A senescent tumor cell-derived nanovesicle directly primes splenic T cells to potentiate cancer radiotherapy

Summary

Radiotherapy (RT)-induced senescent tumor cells (STCs) reinforce an immunosuppressive tumor microenvironment (ITM) and compromise therapeutic outcomes. However, current senolytic strategies lack specificity for STCs and often cause off-target toxicity. Here, we observe that STCs possess enhanced antigen-presenting capacity in patient-derived tumor tissues and murine tumor models. Leveraging this phenomenon, we engineer STC-derived nanovesicles (termed nano-APM) for preserving endogenous antigens and antigen-presenting cues. We demonstrate that systemically administered nano-APMs accumulate in the spleen and establish a pool of STC-specific CD8⁺ T cells. Sequential integration of RT induces local tumor senescence, and nano-APMs then effectively mobilize the STC-specific T cells to stimulate a confined recall response. In murine tumor models, the combination of nano-APM plus RT selectively eliminates STCs, reprograms RT-induced ITM, and elicits durable antitumor immunity. Collectively, this study establishes STC-derived nanovesicles as a practical means to enhance RT efficacy by enabling splenic T cell priming and spatiotemporally confined senolysis.

Graphical abstract

Highlights

• •STCs exhibit potent antigen-presenting activity and directly prime CD8⁺ T cells
• •STC-derived nano-APMs expand STC-specific T cell pools in the spleen
• •Nano-APMs plus radiotherapy drive localized senolysis and immune restoration

Pan et al. engineer senescent tumor cell (STC) membranes into spleen-homing nano-antigen-presenting machinery (nano-APMs) to prime STC-specific T cell immunity, enabling spatiotemporally confined senolytic therapy through sequential combination with radiotherapy.

Introduction

Radiotherapy (RT) is a cornerstone treatment for solid malignancies.¹ RT regresses tumors via DNA damage-driven cell death and eliciting immunostimulatory signals.^2,3,4 Recent evidence suggests that RT-induced senescent tumor cells (STCs) represent a double-edged sword in cancer therapy.^5,6,7 Specifically, STCs can promote tumor suppression by enhancing immune-stimulatory signaling, yet paradoxically foster an immunosuppressive tumor microenvironment (ITM) that supports tumor progression.^8,9,10 This immunosuppressive program is driven in part by upregulation of immune checkpoints such as programmed death ligand 1 (PD-L1),^11,12,13 secretion of immunosuppressive cytokines (e.g., interleukin-1α), and recruitment of regulatory immune subsets (e.g., myeloid-derived suppressive cells).^14,15,16 Thus, selective elimination of STCs has emerged as a promising therapeutic avenue to restore antitumor immunity and improve RT.

To date, the reported strategies for eliminating STCs mainly follow two paradigms: small-molecule senolytics that induce apoptosis by disrupting pro-survival pathways and immune-based approaches that promote antigen-specific clearance, including chimeric antigen receptor (CAR)-T cells and senescence-associated vaccines.^17,18,19 Small-molecule senolytics, including anti-apoptotic protein inhibitors (e.g., navitoclax),²⁰ bromodomain and extraterminal domain (BET) family degraders (e.g., ARV825),^21,22 and p53 activity modulators (e.g., UBX0101),^23,24 lack tumor specificity and cause off-target toxicity in normal cells.²⁵ Despite the promising preclinical output of CAR-T and vaccination therapy, their clinical translation remains limited in solid tumors.^26,27 CAR-T-based senolytic therapy, while more selective, requires complex ex vivo expansion and reinfusion.²⁸ Moreover, once infused, CAR-T cells remain constitutively active, thereby raising concerns for sustained cytotoxicity and immune-related adverse effects.²⁹ Vaccine-based strategies rely on predefined or artificial antigens, which may not fully represent the antigenic landscape of STCs.^30,31 It is essential to develop alternative options for specific elimination of STCs to enhance RT.

In this study, we harness STC-derived nanovesicles to enable spatiotemporally confined immune responses for selective clearance of RT-induced STCs. We first observed that STCs from patient pancreatic and bladder tumors exhibited enhanced antigen-presenting capacity, characterized by upregulation of major histocompatibility complex (MHC) class I and co-stimulatory molecules (CD80 and CD86). Guided by this finding, we engineered a nano-antigen-presenting machinery (nano-APM) from STC membranes, which preserved endogenous tumor antigens and immunostimulatory cues. Upon systemic administration, nano-APMs accumulated in the spleen and directly primed STC-specific CD8⁺ T cells, establishing a dedicated T cell pool. Subsequent RT induced tumor-localized senescence and antigen presentation, which mobilized this pool and triggered a potent recall response restricted to the irradiated site. In murine models of pancreatic ductal adenocarcinoma and bladder cancer, this sequential regimen achieved selective STC clearance, enhanced antitumor immunity, and dramatically delayed tumor progression. Such STC-derived nanovesicles provide a practical and versatile means to potentiate RT via directly priming splenic T cells and enabling spatiotemporally confined elimination of STCs.

Results

STCs in PDAC and BLCA patient tumors exhibit enhanced antigen presentation features

To explore the clinical relevance of tumor senescence, we analyzed The Cancer Genome Atlas (TCGA) cohort and found that patients with high senescence signature scores (high seno-sig cohort) had significantly worse overall survival than those with low scores (low seno-sig cohort) (Figures 1A and S1A). This adverse prognostic association was observed across tumor types with distinct immunogenicity, including poorly immunogenic pancreatic ductal adenocarcinoma (PDAC) and highly immunogenic bladder cancer (BLCA). Interestingly, gene set enrichment analysis revealed that high seno-sig tumors showed significant activation of antigen processing and presentation pathways (Figures 1B and 1C), particularly genes encoding MHC class I molecules (MHC-I, HLA-A/B/C) and co-stimulatory molecules (CD80 and CD86) (Figure 1D).

Figure 1.

STCs upregulated antigen-presentation-associated molecules in human tumor tissues

(A) Survival analysis of low and high senescence signature (seno-sig) cohort of PDAC and BLCA.

(B) KEGG pathway enrichment analysis of the upregulated genes in high seno-sig PDAC and BLCA cohort.

(C) The expression of antigen-presenting-associated genes in low and high seno-sig patients with PDAC or BLCA.

(D and E) Multiplex immunofluorescence (mIF) analysis of MHC-I expression on STCs via staining nuclei (blue), CK19 (purple), P21 (green), and MHC-I (red) in PDAC and BLCA tumor tissue with or without RT treatments.

Data are presented as the mean ± SD. Data in (A) were estimated using the Kaplan-Meier method. Data in (D) were assessed via two-tailed unpaired Student’s t test.

See also Figure S1.

