Cellular senescence was first described in the 1960s, when normal human fibroblasts were observed to undergo a finite number of divisions before entering a state of permanent growth arrest, a phenomenon later termed replicative senescence^[16]. Senescence was subsequently characterized as a stable cell-cycle arrest induced by diverse intrinsic and extrinsic stressors, serving as a critical safeguard against the propagation of damaged or potentially malignant cells^[10,17]. Senescent cells have recently been incorporated into the expanded hallmarks-of-cancer framework as an important component of the tumor microenvironment^[18], underscoring its complex and increasingly appreciated role in tumor biology.

Tumor senescence can be induced by a broad spectrum of cellular stresses, with persistent DNA damage representing a major initiating event [Figure 1A]. These senescence-inducing stimuli can be broadly classified as intrinsic or extrinsic. Intrinsic drivers include telomere attrition, oncogene activation, such as activation of RAS, BRAF^V600E, or MYC, and genomic instability^[19-21]. Alterations in tumor-suppressor pathways, including those involving PTEN, RB, and LKB1, may also trigger senescence through the compensatory activation of cellular stress-response programs^[21,22]. Clinically relevant extrinsic inducers include DNA-damaging chemotherapy and radiotherapy^[23,24]. Selected targeted regimens can also impose a senescence program in defined tumor contexts, whereas oxidative stress can induce premature senescence through persistent stress signaling^[25]. These stimuli mediate sustained DNA damage responses and activate the p53-p21 and p16-Rb pathways^[26], ultimately establishing durable cell-cycle arrest. Notably, the tumor microenvironment itself constitutes an important source of senescence-inducing stress. Hypoxia, nutrient deprivation, inflammatory signaling, and stromal interactions can all contribute to the accumulation of senescent cells within tumors^[27-29], particularly following therapeutic intervention [Figure 1A]. The resulting enrichment of senescent cells may further amplify local stress signals and propagate senescence to neighboring cells, thereby reinforcing the senescent microenvironment through a feed-forward mechanism.

In the tumor microenvironment, senescent cells undergo irreversible cell cycle arrest driven by sustained activation of cyclin-dependent kinase inhibitors, including p16^INK4a, p21^Cip1/Waf1, and p27^Kip1[30]. These senescent cells undergo a broad spectrum of biological alterations involving cellular morphology, chromatin organization, and metabolic regulation [Figure 1B]. Morphologically, senescent cells are typically enlarged and flattened and exhibit increased lysosomal content and elevated senescence-associated β-galactosidase (SA-β-gal) activity^[31]. At the chromatin level, they display extensive remodeling, including the formation of senescence-associated heterochromatin foci (SAHF)^[32] and the accumulation of cytoplasmic chromatin fragments^[33], linking genomic instability to innate immune signaling^[34]. In parallel, sustained DNA damage signaling and cell-cycle arrest induce extensive transcriptional reprogramming^[35], including the activation of pro-survival pathways that enable senescent cells to resist apoptosis and persist within the tumor microenvironment^[36].

Among them, senescent tumor cells (STCs) undergo profound metabolic reprogramming [Figure 1B], characterized by mitochondrial dysfunction, increased production of reactive oxygen species (ROS), and dysregulated autophagy^[37,38]. Moreover, a senescence-associated secretory phenotype (SASP) emerges as a coordinated secretory program driven by upstream transcriptional regulation [Figure 1B]. The SASP comprises a diverse array of bioactive factors, including pro-inflammatory cytokines (e.g., IL-6)^[39], growth factors (e.g., TGF-β)^[25], matrix-remodeling enzymes (e.g., MMPs)^[40], as well as extracellular vesicles and their miRNA cargo^[41,42]. Acting in autocrine and paracrine manners, the SASP reinforces senescence, remodels the tumor microenvironment, and modulates immune responses^[43]. Notably, the composition and functional consequences of the SASP are highly context-dependent, varying with cell type, inducing stimulus, and temporal progression.

