Neurotoxicants driving glial aging: role of astrocytic aging in non-cell autonomous neurodegeneration

Abstract

Astrocytes, the most abundant glial cells in the central nervous system (CNS), play essential roles in maintaining neuronal homeostasis, synaptic regulation, and blood–brain barrier integrity. However, these cells can undergo senescence—a cellular state characterized by irreversible growth arrest and the secretion of proinflammatory factors—in response to aging and pathological stressors, contributing to synaptic dysfunction and neurodegenerative diseases. This review examines the molecular mechanisms driving astrocytic senescence, including oxidative stress, DNA damage, and inflammatory signaling pathways such as NF-κB and the senescence-associated secretory phenotype. A particular focus is placed on the diverse array of known chemical inducers of astrocyte senescence, such as pesticides and heavy metals, which provide critical insights into the processes governing cellular aging in the brain. By analyzing the effects of these inducers, we highlight their implications for neurodegenerative disease progression and brain aging. Understanding astrocytic senescence offers new insights into age-related neuropathology and presents promising avenues for targeted therapies in neurodegenerative disorders induced by environmental toxicants.

Astrocytes play a crucial role in the central nervous system (CNS) and are emerging as key players in neurodegenerative diseases (NDDs). They are one of the most abundant glial cells in the human brain and are necessary for metabolic and trophic support of neurons (Qian et al. 2023). Astrocytes are responsible for many key functions in the CNS, such as forming the bulk of the blood–brain barrier (BBB), where they also mediate the glymphatic system that transports water and ions from the vasculature into and out of the CNS (Jessen et al. 2015; Hablitz and Nedergaard 2021). Through this vital regulation of water transport, astrocytes carry out the extracellular ion homeostasis, modifying synaptic function, as well as the reuptake and recycling of neurotransmitters in the synaptic cleft (Khakh and Sofroniew 2015; Vasile et al. 2017; Escartin et al. 2021). Therefore, the dysregulation of their various homeostatic functions is believed to impact neuronal health significantly (Limbad et al. 2020; Brandebura et al. 2023).

Senescence in aging

Senescence is a process by which cells experiencing acute stress (such as DNA damage, lipid peroxidation, and oxidative stress induced by endogenous or exogenous factors) (Fig. 1) permanently exit the cell cycle (Hayflick 1965; Huang et al. 2022). This transition alters the gene-expression profiles of cells away from their normal functions, shifting to a more pro-inflammatory phenotype (Vernot 2020). This process is believed to be an anticancer mechanism meant to prevent aberrant or damaged cells from continuing to divide in perpetuity to avoid the risk of tumorigenesis (Huang et al. 2022). Failure to eliminate these cells can be problematic in the CNS, as glial cells cease their neurotrophic and homeostatic roles and can acquire a senescence-associated secretory phenotype (SASP) (Coppé et al. 2008; Wang et al. 2024). In this arrested state, cells release matrix proteins, pro-inflammatory cytokines, and important intracellular components such as mitochondrial DNA, generating sub-lethal apoptotic signals (Vernot 2020; Victorelli et al. 2023). SASP factors can lead to the autocrine and paracrine induction of senescence through chronic low-level inflammation as well as the recruitment of peripheral immune cells such as monocytes, natural killer, and T-cells. This recruitment can further perpetuate inflammation in the surrounding microenvironment.

Fig. 1.

Environmental drivers of astrocytic aging and its role on neurodegeneration.

The accumulation of senescent cells throughout the body has been shown to correlate with aging, the primary epidemiological risk factor for many NDDs, which has implicated senescence in the etiology of disorders such as Parkinson’s disease (PD) and Alzheimer’s disease (AD) (Sorokina et al. 2023; Alshaebi et al. 2025). Clearance of these cells typically occurs through the immune system, but as we age, this process becomes less efficient, and the burden of senescent cells increases in tissues associated with age-related pathologies, including the brain and the heart, as seen in humans, primates, and rodents (Ovadya et al. 2018; Martínez-Zamudio et al. 2021). The accumulation of these senescent cells may be due to immune avoidance, caused by the expression of tolerogenic major histocompatibility complex class I variant antigens (such as PD-L1 and PD-L2, which ligate PD-1 on T-cells), and by changes in their antigenic composition that prevent their elimination (Giannoula et al. 2023). This lack of clearance increases the SASP factor burden, and in the CNS, senescence in glial populations may leave neurons vulnerable to environmental stress as their metabolic and structural support is reduced.

