Senescent cell heterogeneity in brain aging and neurodegenerative disease

Summary

Senescent cells in the aging and diseased brain are increasingly recognized as highly heterogeneous catalysts of dysfunction, originating from diverse cell types, and characterized by wide-ranging molecular signatures and functional outcomes. Additionally, technological advances in single-cell transcriptomics and mouse modeling have helped reframe senescence from a static fate to a dynamic trajectory that is heavily influenced by evolving environmental cues. In this review, we identify key hubs of heterogeneity in brain cell senescence. We discuss how differences in senescence induction, cell-cycle arrest mechanism, cell-type biology, and microenvironments contribute to the diverse senescence programs observed in the central nervous system. We also synthesize insights from the recent wave of single-cell transcriptomic studies and discuss how advances in spatial omics technologies could transform our ability to study senescent cells within intact neural circuits. Finally, we argue that integrating multimodal molecular profiling with functional studies in mice will be essential for advancing mechanistic understanding and therapeutic targeting of senescent cells in brain aging and disease.

In brief:

This review by Graves et al synthesizes emerging evidence that senescent brain cells are heterogeneous, dynamic, and context-dependent, highlighting determinants of diverse programs and emphasizing integration of single-cell, spatial omics, and mouse studies to facilitate mechanistic insights and possible therapeutics.

I. Cellular senescence in central nervous system (CNS) aging and disease

A. historical perspective of cellular senescence

One of the first theories of aging was proposed by the German biologist, August Weismann as the “Wear and Tear Theory” in the 1880s. In it, Weismann suggested that organisms simply “wear out” over time due to repeated use, exposure to environmental stressors, and accumulated damage to tissue and cells.¹ Weismann also proposed that reproductive germ cells were potentially immortal while somatic cells had definite lifespans², a notion that Carrel and Ebeling challenged in 1921 when they reported that cells isolated from fetal chicken hearts could be cultured for decades with continuous growth.³ In another twist in 1961, Leonard Hayflick and Paul Moorhead showed that human primary fibroblasts divided a certain number of times before ceasing proliferation in a process they termed replicative senescence, which later became known as “the Hayflick limit”.^4,5 Ultimately, Carrel and Ebeling’s original findings were attributed to continual culture contamination with fresh cells during media changes.

Decades of in vitro cell-culture experiments have provided insights into cellular senescence, defining it as a stable cell-cycle arrest program prompted by diverse endogenous and exogenous stressors including telomere erosion, oncogene activation, oxidative damage, and genotoxic insults.⁶ Core molecular hallmarks of senescence include cyclin-dependent kinase inhibitor upregulation (such as p16^INK4a (hereafter, p16) and/or p21^WAF1/CIP1 (hereafter, p21)), persistent DNA damage (including ɣ-H2AX and telomere-associated foci), chromatin remodeling, characteristic changes in cell morphology, and acquisition of a pro-inflammatory senescence-associated secretory phenotype (SASP).^6,7 The characteristic insights gleaned from senescent cells in culture have laid the foundations for our current understanding of how senescent cells might contribute to the biology of organismal aging. However, a causal relationship was only established a little over a decade ago through genetic mouse models demonstrating that clearance of senescent cells can extend both healthspan and lifespan.^8,9 These findings helped catalyze a rapid expansion of research aimed at defining the role of senescent cells in both aging and age-associated diseases.

Challenges to identifying senescent cells in vivo

Studying senescence in tissues has been challenging for a number of reasons. In vivo senescent cells are rare, heterogeneous, and embedded within complex tissue microenvironments, making them far harder to characterize than their cultured counterparts where the cell type and mode of senescence induction are defined experimentally. Many canonical markers of senescence are neither exclusively senescence-specific nor uniformly expressed in all senescent cells, complicating efforts to distinguish true senescence from other transient cellular processes like quiescence, terminal differentiation, or transient stress responses. Technical barriers in common immunofluorescence and transcriptomics approaches such as non-specific antibodies to core biomarkers of senescence, transcriptomics experimental sensitivity, and selective vulnerability to tissue dissociation-related dropout further obscure the ability to study senescent cells in vivo. Furthermore, better technologies for tracking and characterizing senescent cells are needed in order to assess how senescent cell behavior may change temporally in vivo with age and disease. Because of these challenges, early years of in vivo cellular senescence studies were plagued by inconsistently defined senescent cell phenotypes. To combat this, leaders in the field of senescence research published “Guidelines for minimal information on cellular senescence experimentation in vivo” in 2024. These guidelines call for three conditions to be met by any study on cellular senescence: 1) evidence for expression of at least 3 senescence markers, with one being upregulation of p16 or p21; 2) evidence that reduction of senescent cell abundance leads to a pathophysiological change; and 3) reporting negative results and detailed experimental information to increase reproducibility.¹⁰ Additionally, the Cellular Senescence Network (SenNet) Consortium was established to systematically characterize senescent cells across the body and recently published further recommendations for detecting senescent cells in vivo by summarizing tissue-specific evidence across existing literature.¹¹

When it comes to the aging or diseased brain, studying senescent cells poses a set of extra challenges. Human brain tissue has historically been difficult to obtain and often is available only as post-mortem material, which may not be collected for many hours and typically is not preserved in optimal ways for characterization of senescent cells. Defining senescence in brain cell types that are long-lived, such as neurons and some glia, requires particular care due to the overlap of certain biomarkers between terminal differentiation and senescence. Attempting to study senescent cells in the context of brain disease adds further complexity because many brain diseases involve rampant neuroinflammation that can share features with the SASP and confound interpretation. Brain cells also respond poorly to dissociation in cell-type and region-specific ways, limiting the accuracy of single-cell transcriptomic approaches. Finally, the connectivity of brain cells and the highly organized regional structure of the brain mean that spatial context is essential for understanding the impact of senescent cells on neural circuit function, and ultimately, on cognition. Altogether, these factors create a complex landscape in which to study the cellular senescence in the brain.

Targeting senescent cells in CNS aging and disease

The recognition that senescent cells actively contribute to tissue dysfunction during aging has generated substantial interest in therapeutic strategies aimed at selectively targeting these cells. Two major classes of approaches have emerged: ‘senolytic’ therapies that eliminate senescent cells by inducing apoptosis and ‘senomorphic’ interventions that attempt to suppress the detrimental cell extrinsic effects of senescent cells, with particular focus on the SASP. Early studies demonstrating that genetic or pharmacological clearance of senescent cells can delay age-related pathologies and extend lifespan in mice provided proof-of-principle that senescent cells represent actionable therapeutic targets.^8,9,12,13 These findings sparked rapid growth in the field and led to the development of multiple compounds capable of selectively targeting senescent cells.