To examine MHC-I expression in STCs, we performed multiplex immunofluorescence (mIF) examination on PDAC- and BLCA-patient-derived tumor sections. Tumors collected after RT displayed an increased frequency of CK19⁺P21⁺ STCs compared with tumors without RT (Figures 1E, 1F, and S1D), indicating RT-induced senescence. To further confirm their senescent identity, we profiled additional senescence markers and found that these CK19⁺P21⁺ cells showed concomitant upregulation of P16 and histone H3 lysine 9 trimethylation (H3K9me3) (Figures S1B and S1C). Importantly, MHC-I expression was significantly higher in CK19⁺P21⁺ STCs than in non-senescent tumor cells (CK19⁺P21⁻) in both PDAC and BLCA tumors (Figures 1E, 1F, and S1E). These data supported that STCs in human PDAC and BLCA displayed enhanced antigen-presenting features.

STCs exhibit enhanced antigen-presenting capacity in murine PDAC and BLCA cell models

To further evaluate the antigen-presenting potential of STCs, we established in vitro senescent models of murine PDAC and BLCA cells (sKPC cells and sMB49 cells) under RT.³² Increased β-galactosidase (β-Gal) and elevated p21 expression confirmed the senescent state of both sKPC and sMB49 cells (Figures 2A and 2B). Transcriptomic profiling further revealed upregulation of key senescence markers (e.g., Glb1 and Cdkn2a) and SASP components (e.g., Il1a, Il6, and Cxcl9) (Figures S2A and 2B). Enrichment analysis of upregulated genes showed significant activation of “antigen processing and presentation” process in both senescent models (Figures 2C and 2D). Consistently, antigen-presentation-associated genes, including H-2D1, H-2K1, H-2Q4, H-2T22, H-2T23, B2m, Tapbp, Tap1, and Tap2, were broadly upregulated (Figures S2C and 2D). These data suggested that STCs might upregulate antigen-processing and MHC-I antigen-presentation programs.

Figure 2.

STCs markedly upregulated antigen-presentation-associated molecules in vitro

(A) Western blot analysis of p21 in KPC cells and MB49 cells (n = 3).

(B) Representative micrographs of KPC cells and MB49 cells stained for SA-β-gal activity.

(C and D) KEGG enrichment analysis of the upregulated genes in the STCs (sMB49 cells or sKPC cells).

(E and F) FCM analysis of CD80, CD86, and H-2Kb in sKPC cells and sMB49 cells (n = 3).

(G and H) FCM analysis of CD80, CD86, and H-2Kb in sKPC cells and sMB49 cells after treatments of IFN-γ with various concentrations (n = 3).

(I) FCM analysis of the expression of H-2Kb^SIINFEKL in IFN-γ-amplified senescent or non-senescent KPC^OVA cells and MB49^OVA cells (n = 3).

(J and K) CLSM images of the expression of H-2Kb^SIINFEKL in non-senescent and IFN-γ-amplified senescent or KPC^OVA cells and MB49^OVA cells (n = 3; scale bars, 25 μm).

(L) Schematic illustration for the upregulation of antigen-presentation-associated molecules on STCs.

Data are presented as the means ± SD. Data in (A), (E), (F), (J), and (K) were assessed via two-tailed unpaired Student’s t test. Data in (G)–(I) were assessed via one-way ANOVA with Tukey’s test.

See also Figure S2.

Flow cytometric (FCM) analysis confirmed increased surface expression of H-2Kb (one subtype of mouse MHC-I) together with the costimulatory molecules CD80 and CD86 on both sKPC and sMB49 cells (Figures 2E, 2F, S2E, and S2F), suggesting the acquisition of antigen-presenting and costimulatory cues in STCs. RT has been shown to induce interferon gamma (IFN-γ), which promotes antigen presentation.^33,34,35 Then we measured IFN-γ levels in murine PDAC and BLCA tumor models and observed a significant increase after RT by enzyme-linked immunosorbent assay (Figures S2G and S2H). To model this effect in vitro, we treated KPC and MB49 cells with IFN-γ and found markedly increased surface expression of CD86, CD80, and H-2Kb (Figures S2I and S2J). Furthermore, we employed ovalbumin (OVA)-overexpressed KPC cells (KPC^OVA) and MB49 cells (MB49^OVA) and measured the surface presentation of the OVA-derived SIINFEKL(OVA257-264) peptide via MHC-I. Confocal laser scanning microscope (CLSM) examination showed that SIINFEKL-H-2Kb complex (H-2Kb^SIINFEKL) was significantly higher in tumor cells upon IFN-γ treatment (Figure S2K), supporting IFN-γ-mediated amplification of antigen presentation.

We then applied IFN-γ to STCs and found a markedly stronger upregulation of antigen-presentation-associated molecules. FCM analysis showed that IFN-γ increased H-2Kb expression by 31-fold in sMB49 cells and 63-fold in sKPC cells and also significantly elevated CD80 and CD86 (Figures 2G and 2H). Consistently, under the same IFN-γ stimulation, sKPC^OVA and sMB49^OVA cells exhibited 1.7- and 2.4-fold higher H-2Kb^SIINFEKL, respectively, than their non-senescent counterparts (Figures 2I and S2L). CLSM examination further confirmed elevated H-2Kb^SIINFEKL expression on STCs (Figures 2J and 2K). Together, these results demonstrated that IFN-γ could further amplify antigen presentation in STCs, thereby strengthening signal 1 and signal 2 cues that might support CD8⁺ T cell engagement (Figure 2L).

STCs directly prime CD8⁺ T cells in vitro

To investigate whether STCs can directly activate CD8⁺ T cells, we established a coculture system using sKPC and sMB49 cells with CD8⁺ T cells (Figures 3A and S3A). FCM analysis revealed that sMB49 cells significantly enhanced the proportion of activated CD69⁺CD8⁺ T cells (65.9% ± 0.7%) compared to their non-senescent counterparts (57.5% ± 1.0%) (Figures 3B and 3C). IFN-γ treatment further enhanced CD8⁺ T cell activation to 68.3% ± 1.3%, reaching the level achieved by anti-CD3 plus anti-CD28 antibodies (αCD3/αCD28) (Figure 3C). In contrast, KPC and sKPC cells failed to activate CD8⁺ T cells, with CD8⁺CD69⁺ frequencies of 13.1% ± 0.4% and 11.4% ± 0.4%, respectively (Figures 3B and 3D). Even after IFN-γ stimulation, sKPC cells exhibited only marginal improvement in T cell activation. However, supplementation with αCD28 to enhance co-stimulatory “signal 2” significantly increased the activation frequency of CD8⁺ T cells from 24.9% ± 0.6% to 36.1% ± 0.4% (Figures 3B and 3D), suggesting that inhibitory signaling in sKPC cells might limit T cell activation.

Figure 3.

STCs directly prime CD8⁺ T cells

(A) Schematic illustration for the potential mechanism of STC-mediated direct T cell priming.

(B) Representative FCM plots of activated CD8⁺CD69⁺ T cells incubated with αCD3/αCD28 or tumor cells under various conditions, including pretreated with IFN-γ, RT, sequential RT and IFN-γ, sequential RT, IFN-γ, and α-CD28 antibody, with non-treated controls.

