Microglia are the resident mononuclear phagocytes in the central nervous system (CNS), belonging to the non-neuronal glial system and supporting neuronal functions. They are originated from the myeloid and broadly distributed throughout the brain and spinal cord (Ginhoux et al., 2013). During development, microglia sculpt immature neuronal circuits by phagocytizing and eliminating synaptic structures (axons and dendritic spines) in a process known as synaptic pruning (Paolicelli et al., 2011; Schafer et al., 2012). In CNS parenchyma, microglia account for 5%–20% of the total glial cell population and 10% of the non-neuronal cells (Lawson et al., 1990; Perry, 1998). They play fundamental roles in the control of immune responses and the maintenance of CNS homeostasis (Perry et al., 2010; Salter and Stevens, 2017). They also carry other exclusive characteristics, such as unique motility with their fine processes, which can scan the whole brain parenchyma every few hours (Davalos et al., 2005; Nimmerjahn et al., 2005).

Under physical conditions, microglia phagocytose synaptic structures to remodel the presynaptic environment and release soluble factors in the mature and aging brains (Tremblay et al., 2011; Su et al., 2016). However, under certain pathological conditions or during aging, this function can go beyond physiological control, resulting in an inflammatory state and the progression of neurodegeneration. Understanding age-related microglia response is important for the cure of the severe neurodegenerative diseases (Fuger et al., 2017; Salter and Stevens, 2017).

Neurodegenerative diseases are progressive disorders, characterized by the age-dependent loss of neuronal structures or functions that leads to neuronal death (Ross and Poirier, 2004). These diseases, including AD and PD, are increasingly being realized to share common cellular and molecular mechanisms. They are featured by selective neuronal vulnerability in specific brain regions and intracellular or extracellular deposits of abnormal proteins in or between neurons and other cells (Wong et al., 2002; Nussbaum and Ellis, 2003; Selkoe, 2003; Ross and Pickart, 2004). Emerging evidence points to the inflammatory responses in the CNS as a major cause and a common feature in neurodegenerative diseases. Pro-inflammatory cytokines are chronically increased in the aging brain (Godbout et al., 2005; Sierra et al., 2007), which is correlated with characteristics of the occurrence of chronic neurodegenerative diseases.

In this review, we summarized the current knowledge of microglial autophagy in the chronic neurodegeneration process and discuss the roles of microglial autophagy in disease pathogenesis and its potential as a new therapeutic target.

In general, microglia participate in autophagy by phagocytosis in the CNS. Because of the critical role of autophagy in protein and organelle quality control (Mizushima et al., 2011; Yamamoto and Yue, 2014), the impairment of autophagy will result in accumulation of aggregated proteins and damaged organelles, which are common pathological hallmarks in AD and PD. Accumulating evidence indicates that the autophagy machinery in microglia can contribute to the emergence, acceleration, or amelioration of CNS disease conditions (Keller et al., 2020). So far, two specific mechanisms appear to be relevant to CNS pathology: activation of the inflammasome and increase of autophagy protein-mediated endocytosis/phagocytosis. Autophagy pathways are implicated in the regulation of inflammasome function at various steps by removing triggering agents, inflammasome constituents, or downstream effector molecules (Saitoh et al., 2008; Harris et al., 2011; Liu et al., 2016). As the major cellular component of the innate immune system in the brain, microglia have been found to execute pivotal functions during CNS homeostasis and pathology (Heneka et al., 2014, 2018).

In response to microenvironmental stimuli and factors, microglia can change morphologically and functionally to migrate to the injury sites. They can recognize harmful entities through Toll-like receptors (TLRs) and trigger receptor expression in myeloid cells 2 (TREM2; Blander and Medzhitov, 2004; Takahashi et al., 2005) and signaling pathway activation that leads to reorganization of new phagosomes (Koizumi et al., 2007; Yao et al., 2019). Canonical autophagy pathways require a multistep assembly mechanism. In most cells, ATG proteins regulate the formation of the autophagosome that involves multiple proteins and molecular changes. Initiation of autophagy requires the association of the unc-51-like kinase 1 (ULK1) kinase complex (Shang and Wang, 2011), the phosphorylation of the transmembrane protein ATG9 and the class III phosphatidylinositol 3-kinase (PI3KC3) complex (Young et al., 2006; Russell et al., 2013; Zhou et al., 2017), the recruitment of phosphatidylinositol 3-phosphate (PI3P) effector proteins, and the conjugation of ubiquitin-like human ATG8 to phosphatidylethanolamine (PE) on a nascent phagophore (Dooley et al., 2014). Finally, the mature autophagosome fuses with lysosomes to autolysosomes, resulting in the degradation of cargos (Julg et al., 2020). So far, this kind of autophagy pathway has been found in many important cellular functions, including pathogen defense (Gutierrez et al., 2004; Ogawa et al., 2005), pro-inflammatory response (Saitoh et al., 2008; Houtman et al., 2019), and intracellular aggregates (Choi et al., 2020). In noncanonical autophagy processes, autophagy processes can be deployed to fulfill functions without involving lysosomal delivery of cytosolic cargo. For example, BECN1 is required for conjugating LC3 to the phagosomes or endosomes during this process, and the phagocytic ingestion and subsequent degradation of Aβ plaques (Sanjuan et al., 2007; Heckmann et al., 2019). Deletion of BECN1 in BV2 microglial cells in APP transgenic mouse brain slices resulted in insufficient phagocytosis of Aβ (Lucin et al., 2013). In the meantime, microglia from human AD brains exhibit reduced BECN1 (Lucin et al., 2013). So far, studies related to the canonical autophagy pathways have been widely reported in microglia, while studies related to the noncanonical process still need to be further explored.

