Cell senescence in neuropathology: A focus on neurodegeneration and tumours

Abstract The study of cell senescence is a burgeoning field. Senescent cells can modify the cellular microenvironment through the secretion of a plethora of biologically active products referred to as the senescence‐associated secretory phenotype (SASP). The consequences of these paracrine signals can be either beneficial for tissue homeostasis, if senescent cells are properly cleared and SASP activation is transient, or result in organ dysfunction, when senescent cells accumulate within the tissues and SASP activation is persistent. Several studies have provided evidence for the role of senescence and SASP in promoting age‐related diseases or driving organismal ageing. The hype about senescence has been further amplified by the fact that a group of drugs, named senolytics, have been used to successfully ameliorate the burden of age‐related diseases and increase health and life span in mice. Ablation of senescent cells in the brain prevents disease progression and improves cognition in murine models of neurodegenerative conditions. The role of senescence in cancer has been more thoroughly investigated, and it is now accepted that senescence is a double‐edged sword that can paradoxically prevent or promote tumourigenesis in a context‐dependent manner. In addition, senescence induction followed by senolytic treatment is starting to emerge as a novel therapeutic avenue that could improve current anti‐cancer therapies and reduce tumour recurrence. In this review, we discuss recent findings supporting the role of cell senescence in the pathogenesis of neurodegenerative diseases and in brain tumours. A better understanding of senescence is likely to result in the development of novel and efficacious anti‐senescence therapies against these brain pathologies.

of cell replacement and repair [3]. A few decades later, it is now clear that senescent cells are present in many living organisms, from mice to humans and their presence can either be beneficial or detrimental depending on the biological context [4][5] Figure 1.
Cellular senescence is a survival programme that can be induced by a range of damaging stress signals such as radiation, chemotherapy, replicative stress and oncogenic signalling. Senescent cells are characterised by the stable and irreversible cell-cycle arrest whilst maintaining metabolic activity and viability [4,[6][7][8]. This is distinct from cellular quiescence, which is defined as a reversible proliferative arrest, such as adult stem cells, which can be stimulated to re-enter the cell-cycle by mitogenic signals [9]. In contrast, senescent cells do not proliferate in respond to these signals. However, they can re-enter the cell cycle, mostly in the context of developing cancers, whereby accumulation of genetic or epigenetic alterations results in the disruption of the key molecular pathways maintaining cell-cycle arrest [10][11][12][13].
A hallmark of senescent cells is the activation of a senescence-associated secretory phenotype (SASP), characterised by the synthesis and secretion of a plethora of biologically active molecules (e.g. inflammatory mediators, growth factors, extracellular matrix proteins) [14][15][16]. The SASP underpins the paracrine functions of senescent cells. Senescent cells are involved in essential physiological processes such as embryonic development, immune modulation, tissue regeneration, cell plasticity and reprogramming [17][18][19]. In these contexts, senescent cells are present only transiently to be subsequently eliminated by the immune system [20][21]. In contrast, persistence of senescent cells within tissues results in the deterioration of organ function, which can lead to disease. For instance, age-related conditions such as osteoarthritis, atherosclerosis, fibrosis of the lungs, kidney and heart, sarcopenia, glaucoma, cataracts and type 2 diabetes are all associated with increased numbers of senescent cells [22][23][24]. The repertoire of senescence-associated pathologies has recently been expanded to include neurodegenerative diseases such as Alzheimer's, Parkinson's and multiple sclerosis [25][26][27]. Moreover, in addition to their role in age-related diseases, evidence is mounting that accumulation of senescent cells within tissues may be driving organismal ageing itself [8,[28][29][30][31][32][33][34][35]. in senescence escape, cell-cycle re-entry and tumour cell proliferation [13,44] Figure 2B. However, this view has been challenged,it has been shown that loss of p53 may not be sufficient for senescence escape, but p53-defficient cells are able to bypass the establishment of a senescence programme when targeted to express oncogenic  [51][52] Figure 2E. New compounds capable of selectively killing senescent cells, termed senolytics, have been identified [5]. These chemical drugs inhibit pathways that are essential for survival of senescent cells but are dispensable in non-senescent cellular states. Exploiting these new vulnerabilities in cancer is a very exciting area of research, which is likely to lead to novel, efficacious anti-cancer therapies. Finally, senescent cells can remodel the tumour microenvironment through the SASP creating a permissive setting that allows tumours to progress Figure 2F [37, 53].
In this review we will discuss the role of senescent cells in brain pathologies, in particular age-associated neurodegenerative diseases such as Alzheimer's disease, Parkinson's disease and multiple sclerosis, as well as brain tumours, specifically craniopharyngioma, low-grade glioma, glioblastoma multiforme, medulloblastoma and diffuse midline glioma. We will present evidence accumulated from in vitro and in vivo studies in both mice and humans. The translational implications of such studies will also be discussed. For further reading, we recommend other reviews in the field [4-5, 37, 54-55].