(C and D) The frequency of activated CD8⁺ T cells after coculture with tumor cells or incubated with αCD3/αCD28 (n = 3).

(E) The expression of inhibitory or stimulatory genes in sKPC or non-senescent KPC cells (n = 3).

(F) CLSM examination on PD-L1 expression in senescent or non-senescent KPC cells (n = 3; scale bars, 25 μm).

(G) FCM analysis of PD-L1 expression on KPC, sKPC, IFN-γ-sKPC, MB49, sMB49, and IFN-γ-sMB49 cells (n = 3).

(H) FCM analysis of PD-L1 expression of sKPC cells with PD-L1 knockdown models (n = 3).

(I) FCM analysis of activated CD69⁺ CD8⁺ T cells cocultured with sKPC^KD cells (n = 3).

(J) Schematic illustration for the direct T cell activation via STCs.

Data are presented as the mean ± SD. Data in (E)–(G) were assessed via two-tailed unpaired Student’s t test. Data in (C), (D), (H), and (I) were assessed via one-way ANOVA with Tukey’s test.

See also Figure S3.

To explore potential inhibitory mechanisms, we profiled immunoregulatory molecules in sKPC cells and found significant upregulation of inhibitory checkpoints (Figure 3E). Among them, Cd274 (coding PD-L1 protein) showed the strongest induction, increasing by 23.3-fold, which was further confirmed by CLSM examination (Figure 3F). IFN-γ treatment further elevated PD-L1 expression by 6.1-fold compared to untreated sKPC cells (Figure 3G). Notably, PD-L1 expression was intrinsically higher in KPC than in MB49 cells and remained 2.4- to 8.6-fold higher across RT and IFN-γ stimulation. These quantitative differences in PD-L1 expression correlate with the weaker T cell priming efficacy of sKPC compared to sMB49 cells, implicating PD-L1 as a key brake on STC-mediated activation of CD8⁺ T cells.

We next performed loss-of-function validation using a PD-L1 knockdown model of KPC (KPC^KD) cells (Figure S3B). FCM analysis confirmed reduced PD-L1 expression in sKPC^KD cells and IFN-γ-treated sKPC^KD cells (Figure 3H). Then, we cocultured sKPC^KD cells with CD8⁺ T cells. FCM analysis showed that CD8⁺ T cell activation increased from 12.8% ± 0.5% with sKPC cells to 50.0% ± 0.7% with sKPC^KD cells (Figure 3I). In addition, anti-PD-L1 antibody (α-PD-L1) further enhanced CD8⁺ T cell activation in cultures containing IFN-γ-treated sKPC^KD cells (Figure S3C), supporting PD-L1 as a major inhibitory checkpoint limiting STC-mediated T cell activation. To assess functional cytotoxic differentiation, we measured granzyme B expression in CD8⁺ T cells and found that IFN-γ-amplified sMB49 or IFN-γ-amplified-sKPC^KD cells remarkably increased the frequency of CD8⁺granzyme B⁺ T cells than their non-senescent or without IFN-γ-stimulated counterparts (Figures S3D and S3E). Together, these results suggested that STCs could directly prime CD8⁺ T cells through antigen-presenting machinery, which can be enhanced by IFN-γ and restrained by PD-L1.

STC-derived nano-antigen-presenting machinery directly primes CD8⁺ T cells and specifically accumulates in the spleen

To enable safe and controllable exploitation of STC immunogenicity, we constructed nano-antigen-presenting machinery (nano-APM) by isolating cell membranes from sMB49 and sKPC^KD cells and reassembling them into nanovesicles (Figure 4A). This formulation preserved antigen-presenting functionality while avoiding the carcinogenic risk of intact STCs. Transmission electron microscopy (TEM) and dynamic light scattering (DLS) examination confirmed a uniform spherical morphology with an average diameter of 92–130 nm for both senescent and control nano-APMs (Figures 4B and S4A). Nano-APM also displayed superior colloidal stability upon 24 h of incubation within phosphate-buffered saline (PBS) at 4°C or 37°C (Figure S4B).

Figure 4.

STC-derived nano-antigen-presenting machinery (nano-APM) specifically primes splenic CD8⁺ T cells in vivo

(A) Schematic illustration for engineering senescent tumor cells to nano-antigen-presenting machinery (nano-APM).

(B) DLS and TEM characterization of KPC-cell-derived nano-APM.

(C and D) FCM analysis of activated CD8⁺CD69⁺ T cells cocultured with tumor cell-derived nano-APMs (n = 3).

(E) Cytotoxicity assay detecting the cell viability of KPC or sKPC cells cocultured with nano-APM-primed CD8⁺ T cells or naive CD8⁺ T cells (n = 3).

(F) ELISA assay detecting the secretion of IFN-γ in the coculture system of STCs and nano-APM-primed CD8⁺ T cells (n = 3).

(G) Schematic illustration for conjugating the Cy7 photosensitizer onto nano-APM by a click chemistry strategy.

(H) CLSM examination of T cells incubated with Cy7-linked sKPC^KD APM or Cy7-linked KPC APM (scale bars, 3 μm).

(I) Ex vivo fluorescence imaging of major organs (i.e., the heart, liver, spleen, lungs, and kidneys) and tumors at 2 h post-injection of nano-APM (n = 3).

(J) Semiquantitative analysis of the distribution of nano-APM in major organs and tumors (n = 3).

(K) mIF analysis of splenic CD8⁺ T cells (green), CD11c⁺ cells (DCs, red), nano-APM (purple) (scale bars, 100 μm [left] and 20 μm [right]).

Data are presented as the mean ± SD. Data in (J) were assessed via two-tailed unpaired Student’s t test. Data in (D)–(F) were assessed via one-way ANOVA with Tukey’s test.

See also Figures S4 and S5.

We next investigated whether nano-APMs could activate CD8⁺ T cells in vitro. Compared to their non-senescent counterparts (MB49 APM and KPC APM), sMB49 APM and sKPC^KD APM significantly enhanced CD8⁺ T cell activation by 1.7- and 2.1-fold, respectively (Figures 4C and 4D), indicating that STC-derived nanovesicles retained functional priming capacity of T cells. To demonstrate the specificity of nano-APM-primed T cells, we re-exposed nano-APM-primed T cells to senescent or non-senescent tumor cells. Cytotoxicity assay revealed that T cells primed with sKPC^KD APMs or sMB49 APMs showed minimal killing of KPC or MB49 cells, respectively, but robust cytotoxicity toward STCs (Figures 4E and S4C). Consistently, ELISA analysis showed that coculture of sKPC^KD-APM or sMB49-APM-primed T cells with sKPC or sMB49 cells induced a 2.1- or 2.3-fold higher IFN-γ secretion than that with non-senescent tumor cells, respectively (Figures 4F and S4C). Moreover, OVA-expressing sMB49-derived nano-APMs (sMB49^OVA APMs) effectively stimulated OT-I T cells, expanding IFN-γ⁺ OT-I cells by 8.1-fold over untreated controls (Figure S4D). These results suggested that STC-derived nano-APMs could prime STC-reactive and antigen-dependent T cell responses.

