*Review* **Regenerative Neurology and Regenerative Cardiology: Shared Hurdles and Achievements**

**Dinko Mitreˇci´c 1,2,\*, Valentina Hribljan 1,2, Denis Jageˇci´c 1,2, Jasmina Isakovi´c 3, Federica Lamberto 4,5, Alex Horánszky 4,5, Melinda Zana 4, Gabor Foldes 6,7, Barbara Zavan 8, Augustas Pivoriunas ¯ 9, Salvador Martinez 10, Letizia Mazzini 11, Lidija Radenovic 12, Jelena Milasin 13, Juan Carlos Chachques 14, Leonora Buzanska 15, Min Suk Song <sup>3</sup> and András Dinnyés 4,5,16,17**


**Abstract:** From the first success in cultivation of cells in vitro, it became clear that developing cell and/or tissue specific cultures would open a myriad of new opportunities for medical research. Expertise in various in vitro models has been developing over decades, so nowadays we benefit from highly specific in vitro systems imitating every organ of the human body. Moreover, obtaining sufficient number of standardized cells allows for cell transplantation approach with the goal of improving the regeneration of injured/disease affected tissue. However, different cell types bring different needs and place various types of hurdles on the path of regenerative neurology and regenerative cardiology. In this review, written by European experts gathered in Cost European action dedicated to neurology and cardiology-Bioneca, we present the experience acquired by working on two rather different organs: the brain and the heart. When taken into account that diseases of these two organs, mostly ischemic in their nature (stroke and heart infarction), bring by far the largest burden of the medical systems around Europe, it is not surprising that in vitro models of nervous and heart muscle tissue were in the focus of biomedical research in the last decades. In this review we describe and discuss hurdles which still impair further progress of regenerative neurology and cardiology and we detect those ones which are common to both fields and some, which are

**Citation:** Mitreˇci´c, D.; Hribljan, V.; Jageˇci´c, D.; Isakovi´c, J.; Lamberto, F.; Horánszky, A.; Zana, M.; Foldes, G.; Zavan, B.; Pivoriunas, A.; et al. ¯ Regenerative Neurology and Regenerative Cardiology: Shared Hurdles and Achievements. *Int. J. Mol. Sci.* **2022**, *23*, 855. https:// doi.org/10.3390/ijms23020855

Academic Editors: Masaru Tanaka and Lydia Giménez-Llort

Received: 7 November 2021 Accepted: 9 January 2022 Published: 13 January 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

field-specific. With the goal to elucidate strategies which might be shared between regenerative neurology and cardiology we discuss methodological solutions which can help each of the fields to accelerate their development.

**Keywords:** stem cells; regenerative neuroscience; brain regeneration; neurology; cardiology; myocardial regeneration; clinical trials

#### **1. Introduction**

Regenerative medicine aims at replacing human cells, tissues or organs damaged by disease or aging and restoring their normal functions. Among some of the most promising approaches in regenerative medicine are stem cell-based therapies, which may provide unparalleled possibilities in the treatment of various conditions, including brain and heart diseases [1]. The therapeutic effect, i.e., restoration of function in the damaged tissue, is attained through direct cell replacement, stimulation of endogenous regeneration/repair systems, establishment of a supportive environment for the remaining cells, or a combination of these mechanisms [2]. Besides the progenitor cells from the affected nervous or cardiac muscle tissue, stem cells of most diverse origins have also proved to be candidates with great potential for translation towards clinical trials. Human embryonic stem cells (hESCs) show the highest differentiation potential, yet their use is hampered by numerous ethical controversies, and alternative sources are sought for [3]. Multipotency of mesenchymal stem cells (MSCs), their presence in almost all adult tissues, the fact that they are usually easily accessible and that they can be differentiated into neural and myocardial lineages makes them very appealing in cell-based therapies [4,5]. Lately, great expectations have also been placed on induced pluripotent stem cells (iPSCs) generated from somatic cells that have undergone genetic reprogramming resulting in pluripotency. Most of cell types are easily obtainable and expanded, usually with no need for immunosuppression following transplantation [6]. Over the past two decades a large number of trials have been conducted, and many are currently underway, including those related to demyelinating diseases and spinal cord injuries, amyotrophic lateral sclerosis, stroke, Parkinson's disease, macular degeneration, as well as acute myocardial infarction, ischemic cardiomyopathy, refractory angina and many others [1]. Although different clinical trials in the fields of neurology and cardiology have reported promising benefits of stem cell-based therapies, many challenges still remain.

This review outlines various types of stem cells that are currently available in neurological and cardiovascular regenerative medicine and reports current state of the art in attempts to introduce those procedures in every day practice. Moreover, we analyse the signalling pathways and mechanisms of their action and examine the outcomes that have been reached with their application. In addition, we discuss the use of novel biomaterials as support for 2D and 3D cell growth, as well as the emerging role of exosomes and their cargos in tissue regeneration. Finally, an overview of the main obstacles, some shared between these two fields, and some field-specific, which we have yet to overcome are given.

#### **2. Induced Pluripotent Stem Cells–Flying Start to Boosting In Vitro Models of the Nervous System and the Heart**

The pioneer studies of Yamanaka and his group yielded protocols for obtaining induced pluripotent stem cells (iPSCs), thus providing the opportunity of dedifferentiating any cell to pluripotent state and, equally important, to obtain autologous, patient-specific cells [7,8]. IPSCs are generated from somatic cells, which have been reprogrammed to acquire pluripotency and have the unique capabilities of self-renewal, proliferation, and differentiation [9]. Since iPSCs can give rise to virtually any cellular lineage, an important application of iPSC technology is the in vitro differentiation of specialized cells, like neurons and cardiomyocytes (Figure 1). This can then be used for investigation of a specific tissue, including both fully differentiated cells and their precursors. Such an approach paved the way for promising advances in patient-specific disease modelling, drug screening, and cell-based therapies without the risk of immune rejection [9–12].

**Figure 1.** The circle of stem cells–based technology: from cell isolation to application.

#### **3. iPSC-Derived Cardiomyocytes**

The course of cardiomyocytes differentiation is validated using molecular markers for different stages of development, as well as investigating their beating capacity, electrophysiology and metabolism [13–15]. Most importantly, we developed methods to differentiate iPSC in cardiomyocytes, with the goal to generate a mixed population of cardiac cells, as ventricular-, atrial- and pacemaker-like cardiomyocytes [16]. That protocol consists of two major steps: firstly, growth factors (e.g., activin A, BMP2) or small molecules (e.g., CHIR99021) activate the Wnt signalling, allowing the mesoderm (*Nkx2.5+*, *Gata4+, Mesp1+*) induction [14,17–19]. Secondly, small molecules (e.g., dorsomorphin) or Wnt inhibitors (e.g., IWR-1, IWP-2) are used to enhance the cardiac lineage specification and differentiation (*cTnT+*, *Myh6+*, *Tnni3+*) [17,19,20]. Under these conditions, ventricular-like cardiomyocytes (*Hey2*+, *Mlc2v*+) are predominant than other cardiac cell types [14,21–23]. However, depending on the research aim, there are also methods to purify and isolate an atrial-like cell population (*Kcnj3+, Kcnj5+*, *Cacna1d+*), for example using retinoic acid or BMP antagonist (e.g., Noggin, Gremlin 2), by upregulating atrial-specific genes. [14,23–27]. On the other hand, the pacemaker-like cardiomyocytes are still difficult to obtain *in vitro*. So far, the inhibition of neuregulin1β/ErbB signalling seems the most efficient way to enrich the sinoatrial node cells population (*Hcn4*+, *Tbx3*+, *Tbx18*+), and only recently it has been hypothesized that modulating the Wnt signalling through Nodal inhibition may promote the pacemaker cells fate [28–30].

### **4. Specific Requirements for In Vitro Heart Muscle Model**

The iPSC-derived cardiomyocytes correspond to the foetal-like state concerning their functional and physiological characteristics [31]. A very specific challenge is to obtain more mature cardiomyocytes, and several methods are currently available [32]. Compared to immature counterparts, these adult-like cardiomyocytes metabolise fatty acids, display a high mitochondrial mass, well-arranged sarcomeres, and higher contraction force. Therefore, cardiomyocytes maturation can be achieved by poviding fatty acids to the culture medium [33], using mechanical and electrical stimulation [34] or developing a 3D cellular model [21]. Interestingly, iPSCs obtained from cardiac sources suggest an improved differentiation capacity in vitro and possibly a higher degree of maturation of

cardiomyocytes. In this regard, the epigenetic memory of somatic cell source may play a fundamental role [35,36]. When the current state of the art is taken in account, the most promising approach is cultivating cardiomyocytes in 3D form (Figure 2). It comes very close to heart's unique cytoarchitectural arrangement and to an even higher level of similarity to the original tissue, with the ultimate goal to establish a heart-on-a-chip model [21,37]. This scaffold-based approach can mimic the patient-specific anatomical microstructure and composition of the human heart and vessels as well as generate responsive constructs to study intact tissue-level cardiovascular physiology. The interaction between cells and the cardiovascular extracellular niche and matrix constituents leads to activation of physiological underlying mechanisms and responsiveness to mechanical, electrical and pharmacological cues. Thus, multicellular microtissue may prove useful for many cellbased applications, like cardiotoxicity assessment and modelling myocardial infarction in a dish [38]. However, comparing structural, mechanical, and biological properties of these structures head-to-head with perfused intact tissues like myocardial and vascular slices and wedges is still warranted.

**Figure 2.** Current options offered by stem cell-based technology for regenerative cardiology and neurology.

Another critical cell subtype of the cardiovascular system is those forming the organotypic vasculature [39]. The generation of these endothelial cells should also rely on organ-specific differentiation protocols, where functional readouts can validate the efficacy and quality of the production. Their specific function comprises barrier-forming continuous layers, a specific vasoactive and growth factor secretion profile and thrombogenic properties [39,40]. Most importantly, being more than a passive conduit, prevascularisation by these endothelial cells can support the long-term survival and instruct the contractility and other functions of adjacent cardiomyocytes within the in vitro generated multicellular constructs. To establish vascularisation, pluripotent stem cell-derived endothelial cells show a remarkable capacity to self-organise into functional microvasculature, like cardiac capillaries, thereby providing sufficient perfusion throughout the cell constructs with a substantial thickness [41].