CELLUL AR S ENE SCEN CE AND SA S P
The permanent cell-cycle arrest of senescent cells is mediated principally by the p16 INK4a /Rb and p21 Cip1 /p53 pathways in response to stress stimuli [56][57][58][59]. Expression of p53 due to cellular stress signals activates a multitude of responses, including cell-cycle arrest, which is mediated by the p53 target p21 Cip1 . p16 INK4a mediates cell-cycle arrest by inhibiting CDK4/6 leading to hypo-phosphorylation of RB and inhibition of cell-cycle progression into S phase. p16 INK4 has been termed the master regulator of cell-cycle arrest in senescent cells [57][58]. Activation of the senescence programme leads to further molecular changes: (i) chromatin remodelling (e.g. presence of senescence-associated chromatin foci), (ii) activation of a DNA damage response (e.g. expression of γH2AX), (iii) enlargement of the lysosomal F I G U R E 2 Proposed roles of senescence and SASP in tumourigenesis. (A) Tumour suppressor mechanism. Oncogenic signalling leads to transient proliferation followed by senescence induction (oncogene-induced senescence). Senescent cells activate SASP and attract immune cells that clear them from the tissues, thus preventing subsequent tumour development. (B) Tumour progression by senescence escape or reversion. Following senescence induction, one cell accumulates further mutations (e.g. TP53, encoding p53, or CDKN2A, encoding p16 INK4a ) resulting in senescence escape, cell-cycle re-entry, proliferation and tumour formation. (C) Tumour progression by senescence bypass or evasion. Upon oncogenic signalling and initial proliferation burst, most of the cells become senescent but one cell continues proliferating due to pre-existing mutations in key senescence regulators (e.g. p53 or p16) that prevents the establishment of a senescence response. (D) SASPmediated tumour initiation. Senescent cells, through the SASP, create a pro-tumourigenic microenvironment that leads to cell transformation of a non-tumour cell and tumour formation. (E) Therapy-induced senescence. Following irradiation, chemotherapy and targeted therapy, most of the cells in the tumour bed are killed (e.g. by apoptosis) or induced into senescence with some cells being unaffected (green cell). Senescent cells will eventually contribute to tumour recurrence either in a paracrine manner through SASP-mediated tumour growth, or by senescence escape or bypass. (F) SASP-mediated tumour microenvironment alterations. Senescent cells, through the SASP, can remodel the tumour microenvironment. For instance, by (i) modulating the immune response to create immunosuppressive environment (e.g. M2 macrophage polarisation, T-reg recruitment), (ii) driving extracellular matrix remodelling and tumour vascularisation and (iii) promoting the development of metastatic niches compartment (e.g. increased expression of GLB1 and lipofuscin accumulation); (iv) macromolecular damage (e.g. telomere attrition); (v) deregulated metabolism (e.g. a shift from oxidative phosphorylation to glycolysis); (vi) anti-apoptotic response (e.g. increased expression of BCL-2 family proteins and inhibition of caspase 3) [4,6]. Resistance to apoptotic death is primarily mediated through the stress-induced p53 pathway, which upregulates the expression of anti-apoptotic BCL-2 proteins [60][61]. Additionally, the p53-transcriptional target p21 Cip1 has been shown to be able to directly inhibit caspase 3, hence contributing to apoptotic resistance [62].
The identification of senescence in vivo is difficult, and no single marker can unambiguously be used to define senescent cells.
Initially the expression of β-D galactosidase and its detection in a colorimetric enzymatic assay at acidic pH (SA-β-Gal) was used to define senescent cells. However, this staining can be unreliable in vivo.
Therefore, a consensus has been published, where a combination of markers is recommended to assess cellular senescence [5] Table 1.