During the development of AD, Aβ plaques form in the brain of patients as a result of accumulation of Aβ peptides resulting from the cleavage of amyloid precursor protein (APP) by γ-secretase and β-site APP cleaving enzyme 1 (Hampel et al., 2020). In AD brain, alterations in AMPK signaling are a central issue (Vingtdeux et al., 2010; Zimmermann et al., 2020). As a homeostasis sensor, AMPK activates microglial autophagy in the presence of Aβ, consequently leading to their lysosomal degradation (Cho et al., 2014; Julg et al., 2020). The activated microglial cells are predominantly located surrounding Aβ plaques (Frautschy et al., 1998; Nicoll et al., 2003; Zotova et al., 2013). Many studies have found that the degradation of Aβ plaques is accomplished through the process of autophagy in microglia (Lee and Landreth, 2010; Kitazawa et al., 2011). Activation of peroxisome proliferator-activated receptor-α (PPARA)-mediated induction of microglial autophagy with gemfibrozil and Wy14643 has been associated with amelioration of AD-like phenotype in the APP-PSEN1 mouse model (Luo et al., 2020). Also, disruption of noncanonical autophagy in microglia via targeted deletion of Rubicon and atg5 in vivo leads to the increased deposition of toxic Aβ and subsequent cognitive impairment in the 5 × FAD AD mouse model (Heckmann et al., 2019). Estfanous et al. reported that the ability of microglia to degrade Aβ was inhibited in AD patients compared with normal people (Estfanous et al., 2022). ACAT1/SOAT1 inhibitor K604 can improve the ability of microglia to clear Aβ-42 (Shibuya et al., 2014). These studies demonstrate that microglia can clear Aβ through autophagy process.

In AD patients, another pathological feature is neurofibrillary tangle that is formed by aggregation of abnormal phosphorylation Tau protein, which results in dysfunction of Tau protein, decrease in microtubule stability, and loss of neuronal function (Alonso et al., 2018). Hyperphosphorylated Tau protein in AD patients can be secreted into the extracellular part of neurons (Fiandaca et al., 2015; Jia et al., 2019), and microglia could engulf oligomers of extracellular Tau (Majerova et al., 2014). Other studies have also illustrated that Tau can activate microglia to promote brain inflammation (Jin et al., 2021). Zilkova et al. found that Tau antibodies could promote the absorption of Tau in human microglia (Zilkova et al., 2020). Furthermore, many studies have shown that activated microglia are involved in Tau-mediated lesions, such as Tau aggregation (Gorlovoy et al., 2009; Maphis et al., 2015), transmission (Asai et al., 2015; Wang et al., 2022), and Tau phagocytosis (Brelstaff et al., 2018; Hopp et al., 2018). Researchers have found that inhibition of microglia proliferation can improve Tau-induced neuronal degeneration and loss of function (Mancuso et al., 2019). These studies suggest that Tau and microglia have mutually reinforcing effects.

As mentioned above, both extracellular Aβ aggregation and Tau protein secretion can cause changes in microglia, but how these abnormal proteins activate microglia remains unclear. Under normal physiological conditions, microglia protect the CNS by reducing harmful stimuli, such as pathogenic molecular patterns (PAMPs) and damaging molecular patterns (DAMPs). Some receptors, such as TLRs and nuclear oligomerization domain-like receptors, and viral receptors, are expressed on the surface of microglia. All these receptors belong to pattern recognition receptors (PRRs) and can recognize PAMPs and DAMPs, thus leading to the activation of microglia (Glass et al., 2010). Similarly, Aβ and tau can be recognized by microglial PRRs, such as TLR2, TLR4, cluster of differentiation 14 (CD14), and cluster of differentiation 47 (CD47). Upon binding to these receptors, Aβ and tau can be internalized to induce inflammation through specific pathways involving NLRP3, MYD88, or NF-κB (Taro and Shizuo, 2007; Dansokho and Heneka, 2018; Kabzinski et al., 2019; Lee J. H. et al., 2021; Leng and Edison, 2021), leading to the transcription of pro-IL-β and NLRP3 (Lee J. H. et al., 2021; Leng and Edison, 2021) and the phagocytic activity of microglia. Stimulated and activated microglia also produce pro-inflammatory cytokines, such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), IL-16, and chemokines (Kabzinski et al., 2019). In turn, phagocytosis removes pathological stimuli to maintain brain homeostasis.