#### **5. iPSC-Derived Neurons**

The human brain is comprised of a combination of distinct cellular subtypes with a diverse range of specialized functions such as electrical communication, axonal ensheating and metabolic coupling [42]. These include, but are not limited to neurons, which are the primary functional cells of the brain classified via their associated neurotransmitters, and glial cells, such as astrocytes, microglia and oligodendrocytes, which are all critical for maintaining homeostasis and working function in the CNS [43]. iPSCs can be differentiated into several of these specialized cellular subtypes with functional characteristics that are representative of those found in the brain, such as dopaminergic neurons, cortical neurons and the aforementioned neuroglia [44].

The in vitro neuroectodermal induction of iPSCs, initiated via the dual SMAD inhibition method, results in the efficient generation of neural rosettes comprised of neuronal stem cells (NSCs) (Sox1+, Nestin+) that represent a cross-section of the neural tube (Figure 1). These NSCs can then be passaged, producing neural progenitor cells (NPCs) which can be stably maintained in culture [45]. A neuronal differentiation medium can then be applied to NPCs, which can be plated and further differentiated into more mature neuronal (Map2+, TH+, SLC18A3+) and glial cultures, including astrocytes (AQP4+, s100β+) and oligodendrocytes (NG2, Olig1/2, NBP) [46,47].

Neuronal differentiation of iPSCs provides patient-specific cells of neural lineage, opening possibilities for developing therapeutics, analysing drugs, and studying the underlying mechanisms of neurological pathologies. This is done by differentiating iPSCs into NSCs in a 2D setting which includes primitive and neural rosette-type NSCs [48,49]. Conventional 2D in vitro neural models have enabled vast knowledge enhancements regarding brain cellular subtypes, such as adhesive and migratory cellular attachment sites, formation of spontaneous networks, cell type-specific resting membrane potentials and mechanisms underlying axonal guidance [50].

On the other hand, neural differentiation of iPSCs can also be undertaken in a 3D settings. This can involve the generation of neurospheres, floating 3D NSC cultures that have been widely utilized for in-depth NPC analysis and more closely resemble the in vivo setting than 2D cultures [51,52].

Other 3D methods can utlilize artificial scaffolding or extra-cellular matrix (ECM) materials that are continuously under optimization to recapitulate the anatomical organisation of the brain [53,54]. 3D neural culture models involving cell growth using a hydrogel matrix or synthetic scaffolds are highly desirable, offering systems with intricate and easily calculable architecture with specific functional characteristics. Even though the reproducibility of these 3D models is a current challenge, newer methods that involve laser fabrication and bioprinting offer promising avenues for producing accurate and reproducible 3D in vitro neural cultures. Therefore, iPSCs have been, and will continue to be, utilized for advanced microstructured 3D scaffolds for in vitro disease modelling and for the study of neuronal functionc [55–58].

Persistent advances in the methodology used to obtain in vitro brain tissue from iPSCs led to the development of 3D brain-organoids from embryoid bodies [59]. These organoids have been demonstrated to consist of several distinctive brain regions and heterogenous tissue that can mimic the sophisticated architecture of the central nervous system [60,61]. It is worth noting, however, that as the complexity of these 3D cultures improves, so does their variability and heterogeneity. Therefore, improved methods of high content analysis will be required to determine the phenotypic characteristics of these cultures with multidimensional readouts [62].

There are many issues being investigated concerning the source, quality, stability, safety and scalability of human iPSC and derivative cell production for a variety of uses. Concerning the somatic cell source, pre-existing mutations acquired during the lifetime of the donor are more frequent in skin samples than in bone marrow. This means that very early life stage sources, for example those from umbilical cord blood banks, exhibit these potentially adverse events to a much lesser degree [63–66]. However, during the reprogramming, maintenance and scaling-up of iPSC cultures further mutations, including chromosomal rearrangements, can happen. These need to be monitored, especially in case of further clinical use [67–69]. The process of adaptation to the in vitro culture conditions

favours some chromosomal rearrangements occurring more frequently [70]. Development of culture conditions occurrence, as well as advanced quality control methods, are an important direction of stem cell banking and the key towards clinical applicability. Major public and private entities have created human pluripotent stem cell banks with many cell lines originating from patients of different ethnic groups, yet many of them have not been consented for industrial use, and most of them have not been optimized for clinical grade applications–these are all potential hurdles to overcome if clinical applications are to be considered. The distribution of existing cell lines among ethnic groups is unbalanced, but since more nations are developing their own stem cell banks, we are gradually overcoming this ambiguity.

#### **6. Specific Requirements for the In Vitro Nervous Tissue Model**

While the heart muscle is a rather uniformly structured tissue, generally independent of the microanatomic region, nervous tissue brings inherent complexity stemming from the existence of various regions with a variety of cell subtypes and a multitude of functions. Thus, when cultivating cells of the nervous system in vitro, one can distuingish many types of cultures, existing in a range from mixed spontaneously differentiated and heterogenous cultures to those ones in which selection of one subtype of cells is preferred (e.g., motoric neurons, cholinergic neurons, mixed glia-neuronal cultures, astroyctes, sensoric nerons, etc). Sometimes those experiments even include chimeric interspecies cultures [71]. Another important question which brings this complexity to the next level is whether the nervous system can, if at all, be investigated focusing only on one specific cell type or region, e.g., the cerebral cortex. This is a crucial point to consider since the main function of the nervous system is to achieve a well coordinated interaction between its various regions through receiving and transmission of electrical and chemical signals.

Thus, since the physiology of the brain is rather different than that of the heart muscle, it is crucial to address all the advantages and limitations prior to starting any further development. Two dimensional cultures of nervous tissue brought numerous pioneering discoveries on cellular level, but their value in understanding higher order cellular coordination is very limited. Thus, even more than in the heart muscle, 3D cultures of nervous tissue are required for all the research aiming to elucidate physiological and pathological events occurring in interaction between cells.

#### **7. Brain Organoids**

While stem cells platforms based on 2D culture are being successfully used for modeling of human development and disease at cellular and molecular levels, they lack the conditions imitating spatial and temporal signaling as well as the interactions of the cells in their natural niche. These limitations of in vitro culture might be resolved by the application of biomimetic 3D solutions, especially by combining microenvironmental bioengineering with the intrinsic capacity of pluripotent stem cells to build up 3D structures [72]. This intrinsic ability of pluripotent stem cells to self-organize under 3D in vitro culture conditions into highly structured tissue patterns, opened the era of "brain organoids" [60,73]. Yoshiki Sasai and colleagues were the first to obtain highly patterned neural structures resembling muti-layered brain cortex in vitro from human pluripotent stem cells, using SFEBq (serum-free floating culture of EB-like aggregates with quick re-aggregation) protocol [73]. Further developments from the Jourgen Knoblich group brought advanced brain-like 3D in vitro structures with identified regions of cerebral cortex, retina, meninges and chordoid plexus. These 3D structures all exhibit the major stages of prenatal human brain development with functional nervous tissue cell types and cortical layer architecture, thus offering an unprecedented model for investigating human neurodevelopmental and neurodegenerative diseases [74]. Multimodal Single-Cell Analysis (single cell RT-qPCR and functional-microfluidic linked single cell RT-qPCR) of cerebral organoids cultured for more than nine months revealed a high level of neuronal and glial cell diversity as

well as confirmed their functionality with identified cell-type specific responsiveness to neurotransmitters and spontaneous action potential activity [75].

Brain organoid systems appeared feasible to model early human neurodevelopment and its pathology, however they have anatomical and functional limitations which are impairing their use for studying the later developmental stages due to the lack of the correct neuronal network connectivity and vascularization. Much work in the field has been addressed towards overcoming these limitations with two parallel, but interdependent, directions: the first is focused on developing new protocols for generating replicas of multiple brain regions (development of "directed", region specific organoids), while the second is based on constricting regulatory control of the system through bioengineering approaches.

Apart from diseases modeling, brain organoid technology can be personalized for diagnostic or therapeutic purposes if patient-specific hiPSC are applied (Figure 2) [76,77]. Whole brain (cerebral) patient-derived organoids were used to model microcephaly, macrocephaly (Sandhoff disease), periventricular heteroplasia, schizophrenia, Alzheimer Disease and other neural disorders [76,78,79]. Brain region specific organoids, e.g., forebrain to study autism spectrum disorders, or midbrain to study sporadic or idiopathic form of Parkinson's Disease have been already obtained [80,81]. In addition, those methods are combined with a gene-editing approach with the goal to obtain "healthy/repaired" organoids by producing isogenic CRISPER/CAS9 engineered patient–derived iPSCs, as was shown for Sandhoff disease [82].

#### **8. Sources of Cells for Transplantation into Nervous and Heart Tissue**

Cellular therapy refers to the use of cells as medical product to treat human disorders for which other modalities of therapy either does not exist (e.g., stroke) or they are not efficient (e.g., ALS, heart decompensation). Thus, stem cell therapy has a high measurable potential in the treatment of brain and heart diseases through cell replacement and stimulation of the endogenous repair systems. Stem cells of diverse origins (embryonic stem cells, mesenchymal stem cells, induced pluripotent etc.) are all viable candidates with great potential for translation. Here we focus on two most often used stem cell types for the diseases of the brain an the heart: neural and mesenchymal stem cells.

Neural stem cells are a pluripotent cell population, expressing markers nestin and Nop2 [83], and are, thus, already inclined towards differentiation into neurons and glia. Process of forming adult cells of the nervous system, neurogenesis is a process in which neurons are generated through the division of neuronal precursors cells (NPCs) and their differentiation into neuron-specific progenitors. NPCs subsequently, over various stages of precursors, develop into fully functional and mature neurons which integrate into, and modify, existing neuronal networks. In gliogenesis, NPCs differentiate into glial progenitors, which differentiate into astrocytes, oligodendrocytes and ependymal cells.

