The apolipoprotein E (APOE) ε4 allele is the strongest risk gene variant of AD, especially late-onset AD, and it is widely expressed in astrocytes. Post-mortem studies discovered that APOE ε4 carriers had a higher Aβ plaque burden (Schmechel et al., 1993). Single-nucleus RNA sequencing in the frontal cortex of AD patients with APOE ε4 alleles showed astrocytes with highly expressed APOE ε4 and markers of reactive astrocytes (GFAP, HSP1B, IFITM3, TAPBP, CHI3L1, etc.) (Table 1; Griswold et al., 2021).

Astrocytes carrying APOE ε4 promote Aβ accumulation both in astrocytes and neurons. In vitro, astrocytes from hiPSC carrying APOE ε4 allele increased Aβ precursor protein (APP) levels, Aβ secretion in neurons, and decreased Aβ uptake in astrocytes (Lin et al., 2018; Lee S. I. et al., 2021; de Leeuw et al., 2022; Figure 1). Aβ was produced from APP, which was increased by the formation of lipid rafts induced by APOE ε4 astrocytes (Lee S. I. et al., 2021). Cholesterol signaling also participates in Aβ accumulation regulation (Wang H. et al., 2021). Astrocytes with APOE ε4 increased astrocyte-derived cholesterol (particularly lysosomal cholesterol), resulting in Aβ accumulation (Lin et al., 2018). In addition, astrocytes differentiated from hiPSC with APOE ε4 allele showed activation and released inflammatory cytokines, aggravating the pathological process and neuron death of AD (Table 1 and Figure 1; de Leeuw et al., 2022).

In the AD mouse model, removing APOE ε4 reduced microglial activation and alleviated Aβ deposition in the cortex (Mahan et al., 2022). Interestingly, knocking out APOE ε4 in microglia did not affect Aβ plaque or transcriptional expression compared to controls (Henningfield et al., 2022), indicating that other glial cells, such as astrocytes, play essential roles in Aβ production and accumulation, while microglia appears to maintain Aβ homeostasis. Moreover, APOE ε4 in astrocytes participated in pTau aggregation to accelerate AD progression (Wang C. et al., 2021). BBB destruction also hastens the progression of AD. In a mouse model with knockin human APOE ε4, the number of astrocyte end-feet covering blood vessels in the cortex decreased, and tight junctions of BBB were damaged by increased matrix metalloproteinase 9 (MMP9) (Jackson et al., 2021; Figure 1).

In conclusion, astrocytes carrying APOE ε4 lose their normal functions, resulting in compromised BBB integrity and difficulty in Aβ or pTau clearance, as well as neuroinflammation (Figure 2). With AD progression, reactive astrogliosis occurs in the early stages of AD and precedes the hallmarks of AD (Aβ and pTau deposition).

Presenilin genes (PSEN1 and PSEN2) and APP are responsible for the early onset autosomal dominant inheritance of AD. In the early stages of APP/PS1 transgenic mouse models, astrocytes activate and resist oxidative stress in cellular and extracellular circumstances (Table 1). Astrocytes surrounding Aβ plaque deposits were regulated in K⁺ concentration imbalance to maintain normal functions in neurons and synapses in the early stage (Huffels et al., 2022). PSEN/APP mutant in astrocytes disrupted amino acid transmission between astrocytes and neurons in response to a long-term detrimental stimulus. Active branched-chain amino acids (BCAA) metabolism, impaired leucine metabolism, and neurotransmitter dysfunction through gamma aminobutyric acid (GABA) uptake capacity were discovered in astrocytes carrying PSEN1 or APP pathogenic variants (Salcedo et al., 2021a,b). In the late stage of APP/PS1 mouse models, astrocytes also triggered immune signaling and lack of neuroprotection (Orre et al., 2014). Reactive astrocytes induced neuroinflammation and activated the transcriptional activity of NF-κB to induce inflammatory factors (Cao et al., 2021; Figure 2).

The first causative gene for ALS, Cu/Zn-superoxide dismutase 1 (SOD1) mutation, was discovered in 1993. Protein aggregation and prion-like propagation of misfolded SOD1 protein are the main pathologies caused by SOD1 mutation (Canosa et al., 2022). SOD1 protein is a mitochondrial antioxidant enzyme, and SOD1 mutation increases cytoplasmic stress granules (SG) and ER stress in neurons (Rajpurohit et al., 2020; Figure 1). SOD1 gene loss of function contributes to the process of neuron death (Berdyński et al., 2022). A post-mortem study of SOD1-ALS patients discovered astrogliosis with a high level of C3 as a marker, as well as astrocyte hypertrophy in the motor cortex and spinal cord (Guttenplan et al., 2020).