Astrocytes in aging and neurodegeneration

Astrocytes, in particular, have been shown in past studies to be uniquely sensitive to senescence induction with age in response to oxidant compounds when compared with fibroblasts and other glia (Görg et al. 2018; Lee et al. 2022). An example of this is 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), which astrocytes selectively uptake and convert to its toxic metabolite MPP+ via the monoamine oxidase B enzyme in the cytosol (Ransom et al. 1987; Rappold et al. 2011), making them somewhat vulnerable to its effects, despite the literature preferentially highlighting its toxicity on dopaminergic neurons (Miyazaki and Asanuma 2020). Many claim this is due to dopaminergic neurons having a higher concentration of membrane-bound dopamine transporter protein than astrocytes (Mackie et al. 2018). Despite this, the literature has shown that MPP+ can induce apoptosis (Chuang and Chen 2004) as well as hamper mitochondrial bioenergetics in astrocytes by decreasing intracellular ATP and increasing their burden of reactive oxygen species (Sundar Boyalla et al. 2011).

The concentrations used in Sundar Boyalla et al. (2011) were between 0.1 uM and 1 mM (1 × 10⁻⁷ to 10⁻³), and found that mesencephalic astrocyte apoptosis began at the lowest concentration but did not reach an LD50, and cell death plateaued at 30% after 24 h. Meaning that although lower concentrations can be lethal to some astrocytes, the wider population is resilient to a range of MPP+ concentrations. The exact toxicity of MPP+ depends on the specific cell type, whether primary cells or a cell line. However, studies have demonstrated concentrations as low as 10 μM MPP+ can increase levels of caspase-3 activity by 3.5-fold and decreased dopamine uptake activity by 4-fold in LUHMES cells differentiated to TH+ DAergic neurons (Harischandra et al. 2020). Other studies using iPSC-derived dopaminergic neurons treated with MPP+ demonstrate significant resistance to the toxicant, with viability of cell populations reducing by half at 600 μM (Peng et al. 2013). Thus, the range of concentration of MPP+ used on astrocytes that demonstrate a significant loss of cell viability in Sundar Boyalla et al. (2011) approximates the range of toxicity for dopaminergic neurons.

Astrocytic impairment may therefore be an upstream contributor to dopaminergic neuron loss, as they are the primary cell type in the brain with abundant expression of the key rate-limiting enzyme for MPTP metabolism (Sonninen et al. 2020), and MPP+’s resulting metabolic stress could strain astrocytes’ ability to support midbrain dopaminergic neurons. This may explain why astrocytes are one of the most prominent senescent cells in post-mortem brains of PD patients (Chinta et al. 2018). Other well-characterized mitochondrial toxicants, such as the herbicide paraquat (PQ), which inhibits complex I of the electron transport chain similarly to MPP+, have also been strongly associated with the induction of cellular senescence via in vitro and in vivo mouse models of PD (Chinta et al. 2018), indicating that this process may play an urgent and crucial role in disease pathogenesis, emphasizing the need to understand it.

Astrocytes and senescence

However, even in the worst of cases, only a small number of neurons and astroglia become senescent in normal aging and NDDs, and they are broadly distributed throughout the brain and present heterogeneous phenotypes. Some studies have linked senescence phenotypes in AD to specific cell types, such as excitatory neurons (Dehkordi et al. 2021), whereas others have found that astrocytic senescence may be preferentially induced in the substantia nigra (SN) of post-mortem PD patient brains (Chinta et al. 2018); therefore, the consistency of these patterns across neurodegenerative conditions has not been established. These senescent cells can be sources of chronic inflammation, inducing what some have termed “inflamm-aging” (Franceschi et al. 2000; Jurcau et al. 2024), a term that has largely been used to highlight the effects of continuous low-level, chronic inflammation in the periphery that acts as a persistent source of cellular stress, leading to weakened immune responses and phenotypic aging in humans throughout their lifespan (Jurcau et al. 2024). When applied to the CNS, similar patterns of chronic inflammation are thought to lead to the activation of astroglia, inducing the secretion of pro-inflammatory cytokines (e.g. TNF-α, IL-1β, IL-6) or induction of SASP (Liddelow et al. 2017).

