Immunological consequences of senescence in physiology and pathology Authors: Anastasija Zubova, Federico Pietrocola, Samir Morsli Published: 2026 ABSTRACT BackgroundCellular senescence is a sublethal stress response characterized by a durable cell-cycle arrest and the acquisition of a complex secretory program known as the senescence-associated secretory phenotype (SASP), which can profoundly influence local and systemic immunity. In physiological contexts—including embryonic development, tissue repair, and acute tumour suppression—senescent cells coordinate the recruitment and activation of immune cells, enabling their timely immune-mediated clearance and facilitating tissue remodelling and restoration of homeostasis. However, during aging and chronic disease, immune surveillance mechanisms frequently become compromised, allowing senescent cells to accumulate and persist within tissues.Main bodyThe persistence of senescent cells results in sustained SASP signalling that promotes chronic inflammation, immune dysfunction, and tissue remodelling processes linked to fibrosis, metabolic impairment, tumour progression, and defective tissue repair. In parallel, increasing evidence indicates that immune cells themselves can acquire senescent or senescence-like states, thereby weakening immunosurveillance and generating self-reinforcing feedback loops that further amplify senescent cell accumulation and tissue dysfunction. In this review, we synthesize recent advances in understanding the bidirectional interactions between senescent cells and the immune system across physiological and pathological contexts, with particular emphasis on the mechanisms that govern immune recognition, clearance, and immune evasion.ConclusionsA deeper understanding of the reciprocal interplay between senescent cells and the immune system provides an important conceptual framework for the development of new therapeutic strategies. Interventions aimed at modulating senescence–immune crosstalk — including immune-mediated senolytic approaches, checkpoint modulation, and immune rejuvenation — may offer promising opportunities to restore immune surveillance, limit the detrimental consequences of persistent senescence, and ultimately improve outcomes in age-related diseases. Introduction Cellular senescence has long been viewed as a largely cell-autonomous stress response defined by stable cell-cycle arrest accompanied by extensive metabolic and epigenetic remodelling [1–3]. Increasingly, however, senescence is being recognised as a process embedded within an immunological environment, where the fate of senescent cells and the functional outcomes of senescence responses depend on how senescent cells are sensed and handled by host immunity. Immune responses to senescent cells arise from the balance between signals that stimulate activation and those that impose tolerance, creating a spectrum of outcomes ranging from rapid clearance to long-term persistence within tissues [4]. Accumulating evidence indicates that senescence encompasses a continuum of functional states that emerge across physiological and pathological contexts. In this landscape, the equilibrium between immune stimulatory and immune suppressive cues determines whether senescent cells are eliminated or persist in a tissue. Through a context-dependent senescence-associated secretory phenotype (SASP) together with remodelling of immunomodulatory surface ligands, senescent cells shape the signals that guide or suppress immune recognition [5–8]. The composition, intensity and duration of these signals vary with cell type, stressor and temporal context, distinguishing transient physiological programmes from chronic pathological senescence [6]. The relationship between senescent cells and the immune system is reciprocal. Immune surveillance governs whether the senescence response is resolved or persists, yet immune cells themselves can adopt senescence-associated features that remodel tissue environments and propagate senescence systemically. Age-related decline or dysfunction within immune compartments can amplify inflammatory signalling, shift immune tolerance and generate niches that favour senescent-cell persistence, establishing feedback loops between immune aging and cellular senescence [9, 10]. Together, these observations position senescence not as an isolated cell-intrinsic programme but as a process shaped by continuous dialogue with the immune system. The strength of this senescence–immune crosstalk is shifting the therapeutic paradigm from classical senolytics toward immuno-senolytic strategies aimed at reactivating endogenous immune surveillance or deploying engineered immune cells to selectively eliminate senescent populations. Determinants of immune recognition of senescent cells by host immune cells Immune recognition of senescent cells emerges from coordinated remodelling of both the secretome and the cell surface, which together determine how these cells are sensed by innate and adaptive immune compartments. Rather than relying on a single pathway, senescent cells display a composite molecular signature shaped by cytokine gradients, danger signals and altered immunomodulatory ligands that regulate immune recruitment and cytotoxic engagement. Within this framework, the SASP acts as a central interface linking intrinsic stress signalling to immune-cell recruitment and activation, orchestrating the spatial dynamics of immunosurveillance [7]: early SASP programmes favour immune recognition and clearance, whereas chronic or dysregulated SASP signalling shifts tissues toward immunosuppressive inflammation that enables senescent-cell persistence (Fig. 1). Reactivation of antiviral sensing pathways represents a prominent feature of senescence-associated immune engagement. The cyclic GMP-AMP synthase - stimulator of interferon genes (cGAS–STING) cytosolic nucleotide-sensing pathway, together with type I interferon (IFN-I) signalling, drives inflammatory programmes that promote immune interaction [11, 12]. Cytoplasmic chromatin fragments (CCFs) are observed across multiple mouse and human senescence models, including oncogene-induced senescence (OIS) and DNA-damage-induced senescence (DIS) [13, 14] and STING-deficient mice show impaired immunosurveillance following hydrodynamic tail-vein delivery of oncogenic Neuroblastoma RAS (NRas), an OIS model, together with reduced inflammatory responses after ionising radiation [13]. In OIS- and DIS- fibroblasts, topoisomerase-1 cleavage complexes (TOP1cc) promote the formation of cytoplasmic chromatin that activates cGAS–STING signalling [15]. Pharmacological induction of TOP1cc using camptothecin enhances cGAS signalling and amplifies the SASP in cisplatin-induced senescent ovarian cancer cells, resulting in reduced tumour growth and improved survival. Release of mitochondrial DNA and RNA further reinforces this pathway [16–18], as shown by diminished SASP factor production and reduced immune recruitment in Bax/Bak-deficient mice models of metabolic dysfunction-associated steatohepatitis (MASH) [17] and liver irradiation [18]. Senescent cells also activate anti-viral sensing response through de-repression of endogenous retroelements. Activation of Long Interspersed Nuclear Element-1 (LINE-1, henceforth L1) retrotransposable elements induces IFN-I signalling in senescent cells, and pharmacological suppression of L1 activity in aged mice reduces tissue inflammation [19]. In Sirtuin 6 (SIRT6)-deficient and wild-type aged mice, cytoplasmic accumulation of L1 cDNA activates IFN-I responses via cGAS–STING [20], while endogenous retroviruses (ERVs) further amplify antiviral signalling, as shown in experimental settings of OIS where loss of epigenetic silencing leads to accumulation of double-stranded RNA and activation of mitochondrial antiviral pathways [21]. A subclass of senescence-associated ERVs (SA-ERVs) promotes Activating Transcription Factor 3 (ATF3)-dependent dsRNA generation in senescent fibroblasts, reinforcing IFN-I responses observed in aged tissues [22]. More recently, activation of the human endogenous retrovirus HERV-K (HML-2) has been shown to induce senescence and produce retrovirus-like particles (RVLPs) in primary human fibroblast and mesenchymal progenitor cells, which can propagate senescence phenotypes to neighbouring cells. HERV-K activation is observed in aged tissues in primates and humans, and its suppression alleviates cellular senescence and tissue degeneration, suggesting that retroelement reactivation may actively contribute to age-associated tissue decline [23]. Additional layers of SASP regulation contribute to immune activation. In murine models of OIS, inflammasome activation shapes the secretory programme by promoting cytokine maturation and reinforcing paracrine senescence [24]. Moreover, in OIS, the innate immune receptor Toll-like receptor 2 (TLR2) further amplifies inflammatory signalling through sensing of senescence-associated damage-associated molecular patterns (DAMPs) and promotes immune recruitment [25]. Senescence induction itself influences early secretory outputs. Following Kirsten Rat Sarcoma Viral Oncogene Homolog (KRAS) -driven OIS, the cyclin-dependent kinase inhibitor 1 (henceforth referred to as p21) initiates an immediate secretome (PASP) through retinoblastoma (Rb) protein-dependent transcription [26]. PASP factors including CXC motif chemokine ligand 14 (CXCL14) drive macrophage recruitment, enabling immune surveillance of p21⁺ cells in vivo. In the mouse liver, macrophages transiently monitor oncogene-induced p21⁺ cells and disengage when p21 levels normalise, whereas sustained p21 expression promotes M1 polarisation and recruitment of cytotoxic T lymphocytes before classical senescence markers appear [26]. Beyond secreted cues, senescent cells remodel their surfaceome to enhance immunogenicity. Upregulation of natural killer group 2D (NKG2D) ligands such as MHC class I-related chain A (MICA) and UL16 Binding Protein 2 (ULBP2) promotes natural killer (NK) cell–mediated cytotoxicity toward DIS, OIS and replicative-induced senescent (RIS) fibroblasts [27]. NK cell killing of senescent cells through ER-stress–dependent exposure of ecto-calreticulin (e-CRT), via NKp46 engagement, was also shown in therapy-induced senescent (TIS) human and murine lung cancer models [28]. Emerging evidence indicates that this surface remodelling also expands recognition beyond conventional NK responses. Lipid-activated invariant natural killer T (iNKT) cells preferentially eliminate pathological senescent cells in vivo in a CD1d-dependent manner. Activation with α-galactosylceramide reduces senescent preadipocytes in obese mice and limits bleomycin-induced lung fibrosis while improving metabolic and survival outcomes [29]. In parallel, human and mouse Vγ9Vδ2 γδ T cells selectively recognise TIS human and mouse fibroblasts in vitro through stress-induced ligands such as Butyrophilin subfamily 3 member A1 (BTN3A1) [30] and cooperate with NKG2D-dependent signalling to mediate cytotoxicity, reducing senescent burden and improving disease outcomes in mouse models of lung fibrosis [31]. Importantly, senescent cells retain the ability to stimulate cluster of differentiation 8 (CD8)-dependent cytotoxic activity. This is achieved via type I interferon–dependent upregulation of MHC I complexes, presentation of unique antigenic epitopes, and promotion of dendritic-cell activation through release of DAMPs, including Adenosine triphosphate (ATP) and CRT [32]. Accordingly, vaccination with live TIS cancer cells elicits strong prophylactic and therapeutic antitumour immunity [32, 33]. In liver cancer models, TIS increases IFN-γ receptor expression and antigen-presentation machinery, and disruption of IFN-γ sensing impairs immune-mediated clearance without directly affecting the senescence programme [34]. In aged human skin, senescent fibroblasts can be selectively eliminated by cluster of differentiation 4 (CD4⁺) cytotoxic T lymphocytes (CTLs) through a virus–immune axis linking senescence to antiviral surveillance [35]. Although senescent fibroblasts are more abundant in old compared with young skin, their numbers do not continue to increase in the elderly, suggesting the presence of active immune control mechanisms. This process is associated with increased Chemokine C-X-C motif ligand 9 (CXCL9) expression and recruitment of cytotoxic CD4⁺ T cells, which inversely correlates with senescent fibroblast abundance. Mechanistically, senescent fibroblasts upregulate HLA class II molecules and reactivate human cytomegalovirus glycoprotein B (HCMV-gB), thereby becoming direct targets for skin-resident CD4⁺ cytotoxic T lymphocytes in an HLA-II–dependent manner [35]. HCMV-gB expressed by senescent cells activates local CD4 CTLs, mirroring antiviral immune responses observed in HCMV-infected fibroblasts [36], and suggesting that senescence can expose latent viral antigens that license immune recognition and immune-mediated clearance. CD4⁺ T cells differentiate into Eomesodermin⁺CCL5⁺ cytotoxic lymphocytes (CD4-Eomes) when exposed to senescent-cell–rich environments in liver cirrhosis mouse models and with age [37]. Reduction of senescent-cell burden using senolytics limits CD4 T cell differentiation, whereas selective deletion of CD4-Eomes cells in aged animals increases senescent-cell burden and accelerates functional decline. In line with this, ablation of CD4-Eomes cells exacerbates fibrosis and worsens pathological outcomes in murine models of liver cirrhosis, mirroring the increased burden of senescent cells in the hepatic parenchyma [37]. Overall, the mechanisms presented in this section depict senescent cells as immunologically active nodes in which cell intrinsic damage sensing, SASP regulation, and immune cell engagement intersect to shape cellular fate. By influencing immune recognition and clearance efficiency, these pathways determine whether senescence supports tissue homeostasis or contributes to chronic disease, highlighting opportunities for therapeutic intervention. Noteworthy, recent work suggests that senescence-associated immunogenicity is highly context dependent [38]: distinct cell types and