Mesenchymal stem cells (MSC) are defined as a heterogeneous subset of stromal cells that can be easily isolated from many adult tissues and possess multilineage potential, i.e., ability to differentiate into cells of the mesodermal linage, such as adipocytes, osteocytes, chondrocytes, and myocytes [84]. Actually, the multilineage potential od MSCs allows them to differentiate into neuron-like cells, which exibit molecular and cellular characteristics of neurons. MSCs can give rise to derivatives of both ectodermal and mesodermal lineages. For example, MSCs derived from dental ligament can easily be differentiated into neurons and cardiomyocytes, opening up possibilites in treatment of neuromuscular diseases by tackling different aspects of such a complex pathophysiology [85–89].

#### **9. Extracellular Vesicles–Desired Cellular Product on a Way towards Clinical Application**

Extracellular vesicles (EVs) represent a modality for intercellular communication by acting as plasma membrane enclosed containers for a wide array of signalling molecules and ensuring transfer of biological information over long distances throughout the organism [90,91]. EVs are secreted by all types of cells and their molecular cargo reflects origin and physiological (or pathological) state of the producing cell [92–94]. This dichotomy

is most apparent in the central nervous system (CNS), where EVs are involved in the propagation and spread of several neurodegenerative diseases [94,95]. At the same time, EVs isolated from different types of, healthy" cells can act as effective suppressors of pathological processes [93,96–98]. Since these particles bring therapeutic potential, it is important to develop methods for their effective labelling and follow up. Direct labelling is the simplest and most preferable method for the experiments addressing the effects of external EVs, whereas precise monitoring of behaviour and fate of cell-specific EVs within heterogeneous and 3D cultures requires more sophisticated indirect labelling techniques. We refer the readers to excellent and comprehensive reviews that provide an in depth coverage of the topic [99–101]. Direct EV labelling is most often performed with lipophilic dyes by inserting lipid-anchored fluorophores into the EV membranes. Many commercial dyes such as PKH26, PKH67, DiI, DiD, Dir have been developed and extensively used for EV labelling and tracking in vitro by fluorescence imaging. However, lipophilic fluorophores have several important limitations. First of all, most EV preparations isolated from different sources such as serum, or cell culture supernatants, are contaminated with lipoproteins that can also incorporate lipophilic dyes, thus leading to misinterpretation of EV uptake experiments [102]. Some lipophilic dyes also tend to aggregate, forming nanoparticles with similar size to the EVs (100 nm), that also can be taken up by the cells [102]. In addition, PKH dyes can increase EV size by enhancing clustering and aggregation [103]. These limitations can be, at least partially, overcome with the use of recently introduced Mem lipophilic dyes that did not aggregate or change the size of EVs [104]. In conclusion, although simple and straightforward, direct EV labelling with lipophilic dyes has important limitations and therefore requires careful interpretation to avoid misleading results. Indirect labelling can be achieved by CFSE (carboxyfluorescein diacetate succinimidyl ester) fluorescent dye. It is activated by esterases and covalently binds to free amines inside the cells, or vesicles. Interestingly, after indirect labelling of cells, CFSE-positive EVs were detected only in the pellets after 10,000× *g* centrifugation (corresponding to microvesicular fraction originating from plasma membrane) indicating that indirect CFSE labelling may help to distinguish between microvesicular and exosomal fractions [105]. However, large concentrations of CFSE could be necessary to obtain vesicles with sufficient fluorescence and such high dye concentrations can be detrimental to the EV-producing cells [106]. Another promising study recently used hydrophobic insertion of maleimide (Mal) into the EV membranes [107]. Other strategies are focused on RNA imaging using chemical dyes such as Alexa Fluor 488-labeled siRNA, or Cy5-siRNA, or other membrane permeable dyes that are selective for RNA that do not require conjugation, such as including E36, Styryl-TO and SYTORNAselect [108].

When coming to the topic of application of EVs, the major advantage they possess is that they can easily cross the blood brain barrier (BBB) and enter into the brain [109]. However, pharmacokinetic studies in vivo have shown that EVs can be very quickly removed from the bloodstream, with a majority of them being = entrapped in the liver and the lungs [110]. Accordingly, several groups investigated alternative delivery via minimally invasive intranasal route [111,112]. The EVs secreted by mouse macrophages were permeabilized, loaded with antioxidant enzyme catalase and applied intranasally to 6-hydroxydopamine (6-OHDA) mice [113]. Study demonstrates that EVs associated with microglia cells reduced their inflammatory activity and improved the apomorfin test results [113]. Another study compared how EVs stained with lipophilic dyes, or labelled with gold nanoparticles, distribute in the CNS after intranasal application [114]. Gold nanoparticles-marked EVs allow live observation of particle distribution in the brain by using accurate computer tomography methods. Interestingly, another study demonstrated similar distribution patterns of EVs labelled by both methods [9,114]. More importantly, EVs selectively accumulated in the affected areas of the brain. For example, after intranasal application to the 6-OHDA-treated mice (PD model), exosomes selectively accumulated in the damaged striatum areas even up to 96 h [114]. These findings confirm the potential of EVs as a therapeutic tool against various diseases and disorders of the CNS.

#### **10. Cell Transplantation for Heart Ischemia**

Heart failure and its direct consequences represent the leading cause of death worldwide [115]. Although heart transplantation developed substantially in the last decades, there are not enough donors which would satisfy the existing needs. Moreover, heart transplantation is a very complex and expensive procedure that afterwards require lifelong immunosuppression. The mechanism by which transplantation of stem cells into the infarcted heart leads to health improvement is not yet completely understood. The most straightforward expectation would be that transplanted stem cells form new myocardial cells with the capability to contract. However, preclinical and clinical trials revealed at least two obstacles in this theoretically simple approach: first, transplanted cells survive very briefly, so differentiation into myocardial cells is not sufficient. Second, if maturation occurs, coupling with the host myocardium is not successful. As a result, arrhythmia is a very common side effect of such an approach [116].

Preclinical studies focusing on acute infarction, e.g., with interventions within 4 weeks after the incident, reported beneficial effects [117]. On the other hand, studies which were aiming to improve condition several months after the ischemic incident were not so successful [118]. With that being said, recently, the attention has shifted from the potential of transplanted stem cell to differentiate into cardiomyocytes towards secreting factors that improve the condition of damaged myocardium [119]. Reported mechanisms include immune modulation which promotes endogenous cardiac repair [120]. Additionally, it has been shown that stem cells transplanted into the heart secrete cytokines, with rather significant anti-apoptotic effect. One of the most positive effects observed after myocardial infarction is achieved by IL-10, which improves survival and function of myocardial muscle [121].

There are many clinical trials which assessed the efficiency of stem cells for acute myocardial infarction. However, their results are rather heterogenous. Those which focused on myocardial contractility and ventricular remodeling did not find statistically significant improvement. However, significant improvements were found when a longer followup was taken into account, ranging from one to three years [122,123]. Most importantly, ejection fraction was regularly improved and even ventricular remodeling was shifted in a positive direction.

Even after the transplanted cells disappear, beneficial effects can be followed for months and years. Thus reduced inflammation and stimulated vascularization can be detected for a long period, reaching up to few years [124]. Thus, it became clear that, unlike pharmacologic and surgical approaches, cell therapy can stimulate endogenous tissue regeneration to reverse worsening cardiac dysfunction. Some of the most commonly reported benefits of stem cells based clinical trials are listed in the Table 1.


**Table 1.** Overview of pathological entities and reported beneficial effects of cells.

### **11. Specific Requirements for Further Improvement of Cell-Based Therapy of Heart Diseases**

Future developments needed to boost cell-based therapy of heart diseases include nanotechnologies and bioengineered platforms, where stem cells are preconditioned to resist their implantation into a highly stressed myocardial tissue. Basically this approach consists of the development of bioactive membranes made of two integrated materials: (a) one nanofiber matrix made out of self-assembling peptides with molecule-release capacity (for growth factors such as VEGF and FGF), and (b) contained in a microscale elastomeric scaffold that provides the mechanical framework (elastic, loading) that will match the cardiac tissue mechanics. Both are essential to promote local angiogenesis in a necrotic affected tissue as well as its regeneration.

In many congenital heart diseases neonatal ventricles demonstrate a number of intrinsic pathologic modifications, including relative immaturity of the extra-cellular matrix, inappropriately low transcription factor expression and increased myocyte apoptosis, this should open the way for the evaluation of treatments associating tissue engineering with cells implants. The main mechanisms by which cell transplantation and tissue engineering can bring functional benefits in myocardial diseases is the combination of cells and scaffolds, which limit the spread of the pathologic area, preventing excessive remodeling and dilatation of the ventricle [143–146].

Emerging biomimetic technologies include 3D printing and additive manufacturing [147]. For heart healing applications, 3D-printed porous poly-caprolactone (PCL) elastomeric scaffolds represent a promising material functionalized with bio-additives such as stem cells, exosomes and angiogenic growth factors. Cardiopatch and Cardiowrap ventricular support bioprostheses were able to integrate in the damaged myocardium and the adjacent healthy heart, becoming artificial extracellular matrix that offers adequate cell niches for the homing of stem cells. These approaches contribute to the generation of Bioartificial Myocardium, offering posibility that the need for heart transplantation in the future will be reduced [127,148].

#### **12. Cell Transplantation for Diseases of the Nervous System**

The limited neurogenesis capacity in the brain makes neurological conditions difficult to treat. That's why cell transplantation approach is intensively being tested for neurological diseases.

Post-ischemic acute brain injury typically peaks within 24 h of the insult, and reaches its highest point within 48 h [149]. Due to this quick onset and short duration of acute brain injury, potential neuroprotective therapies need to be administered early, i.e., within 3–6 h of the onset. This has proven to be challenging in the clinical practice. Any treatment outside of the 48 h window will offer limited neuroprotection, and could instead be mainly restorative, targeting angiogenesis, vasculogenesis, neurogenesis, and synaptogenesis [128,150]. Finding a therapeutic approach that would delay the progressive secondary neurodegeneration will also benefit stroke survivors. To date, most cell transplantation studies have been conducted on animals during acute phase of post-ischemic injury, leaving chronic time points understudied. It has already been shown that in addition to anti-inflammatory, anti-oxidative and anti-apoptotic effects, transplanted cells also secrete various factors acting neurotrophically exhibiting neuroregenerative effects [130,151].