Astrocytes derived from human mutant SOD1-overexpressing mice specifically damaged MNs, while other types of cells such as interneurons, GABAergic neurons, or dorsal root ganglion neurons were unaffected (Nagai et al., 2007; Bunton-Stasyshyn et al., 2015; Harlan et al., 2019). Meanwhile, in co-culture, microglia and fibroblasts with SOD1 mutant did not affect the MNs viability (Nagai et al., 2007). Astrocytes with SOD1 mutation reduced MNs viability in co-culture by inducing nitroxidative stress, transporting oxidative stress via secretomes (Rajpurohit et al., 2020), and producing hyperexcitability through dysregulated AMPA receptors and extracellular glutamate secretion to MNs (Van Damme et al., 2007; Rojas et al., 2014; Mohamed et al., 2019; Figure 1). Astrocytes derived from hiPSCs with SOD1 mutant showed astrocytes activation, increased levels of cytokines and the pro-inflammatory transcription factor NF-κB, and elevated SG. Moreover, astrocytes derived from SOD1-mutant hiPSCs stimulated mechanistic target of rapamycin (mTOR) activation induced by increased insulin-like growth factor 1 receptor (IGF1R) levels. IGF1R inhibition in astrocytes was found to be neuroprotective (Granatiero et al., 2021).

Animal models for studying astrocytes with SOD1 mutation investigated astrocytes changes at different stages. In the early and prodromal stages of ALS, astrocytes in the motor cortex and spinal cord of SOD1^G93A mice demonstrated different vulnerability to oxidative stress as astrocytes in the motor cortex of the SOD1^G93A mouse model had increased the oxidative stress, decreased the antioxidant capacity, and a relative mitochondria respiratory uncoupling, whereas the astrocytes in the spinal cord showed a higher endurance against oxidative damage through an increased antioxidant defense and a preserved mitochondrial respiratory function (Marini et al., 2021). The different responses of astrocytes from the motor cortex and spinal cords to oxidative stress indicate selective damage in ALS progression. Inflammation associated with astrocytes carrying SOD1 mutant also participates in the prodromal stage of ALS. The increasing level of astrocytic-specific TGFβ1 accelerated the disease progression in the SOD1^G93A mouse model (Endo et al., 2015). The knockout inflammatory factors (IL-1α, TNFα, and C1q) slowed the progression along with reducing toxicity and activation of astrocytes in early mouse models (Guttenplan et al., 2020). Interestingly, decreasing NF-κB in astrocytes from the cortex of SOD1^G93A mice in the prodromal stage inhibited cortical inflammation (Gomes et al., 2019). Despite the fact that astrocytes in the motor cortex and spinal cord had different endurance in the early stage of the disease, widespread and excessive oxidative stress in astrocytes was activated in the late stage of the SOD1^G93A mouse model (López-Blanch et al., 2021). In the SOD1^G93A mouse models, astrocytes from both the motor cortex and the spinal cords were characterized by ER-mitochondrial impairments, which were more pronounced in the mutated motor cortex than in the spinal cord cells (Marini et al., 2021). NF-κB increased in the symptomatic stage as well (Gomes et al., 2019), and astrocyte NF-κB-dependent activation accelerated the disease progression (Ouali Alami et al., 2018).

These findings suggest that SOD1 mutations participate in reactive astrogliosis in the early stage of ALS. Reactive astrocytes are primarily neurotoxic, causing oxidative stress, excessive excitability, and neuroinflammatory activation (Figure 3).