PRRs-mediated signal transduction promotes the release of inflammatory cytokines from microglia. TLR is expressed in cell membranes and the cytoplasm, and the TLR in cytoplasm can detect the nucleic acid of viruses and bacteria. Some TLRs can bind to other molecules, such as CD14 and CD36, as co-receptors (Dansokho and Heneka, 2018).

TLR-mediated signaling pathways lead to the production of type I IFNs and pro-inflammatory cytokines from microglia (Figure 1), and there are two ways to contribute to the release of inflammatory cytokines, MyD88 and TRIF-dependent pathways, respectively (Taro and Shizuo, 2007). It has pointed out that the activation of TRL4 produces beneficial (Michaud et al., 2013) and harmful (Zhang et al., 2015) inflammatory reactions in AD. TLR2 has been reported to produce a severe inflammatory response to Aβ (Nixon et al., 2005). However, activation of TRL9 facilitates the clearance of Aβ (Scholtzova et al., 2014). These studies suggest that activation of different TLRs produces different inflammatory responses, and that the same TLR also produces different inflammatory responses to different stimuli. It has also been shown that the activation of TLR2 and TLR4 in the early stage of AD contributes to the clearance of Aβ, while long-term activation may lead to the accumulation of Aβ (Lee S. et al., 2021). Therefore, the inflammatory response mediated by TLRs is complex, and we need to further understand the role of different TLRs in the signaling pathways.

Inflammasomes are a group of polyprotein complexes, which exist in microglia and can also recognize PAMPs or DAMPs. Therefore, inflammasomes are also considered to belong to intracellular PRRs-NOD-like receptors (NLRs) (Strowig et al., 2012). Inflammasome has various forms, including NLRP1, NLRP3, NLRP6, NLRP7, NLRP12, etc. (Dansokho and Heneka, 2018). Under normal physiological condition, the NLRP inflammasome forms vesicle-autophagosome with PAMPs or DAMPs, which then fuses with lysosome resulting in degradation of the cargo. Nevertheless, excessive activation of NLRP3 inflammasome can contribute to development of inflammatory diseases (Biasizzo and Kopitar-Jerala, 2020). NLRP inflammasome plays a certain role in the pathogenesis of AD. In monocytes of AD patients, the mRNA and protein levels of NLRP3 are significantly increased (Saresella et al., 2016), and NLRP1 is activated in the hippocampus of AD patients (Španić et al., 2022). The abnormal Aβ and Tau induce inflammation through specific pathways involving MYD88 or NF-κB (Yang et al., 2020; Barczuk et al., 2022), leading to the transcription of pro-IL-β and NLRP3 (Lee J. H. et al., 2021; Leng and Edison, 2021). Increasing evidence has shown that soluble Aβ oligomers can activate the NLRP3 inflammasome in microglia (Lučiūnaitė et al., 2020), while deletion or inhibition of NLRP3 in APP/PS1 in mice will reduce spatial memory loss and Aβ aggregation (Michael et al., 2012; Dempsey et al., 2017; Feng et al., 2018). In addition, p-Tau and aggregated Tau can also activate the NLRP3 inflammasome in microglia and further aggravate Tau lesions (Jiang et al., 2021). Similarly, inhibition of NLRP3 in TauP301S transgenic mice reduces Tau phosphorylation and Aβ accumulation in the hippocampus (Stancu et al., 2019). In conclusion, these studies show that NLRP3 inflammasome plays an important role in the development of AD, perhaps because the excess NLRP3 inflammasome in AD cannot be degraded by autophagy, leading to a more severe inflammatory response.

Autophagy and its dysfunction are associated with a variety of human pathologies, including aging, neurodegenerative disease, heart disease, cancer, and metabolic diseases, such as diabetes. A plenty of drugs and natural products have been found to modulate autophagy function through multiple signaling pathways. Small molecules or nanomedicine that can regulate autophagy seem to have great potential to intervene in neurodegenerative diseases that are largely due to the accumulation of misfolded proteins. Results from the present study indicate a correlation between microglial autophagy capacity and severity of neurodegeneration, and also provide important resources to better understand the pathogenesis of these important brain diseases. However, there are a few issues that need to be paid attention to. First, because autophagy function is impaired by aging (Gabande-Rodriguez et al., 2020), it is necessary to improve autophagy capacity in the aged brain. Second, it is also important to increase the efficiency of drug to cross the blood–brain barrier. Last but not the least, a deeper understanding and accurate detection of the early pathological changes in neurodegenerative diseases is important for developing more effective therapeutic methods.

All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.

The research presented in this article was supported by National Key Research and Development Program of China (2021YFF0702201), Natural Science Foundation of Guangdong Province (2022A1515012651, 2022A1515012301), the National Natural Science Foundation of China (32070534, 81830032, and 31872779), Guangzhou Key Research Program on Brain Science (202007030008), and Department of Science and Technology of Guangdong Province (2021ZT09Y007 and 2020B121201006).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.