Introduction

AD is a neurological condition characterised by progressive decline in cognition, with concomitant functional decline. The main pathological hallmarks of Alzheimer’s disease are the aggregation of beta amyloid peptides into extracellular plaques and hyperphosphorylated tau proteins into intracellular neurogenic fiber tangles, accompanied by neuroinflammation, gliosis and neurodegeneration (1). Based on 2020 Alzheimer’s Disease Facts and Figures (2), more than 5.8 million Americans have AD dementia today, which is the most common form of dementia worldwide (2). Despite remarkable advances in unraveling the biological underpinnings of AD during the last 25 years, no drugs have been found that can slow the progression of AD while promising preclinical trials have repeatedly ended in failure to translate into treatments for AD patients (3).

Cell death is a critical phenomenon for physiological functions such as immunity, development, and tissue homeostasis. Programmed cell death is a distinct biochemical pathway in which cells die to elicit various physiological outcomes and functions as a defense mechanism against various infections, diseases, and microorganisms (4). Programmed cell death is well known as a type of cell death during the earlier period. Recently, regulated pathways of cell death, including necroptosis, pyroptosis, autophagy, and ferroptosis (5), have been discovered that have unique biological functions and pathophysiological characteristics (6, 7). Cell death is inseparable from the progression of inflammation in AD. Regardless of the types of death, they all promote the occurrence and development of inflammation. Hyperphosphorylated tau protein aggregates into NFTs, which is a major pathological hallmark of AD. It is proved that hyperphosphorylation of tau may result in dysfunction of autophagy, ferroptosis and pyroptosis (8–10). On the contrary, the damage of cell, including autophagy, ferroptosis and pyroptosis, can exacerbate the progression of tau hyperphosphorylation, which aggravates the inflammatory state of AD.

Microglia are immune cells in the brain that participate in maintaining immune defense of the nervous system (11, 12). Microglia are modified macrophages that compose approximately 10%–12% of total cells (13) and are considered instigators of damage and guardians of brain homeostasis, playing vital roles in both neuroprotection and neurodegeneration (14). However, there is no systematic description of the effect of programmed microglial death on the nervous system.

In this review, we evaluate the association between programmed microglial death and AD, and we discuss the role of NFTs and tau in explaining how cell death might contribute to AD. The findings are contextually framed by evidence that activation of the innate immune system alters central nervous system mechanisms and increases cell damage, possibly driving onset and progression of AD. Inhibiting the progressive of cell death could be targets for the prevention of AD.

Autophagy

Three types of autophagy have been reported: macroautophagy, micro autophagy, and chaperone-mediated autophagy. Among them, macro autophagy is the most widely studied and well described of the three types (15). Some studies have shown that macro autophagy is the main pathway of intracellular degradation (16). Autophagy, which acts as a protective mechanism, maintains normal cellular function and homeostasis by degrading engulfed substances (17). Autophagy is involved in the degradation and elimination of damaged, denatured, aged or dysfunctional organelles, denatured proteins, and other biological macromolecules through a process that provides the energy necessary for cell survival and repair (18). Autophagy is a cellular catabolic process involving the sequestration of misfolded proteins and damaged cytoplasmic organelles by a double-membrane structure known as an autophagosome and the degradation of the engulfed contents via fusion with a lysosome (12). Therefore, this autophagy-lysosomal pathway plays an important role in maintaining normal cellular function and intracellular homeostasis. Accumulating evidence indicates that autophagy is closely associated with inflammasomes, and these factors mutually influence each other. In neurodegenerative diseases, microglial autophagy is impaired and downregulated (19). The transformation from the M1 to the M2 phenotype can protect the body from excessive inflammatory injury, and one of the mechanisms affecting macrophage polarization is autophagy (20). The phosphatidylinositol 3-kinase (PI3K)-mTOR pathway is important in regulating macrophage polarization. Activation of the

PI3K-mTOR pathway can increase M2 polarization while inhibition of PI3K or mTOR exerts the opposite effect (21). However, microglial autophagy transforms microglia from a proinflammatory state to a favorable anti-inflammatory state and inhibits NOD-like receptor family, pyrin domain containing 3 (NLRP3) inflammasome-mediated inflammatory responses, acting as negative regulators (22). Ma’ (23) results indicate that the overactivation of autophagy is responsible for the M1 polarization of microglia and promotes microglial apoptosis. Emerging evidence indicates that overactivation of the NLRP3 inflammasome in microglia is a driving factor that exacerbates pathology and ultimately accelerates neuronal death and the progression of neurodegenerative diseases (24, 25). In mammalian cells, microglial autophagy has been demonstrated to be critical for microglial activation in vitro, and the inhibition of microglial autophagy results in the upregulation of proinflammatory cytokines, causing increased M1 microglial activation (26–28), (Figure 1).