Upon optimized dose regime and the route of administration, the use of stem cells shows benefits in both the acute and subacute phase, as well as in the chronic phase of cerebral ischemia [131,132]. Similar has been observed in other diseases with neuroinflammatory componente, like amytrophic lateral sclerosis or multiple sclerosis. Since a higher degree of neuroinflammation is present in the acute and subacute phase of cerebral ischemia, in these phases it is necessary to use higher doses (10–1200 million cells) and to choose less invasive ways of stem cell application, such as intravenous, intra-arterial, intranasal and intraperitoneal [131,133,134]. In these phases, various stem cells have shown positive effects so far. In the acute phase (1–3 days after stroke): mesenchymal stem cells (MSCs) and

human mononuclear cells (MNCs), human embryonic stem cells (hESCs), human neural stem cells (hNSCs), and multipotent adult progenitor cells (MAPC) were used [131,152,153]. In the subacute phase (7 days after stroke): autologous CD34+ stem/progenitor cells and bone marrow stem cells (BMSCs) were used [131,154]. In the chronic phase (weeks, months, years) after stroke the smaller doses of stem cells were used (1–5 million cells), albeit with more invasive application methods (intracerebral and intraventricular) in order to allow greater bioavailability of injected cells near the affected brain region [128,155].

In the last two decades more than 70 clinical trials with stem cells for brain diseases have been successfully finished, but no definitive efficacy trials have been concluded. As such, there is currently still no approved cell therapy for neurological diseases. When talking about stroke, as the most common disease of the brain, various approaches have been taken thus far. Not entering into details of various type of stem cells and routes of cell delivery, all trials of Phase 1 and 2 reported safety and visibility. It is interesting to mention that one of the very first trials performed in 2005 in South Korea with 30 patients with cerebral infarct, who received IV infusion of autologous MSCs, reported a significant reduction in mortality within five years of stroke incidence compared to patients who did not receive MSC transplantation [135]. In clinical settings, the recipients of allogeneic MSCs demonstrated long-lasting or transient neurological improvement. Additionally, allogeneic MSCs infusion was associated with a short term decrease in circulating T cells and inflammatory cytokines [136]. The implantation of SB623 to the sites surrounding the subcortical stroke region was safe and accompanied by improvements in neurological recovery in 12 patients in a 2-year study [137]. At this stage, clinically confirmed beneficial effects were shown by CTX0E03 cells (hNSCs), administered one month after cerebral ischemia (a single intracerebral dose of up to 20 million cells), and SB623 (allogeneic MSCs), administered several times with 2.5, 5, and 10 million cells for a period of 6–60 months after stroke [129,131]. As the systemic inflammatory response is a major pathological component in secondary post-ischemic cell death [156], including some specific types of death, like necroptosis [157], stem cell transplantation should to be the therapy of choice to reduce neuroinflammatory effects and help stroke outcomes. Considerable numbers of clinical trials with stem cell therapy for stroke are currently underway. Clinical trials should include patient's co-morbidities which also can affect the efficacy and effectiveness of cell therapy.

MSCs are a population of cells which can be safely harvested from patients. Due to their low immunogenicity and reported benefits, they are already being recognized as approved therapeutic product in some countries [158]. Additionally, MSCs are capable of migrating towards lesioned areas upon receiving attraction signals by certain chemokines, suggesting their potential use as vehicles for therapeutic agent's delivery [159]. Therefore, as therapeutic agents, MSCs have multiple modes of action, including cell replacement, immunologic and metabolic properties; showing a pleiotropic activity that modify the tissues response to injuries and activate restorative mechanisms that improve organ function. Intense interchange of active cellular products between MSCs and resident cells have been proven, demonstrating the potential of MSCs secretome to achieve various paracrine effects, including immunomodulation [160]. Moreover, organelle interchange has been proven, including vesicular traffic (exosomes, microvesicles, etc), where, in addition to the vesicular cargo, MSCs inject membrane (carrying protein membrane complexes, receptors, ion channels, etc.) into host cells [161].

MSCs from the bone marrow had been widely used in clinical trials for neurological diseases. They were demonstrated to be safe but their effects were not always consistent, as preclinical studies suggested. This may be due to poor survival in disease environments and/or because inappropriate therapeutic dosage and route of delivery or inconsistent trial design [162–165].

In some studies, ALS patients treated with MSCs displayed a slight and transient decline in disease progression [138] Interestingly, postmortem evaluation of ALS patients treated with MSCs showed that a more significant number of motor neurons were preserved at the location in the spinal cord where the cells were administered, compared to other spinal sites [139]. Some of the most commonly reported benefits of stem cells based clinical trials are listed in the Table 1.

### **13. Specific Requirements for Further Improvement of Cell-Based Therapy of Brain Diseases**

More than 300 papers have been published in the last 20 years reporting transplantation of cells in animal models and more than 70 clinical trials have been conducted in humans with neurological diseases with some common breakthroughs and some common obstacles. First of all, dogma that transplanted cells need to integrate and survive for a longer period is not only seen as obsolete, but in some cases is even overly stressed. Therefore, one needs to focus on cell products which are, nevertheless, being secreted in large quantities by many cell types. In addition, modification of these products can be achieved by genetic modifications of the stem cells [166]. Secreted growth factors, short sequences of RNA in various forms and still yet to be discovered components, often packed in the form of extracellular vesicles, obviously have a very strong and beneficial influence. So, it became clear that we need to focus on recognizing those beneficial products, to discover mechanism by which they improve regeneration, and then on methods how to deliver them in sufficient quantities. Apart from direct transplantation, intravascular delivery, based on positive results, deserves our attention [167]. Moreover many methodological gaps in clinical translation must be recognized. Well-designed, biomarker oriented endpoints and comparative trials are needed to address specific issues such as type of cells, cell doses, responsive phenotypes and time window of efficacy.

When thinking about side effects of cellular therapy, it is important to notice that transplantation of stem cells into brain tissue very rarely brings any significant obstacles from that side. Probably the most well defined are those linked to dyskinesia, mostly observed in transplantation to patients suffering from Parkinson's disease. However, methods to predict which patients are more prone to those side effects have been already developed. It is interesting to notice that no serious effects coming from uncontrolled electrical activity (e.g., partial or generalized seizures) of such cells have been reported. On the other hand, common obstacles observed are a limited period of activity of such cells, with very time-limited secretion of needed molecules. Thus, the main focus is in securing longer and more substantial effects of the secretome.

#### **14. Conclusions Remarks**

In this review we gathered experience from the last few decades dealing with attempts to treat diseases of the heart and the brain (primarily ischemic in its nature) by using stem cells and their products (Figure 2). When we make a general overview of what has been achieved with these replacement strategies, i.e., the approach in which transplanted cells will replace lost ones in the host tissue, results are rather limited. Nevertheless, replacement therapy seems to be very promising in the case where a very specific subpopulation of neurons, in limited regions, are involved. This can be seen in positive, albeit transient, results in clinical trials including patients with Parkinson's disease [141]. In all other cases, especially in brain ischemia (stroke) and myocardial infarction, transplanted cells can still hardly replace what has been lost. It is very interesting to notice that we expected probably much more from this approach in the heart tissue, which is, in theory, much less complex, than the neural one. However, cells which succeeded to survive in the cardiac muscle for a longer period, could hardly coordinate their activity with the rest of the healthy muscle and, most interestingly, often cause problematic arrhythmias. It is important to notice that arrhythmias in the heart muscle are a much more common problem of stem cell transplantation than uncontrolled electric activity of the transplant in the brain. In the same time, several decades of stem cells - based attempts to treat those diseases brought a huge progress in understanding of complexity of the tissue affected by pathological process. Although the ultimate goal is to discover and launch new drugs and/or procedures for human diseases, fragments of knowledge which we are collecting are without doubts constantly improving medicine.

When we take a look into the effects transplanted cells achieved with their secretome, and considering the experience in treating both the heart and the brain, this strategy emerges as a promising one. This idea has been boosted even further by the discovery of several types of extracellular vesicles which carry short sequences of RNA, peptides, growth factors, etc. In both organs, products of transplanted cells clearly influence inflammation and, in most of the cases, decrease damage with measurable effects. This is the case with neurodegenerative diseases such as ALS [168] or Alzheimer's disease [169]. One of the probably most surprising observations, again seen in both the heart and the brain, is that those effects are often more pronounced in chronic than in acute phases. Thus, overall survival and improvement in major parameters demonstrate statistically significant differences when patients are followed after 6, 12 or 48 months [122,123,137]. How is this possible if majority of cells disappear within a few weeks after transplantation? We can think of two possible explanations: first, those cells which remain, although in small numbers, are naturally selected as those which succeed to achieve substantial positive effects. So here we obviously have an example of supreme quality ruling over quantity. Another element adding to this explanation might be that a combination of positive effects achieved by all cells, before they disappear within a few weeks after transplantation, triggers a positive chain of events which requires a lot of time to pass the threshold which is then recognized as a positive therapeutic effect. Another common point where research into the brain and the heart yielded mutual benefits for both fields is a piece of knowledge about the need for standardization of products secreted by stem cells. Standardization is not only needed in order to cause more comparable results, but also to better define routes of delivery. When this will be achieved, and many efforts are currently being undertaken in that direction, one can imagine repetitive injection of solutions with extracellular vesicles, which will improve regeneration of either neural or cardiac muscle tissue. This review could not cover all parts of this complex field, so, for example, here we did not take in consideration numerous options of genetic engineering, which offers advantages of genetically modified cells. In addition, bioengineering field based on biomaterials is progressing even faster than stem cells. By taking all this in consideration, one of the factors which slow down the progress is complexity of all these elements which requires truly multidisciplinary approach. A very wide and multidimensional perspective is needed in order to pass the threshold of success in clinical trials. To conclude, the major advice we can get from the experience collected thus far is that more standardized methods of transplantation, either with well defined populations of cells or with extracellular vesicles are needed. In addition, transplanted cells need time to bring positive effects. Clinical trials need to plan prolongued follow up of the patients and, whenever possible, account for repeated therapeutic procedures based on cells and/or their products. When such a protocol enters routine practice, we will be witnessing the final confirmation of the value of regenerative medicine in the treatment of major human diseases.