Evidence supporting this strategy is now beginning to emerge within the nervous system. In mouse models of tauopathy, pharmacological or genetic elimination of senescent cells reduces neuroinflammation, attenuates tau pathology, and improves cognitive performance.^14–16 Similarly, studies in Alzheimer’s disease models have demonstrated that removal of senescent oligodendrocyte progenitor cells decreases amyloid burden and restores aspects of cognitive function.¹⁷ These findings suggest that senescent cells can contribute directly to disease progression in the brain rather than simply representing passive byproducts of aging.

Despite these promising observations, therapeutic targeting of senescent cells in the CNS presents several challenges. First, the blood–brain barrier restricts access of many senolytic compounds to neural tissue, complicating drug delivery and limiting therapeutic efficacy.^18,19 Second, many neural cell types are long-lived and play essential roles in maintaining neural circuit function. Eliminating senescent cells in the brain therefore raises important questions about whether the removal of these cells might inadvertently disrupt tissue homeostasis.^20,21 In contrast to rapidly renewing peripheral tissues, where senescent cells can often be replaced through cell proliferation, neuronal and glial populations in the adult CNS have limited regenerative capacity. As a result, strategies that eliminate senescent cells may have fundamentally different consequences in the brain than in other organs.

Another major challenge arises from the heterogeneity of senescence programs across neural cell types and disease contexts.^22,23 As discussed later, senescence phenotypes differ substantially between microglia, astrocytes, oligodendrocyte lineage cells, neurons, and vascular cells. Moreover, the molecular pathways that drive senescence may vary depending on the initiating stressor, such as DNA damage, chronic inflammation, metabolic dysfunction, or proteotoxic stress. These differences raise the possibility that therapies targeting one senescence pathway may be effective only in specific cellular or disease contexts. Consequently, future efforts to therapeutically target senescent cells in the nervous system will likely require precision approaches that account for both cellular identity and disease state.

Emerging technologies are beginning to provide the tools necessary to develop such strategies. Single-cell and spatial transcriptomic approaches now allow investigators to map senescence-associated gene expression patterns across diverse cell populations and anatomical regions. These approaches are revealing that senescent cells in the central nervous system (CNS) may not represent a single uniform phenotype but rather a collection of related cellular states with distinct molecular signatures and functional consequences. Identifying which of these states contribute most strongly to disease progression will be essential for designing effective therapeutic interventions.

Aim of this review

In this review, we explore the evidence for cellular senescence in neural tissues and the technologies researchers use to generate these important data. We discuss how differences in induction mechanisms, intrinsic cell-type biology, and regional microenvironments contribute to the diversity of senescence programs observed in the brain and spinal cord. We also discuss how recent advances in single-cell and spatial omics technologies are transforming our ability to study senescent cells within intact neural circuits. Finally, we consider the experimental and conceptual challenges associated with identifying senescent cells in vivo, particularly within complex neural tissues, and how the laboratory mouse remains a powerful tool for overcoming these challenges. By integrating insights from molecular biology, neuroscience, and aging research, we aim to provide a framework for studying senescence in the nervous system and highlight key questions that will guide future studies in this rapidly evolving field.

II. Sources of heterogeneity in cellular senescence

The CNS contains a diverse set of specialized cells with distinct developmental origins, metabolic demands and response to injury. As a result, senescence in the brain and spinal cord does not represent a single conserved program but rather a spectrum of context-dependent cellular states. Evidence from experimental models and human tissues suggest that this arises from at least four major sources: inducer heterogeneity, cell cycle arrest program heterogeneity, cell-type specific heterogeneity, and microenvironmental heterogeneity. Studies consistently point to a growing number of recognizable senescent cell types found in the aging and diseased CNS. We’ve summarized the most well-characterized subsets in Table 1 and Table 2 and discuss them in the following subsections.

Table 1.

Cell-type and context-specific features of senescence across the aging CNS

Brain Region	Cell Type	CDKi Status	Morphology	Core SASP	SA-β-Gal	Publication
Corpus callosum, pons, medulla oblongata, spinal cord	Microglia	↑p16	NA	IL-1β, Cxcl10	NA	Matsudaira et al., 2023
Neocortex & Hippocampus	Microglia	↑p16 –p21	NA	IL-1β, TNF-α, TGF-β	Positive	Stojiljkovic et al., 2019
Whole brain	Microglia	↑p16 ↑p21	NA	NA	NA	Rachmian et al., 2024
Forebrain	Microglia	↑p16 –p21	NA	IL-6, IL-1β, TNF-α	Positive	Xu et al., 2020
Hippocampus	Microglia OPC	↑p16	NA	ANGPTL-1, IGF-1, IL-1β, MIF, PLAUR, TIMP2	NA	Ogrodnik et al., 2021
Hippocampus	OPC	↑p16 ↑p21	Increased number of branches & processes	Ccl3, Ccl4, Ccl5	Positive	Chen et al., 2025
Fimbria	OPC	↑p16	Increased convex hull area and perimeter	SerpinA3N	NA	Gomez et al., 2024
Dorsal root ganglion	Neuron	↑p16 ↑p21	NA	IL-6	Positive	Donovan et al., 2025
Cerebellum	Neuron	NA	NA	IL-6	Positive	Jurk et al., 2012
Whole brain	Endothelium	↑p16	NA	IL-6, TNF-α	NA	Csik et al., 2025

Table 2.