ANTI -S ENE SCEN CE AND ANTI -SA S P S TR ATEG IE S
After the discovery of the detrimental role that senescent cells play in ageing and in numerous pathologies, it soon became relevant to develop specific targeted strategies. The first proof-of-concept that the ablation of senescent cells was beneficial and reduced ageingassociated disorders was published in 2011 [28,77]. This was followed by studies showing that the selective killing of senescent cells using chemical compounds improves organ function in ageing mice [33,[78][79]. A non-exhaustive list of current and promising strategies is presented in Table 2. There are four main approaches of antisenescence and anti-SASP strategies currently being investigated.

Prevention of senescent cell accumulation
The chronic reduction of total calorie intake has been reported to counteract several age-associated alterations, through molecular and physiological effects, including prevention of senescent cells accumulation [80][81][82]. As a consequence, caloric restriction mimetics are studied in the context of ageing, particularly among them the modulation of glucose metabolism by 2-deoxy-D-glucose, which has been shown to reduce degeneration of dopaminergic neurons in a Parkinson's disease mouse model [83]. Resveratrol and other polyphenols are also able to suppress the formation of reactive oxygen species (ROS) and to limit cellular senescence in neurons [84][85]. Cells treated with caloric restriction mimetics express molecular pathways similar to cells affected by long-term calorie restriction or short-term fasting, including the autophagy pathway. The crosstalk between autophagy and SASP production is an important element to investigate to better understand the regulation of cell senescence by these drugs.

Ablation of senescent cells: senolytics
Among the senescence hallmarks, the anti-apoptotic programme is not only required for senescent cell survival, but also the easiest to target. Thus, the first senolytic drugs that have been reported are inhibitors of the anti-apoptotic B cell lymphoma 2 (BCL-2) protein family [33, [78][79]. Two of these promising drugs, ABT-263 and ABT-737, have been shown to be capable of selective elimination of senescent cells and causing therapeutic benefits in several physiological and disease contexts (e.g. regeneration [86], cancer [87], type 1 diabetes [88], and atherosclerosis [89]. Other anti-apoptotic pathways have been investigated, in particular the inhibition of the MDM2/p53 interaction (e.g. UBX0101 [90] and USP7 inhibitor [91]. In mouse models, UBX0101 is able to attenuate the development of osteoarthritis by selective clearance of senescent cells, however a phase II trial did not replicate these results (NCT04129944) [92]. Recently, high throughput drug screenings have uncovered the senolytic activity of cardiac glycosides, through a process mediated by induction of the pro-apoptotic BCL2-family protein NOXA [93][94]. Another class of senolytics take advantage of the high lysosomal β-galactosidase activity of senescent cells to deliver more specifically cytotoxic drugs to senescent cells and reduce the toxic side effects [95][96].

Making senescent cells harmless: SASPmodulating drugs
An additional approach to interfere with the detrimental effects of senescent cells is the modulation of their secretome, either by disrupting  More research is needed to identify safe and efficacious anti-senescence therapies able to counteract the detrimental effects of ageing and cancer. As previously mentioned, senescent cells are highly heterogeneous and the activation of specific transcriptomic programmes is dependent on cell type, stress inducers and duration of senescent induction [75,105]. Therefore, the identification of the best anti-senescence approach may need to be tailored to the specific cellular context, whether ageing, specific degenerative disease or cancer.