**Funding:** DM research is funded by Croatian Science Foundation project Orastem (IP-16-6-9451) and by European Union through the European Regional Development Fund, as the Scientific Centre of Excellence for Basic, Clinical and Translational Neuroscience under Grant Agreement No. KK.01.1.1.01.0007, project "Experimental and clinical research of hypoxic-ischemic damage in perinatal and adult brain". VH and DJ are supported by PhD grants by Croatian Science Foundation. LM research is partly funded by the AGING Project for Department of Excellence at the Department of Translational Medicine (DIMET), Università del Piemonte Orientale, Novara, Italy. LR and JM research is supported by Grants no 451-03-68/2020-14/200178 and no 451-03-9/2021-14/200129 of the Ministry of Education, Science and Technological Development of Republic of Serbia. GF was supported by Medical Research Council [MR/R025002/1], NIHR Imperial Biomedical Research Centre and the Hungarian National Research, Development and Innovation Fund [2020-1.1.6-JÖVO-2021- ˝ 00013 and K128369]. A.H, F.L., A.D were supported by EU Horizon 2020 Marie Skłodowska-Curie grant agreement No 812660 (DohART-NET) and Grant Agreement No. 739593 (HCEMM) (for A.D) and Chinese-Hungarian Bilateral Project (2018-2.1.14-TÉT-CN-2018-00011, Chinese No. 8-4 (for A.D.) SM is founded by Spanish State Research Agency SAF2017-83702-R and the TERCEL (Instituto de Salud Carlos III (RD16/001/0010 and PID2020-11817RB-100). LB is supported by National Science Centre, Poland, grant # 2019/35/B/NZ3/04383.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors acknowledge support within the framework of COST Action BIONECA CA 16122—Biomaterials and Advanced Physical Techniques for Regenerative Cardiology and Neurology.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **Iron Homeostasis Disorder and Alzheimer's Disease**

**Yu Peng 1, Xuejiao Chang <sup>1</sup> and Minglin Lang 1,2,\***


**Abstract:** Iron is an essential trace metal for almost all organisms, including human; however, oxidative stress can easily be caused when iron is in excess, producing toxicity to the human body due to its capability to be both an electron donor and an electron acceptor. Although there is a strict regulation mechanism for iron homeostasis in the human body and brain, it is usually inevitably disturbed by genetic and environmental factors, or disordered with aging, which leads to iron metabolism diseases, including many neurodegenerative diseases such as Alzheimer's disease (AD). AD is one of the most common degenerative diseases of the central nervous system (CNS) threatening human health. However, the precise pathogenesis of AD is still unclear, which seriously restricts the design of interventions and treatment drugs based on the pathogenesis of AD. Many studies have observed abnormal iron accumulation in different regions of the AD brain, resulting in cognitive, memory, motor and other nerve damages. Understanding the metabolic balance mechanism of iron in the brain is crucial for the treatment of AD, which would provide new cures for the disease. This paper reviews the recent progress in the relationship between iron and AD from the aspects of iron absorption in intestinal cells, storage and regulation of iron in cells and organs, especially for the regulation of iron homeostasis in the human brain and prospects the future directions for AD treatments.

**Keywords:** Alzheimer's disease; iron homeostasis disorder; iron homeostasis regulators; β-amyloid; tau; APP; central nervous system; oxidative stress; pathogenesis; genetic intervention

### **1. Introduction**

The transition metal element iron is the second most abundant metal element in the earth's crust behind, aluminum. It is also an essential trace element and an important component of metalloprotein for human body [1,2]. Due to its unique chemical reaction characteristics, it plays an important role in maintaining normal physiological function and metabolism, such as oxygen transport, DNA synthesis, iron sulfur cluster synthesis, neurotransmitter synthesis and electron transfer in respiratory chain [3–5]. The adult human body contains 3–5 g of iron [2]. In the normal metabolism of the human body, iron ions are absorbed into the blood through the small intestine and transported to the parts of the body requiring iron. Although the body strictly regulates the regulation of iron metabolism, changes with age, genetics and the environment will lead to iron metabolism disorders [6]. The disorder of iron metabolism in the body will catalyze the formation of reactive oxygen species (ROS) through Fenton and other chemical reactions, attack DNA, protein and lipid molecules, and lead to cell damage [7,8]. In recent years, more and more research teams on the pathogenesis of Alzheimer's disease (AD) have shown that the oxidative stress induced by iron metabolism disorder and the production of ROS are related to the pathological process of AD [7,9]. Alzheimer's disease is an age-related neurodegenerative disease with clinical symptoms of memory decline, cognitive impairment and learning impairment [10–12]. With the increasing human life span, the incidence rate of AD is also increasing, and has become one of the most important fatal diseases [5,13,14]. The pathological features of AD in the brain are the extracellular deposition of Aβ proteins forming

**Citation:** Peng, Y.; Chang, X.; Lang, M. Iron Homeostasis Disorder and Alzheimer's Disease. *Int. J. Mol. Sci.* **2021**, *22*, 12442. https://doi.org/ 10.3390/ijms222212442

Academic Editor: Masaru Tanaka

Received: 16 October 2021 Accepted: 10 November 2021 Published: 18 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

insoluble senile plaques and the intracellular accumulation of hyperphosphorylated tau proteins forming neurofibrillary tangles (NFTs), which result in a large degree of neuronal cell death [11,15,16]. Thus far, the main causes and pathogenesis of AD have not been fully clarified. Many research teams have found that there is regional deposition of iron in the brain of AD patients [17–19]; treatment with an iron chelator can effectively alleviate the symptoms of AD [9], suggesting that iron metabolism disorder has a close relationship with AD.

This paper reviews the relevant research progress in the field of iron and AD in recent years, focusing on the oxidative stresses induced by normal iron metabolism and its metabolic disorders, especially for abnormal expression of the iron transporters, transferrin receptors, divalent metal transporters, and their relationships with the AD pathological mark proteins, such as Aβ and tau proteins. Relevant contemporary AD treatment measures have also been discussed and prospected. The iron homeostasis on AD provides a theoretical basis for the prevention and treatment of neurodegenerative diseases and an effective drug screening target.

#### **2. Physiological Function and Metabolic Process of Systemic Iron**

#### *2.1. Physiological Function of System Iron*

Iron is an essential trace metal element and an important component of metalloprotein [2]. Due to its unique chemical reaction characteristics, iron plays an important role in oxygen transport, DNA synthesis and repair, energy generation and enzyme function, such as the formation of a variety of coordination compounds with organic ligands and redox reactions by the mutual conversion of divalent iron and trivalent iron [3,6,8,20].

### *2.2. Metabolic Process of System Iron*

#### 2.2.1. System Iron Absorption

As we know, the adult human body contains about 3–5 g of iron [2], for individuals without blood transfusion, a part of the iron in the system comes from intestinal cells absorbed from food, and the other part comes from macrophages [6,21]. As shown in Figure 1, the absorption of iron from food is Fe3+, which is reverted to Fe2+ by DCYTB (duodenal cytochrome-b-like protein); then, the divalent metal transporter 1 (DMT1) on the surface of the intestinal cell membrane combines the ferrous iron and transports it into the intestinal epithelial cells [20]. The ferrous iron entering intestinal epithelial cells can be transported to mitochondria for heme molecule synthesis, or oxidized to ferric iron and stored in ferritins [22,23]. Excess Fe2+ is released into plasma by FPN (ferroportin) which is located on the basal intestinal cell membrane where it is again oxidated to Fe3+ by the same situated hephaestin [24–26].

As shown in Figure 2, Fe3+ in plasma can bind to transferrin (TF), which is transported through blood in the form of TF-Fe complex. The complex then bind to transferrin receptor 1 (TfR1) that highly expressed on the surface of iron demanding cell membrane, and it enters iron demanding cells through clathrin-mediated endocytosis [27,28]. Fe3+ in endocytic vesicles is reduced to Fe2+ by STEAP (six-transmembrane epithelial antigen of prostate) and released into cells by divalent metal ion transporter DMT1 after separation in low pH environment of endocytic vesicles [29–32]. In addition, ZIP14 (member of the Zrt/IRT family) was initially identified as a transporter of Zn. In subsequent studies, it was found to be involved in the transport of ferrous iron released from endocytic vesicles into the cytoplasm [1,28,33]. TF and TfR1 separated from Fe3+ enter the plasma and are redistributed to the surface of cell membrane to participate in iron transport and the next round of iron absorption, respectively [1]. Fe3+ in plasma can also combine with citrate, ATP and ascorbate to form small-molecule complexes [6].

#### 2.2.2. Storage and Loss of System Iron

In the body, iron is mainly stored in liver cells and macrophages. Macrophages phagocytize the aging red blood cells and release the iron ions inside red blood cells; then, the released irons are stored in ferritin proteins in the macrophages [34]. When the body is in a state of iron demand, macrophages secrete ferritin protein into the serum circulatory system; therefore, the concentration of ferritin protein in serum can reflect the state of iron content in the body [35]. Ferritin protein plays an important role in iron storage and antioxidation in cells [36]. Ferritin protein contains two subunits of H-ferritin and L-ferritin, which exhibit ferrous oxidase activity and iron storage function, respectively [37]. Fe2+ in cells is oxidized by H-ferritin and stored in L-ferritin. Each ferritin protein can store 4500 iron atoms, which can considerably reduce the cell level of free iron ions and prevent the damage caused by free-iron-induced oxidative stresses; thus, it has antioxidant effects [38]. When the concentration of iron in cells decreases, ferritin protein is decomposed into hemoxanthin by lysosomes. Hemoxanthin and ferritin protein can be detected by Prussian blue staining [39]. In addition to ferritin protein, iron entering the cell can enter mitochondria to synthesize heme, as well as the iron sulfur cluster, and participate in the process of aerobic respiration as a cofactor of mitochondrial respiratory chain protein. It can also combine with some small molecular substances in the cell, such as citric acid, ATP, AMP and pyrophosphate to form an intracellular free iron pool [6,8,40]. The amount of pooled free iron can reflect the change in iron content in cells, which can be detected by some fluorescence techniques [8]. Increasing the pool content will produce harmful substances through redox reactions, causing damage to cells, which could even lead to cell death when it is serious [8,41,42]. Iron entering the blood can also be ingested and utilized by iron cells and iron storage cells. Most of the iron in the blood is used by red blood cells to participate in the transport of oxygen. About 20–30% of the iron is stored in the liver and macrophages, and some iron is involved in the formation of myoglobin, cytochrome and iron-containing enzymes [8].