Cellular senescence within tumors is not confined to malignant cells but also arises across stromal, vascular, and immune compartments^[44] [Figure 1C]. Senescent CD8⁺ T cells are characterized by increased expression of p16^INK4a, p21^Cip1/Waf1, CD57, and KLRG1, elevated senescence-associated β-galactosidase (SA-β-gal) activity, and loss of the costimulatory molecules CD27 and CD28^[45]. Several exhaustion-associated inhibitory receptors, including PD-1, TIM-3, and LAG-3, may also be highly expressed in senescent T cells. Senescent and exhausted T cells also share functional features, including impaired proliferative capacity and reduced cytotoxicity^[46]. This phenotypic and functional convergence raises the possibility that activation of senescence-associated signaling contributes to the initiation or stabilization of T-cell exhaustion, although the two states remain biologically distinct^[47].

Tumor-infiltrating myeloid cells, particularly tumor-associated macrophages (TAMs), can similarly acquire senescent or senescence-like states^[48] [Figure 1C]. Senescent TAMs are characterized by lysosomal alterations, metabolic reprogramming, and immunosuppressive activity^[49]. These myeloid populations suppress CD8⁺ T-cell and natural killer (NK)-cell responses through mechanisms that include the upregulation of inhibitory molecules such as PD-L1 and the reduced expression or responsiveness of innate immune pattern-recognition receptors, including Toll-like receptors^[49].

Senescent CAFs are generally characterized by combined p16/p21 expression, SA-β-Gal activity^[50], and a SASP enriched in IL-8, CXCL1, TGF-β, MMPs, and FGF2^[51] [Figure 1C]. Their excessive extracellular matrix deposition, induction of hypoxia, and lactate-mediated metabolic-epigenetic signaling restrict cytotoxic lymphocyte infiltration^[50] and promote tumor plasticity, invasion^[52], and lymphatic metastasis^[53]. Senescent endothelial cells similarly exhibit canonical senescence-associated features, including increased expression of p16^INK4a and p21^Cip1/Waf1 and elevated SA-β-gal activity^[54]. They also undergo substantial functional alterations characterized by increased oxidative stress and dysregulated adhesive and chemotactic signaling, including VCAM-1- and CXCL11-enriched secretory programs^[54-56]. The resulting loss of endothelial junction integrity increases vascular permeability, disrupts immune-cell trafficking, and facilitates tumor-cell dissemination.

Despite substantial progress, the definition and identification of senescent cells rely on combinatorial assessment of multiple features. Emerging high-throughput technologies, such as single-cell omics and CRISPR-based screens, are expected to refine the molecular taxonomy of senescence and elucidate the mechanisms by which systemic factors and microenvironmental cues (e.g., neural signaling) induce senescence.

Senescent cells can stimulate antitumor immunity via enhancing immunogenicity and recruiting cytotoxic immune cells. Senescence-inducing CDK4/6 inhibition can reprogram mammary tumors toward a more inflamed state by increasing chemokine-mediated T-cell recruitment^[58].

STCs also undergo cell-intrinsic alterations that further enhance their immunogenicity. Notably, the upregulation of antigen presentation machinery, including MHC class I molecules, together with increased expression of IFN-γ receptor 1, enhances their susceptibility to immune recognition and elimination^[13]. In addition, senescent cells express higher levels of stress-induced ligands, such as MICA, ULBP2, and ICAM-1, which engage activating receptors such as NKG2D on NK cells, thereby amplifying cytotoxic responses^[59]. Senescent endothelial compartment can increase vascular permeability and accessibility, thereby enhancing the intratumoral penetration of chemotherapeutic agents and facilitating immune-cell infiltration into tumors^[60,61]. This immunostimulatory function represents an early, protective role of senescence, in which cell-intrinsic immune signaling cooperates with SASP-mediated recruitment and activation of effector cells to establish an effective antitumor surveillance network.

When STCs persist within the tumor microenvironment and evade immune clearance, their immunoregulatory function shifts from immunostimulatory to immunosuppressive. This transition is driven by multiple, interconnected mechanisms, including the recruitment and expansion of immunosuppressive cell populations, the upregulation of immune checkpoint molecules, and the establishment of a senescent niche that sustains local immune suppression.