S ENE SCEN CE IN NEURODEG ENER ATIVE DISE A SE S
Age is the most common risk factor for neurodegenerative diseases [106]. The incidence of conditions such as Alzheimer's and Parkinson's disease, which are characterised by cognitive decline and loss of neurons and synaptic connections, increases with age [107]. Age is also a risk factor for inflammatory diseases such as multiple sclerosis, which show loss of axons, dendrites, and neurons [108]. Senescent cells have been identified in different cell types of the nervous system, including neural stem cells, neurons, astrocytes, oligodendrocytes and microglia [109][110][111][112][113][114][115]. Although neurons are characterised by permanent exit of the cell cycle, they have been shown to accumulate DNA damage and acquire additional features that typify senescence, including SASP activation [113]. These senescent cell types have been implicated in the pathogenesis of Alzheimer's disease, Parkinson's disease, multiple sclerosis, frontotemporal dementia and ischaemia/stroke. Cellular senescence may contribute to the initiation and/or progression of neurodegenerative diseases by promoting chronic inflammation, causing loss of regenerative properties and enhancing age-related decline in the blood-brain barrier and micro-vasculature [116].

Alzheimer's disease
Alzheimer's disease (AD) is the most common neurogenerative disease with an incidence of 11.08 per 1000, doubling every 5 years stem cells of the APP/PS1 AD mouse model [125]. In this study, Aβ fibrils can accelerate neural stem cell senescence via activation of the MAPK pathway, ultimately leading to loss of neurogenesis.

Senescence hallmarks Markers Limitations Biological consequences
Resistance to apoptosis Increased expression of BCL-2 family members Anti-apoptotic protein upregulation

TRAIL-Decoy Receptor DcR2 over expression
Markers not present in mice Hiding from Immune system NKG2D ligands over expression  Higher levels of pro-inflammatory SASP factors has been reported in aged human and mouse brains compared with younger controls [126]. Inflammation is a key feature which contributes to the initiation, severity and progression of most neurodegenerative diseases including AD [127]. Expression of SASP factors, e.g. IL6, IL1B, TGFβ, TNFα and MMP-1, −3 and −10 and activation of the p38MAPK pathway are upregulated in human AD samples and murine models [72,[128][129][130][131][132].
Microglia, the resident macrophages of the central nervous system (CNS) whose functions are tightly regulated by their microenvironment, can secrete SASP factors [133]. Ageing or neurodegenerative accumulation of misfolded protein induces microglia proliferation and promotes an activated state. This state is known as microglia priming and initiates the reactive defence program characterised by phagocytosis and increased release of cytokines, tumour necrosis factor (TNF) and nitric oxide [134]. Primed microglia are also prone to be stimulated by secondary sources of inflammation, triggering an exaggerated and chronic inflammatory response in the CNS [135]. Both aged and AD brain samples show microglia priming and an increase of their pro-inflammatory response [134].
Ex vivo and in vitro studies have revealed that aged microglia secrete higher levels of SASP factors such as IL6 and TNFα compared with young microglia. Aged microglia lose their ability to phagocytose Aβ fibrils and undergo replicative senescence due to telomere shortening [110,[136][137]. Evidence linking age-related senescence and AD pathogenesis has been provided by a study in which aged rat microglia were isolated and treated with Aβ oligomers in vitro. Upon treatment, these activated microglia become senescent, shown by SA-βgal staining and production of IL1B, TNFα and MMP2 [137].
This suggests that age-related senescence in AD microglia may play a role in disease progression.
Primed microglia and neuroinflammation are considered to play key roles in the initiation and progression of AD. An increase in numbers of primed microglia correlates with AD disease progression in humans [138], however, the mechanisms by which these cells could detrimentally affect AD pathogenesis are not yet fully elucidated.
Primed microglia release IL1B and IL18 [139], and in a study on human AD, it has been shown that IL1B induces the secretion of TNFα, promoting the formation of amyloid plaques [140].
Another heterogenous cell population implicated in AD are astrocytes [141]. These cells have diverse homeostatic roles in the CNS including neurotransmitter uptake/recycling, synaptic activity, maintenance of the blood brain barrier and inflammation [142]. Single cell sequencing from wild type and AD mouse models has identified disease-specific astrocytes that are apparent before the onset of neurological phenotypes and are increased with disease progression [138]. These astrocytes express an inflammatory and neurotoxic gene profile that is analogous to that observed in aged wild-type astrocytes (i.e. upregulation of genes involved in development and differentiation, metabolic pathways of lipid and cholesterol, response to toxic compounds and inflammatory signalling, including NfκB signalling and ROS). Furthermore, these upregulated genes have been identified in aged human brain samples from AD post-mortem samples, confirming previous studies in which overexpression of IL6 in murine astrocytes results in the appearance of AD-like neurological symptoms [143] and in the formation of amyloid plaques that are similar to those observed in human AD patients [144][145]. These studies provide evidence that neuroinflammation, caused by the secretion of chemokines and cytokines commonly found in the SASP, contributes to the initiation and progression of AD.