Key markers of astrocyte senescence include elevated levels of glial fibrillary acidic protein (GFAP), vimentin, β-galactosidase, and extracellular lamin B1 (Cohen and Torres 2019). However, astrocytes also have basal immune properties, as they can be activated through gliosis via inflammatory signals from activated microglia through a combination of IL-1α, TNF, and C1q (Liddelow et al. 2017). In these activated states, astrocytes can also perpetuate inflammatory signals in the CNS to combat pathogens or tissue injury. Critically, these pro-inflammatory factors secreted by activated astrocytes often overlap with factors secreted by senescent astrocytes (Boisvert et al. 2018; Miller et al. 2022), indicating that the difference between the 2 phenotypes is ill-defined and requires further investigation to conclusively differentiate (Cohen and Torres 2019; Han et al. 2020) (Fig. 2).

Fig. 2.

Venn diagram highlighting unique and shared features of astrocyte senescence and astrogliosis.

A systemic review by López-Teros et al. (2024) has attempted to compare inflammatory cytokine profiles and confirmed that astrocytes are a primary contributor to neuroinflammation in aging and environmental exposures through the secretion of IL-18, IL-6, 1β, TNF-α, and CXCL2. However, the authors remained unable to clearly distinguish between the contribution of activated astrocytes in gliosis and senescent astrocytes. The paper claims that these highlighted factors are also important for microglial activation and the recruitment of peripheral immune cells through permeabilization of the BBB. These interactions, which are driven by the above cytokines, are thought to play important roles in neurodegenerative conditions such as PD and other age-related illnesses. Despite the importance of these systemic analyses, major limitations in the field, such as the dearth of biochemical markers to clearly distinguish between activated and senescent phenotypes in astrocytes, has greatly impacted our ability to neatly characterize, define, and differentiate between these cellular states with any certainty. Although some proliferation markers, such as Ki67 and Cyclin D1, may be helpful in identifying actively dividing cells (such as gliotic astrocytes) and distinguishing them from senescent cells, quiescent cells may also lack such markers and could therefore complicate this identification. The potential for mixed populations of cells in culture makes it more difficult to use proliferation factors as markers for senescence alone, given that off-target cell types may express these factors as well (O’sullivan et al. 2024). Additionally, Ki67 has been shown to vary greatly within a single population depending on the specific phase of the cell cycle, as it is degraded continuously during G0 and G1 and is produced from the S phase until mitotic exit (Miller et al. 2018). Quiescent cells, for example, still express Ki67 in a time-dependent manner from their last cell cycle, showing that binary scoring of this marker can overly simplify the proliferative state of a given population. Therefore, any current comparison between astrocyte gliosis and senescent states can be characterized as tentative, if not speculative.

López-Teros et al., while discussing important findings in the field, also urge future research to focus our collective efforts on identifying molecular and biochemical markers before attempting to parse out astrocyte secretory profiles for the sake of advancing our understanding of astrocytic functioning. Although it may be exciting to pursue large-data analyses of senescence-associated secretory profiles, without a foundational basis distinguishing these 2 phenotypes, these data cannot address the questions we seek to answer. Instead, these data can complicate our understanding of senescence, as we do not fully agree on what markers meaningfully differentiate the diversity of these cellular mechanisms. Just like in the neuroinflammation field, we are likely to see an array of phenotypic senescence profiles that differ based on variable induction mechanisms that may be unique to cell types, age, availability of metabolic resources, and environmental exposures.

Atrocytic vulnerability to metals

Environmental exposure to metals has traditionally occurred through food and water contamination but increasingly has begun to include modern industrial consequences such as air pollution and occupational exposures (Table 1). Through oral intake or inhalation, metal ions are capable of translocating across membranous barriers and entering organ systems, where they are capable of influencing cellular functions. Specifically, metal ions are capable of passing through the BBB and entering the CNS through the use of protein transporters mainly found in astrocytes. In comparison to neurons (and other CNS cells), astrocytes are more sensitive to changes in metal ions in the extracellular environment (Li et al. 2024). Although neurons are sensitive to metal exposure at very high concentrations, astrocytes are the first to respond to exogenous metals, which can also leave them vulnerable to overexposure in pathological contexts. This astrocytic sensitivity can be attributed to 2 key features: location and channel expression.