Figure 1
figure 1

Autophagy pathway

The energy sensors mTORC1 and AMPK control autophagy activation via the ULK1 complex. Various stress conditions stimulate ULK1 complex phosphorylation of the VPS34-Beclin-1 complex. Following activation, the ULK1 and PI3KC3 complex regulate the formation of the autophagosomes. The resulting LC3-I is activated and binds on the autophagosome and is esterified to form LC3-II. Finally, the mature autophagosome fuses with lysosomes to autolysosomes, and regulates the polarity of microglia.

Full size image

Cognitive Decline

One of the reasons causes cognitive decline of AD is the sequential functional modifications of the mitochondrial dysfunction, including inflammation, impaired energy metabolism, oxidative stress and synaptic dysfunction (29, 30). The significance of autophagy dysfunction in AD pathophysiology is now appreciated due to evidence reporting dysfunctional autophagy in the AD leading to neuronal degeneration (31–33).

In autophagy-deficient AD mice, Nilsson et al, reported an important reduction associates with intraneuronal Aβ accumulation increasing in Aβ metabolism, which further enhances neurodegeneration and memory impairment (34). Abubakar’s research revealed that improved memory function, demonstrated via Morris water maze test, was associated with reduced Aβ levels and related pathology (35). Glatigny et al. showed that the induction of novel stimuli increases autophagy by fostering synaptic plasticity in hippocampal neurons. In addition, they showed that restoring autophagy levels in old hippocampi reverses memory deficits (36).

Given the extensive observations showing autophagic alterations in the brain of AD patients or in different mouse models of AD, it is now well accepted that improving of the autophagic flux ameliorates cognitive impairment in AD mouse models (37–39). Restoration of neuronal mitophagy was found sufficient to ameliorate the cognitive decline in an AD mouse model by preventing synaptic failure (40). Thus, the cognitive decline of AD is related to the autophagy closely, which may inhibit the progress of cognitive decline, in brain.

NFTs and tau

Extracellular amyloid plaque deposits, composed of Aβ peptides, as well as intracellular accumulation of NFTs, consisting of hyperphosphorylated tau (p-tau) protein are regarded as the characters of AD (41). Major pathological hallmarks of AD include intracellular deposition of neurofibrillary tangles (NFTs), which were associated with paired helical filaments (PHF), p-tau, and extracellular accumulation of Aβ peptide in the senile plaques (42, 43). Data show that Aβ peptides and tau protein accumulation, the principal hallmarks of AD, can be influenced due to autophagy dysregulation (44). Studies suggest that dysfunction of the autophagy-lysosome system promotes the formation of tau oligomers and accumulation of insoluble tau species (9).

It has been reported that tau secretion is promoted by autophagy inducers and knockdown of beclin-1 (45). Besides, the accumulation of hippocampal phosphorylated tau is responsible for abnormal mitophagy function, mitochondrial dynamics hippocampal-based learning and memory impairments in tau mice (46). It has been reported that the phosphorylated tau can also interact with VDAC1 and Drp1, likely leading to mitochondrial dysfunction and abnormal mitophagy, ultimately possibly leading to damage and cognitive decline (47, 48).

Aβ

Autophagy is closely related to the metabolism of Aβ, and autophagy is an important regulator of its production and clearance. The precursor protein of Aβ, APP, and the γ-secretase that cleaves APP are highly enriched in autophagosomes. Many studies have shown that Aβ secretion is significantly reduced in autophagy-deficient mouse models (49). This result was also verified in AD patients, where patients with severely impaired autophagy showed a significant decrease in learning ability and cognitive function (50). Induction of autophagy recovery accelerates Aβ clearance and restores cognitive function. Autophagy can influence Aβ clearance by affecting multiple stages of Aβ (51). It was shown that healthy microglia can wrap and degrade Aβ through autophagosomes, but the expression of autophagy-related proteins NBR1 and ATG7 was reduced in microglia of 5XAD mice, and the degradation of Aβ was impaired in 5xFAD microglia, further suggesting the necessity of autophagy in microglia for the degradation of Aβ (52).