TAR DNA-binding protein 43 (TDP-43) mutation results in abnormal TDP-43 inclusion in the cytoplasm of MNs and glial cells, which is a pathological hallmark of most sALS and fALS, as well as in FTD (Keating et al., 2022). TDP-43, as a DNA-/RNA-modulating protein, is involved in RNA splicing, transport, and translation, as well as cellular dysfunction and toxicity (Keating et al., 2022). Normal TDP-43 maintains the protective functions of MNs, loss of function of which causes hippocampal and cortical synaptic deficits, as well as dysfunction of RNA metabolisms (Ni et al., 2021). TDP-43 cytoplasmic aggregation in astrocytes is a key feature in fALS with TDP-43 mutant, sALS, and FTD. TDP-43 positive cytoplasmic inclusions were predominant in astrocytes of the anterior horn of the spinal cord, and quite rare in neurons in a patient with fALS who carried TDP-43 mutant (Takeda et al., 2019). RNA-seq transcriptomes on post-mortem frontal, temporal cortex, and cerebellum tissue from FTD patients with TDP-43 cytoplasmic aggregation discovered upregulated markers of astrocytes (GFAP) in the frontal cortex (Hasan et al., 2021).

TAR DNA-binding protein 43 mutation, like SOD1 mutation, is toxic to astrocytes. Astrocytes with TDP-43 mutant showed cytoplasmic mislocalization of TDP-43 protein and decreased astrocyte survival (Serio et al., 2013). When primary astrocytes with TDP-43 mutation were co-cultured with MNs, MNs death existed as a result of oxidative stress and dysregulated sodium channels (Rojas et al., 2014). On the other hand, normal astrocytes provide neuroprotective and metabolic support to neurons even when TDP-43 aggregation is mislocated. Astrocytes derived from sALS iPSCs were found to reduce cytoplasmic TDP-43 mislocalization from MNs to astrocytes (Smethurst et al., 2020). However, neuronal protection from astrocytes is limited. Primary astrocytes with TDP-43 inclusions accumulated more lipid droplets, activated aerobic glycolysis, and downregulated lactate transporters, resulting in decreased metabolic support for MNs and neurotoxicity (Velebit et al., 2020). TDP-43 inclusion in the cytoplasm also induced astrocyte inflammation and activation, secreting pro-inflammatory factors (IL-1β, IL-6, and TNFα) and causing neurodegeneration and neuroinflammation (Lee et al., 2020; Kim et al., 2021).

A mouse model with a human TDP-43 mutant that was restricted to astrocytes showed progressive loss of MNs, denervation of skeletal muscles, and consequent paralysis. TDP-43 mutant activated astrocytes, and glutamate transporters GLT-1 and GLAST in astrocytes were reduced, resulting in MNs neurotoxicity from high glutamate levels (Tong et al., 2013). Besides, astrogliosis and neuroinflammation were found in the majority of the spinal cord and cortex of TDP-43 transgenic mice (Yang et al., 2022). Levels of mutant TDP-43 expression in astrocytes determined the degree of injury to MNs (Yamanaka and Komine, 2018). In transgenic mice with TDP-43 mutant, reactive astrocytes secreted neurotoxic factors, lipocalin 2 (lcn2), to specifically cause neuron death, while other glial cells were unaffected (Table 1; Bi et al., 2013). Oxidative stress and abnormal ATP accumulation in astrocytes were also elevated in TDP-43 transgenic mice, along with glutathione and L-glutamate uptake deficits (Moujalled et al., 2017; Barton et al., 2020). TDP-43 overexpression may also activate astrocytes (Table 1). TDP-43 overexpression in astrocytes triggered inflammation in the brain, resulting in BBB permeability disruption (Zamudio et al., 2020).

These findings suggest that the balanced TDP-43 expression is essential to maintain astrocyte homeostasis. Astrocytes containing mutant TDP-43 destroy and transform the protective and supportive function of astrocytes into neurotoxicity through secreting pro-inflammatory factors, inducing oxidative stress and excitatory glutamate toxicity (Figure 3).

Hexanucleotide repeat expansions in C9orf72 comprise the most common causative gene of ALS and FTD in Caucasians. ER-mitochondrial signaling, protein homeostasis, and RNA metabolisms were found to be dysfunctional in cases with C9orf72 expansions (Gomez-Suaga et al., 2022). Dipeptide repeat (DPR) polypeptides translated by C9orf72 repeat expansions and deficiency of C9orf72 expression are detrimental to neurons and microglia (Lall et al., 2021). Post-mortem analysis of C9orf72-ALS patients revealed astrogliosis and astrocyte senescence (Table 1; Guttenplan et al., 2020). A group of enriched proteins specifically expressed from astrocytes was significantly increased when ALS patients with the C9orf72 expansion were compared to sALS patients (Umoh et al., 2018). GFAP expression was elevated in the motor cortex and hippocampus of transgenic mouse models expressing GFP-poly (GR)₁₀₀ and poly (GA)₅₀ (Table 1; Zhang et al., 2016; Zhang Y. J. et al., 2018), whereas GFAP level was not changed in the spinal cord (Schludi et al., 2017). C9orf72 BAC mouse models revealed that astrogliosis preferred to exist in the end-stage of the disease (Liu et al., 2016). However, there was no difference in plasma GFAP expression between the presymptomatic and symptomatic stages of C9orf72-FTD (Heller et al., 2020). Therefore, further research into the specific role of astrocytes with C9orf72 expansion is required.