Current senolytic strategies can be broadly categorized into the direct pharmacological induction of STC apoptosis, sequential “one-two punch” therapy, and immune-mediated clearance of STCs [Figure 4]. Senolytics exploit intrinsic vulnerabilities of senescent cells, most notably their reliance on senescence-associated anti-apoptotic pathways (SCAPs). STCs frequently upregulate pro-survival proteins, such as BCL-XL and MCL-1, that enable their persistence despite cellular damage^[99]. Pharmacological inhibition of these pathways using agents, such as navitoclax (ABT-263)^[100], dasatinib plus quercetin (D+Q)^[101], and fisetin^[36], can selectively induce apoptosis of senescent cells. Clinical translation of senolytics remains early and has largely occurred outside oncology. D+Q has undergone a small randomized phase I pilot study in idiopathic pulmonary fibrosis^[102], fisetin is being evaluated in a phase II sepsis trial^[103]. These studies do not yet establish clinical clearance of STCs or improved cancer outcomes.

Senolytic strategies are particularly attractive when administered sequentially after senescence-inducing therapies, forming the so-called “one-two punch” approach. In this therapeutic paradigm, tumor cells are first driven into a senescent state by treatments such as chemotherapy, radiotherapy, or CDK4/6 inhibitors, and are subsequently eliminated using senolytic agents^[104]. This sequential strategy can promote the clearance of residual STCs while limiting the long-term detrimental consequences of their persistence, including metastatic progression, immune evasion, and therapeutic resistance^[57,104]. However, a major challenge for this sequential strategy is that the optimal interval between senescence induction and senolytic administration is unlikely to be fixed across tumor types, inducing agents, and cellular compartments. Longitudinal imaging of senescence-associated signals may therefore help identify the emergence and persistence of senescent-cell populations and guide the timing of the second therapeutic intervention.

An alternative and increasingly explored strategy involves enhancing the immune-mediated clearance of senescent cells. Given that immune surveillance represents a physiological mechanism for STC elimination, therapeutic approaches aim to potentiate this process. For instance, chimeric antigen receptor (CAR)-T or CAR-NK cells targeting senescence-associated surface antigens ( e.g., uPAR) have demonstrated selective elimination of STCs in tumor models^[105]. Similarly, vaccines directed against senescence-associated antigens (e.g., CD153 or GPNMB) can induce antibody-dependent cellular cytotoxicity and facilitate immune clearance^[106]. Although these approaches offer high specificity, their clinical translation remains constrained by the limited availability of robust and universal senescence-specific targets, as well as concerns regarding immune-related toxicity.

In contrast to senolytics, senomorphic strategies aim to modulate the functional phenotype of senescent cells, particularly the production and activity of SASP components, thereby mitigating their detrimental effects on the iTME. The SASP production is governed by multiple interconnected signaling pathways, including mTOR, MAPK, cGAS-STING, JAK-STAT, and NF-κB signaling pathways^[107-109]. Pharmacological targeting of these pathways using agents such as rapamycin or everolimus (mTOR inhibitors), losmapimod (p38/MAPK inhibitor), metformin, or vitamin D has been shown to suppress the production of SASPs^[110-113]. Moreover, direct neutralization of individual SASP components using antibodies against IL-11, IL-6, or IL-1β has demonstrated therapeutic potential in models of aging-related diseases^[114]. By attenuating chronic inflammatory profile, these interventions can partially restore immune surveillance and enhance antitumor efficacy.

Notably, SASP modulation does not necessarily require its global suppression. Emerging evidence suggests that selective reprogramming of the SASPs rather than its complete inhibition may enhance antitumor immunity^[115]. For instance, certain therapeutic combinations can induce more immunostimulatory SASPs, promoting recruitment of NK cells and T cells while limiting immunosuppressive signaling^[116]. These findings underscore the context-dependent nature of senomorphic interventions and suggest that preserving beneficial immune-activating signals may be critical to achieving optimal therapeutic outcomes.