Table 1.

Environmental factors and their role in senescence.

Environmental toxicant	Cell type or tissue	Senescence marker	Pathways involved	References
Iron	Embryonic mouse fibroblasts, and mouse hepatic tissue	Ferritin accumulation, upregulation of microtubule-associated protein light chain 3 (LC3-II)	Fe acquisition and storage, Fe-mediated apoptosis pathways (ferroptosis), perceived cellular Fe-deficiency due to excessive ferritin chelation of Fe	Masaldan et al. (2018)
	Neural progenitor cells and neurons		L-ferritin gene mutation	Cozzi et al. (2019)
	Retinal pigment epithelial cells		cGas-Sting pathway	Li et al. (2024)
Manganese	human colorectal cancer cells (HCT116)	Sa-β-gal	P53 induced senescence, decreased proliferation	Behrend et al. (2005)
Paraquat	Primary mouse cortical astrocytes and	SA-β-Gal, increased IL-6,		Chinta et al. (2018)
	Neural progenitor cells, mice (in vivo)		Decreased proliferation, decline in TH+ neurons in the SNpc,	Chinta et al. (2018)
	Pulmonary Epithelial cells		Increased proliferation of epithelial cells (fibrosis),	Huang et al. (2022a)
Rotenone	human astrocyte cell line	SA-β-Gal, IL-1α, IL-1β, IL-6, IL-8, P21	mTOR and DNA-damage reduction pathways were elevated in treated cells.	Simmnacher et al. (2020)
	iPSC-derived midbrain neurons (controls and SNCA duplication mutants)		TH+ survival decline (only in SNCA mutants)	Simmnacher et al. (2020)
	Dermal fibroblasts	IL-6, IL-1β, CXCL10, MMP9	Mitochondrial dysfunction, superoxide imbalance,	da Cruz et al. (2023)

Iron

The brain iron (Fe) homeostatic system is one of the better-explored metal homeostatic systems due to the brain’s high Fe content and its potential implication in diseases such as AD. This Fe homeostatic mechanism describes the role of astrocytes in the accumulation, storage, and trafficking of Fe. It is important to note that Fe in biological systems are ions, usually with a ²⁺ or ³⁺ charge. Fe²⁺ and Fe³⁺ play an important role in mitochondrial respiration by forming Fe–sulfur clusters and are thus key in energy generation and oxidation–reduction reactions. Astrocytes are capable of internalizing Fe ions using numerous ion transporters, most commonly divalent metal ion transporter 1 (DMT1) (Moos et al. 2007; Rathore et al. 2012; Li et al. 2024). In states of high extracellular Fe, astrocytes increase their expression of DMT1 and sequester Fe in their mitochondria, acting as an Fe sink (Rathore et al. 2012). Fe dyshomeostasis can promote the accumulation of proteins, such as amyloid-β, a hallmark of many NDDs (Rathore et al. 2012; Baringer et al. 2023). Dyshomeostasis can also promote cellular pathology and oxidative stress, proving the sensitivity of the extracellular environment to metal ions. This sensitivity of the neuronal environment to Fe and other metal ions demonstrates the importance of astrocytic maintenance of metal balance.

Although Fe-induced senescence has been understudied in astrocytes specifically, recent studies have highlighted this mechanism in other cell types. Research in primary mouse embryonic fibroblasts and human primary diploid fibroblasts has shown that senescence (induced by irradiation, replicative stress, and oncogenic mutations) alters Fe acquisition and storage as well as Fe-mediated pathways for cell death (Masaldan et al. 2018). These cells accumulate up to 30 times the amount of Fe than their homeostatic counterparts, with ferritin acting as a critical marker for predicting the induction of senescence. This effect was not altered by the use of chelators such as deferiprone. This phenotype hampered ferritin-driven ferroptosis, as the lysosomal degradation of this protein was disrupted, which was confirmed through several markers for autophagosome number, primarily via the upregulation of microtubule-associated protein light chain 3 II. These senescent cells were resistant to ferroptosis, and the induction of autophagy using rapamycin was able to reverse intracellular ferritin Fe-hoarding. These results were replicated in cultured primary-aged hepatic mouse tissue, showing that senescence preceded the accumulation of intracellular Fe. Similarly, another study looking at neuroferritinopathy (a mutation in the L-ferritin gene that causes movement disorders) found that non-ferritin-bound Fe is sufficient to cause both cellular senescence and ferroptotic cell death in neurons and human fibroblasts, which highlights the importance of Fe homeostasis in NDDs and neuronal aging (Cozzi et al. 2019).