Ferroptosis

Ferroptosis, which mainly occurs in the brain (53, 54), is defined as an iron-dependent form of regulated cell death that occurs through the lethal accumulation of lipid-based reactive oxygen species (ROS) when glutathione (GSH)-dependent lipid peroxide repair systems are compromised (55). Cellular accumulation of lipid peroxidation products and lethal ROS are characteristics of ferroptosis and ultimately result in oxidative stress and cell death. The peroxidation of proteins, nucleic acids, and lipids is promoted when intracellular iron accumulation generates ROS and causes oxidative stress via the Fenton reaction, which is the key process that propagates ferroptosis (56–58). Morphologically, ferroptosis is mainly characterized by shrunken mitochondria with increased membrane density and normal-sized nuclei, as well as diminished or vanished mitochondrial cristae and ruptured outer membranes (59). A wealth of evidence underscores the tight link between ferroptosis and the nervous system. There are three hallmarks of ferroptosis that were highlighted by Dixon and Stockwell: the loss of lipid peroxide repair capacity by the phospholipid hydro peroxidase GPX4, which was identified as a key regulator of ferroptosis in cancer cells (60); the availability of redox-active iron; and the oxidation of PUFA-containing phospholipids (61). Among them, it is accepted that the fundamental characteristic of ferroptosis is lipid peroxidation (62). As GPX4 is a central protector against the formation of lipid hydroperoxides (63), and its degradation is a mandatory signaling event in the execution of ferroptotis cell death (60, 64, 65). Ferroptosis widely exists in various parts of the central nervous system, such as the cerebral cortex, hippocampus, striatum, and spinal cord (59). Ferroptosis has been implicated in several neurological diseases, such as neurodegenerative diseases, hemorrhagic stroke and ischemic stroke (66, 67).

Recent studies have shown that iron accumulation in microglia also contributes to microglial dysfunction and Aβ accumulation (68). (Figure 2)

Figure 2
figure 2

Ferroptosis

System xc- transports intracellular Glu to the extracellular space and extracellular Cys2 into the cell, which is then transformed into Cys for GSH synthesis. Excess irons are the basis for ferroptosis execution. Circulated iron was combined with transferrin in the form of Fe3+, and then it entered into cells by TFR1. Iron in Fe3+ form was deoxidized to iron in Fe2+ by iron oxide reductase STEAP3. Ultimately, Fe2+ was released into a labile iron pool in the cytoplasm from the endosome mediated.

Full size image

Cognitive Decline

Several studies have observed the excess brain iron accumulation was associated with accelerated cognitive decline in AD patients (69). Iron rises in the brain with aging and may be pathological because it is associated with cognitive decline prior to disease (70, 71). The increasing of iron occurs as early as the mild cognitive impairment stage of AD and contributes to longitudinal outcomes (72, 73). It was found out that there is a large amount of activated iron-containing microglia around Aβ plaques (74, 75). The brain iron burden is elevated in AD patients, combined with Aβ positron emission tomography (PET), which indicates that brain iron load is positively associated with Aβ deposition-related cognitive decline, suggesting that cognitive function damage may exacerbate due to iron combine with Aβ. The pathologic mechanism could be that iron promotes the production of free radicals and oxidative stress and possibly also involves ferroptosis (76, 77).

The maintenance of glutathione GSH is a key antioxidant element in brain redox homeostasis (78). In the AD model, N-acetyl cysteine (NAC) can protect neurons’ function and improving learning and memory deficits via increasing GSH levels (79). A recent study indicated that deficiency of ferroportein (Fpn) in principal neurons in neocortex increased iron levels and induced AD-like hippocampal atrophy and memory deficits, while restoring Fpn expression could effectively ameliorate the memory loss and ferroptosis in AD model mice (80).

NFTs and tau

Aging and changes in iron metabolism are associated with the development of Aβ plaques and NFTs (81). Iron and ferritin are found within plaques, NFTs, and blood vessels in AD (82). Svobodová et al. (83) demonstrated in an APP/PS1 transgenic mice model that free iron and ferritin accumulation follows amyloid plaque formation in the cerebral cortex area. Actually, iron deposition has been involved in the misfolding process of the Aβ plaques and NFTs (8). Additionally, iron is related to the development of tau protein, which is present through the induction and regulation of tau phosphorylation (8, 84). It is possible that this surface trafficking of APP may be impaired by the hyperphosphorylation and aggregation of tau (thus lowering the soluble fraction of tau) during AD pathogenesis (8, 85). Tau loss preceded iron accumulation, and APP treatment lowered iron (86). The evidence suggests that iron interacts with tau to cause neurodegeneration in AD and related conditions; conversely, tau maintains cellular iron homeostasis, however a putative role of an iron-tau interaction in ferroptotis stress needs further investigation.