Motor neurons undergo cell senescence due to oxidative stress and neurotoxicity when co-cultured with fibroblast-derived astrocytes from C9orf72-ALS patients (Birger et al., 2019; Zhao et al., 2020). Both astrocytes and neurons in cortical organoids derived from hiPSCs of C9orf72 cases existed in DPR polypeptides and expanded RNA foci via disrupted autophagy and proteostasis (Mizielinska et al., 2013; Conlon et al., 2016; Zhao et al., 2020; Szebényi et al., 2021). Astrocytes from hiPSCs with C9orf72 expansion caused voltage-activated currents and action potential output loss in MNs, resulting in disrupted action potentials (Zhao et al., 2020). There was a lack of energy production due to adenosine to inosine deamination defect, reduced glycogen metabolism, and mitochondrial respiration dysfunction in astrocytes in familial C9orf72 cases (Allen et al., 2019a,b). In addition, intracellular glutamate level was elevated in astrocytes of patients with C9orf72 expansion (Fomin et al., 2018). In contrast to SOD1-ALS astrocytes, glutamate secretion was not increased in C9orf72-ALS astrocytes in vitro (Mohamed et al., 2019). These results indicated that different glutamate secretion and uptake mechanisms exist in astrocytes with SOD1 mutant and C9orf72 expansion.

In conclusion, C9orf72 expansions are considered the primary cause of toxicity in astrocytes by disrupting energy supply and mitochondrial respiration. Astrocytes dysfunction results in neurotoxicity to MNs and, as a result, ALS pathogenesis (Figure 3).

Fused in sarcoma gene (FUS) mutation mainly exists in early onset ALS and FTD. FUS is a multifunctional DNA/RNA-binding protein involved in RNA processing and metabolism. Cytoplasmic FUS inclusion is another pathological hallmark of ALS. FUS protein mislocalization in mitochondria caused by FUS mutation causes mitochondrial death in neurons (Deng et al., 2015). FUS mutant also impaired neuromuscular junctions and caused mitochondrial dysfunction in neurons and skeletal muscle (Yu et al., 2022). An ALS patient carrying the FUS variant showed increased expression of FUS protein in reactive astrocytes and astrogliosis accumulated surrounding MNs with RNA foci (Wongworawat et al., 2020). Another patient with FTD carrying the FUS variant had FUS-positive cytoplasmic or intranuclear inclusion (Murakami et al., 2021). However, FUS inclusion remained mainly in oligodendrocytes rather than astrocytes in other ALS cases with the FUS variants (Mackenzie et al., 2011). These disparities were most likely caused by the small number of autopsies performed on ALS/FTD patients carrying FUS variants.

Silencing FUS expression in the brain by the AAV vector system provoked a proliferation of astrocytes and astrogliosis in marmosets (Table 1; Endo et al., 2018). FUS mutant astrocytes triggered MNs susceptible to excitotoxicity via AMPAR-mediated cell death (Kia et al., 2018). Clock and clock-controlled genes altered in FUS-ALS iPSC-derived astrocytes, contribute to metabolic and redox impairment (Killoy et al., 2021). In SOD1-mutant mouse models, misfolding cytoplasmic FUS accumulation in reactive astrocytes caused MN degenerative death via pro-inflammatory and neurotoxic pathways such as TNF-α with neutralizing antibodies (Li et al., 2016). These findings indicate that FUS deficiency and mislocalization are toxic to astrocytes. Meanwhile, astrocytes with FUS mutation reduced neuron viability through neuroinflammation and metabolic dysfunction (Figure 3).

The microtubule-associated protein tau (MAPT) is a causative gene in tauopathy diseases such as FTD, and progressive supranuclear palsy (PSP). MAPT encodes tau protein and produces six different isoforms of tau, primarily 3R and 4R in approximately 1:1 ratio in adult brains. Post-mortem neuropathological examination revealed that the presence of astrogliosis in patients with MAPT P301T variant, with these astrocytes retaining tau protein inclusion (Table 1; Erro et al., 2019).