Cell-type and context-specific features of senescence across the diseased CNS

Context	Location	Cell Type	CDKi Status	Morphology	Core SASP	SA-β-Gal	Publication
Ischemia	Whole brain	Microglia	↑p21	NA	IL-6, IL-1β, Cxcl2, Cxcl10	NA	Yang et al., 2025
	Neocortex	Microglia	↑p16 –p21	NA	IL-6, IL-1β, TNF-α	Positive	Baixauli-Martin et al., 2025
		Neuron	↑p16 –p21	NA
	Striatum	Microglia	↑p16 –p21	NA	IL-6, IL-1β, TNF-α	Positive
		Neuron	↑p16 –p21	NA
	Neocortex	Microglia	↑p16 ↑p21	Cytoplasmic p16 expression; Membranous p21 expression	Cxcl11, NF-κB	Negative	Torres-Querol et al., 2021
		Neuron	↑p16 ↑p21	Cytoplasmic p16 expression
Aβ Plaques	Neocortex	OPC	↑p16	Cytoplasmic Olig2 expression; Accumulation of autolysosomes	IL-1β, TNF-α	Positive	Zhang etl al., 2019
	Neocortex	Endothelium	↑p21	NA	NA	Positive	Ting et al., 2023
		Pericyte	↑p21	NA	NA	Positive
	Neocortex	Microglia	↑p16 ↑p21	No change in cell body size or surface area	IL-6, IL-1β	Positive	Hu et al., 2021
	Whole brain	Microglia	↑p16 ↑p21	NA	NA	NA	Rachmian et al., 2024
	Cerebellum	Astrocytes	–p16	NA	NA	NA	Bhat et al., 2012
	Neocortex	Astrocytes	↑p16	NA	IL-6, IL-8, Ccl5, MMP-1	Positive
	Hippocampus	Astrocytes	↑p16	NA	IL-6, IL-1β, SerpinA3N	Positive	Han et al., 2024
Tauopathy	Cortex & Hippocampus	Microglia	↑p21	Mixed population of hypertrophic and dystrophic microglia	IL-1β, Ccl2, Ccl3, Ccl4	NA	Ng et al., 2023
	Whole brain	Microglia	↑p16 ↑p21	NA	NA	NA	Rachmian et al., 2024
	Cortex & Hippocampus	Microglia	↑p16 ↑p21		PAI-1, IL-6	Positive
		Astrocytes	↑p16 ↑p21	NA	IL-6	Positive	Bussian et al., 2018
		Oligo-dendrocytes	–p16 ↑p21		No significant changes	Negative
	Cortex	Neurons	↑p16 ↑p21	NA	IFN-γ, TNF, TLR4, IL1-β, CXCL1, NFκB	Negative	Musi et al., 2018
	Pre-frontal cortex	Neurons	↑p19	Increased nuclear size	NA	Positive	Dekhordi et al., 2021
Synuclein-opathy	Midbrain	Astrocytes	↑p16	NA	IL-6, IL-α, IL-1β, MMP-3, MMP-9	Positive	Jiang et al., 2023

Inducer heterogeneity

In cell culture, senescence is often categorized according to the initial stressor, which can include replicative senescence driven by telomere erosion, oncogene-induced senescence, and stress-induced senescence from oxidative damage, DNA damage, or mitochondrial dysfunction.^5,6,24,25 The stresses encountered by cells in the CNS with age and disease can include chronic inflammation, ischemic injury, radiation exposure, metabolic dysregulation, and accumulation of misfolded proteins. This frequently results in distinct molecular and functional phenotypes even within the same cell type^14,15. For example, models of acute injury, such as ischemia or traumatic injury, induce robust DNA damage responses and oxidative stress,^26,27 whereas neurodegenerative disease models are more strongly associated with chronic inflammatory signaling and proteostatic stress.^28–30 Consequently, senescence phenotypes observed across experimental systems likely differ depending on the initiating insult, which may or may not be known.

These differences complicate direct comparison across studies. Replicative exhaustion in cultured cells captures aspects of telomere-driven senescence but does not recapitulate the inflammatory and metabolic stresses present in the aging CNS.^{5,6,10,24,31–34} Conversely, acute injury models may induce rapid senescence-like states that differ from those that emerge during gradual aging.^{10,26,32–37} Understanding how these distinct stressors shape senescence programs remains essential for interpreting experimental findings across CNS models.

Cell cycle arrest program heterogeneity

Another important source of senescent cell heterogeneity is the relative contribution of the canonical senescence effectors of p16 and p21. Both molecules can enforce durable cell cycle arrest, yet increasing evidence indicates that they may define biologically distinct senescent cell states.^31,38,39 For example, multi-tissue analyses suggest that p16-high and p21-high expressing cells exhibit different transcriptional programs, immune interactions and secretory profiles.^33,40,41 These distinctions may be relevant in the brain and spinal cord, where the initiating stressor and cell type strongly influence which pathway predominates.^33,42

Studies in models of cerebral ischemia illustrate this complexity. Several investigations report robust induction of p16 expression in glia and vascular cell populations following ischemic injury, suggesting that p16-mediated arrest may dominate in this context.³⁵ In contrast, other studies detect more prominent p21 induction, particularly during early DNA damage response following ischemia or oxidative stress.²⁷ These discrepancies likely reflect differences in the temporal dynamics of senescence induction, as p21 is often activated downstream of p53 during acute stress responses, whereas p16 tends to accumulate during persistent or chronic cellular stress. These temporal differences have been well characterized in replicative senescence models, p21 induction early in senescence is tightly linked to interactions with specific cyclin-CDK complexes (i.e. Cdk2, Cdk4/6), and p16 induction later in senescence is associated with durable inhibition of CDK4/6, thus stabilizing the irreversible arrest.^43–45 However, this framework must be interpreted cautiously in CNS tissue, which consists largely of post-mitotic or slowly dividing cells. In these contexts, p16/p21 axes may regulate stress responses, inflammatory signaling or metabolic adaptation rather than strictly enforcing proliferative arrest.^36,43

Single-cell transcriptomic analyses from mice and human tissues further highlight this complexity. Pseudotime analyses across multiple tissues, including brain, suggest that p16-positive and p21-positive cells rarely co-express both markers simultaneously and instead follow divergent transcriptional trajectories.³³ Thus, differences observed between acute injury models and aging studies may reflect distinct senescence trajectories rather than conflicting findings.

Cell-type specific heterogeneity

A major determinant of senescent cell characteristics is the intrinsic biology of each cell type. The CNS comprises a diverse set of specialized cells including microglia, astrocytes, oligodendrocyte lineage cells, and vascular cells, that differ markedly in replicative capacity, metabolic demands, and susceptibility to environmental stress. These intrinsic properties shape both the likelihood that a cell will undergo senescence and the molecular pathways that govern the resulting phenotype (Figure 1).

Figure 1. Context-dependent heterogeneity of senescence in the CNS.

Diverse inducers, cell types, anatomical regions, and cell cycle arrest programs converge to generate heterogeneous senescence phenotypes in the central nervous system. Each line is color-coded to a biological context that leads to senescence in the nervous system and the various regions, cell types and senescence phenotypes that context has been shown to be involved. Lines with lighter shaded colors denote region, cell-type or phenotype that biological context has been implicated in.

Microglia

Microglia, the resident immune cells of the CNS, exhibit a complex relationship with senescence due to their intrinsic immunological plasticity. These cells continuously transition between activation states in response to injury, infection or protein aggregation, making it challenging to disentangle bona fide senescence from chronic inflammatory activation.^46,47 In vitro models of replicative senescence in cultured microglia demonstrate many canonical features of senescence, including telomere shortening, increased SA-β-Gal, and upregulation of p16, p21 and p53,⁴⁸ providing evidence that microglia have the potential to exhibit canonical senescence features. However, the extent to which these features manifest in physiologically relevant contexts is still uncertain.