**Figure 1.** Nonheme iron intestinal absorption and transport by intestinal cells. Food Fe3+ is reduced to Fe2+ by DCYTB, which binds to the divalent metal transporter DMT1 on the surface of the intestinal cell membrane and transported into the intestinal epithelial cells. The Fe2+ that enters the intestinal cells can enter the mitochondria for the synthesis of heme. It can also be oxidized to Fe3+ and then stored in ferritin. The excess Fe2+ is released from FPN into the plasma and then oxidized to Fe3+ by hephaestin. Each molecule of Apo-transferrin in the plasma combines with two Fe3+ ions to form Holo-transferrin-Fe. The complex transports iron in the blood to the organs in the body that require iron.

**Figure 2.** Somatic cell absorption and the transport of iron ions. Fe3+ in plasma can bind to Apotransferrin (Tf), which forms a Tf–Fe complex; then, it is transported through the blood to bind to the transferrin receptor (TfR1) that requires high expression on the surface of iron cell membranes, which enters the iron-requiring cells through endocytosis mediated by clathrin. In the endocytic vesicles, Fe3+ is reduced to Fe2+ by the six-transmembrane epithelial antigen of prostate (STEAP), After separation in a low-pH environment, Fe2+ is released into the cell by the divalent metal ion transporter DMT1. The iron ions in the cell can enter the mitochondria to participate in the redox reaction and can also be stored in the ferritin protein. When the body is in a state of iron limiting, Fe2+ can be transported to the outside of the cell through FPN and oxidized by hephaestin to Fe3+, and combines with Apo-transferrin to form Holo-transferrin.

The normal human body loses about 1–2 mg of iron every day [36,43]. Iron in the body is mainly excreted from intestinal mucosa, skin cells, sweat and urine [4,30,44].

#### 2.2.3. Regulation of Iron in Cells

Iron regulatory proteins (IRPs) combine with iron regulatory elements (IREs) in the 3 or 5 untranslated region of mRNA transcripts of iron-metabolism-related genes to regulate the iron concentration in cells [43,45,46]. The IRE region contains a loop of 5 cagugn-3 folded by 30 nucleotides (in which the hydrogen bond formed between G and C stabilizes its structure), without pairing to form hydrogen bonds which will destroy this structure [8,46–48]. As shown in Figure 3, IREs are located at the 3 -UTR and 5 -UTR areas of TfR1 and DMT1 mRNAs, and ferroportin and ferritin mRNAs, respectively, although the binding of IRPs to IREs could finely regulate the iron concentration in cells [49]. When the concentration of iron ions in cells is too high, on the one hand, it will induce conformation changes in the untranslated region of TfR1 and DMT1 mRNAs, so that IRPs cannot bind to the IRE region, and those mRNAs are degraded, whereas the expression levels of ferroportin and ferritin proteins are increased; on the other hand, iron ions can bind to IRP1, forming iron sulfur clusters in IRP1 that exhibit cytoplasmic aconitase activity. In contrast, when the concentration of iron ions decreases, the binding ability between IRP and IRE is enhanced, the expression level of ferroportin and ferritin proteins will decrease, and the expression level of TfR1 and DMT1 will increase [48,50–52].

**Figure 3.** Regulation of iron homeostasis in cells. IREs are located in the 3 -UTR region of TfR1 and DMT1 mRNAs, whereas they are located in the 5 -UTR region of ferroportin and ferritin mRNAs. The combination of IRP and IRE regulates the iron ion concentration in the cell. When the iron ion concentration in the cell is too high, it will induce conformation changes in the untranslated region of mRNAs, making IRPs unable to bind to the IRE region; then, mRNAs of TfR1 and DMT1 are degraded, and the expression level of ferroportin and ferritin increases. On the other hand, iron ions bind to IRP1, and can form iron–sulfur clusters in IRP1 that exhibit cytoplasmic aconitase activity. In contrast, when the iron concentration in the cell decreases, the binding ability of IRP and IRE is enhanced, which leads to a decreased expression of ferroportin and ferritin proteins, and an increased expression of TfR1 and DMT1.

Iron regulatory proteins IRP1 and IRP2 are intracellular iron sensors. These are two proteins that are homologous proteins and belong to the iron–sulfur cluster isomerase family [53]. IRP1 can form a cis-aconitase-type iron sulfur cluster (4Fe-4S), which not only determines its functional mode, but also serves as an important regulatory site. IRP1 forms an iron sulfur cluster only when cells are rich in iron, in which IRP1 can display cis-aconitase activity in cytoplasm; however, it reduces the ability of IRP1 to bind to IRE. Low concentrations of iron in cells induce the depolymerization of iron sulfur clusters in IRP1 and enhance the ability of IRP1 to bind to IRE, although the mechanism of iron sulfur cluster depolymerization in IRP1 has not been fully illuminated. In addition, the increase in NO and H2O2 concentration in cells will activate the activity of IRP1 and promote its binding to IRE [38].

Iron ions and oxygen regulate the synthesis of IRP2 in cells through post-translational mechanisms. IRP2 has lost the activity of aconitase in the process of evolution. The decrease in intracellular iron ions and oxygen concentration promote the synthesis of IRP2 and maintain its stable state. In contrast, the increase in iron ion and oxygen concentration will accelerate the degradation of IRP2. The N-terminal 73 amino acid sequence of IRP2 is characteristic of IRP2. This highly conserved 73 amino acids is encoded by a determined exon and is related to the iron-dependent degradation of IRP2 [8].

#### 2.2.4. Regulation of Iron in the System

Iron ions exported from the intestine are absorbed by the iron-demanding tissues and organs of the body through blood circulation. The liver is the main organ for the regulation of iron balance, which plays an important role in the regulation of whole-body iron balances [34]. The liver produces and secretes the hepcidin hormone [25,54], which is a short polypeptide composed of an 84 amino acid sequence encoded by the HAMP gene and a 25 amino acid sequence hydrolyzed by basic amino acid protein hydrolase [30,34,42,55–58]. When the iron in the body is in a high-concentration state, hepcidin combines with FPN protein and JAK2 on the intestinal epidermal cell membrane to form a complex, which is phosphorylated before the endocytosis of FPN. FPN is endocytosed into the cell and degraded in the lysosome after ubiquitination to reduce the concentration of iron in the blood [30,46,55,59]. In contrast, when the body is in a state of iron deficiency, hypoxia, inflammation and erythrocyte synthesis, the expression of hepcidin decreases. Some studies have shown that hepcidin can also be produced by other organs and tissue cells, such as the heart, alveolar macrophages and spleen macrophages [42,60–62]. In addition to the liver, red blood cells and macrophages participate in the iron metabolism of the body. For example, iron in red blood cells participates in the synthesis of hemoglobin, and macrophages can phagocytize aging red blood cells to release iron; therefore, macrophages could participate in iron circulation when the body is in a state of low iron concentration [42,63].

#### *2.3. Roles of Microbiota in Iron Homeostasis and Neurodegenerative Diseases*

In mammals, iron ions are absorbed mainly through the duodenum, and there is a strict regulation mechanism for iron ion absorption. Iron ions that are not absorbed into the duodenum end up in the colon cavity, which is home to a host of microbes called the gut microbiome. Iron plays an important role in the growth of intestinal micro-organisms because it plays an important role as a ferritin cofactor in redox reactions, metabolic pathways and electron transport chains of microorganisms. Therefore, the content of iron ions in the colonic lumen will affect the composition, growth and living status of intestinal microbes, and conversely, the changes of intestinal microbes will also affect the health status of the host [64]. A growing number of studies have shown that the gastrointestinal tract and the central nervous system interact through the gut-brain axis, including neuronal, immune and metabolite-mediated pathways. Preclinical and clinical studies have shown that gut microbiome plays a key role in the gut-brain interaction, and that disturbances in the composition of gut microbiota are associated with the pathogenesis of neurological diseases, especially the neurodegenerative diseases [65]. Maternal immune activation (MIA) increases the risk of autism spectrum disorder (ASD) in offspring. Dysregulation of microorganisms is associated with ASD symptoms. In lipopolysaccharide (LIP) -induced MIA progenies, MIA progenies exhibited an abnormal brain-gut-microbiome axis compared with that of the control progenies, which were characterized by social behavioral deficits, anxiety-like and repetitive behaviors, low myelination, and ASD-like microbiome [66]. Studies have shown a potential link between host microbiome (such as gut and oral bacteria), neuroinflammation, and dementia, which may be caused by bacterial invasion of the brain due to barrier leakage, toxin and inflammation factor production, or indirectly by modulating immune responses, and moreover, the composition of microbiota affected the deposition level of Aβ in the cerebral cortex of APP/PS1 mice [67], suggesting a critical role of iron in these processes.

### **3. Brain Iron Metabolism**

### *3.1. Brain Iron Absorption*

The brain is composed of neurons and glia. Ferritin is also the main iron storage protein in neurons, and neuromelanin has been found to storeiron ions for a long time. In glial cells, astrocytes and microglia synthesize L-ferritin to store iron ions, and L- and H-ferritin are expressed in oligodendrocytes [68]. Cells in the CNS are not in direct contact with nutrients, including iron ions. The blood–brain barrier (BBB) and blood–brain spinal cord barrier (BBSCB) separate the CNS from the system circulation. BBB is a special structure, which is composed of auxiliary feet of capillary endothelial cells, peripheral skin cells and astrocytes, and it strictly regulates the substances entering the CNS [69,70]. The hydrophobic BBB prevents the hydrophilic holo-TF from entering the nervous system. Holo-TF must pass across the BBB through the brain capillary endothelial cells. Holo-TF in the blood circulation binds to the TF receptor TfR1 on the luminal surface of the brain capillary endothelial cells and enters the cells. The FPN on the abluminal surface transports ferrous iron out of the capillary endothelial cells, where Fe2+ are oxidized to Fe3+ by ceruloplasmin (CP) [71,72]. CP is expressed in astrocytes and promotes the transport of FPN-exported ferrous iron [24,73]. The binding of iron ions into intercellular fluid and cerebrospinal fluid is secreted by nerve cells, especially TF, synthesized and secreted by oligodendrocytes, and choroid plexus cells, which diffuse through brain parenchymal tissue and bind to the TfR1 receptor on the surface of nerve cell membranes. After releasing iron ions, apo-TF enters the blood circulation through arachnoid villi [71,74]. FPN is regulated by hepcidin in the system, although the source of hepcidin in the brain is unknown. It may enter the brain across the BBB for iron metabolism regulation [68].