To visualize nano-APM-T cell interactions, sKPC^KD and KPC^KD APMs were labeled with Cy7 via our reported click chemistry strategy³⁶ (Figures 4G and S5A–S5C). CLSM imaging confirmed efficient labeling of Cy7 on the membrane of STCs (Figure S4E). Subsequent FCM analysis validated the retention of CD80 and H-2Kb on Cy7-sKPC^KD APMs (Figures S4F and S4G). Co-incubation of Cy7-linked nano-APM with CD8⁺ T cells showed markedly higher colocalization between Cy7 signal and H-2Kb in the Cy7-sKPC^KD APM group compared to that in KPC^KD APM, suggesting the affinity between STC-derived nano-APM and T cells (Figure 4H).

We then explored the in vivo biodistribution of nano-APMs in a PDAC-tumor-bearing mouse model. Following systemic administration, ex vivo fluorescence imaging of major organs showed a 2.2-fold enrichment of Cy7-labeled sKPC^KD APM in the spleen compared to KPC-derived controls, while accumulation in tumors and other major organs remained minimal (Figures 4I and 4J). mIF confirmed this selective accumulation in splenic tissue (Figures 4K and S4H). Moreover, the colocalization between nano-APM and CD8⁺ T cells was dramatically higher in Cy7-linked sKPC^KD APM than its counterpart (Figures 4K and S4I), supporting direct in vivo T cell engagement. Partial colocalization of nano-APM with CD11c⁺ dendritic cells (DCs) was also observed (Figure 4K), suggesting additional interaction with professional antigen-presenting cells. Together, these findings demonstrated that STC-derived nano-APMs preferentially accumulated in the spleen and effectively engaged T cells in vivo.

Spleen-targeted nano-APM expands STC-specific CD8⁺ T cells and establishes immune memory

Given the preferential accumulation of STC-derived nano-APMs in the spleen, we next investigated their capacity to drive CD8⁺ T cell priming in vivo. Healthy mice were intravenously (i.v.) injected with nano-APMs, and spleens were collected at 3, 5, or 7 days post-administration (Figure 5A). FCM analysis revealed no significant change in DC and mature DC frequency (Figures 5B, 5C, S5D, and S5E) but showed a progressive increase in CD3⁺ T cells (Figures S5F and S5G). Notably, the frequency of CD8⁺ T cells increased from 33.0% ± 1.2% to 38.0% ± 1.0% (Figures 5D and 5E), accompanied by a 1.5-fold rise in activated CD8⁺ T cells by day 7 (Figure 5F). Nano-APM immunization also promoted the expansion of effector memory CD8⁺ T cells (T_EM), with their frequency increasing from 15.8% ± 1.5% to 33.6% ± 2.8% (Figures 5G and 5H), and elevated the ratio of T_EM to central memory T lymphocytes (T_CM) by 2.3-fold higher than that without nano-APM treatments (Figures 5I, S5H, and S5I). These results demonstrated that nano-APM induction expanded splenic CD8⁺ T cells and established immunological memory.

Figure 5.

STC-derived nano-APM establishes a bank of STC-specific CD8⁺ T cells in the spleen

(A) Schematic illustration for i.v. injection of nano-APM into mice at predetermined time points.

(B and C) Representative FCM plots and frequency of mature DCs (CD11c⁺CD80⁺CD86⁺) in the spleen (n = 5).

(D and E) Representative FCM plots and frequency of CD8⁺ T cells and CD4⁺ T cells (n = 5).

(F) Number of activated CD8⁺ T cells (CD8⁺CD69⁺) per 5 × 10⁵ cells (n = 5).

(G and H) Representative FCM plots and the frequency of splenic T_EM (CD45⁺CD3⁺CD8⁺CD44⁺CD62L⁻) and T_CM (CD45⁺CD3⁺CD8⁺CD44⁺CD62L⁺) (n = 5).

(I) The ratio of T_EM/T_CM over time after nano-APM treatment.

(J) H&E analysis of spleen over time after nano-APM treatment (scale bars, 200 μm).

(K) The largest diameter of lymphoid follicles within the spleen over time post-nano-APM treatment (n = 20 from five mice).

(L) mIF examination of CD8 in the spleen at 0 and 7 days post-nano-APM treatment (scale bars, 200 μm [left] and 50 μm [right]).

(M) Splenic tetramer⁺ CD8⁺ T cells of mice after treatments of MB49^OVA or sMB49^OVA APM (n = 3).

(N) Schematic illustration for the harvest of nano-APM-immunized CD8⁺ T cells and the establishment of coculture model of nano-APM-immunized CD8⁺ T cells with KPC or sKPC cells.

(O and P) Representative FCM plots and the frequency of CD8⁺IFN-γ⁺ and CD8⁺TNF-α⁺ T cells (n = 3).

Data are presented as the means ± SD. Data in (C), (E), (F), (H), (I), (K), (M), and (P) were assessed via one-way ANOVA with Tukey’s test.

See also Figure S5.

Histological analysis of spleen sections further supported splenic activation after nano-APM immunization. Hematoxylin and eosin (H&E) revealed progressive enlargement of germinal centers within the lymphoid follicles from 309.5 ± 83.9 μm to 619.7 ± 159.2 μm, along with an increased abundance of vascular structures within the follicles (Figures 5J and 5K). Consistently, mIF examination showed increased CD8⁺ T cells, CD20⁺ B cells, and blood vessels within periarteriolar lymphoid sheath of immunized spleens (Figure 5L). These observations suggested that nano-APM immunization reshaped the splenic niche to facilitate immune cell mobilization and trafficking.

To directly assess antigen specificity, sMB49^OVA APMs were intravenously injected into healthy mice. FCM analysis of splenic CD8⁺ T cells showed a 27.5- and 4.1-fold increase in SIINFEKL-specific Tetramer⁺ CD8⁺ T cells (OVA-specific) compared to the control group and MB49^OVA APM group (Figures 5M and S5J), respectively. Functional specificity was further validated by coculturing tumor cells (KPC or sKPC cells), with CD8⁺ T cells isolated from the spleens of mice immunized with sKPC^KD or KPC APMs (Figure 5N). FCM analysis showed that only CD8⁺ T cells from sKPC^KD-APM-immunized spleens elicited robust effector responses against sKPC cells, as indicated by 3.6- to 11.7-fold increase in CD8⁺IFN-γ⁺ T cells and a 2.6- to 5.6-fold increase in CD8⁺TNF-α⁺ T cells compared to control groups (Figures 5O and 5P). Collectively, these findings demonstrated that spleen-targeted nano-APMs expanded functional STC-reactive CD8⁺ T cells and promoted immune memory.