Microglia isolated from aged brains reveal more modest differences in classical senescence markers and retain proliferative capacity, suggesting that aging-associated microglial dysfunction may not fully correspond to classical senescence.^34,48–51 Evidence from human aging studies similarly indicates that microglia undergo a progressive decline in homeostatic function with age, characterized by reduced motility, impaired phagocytosis, and altered inflammatory signaling.⁴⁷ Morphologically, aged microglia frequently exhibit dystrophic features, including fragmented processes, cytoplasmic beading and shortened ramifications.⁵² These dystrophic microglia accumulate in aging human brains and in neurodegenerative disorders, and have been proposed to represent a senescence-like phenotype distinct from microglial activation.^52,53 In other words, aging-associated microglial dysfunction may be better conceptualized as a distinct, senescence-like state that diverges from classical definitions (Table 1).

In mouse models of tauopathy, senescent glia—including microglia—accumulate in regions with tau pathology and exhibit increased expression of p16, p21 and SASP.¹⁴ Genetic or pharmacological clearance of these senescent cells reduce tau pathology and improves cognitive function, suggesting that senescent glia contribute directly to tau disease progression.^14,16 At the transcriptional level, microglial senescence programs may overlap with other recently described microglial states identified by single-cell transcriptomics. For example, disease-associated microglia (DAM) emerge in response to amyloid pathology and are characterized by distinct transcriptional programs involving lipid metabolism, phagocytic pathways and inflammatory signaling.⁵⁴ Although DAM states are not synonymous with senescence, some DAM-associated transcriptional features overlap with senescence-associated inflammatory signaling pathways.^50,53,55,56 These observations support a model in which microglial senescence exists along a continuum that partially overlaps with other activation states, complicating efforts to define and therapeutically target senescent microglia in disease (Table 2).

In contrast to the more ambiguous phenotypes observed in aging and disease, DNA damage-inducing stressors such as irradiation elicit a more canonical senescence program in microglia. Ionizing radiation leads to increased SA-β-Gal+ microglia, elevated p16 expression and persistent inflammatory cytokine production.^57,58 In these models, senescence appears to be driven primarily through activation of the p16-RB pathway with comparatively weaker induction of p21 signaling, highlighting how the relative contribution of p16 and p21 pathways depends strongly on the initiating insult (Table 2).

Astrocytes

In contrast to microglia, astrocytes often exhibit a more consistent senescence phenotype across a wide range of contexts that involve amyloid-β (Aβ) plaques, tau aggregates, oxidative stress, aging, or irradiation. One of the earliest demonstrations of astrocytic senescence reported an increase in number of p16-positive astrocytes in both aged and Alzheimer’s disease brains, as well as demonstrated that Aβ exposure induces p16 expression and SA-β-Gal activity in cultured human astrocytes.⁵⁹ These findings established astrocytes as a cell type capable of mounting a canonical senescence response in both physiological aging and neurodegenerative disease contexts (Table 1, Table 2).

Subsequent studies across multiple experimental models have confirmed that astrocyte senescence involves increased expression of p16, p21, DNA damage and SA-β-Gal activity, accompanied by acquisition of pro-inflammatory secretory phenotype and reduced neurotrophic support.^59–64 Although the composition of the astrocytic SASP varies across contexts³³, a core set of NF-κB-regulated cytokines, including IL-1β, IL-6, TNF-α, are consistently upregulated.^59–63 Functionally, this shift is associated with impaired neuronal support and increased propagation of inflammatory signaling within the CNS microenvironment.

Interestingly, in vitro studies across different disease models show that treatment of senescent astrocytes with neurotrophic factors can partially restore their neuroprotective capacity across diverse neuronal populations, attenuate inflammatory signaling, and downregulate p21 expression.^60,61,64,65 These findings suggest that, unlike other senescent cells, astrocyte senescence may retain a degree of functional plasticity, raising the possibility that aspects of this state are modulable.

Altogether, these observations support a model in which astrocytes adopt a relatively stereotyped senescence program that shifts them from neuro-supportive cells towards inflammatory amplifiers, while still retaining partial plasticity that may be therapeutically targetable.

Oligodendrocyte Progenitor Cells

Unlike most mature cell types in the CNS, oligodendrocyte progenitor cells (OPCs) retain proliferative capacity throughout adulthood and are responsible for generating new oligodendrocytes during homeostasis and in response to demyelinating injury. Because of this proliferative potential, OPCs might be expected to undergo classical replicative senescence. However, several experimental observations suggest that senescent OPC biology diverges from canonical senescent fibroblast models.

Early work demonstrated that OPC proliferation is governed by intrinsic developmental timing mechanisms rather than by telomere-dependent replicative exhaustion. In vitro studies show that OPCs can undergo multiple rounds of division when maintained in the presence of mitogenic signals such as PDGF and FGF without entering stable proliferative arrest.^66–68 Instead of undergoing irreversible arrest, OPCs often exit the cell cycle through differentiation into mature oligodendrocytes or through transitions into reversible quiescent states regulated by environmental signals.^67,68 These observations suggest that OPC biology appears to be incompatible with classical replicative senescence, where cell fate decisions favor differentiation or quiescence over irreversible arrest (Table 1).

Evidence for OPC senescence has emerged relatively recently. In Alzheimer’s disease models, OPCs located near amyloid plaques exhibit classical markers of cellular senescence, including p16, p21 and SA-β-Gal, whereas neighboring mature oligodendrocytes and astrocytes do not display comparable signatures.¹⁷ Aggregated Aβ can directly induce senescence in cultured OPCs, suggesting that disease-associated proteotoxic stress selectively targets this lineage. Furthermore, selective elimination of senescent OPCs in APP/PS1 mouse models of Aβ pathology reduces neuroinflammation, decreases amyloid burden, and improves cognitive performance, indicating that OPC senescence may contribute to disease progression (Table 2).

Beyond neurodegenerative disease, stress-induced dysfunction of OPCs has also been observed in aging and inflammatory contexts. Aging OPC populations show reduced proliferative capacity and impaired differentiation into mature oligodendrocytes, changes that correlate with diminished remyelination efficiency in aged CNS tissue.⁶⁹ These deficits are often accompanied by increased expression of cell-cycle inhibitors, suggesting activation of senescence-associated pathways.⁷⁰ A recent study demonstrated that defects in autophagy within aged OPCs, trigger expression of p16, p21 and SASP that impair myelination, long-term potentiation and cognitive behavior in aged mice; notably, senolytic treatment or neutralization of the SASP components, CCL3 and CCL5, partially rescued these deficits.⁷¹ Together, these findings suggest that OPC senescence emerges not only in response to disease-specific proteotoxic stress but also during normative aging, where defects in cellular homeostasis, such as impaired autophagy, can drive senescence-associated programs that compromise remyelination and neuronal function.

