. One of the limitations to use glial cells as models is their slow growth rate, challenging their obtaining. For ferroptosis studies, astrocytes obtained from angiotensin II-stimulated mouse primary cortex are used. The stimulation with angiotensin II increased the level of inflammatory markers (COX-2, IL-6, IL-1b), as well as free radicals, reducing Glutathione levels via downregulation of GPX4. Adding ferrostatin-1 shows dose-dependent reversing of the effects.

Microglia contribute to the establishment of neuronal networks, injury repair, clearance of the extracellular space, and immune response. In pathological states, microglia produce pro-inflammatory cytokines and contribute to neuroinflammation. Like the astrocytes, microglia also contribute to ferroptosis. Graphene quantum dots (GOD)-induced ferroptosis was evaluated on the microglia cell line. The application of ferroptosis inhibitors (Ferrostatin-1, DFO) reversed all the pathological effects. Hence no evidence suggests microglial ferroptosis in the development of AD.

Oligodendrocytes are common myelin-producing cells. They have high levels of iron, which creates a fertile ground for the development of ferroptosis. These characteristics along with their prolonged survival in vitro make oligodendrocytes a good model to study ferroptosis. Cells treated with hemin showed ferroptosis activation manifested with increased levels of free radical and lipid peroxidation product—malonyl dialdehyde. To reveal type of the cell death, apoptosis, necroptosis, ferroptosis, and autophagy inhibitors were applied. Ferroptosis inhibitors solely reversed cell death [72]. Currently, no data is available for the use of oligodendrocytes as a model for studying ferroptosis in AD, hence the model has great potential

Primary Neuronal Culture-Based Models

Primary neuronal cultures are widely used to study mechanisms of various brain diseases, including neurodegenerative diseases. Primary neurons are obtained from mouse embryonic brain tissue. Differentiation and synaptogenesis take place in the culture giving wide opportunities to study these processes. The primary neuronal culture was exposed to erastin, a ferroptosis inducer for 48 h (50 μM) was added to the p. It resulted in decreased viability, higher production of free radicals, and downregulation of GPX4 and Xc system [73]. Pretreatment with deferoxamine, which is a chelating agent to remove excess iron and aluminum [74] reversed all the observed effects. This suggests the potential use of primary neuronal cultures from hippocampal and cortical regions to study ferroptosis in AD.

In vitro, models make it possible to study pathological mechanisms of ferroptosis, such as disrupted iron metabolism, disbalance between free radical generation and neutralization systems, lipid peroxidation, etc. In vivo, models are a good and regulated target to observe ferroptosis initiation, processing, and reversing.

Hence to observe organism-level changes in vivo models are required.

In vivo, models for studying ferroptosis in AD are designed primarily on rats and mice. These are valuable tools to observe functional and organic alterations on the organism level. AD is a rather complex pathology, integrating cognitive decline and behavioral alteration on the context of neurodegeneration. The in vivo models aim to reproduce these deficits as similarly as possible.

Transgenic Mouse Models

APP/PS1 Mice. The first transgenic murine models of AD were designed almost 3 decades ago. These animals were overexpressing mutated APP. This model made it possible to track amyloid beta accumulation in the brain with subsequent microglial activation, neuronal death, and synaptic dysfunction. Despite the hyperphosphorylation of tau was observed in the existing transgenic models, no neurofibrillary tangles were generated. The next step of the transgenic model was the incorporation of the mutated MAPT gene and the development of mice that both produce amyloid-beta plaques and neurofibrillary tangles [75].

Another gene implicated in the pathogenesis of AD is the presenilin genes (PS1 and PS2), encoding subunits of gamma-secretase, responsible for Amyloid beta production. Implementation of mutant genes of presenilin showed altered processing of APP, but no plaques were developed. This approved that APP has a crucial role in plaque formation [76].

Further, it was found that human and mouse Amyloid-beta have differences in amino acids, which prevents formation of the plaques [77]. The APP/PS1 mice model was used to study the involvement of ferroptosis in AD-associated neurodegeneration. The increased permeability and reduction of the pericyte number was observed. Authors stated that ferroptosis is involved in pericyte mitophagy [78].

Currently, there are several lines of transgenic mice models of AD, which differ in characteristics and composition of plaques, presence/absence of angiopathies, etc.

Sporadic AD models. Familial AD encompasses mutation in the gene of Amyloid precursor protein (APP), leading to the overproduction of insoluble beta-amyloid plaques and neurofibrillary tangles. Sporadic AD symptomatically is similar to AD but has unknown etiology. Some authors consider sporadic AD as a brain form of diabetes mellitus and insulin resistance, accompanied by impaired cognitive status.

Streptozotocin is widely used for type-1 diabetes modeling. Parenteral administration of high doses of the drug causes the destruction of pancreatic beta-cells, causing type-1 diabetes. Continuous administration of parenteral low doses of STZ causes damaged insulin receptor signaling, with a following insulin resistance. The ICV administration of sub diabetic doses of STZ leads to the development of an insulin-resistant brain state, which is common for sporadic AD. The experimental animals develop insulin-resistant brain states, common for sporadic AD and accompanied by cognitive decline, impaired cholinergic transmission, energetic deficits, and oxidative stress [79].

FAB animal model of AD. Among the environmental models of AD ferrous amyloid buthionine (FAB) model is considered as one which corresponds to a sporadic form of the disease. The FAB cocktail consists of different agents with the potential neurodegenerative effect. Amyloid aggregates (Ab1-42) are used as the main trigger of neurodegeneration. Ferrous sulfate and buthionine sulfoximine are used to initiate oxidative stress by triggering ROS formation and inhibition of GSH synthesis. In most of the studies that used the FAB model chronic injection into rat brain was described. Thus, both of these features, chemical composition and time period of infusion, best mimics natural phenotype of Alzheimer’s disease and evidence the role of ferroptotic mechanism of neurodegeneration. Interestingly, to our knowledge FAB animal model was described in 2005 by Lecanu et al., whereas ferroptosis was first defined in 2012 by Dixon et al., FAB-treated animals, but not Aβ alone or in combination with iron, showed impaired memory, increased levels of hyperphosphorylated Tau protein in cerebrospinal fluid, neuronal loss, and gliosis [80]. Moreover, histological studies display amyloid plaque deposits and neurofibrillary tangles in the hippocampus and cortex [81]. Another study reported that FAB-treated rats exhibit functional alteration in pyramidal cells of the hippocampus, particularly the increased activity of dopamine neuron populations and memory deficit [82]. In vitro studies showed the aggregated Ab 1-42 alone is significantly less toxic to hippocampal cells, while FAB more completely recapitulates neuronal damage in vitro [83].

Thus, simultaneous damage of pro- and anti-oxidant systems makes the FAB animal model more real and powerful for studying Alzheimer’s disease-like neurodegeneration.

CONCLUSION

Animal models to study ferroptosis in Alzheimer’s disease mechanisms are gaining weight. Combination of in vitro and in vivo experimental models may enable providing an evidence to study ferroptosis on molecular, cellular and organism levels in the neurodegeneration context. This makes the FAB model a good tool for complex understanding of ferroptosis in AD.