senescence inducers generate divergent immunopeptidomes and immune responses, with only specific subsets effectively activating T cells and NK-cell cytotoxicity, underscoring the need for cell-type- and context-specific immunotherapeutic strategies [4]. Consequences of senescent cell immune surveillance in physiology In physiological settings, the beneficial effects of senescence are tightly linked to effective immune-mediated clearance. Rather than representing a static endpoint, senescence often functions as a transient programme whose resolution depends on timely immunosurveillance. Successful elimination of senescent cells limits prolonged SASP exposure — thereby preventing immunosuppressive inflammation — restores tissue homeostasis and guides processes such as tissue remodelling, regeneration and tumour suppression [39]. Developmental and morphogenetic senescence During embryogenesis, senescence operates as a programmed physiological process that guides morphogenesis rather than arising as a response to cellular stress [40, 41]. In a p53-independent manner, p21⁺ cells sculpt developing tissues by facilitating the removal of transient structures and maintaining the balance of progenitor populations. Loss of p21 disrupts this process, leading to morphogenetic abnormalities that cannot be compensated by apoptosis alone. A defining immunological feature of developmental senescence is its coordination with macrophage-mediated clearance. Senescent embryonic cells recruit and become surrounded by F4/80⁺ macrophages prior to the onset of apoptosis [40, 41]. Whether these cells undergo apoptosis before clearance or are directly phagocytosed remains unresolved. Nonetheless, both studies indicate that senescent cells, together with macrophage recruitment, actively contribute to tissue patterning during embryonic development. In keeping with these findings, senescent-like programmes have also been described during development in several non-mammalian organisms [42–44]. During pregnancy, the placenta acts as an immunologically dynamic interface where senescent cells coexist as tolerogenic entities within tightly regulated maternal–foetal immune interactions [45, 46]. Emerging evidence indicates that cellular senescence is not merely a pathological feature of placental aging but also a physiological component of human development, possibly contributing to the protection of the foetus from maternal immune-mediated damage [47]. In line with this speculation, the placenta expresses several immunosuppressive molecules found up-regulated in senescent cells, including Programmed Death-Ligand 1 (PD-L1), Programmed Death-Ligand 2 (PD-L2), Fas-ligand (Fas-L) and Galectin 9 (LGALS9) [48, 49]. During normal placentation, extravillous trophoblasts undergo genome amplification and transition into a senescent-like state upon invasion of the decidua, characterised by metabolic rewiring, acquisition of a SASP and loss of proliferative potential [50]. This programme appears to function as a ploidy-limiting mechanism that stabilises differentiation and constrains excessive trophoblast expansion. Disruption of this balance, as observed in hyperplastic pregnancy disorders such as androgenic complete hydatidiform moles, is associated with reduced senescence and aberrant polyploidisation, suggesting that controlled senescence contributes to placental homeostasis. Consistent with this concept, pathological pregnancies are increasingly linked to dysregulated placental senescence. In mouse models of the developmental disorder Cornelia de Lange syndrome, mutations in cohesin-associated genes — which normally preserve genome integrity and coordinate chromosome segregation, transcriptional control and DNA repair [51]— lead to excessive DNA damage and persistent placental senescence accompanied by embryonic developmental defects [52]. These findings suggest that intact cohesin function is required to maintain placental homeostasis and prevent maladaptive senescence. Notably, transplantation of cohesin-mutant embryos into wild-type placentas, or pharmacological inhibition of inflammatory cytokine signalling, partially rescued developmental abnormalities, indicating that excessive senescence-associated inflammation within the placental niche actively contributes to embryonic pathology. Complementary human studies further show that exaggerated trophoblast senescence and SASP signalling precede clinical manifestations of disorders such as preeclampsia and peripartum cardiomyopathy, where elevated circulating SASP factors — including activin A and Plasminogen Activator Inhibitor 1 (PAI-1) — correlate with maternal cardiovascular dysfunction and may serve as early predictive biomarkers [53–57]. Regeneration & wound healing Mounting evidence suggests that senescence represents an evolutionarily conserved programme that supports regenerative repair by coordinating inflammatory and immune responses with tissue remodelling. In non-mammalian vertebrates such as zebrafish and salamanders, senescence is rapidly induced following injury and is required for effective regeneration [58–61]. Similarly, in murine models of hepatic injury, activated stellate cells undergo senescence to halt fibrogenesis, while their SASP recruits immune effectors — particularly NK cells — that recognise and eliminate these cells once repair is initiated [62]. In a mouse model of oxygen-induced retinopathy, senescent vascular endothelial cells generate sterile inflammation and secrete factors that recruit neutrophils and promote neutrophil extracellular trap (NET) formation. NET-mediated clearance of senescent vasculature drives vascular pruning and regenerative remodelling, whereas disruption of NETosis impairs senescent cell removal and compromises tissue recovery [63]. A similar principle operates during cyclical regeneration of the endometrium. Following ovulation, activation of forkhead box O1 (FOXO1) drives differentiation of endometrial stromal cells (EnSCs) into decidual cells that regulate implantation. As part of this process, FOXO1 signalling induces a subset of decidualising cells to enter a transient senescent state, reinforced by interleukin-8 (IL-8), as demonstrated using human EnSCs in vitro. These senescent cells generate a localized inflammatory niche that promotes tissue remodelling and endometrial receptivity. Using in vitro co-culture assays of human endometrial biopsies across the menstrual cycle, uterine natural killer (uNK) cells are activated by interleukin-15 (IL-15) during the luteal phase, and subsequently eliminate senescent decidual cells, thereby restoring tissue homeostasis [64].When clearance fails, senescent cells persist, acquire progesterone resistance and secrete extracellular-matrix remodelling factors that disrupt tissue architecture, a feature observed in recurrent pregnancy loss [65]. Endometrial assembloid models further show coordinated emergence of differentiated and senescent states across epithelial and stromal compartments, where acute senescence supports implantation signalling but persistent senescence compromises tissue