Tau is expressed by astrocytes and oligodendrocytes, though at lower levels than in neurons (Chung et al., 2021). However, the high expression of phosphorylated tau in astrocytes in the pathological states suggests that astrocytes may also be involved in tauopathy. The overexpression of ptau in astrocytes was sufficient to cause cognitive decline in a mouse model (Richetin et al., 2020). Astrocytes derived from asymptomatic MAPT 10 + 16 intronic mutation iPSCs had a higher ratio of 4R:3R-tau transcript and protein than neurons (Setó-Salvia et al., 2021). Reactive astrocytes stimulated by tau express neurotoxic factors and respond to oxidative stress in vitro by producing ROS and membrane activation (Esteras et al., 2021; Ungerleider et al., 2021). Tau inhibited mitochondrial calcium efflux in both neurons and astrocytes derived from 10 + 16 MAPT hiPSC (Britti et al., 2020). In the presence of mitochondrial dysfunction, astrocytes and neurons derived from 10 + 16 MAPT hiPSC were more susceptible to calcium-induced caspase 3 activation and cell death (Britti et al., 2020).

Parkinson’s disease is the second most common neurodegeneration, characterized by dopaminergic neurons (DA neurons) death in the substantia nigra (SN) and dopamine (DA) deficiency in the striatum. Astrocyte response to acute and chronic stress earlier than neuron in PD. In contrast to the widespread astrogliosis seen in AD and ALS, the cortex of sporadic PD lacked or contained only mild GFAP-positive astrocytes in post-mortem examination (Tong et al., 2015; Cardinale et al., 2021). Patients with PD showed increased reactive astrocytes in the brainstem in the early stage of PD, while decreased reactive astrocytes in the cortex and brainstem in the middle and late stages of the disease, according to an in vivo PET imaging study (Wilson et al., 2019). These findings indicate different conditions of astrocytes during the PD process. Astrocytes provide neuroprotection via releasing antioxidants in response to oxidative stress in the early stages of PD (Takahashi and Mashima, 2022). However, pro-inflammatory factors, energy deficiency, and synapse dysfunction from reactive astrocytes accelerate neuron death in the late stages of PD (Liddelow et al., 2017; Kam et al., 2020). Pathogenic and risk genes associated with PD may assist in understanding the role of astrocytes in PD.

Leucine-rich repeat kinase 2 (LRRK2) is involved in sporadic and familial PD and is a link to lysosomal and mitochondrial functions. Excessive activation and phosphorylation of LRRK2 protein impaired autophagy in patients with PD (Di Maio et al., 2018; Pang et al., 2022). Astrogliosis and neural loss were discovered in the SNs of patients with the LRRK2 variant (Table 1; Takanashi et al., 2018).

Astrocyte atrophy was found in the derived astrocytes from PD patients with LRRK2^G2019S mutation (Table 1; Ramos-Gonzalez et al., 2021). LRRK2^G2019S mutated astrocyte from hiPSC appeared to cause nutritional damage to the DAergic neurons, damage the ATP supply, low-mitochondrial density, and ER disorder (de Rus Jacquet et al., 2021). Moreover, autophagy dysfunction was observed in astrocytes derived from LRRK2^G2019S mutated iPSCs (di Domenico et al., 2019; Figure 1). LRRK2^G2019S mutated astrocytes decreased the capacity to internalize and degrade fibrillar α-synuclein via the lysosomal pathway (Streubel-Gallasch et al., 2021). These disorders lead to a lack of α-synuclein internalization and clearance in astrocytes, as well as α-synuclein accumulation.

Primary cilia loss and astrocytic atrophy are the main characteristics of astrocytes in LRRK2 mutant models (Table 1; Khan et al., 2021), with astrocytes losing physical functions, such as ER stress, mitochondrial dysfunctions (Lee J. H. et al., 2021), etc. However, primary astrocytes transformed into reactive astrocytes in LRRK2 transgenic mice when exposed to additional oxidative stress, such as a toxic dose of MPTP (Arbez et al., 2020). In LRRK2 mutation mouse models, reactive astrocytes protect neurons through anti-inflammatory functions. Astrocyte activation of Nrf2 inhibited neural degeneration in transgenic mouse models by antagonizing LRRK2^G2019S-induced Mad/Smad signaling (Lin et al., 2021).