Aβ

In AD, neuroinflammation exacerbates Aβ deposition. Lipid metabolites in iron death trigger inflammation, which in turn mutually promotes iron death by increasing iron deposition. Iron is involved in Aβ plaque deposition, and iron accumulation accelerates Aβ plaque deposition (69). Iron dysregulation increases the production of ROS and is one of the prominent manifestations of AD pathology. Excessive increase in ROS promotes the development of Aβ. As reported by Kristin et al. Aβ induces strong ROS production in BV2 microglia via NADPH oxidase. Intracellular iron depletion inhibits Aβ-induced ROS. Aβ plaques are surrounded by microglia overexpressing HO-1 (87). Fernández-Mendívil’s study showed that (88) overexpression of HO-1 in microglia under inflammatory conditions leads to toxic accumulation of iron leading to iron death. In addition, mitochondrial contraction serves as one of the unique markers of iron death, and mitochondrial dysfunction is associated with Aβ accumulation. The level of iron in the brain is closely linked to both translation and processing of APP.

Pyroptosis

Pyroptosis is a proinflammatory form of programmed cell death that triggers an inflammatory response upon infection or other stimuli (89). Pyroptosis features rapid plasma membrane rupture and the release of proinflammatory intracellular contents such as interleukin 1β (IL-1β) and interleukin 18 (IL-18) (90). Pyroptosis is regarded as a critical host defense mechanism against intracellular pathogenic bacteria by releasing these organisms into the extracellular environment, where they can be killed by neutrophils (91–93). Pyroptosis is dependent on inflammatory caspases (caspase-1 and caspase-4/5/11) and is accompanied by inflammation. Canonical pyroptosis is executed by cleaved caspase-1, which not only causes cell lysis but also mediates the proteolytic cleavage and release of IL-1β and IL-18 (94). Chang’s (89) data demonstrated that intensive pyroptosis and increased caspase-1 activity indeed occurred in activated microglia after cardiac arrest, along with elevated levels of IL-1β and IL-18. This finding shows the inseparable connection between pyroptosis, caspase-1, IL-1β, and IL-18. Furthermore, recent studies have reported that gasdermin D (GSDMD) is the executioner of pyroptosis (95). After being cleaved by caspase-1, the GSDMD N-terminal domain (GSDMD-N) can form pores in the plasma membrane. Holes in the cell membrane cause the loss of cell integrity. IL-1β and IL-18 are released through these pores, thereby perpetuating the inflammatory response.

In addition, ATP acts as a canonical activator that induces NLRP3 inflammasome activation in macrophages, leading to caspase-1/GSDMD-mediated pyroptosis. Zeng’s (93) study showed that ATP was able to induce alternative pyroptosis in macrophages in which NLRP3-mediated rapid pyroptosis was blocked, highlighting another interplay between the pyroptosis and apoptosis pathways. When the oxygen level decreases, mitochondria produce a large amount of ROS in response to hypoxia, which is a key stimulus that promotes the activation of NLRP3 inflammasomes. NLRP3 inflammasome activation can also activate caspase-1, resulting in pyroptosis. Poh et al. (96) provided evidence to support that the inflammatory response induced by inflammasome activation through proinflammatory mediators such as both IL-1β and IL-18, damage-associated molecular patterns (DAMPs) (i.e., HMGB1 and IL-1α) and inflammasome components released into the extracellular environment caused pyroptosis in microglial cells (Figure 3).

Figure 3
figure 3

Pyroptosis

In the canonical model of pyroptosis, inflammasome sensor proteins recognize cellular stressors, including those from bacteria, viruses, toxins etc. These stressors activate inflammasome sensors indirectly, such as NLRP3. NLRP3 subsequently activates caspase-1 via the adaptor protein ASC. Caspase-1 processes and activates IL-1β and IL-18, and also cleaves GSDMD to release the membrane pore-forming GSDMD-N domain. GSDMD-N pores promote the release of activated IL-1β and IL-18. Cytosolic LPS binds Caspase-4/5/11 to trigger their cleavage of GSDMD, but not IL-1β and IL-18.

Full size image

Cognitive Decline

In the AD brain and AD transgenic mice, the expression of several pyroptosis-related proteins, including NLRP1, NLRP3 and caspase-1, is increased, and inhibiting pyroptosis alleviated the recognition dysfunction of APP/PS1 mice (97, 98). NLRP3 deficiency in AD model results in the rescue of memory deficits and a decrease of Aβ deposition (99). So, NLRP3 inflammasome-mediated pyroptosis may provide a progressive memory loss of AD. It has been reported that NLRP3 inflammasome activation is caused by the action of cathepsin B released from the lysosome rather than the direct actions of Aβ (100). These findings were confirmed by a separate study showing that cathepsin B inhibitors improve the memory deficit in transgenic AD mice (101).