Nano-APM-enhanced RT potentiates antitumor immunity and suppresses BLCA growth

Building on the robust induction of STC-specific CD8⁺ T cells in the spleen, we next evaluated the antitumor efficacy of a sequential regimen combining nano-APM with RT in a murine BLCA model. Tumor-bearing mice were randomly grouped into the following: control (G1), RT (G2), sMB49 APM (G3), MB49 APM plus RT (G4), or sMB49 APM plus RT (G5) with identical RT dosing and nano-APM dosing (Figure 6A). Tumor growth was modestly inhibited by RT or sMB49 APM monotherapy, with 34.1% and 18.0% inhibition, respectively (Figures 6B and S6A). In contrast, the sequential combination of sMB49 APM plus RT exhibited persistent inhibition on tumor growth, achieving a 74.3% tumor inhibition efficiency by the 19^th day (Figures 6B and S6A). The tumor images and weight of sMB49 APM plus RT displayed a significant decrease by 7.5-, 1.8-, and 1.3-fold than that of RT, sMB49 APM, and MB49 APM plus RT, respectively (Figures 6C and S6B). The median survival time of mice was also significantly extended from 21.5 days in the control group to 27.5 days in the RT group, and 29.5 days in MB49 APM plus RT, to more than 43 days in sMB49 APM plus RT group (Figure 6D). Negligible body weight changes were observed during the experimental period of nano-APM plus RT, suggesting a favorable safety profile of such strategy (Figure S6C).

Figure 6.

Sequential combination of nano-APM and RT activates tumor immunity and suppresses growth of BLCA

(A) Experimental schedule for sMB49 APM treatments in the BLCA model.

(B and C) Averaged tumor growth profiles and images of BLCA model (n = 6).

(D) Survival plots of tumor-bearing C57BL mice (n = 6).

(E) H&E analysis of tumor sections after various treatments (scale bars, 50 μm).

(F) IHC analysis of the expression of p21 in tumor sections after various treatments (scale bars, 50 μm).

(G) FCM-determined frequency of mature DCs in the tdLNs (n = 5).

(H and I) Representative FCM plots of the tumor-infiltrating CD8⁺ and CD4⁺ T cells and the frequency of the tumor-infiltrating CD8⁺ T cells (n = 5).

(J and K) The frequency of intratumoral CD8⁺ Granzyme B⁺ and CD8⁺ IFNγ⁺ T cells (n = 5).

(L) IF staining for CD8 and Foxp3 in tumor sections after treatments (scale bars, 50 μm).

(M) The frequency of splenic T_EM cells (CD3⁺ CD8⁺ CD44⁺ CD62L⁻) (n = 5).

(N) Schematic illustration of the mechanism of sMB49-APM-mediated activation of antitumor immunity in RT controllable manner.

Data are presented as the means ± SD. Data in (J) were assessed via two-tailed unpaired Student’s t test. Data in (G), (I)–(K), and (M) were assessed via one-way ANOVA with Tukey’s test. The statistical significance of the data in (D) was calculated by survival curve comparison with log rank (Mantel-Cox) test.

See also Figure S6.

To explore the mechanism underlying this enhanced efficacy, H&E analysis and terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining confirmed the significantly increased necrosis of tumor tissues in MB49 APM plus RT (Figures 6E and S6D). IHC analysis further demonstrated that p21⁺ cells were markedly increased after both RT and non-senescent MB49 APM plus RT, whereas sMB49 APM plus RT reduced p21 expression, indicating efficient clearance of RT-induced STCs by the sequential regimen (Figure 6F).

Given the pronounced tumor inhibition and STC reduction, we next assessed whether nano-APM-based senolysis enhanced antitumor immunity. FCM analysis of tumor-draining lymph nodes (TdLNs) showed that sMB49 APM plus RT elicited DC maturation by a percentage of 29.3% ± 1.5%, significantly higher than that of the control (10.7% ± 2.2%), RT (23.2% ± 2.1%), and MB49 APM plus RT (12.9% ± 0.8%) (Figures 6G and S6E). The enhancement of DC maturation might be attributed to the synergistic immuno-stimulatory effects of combining sMB49 APM with RT. Moreover, the frequency of CD8⁺ T cells was significantly increased from 2.8% ± 1.1% (control), 9.9% ± 1.8% (RT), 6.1% ± 1.6% (sMB49 APM), and 14.9% ± 1.0% (MB49 APM plus RT) to 27.0% ± 4.8% (sMB49 APM plus RT) (Figures 6H and 6I). Furthermore, analysis of effector CD8⁺ T cells demonstrated that sMB49 APM plus RT remarkably increased the frequency of CD8⁺Granzyme B⁺ T and CD8⁺ IFN-γ⁺ T cells by 1.8- to 7.2-fold and 1.5- to 7.2-fold higher than that of its counterparts, respectively (Figures 6J, 6K, and S6F). IF analysis of CD8 (green) and Foxp3 (red) corroborated increased CD8⁺ T cell infiltration together with reduced regulatory T cells (Tregs) after sMB49 APM plus RT (Figure 6L). In addition, sMB49 APM plus RT increased the T_EM fraction to 66.7% ± 2.6%, representing a 2.1-fold increase over controls (Figures 6M and S6G), implying their potential for long-term tumor inhibition. Collectively, these data indicated that STC-derived nano-APMs potentiated RT efficacy by promoting STC clearance and strengthening antitumor immunity (Figure 6N).

Combination of nano-APM and RT outperforms conventional senolytic therapy in murine PDAC models

Given the potent efficacy observed in the BLCA model, we next tested nano-APM therapy in an immunosuppressive tumor model of PDAC and benchmarked it against a conventional pharmacologic senolytic regimen using RT followed by ABT-263 (navitoclax).⁵ KPC-tumor-bearing mice were randomized into five groups: control (#1), sKPC^KD APM (#2), RT (#3), RT plus ABT-263 (#4), and sKPC^KD APM (#5) plus RT with identical RT and nano-APM dosing (Figure 7A). Tumor volume profiles showed that tumor growth was significantly delayed in sKPC^KD APM, RT, and RT plus ABT-263 during treatment period but subsequently relapsed 1 week post-treatments (Figures 7B and S7A). In contrast, sKPC^KD APM plus RT persistently inhibited tumor growth with an 84.0% tumor inhibition efficiency by the 25^th day, which was 2.7-, 1.7-, and 1.9-fold higher than that of sKPC^KD APM, RT, and RT plus ABT-263 groups, respectively (Figures 7B and S7A). Representative tumor images corroborated the enhanced efficacy of sKPC^KD-APM-augmented RT (Figure 7C). Survival was also markedly extended, increasing from ∼26 days in controls to ∼29 days with sKPC^KD APM and ∼40 days with RT or RT plus ABT-263 and exceeding 50 days with sKPC^KD APM plus RT (Figure 7D).

Figure 7.

Combination of nano-APM and RT demonstrated superior tumor suppression and ITM reprogramming in PDAC tumors

(A) Experimental schedule for treatments of sKPC^KD APM or conventional senolytic ABT-263 in the PDAC model.