These findings suggest that LRRK2 mutant trigger astrocyte senescence, resulting in neural death, while reactive astrocytes may also play anti-inflammation roles in PD patients with LRRK2 mutant (Figure 4).

The α-synuclein gene (SNCA) was discovered in 1997 as the first familial PD gene. Point mutations (such as A53T, A30P, E46K, and H50Q) and copy number variation were common pathogenic variants in PD (Polymeropoulos et al., 1997). Insoluble α-synuclein accumulation is an important pathologic marker in PD. Astrocytes in post-mortems were found to internalize a significant amount of α-synuclein fibrils (Kovacs et al., 2014), and to participate in the spread of α-synuclein via extracellular vesicles or exosomes (Rostami et al., 2020).

Extensive reactive astrocytes were discovered in SNCA A53T transgenic mouse models (Table 1; Yang et al., 2015). In astrocytes, SNCA A53T and A30P variants remained in the impaired ER stress. Interestingly, rotenone-induced astrocyte senescence led to axonal degeneration of midbrain neurons with SNCA locus duplication (Simmnacher et al., 2020). Cytoplasmic α-synuclein in neurons recruits and activates astrocytes via a prion-like process, which is a common pathology in patients with PD (Sznejder-Pachołek et al., 2017; Rizor et al., 2019). Astrocytes reduced α-synuclein proteins misfolding, prevented neurotoxicity from α-synuclein aggregation, rescued damaged DA neurons, and then protected DA neurons at an early stage of PD (Jewett et al., 2018; Zhang Z. et al., 2018). Long-term α-synuclein stress, on the other hand, disrupts the physiological functions of astrocytes. Dysfunction of the ER-Golgi compartment α-synuclein overexpressing astrocytes may lead to a decrease in glial cell-derived neurotrophic factor (GDNF) level, which would suppress the neurite outgrowth (Liu et al., 2018). Besides, exogenous overexpression of α-synuclein proteins in astrocytes damaged the autophagy-lysosomal pathway and promoted astrocyte apoptosis (Erustes et al., 2018). Neuroinflammation induced by astrocytes with α-synuclein aggregation also hastens the progression of PD (Kim et al., 2018; Kwon et al., 2021).

In conclusion, SNCA variants cause astrocyte reaction and atrophy, as well as ER dysfunction. Aggregation of α-synuclein is toxic to astrocytes via multiple pathways, resulting in astrocyte dysfunction and accelerating α-synuclein accumulation (Figure 4).

PTEN-induced putative kinase 1 (PINK1), parkin (PRKN), and DJ-1 genes are all known to play a role in early onset PD (EOPD) (Bonifati et al., 2003; Dawson and Dawson, 2010; Kawajiri et al., 2011). Mitophagy is mediated by these genes via the PINK1/parkin pathway. PINK1 protein is a mitochondrial kinase, parkin is an E3 ubiquitin ligase, and DJ-1 protein participate in proteasome degradation (Panicker et al., 2017; Li et al., 2021).

Reactive astrocytes in iPSCs induced from patients with homozygous PRKN variants were reduced (Table 1; Kano et al., 2020). Deletion of the PRKN leads to abnormal astrocyte function, resulting in DA neurons being vulnerable to oxidative stress (Solano et al., 2008). Parkin controls neuronal homeostasis by regulating astrocyte ER stress and inflammation. In response to ER stress, parkin deficiency astrocytes increased ER stress and released cytokine, and decreased neural support (Singh et al., 2018).

Reactive astrocytes increase PINK1 protein expression (Table 1; Choi et al., 2016; Jarazo et al., 2022). Astrocytes with PINK1 loss of function increased neuroinflammation and lacked physiological support to neurons (Leites and Morais, 2021). PINK1 deficiency disrupted mitophagy, altered PINK1-dependent ubiquitin phosphorylation, and increased nitric oxide production in astrocytes (Sun et al., 2018; Barodia et al., 2019; Komilova et al., 2021).