integrity [66]. Genetic and injury-induced reprogramming studies further suggest that senescence-associated inflammatory signalling contributes to tissue plasticity in vivo. Transient expression of the pluripotency-inducing factors octamer-binding transcription factor 4 (Oct4), SRY-box transcription factor 2 (Sox2), Krüppel-like factor 4 (Klf4) and c-Myc (OSKM) in mice promotes a senescence response that enhances cellular reprogramming via IL-6–dependent signalling, whereas genetic suppression of interleukin-6 (IL-6) markedly impairs this process [67]. In parallel, senescent cells promote dedifferentiation and lineage conversion in neighbouring cells via paracrine signalling in mouse models of tissue injury, as in the case of lung injury induced by bleomycin [67]. In skeletal muscle, injury-associated accumulation of senescent cells enables in vivo reprogramming via IL6-dependent Paired Box 7 (Pax7)+ stem cell activation, while IL6 blockade or senescent cell removal by navitoclax treatment reduces reprogramming efficiency [68]. Senescence also represents a transient component of the acute cutaneous wound-healing response. Shortly after injury, cyclin-dependent kinase inhibitor 2 A (hereafter referred to as p16) expressing senescent fibroblasts and endothelial cells accumulate at the wound edge and release SASP factors, including platelet-derived growth factor-AA (PDGF-AA), that promote myofibroblast differentiation and granulation tissue formation. Consistently, genetic ablation of p16+ cells delays wound closure and hinders myofibroblast differentiation [69]. PDGF-AA additionally supports recruitment of innate immune cells that coordinate the transition from inflammation to proliferation during wound closure [70]. Recent work highlights functional heterogeneity among senescent populations: in murine dorsal wounds, p21⁺ cells dominate the inflammatory phase and amplify Nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB)-dependent signalling, whereas p16+ cells are more restricted to stromal compartments. Both targeted inhibition of NF-κB signalling in p21⁺ cells and p21+ cell ablation accelerate wound resolution and limit fibrosis, underscoring how distinct senescence programmes shape repair dynamics [71]. Together, these studies reveal that senescence can function as a transient, immune-guided programme that supports tissue plasticity and repair when its induction and clearance are finely coordinated. Mechanisms of immune evasion by senescent cells Although senescent cells can be eliminated by host immune cells, their accumulation during aging and disease indicates that clearance is frequently hampered. Senescent cells actively reshape their environment, engaging intrinsic programmes and local immune circuits that blunt cytotoxic responses while sustaining SASP signalling and chronic inflammation linked to fibrosis, tumour progression and tissue dysfunction (Fig. 1). Understanding how senescent cells evade immune regulation mechanistically will be central to developing strategies that target pathological senescence without disrupting its physiological functions [4]. Cell-intrinsic mechanisms: surface ligands suppressing cytotoxic immunity A first layer of evasion occurs at the surface of senescent cells themselves, where upregulation of inhibitory ligands directly suppresses NK- and T-cell-mediated killing. The best-characterised example involves engagement of inhibitory checkpoint pathways. Senescent cells heterogeneously express the immune checkpoint ligand PD-L1, and PD-L1⁺ subsets preferentially accumulate with age in vivo, suggesting that checkpoint engagement contributes to senescent cell persistence [72, 73]. Single-cell analyses indicate that PD-L1 expression associates with heightened SASP activity, and PD-L1⁺ senescent cells display increased resistance to T-cell–mediated clearance compared with PD-L1⁻ counterparts. Consistently, PD-1 blockade in naturally aged mice, as well as in models of liver aging and non-alcoholic steatohepatitis, reduces the burden of p16⁺ cells through activation of CD8⁺ T cells, ameliorating multiple aging-associated phenotypes. These findings support a model in which checkpoint signalling selectively protects highly inflammatory senescent subsets from immune surveillance, thereby promoting their accumulation during aging [72]. Notably, in murine models of lung inflammation, where p16 further stabilises PD-L1 on senescent macrophages and epithelial cells, PD-L1 blockade selectively depleted p16⁺ cells and improved tissue inflammation; however, PD-1 inhibition failed to reproduce this effect, suggesting that antibody-dependent cytotoxicity rather than restoration of proficient immunosurveillance may mediate clearance in this context [73]. Recently, a blinded cross-laboratory replication demonstrated that PD-1 blockade did not reduce p16⁺ cells, nor improved aging-related parameters in mice [74]. The authors of the original study highlighted the limitation of relying on bulk p16⁺ expression that may not accurately quantify exact senescence burden [75]. Together, these emphasise that caution is required when interpreting these results — which should be further evaluated prior to therapeutic translation. Senescent cells can also evade immune surveillance through expression on their surface of the non-classical MHC molecule HLA-E. TIS, OIS and RIS dermal fibroblasts upregulate HLA-E in response to SASP-associated inflammatory cues via p38 MAPK signalling, engaging NKG2A on NK and highly differentiated CD8⁺ T cells to suppress cytotoxicity [76]. Elevated HLA-E expression in aged human skin supports a role for this pathway in maintaining senescent cell persistence in vivo, while disruption of the HLA-E–NKG2A interaction restores immune-mediated killing and highlights a potential therapeutic avenue [76]. The co-expression of these two inhibitory programmes — PD-L1 and HLA-E — is well illustrated in pathological contexts such as multiple sclerosis lesions and experimental demyelination, where senescent oligodendrocyte precursor cells and microglia exploit both ligands alongside SASP-driven impairment of remyelination; senolytic treatment or combined blockade of PD-1 and NKG2A reduces senescent glia burden and improves functional recovery [77]. A third surface-level mechanism operates through glycolipid signalling rather than classical checkpoint pathways. RIS human lung fibroblasts and DIS mouse embryonic fibroblasts increase expression of the glycosyltransferase ST8SIA1, leading to accumulation of the disialoganglioside GD3 at the cell surface, where it engages Siglec-7 on NK cells and ultimately inhibits NK-cell degranulation [78]. GD3 expression increases across multiple aging tissues and associates with reduced senescent cell immunosurveillance; targeting GD3 restores NK-cell activity, attenuates lung and liver fibrosis, improves age-related bone remodelling and extends lifespan in mouse models [78]. GD3 has also been identified as a marker of senescence in osteoarthritis using a human and murine in vitro 3D organotypic cartilage model [79], reinforcing the tenet