In addition to targeting SASP regulatory pathways, immune checkpoint blockade represents a complementary strategy for counteracting the immunosuppressive phenotype of STCs. STCs frequently upregulate inhibitory ligands, including PD-L1, PD-L2, and HLA-E, which suppress cytotoxic immune responses and facilitate immune evasion^[117]. Antibodies targeting the PD-1/PD-L1 or NKG2A/HLA-E axis can restore T-cell and NK-cell activity^[118], thereby promoting the immune-mediated clearance of STCs and potentially enhancing the efficacy of immunotherapy.

Given the involvement of multiple cellular compartments in senescence-mediated immune remodeling, therapeutic strategies should extend beyond targeting STCs alone to encompass the broader senescent TME. As a dominant stromal population within tumors, senescent CAFs are one of the major sources of immunosuppressive SASPs^[119]. The selective elimination of senescent CAFs or suppression of their SASPs has been shown to restore immune surveillance and improve responses to chemotherapy^[120]. Similarly, targeting senescent macrophages may alleviate local immune suppression, restrain early tumorigenesis, and promote the intratumoral accumulation of cytotoxic T cells. Several senolytic and senomorphic agents, including navitoclax, D+Q, and niacin, have demonstrated therapeutic potential against senescent macrophages in preclinical studies and early clinical investigations^[121-124].

Senescent T cells are characterized by diminished cytotoxicity and impaired responsiveness to immunotherapy. Unlike strategies targeting senescent stromal cells, therapeutic approaches to T-cell senescence primarily focus on functional rejuvenation rather than cellular elimination. Pharmacological inhibition of PI3Kδ with leniolisib or p38 MAPK with BIRB 796 has been reported to attenuate T-cell senescence in several oncological contexts, including hematologic malignancies and solid tumors following chemotherapy^{[110,125,126]}. Genetic or epigenetic modulation of p53- or SIRT1-associated pathways in CD8⁺ T cells may similarly reduce senescent features while enhancing their migratory capacity and cytotoxic function^[127]. Beyond these cell-intrinsic pathways, metabolic reprogramming represents another promising therapeutic strategy. Modulation of the FOXO1-SIRT1 axis can regulate glycolytic metabolism and granzyme B production, whereas targeting group IVA phospholipase A2-mediated lipid-droplet accumulation may help preserve metabolic homeostasis^[128,129]. Neuroendocrine signals mediated through β-adrenergic receptors can also influence the metabolism and function of senescent T cells, identifying catecholamine-β-adrenergic signaling as a potential therapeutic target^[130].

Collectively, these findings support a paradigm shift in senotherapy from indiscriminate targeting of senescent cells toward a multi-level intervention strategy that integrates signaling pathway modulation, epigenetic reprogramming, metabolic regulation, and immune niche remodeling. Collectively, senomorphic modulation and cell-type-specific targeting of senescent stromal and immune populations provide complementary strategies for suppressing deleterious SASP signaling, restoring immune-cell function, and remodeling the senescent iTME [Figure 5].

To address these challenges, prodrug strategies have emerged as a promising solution. Prodrugs are inactive compounds designed to respond to specific tumor microenvironment stimuli, such as pH, redox conditions, enzymes, or hypoxia^[134]. The integration of conventional senolytic agents with prodrug strategies holds promise for achieving precise senescence intervention. For instance, González-Gualda et al. developed Nav-Gal, a galactose-conjugated navitoclax prodrug activated by SA-β-Gal in senescent cells. In non-small cell lung cancer (NSCLC) models, Nav-Gal combined with cisplatin reduced tumor burden more effectively than monotherapy, while exhibiting reduced navitoclax-related platelet toxicity^[135]. Similarly, Guerrero et al. designed a galactose-modified duocarmycin (GMD) prodrug that triggers senescent cell apoptosis via SA-β-Gal. In a craniopharyngioma mouse model, GMD selectively eliminated β-catenin-accumulating senescent clusters without affecting normal pituitary cells, demonstrating its potential for targeted elimination of preneoplastic senescent cells^[136]. However, the repertoire of prodrug activation strategies that enable specific responsiveness to senescent cells remains limited, warranting further development of novel approaches tailored to the unique features of the senescent microenvironment.