(B and C) Averaged tumor growth profiles and images of PDAC model (n = 6).

(D) Survival plots of tumor-bearing C57BL mice (n = 6).

(E) TUNEL staining and analysis of tumor sections after various treatments (scale bars, 50 μm).

(F) IHC examination and analysis of tumor section on p21 expression after various treatments (scale bar, 50 μm).

(G) FCM-determined frequency of mature DCs in the tdLNs (n = 5).

(H and I) Representative FCM plots of the tumor-infiltrating CD8⁺ and CD4⁺ T cells and the frequency of the tumor-infiltrating CD8⁺ T cells (n = 5).

(J and K) The frequency of intratumoral CD8⁺ Granzyme B⁺ and CD8⁺ IFNγ⁺ T cells (n = 5).

(L) The ratio of the frequency of CD8 T cells to that of Tregs (CD8/Treg ratio).

(M) IF staining for CD8 and Foxp3 in tumor sections after treatments (scale bars, 50 μm).

(N) Schematic illustration of the superior advantages of sKPC^KD-APM-enhanced RT over the conventional senolytic therapy.

Data are presented as the means ± SD. Data in (G) and (I)–(L) were assessed via one-way ANOVA with Tukey’s test. The statistical significance of data in (D) was calculated by survival curve comparison with log rank (Mantel-Cox) test.

See also Figures S7 and S8.

H&E and TUNEL staining showed increased necrosis and a higher abundance of TUNEL⁺ cells in tumors treated with sKPC^KD APM plus RT compared with all other groups (Figures 7E and S7C). Importantly, p21⁺ STCs were more effectively reduced by sKPC^KD APM plus RT compared to the conventional senolytic therapy (RT plus ABT-263) (Figure 7F). These results supported that nano-APM-mediated immune senolysis enabled more effective STC elimination than pharmacologic apoptosis-based senolysis.

We next investigated whether nano-APM-based immune senolysis reprogrammed the RT-induced ITM in PDAC. FCM analysis of TdLNs demonstrated a significant promotion of sKPC^KD APM plus RT in driving DC maturation compared with all other groups (Figures 7G and S7D). FCM analysis of intratumoral immune cells showed that the frequency of CD8⁺ T cells was significantly increased to 77.3% ± 4.4%, which was 1.2- to 1.7-fold higher than that of groups of control, sKPC^KD APM, RT, or RT plus ABT-263 (Figures 7H and 7I). Furthermore, analysis of effector CD8⁺ T cells demonstrated that RT plus sKPC^KD APM remarkably increased the frequency of CD8⁺Granzyme B⁺ T and CD8⁺IFN-γ⁺ T cells by 1.7- to 9.2-fold and 1.8- to 8.0-fold higher than that of its counterparts, respectively (Figures 7J, 7K, and S7E).

Conversely, the frequency of intratumoral Tregs was markedly decreased to 8.1% ± 2.0% upon sKPC^KD APM plus RT compared with RT plus ABT-263 (25.1% ± 3.5%) and RT (33.1% ± 1.2%), which was significantly higher than that of control (24.6% ± 1.6%) or sKPC^KD APM (19.8% ± 1.9%) (Figure S7F). Accordingly, the ratio of CD8⁺ T cell to Treg of sKPC^KD APM plus RT was increased by 3.8- to 21.5-fold higher than its counterparts (Figure 7L), indicating a pronounced shift toward adaptive antitumor immunity. IF analysis of CD8 (red) and Foxp3 (green) further validated the increase of CD8⁺ T cells but decrease of Tregs in sKPC^KD APM plus RT group (Figure 7M). In addition, sKPC^KD APM plus RT increased the frequency of splenic T_EM by 37.2% ± 4.8%, which was 3.1-fold higher than that in the control group (Figure S7G), confirming its therapeutic effects of long-term tumor inhibition over conventional senolytic therapy.

To further evaluate immune modulation in a clinically relevant setting, we established an orthotopic PDAC model and randomized mice into control, RT, α-PD-L1 plus RT, and sKPC^KD APM plus RT groups under identical RT dosing (Figure S8A). FCM analysis showed that sKPC^KD APM plus RT reduced p21⁺ STCs more effectively than the other groups (Figures S8B and S8C). Immune profiling further revealed increased tumor-infiltrating CD8⁺ T cells, with concomitant reductions in Tregs after sKPC^KD APM plus RT compared with α-PD-L1 plus RT (Figures S8D–S8G). Moreover, the frequency of natural killer (NK) cells also significantly upregulated upon sKPC^KD APM plus RT (Figures S8H and S8I), supporting coordinated activation of adaptive and innate immunity. This may reflect broader immune-microenvironment reprogramming and potential antigen spreading^37,38,39 following nano-APM-primed STC-reactive immunity.

Biosafety assay with hematology and serum biochemical analysis displayed comparable biochemical parameters between the nano-APM and PBS groups (Figures S8J and S8K). H&E staining of major organs (heart, liver, spleen, lung, and kidney) showed no overt pathological abnormalities (Figure S8L), supporting the biosafety of nano-APMs. Taken together, nano-APM-mediated immune senolysis outperformed pharmacologic senolysis by achieving durable and selective STC clearance and mitigating RT-induced ITM in PDAC, resulting in sustained tumor inhibition (Figure 7N).

Discussion

The contribution of STCs to RT-induced ITM is increasingly recognized, yet current senolytic strategies face critical limitations, including insufficient selectivity and safety concerns. To address these challenges, we proposed a clinically actionable strategy by sequentially combining STC-derived nano-APM and RT. In this study, we identified STCs as unconventional antigen-presenting cells capable of directly priming CD8⁺ T cells and developed STC-membrane-derived nano-APM to safely harness this capacity. When combined with RT, this strategy enabled spatiotemporally confined activation of STC-specific immunity, leading to effective tumor suppression in both BLCA and PDAC models.

Activation of tumor-reactive T cells has been conventionally identified to be dependent on professional APCs, which capture and process tumor antigens.^40,41 Previous studies hinted that senescent cells can modulate immune recognition, for example by engaging CD4⁺ T cells via MHC-II or by evading CD8⁺ T cell killing through nonclassical MHC-I.^42,43 Here, we identified STCs as unconventional antigen-presenting sources capable of directly priming CD8⁺ T cells through upregulated MHC class I and costimulatory molecules CD80 and CD86 (Figures 1 and 2). This finding reshapes the current view of tumor senescence, highlighting that STCs are not only immunosuppressive mediators but can also serve as active initiators of antitumor immunity.

Autologous tumor-cell-derived bioactive formulations present notable advantages in biocompatibility and personalization, yet their direct application is limited by inherent oncogenic risk. Strategies such as genetic engineering of tumor cells,⁴⁴ inactivated whole-cell vaccines,⁴⁵ and tumor cell-derived biomimetic nanoparticles^46,47,48 have attempted to address this issue but often compromise either functional fidelity or safety. In this study, we engineered STCs into cell nano-APMs, which preserved native antigen-presenting features while eliminating the risks associated with live cells and efficiently induced STC-specific T cells in the spleen (Figures 4H and 5D). Importantly, nano-APMs can be readily manufactured from patient-derived tumor specimens obtained during surgery or biopsy, underscoring their translational potential for personalized RT enhancement.