Senolytic therapy in AD
A recent report has identified senescent oligodendrocyte precursors with AD using this senolytic combination [146].
In a study using micro-dissected post-mortem human AD, a senescent transcriptomic profile has been identified in neurons containing neurofibrillary tangles (NFTs) of aggregated tau protein [147].
NFT-accumulating neurons in different AD murine models display a senescent phenotype, evidenced by expression of CDKN2A mRNA.
Treatment with Dasatinib and Quercetin can kill these senescent cells resulting in reduction of both NFT density and neuronal loss.
Collectively, these findings indicate a strong association between the presence of cellular senescence in the brain and neurodegeneration, which is supported by mechanistic studies in murine models. Potential therapeutic avenues that selectively kill senescent cells could revolutionise AD treatments.

Parkinson's disease
Parkinson's disease is the second most common neurodegenerative disease after AD, affecting about 7-10 million people worldwide over the age of 65. It is characterised by the progressive loss of dopaminergic neurons in the substantia nigra pars compacta of the midbrain, leading to progressive motor degeneration. A key pathological feature is the presence of Lewy bodies, composed of aggregates of α-synuclein, a protein involved in DNA damage repair [148]. PD symptoms manifest once 80% of the dopaminergic neurons are lost [149]. Currently there are no chemical treatments that can prevent disease progression.

Evidence of senescence in PD has been shown in various studies.
Higher expression levels of p16 INK4a , p21 Cip1 and inflammatory markers (such as IL6) have been identified in PD patients compared with healthy controls. Increased expression of these factors is associated with faster cognitive decline in the patients [27]. A recent study has revealed that a DNA binding protein, STAB1, which is associated with PD, prevents cellular senescence in dopaminergic neurons in vivo [150]. Senescent astrocytes have been observed in both human PD samples and a PD murine model [115]. Sporadic PD in humans has been associated with exposure to the herbicide paraquat (PQ), and PQ administration to mice is sufficient to induce PD-like phenotypes.
Human astrocytes cultured in vitro with PQ show positive SA-βgal staining and reduced proliferation. Conditioned medium from these senescent astrocytes reduces human DA neurone viability and decreases neural stem cell proliferation [115]. The genetic ablation of

Multiple sclerosis
Multiple sclerosis (MS) is an autoimmune disease causing severe physical incapacitation and neurological damage, affecting around 2.5 million people worldwide [156]. The debilitating causes of MS are due to CNS demyelination and neurodegeneration with limited remyelination. Infiltrating macrophages and lymphocytes cause multifocal inflammation and oligodendrocyte cell death, which lead to demyelination, neuronal and axonal loss, and generation of CNS plaques that contain inflammatory cells and demyelinated axons.
The aetiology of MS remains unknown, but certain genetic and environmental factors might influence the likelihood of developing MS [108]. Several immunosuppressive and immunomodulatory treatments are available,however, disease progression is still common.
The presence of senescent cells with activated SASP has been observed in mouse models and human MS samples. Using a gliotoxin-induced demyelination MS murine model, it has been reported that aged mice show slower rates of remyelination than younger mice, suggesting that age-related senescence could play a role in the onset and progression of this disease [157]. In another mouse model, in which demyelination is induced by feeding the mice cuprizone, increased numbers of SA-βgal-positive glial cells have been identified in demyelinating fibres of the corpus callosum [158]. In comparison with age-matched control tissue, demyelin-  [166][167]. Like human ACP, the mouse tumours contain cell clusters expressing senescence markers.
Interestingly, genetic tracing in the ACP mouse models has revealed that the clusters derive from SOX2+ve stem cells expressing oncogenic β-catenin. In contrast, the tumours are derived from a different cell lineage and do not express oncogenic β-catenin [167]. This initial study led to propose a paracrine model of tumourigenesis, whereby oncogenic SOX2+ stem cells give rise to senescent clusters that induce tumour formation in a cell non-autonomous manner [167][168].
More recently, a mechanism for this paracrine model has been pos-