Targeted protein degradation (TPD) technologies, such as proteolysis-targeting chimeras (PROTACs), enable selective elimination of disease-relevant proteins, offering potential advantages including reduced off-target toxicity and sustained therapeutic effects^[137]. PROTACs are bivalent small molecules consisting of a ligand that recognizes the target protein linked to an E3 ubiquitin ligase, thereby facilitating selective protein degradation^[138,139]. Recent studies have leveraged PROTAC technology to transform navitoclax into a platelet-sparing targeted chimera, addressing navitoclax-induced thrombocytopenia while enhancing senolytic activity^[137,140]. A VHL PROTAC DT2216 targeting BCL-XL could induce STC apoptosis while minimizing platelet toxicity due to the low dependence of platelets on VHL-mediated degradation^[137]. DT2216 has demonstrated excellent antitumor performance in drug-resistant triple-negative breast cancer and patient-derived T-ALL models^[137], underscoring the high therapeutic potential of TPD in tumor senotherapy. Beyond PROTAC-based approaches, expanding the application of TPD technologies, including molecular glue, lysosome-targeting chimeras (LYTACs), and antibody-based degraders, holds great promise for developing next-generation senescence-targeting therapies.

Drug delivery systems (DDS) have attracted considerable interest in senotherapy, offering the ability to actively target senescent cells or specific subsets thereof, prolong drug half-life, and minimize off-target distribution^[141-143]. Mesoporous silica nanoparticles (MSNs) possess favorable chemical properties, thermal stability, and biocompatibility, promoting their application in senotherapy^[144]. Several studies have developed MSN-based formulations that exploit elevated SA-β-Gal activity in senescent cells for targeted delivery of navitoclax^[145-149]. These systems, capped with galacto-oligosaccharide derivatives, enable selective release within senescent cells, thereby reducing navitoclax-related thrombocytopenia while enhancing senolytic efficacy^[145,150]. In preclinical models, such formulations have demonstrated improved tumor growth inhibition and reduced metastasis when combined with senescence-inducing therapies (e.g., palbociclib or doxorubicin), while mitigating off-target toxicities including endothelial dysfunction and cardiotoxicity^[150,151].

The SASP produced by STCs exerts pleiotropic effects, including immune cell recruitment and enhanced vascular permeability, which can be harnessed for drug delivery^[152,153]. Leveraging these properties, cell membrane coating technology has emerged as an innovative approach enabling immune evasion, prolonged circulation, and targeted delivery. Pan et al. demonstrated that STCs exhibit enhanced antigen-presenting capacity and engineered their membranes into nanovesicles (nano-APM) to selectively prime splenic CD8⁺ T cells. Sequential combination of nano-APM with RT achieved robust antitumor efficacy, as evidenced by significantly delayed tumor growth and prolonged survival^[154]. Yang et al. engineered STC membranes into biomimetic nanovaccines. Such formulations enhance dendritic cell internalization and lymph node targeting, promoting antitumor immune responses. In murine melanoma models, these nanovaccines combined with ICIs significantly suppressed tumor growth^[155]. A complementary approach utilizes senescent red blood cell membranes fused with tumor cell membrane-associated antigens. These nanoformulations accumulate in the spleen, activate antigen-presenting cells, and synergize with anti-PD-L1 therapy to achieve durable antitumor effects^[156]. Venetoclax is a selective BCL-2 inhibitor with established activity in hematologic malignancies, although its role as a broad-spectrum senolytic remains unproven^[157]. Senescent-cell vaccination has also reduced tumor growth in several preclinical solid-tumor models^[158].

Collectively, current advanced senotherapies show promise in senescence intervention. Integrating prodrug strategies, TPD technologies, and advanced DDS is anticipated to pave the way for next-generation senotherapies. The convergence of these approaches may ultimately enable safe, effective, and broadly applicable senotherapeutic interventions.