that these surface-level checkpoint mechanisms are not exclusively restricted to aging tissues but are also present in chronic inflammatory disease. These intrinsic surface-ligand programmes are not confined to aging and inflammatory disease contexts but are equally operative in tumours, where therapy-induced senescence engages additional and overlapping checkpoint mechanisms that further shield senescent cancer cells from immune elimination. Chemotherapy-induced senescence drives strong upregulation of PD-L2 in multiple murine and human cancer cell lines — dispensable for senescence induction but required for immune escape. In an orthotopic pancreatic tumour model, loss or blockade of PD-L2 accelerates clearance of TIS tumour cells, reduces SASP-associated chemokines such as CXCL1 and CXCL2, limits MDSC recruitment and enables CD8⁺ T-cell-mediated tumour regression; combining chemotherapy with anti-PD-L2 antibodies promotes durable tumour remission in a murine mammary cancer model [80]. Transcriptomic analyses of human lung adenocarcinoma have similarly identified immunomodulatory subsets of senescent tumour cells characterised by NF-κB- and STAT3-driven cytokine programmes — including CXCL1, CXCL2 and CXCL8 — and enriched checkpoint interactions such as LGALS9–TIM3, which dampen cytotoxic T-cell activity and associate with poor responses to immunotherapy [81]. In mouse breast cancer models and patient samples, chemotherapy-induced senescent tumour cells were also found to activate distinct immune-evasion programmes characterised by interferon-driven PD-L1 expression and p53-dependent CD80 signalling; targeting these pathways in mice enhances T-cell infiltration and therapeutic response, though tumours frequently persist, highlighting the redundancy of immune-inhibitory circuits that protect senescent tumour cells during residual disease [82]. Beyond checkpoint ligands, TIS cancer cells activate complement protein C3 alongside complement regulatory proteins (CRPs) in vitro [83]. In murine triple-negative breast cancer (TNBC) models, complement factor expression is further upregulated by tumour fibroblasts following chemotherapy, skewing the anticancer immune response towards immunosuppression [84]. Viewed together, these findings in senescent tumour cells reveal that the evasion mechanisms present in senescent stromal compartments are mirrored and reinforced in the tumour cell compartment itself, creating overlapping and partially redundant barriers to immune clearance. Cell-intrinsic mechanisms: chromatin regulation limiting innate immune sensing Beyond surface ligand expression, senescent cells can limit their own immunogenicity at the level of chromatin regulation. SMARCA4, a SWI/SNF chromatin remodeller upregulated during senescence, restrains the immunogenicity of TIS- and OIS-cancer cells by limiting activation of innate sensing pathways [85]. Genetic or pharmacological inhibition of SMARCA4 derepresses repetitive elements and amplifies cGAS–STING and MAVS–MDA5 signalling, enhancing SASP output and promoting NK-cell–mediated clearance. Consistent with this enhanced immunogenicity, in immunocompetent murine ovarian cancer models, a PROTAC targeting SMARCA4 synergized with cisplatin to increase infiltration of activated NK cells and CD8⁺ T cells, resulting in more efficient elimination of therapy-induced senescent cells [85]. Together with the surface-ligand mechanisms described above, SMARCA4-mediated dampening of innate sensing reveals how senescent cells exploit complementary intrinsic programmes to reduce their visibility to the immune system. Cell-extrinsic mechanisms: microenvironmental remodelling in tumours In tumours, the evasion strategies described above are compounded by extrinsic, paracrine mechanisms through which senescent stromal cells remodel the immune microenvironment. While the intrinsic mechanisms outlined above operate broadly across aging and disease contexts, the tumour microenvironment represents a setting of particular complexity, where senescent stromal and cancer cells interact with immune populations in ways that are further shaped by therapeutic intervention. TIS-cancer-associated fibroblasts (CAFs) illustrate this principle: p16⁺ myofibroblast CAFs deposit extracellular matrix that constrains NK-cell cytotoxicity, whereas their ablation restores tumour control [86]. Senescent stromal cells also recruit immature myeloid populations through IL-6–dependent chemokine signalling, establishing inflammatory niches that resist clearance and promote tumour growth, reversible by IL-6 blockade [87]. In addition, Fas ligand expression on TIS- and OIS-fibroblasts induces apoptosis of infiltrating T and NK cells, reinforcing immune escape and tumour progression in an iPSC-derived humanised mouse model [88]. SASP factors secreted into the tumour microenvironment can further consolidate immunosuppression. DNA-damage–induced stromal senescence promotes secretion of amphiregulin (AREG), which induces PD-L1 expression in neighbouring tumour cells and dampens cytotoxic lymphocyte activity, while targeting AREG restores immune competence and reduces chemoresistance in preclinical models [89]. Reciprocally, tumour-derived signals can themselves induce stromal senescence: in lung adenocarcinoma, exosomal 3′tiRNA-AlaCGC triggers a SASP programme in fibroblasts through FOXO3 inhibition and TGF-β/Smad3 activation, promoting LGALS9 secretion and CD8⁺ T-cell dysfunction, thereby contributing to resistance to anti-PD-L1 therapy [90]. These reciprocal interactions — senescent stroma shaping tumour immunity and tumour-derived signals inducing stromal senescence — highlight how the immunosuppressive niche is actively maintained and self-reinforcing. Myeloid-derived suppressor cells (MDSCs) further shape senescence outcomes within this microenvironment. In murine PTEN-null prostate tumours, CD11b⁺Gr-1⁺ myeloid cells accumulate alongside senescence activation and antagonize the immunostimulatory SASP through secretion of IL-1RA, thereby sustaining tumour proliferation; disruption of this circuit enhances senescence induction and improves responses to docetaxel therapy [91]. In parallel, TIS can skew the SASP towards an immunosuppressive profile in PTEN-null tumours through JAK2–STAT3 signalling, whereas pharmacological reprogramming of this pathway restores immunogenicity and enhances chemotherapy efficacy [92]. Collectively, these studies show that senescent cells employ multiple, context-specific immune-evasion strategies — spanning checkpoint signalling, chromatin regulation and microenvironmental remodelling — that allow their persistence and highlight immune pathways as therapeutically actionable points for clinical intervention. Dysfunctional adaptive immunity and senescence in disease Loss of immune fitness contributes to the active propagation of senescence systemically. Metabolic collapse within T cells induced in mouse by deletion of the mitochondrial transcription factor Mitochondrial transcription factor A (TFAM) rewires T-cell metabolism toward glycolysis and helper Type I (Th1) polarization, generating a cytokine milieu dominated by IFN-γ and Tumor Necrosis Factor-alpha (TNF-α) that drives paracrine senescence across multiple tissues. Restoration of nicotinamide adenine dinucleotide (NAD⁺) levels or TNF-α blockade mitigates inflammation, preserves tissue function and delays decline, illustrating how immune metabolic stress can act upstream of systemic senescence [93]. Defective surveillance by cytotoxic lymphocytes further amplifies this phenomenon. In perforin-deficient mice, impaired NK- and CD8⁺-mediated killing leads to widespread accumulation of p16⁺ and SA-β-Gal⁺ senescent cells accompanied by fibrosis, chronic inflammation and reduced lifespan [9]. Pharmacological elimination of senescent cells using the B-cell lymphoma 2 (BCL-2) inhibitor navitoclax reduces tissue deterioration and extends survival, directly linking defective immunosurveillance to pathological senescence accumulation. Immune cells themselves are also susceptible to senescence-inducing stress, leading to progressive loss of genomic integrity, mitochondrial fitness and cytotoxic capacity. DNA-damage accumulation confined to hematopoietic compartments is sufficient to induce systemic aging. In Excision Repair Cross-Complementation Group 1 (Ercc1)-deficient mice, T cells, B cells, macrophages and NK cells acquire senescence features marked by p21/p16 induction and inflammatory SASP expression, preceding peripheral tissue dysfunction [10]. Transfer of aged or Ercc1-deficient splenocytes propagates organ damage in young recipients, whereas transplantation of young immune cells reverses these effects. Similarly, human CD8⁺ effector-memory T cells develop senescence signatures with age and disease, displaying impaired proliferation, cytokine production and reduced ability to eliminate senescent targets, thereby limiting immunotherapy efficacy [94]. Telomere attrition provides a further DNA-maintenance mechanism linking immune ageing to senescence. Repeated antigen-driven clonal expansion progressively shortens telomeres in lymphocytes, particularly late-differentiated CD8⁺CD28⁻/CD57⁺ T cells, where critically short or dysfunctional telomeres activate a persistent DNA-damage response and limit proliferative reserve [95]. This contributes to impaired immune renewal, reduced responses to infection and vaccination [96–98]. Beyond telomerase-mediated elongation, a subset of naïve and central memory T cells can acquire telomere vesicles directly from antigen-presenting cells at the immunological synapse — a telomerase-independent mechanism through which Rad51-mediated telomere fusion lengthens chromosome ends prior to clonal division, conferring durable protection from senescence [99]. Conversely, preserved telomere length and inducible telomerase activity in T cells associate with healthier immune ageing, including high-functioning centenarian phenotypes, supporting telomere maintenance as a determinant of immune fitness [100]. In a murine model of posttraumatic osteoarthritis, anterior cruciate ligament transection induces a type-17 immune response in the articular compartment and draining lymph nodes that parallels the accumulation of senescence markers p16 and p21. Innate lymphoid cells, γδ T cells, and CD4⁺ T cells all contribute to IL-17 production, and intra-articular neutralization of IL-17 reduces joint degeneration and senescence marker expression. Th17 cells promote senescence induction in fibroblasts, while senescent cells reciprocally skew naïve T cells toward Th17 or Th1 phenotypes depending on TGF-β availability, establishing a feed-forward inflammatory circuit that promotes tissue degeneration [101]. Beyond T-cell dysfunction, dysregulation of humoral immunity can also promote senescence. Spatial transcriptomic analyses identified immunoglobulin-rich inflammatory niches across multiple aged mice tissues characterised by complement activation and immune-cell infiltration [102]. Exposure to Immunoglobulin G (IgG) induces senescence markers in macrophages and microglia, and systemic administration of IgG accelerates systemic senescence, indicating that antibody accumulation can directly reshape innate immune aging [102]. In parallel, B cells further contribute to immune dysfunction during aging by actively impairing CD4⁺ T-cell function and promoting maladaptive B-cell–intrinsic insulin receptor signalling. Through MHC class II–dependent interactions, B cells promote loss of naïve CD4⁺ T cells, clonal restriction and differentiation toward dysfunctional senescent-like subsets. Consistently, mice lacking B cells maintain a more youthful CD4⁺ T-cell compartment, display improved immune fitness, and exhibit extended healthspan and lifespan, further implicating maladaptive humoral signalling as a driver of age-associated immune decline [103]. Dysfunctional innate immunity and senescence in disease Senescent macrophages represent a functionally distinct immune population that influences tissue function during aging through increased persistence, SASP factor release, and altered phagocytic and antigen-presenting functions that sustain inflammation [104]. Consistent with this concept, senescent macrophages have been shown to actively promote disease progression in multiple pathological contexts. In an oncogenic KRAS-driven lung mouse model, p16⁺ macrophages accumulate within premalignant lesions, suppress cytotoxic T-cell responses and promote tumour progression; senescent cell clearence delays tumour initiation and reduces tumour burden [105, 106]. Similarly, senescence-like tumour-associated macrophages (TAMs) accumulate in orthotopic mouse models of triple-negative breast cancer following exposure to Adriamycin [107]. These TAMs promote chemotherapy resistance via IL-6 signalling, and senolytic ablation sensitized the tumours to chemotherapy. Notably, CD163⁺ TAMs expressing senescence markers were also identified in human breast cancer tissues after chemotherapy treatment [107]. Lineage-tracing studies in mice identified p16⁺ macrophages as key contributors to fibrotic progression in the liver [108]. TREM2⁺p21⁺ macrophages accumulate with age and in mouse models of Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD), where their selective clearance by ABT-263 limits fatty liver disease progression [109]. Consistent with this, excess cholesterol is sufficient to induce mouse bone-marrow derived macrophage senescence in vitro, suggesting a mechanistic basis for senescent macrophage accumulation in the context of metabolic dysfunction [109]. Furthermore, macrophages derived from aged bone marrow appear central in sensitizing young animals to pulmonary fibrosis following aged bone-marrow transplantation [110]. In addition, senescent macrophages were shown to promote ferroptosis-associated muscle atrophy in an osteoarthritis mouse model [111] and contribute to white adipose tissue fibrosis in obesity via osteopontin-dependent signalling [112]. Environmental stressors also induce tissue-specific macrophage senescence. Heavy-metal exposure promotes FOXO3-dependent senescence of epididymal macrophages, leading