Neurons and Mature Oligodendrocytes

Both, neurons and mature oligodendrocytes, are terminally differentiated cells that do not normally undergo cell division in the adult CNS. This raises the important conceptual question: can post-mitotic cells exhibit senescence-related alterations independent of cell cycle arrest?

Early studies demonstrated that neurons exposed to genotoxic stress, oxidative damage or irradiation develop hallmark features of senescence, including SA-β-Gal, DNA damage foci, and upregulation of p21 and p16.^72–74 In particular, Jurk and colleagues showed that neurons subjected to DNA damage accumulate double-strand DNA breaks and exhibit a senescence-like transcriptional profile associated with chronic inflammatory signaling.⁷³ In aging brain tissue, neurons can accumulate senescence features while remaining metabolically active and functionally integrated within neural circuits.^72–74 However, interpretation of senescence markers in neurons is complicated by observations that certain neuronal populations, including cells in the developing and adult olfactory bulbs, exhibit SA-β-Gal activity under physiological conditions.^75–77 These findings suggest that SA-β-Gal activity in neurons may sometimes reflect elevated lysosomal activity rather than bona fide cellular senescence, highlighting the importance of combining multiple molecular markers when defining senescent neuronal states (Table 1, Table 2).

Evidence for similar stress-associated states in mature oligodendrocytes, myelin-producing cells that enable saltatory conduction and aids in neuronal metabolic support, has emerged more recently. Experimental models in which oligodendrocytes are exposed to inflammatory signaling, oxidative damage or dysregulated NF-kB activation demonstrate induction of senescence markers including p16, p21, p53 and SA-β-Gal, along with transcriptional programs associated with DNA damage response and metabolic dysfunction.^78,79 These changes are accompanied by impaired myelin maintenance and white matter degeneration suggesting that senescence-like oligodendrocyte states may contribute to age-associated white matter decline. In aging studies, oligodendrocytes that accumulate oxidative DNA damage, mitochondrial dysfunction, and altered inflammatory signaling correlate with impaired myelin maintenance and reduced support for axonal function.^78,79 While these features do not necessarily represent senescence, they overlap with many of the molecular hallmarks observed in other senescent cell types and are consistent with a broader stress-induced senescence phenotype.

Altogether, available evidence suggests that neurons and oligodendrocytes can adopt senescent-like states, though the biological significance of these states likely varies on the initiating stressor and disease context. Interpretation remains challenging, because many studies rely on a limited number of senescence markers making it difficult to distinguish senescence from other forms of cellular dysfunction. Consequently, defining the molecular mechanisms that govern these senescence-like states in post-mitotic cells remains an important challenge for understanding how senescence contributes to CNS aging and disease.

Endothelial Cells and Pericytes

Vascular cells represent another important but relatively understudied source of senescence within the CNS. Endothelial cells and pericytes are critical for maintaining blood-brain barrier (BBB) integrity, regulating cerebral blood flow and coordinating communication between vascular and neural compartments. Because they are continuously exposed to systemic stressors, vascular cells may be particularly susceptible to senescence during aging and disease. Increasing evidence suggests that senescence within the neurovascular unit contributes to BBB dysfunction, neuroinflammation, and impaired tissue repair, although the molecular features of endothelial and pericyte senescence appear to vary depending on the initiating stressor and anatomical context.^16,80–82

In vitro studies of brain microvascular endothelial cells, which line vessel walls and form the primary BBB, demonstrate that repeated passaging or exposure to oxidative stress induces classic senescence features of increased p16 and p21, SA-β-Gal and persistent DNA damage response. These changes are accompanied by alterations in endothelial barrier properties such as reduced tight junction integrity and increased permeability.⁸² In vivo, aging brain vasculature shows increased expression of senescence markers in endothelial cells and increased transcriptional signatures associated with inflammatory signaling and oxidative stress, suggesting that endothelial senescence may contribute to gradual decline in BBB integrity observed during aging.^32,80 Furthermore, genetic deletion of p16 in endothelial cells has shown to improve clinical decline and BBB permeability in tauopathy models.¹⁶

Pericytes, which reside along the abluminal surface of capillaries and provide structural support to the BBB, also exhibit senescence-associated phenotypes during aging and disease. Aging studies have shown that pericyte density declines in the aged brain, accompanied by increased expression of p16, p21 and inflammatory transcriptional signatures.^33,83,84 These changes are often accompanied by morphological alterations including pericyte hypertrophy, cytoplasmic vacuolization and detachment from endothelial cells, all of which can compromise vascular stability. In models of Alzheimer’s disease, pericyte dysfunction is associated with increased senescence marker expression, reduced clearance of toxic metabolites from the brain and impaired vascular integrity.³⁰

Despite these emerging observations, the molecular definition of endothelial and pericyte senescence in the CNS remains incomplete. Many studies rely on limited number of markers, and the relative contribution of p16 or p21 mediated pathways remains poorly defined across different disease models. Because vascular cells are highly responsive to environmental signals, it can be challenging to distinguish stable senescence programs from transient stress responses or inflammatory activation.

Microenvironmental heterogeneity

Cells of the CNS reside in specialized anatomical niches that differ markedly in metabolic demand, vascularization, contact rate with immune cells, and regenerative capacity, all of which collectively determine the nature of stresses they experience. These regional differences likely influence how senescence programs are induced and maintained across the nervous system.^73,74

Within the brain, several studies suggest that senescence-associated molecular programs are not uniformly distributed across anatomical regions. For example, analyses of aging cortical tissue frequently report increased expression of DNA damage response genes and p21-associated pathways, consistent with activation of p53-mediated senescence programs in neuron and glia.⁷³ In contrast, cerebellar aging appears to involve comparatively lower levels of inflammatory activation and senescence-associated gene expression, suggesting that neuronal and glial populations within this region may be less susceptible to certain senescence-inducing stressors such as chronic inflammation.⁷⁴ The hypothalamus also has its own distinct senescent environment. In this region, inflammatory signaling from microglia and astrocytes can trigger NF-kB-dependent pathways that drive systemic metabolic aging.⁸⁵ Experimental manipulation of hypothalamic inflammatory pathways influences lifespan in mice, indicating that inflammatory signaling within this region plays a central role in systemic aging. Because chronic inflammation is a key component of SASP, these findings raise the possibility that senescence-like inflammatory programs within hypothalamic glial populations may contribute to organismal aging. These observations may suggest that distinct molecular mediators of senescence may predominate in different brain regions with DNA-damaged associated pathways enriched in cortical tissues, and inflammatory signaling pathways playing a more prominent role in hypothalamic aging.