Manganese

Manganese (Mn) is an essential metal for the function of numerous enzymes and metabolic pathways. Excess Mn has been demonstrated to accumulate in the brain and result in neurological diseases such as Manganism. Given Mn utilization of common metal transporters that are highly expressed by astrocytes, Mn can preferentially enter astrocytic mitochondria and then further accumulate within the mitochondrial matrix.

Mn exposure has been strongly associated with the progression of PD and AD, making it an increasingly relevant candidate to study senescence induction (Vielee and Wise 2023). Recent epidemiological studies have highlighted the increased presence of senescent neurons, astrocytes, and microglia early in the progression of AD (Bhat et al. 2012; Yu et al. 2024), highlighting the importance of senescence in NDDs. Critically, Mn from whole blood serum has been shown to positively correlate with Clinical Dementia Rating Scale score and Aβ peptide concentrations, and these findings have recapitulated in both in vitro and in vivo studies (Tong et al. 2014). Therefore, it cannot be ruled out that Mn may be involved in the senescence induction of these astrocytes.

Pesticides and senescence

Acute pesticide exposure has long been epidemiologically associated with the development of NDDs, especially those that specifically target complex I of the mitochondrial ETC, like PQ, rotenone, and maneb. Given that one of the key pathological hallmarks of most NNDs includes oxidative stress, reductions in bioenergetic output, and dysfunctional mitophagy, it is logical to assume these effects could influence senescence development as well.

Paraquat

PQ is an industrial herbicide used as a non-selective desiccant to help clear invasive weeds, as well as excess plant material like vines before the mechanical harvesting of cotton, soy, and corn, among other crops. Its use has long been correlated with cases of PD amongst farm workers and within rural communities (Dinis-Oliveira et al. 2006), and recent studies have reinforced this relationship. A recent study by Dr Beate Ritz’s group measured a 72% increase (1.72-fold) in the odds ratio for PD in individuals who lived within 500 m of application sites in California’s Central Valley (Paul et al. 2024). Although these data are not causational, they highlight the correlational risks of PQ use in nearby rural areas, which is valuable when considering environmental candidates for senescence induction.

A previous study by Chinta et al., published in Cell Reports in 2018, found that PQ treatment can induce senescence in primary cortical mouse astrocytes in vitro. These cells showed significantly upregulated levels of SA-β-gal and presented a SASP phenotype with increased secretion of IL-6. The effects of astrocytic aging and neurodegeneration were abrogated in an in vivo p16-3MR mouse model that eliminates all the senescent cells. These findings demonstrate the potential for PQ to induce senescence in astrocytes, which can cause non-cell autonomous neurodegeneration in murine models and human neural progenitor cells.

Very little research has been done to follow up on Chinta et al.’s (2018) study and its implications, which establish PQ as a key environmental factor capable of driving senescence induction in astrocytes. However, recent literature has looked at senescence in other cell types. One such study by Huang et al. (2022), published in Environmental Toxicology, looked at PQ-induced (20 mg/kg in 100 μl of saline) senescence in pulmonary epithelial cells in C57BL6/J mice. PQ is known to target the lung and is mainly absorbed by alveolar epithelial cells, which induces oxidative stress damage and causes rapid and irreversible pulmonary fibrosis (Tomita et al. 2007). These results showed that epithelial senescence can drive further proliferation and differentiation of pulmonary myofibroblasts that promote further fibrosis in the lungs. Collectively, these data show the efficacy with which PQ can induce senescence across cell types. This further emphasizes the need to continue investigating how environmental factors influence the progression of NDDs through this cellular mechanism.