Low-grade gliomas
Low-grade gliomas (LGGs) are a diverse group of benign brain tumours (WHO grade I and II). Symptoms are variable and largely at- Senescence has been more thoroughly investigated in paediatric than adult LGG. Pilocytic astrocytoma (PA; WHO grade I) is the most prevalent paediatric LGG and the most frequent paediatric brain tumour in children. Constitutive activation of MAPK pathway, by genetic mutations, is detectable in nearly all cases [172], which leads to oncogene-induced senescence, as shown by β-galactosidase activity and induction of p16 INK4a expression in up to 90% of primary PA samples [173]. SASP factors (e.g. IL1B and IL6) are found to be upregulated in primary human tumours as well as in a PA mouse model [174]. SASP expression in PA tumours is associated with favourable prognosis whereas anti-inflammatory treatment with dexamethasone inhibits the SASP and induces regrowth of senescent cells.
These results highlight the importance of paracrine propagation and maintenance of senescence in paediatric LGG. Of relevance, senescent PA cells can be ablated using senolytics (i.e. ABT-263 and ABT-737), paving the way to a new type of treatment for these patients.
Homozygous deletion of CDKN2A (encoding p16 INK4a ) can be observed with low frequency in paediatric LGG [175], but is more common in higher-grade tumours, such as pleomorphic xanthoastrocytoma and anaplastic astrocytoma with piloid features, suggesting that it probably acts as a second oncogenic hit, promoting senescence escape and facilitating transformation into high-grade glioma [176][177]. Secondary alterations involving homozygous or hemizygous losses of CDKN2A and TP53 are more characteristic in adult LGG. Increased survival has been associated with absence of mutations in CDKN2A and TP53, suggesting that senescence escape may promote tumour progression [178][179].
Together, this research area has highlighted the presence of a large number of senescent cells in LGG. These cells through the SASP seem to play a critical role in tumour control, preventing the progression of the tumour to a more aggressive cancer. However, senescent cells in LGG are susceptible to a second oncogenic hit, promoting senescence escape and tumour progression.

Glioblastoma multiforme
Glioblastoma multiforme (GBM; WHO grade IV) is one of the most common and aggressive primary brain tumours accounting for 60% of brain tumours in adults. They are highly infiltrative and have an average survival of less than 25% after two years due to the high recurrence rate [180]. Evidence of therapy-induced senescence in GBM has been shown following TMZ treatment and radiotherapy [182][183].
Culture of GBM cell lines in the presence of TMZ induces senescence through a DNA damage response pathway and expression of p21 Cip1 . Subsequently, the NF-κB pathway is activated, accompanied by the production of the SASP components IL6 and IL8 [183][184].
Confirming the in vitro data, orthotopic transplantation of GBM cell lines into immunodeficient mice followed by oral administration of TMZ, leads to a senescence response evidenced by p21 Cip1 expression and NF-κB pathway activation in the tumour.
It is thought that radiotherapy in GBM leads to increased recurrence rates due to the induction of a tumour-promoting microenvironment [182,[185][186][187]. The DNA damage caused by irradiation results in the induction of senescence and SASP in both tumour cells and/or non-tumour cells in the microenvironment, which as previously discussed can be pro-tumourigenic and lead to recurrence [7,15,52]. It  [188].
A recent study has demonstrated that GBM cell lines can be driven into senescence, by either TMZ treatment or irradiation, to subsequently be selectively ablated with Navitoclax (ABT-263) as a senolytic [189]. Since the induction of senescence and SASP, caused into eight different types) [190][191]. These groups differ not only in their gene expression but also their methylation patterns, histology, clinical characteristics, metastatic potential, incidence and rate of recurrence. Despite the extensive clinical treatment stratification, outcomes of therapy can still be poor due to recurrence [192]. promoter methylation, supporting a senescence evasion mechanism.
Additional evidence of senescence has been proposed from in vitro studies using the cell lines DAOY and ONS-76 [193]. Knockdown of citron kinase protein (CITK), which is required for normal prolifer-  [194]. Expression of senescence markers, such as p16 INK4a , is very low in this tumour type, and this is probably due to the oncogenic driver mutations' ability to represses the CDKN2A locus [195]. In contrast to the tumour cells, p16 INK4a +ve cells are often found in the tumour microenvironment (up to 80% of tumours [196], suggesting that these potentially senescent cells could have a role in tumourigenesis and/or treatment resistance.
Conventional clinical management by radiotherapy or new targeted therapies could be used to trigger TIS in DMG tumours, as suggested by in vitro studies on patient-derived DMG cell lines. The combination of radiation and the mTOR inhibitor AZD2014 has been shown to result in a strong synergistic antitumour activity preclinically [197], suggesting that the use of senolytics or SASP modulators could be of therapeutic relevance. Likewise, a recent study has proposed a new model where senescence is induced in DMG tumour cells by inhibition of BMI1. In vivo, the clearance of these treatment-induced senescent cells with ABT-263 attenuates tumour growth and prolongs animal survival [198].