Regional heterogeneity in senescence programs is also evident in the spinal cord. Compared with many brain regions, the spinal cord contains large motoneurons with high metabolic demands and extensive axonal projections, making these cells particularly vulnerable to oxidative stress and mitochondrial dysfunction. In models of spinal cord injury, senescence associated-markers, including p16 and p21, accumulate in astrocytes and microglia surrounding lesion sites, suggesting that senescence in this context is largely driven by inflammatory and tissue damage signals.⁸⁶ Consistent with a functional role for these senescent cell populations, recent work by Mannarino and colleagues demonstrated that senolytic treatment alleviates chronic back pain, supporting a link between senescence-associated pathways and central sensitization.⁸⁷ In neurodegenerative diseases affecting the spinal cord, such as amyotrophic lateral sclerosis (ALS), spinal cord tissues display increased expression of senescence markers in both neuronal and glial populations together with inflammatory cytokines and oxidative stress pathways, indicating that multiple pathways, including mitochondrial dysfunction, may contribute to senescence-like states in this disease.⁸⁸ Recent transcriptomic analyses of aging rat spinal cords revealed that aging-associated transcriptional changes were localized to distinct anatomical domains across dorsal and ventral spinal cord compartments.⁸⁹ Notably, regions enriched for oligodendrocyte lineage cells showed increased expression of lipid metabolism and ferroptosis-associated gene signatures, suggesting that oxidative lipid damage may represent a dominant stress pathway in these areas. Because lipid peroxidation and metabolic stress are also associated with senescent inflammatory signaling, these findings raise the possibility that pathways that confer ferroptosis-resistance may converge with senescence-associated pathways in the myelin-rich spinal cord.

Another useful framework for understanding spatial variation in senescence across the CNS is the distinction between gray matter and white matter environments, which differ in cellular composition and metabolic stressors. White matter tracts are enriched for oligodendrocyte lineage cells and contain large quantities of lipid-rich myelin, creating conditions that favor oxidative lipid damage and metabolic stress. Single cell transcriptomic analysis has identified oligodendrocyte populations with aging-associated transcriptional states enriched for stress response genes, inflammatory mediators and metabolic dysfunction.⁴⁶

Microglia in white matter regions similarly adopt transcriptional programs characterized by increased lipid metabolism and phagocytic activity, reflecting their roles in myelin turnover and debris clearance.⁹⁰ These white matter-associated microglial states interact closely with oligodendrocyte lineage cells and may promote senescence-like phenotypes in oligodendrocyte progenitors, particularly under conditions of chronic inflammation or demyelination.⁹¹ In contrast, gray matter microglia interact primarily with neurons and synaptic structures and may therefore influence neuronal stress responses rather than myelin maintenance.^92,93 Consistent with this distinction, studies of aging non-human primate spinal cord and human spinal cord biopsies have identified gray matter-associated microglial populations expressing CHIT1 as essential to driving motoneuron aging that could be partially rescued by administration of ascorbic acid.⁹⁴ These findings suggest that senescence-associated inflammatory signaling may differ depending on whether microglia are embedded within oligodendrocyte-rich or neuron-rich environments.

Evidence of regional differences in senescence is also emerging in the peripheral nervous system (PNS). Transcriptomic analyses of aging dorsal root ganglia (DRG) have shown increased expression of genes associated with inflammatory signaling, oxidative stress and cellular damage responses.⁷² These changes suggest that peripheral sensory neurons accumulate stress associated molecular features that overlap with senescence pathways, although the specific molecular mediators differ from those observed in CNS neurons. In addition to neuronal changes, Schwann cells in peripheral nerves can adopt senescence-like phenotypes in response to injury or chronic inflammation, characterized by inflammatory signaling and altered support for axonal regeneration.⁹⁵

III. Single-cell resolution technologies to characterize ‘senotypes’ in CNS aging and disease research

Single-cell and spatial transcriptomics to evaluate cell-type specificity of senescence

In the last decade, single-cell RNA-sequencing (scRNA-seq) and single-nucleus RNA-sequencing (snRNA-seq) have emerged as powerful tools for characterizing rare and heterogeneous populations of cells, making them especially advantageous for studies on senescent cells. Traditional bulk transcriptomic approaches obscure the senescent cell diversity and may fail to detect a senescence signature altogether due to low total abundance of senescent cells. Single-cell and single-nucleus platforms enable high-resolution profiling of individual cells in diverse cell populations and have been critical for characterizing senescent cells in aged and diseased brains. Perhaps the most powerful insights provided by these single-cell resolution technologies revolve around understanding the cell-type specificity of senescent cells in different CNS contexts.

To date, single-cell and single-nucleus transcriptomic studies have revealed profound age-induced and cell-type specific changes in the aging mouse^{32–34,80,96–101} and human^33,102–104 brain. These studies demonstrate that brain aging is accompanied by cell-type-specific transcriptional shifts that are reminiscent of senescence, including cell cycle arrest, upregulation of inflammatory pathways, metabolic dysregulation, and changes to cellular functions in many different CNS cell types. Many of these studies highlight aging-induced senescence signatures particularly in glial cells such as microglia^{32,34,96,97,100,104,105}, astrocytes^80,104,106, and oligodendrocyte lineage cells^97,98,105. Interestingly, microglia with aging-induced senescence signatures are also associated with increased inflammatory signaling and transcriptional states similar to those of DAM.^34,100 Other studies explore age and senescence signatures in endothelial cells or ependymal cells, hypothesizing that these could underlie age-related deterioration in the blood-brain- and blood-cerebrospinal-fluid-barriers.^{32,80,96,97,99,104,105} The majority of single-cell transcriptomic studies in aged brain find no senescence-related changes in neuronal cell populations, but a handful of studies have reported evidence of senescence in specific neuronal subtypes.^98,103 Several studies have also performed spatial transcriptomics on aged mouse brains which help isolate regionality of age-related changes. These studies find that age-related changes overlap with senescence signatures most abundantly in inflammatory hotspots in white matter tracts, further validating that glial cells may be the most vulnerable brain cell types to undergo senescence.^96,105,106

Single cell transcriptomic studies in mouse models of brain disease predominantly point toward microglia as the senescent cell type driving neuroinflammation and Alzheimer’s-like pathology.^55,107,108 In an LPS-induced neuroinflammation model, microglia exhibit transcriptional signatures of senescence, dysregulated phagocytosis, and lysosomal stress along with histological indications of senescence that are absent in astrocytes or neurons.¹⁰⁸ Importantly, pre-treatment with the senolytic ABT-737 prevents inflammatory changes and preserves cognition in this model, indicating that senescent microglial likely exacerbate damaging neuroinflammation.¹⁰⁸ Additionally, single-cell transcriptomics have implicated senescent microglia in both an amyloid-plaque model and a tau-tangle model of Alzheimer’s disease (AD).^55,107 These studies also suggest that senescent microglia are a subtype of DAM, which have long been causally linked to neurodegeneration.