Rotenone

Another well-characterized mitochondrial complex I inhibitor is the pesticide rotenone. Exposure to rotenone, much like PQ, has long been epidemiologically associated with an increased risk of PD (Tanner et al. 2011). Rotenone exposure has been used as a preclinical PD animal model that effectively reproduces pathological hallmarks, including loss of dopaminergic TH+ neurons in the SN and locomotor dysfunction (Cannon et al. 2009). Therefore, rotenone is an excellent candidate to study for chemical inducers of senescence in NDDs.

Simmnacher et al. demonstrated novel signatures of senescence in cultured human astrocytes using rotenone or hydrogen peroxide as their sub-lethal chemical stimuli. They found that rotenone-treated astrocytes had a dose-dependent increase in SA-β-gal-positive cells and SASP factors. P21 transcripts, a senescence marker responsible for the cell cycle inhibition characteristic of this phenotype, were also significantly increased post rotenone exposure, establishing that this neurotoxic pesticide can definitively cause senescence-like phenotypes in cultured human astrocytes (Simmnacher et al. 2020). This study also demonstrated that rotenone-induced astrocytic senescence leads to non-cell autonomous neurodegeneration in human iPSC-derived midbrain neurons from PD patients. Other recent research has also shown that senescent astrocytes, induced by aging Swiss mice, may have a greater vulnerability to rotenone-induced mitochondrial insult in astrocytes (Diniz et al. 2024). Other studies have demonstrated rotenone’s ability to induce senescence in human dermal fibroblasts due to mitochondrial dysfunction, superoxide imbalance, and DNA damage, resulting in increases in SA-β-gal and SASP factor (IL-6, IL-1β, CXCL10, and MMP9 expression) (Cruz et al. 2023). It is important to study the multifactorial risk factors in neurotoxicology, as exclusively analyzing genetics, basic mechanisms of aging, and environmental exposures is insufficient to understand the real-world contexts of health and disease.

Others

Maneb is a Mn-bound fungicide, and mancozeb shares that same chemical structure with the addition of zineb, a zinc-containing pesticide. Although in isolation they do not induce senescence, it has been shown that in low-dose mixtures (nanomoles to a few hundred nanomoles) of 7 pesticides of different chemical families (including chlorpyrifos-ethyl, dimethoate, diazinon, iprodione, and imazalil), they are capable of inducing senescence of normal mesenchymal stem cells (MSCs) in vitro and enhance tumorigenesis in vivo (Hochane et al. 2017). These data reveal that chronic low-dose exposure to diverse pesticides can work synergistically to induce senescence (as well as other adverse effects) to hasten age-related pathological conditions.

A follow-up study using the same low-dose mixture of pesticides showed that bone marrow MSCs (BM-MSC) exposed for 21 days had lowered cell proliferation and increased markers of DNA damage and senescence (Foucault et al. 2021). These BM-MSCs also showed a concomitant decrease in aldehyde dehydrogenase 2 transcription, translation, enzymatic activity, and increases in acetaldehyde levels. These resulted in a reduced capacity for BM-MSCs to carry out their homeostatic functions, mainly promoting hematopoiesis.

Although these mixture studies are increasingly difficult to conduct and parse out, we must push research in Toxicology and its many subfields to work toward standardizing such methodologies. Humans are exposed to a variety of environmental toxicants daily that interact with each other in critical ways. Although it is enticing to look at these chemicals in isolation to establish mechanisms of toxicity, in environmentally relevant contexts, we are not exposed to just 1 chemical at a time. We are concomitantly exposed to hundreds, if not thousands, of toxicants, pharmaceuticals, metals, and solvents in tandem. These may act to accentuate or counteract cellular and pathological outcomes that need to be studied holistically for scientists to clearly define their collective effects on human health. It is also critical for us to understand that the composition of these exposures can vary simply based on socioeconomic background and geography. Therefore, we must not only study single mixtures but also characterize the plethora of mixtures present around us. Although this is a daunting task, the next step in Toxicology research should be to embrace the breadth of the exposome. Otherwise, we risk overvaluing our mechanistic research on individual toxicants as relevant to human health, when in reality we are relying on piecemeal, reductive depictions of real-world exposures.