Current senolytic strategies, including CAR-T cells, pro-senescence therapies combined with immune checkpoint blockade,^49,50,51,52 and the conventional “one-two punch” strategy of RT-induced senescence followed by small-molecule senolytics (e.g., ABT-263), are limited by inefficiency, systemic toxicity, or procedural complexity. To address above challenges, we herein proposed a sequential regimen in which nano-APMs first prime STC-specific CD8⁺ T cell responses in the spleen, followed by RT-induced senescence that locally reactivates this pool for targeted elimination of STCs. This priming-reactivation approach achieved spatiotemporally confined senolysis with potent antitumor immunity in both BLCA and PDAC models (Figures 6 and 7), offering a paradigm for developing an effective and safe senolysis strategy and a clinically translatable means to improve RT efficacy.

In summary, we demonstrated that STCs can function as antigen-presenting sources and developed STC-derived nano-APM to directly prime CD8⁺ T cells in the spleen. Sequential combination with RT enabled spatiotemporally confined reactivation of these T cells, resulting in efficient clearance of STCs and robust tumor suppression. Our results offer mechanistic insights into RT-induced senescence and establish nano-APM-based senolysis as a clinically promising approach to enhance RT efficacy.

Limitations of the study

Despite satisfactory therapeutic effects, our study has several limitations. First, the therapeutic efficacy of nano-APM should be further investigated in larger animal models with varying dosages and administration schedules. Second, it remains uncertain whether the safety and antitumor efficacy demonstrated in mice can be replicated in human patients, due to the significant differences in host and cancer genetics between species.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Prof. Yi Lai (laiyi@simm.ac.cn).

Materials availability

All reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• •The sequencing data generated in this study are available in the NCBI Gene Expression Omnibus (GEO) under accession code GSE306978.
• •The published article does not report custom computer code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

Financial support from the New Organizational Model Initiative (2025ZD0552300), National Natural Science Foundation of China (32571707, 82203041, 32411530049, U22A20328, 22474042, and W2412035), Science and Technology Commission of Shanghai Municipality (23ZR1475000), Shanghai Oriental Talents Program (BJKJ2024046), Shanghai Post-doctoral Excellence Program (2025451), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB1060000), and State Key Laboratory of Chemical Biology is appreciated. We thank the staff members of the Integrated Laser Microscopy System (https://cstr.cn/31129.02.NFPS.CLMIS) at the National Facility for Protein Science in Shanghai (https://cstr.cn/31129.02.NFPS) for providing technical support and assistance in data collection and analysis. Additionally, the TEM data were collected at the Cryo-Electron Microscopy Research Center, Shanghai Institute of Materia Medica, Chinese Academy of Sciences. All animal procedures were conducted in accordance with guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of the Shanghai Institute of Materia Medica, CAS (Approval No. 2024-09-YHJ-10).