Recently, studies have shown that neuroinflammation inhibition can regulate the cognitive function of AD (102). Several studies have demonstrated that pyroptosis plays an important role in mediating the occurrence of neuroinflammation (103). To evaluate whether disturbing pyroptosis might affect cognitive function, Li et al. assessed spontaneous alternation in the Y-maze as a measure of spatial memory, and a significant improvement in spatial memory disorder was shown in the caspase inhibitor treated models (10).

NFTs and tau

Numerous studies have pointed out that overexpression of IL-1β aggravates AD pathogenesis, owing to tau hyperphosphorylation (104), which inhibits long-term potentiation and affects synaptic plasticity (105). IL-1β can increase the phosphorylation of tau proteins (106). IL-18 shares structural similarities with proteins in the IL-1 family, which can also modulate the hyperphosphorylation of tau protein by increasing the expression of kinases involved in tau phosphorylation (10). Recently, a study by Ising and co-workers showed that NLRP3 acted as a link between Aβ plaques and NFTs formation (25). These authors demonstrated that the injection of Aβ-containing APP/PS1 brain homogenates induced tau hyperphosphorylation in tau22 mice, but not in tau22 mice deficient for NLRP3, indicating that NLRP3 was an important mediator of Aβ-induced tau pathology. What’s more, li’s study also discovered that suppressing the activity of caspase resulted in a remarkable reduction of pyroptosis-related proteins and significantly inhibited the hyperphosphorylation of tau proteins (10).

Aβ

A previous study described that Aβ accumulates in AD brain to form characteristic plaques, which then activate NLRP3 inflammatory vesicles that trigger cellular scorching via the NLRP3/caspase-1/GSDMD signaling pathway, ultimately leading to elevated levels of inflammatory factors IL-1β and IL-18 (107). In addition, Aβ secretion of IL-1β is dependent on NLRP3, ASC and caspase-1 activity and requires release of histone B from damaged lysosomes (108). In addition, aberrant activation of microglia-specific NLRP3 leads to microglia Aβ phagocytosis dysfunction during the pathology of AD. In a mouse model of AD, intracellular activation of NLRP3 inflammasomes was observed to promote M1 phenotype microglia activation, leading to Aβ accumulation and higher cognitive impairment[109]. Inhibition of NLRP3 inflammatory vesicle activation reduces Aβ-induced neuroinflammation in microglia and thus treats AD (110).

Conclusion and Perspectives

Microglial death is a complex process involving physiological and pathological activities in the body and is regulated by a variety of factors. And AD, as one of the most commonly diseases around the world, has bothered human all the time. We are just beginning to understand how microglia function in health and are altered in AD. The way of microglial death perhaps plays a significant role in the proceeding of AD. In the early stage of the disease, the excess brain iron accumulation is occurred with accelerated cognitive decline in AD. During the development of AD, microglial autophagy plays an important role in the removal of misfolded protein aggregates, the clearance of damaged mitochondria and their resultant ROS, and the degradation of the NLRP3 inflammasome or its components. Additionally, NLRP3 inflammasome-mediated pyroptosis may also provide a progressive memory loss of AD. Studies have indicated that the targeted inhibition of NLRP3 inflammasome overactivation at different levels by NLRP3 inflammasome inhibitors or autophagy inducers could inhibit the occurrence or development of cognitive decline caused by AD. Intracellular accumulation of neurofibrillary tangles (NFTs), consisting of p-tau protein are regarded as the characters of AD. It is reported that iron deposition has been involved in the misfolding process of the Aβ plaques and NFTs. The accumulation of phosphorylated tau is responsible for abnormal mitophagy function, mitochondrial dynamics hippocampal-based learning and memory impairments in tau mice. The overexpression of IL-1β aggravates AD pathogenesis, owing to tau hyperphosphorylation. On the contrary, IL-1β can increase the phosphorylation of tau proteins, which would increase the severity of AD. Studies also discovered that suppressing the activity of inflammation resulted in a remarkable reduction of pyroptosis-related proteins and significantly inhibited the hyperphosphorylation of tau proteins. These microglial death modes are inextricably linked with the phenotypic conversion of microglia, energy metabolism, and the occurrence and development of neurological diseases, like AD.

The modes mentioned in this review of cell death attract widespread attention. To sum up, programmed cell death might be a risk factor for AD. Programmed cell death may be a biologically plausible mechanism for AD pathogenesis, and the effects of cell damage about AD provide a framework with which to understand how microglia might contribute to the early, and possibly the later, course of AD. Therefore, further understanding of microglial death can provide more reliable evidence for the clinical treatment of AD.