CON CLUDING REMARK S AND PER S PEC TIVE S
There is sufficient evidence to support the idea that senescent cells play a critical role in the pathogenesis of neurodegenerative conditions and brain tumours. The ablation experiments using genetic and chemical approaches have fuelled the interest in anti-senescence therapies as potential treatments against these pathologies.
However, several questions still remain that should be addressed to support further the development of senotherapies. in premature death due to hepatic dysfunction [199]. This study highlights that the balance between the beneficial and detrimental functions of senescence must be thoroughly understood.
The ablation experiments in neurodegeneration mouse models suggest that senescent cells are not just bystanders, but they contribute to disease progression and cognitive loss. It will be interesting to assess whether such a role is preserved in humans. Another cerebral disease highly associated with old age is ischaemia/stroke and aneurysms [200].
During the acute phase of ischaemia in humans, pathogenic processes such as neuroinflammation (cytokines and chemokines) and oxidative stress have been shown to be upregulated. Furthermore, aged murine models have demonstrated a higher inflammatory response during the acute phase of ischaemia, which results in increased cerebral injury compared to young animals [200]. It will be important to define whether senotherapies are only able to prevent disease progression, or in addition, senolytics can improve cognitive decline and restore brain function in patients with advanced disease. These questions can be addressed in human trials, as those already running to test the efficacy of senotherapies against other human conditions.
Senescence is postulated to be a cell autonomous barrier against cancer that maintains potential cancer-initiating cells in a benign, One of the main problems with current anti-cancer therapies is tumour recurrence. It is thought that senescent cells within the tumour bed are therapy-resistant and will eventually re-enter the cell cycle and give rise to a relapsed tumour. It has been shown that passing through a senescent state, even if transiently, can bestow features of stemness upon tumour cells making them more aggressive and malignant [12]. Therefore, there is a strong rationale to use senotherapies as adjuvant treatments to eliminate senescent cells prior to tumour recurrence. This is a promising approach, whereby current effective senescence-inducing treatments, such as cytostatic chemotherapy, radiotherapy or specific targeted therapies, could be combined with senolytics in order to ablate the senescent cells prior to senescence escape and progression to recurrence. Preclinical research using suitable models of brain tumours will facilitate the development of clinical trials to test these combination therapies. We wonder whether Hayflick thought that his initial observations would ever become the catalyst that fuelled a vast research field, which potentially could improve clinical outcomes for the most relevant human diseases or even prolong a healthy life span. Despite current limitations and unknowns, it is difficult to not be affected by an encouraging optimism towards the translational potential of anti-senescence therapies against brain pathologies and cancer. Future research will reveal key mechanistic insights into how senescent cells contribute to human disease paving the path to novel anti-senescence treatments.

E TH I C S S TATEM ENT
Ethics approval was not required since this paper does not concern animal experimentation or the use of human volunteers.

ACK N OWLED G EM ENTS
The authors wish to thank Prof. T. Jacques and Prof. Rick Livesey for their comments and suggestions. We also thank Dr J. Grey for his feedback.