In post-mortem brains of humans with AD, results from single-cell transcriptomic studies are mixed with evidence for senescence in glia, vascular cells, and neurons.^{103,107,109–112} Similar to findings in mouse models of brain disease, several human studies present strong transcriptional evidence for senescence in microglia.^110–112 Furthermore, senescence signatures are reduced in microglia from patients with a TREM2 variant that is associated with reduced microglial responsiveness¹¹⁰ and increased with myelin treatment in cultured microglia¹¹¹, suggesting that cellular stress may push microglia in a diseased brain into a senescent state. These studies also identify less-pronounced features of senescence in other non-neuronal cell types such as astrocytes, oligodendrocyte-lineage cells, and/or endothelial cells with cell-type specific differences in senescent signatures.^110–112 While the majority of single cell transcriptomic studies point toward senescence in non-neuronal cells types, there are a couple of studies in human AD brains with evidence for senescence in certain populations of excitatory neurons.^103,109 Mechanistically, these senescent neurons may be the result of transcriptional re-commitment to cell cycle¹⁰³ and be able to induce paracrine senescence in glial cell types in culture¹¹³.

Senescent cells as dynamic and heterogenous rather than a static state

The SenNet consortium, a collaborative research initiative to identify, map, and characterize senescent cells, recently published an emerging framework which they term ‘senotype’ to describe the dynamic and context-dependent trajectory shaped by diverse molecular and environmental cues influencing senescent cell heterogeneity.¹¹⁴ Differing senotypes likely exhibit unique combinations of transcriptional identity, immune interactions, SASP, chromatin state, and metabolic activity. This trajectory-based view positions senescence not as the endpoint of a binary fate decision, but as a continuum influenced by cell types, inducing stressor, tissue microenvironment, and time. Building out the senotype framework with integrated multi-omic studies that combine genomic, epigenomic, transcriptomic, proteomic, metabolomic, and spatial data offers a powerful path to resolve senescent cell diversity with greater precision. Precise application of this framework and these tools to the brain will be essential for enabling a more nuanced understanding of how senescence contributes to brain aging and disease.

In summary, single-cell resolution transcriptomics have been instrumental in identifying which cell types may be senescent in aged and diseased brains. These technologies have also characterized context-specific senescence-associated transcriptional signatures. which help to generate hypotheses about how different senescent cell types may ultimately contribute to cognitive dysfunction. Additionally, integration of single-cell and/or single-nucleus transcriptomics with single-cell epigenomics or spatial transcriptomics has the potential to provide more nuanced insights into the role of senescent cells in microenvironmental disturbances. Such granular characterization not only advances our understanding of brain aging but also identifies potential therapeutic targets for mitigating senescence-driven pathology in prevalent and devastating brain disorders.

IV. In vivo animal studies to characterize the functional relevance of senescent cells in CNS aging and disease

Advances in single cell technologies have allowed researchers to explore cell type specificity of senescent cells in the aged and diseased brain in service of better understanding how senescence mechanistically contributes to cognitive dysfunction. Despite the recent advances in this space, several large questions remain. Are some brain cell types truly senescent or just immune-activated or terminally differentiated? Are senescent cells causal drivers of brain dysfunction or just secondary amplifiers? Resolving these questions requires experimental systems that move beyond molecular description, with the laboratory mouse remaining a uniquely powerful platform for validating senescent cell biology and defining their functional consequences in vivo.

Senescence, activation, or terminal differentiation? Simple question, complicated answer

One particularly challenging aspect of studying senescent cells in the brain is defining senescence in immune-like cell types, like microglia, or terminally differentiated brain cell types, like neurons, because these cells naturally share transcriptional, functional, and morphological features with senescence programs. This ultimately makes it difficult to distinguish senescence from baseline activation or differentiation signatures.

The classic example of the activation versus senescence conundrum, is the DAM cell type. As single-cell technologies developed, DAMs were first described in neurodegenerative contexts as a distinct microglia population marked by down regulation of homeostatic transcriptional programs and upregulation of phagocytosis, lipid-metabolism, and inflammatory states.¹¹⁵ DAMs display an array of partially overlapping phenotypes with senescent cells that complicate studies in this space. For example, DAMs exhibit upregulation of inflammatory molecules that resemble the SASP. To further complicate the matter, recent scRNAseq experiments have identified DAM subsets that also show signs of senescence in various mouse models of plaque or tangle-based neuropathology.^55,107 Several studies also propose mechanisms whereby disease induces cellular stress which then triggers the senescent state in microglia^110,111, suggesting that it’s possible that senescent microglia actually result from certain DAM states. Interestingly, overlapping DAM and senescence transcriptional signatures have also been in reported in microglia from non-diseased aged mice.^34,100 More work is needed to truly define the existence and consequences of senescent microglia in brain disease and aging. Distinguishing true microglial senescence will require careful integration of transcriptional markers, morphological criteria, metabolic shifts, and contextual cues.

Another example of a cell type with senescence features that are difficult to resolve is the terminally differentiated neuron. Neurons and senescent cells share the core characteristic of cell-cycle withdrawal, as well as related metabolic and transcriptional adaptation, making them somewhat difficult to distinguish from one another. Neurons are also long-lived cells that naturally accumulate DNA damage, metabolic stress, lipofuscin, and protein aggregates over time, all of which can resemble senescence phenotypes without necessarily reflecting a true senescence program.¹¹⁶ Despite these challenges, a handful of studies have identified subpopulations of excitatory neurons as senescent in the brains of humans with AD.^103,109 However, the gene sets used to define senescent neurons in one of those studies were applied to an independent human AD where they instead pointed toward microglia, rather than neurons, as exhibiting senescence.¹¹⁰ These discrepancies in defining neuronal senescence in AD are puzzling and require careful future study.

Are senescent cells causal drivers or secondary amplifiers in neurodegenerative disease?

Senescent cells are increasingly implicated in neurodegenerative disease and age-related cognitive decline, but whether they act as primary causal drivers or as secondary amplifiers of ongoing pathology remains a central question. Evidence from studies on Alzheimer’s disease suggests that cellular senescence can emerge directly from disease-initiating insults. In particular, studies have shown that amyloid-beta treatment can induce senescence features such as SA-β-gal activity, cell cycle arrest, and SASP production in cultured astrocytes^59,117 and neurons¹¹⁸. Other studies have shown that exposure to various forms of tau in culture can similarly trigger senescence features in astrocytes^119,120 and microglia¹²¹. These findings suggest that cellular senescence might come secondary to disease. However once induced, senescent cells also appear to acquire various levels of cellular dysfunction which, in turn, exacerbate disease. For instance, microglia with senescent like qualities induced by tau or amyloid-beta have been shown to also exhibit reduced phagocytic capacity¹²² and ability to clear tau in culture¹²¹. In this way, senescent cells may be induced by early pathogenic insults and then go on to act as secondary amplifiers of disease. This mechanism fits with transcriptomic-based hypotheses that senescent microglia arise from DAM populations^110,111 and further exacerbate a volatile microenvironment.