DJ-1 is a neuroprotective protein in PD, which deficiency impaired glutamate uptake into astrocytes by altering EAAT2 expression in vitro (Kim et al., 2016). DJ-1 was found to be highly expressed in reactive astrocytes in vivo (Frøyset et al., 2018). DJ-1 is a positive regulator of STAT3 activation, the most important astrogliosis mediator (Choi et al., 2018b). DJ-1 overexpression in astrocytes protected neurons from multiple PD processes (de Miranda et al., 2018), and regulated several proteins that support physiological functions on astrocytes and neurons, including redox regulation, anti-inflammation, and mitochondrial respiration (Ashley et al., 2016; Frøyset et al., 2018). DJ-1 dysfunction resulted in the harmful inflammatory response in PD development (Choi et al., 2019). DJ-1 knockout astrocytes may provide less neuroprotection to surrounding neurons due to changes in pro-inflammatory mediator expression (Table 1; Ashley et al., 2016). DJ-1 knockout mice had defective astrogliosis caused by decreased CCL2, Sox9 expression, and reduced monocyte infiltration, which disrupted recovery from CNS injury and accelerated PD progression (Table 1; Choi et al., 2018a,2020).

Taken together, since PRKN/PINK1/DJ-1 genes are associated with mitophagy and ubiquitin functions, their loss of functions of these genes in astrocytes participates in excessive oxidative stress, neuroinflammation, and mitochondrial dysfunctions, all of which contribute to the PD process (Figure 4).

β-glucocerebrosidase gene (GBA), a pathogenic gene for Gaucher’s disease (GD) (Neumann et al., 2009), is the most common genetic risk factor of PD and is associated with an autophagic-lysosomal pathway (Senkevich and Gan-Or, 2020). GBA deficiency may cause α-synuclein aggregation and alter neuronal susceptibility to pathology (Henderson et al., 2020; Pang et al., 2022).

Extensively reactive astrocytes with GFAP, S100β, and severe cytoskeletal hypertrophy in astrocytes were discovered from iPSCs from GD patients (Table 1; Aflaki et al., 2020). Furthermore, residual GCase activity appeared to determine the degree of astrogliosis, inflammatory response, Ca⁺ dysfunction, and ability to process α-synuclein (Aflaki et al., 2020; Sonninen et al., 2020). Autophagy and lysosomal storage disorder in reactive astrocytes, combined with inflammation, disrupt mitochondrial homeostasis and cause α-synuclein aggregation in the cortex (Di Malta et al., 2012; Osellame et al., 2013; Booth et al., 2017; Sanyal et al., 2020). These findings suggest that GCase deficiency in GBA mutant astrocytes is likely a trigger of reactive astrogliosis through inflammation, impaired autophagy, etc., and as a result of PD progression (Figure 4).

As previously discussed, reactive astrogliosis is common in neurodegenerative diseases with pathogenic or risk gene variants, indicating that astrocytes activate in the early stages and precede hallmarks, and exert different effects throughout the disease progress. New technologies have emerged to assist researchers in directly observing the progression of reactive astrogliosis in patients. The detection of reactive astrocytes is gradually being implemented to detect the disease at its early stages and track its progression. A selective monoamine oxidase-B (MAO-B) tracer is also used to detect reactive astrogliosis since MAO-B is overexpressed in reactive astrocytes. The tracer identified the reactive astrogliosis in mild cognitive impairment (MCI) and AD (Vilemagne et al., 2022b) when compared to controls, and it was detectable at the preclinical stages of Aβ accumulation (Vilemagne et al., 2022a). Markers secreted by reactive astrocytes could also be implemented as a diagnostic tool in patients. Plasma GFAP differs in FTD and AD, which is useful to distinguish FTD and AD and predicting cognitive decline when combined with plasma Nfl detection (Zhu et al., 2021). Salivary GFAP is also considered a potential biomarker for the diagnosis of MCI and AD (Katsipis et al., 2021). These findings indicate that astrocytes are useful to recognize neurodegenerations, though the specificity and sensitivity of reactive astrocytes for diagnosis need to be improved since shared mechanisms induced by astrocytes exist in various neurodegenerative diseases. Moreover, given the neurotoxic roles of reactive astrocytes, regulating pro-inflammatory factors, synapse dysfunctions, high BBB permeability, etc. induced by reactive astrocytes probably reduce the neurotoxicity. Targeting molecular or pathways associated with astrocytes such as STAT, EAAT, GFAP, and connexin 43 via adeno-associated virus (AAV) reduced Aβ accumulation in mouse models (Price et al., 2021). Researchers also attempted to restore physiological capabilities in ALS patients by transplanting astrocytes derived from human embryonic stem cells (ClinicalTrials.gov Identifier: NCT03482050).