to impaired sperm motility, whereas selective clearance of these cells restores reproductive function in mice [113]. Emerging evidence shows that senescent myeloid populations can mediate neurodegeneration. In Langerhans cell histiocytosis–associated neurodegeneration, circulating BRAF^V600E⁺ myeloid cells differentiate into senescent, inflammatory macrophages that infiltrate the brain following blood–brain barrier disruption, driving neuroinflammation and neuronal injury; pharmacological inhibition of MAPK signalling or senolytic treatment with navitoclax ameliorates disease in mice [114]. However, the impact of macrophage senescence is context dependent. In the bone marrow, genetic ablation of p16⁺ macrophages does not prevent bone density decline in aged mice, whereas targeting p16⁺ mesenchymal populations elicits a protective effect [115]. In settings of kidney transplantation, senescent p21+ macrophages dampened CD8⁺ T-cell activity through a Zinc Finger Protein 36 (Zfp36)– interleukin-27 (IL-27) signalling axis, thereby limiting allograft rejection [116]. Similarly, p16⁺ immune cells, including macrophages, display few classical senescence markers and instead serve as critical mediators of disease tolerance — a complementary host defence mechanism that limits tissue damage independently of pathogen clearance [117]. Here, p16+ immune subtypes are induced via TLR-7 activation or low-level STING signalling, after which they reduce adenosine accumulation, in part through nicotinamide N-methyltransferase (NNMT)-dependent regulation, and preserve tissue homeostasis. Accordingly, p16+ immune cells mediated protection against endotoxin shock, bacterial sepsis and ionising irradiation in mice, and their expansion — either via genetic deletion of Ifih1 or through an mRNA Coronavirus disease 2019 (COVID-19) vaccination — improves resilience to severe inflammation and delays age-related organ deterioration.Together, these findings indicate that the functional impact of macrophage senescence varies across tissues and disease settings and caution against indiscriminate senolytic approaches in favour of cell-type–specific targeting strategies. Senescence-like programmes also arise in neutrophils. In mouse models of prostate cancer, tumour-derived Apolipoprotein E (APOE) engages Triggering Receptor Expressed on Myeloid cells 2 (TREM2) on infiltrating neutrophils, driving a persistent senescence-like phenotype that suppresses anti-tumour immunity and correlates with poor prognosis in patients [118]. In breast cancer, senescent neutrophils release exosomes enriched in piRNA-17,560 that promote epithelial–mesenchymal transition and chemotherapy resistance in a mouse xenograft assay [119]. Collectively, these findings suggest that immune dysfunction and senescence-like immune states are not merely by-products of aging or pathology but active contributors to systemic senescence and disease progression. Immunotherapeutic strategies targeting senescent cells The growing body of preclinical evidence indicating that modulation of senescence–immune interactions can provide therapeutic benefit in age-related diseases paves the way for multiple clinical applications. First-generation senolytics, while instrumental in establishing that reducing senescence burden can yield clinical benefits, have shown limited therapeutic efficacy in patients [120, 121]. Further exploration of alternative approaches to reduce pathogenic senescence burden through modulation of the immune system could bolster established therapeutic regimens and limit treatment resistance. Immunotherapy-induced senescence surveillance Rather than focusing on individual pathways in isolation, several studies now converge on the concept of senescence-associated immune checkpoints - inhibitory signalling axes that restrain cytotoxic responses and allow senescent cells to persist. These include the abovementioned ligand–receptor interactions involving HLA-E–NKG2A, PD-1–PD-L1 and glycolipid-mediated signalling exemplified by the ST8SIA1–GD3–Siglec-7 axis [72, 76, 78]. As outlined above, disruption of these pathways enhances NK- and T-cell–mediated clearance of senescent cells and mitigates fibrosis, inflammation and age-associated dysfunction. Anti-NKG2A antibody (monalizumab) therapy has completed a phase III trial in head and neck cancer (NCT04590963), and multiple PD-1/PD-L1 inhibitors are approved clinically for a wide range of malignancies [122]. This raises the possibility that established immunotherapeutic frameworks used in oncology could be repurposed in the future — as standalone interventions or in combination — to target maladaptive senescence. In parallel, the antigenic landscape of senescent cells offers additional opportunities for immunotherapeutic intervention. Senescent cells present unique antigenic epitopes in association with MHC class I complexes, suggesting that immunogenic peptides could be harnessed to elicit protective immune responses through vaccination strategies. Immunisation with senescent cancer cells elicited robust anti-tumour immunity in preclinical studies [32], supporting the feasibility of antigen-guided approaches. Although direct vaccination with senescent cells may pose translational challenges, identification of defined seno-antigens could enable cell-free vaccination platforms. As a proof of concept, targeting the seno-antigen GPNMB enhanced senescence immunosurveillance and reduced atherosclerotic and aging-associated pathologies in mice [123]. Immunotherapeutic targeting of senescence surface markers Immunotherapeutic strategies targeting senescence-associated plasma membrane markers are rapidly emerging. Preclinical studies demonstrate that engineered chimeric antigen receptor T (CAR-T) cells can be directed against proteins selectively upregulated on senescent cells. CAR-T cells targeting the urokinase plasminogen activator receptor (uPAR), a marker broadly expressed across senescent populations and correlated with SA-β-gal activity, efficiently eliminated senescent cells in mouse models of lung adenocarcinoma and liver fibrosis, reducing tumour burden, fibrosis and inflammatory cytokine production [124]. In aged or high-fat-diet fed mice, a single administration of uPAR-CAR-T cells produced durable reductions in senescent cell burden accompanied by improvements in metabolic function and physical fitness [125]. An additional study also demonstrated effective clearance of intestinal senescent cells coupled with enhanced tissue regeneration following irradiation and attenuation of colitis in aged animals [126]. Despite promising pre-clinical results, no clinical trials have yet evaluated CAR-T therapy for senescence. In oncology however, CAR-T cells are clinically approved for the treatment of some in haematological malignancies [127], and clinical trials are underway for CAR-T therapy in solid tumours also (NCT05089266; NCT06101082; NCT04581473). Complementary approaches harness innate-like recognition pathways to achieve broader senolytic activity. NKG2D-CAR-T cells selectively eliminated senescent fibroblasts induced