#### *3.2. Brain Iron Regulation*

The regulation of brain iron homeostasis at the cellular level involves IRPs regulating the expression of related proteins [9,75,76]. The decrease in the IRP2 expression level will lead to the imbalance of brain iron, but it has little effect on myelin iron. Mutations in genes controlling brain iron homeostasis will lead to the disorder of brain iron metabolism and affect the synthesis of myelin. It is unclear whether hepcidin plays a key role in the mediation of brain iron metabolism; whether hepcidin is synthesized in the brain or passes through the BBB after its synthesis in the liver has not been revealed. Recent results show that inflammation activates microglia and promotes the release of hepcidin by astrocytes in the model of signal cascade between inflammatory cells; this signal prevents the release of iron ions in neurons and eventually leads to neuronal death. At the same time, it will also lead to the release of anti-inflammatory and pro-inflammatory factors. Normal human microglia are not activated, and there is no intercellular signal cascade [36,72,77].

#### *3.3. Brain Iron Accumulation and Toxicity*

Iron ions accumulate in the brain with age [9,78,79]. Iron ions mainly bind to ferritin protein and substantia nigra [80–82]. The accumulation of iron ions can induce neurotoxicity through different mechanisms. The excessive accumulation of iron ions will increase the permeability of the BBB, induce inflammation, affect the redistribution of iron ions in the brain, and then change brain iron metabolism [47]. Iron ions can act as both electron acceptors and electron donors; therefore, when iron ions accumulate in the brain, they will produce reactive oxygen free radicals through Fenton and Haber–Weiss chemical reactions [41,83,84]. Free radicals are highly active substances, which may promote protein oxidation, membrane lipid peroxidation and nucleic acid modification. When the levels of ROS exceed the antioxidant capacity of organelles, this will induce oxidative stress and damage neurons [38,85,86], leading to tissue degradation in severe cases.

$$\text{Fe}^{2+} + \text{H}\_2\text{O}\_2 \rightarrow \text{Fe}^{3+} + \text{OH}^\bullet + \text{OH}^- \text{ (Fenton)}$$

$$\text{Fe}^{3+} + \text{O}\_2\text{4}^{\bullet-} \rightarrow \text{Fe}^{2+} + \text{O}\_2$$

$$\text{O}\_2\text{4}^{\bullet-} + \text{H}\_2\text{O}\_2 \rightarrow \text{O}\_2 + \text{OH}^\bullet + \text{OH}^- \text{ (Haber-Weiss)}$$

#### **4. Iron Metabolism and AD**

### *4.1. Effect of Iron Metabolism Disorder on AD*

AD is the most common cause of dementia, which is characterized by impaired cognitive function and decreased ability of learning, memory and reasoning [24,87]. It was originally described by Dr. Alois Alzheimer, a German doctor. Patients with this kind of disease exhibit strange behavioral symptoms, memory loss and motor loss. Its histopathological characteristics are amyloid plaques deposited outside the cells, and the excessive phosphorylation of tau protein related to the cytoskeleton which forms neurofibrillary tangles in the cells [88–90]. With the increase in age, iron ions in the brain tend to accumulate, especially in the cortex, globus pallidus, red nucleus, dentate nucleus and substantia nigra; however, the related molecular mechanisms are not clear at present [9,74,79]. The emerging evidence shows that iron with high redox activity is related to the deposition of amyloid plaques and the formation of nerve fiber tangles, suggesting it may be one of the main causes of AD [91–94].

The postmortem brain anatomy of AD patients showed that there was more Aβ deposition and neurofibrillary tangles in the hippocampal region of the patients [95–97]. Moreover, by detecting the level of antioxidant protein in the hippocampus and amygdala, the level of oxidative stress in these two regions was found to be much higher than other regions. Moreover, the oxidative stress caused by iron accumulation will enhance the activity of IRP1, resulting in the enhancement of iron absorption through TfR1 and the increase in intracellular free iron level by reducing the concentration of ferritin-H and ferritin-L, which further enhances intracellular oxidative stress [93,98]. Based on magnetic resonance imaging (MRI) technology [99], it was found that iron accumulation may further lead to the deposition of Aβ amyloid and the formation of neurofibrillary tangles in the brain of AD patients. Considerable studies have shown that iron metabolism disorder can affect Aβ misfolding and tau hyperphosphorylation, and the resultant oxidative stress and metal toxicity of iron ions may lead to AD [100–103].

Even more evidence supports a key role of ROS and RNS (reactive nitrogen species) in leading to AD, which are toxic and related to the formation of oxidative stress in the brain of AD patients [104]. The oxidative stress was more obvious with the increase in iron concentration, and the oxidation of protein, lipid and DNA in Aβ aggregation area was more significant [105,106]. The free radicals produced at regions of Aβ aggregation will destroy the adjacent neurons, resulting in a decline in cognitive and memory functions. The accumulation of tau protein in neurofibrillary tangles is also related to the induction of heme oxygenase-1 (HO-1). Overexpression of HO-1 can lead to the increase in iron content and accumulation of tau proteins in the mouse brain. In AD patients or patients with slight cognitive impairment, the concentration of HO-1 in the hippocampus and frontal cortex increased [86,107,108]. Increased levels of iron-bound melanin transfer protein were detected in the serum of AD patients, indicating that there may be abnormal binding of iron in the brain of AD patients. It was also found that iron ions accumulated in regions of Aβ deposition and neurofibrillary tangles formed by hyperphosphorylation of tau protein, and which were distributed in hippocampus, parietal cortex and motor cortex [93,106,109–112]. The Aβ amyloid is a segment of amyloid precursor protein (APP) cleaved by secretory enzymes [113]. APP is a transmembrane protein mainly expressed in the nervous system. At present, the physiological function of APP is not fully understood, and it may play a role in brain development, memory and synaptic plasticity [114]. In nerve cells, the concentration of iron ions regulates expression of the APP gene. The mechanism is shown in Figure 4. There is a loop ring formed by 11 bases in the 5 -UTR region of APP mRNA, which is called IRE. IRPs combine with IRE to regulate the synthesis of APP. High concentrations of iron in cells will combine with IRP1 to form iron sulfur clusters; at the same time, high concentrations of iron will also induce conformational changes in the IRE region of APP mRNA, increasing the expression of APP. In contrast, when the cell iron concentration is at a low level, IRP1 will bind to IRE and the expression of APP will decrease [97,106,115]. Under the action of different secretory enzymes in nerve cells, APP can undergo two different cleaving pathways, including the amyloidosis pathway and non-amyloidosis pathway. In the normal physiological state, APP is cleaved through the non-amyloidosis pathway, in which APP is firstly cleaved by α secretory enzyme, producing a segment called sAβPPα; then, the fragments undergo β and γ secretase cleavage to form non-toxic fragments of P3, Aβ<sup>16</sup> and Aβ17–40/42, respectively. The high concentrations of iron in cells promote the cleaving of APP through amyloidosis pathway, in which APP undergoes β and γ secretases cleavage to form Aβ1–40 and Aβ1–42 fragments. The Aβ1–42 fragment is precipitated by Ile41, and the three histidines at its N-terminal

can combine with Fe2+ to induce oxidative stress, resulting in Aβ1–42 damage to cells at deposition [93,106,109–111,115].

**Figure 4.** High concentrations of iron in neurons induce Aβ formation. The 5 -UTR region of APP mRNA has an 11-base loop called IRE. The combination of IRPs and IRE regulates the synthesis of APP. The high concentration of iron in the cell will combine with IRP1 to form iron–sulfur clusters, and make IRP1 lose the ability to bind to IRE. At the same time, high concentrations of iron will also induce conformational changes in the IRE region of APP mRNA, which increases the expression of APP; in contrast, when the iron concentration in the cell is at a low level, IRP1 will bind to the IRE of APP mRNA, resulting in the decreased production of Aβ42. Aβ1–42 aggregates to form amyloid plaques.

> In addition, the deposition of Aβ1–42 can induce the hyperphosphorylation of tau protein, although the specific mechanism is not clear. At the same time, it can also lead to the disorder of energy metabolism, the activation of immune cells and the disorder of normal function of nerve cells, resulting in cell damage and death [50]. NFTs formed by the hyperphosphorylation of tau proteins and the combination of cytoskeleton mean that the cells are unable to maintain their normal structure. Many neurons in AD patients are affected by NFTs. A large number of NFTs were found in the hippocampus of patients with AD, and the hippocampus participates in the processing of experience and precedes the storage of permanent memory. In the early stages of AD, the clinical manifestations are the decline of learning ability, the ability to form new memory and the memory storage ability. At the same time, the basal forebrain, which provides the innervation activity of cholinergic neurons for the cortex, will also be affected, resulting in the reduction in cholinergic neurotransmitters. Generally, cholinergic enzyme inhibitors can be used to treat the reduction in cholinergic neurotransmitters. A Canadian butylcholinesterase inhibitor exhibited good performance for the treatment of AD symptoms. In clinical treatment, it

has been shown that this drug is suitable for the improvement of mild and moderate AD symptoms [116].

### *4.2. Relationship between Iron-Homeostasis-Related Proteins and AD*

Oxidative stress can lead to neuronal damage; it has been observed that the disorder of iron metabolism and the expressional change in iron regulatory proteins in the iron metabolism pathway could lead to the accumulation of iron ions in the brain and induce oxidative stress, resulting in the damage of neurons [107]. Many experimental results have showed that iron accumulation in the brain of AD patients is one of the sources of brain oxidative stress, and this has a close relationship with the disorder of brain iron metabolism and some key iron homeostasis regulators, such as ferritin protein, transferrin protein, FPN, etc. [100].