Another model for how senescent cells contribute to brain disease is that senescent cells accumulate with age in the brain environment before disease onset and actively promote the conversion to disease state. This hypothesis is supported by the overwhelming evidence that senescent cells accumulate in normative aging in many tissues, including the brain.^123,124 In this way, senescent cells may actually act as causal drivers of disease by seeding the brain environment with pro-inflammatory and dysfunctional cells.

Thus, current evidence supports a dual role: senescence can be both a downstream response to early neurodegenerative triggers and a potent feed-forward mechanism that accelerates disease progression, making it an attractive therapeutic target regardless of its temporal position in the pathogenic hierarchy. However, dialing in whether senescent cells truly drive in the initiation of neurodegenerative disease or simply partner in its cascade will be imperative for developing effective therapeutic strategies. To both answer these outstanding questions and to test therapeutic potential, studies must move from culture-based systems and post-mortem human brain tissue interrogation into pre-clinical models like the laboratory mouse.

Moving beyond description: mouse modeling as a cornerstone tool for addressing causality and therapeutic relevance

The laboratory mouse has long served as a cornerstone of biomedical research due to its unique combination of genetic tractability, physiological similarity to humans, and suitability for controlled experimental manipulation. Its relatively short lifespan, well-characterized genome, and the availability of sophisticated genetic engineering tools enable precise modeling of human diseases and mechanistic interrogation of complex biological processes, such as cellular senescence. Mouse models overcome several major hurdles inherent in studying senescent cells. They help to bridge high-dimensional molecular findings and biological relevance by acting as a backdrop for validation of large-scale transcriptomic findings. They also help to validate the presence and biological consequences of senescent cells in various contexts and offer a crucial mechanism to test causality via senolytic treatment paired with functional outcomes testing.

Although scRNAseq has greatly advanced the ability to resolve senescent cell heterogeneity, these approaches remain constrained by technical limitations including drop-out effects, under-detection of low-abundance transcripts, and the loss of spatial context. Such challenges are particularly acute for senescent cells, which are large, fragile, and therefore disproportionately lost during tissue dissociation. This means that scRNAseq-defined senescent populations likely represent only a subset of those present in vivo. Spatial transcriptomic technologies can mitigate issues related to tissue dissociation and drop-out effects, yet they currently lack true single-cell resolution and remain comparatively low-throughput and costly. Moreover, transcriptomic data alone typically cannot meet field-wide guidelines for studying cellular senescence, which emphasize the detection of at least three senescence markers and demonstration that reducing senescent cell burden modifies the condition of interest.¹⁰ To address these shortcomings, in vivo experimentation is required to validate omics characterizations of senescent cells and establish their functional relevance in the aging and diseased brain.

Studies that pair large-scale omics with in vivo interrogation using the laboratory mouse have yielded some of the most nuanced insights into the role senescent cells play in the aging and diseased brain. This integrative approach not only strengthens causal inference but also enables researchers to pinpoint the specific cell types and brain regions most enriched for senescence, critical information for future development of targeted therapeutics. Using this discover-and-validation framework, two independent studies have identified white matter senescent cells as the primary contributors to age-induced senescence in the brain by combining omics-based characterization with techniques best done in live mice and/or with fresh tissue such as immunofluorescence-based staining, MRI-based imaging, and primary cell culture.^34,106 Another investigation demonstrated senescent microglia promote neuroinflammation-induced cognitive dysfunction through characterization by scRNAseq and follow-up mouse experiments including senescent cell elimination, microscopy, electrophysiology, and behavioral testing.¹⁰⁸ And perhaps most clinically significant, the combination of scRNAseq and in vivo validation techniques has also identified a population of TREM2-expressing senescent microglia in a mouse model of AD as targetable contributors to neuroinflammation and cognitive impairment.⁵⁶ These studies are powerful examples of the ability of in vivo validation to transform omics-derived signatures into biologically and therapeutically meaningful insights.

Some of the most compelling evidence for the functional relevance of senescent cells in brain aging and disease comes from mouse studies with direct testing of causality through senolytic interventions paired with functional outcomes testing. These studies in mice allow researchers to link the removal of senescent cells with measurable improvements in physiology and behavior, largely through various cognitive performance tests. With this methodology, researchers have been able to show that senescent cell removal with various senolytics improves cognitive function in aged mice^{80,98,125,126} and in mouse models of Alzheimer’s disease^{14,17,56,127,128}. Additionally, this approach has allowed researchers to identify context-specific SASP that is modifiable with senescent cell targeting therapies in vivo.^71,98 In addition to pharmacological agents, senescent cells can efficiently and specifically be targeted in a number of transgenic mouse models reviewed in depth by others.¹²⁹ Combining targeted senescent cell clearance and functional outcome assessment offers a uniquely powerful framework for identifying if senescent populations are truly pathogenic and for validating their roles in vivo.

Mouse modeling is an indispensable platform at the intersection of cellular senescence and brain aging research. Mouse studies enable validation of putative transcriptional senescence signatures generated from omics workflows and, crucially, allow researchers to test causality through senolytic interventions paired with functional, behavioral, and molecular readouts. By integrating discovery-driven molecular profiling with in vivo mechanistic experimentation, the laboratory mouse continues to provide an unparalleled level of physiological relevance and experimental control in rigorous biomedical science.

V. Conclusion

Cellular senescence in the brain is a spectrum of context-dependent states that vary by cell type, trigger, microenvironment, and functional consequence. Single-cell transcriptomic technologies have been pivotal for describing this heterogeneity across brain aging and neurodegenerative disease, exposing senescence programs in glia, vascular cells, and neurons. However, technical limitations and lack of spatial context mean that omics experiments alone cannot fully define senescence or its consequences in the living brain. The laboratory mouse remains indispensable for studying how senescent cells impact brain biology; it provides greater tissue abundance, specialized tissue preparation for precise characterization of senescent cells, and therapeutic intervention with functional outcomes testing.

At the crossroads of cellular senescence and brain health, the most substantial advances have, and will continue to, come from the tight coupling of large-scale omics-based discovery with in vivo validation, mechanistic exploration, and therapeutic testing in preclinical mouse models. By iterating between single-cell transcriptomics, ideally from human brains, and rigorously controlled mouse studies, our field can move from cataloging senescent states to pinpointing which ones are targetable drivers of brain aging and disease. This ultimately will inform therapeutic development with meaningful clinical outcomes.