by diverse stressors and persisted in vivo, reducing senescence markers and SASP gene expression while improving muscle, bone and functional outcomes in accelerated aging mouse models. Similar benefits were observed following administration of autologous NKG2D-CAR-T cells to aged non-human primates, which lowered senescence signatures without overt toxicity [128]. In oncology, a phase I trial of NKG2D-CAR NK cells (NCT05213195) and the AERIAL trial (NCT06193902) provide early safety data, but further trials are required to assess clinical efficacy in patients when directly targeting senescence with this strategy. γδ T cells represent an additional clinically tractable strategy, that bypasses peptide–MHC restriction; human Vγ9Vδ2 T cells selectively targeted senescent stromal and endothelial populations, and adoptive transfer reduced senescent cell burden, inflammation and fibrosis in pre-clinical pulmonary injury models [31]. Allogenic γδ T cell therapy is currently in phase 1 clinical trials (NCT06364800; NCT06193486; NCT06069570) for oncological indications, but this approach for targeting senescence remains pre-clinical at this stage. The preclinical success of cell-based therapies targeting surface markers on senescent cells suggests that identifying markers unique to maladaptive senescent populations may enable more precise therapeutic strategies. Rather than relying on broad senolysis or nonspecific immune activation, defining cell type–specific senescence signatures could support targeted interventions that selectively eliminate pathogenic senescent cells while preserving beneficial or homeostatic variants. In parallel, approaches that preserve or restore immune fitness may offer complementary avenues to limit senescence burden. Strategies aimed at maintaining barrier integrity, preventing chronic inflammation, or rejuvenating haematopoietic stem cells to sustain a youthful immune repertoire could help reinforce endogenous immunosurveillance and restrain the accumulation of senescent cells. Together, these concepts point toward a future in which precision targeting of senescence is combined with immune rejuvenation to counteract age-associated pathology more effectively. Discussion It is becoming increasingly clear that senescence represents an umbrella term encompassing multiple functionally distinct states that coexist across tissues and time. Homeostatic or structural senescence supports normal tissue architecture and basal remodelling [129, 130], whereas adaptive senescence represents a transient stress-responsive programme that facilitates repair and immune-mediated resolution [69]. In contrast, maladaptive senescence reflects persistent or dysregulated states that sustain SASP signalling, promote immune dysfunction and chronic inflammation, and drive fibrosis, metabolic decline and tumour progression. Future therapeutic strategies must therefore distinguish beneficial from pathological senescence. Unselective clearance risks disrupting tissue integrity or regenerative capacity, whereas failure to remove maladaptive senescent populations promotes disease progression. Precision approaches — including intermittent senolysis, inducible or self-limiting CAR-T systems, or vaccines directed against pathology-restricted seno-antigens — may help preserve physiological senescence while enabling targeted elimination of harmful subsets. Systematic interrogation of existing single-cell and spatial datasets will be essential to determine whether candidate immunotherapy targets are expressed during wound healing, development or other physiological contexts. As a note of caution, translation of senescence-directed immunotherapies will require careful attention to safety and patient stratification. Clinical experience with CAR-T cell therapies highlights risks of systemic immune activation, including cytokine-release syndrome and immune effector cell–associated neurotoxicity, even when antigens appear disease-restricted [131]. Immune checkpoint blockade carries additional risks of immune-related adverse events such as colitis, pneumonitis, hepatitis and endocrinopathies, reflecting generalized immune disinhibition [132]. Together, these observations suggest that future senescence immunotherapies will need biomarker-driven patient stratification, controllable or localised delivery strategies, and built-in safety switches to balance senescent-cell clearance with preservation of immune homeostasis. Fig. 1Pro- and anti- immunosurveillance mechanisms of senescent cells. Senescent cells can either be recognized and eliminated by the immune system or evade immune clearance and persist in tissues. During acute/adaptive senescence (left panel), cytosolic nucleic acid sensing pathways including cGAS–STING, mitochondrial nucleic acids and endogenous retroelements activate type I interferon signalling and immunostimulatory SASP programs. These responses promote antigen presentation (MHC-I, MHC-II, CD1d), expression of stress ligands such as NKG2D ligands and BTN3A1/BTN2A1, and release of DAMPs including ATP and calreticulin (CRT), facilitating NK-cell and T-cell–mediated immunosurveillance. In contrast, chronic/maladaptive senescence (right panel) promotes immune evasion through expression of inhibitory ligands and immune checkpoints including HLA-E, PD-L1, PD-L2, FasL and GD3, together with activation of immunosuppressive signalling pathways such as JAK–STAT and TGF-β–SMAD. These changes reshape the SASP towards pro-inflammatory but immunosuppressive factors including IL-6, galectin-9, amphiregulin (AREG) and extracellular matrix (ECM) components, enabling senescent cells to evade immune clearance and accumulate during ageing and disease. Created in BioRender. Pietrocola, F. (2026) https://BioRender.com/ks4vpfk Pro- and anti- immunosurveillance mechanisms of senescent cells. Senescent cells can either be recognized and eliminated by the immune system or evade immune clearance and persist in tissues. During acute/adaptive senescence (left panel), cytosolic nucleic acid sensing pathways including cGAS–STING, mitochondrial nucleic acids and endogenous retroelements activate type I interferon signalling and immunostimulatory SASP programs. These responses promote antigen presentation (MHC-I, MHC-II, CD1d), expression of stress ligands such as NKG2D ligands and BTN3A1/BTN2A1, and release of DAMPs including ATP and calreticulin (CRT), facilitating NK-cell and T-cell–mediated immunosurveillance. In contrast, chronic/maladaptive senescence (right panel) promotes immune evasion through expression of inhibitory ligands and immune checkpoints including HLA-E, PD-L1, PD-L2, FasL and GD3, together with activation of immunosuppressive signalling pathways such as JAK–STAT and TGF-β–SMAD. These changes reshape the SASP towards pro-inflammatory but immunosuppressive factors including IL-6, galectin-9, amphiregulin (AREG) and extracellular matrix (ECM) components, enabling senescent cells to evade immune clearance and accumulate during ageing and disease. Created in BioRender. Pietrocola, F. (2026) https://BioRender.com/ks4vpfk