### 4.2.1. Apolipoprotein E and AD

Apolipoprotein E (ApoE) is involved in the transport of cholesterol and other substances from the brain to the blood, including the discharge of Aβ protein from the brain to the blood. ApoE has three different conformations, which are encoded by *ApoE2*, *ApoE3* and *ApoE4* genes [117]. These three conformations are due to the differences in amino acid composition, resulting in differences in the structure, binding properties and multiple functions of lipoproteins. Among the three conformations, ApoE4 can lead to AD [18,118–122]. It can be seen from the extant literature that high concentrations of iron in cells will induce oxidative stress and cause damage to lipids, proteins and nucleic acids. Among them, lipid peroxidation will induce the production of 4-hydroxynonenal (4-HNE) molecules with high activity and neurotoxicity. It can combine with cysteine residues, lysine residues and histidine residues to reduce its damage to other molecules. Compared with ApoE2 and ApoE3, ApoE4 lacks cysteine amino acid and cannot clear HNE, resulting in the oxidative modification of proteins in neurons and neuronal death, increasing the risk of AD [123,124].

### 4.2.2. Ferroptosis and AD

Ferroptosis is an iron-dependent programmed cell death, which can lead to many diseases [125]. Ferroptosis was first described by Dixon in 2012, and is characterized by the accumulation of lipid reactive oxygen species. The experimental results show that GPX4 knockout mice exhibit neuronal necrosis, which will become more serious due to the lack of vitamin E (iron death inhibitor) in food. In contrast, inhibiting iron death can effectively improve the symptoms of AD. GPX4 is an anti-peroxidase that inhibits lipid peroxidation [101,126]. Moreover, lipid peroxidation products and 4-HNE in the AD brain have been significantly increased, indicating that ferroptosis will increase the risk of AD [127–129]. Iron induces oxidative stress, directly affecting lipids, DNA and proteins. Lipid peroxidation and iron metabolism disorder and accumulation in AD brain are also necessary conditions for ferroptosis [130]. In addition, iron ions interact with Aβ and Tau to induce ROS, which also leads to ferroptosis [101,129].

### 4.2.3. Iron Homeostasis Key Regulators and AD

Through the utilization of Western blot technology, researchers have found that in comparison with ferritin protein in the normal brain, the expression levels of ferritin protein in the brains of AD patients were increased significantly, including L-ferritin and H-ferritin proteins [36]. The ELISA results showed that the concentrations of H-ferritin and L-ferritin in the hippocampus of AD patients were three times higher than those in normal human brains. Moreover, the increases in H-ferritin and L-ferritin protein concentrations were not consistent with the increase in iron concentration, which was about 50% of the increase in iron concentration. Compared with the normal brain, the expression levels of ferrous oxidase CP increased significantly [131]. Results obtained from immunohistochemical experiments showed that the expression levels of transferrin proteins in the AD brains were also found to be increased compared with those in the normal brain [100].

However, by using Weston blot technology, it was found that the expression levels of DMT1 and FPN decreased in the AD brains compared with those of normal human brains. Due to the abnormal expression of genes related to iron metabolism, iron accumulates in AD brain and induces oxidative stress, which may damage brain neurons [36,131].

### 4.2.4. Furin and AD

Furin is associated with iron and Aβ metabolism [132]. Low concentrations of iron enhance furin enzyme activity, whereas high concentrations of iron reduce furin enzyme activity. Furin can enhance the activity of α secretory enzymes, and high concentrations of iron in cells reduce furin enzyme activity, leading to the amyloidosis pathway of APP cleaving. Recent experimental results also showed that the expression levels of furin mRNA in the brain of AD patients are lower than those of normal human brains [9,110,133].

### **5. Strategies for Treating AD**

### *5.1. Iron Chelation in the Treatment of AD*

Iron chelation strategy is the most direct method for limiting and redistributing iron in the system. At present, the most commonly used chelating agents are deferoxamine, deferrone and ferrite [2,134,135]. Deferoxamine is a chelating agent, recently found to exhibit good clinical manifestations. Although these chelating agents can improve the symptoms of AD caused by iron excess to a certain extent, they can also have toxic effects on the human body, such as allergic reactions, liver and kidney failure, etc. [1,136].

#### *5.2. Regulating Iron Metabolism Pathway Proteins to Improve AD Symptoms*

Fursultiamine is a small molecular substance called thiamine tetrahydrofuran disulfide, which can bind to cys326 amino acid residues of brain FPN and protect hepcidin from the endocytosis of FPN, thus improving the efflux of brain iron through this ferrous transporter. However, fursultiamine has limited functions in the body, because it can be quickly converted to ammonium sulfate, resulting in reduced iron contents in the body [137].

The anti-ferroportin antibody ly2928057 was successfully tested *in vitro*, and it has also been tested for its potential to effectively reduce iron concentrations in vivo by interfering with the potential regulatory mechanism of hepcidin. The specific mechanism is to regulate the BMP6 (bone morphogenetic protein 6)–SMAD signal pathway and prevent the binding of BMP6 to its receptor BMP6R [1,138]. Another way is to block the phosphorylation of SAMD with doxomorphine, so as to reduce the production of hepcidin induced by BMP6R [139]. The body has its own regulatory mechanism; therefore, the treatment of FPN or hepcidin interference is a great challenge, which is not conducive to long-term treatment.

Similar to glutathione peroxidase, ebselene, a drug containing selenium, also exhibits antioxidant effects. This drug can inhibit the absorption of iron ions through DMT1; however, it can cause cardiomyopathy [140]. Recent studies have shown that pyrazole derivatives and benzyl isothiourea have inhibitory effects on DMT1 both in vitro and in vivo [141].

### *5.3. Antioxidant Therapy Improves AD Symptoms*

Brain iron excess induces oxidative stress through Fenton chemical reactions, which cause damage to protein, lipid and DNA [7], and lead to ryanodine-receptor-mediated calcium release under the stress, resulting in neurotoxicity [105]. Small molecular substances have been designed for ROS scavenging at fixed sites. These kinds of antioxidants enter the mitochondrial matrix driven by the mitochondrial intimal potential to scavenge active free radicals in the matrix [142]. In addition, antioxidants in food, such as tea polyphenols, can effectively improve AD symptoms through scavenging oxygen free radicals, chelating iron ions and their anti-inflammatory effects [143–145]. Other native neuroprotective compounds or species include resveratrol, curcumin, pinocembrin, caffeine, the combination of Panax ginseng, ginkgo biloba, crocus sativus [146–150]. The anti-inflammatory and

antioxidant properties of catechins in tea have been reported in vivo and in vitro, with potential for the prevention of AD symptoms [151].

Acetylcholinesterase inhibitors (AChEIs) have also been found to exhibit antioxidant effects. In 2010, Sinem et al. showed that ACHEIs can reduce the levels of lipid oxidation, blood markers and nitric oxide in AD patients [152]. ACHEI is the main drug for the treatment of AD, but it also has certain limitations [153].

#### **6. Conclusions and Prospect**

Iron is a rich metal element in the earth's crust. The unique redox properties of iron allow for efficient electron transfer, which is beneficial to many diverse biological reactions [154]. However, when iron metabolism in the body is unbalanced, such reactive properties of iron may also promote the generation of ROS, which will lead to the excessive accumulation of iron ions in the body [155,156]. As a result, there are fine regulatory mechanisms for iron absorption, storage and distribution in organisms. The excessive accumulation of iron induces oxidative stress reactions, which, in large doses, can be damaging to intracellular systems, including the tissues and organs of the body. Moreover, iron plays an important role in the formation of a myelin sheath in the brain and aerobic respiration in mitochondria. When brain iron metabolism is disordered, iron will be enriched in different regions of the brain, and the enriched iron will cause oxidative stress, mediate APP undergoing the amyloidosis pathway, and finally lead to the development of AD. In AD, oxidative stress caused by brain iron accumulation promotes the deposition of amyloid protein and the hyperphosphorylation of tau, which causes damage to neurons, resulting in declines in motor, cognitive and memory functions, etc. [133]. Although using iron-chelating strategies has achieved some positive results for improving the symptoms of AD, there is still much research needed in order to translate the research into practice for the clinical treatment of AD.

Nevertheless, there have been few studies on iron-reducing strategies in AD patients through genetic methods, and excessive emphasis has been put on the amyloid-reducing strategies, which have been disappointing thus far. Given that more and more ironchelating compounds have potential disease-improving effects, as well as the availability of biomarkers of iron load in MRI and cerebrospinal fluid, there is considerable room for exploring this type of treatment to avoid its side effects as far as possible. In addition, genetic studies on the regulation of some key genes in iron homeostasis in model animals have shown potential for more effective and precise treatment [14,157].

Furthermore, AD is characterized by the progressive dysfunction and death of thecortical and hippocampal neurons; the main hypothetical mechanisms are the hyperphos-phorylation of tau protein to form NFTs and the deposition of Aβ protein to form SPs [158,159]. However, a large number of clinical trials of drugs based on these two hypotheses all over the world have ended in failure; there is currently no effective treatment method. In addition, these two assumptions are facing increasing challenges [160,161]. In fact, the involvement of iron in the pathogenesis of AD has been widely accepted. Iron not only aggravates the accumulation of toxic Aβ and hyperphosphorylated tau, but also directly induces neuronal oxidative damage [162]. Considering the particularity and importance of iron role in the process of ferroptosis, it is essential to uncover how does ferroptosis play in the molecular pathophysiology of AD in the future research, which may provide new insights into the disease [163,164] and new ideas for the treatment [101]. Combing with the recent finding of a potential link among iron, host microbiome and AD, therefore, by deeply studying the mechanism of iron metabolism in the body and brain, it is expected to find new effective targets and therapeutic measures to improve or cure the disease.

**Author Contributions:** Concept and design: M.L. Writing, review of manuscript: Y.P. and M.L. Revision of manuscript: M.L., Y.P. and X.C. Display item design: Y.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Fundamental Research Funds for the Central Universities, Beijing Municipal Natural Science Foundation (7202129), the Class B Breeding Program of Special Projects for Leading Science and Technology of the Chinese Academy of Sciences (XDPB16), the National Natural Science Foundation of China (31571042), the Key Basic Research Project of Applied Basic Research Program of Hebei Province (18966315D), and One Hundred Outstanding Creative Talents Support Program of Hebei (BR2-218).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **Abbreviations**


### **References**