Senolytic drugs

A forthcoming strategy for mitigating the negative effects of senescence is to broadly eliminate senescent cells to prevent their accumulation and subsequent chronic inflammation. Paradoxically, this senescence-associated chronic inflammation is believed to be oncogenic despite senescence being an anticancer mechanism under ideal physiological conditions. Therefore, senolytics are considered for treating and preventing various cancers, which are believed to arise from senescent cells and the harmful microenvironments they produce (Bousset and Gil 2022; Qi et al. 2022). However, much of this work has been done in preclinical models and still requires rigorous clinical testing. Although this approach has gained traction recently, eliminating senescent cells may have its own risks in the CNS due to the complex relationship between senescence and cancer. Given the importance of astrocytes and other glial cells in maintaining neuronal homeostasis and overall health in the brain, potential off-target effects could deplete a vital population of glial cells.

Major senolytics currently in clinical trials include quercetin, fisetin, dasatinib, and navitoclax. Though some of these drugs are FDA-approved dietary supplements and could therefore be more readily available as potential therapies for patients, they have their own limitations. Quercetin may have off-target effects that could inhibit not only the senescent cells but also healthy cells (Yuan et al. 2015). Senolytics has also been used to attenuate microglia activation, disease severity, and demyelination in an immune-mediated model of multiple sclerosis, but GFAP expression remained unchanged between the control and the senolytic (ABT-263)-treated mice (Drake et al. 2024).

Alternative strategies

Although bulk-application senolytics and senescence inhibitors have been developed for some time, novel targeted approaches have recently been discussed as potential therapeutic interventions. One key shortcoming of these medications is that their therapeutic benefits require repeated or continued administration for optimal results.

The use of chimeric antigen receptor (CAR) T cells, a targeted immunotherapy popular in cancer research for its selective removal of cancerous cells using unique surface antigens to avoid harming healthy cells, may have potential applications for targeting senescent cells. In Amor et al. (2024), researchers created a CAR T-cell line that targets the senescent-associated urokinase plasminogen receptor (uPAR) to clear senescent cells across different tissues (including the pancreas, liver, and adipose tissue) in mice with a single injection. Amor et al. were able to show an improvement in aging-associated metabolic dysfunction (through maintained glucose tolerance) and increased exercise capacity.

This study is exploratory and comes with many limitations. Only a subset of the uPAR-positive cells co-express SA-β-Gal, considered to be one of the gold standard markers for senescence. uPAR is a protein homeostatically expressed in a subset of myeloid cells that promotes the remodeling of extracellular matrices involved in preventing blood clots, aiding in wound healing and tumorigenesis. Previous work by this group published in Nature has shown that uPAR is upregulated during senescence in multiple cell types and can be cleared without harming healthy tissues (Amor et al. 2020). However, this lack of consistency between markers could be indicative of various possibilities. It may be that there is a time- or age-dependent expression of senescence markers and that uPAR is expressed early in senescence induction, or that these different marker profiles represent distinct populations of senescent cells across cell types and tissues. A lack of statistical power also prevents assertions on the therapy’s effect on lifespan longevity, despite the therapeutic benefits lasting for over a year (half the typical lifespan of a mouse).

Additionally, these studies did not focus on senescence in the brain and whether the clearance of senescent glial cells would be possible using this methodology. However, CAR T cells have a clear mechanistic logic (unlike many pharmaceuticals with incomplete mechanisms), which allows for the development of more specific antigens that could potentially target senescent glial cells to yield therapeutic cognitive benefits or confer protection against neurological disease in the long term.

Looking beyond astrocytes

Senescence has been studied in various glial cell types in the context of aging and disease. A dysregulation of astrocyte and microglial gene expression can act as an important predictor of aging, with high-energy brain regions like the SN and hippocampus being the most susceptible to these changes (Alshaebi et al. 2025). This is in part due to these regions experiencing greater glucose deprivation, hypoxia, and oxidative stress related to their high bioenergetic demands. Aged microglia from healthy patients become dystrophic through the shortening or loss of fine branches and cytoplasmic fragmentation. These microglia tend to have lowered catabolic ability, as oxidized proteins and lipid waste begin to aggregate. Phagocytic activity is also impaired with aging, along with motility and migratory velocity. These functional disruptions, as they increase in severity with age, begin to increase the brain’s susceptibility to disease.

Due to their developmental origin, microglia have an inherently limited replicative capacity in comparison to other cell types in the brain (Najafi et al. 2018; Rim et al. 2024). Microglia are derived from hematopoietic stem cells in the embryonic yolk sac that invade the brain parenchyma during development prior to the formation of the BBB. These cells then mature and develop in the brain, differentiating into mature microglia with an extended lifetime that can span between 15 months and an entire lifetime (Rim et al. 2024). This prolonged lifespan can lead to the accumulation of damage to cellular machinery that can impair homeostatic functions. Although they can replicate to replenish microglial populations, the resulting daughter cells will have distinct transcriptomic profiles and increased pro-inflammatory activity, which suggests changes in cellular function (Najafi et al. 2018).

Aged microglia can accumulate myelin debris, which can induce a loss of phagocytic capacity and lysosomal function and can trigger lipid droplet formation. This accumulated damage and impaired function, along with exposure to environmental factors influencing oxidative stress, can cause microglia to have chronic inflammation and pronounced genomic instability. This instability can manifest through double-stranded DNA breaks, leading to a loss of epigenetic information, genotoxic mutations, and upregulation of senescence factors such as p16 and p53, which arrest the cell cycle.

Microglial vulnerability can also be influenced by the activity of other cell types. For example, astrocytes normally release factors that aid in the differentiation of oligodendrocyte precursor cells (OPCs). However, under pathological contexts such as AD, senescent astrocytes fail to release these factors and instead may release HMGB1, which can cause OPC dysfunction (Lau et al. 2023). This sequence of events can lead to a reduction in oligodendrocyte population and in the rate of remyelination of neurons, as well as an increase in myelin debris. This debris, rich in cholesterol, is then phagocytosed by microglia, which poses a repeated phagocytic challenge in cholesterol metabolism via the receptor triggering receptor expressed on myeloid cell 2. These lipids can contribute to lysosomal dysfunction, accumulation of lipid droplets characteristic of an aging brain, and senescence induction. Senescent microglial accumulation is suspected to be a major determinant of late-onset AD progression, and potentially of other NDDs. There are many pathological classifications of microglia, many of which share traits with senescent phenotypes, making it challenging to differentiate each population. This is especially challenging given that senescent microglia are predicted to have distinct markers of senescence depending on their regional niche in the brain. This fact continues to exemplify the importance of characterizing senescence phenotypes and distinguishing them from each other and other pathological cellular states in order to specifically target them for treatment.

Future directions

Senescent astrocytes are enriched in NDDs, and changes to their homeostatic functions induce the dysfunction and death of neurons that depend on their metabolic and mechanical support. Novel research using a nationwide medical record dataset on the incidence of nearly 1,000 diseases over 50 million life-years has demonstrated that aging is associated with the accumulation of senescent cells throughout the body, but most consequentially in tissues known to have age-related pathologies (Katzir et al. 2021). However, many senolytics and senescence inhibitors can treat age-related as well as genetically and environmentally induced senescence, suggesting a shared mechanism (Xia et al. 2020). Genetic mutations for NDDs, such as LRRK2 G2019S in PD (when penetrant), are known to increase the rate of induction and accumulation of senescent astrocytes in the brain (Olsen and Feany 2021). This emphasizes that the main correlative risk factor for most NDDs (including PD), aging, may be accelerated by known pathogenic mutations and environmental toxicants (Chinta et al. 2018).

Researchers in the neuroscience, neurodegeneration, and neurotoxicology fields have created a foundational understanding of senescence, its basic hallmarks, markers, and its general secretory phenotype. Generally, research has focused more broadly on genetic, mechanical, or environmental stressors and how they influence senescence in astrocytes and CNS cells. However, future studies to explore the combinatory role of genetic and environmental factors on NDDs is warranted. Given that NDDs are multifactorial in origin, including PD, these factors are likely acting in combination and likely not all at once but throughout the lifetime of an individual. Therefore, the interplay between them should be studied under the assumption that exposure to these stimuli may differ based on time, duration, and intensity, which should all be considered as modulatory factors that could affect the resulting astrocytic senescent phenotype.

Funding

This work is supported by National Institutes of Health grants R00ES033 723 to SS and P30ES001247, T32-ES007026 and R25GM140202.

Conflicts of interest. None declared.