Author contributions

J.P., S.Z., and Y.X. contributed equally to this study. H.Y., Y.C., Z.X., and Y.L. conceived the project, and J.P, S.Z., X.C., H.Y., and Y.L. designed the study. J.P., S.Z., Y.X., and Y.L. performed the experiments and prepared the initial manuscript. Y.X., B.G.D.G., M.L., Z.P., B.H., X.Q., A.J., and Z.L. helped revise the manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-mouse CD86 antibody (Clone GL-1), APC conjugated	BD Biosciences	Cat# 558703; RRID: AB_2075114
Anti-mouse CD80 antibody (Clone 16-10A1), PE conjugated	Biolegend	Cat# 104708; RRID: AB_313129
Anti-mouse CD80 antibody (Clone 16-10A1), FITC conjugated	Biolegend	Cat# 104706; RRID: AB_313127
Anti-mouse CD11c antibody (Clone N418), FITC conjugated	Biolegend	Cat# 117305; RRID: AB_313774
Anti-mouse CD45 antibody (Clone 30-F11), FITC conjugated	BD Biosciences	Cat# 103108; RRID: AB_312973
Anti-mouse CD45 antibody (Clone 30-F11), PE conjugated	BD Biosciences	Cat# 103106; RRID: AB_312971
Anti-mouse CD25 antibody (Clone PC61.5), APC conjugated	Invitrogen	Cat# 17-0251-81; RRID: AB_469365
Anti-mouse CD3 antibody (Clone 17A2), Percp-Cy5.5 conjugated	Biolegend	Cat# 100218; RRID: AB_1595492
Anti-mouse CD4 antibody (Clone GK1.5), APC-Cy7 conjugated	BD Biosciences	Cat# 552051; RRID: AB_394331
Anti-mouse Foxp3 antibody (Clone FJK-16s), PE conjugated	Invitrogen	Cat# 12-5773-82; RRID: AB_465936
Anti-mouse CD8α antibody (Clone 53-6.7), PE conjugated	BD Biosciences	Cat# MA1-10304; RRID: AB_11156191
Anti-mouse CD8α antibody (Clone 53-6.7), PE/Dazzle 594 conjugated	Biolegend	Cat# 100762; RRID: AB_2564027
Anti-human/mouse Granzyme B antibody (Clone GB11), FITC conjugated	Biolgend	Cat# 515403; RRID: AB_2114575
Anti-mouse IFNγ antibody (Clone XMG1.2), APC conjugated	BD Biosciences	Cat# 554413; RRID: AB_398551
Anti-mouse CD44 antibody (Clone IM7), APC conjugated	BD Biosciences	Cat# 559250; RRID: AB_398661
Anti-mouse CD62L antibody (Clone MEL-14), FITC conjugated	Biolgend	Cat# 104406; RRID: AB_313093
Anti-CD127 antibody (CloneSB199), PE conjugated	BD Biosciences	Cat# 552543; RRID: AB_394417
Anti-mouse TNF-α antibody (Clone MP6-XT22), APC conjugated	BD Biosciences	Cat# 561062; RRID: AB_398553
Anti-mouse H-2Kb antibody (Clone AF6-88.5), PE-Cy7 conjugated	Biolgend	Cat# 116520; RRID: AB_2721684
Anti-mouse H-2Kb antibody bound to SIINFEKL (Clone 25-D1.16), PE conjugated	Biolgend	Cat# 141603; RRID: AB_10897938
Anti-mouse CD69 antibody (Clone H1.2F3), APC conjugated	Biolegend	Cat# 104513; RRID: AB_492844
Anti-mouse/human CD11b antibody (Clone M1/70), Percp-Cy5.5 conjugated	Biolgend	Cat# 101228; RRID: AB_893232
Anti-mouse NK1.1 antibody (CloneS17016D), APC conjugated	Biolegend	Cat# 156513; RRID: AB_2888852
Anti-GAPDH Mouse mAb	Abbkine	Cat# ABL1020; RRID: AB_3714704
Anti-FoxP3 Rabbit mAb	CST	Cat# 12653; RRID: AB_2797979
Anti-CD8 alpha Rabbit pAb	Servicebio	Cat# GB114196; RRID: AB_3064847
Anti-CD11c Rabbit pAb	Servicebio	Cat# GB11059; RRID: AB_2905514
Anti-P21 Rabbit mAb	Abcam	Cat# ab109199; RRID: AB_10861551
Anti-PD-L1 Rabbit mAb	Abcam	Cat# ab213480; RRID: AB_2773715
Anti-MHC class I Rabbit mAb	Abclonal	Cat# A8754; RRID: AB_2863602
Anti-Cytokeratin 19 (CK19) Rabbit mAb	Abcam	Cat# ab52625; RRID: AB_2281020
Anti-CDKN2A/p16INK4a Rabbit pAb	Abcam	Cat# ab189034; RRID: AB_2737282
Anti-TriMethyl-Histone H3-K9 Rabbit mAb	Abclonal	Cat# A22295; RRID: AB_3066640
Dylight 594, Goat Anti-Rabbit IgG	Abbkine	Cat# A23420
Dylight 488, Goat Anti-Rabbit IgG	Abbkine	Cat# A23220; RRID: AB_2737289
HRP, Goat Anti-Rabbit IgG	Abbkine	Cat# A21020; RRID: AB_2876889
HRP, Goat Anti-Mouse IgG	Abbkine	Cat# A21010; RRID: AB_2728771
Chemicals, peptides, and recombinant proteins
Dulbecco’s Modified Eagle Medium	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0212
Roswell Park Memorial Institute (RPMI) 1640 medium	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0215
Penicillin-Streptomycin	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0110
0.25% trypsin-EDTA	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0232
HEPES buffer solution (2×)	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0036
Fetal Bovine Serum	Gibco	Cat# 10099158
Deoxyribonuclease I (DNase I)	YEASEN Biotechnology (Shanghai, China)	Cat# 10608ES25
collagenase IV	YEASEN Biotechnology (Shanghai, China)	Cat# 40510ES60
Hyaluronidase	YEASEN Biotechnology (Shanghai, China)	Cat# 20426ES60
Anti-mouse PD-L1 mAb	BioXCell	Cat# BE0101
mouse IL-2 protein	MedChemexpress (MCE, Shanghai, China)	Cat# HY-P7831
Mouse IFN-γ protein	MedChemexpress (MCE, Shanghai, China)	Cat# HY-P7071
mAb anti-mouse CD3	BioXCell	Cat# BE0001-1
mAb anti-mouse CD28	BioXCell	Cat# BE0015-1
4′,6-Diamidino-2-phenylindole dihydrochloride (DAPI),	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0128
pre-stained rainbow protein marker (10–190 kDa)	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0342
Bicinchoninic Acid (BCA) protein quantification kit	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0082
TBST buffer solution (10×)	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0091
ECL luminescence reagent	Meilun Biotech Co., Ltd (Dalian, China)	Cat# MA0186
2-(7-Azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HATU)	J & K Scientific Ltd (Beijing, China)	Cat# F023926
N, N-dimethylformamide (DMF)	J & K Scientific Ltd (Beijing, China)	Cat# 937230
Cy7	Bide Pharmatech Co., Ltd.	Cat# BD13808263
DBCO-acid	Bide Pharmatech Co., Ltd.	Cat# BD00805279
Critical commercial assays
EasySep Mouse CD8a Positive Selection Kit II	STEMCELL Technologies	Cat# 18953
Deposited data
Raw sequencing data: RNA-seq	This paper	GEO: GSE306978
Experimental models: Cell lines
KPC	Gift from Lab of Dr. Y. Huang	N/A
MB49	Gift from Lab of Dr. Longcheng Li	N/A
Experimental models: Organisms/strains
Mouse:C57BL/6J	Shanghai Experimental Animal Center (Shanghai, China)	N/A
Software and algorithms
GraphPad Prism software	GraphPad Software, Inc	https://graphpad.com/scientificsoftware/prism/
FlowJo-10.7	Tree Star Inc.	https://www.flowjo.com/solutions/flowjo
RStudio Server	RStudio, PBC	https://www.rstudio.com/products/
R-4.1.2	R-project	https://www.r-project.org/

Experimental model and study participant details

Patient-derived tumor samples

Pancreatic and bladder cancer tissues were obtained from Xinhua Hospital (No. XHEC-D-2025-026, XHEC-STCSM-2021026) under guidelines approved by the Review Board of the Medical Ethics Committee of Xinhua Hospital, Shanghai Jiao Tong University School of Medicine.

Animal

Female C57 BL/6 (6–8 weeks, 18–20 g) were obtained from Shanghai Experimental Animal Center (Shanghai, China). All mice were kept under the pathogen-free condition and used following the guidelines approved by the Institute of Animal Care and Use Committee, Shanghai Institute of Materia Medica, Chinese Academy of Sciences.

Cell lines

The MB49 cell (mouse bladder cancer cell line) obtained from Dr. Longcheng Li from Peking Union Medical College Hospital of Chinese Academy of Medical Sciences. KPC pancreatic tumor cells were kindly provided by Prof. Y. Huang from Shanghai Institute of Materia Medica of Chinese Academy of Sciences. KPC cells were obtained from the primary KPC tumors in Pdx-Cre; Kras^LSL−G12D/+; Trp53^R172H/+ mice. An ovalbumin (OVA) expressing MB49 (MB49^OVA) cell line and KPC (KPC^OVA) cell line were generated as described.⁴⁹ OT-1 T lymphocytes were kindly provided by Dr. L. Gong from SIMM of CAS, China. All tumor cells were maintained in DMEM medium with 10% FBS (V/V), penicillin G sodium and streptomycin sulfate (100 U/mL). All lymphocytes were maintained in RPMI 1640 with 10% FBS (V/V), penicillin G sodium and streptomycin sulfate (100 U/mL), 1% Non-Essential Amino Acids Solution (NEAMM), 0.1% β-Mercaptoethanol (100×), 1% 1M HEPES, and 1% Sodium Pyruvate(100mM). All cells were incubated at 37°C with 5% CO2 supply. All cellular experiments in vitro were conducted during the logarithmic phase of cell growth in vitro.

Method details

Western blot assay

The cells were collected and lysed with RIPA lysis buffer containing 1% protease inhibitor. Equal amounts of protein were loaded and separated using SDS-PAGE gels, and then transferred to PVDF membranes. Blocking with 5% BSA in TBST for 1h, membranes were conjugated to primary antibodies overnight at 4°C. After incubation with HRP-conjugated secondary antibodies for 1h at room temperature, signal visualization was carried out using gel imager (Tanon Image, Shanghai) and analyzed by ImageJ (NIH, USA).

Bioinformatics analysis

To investigate the clinical relevance of tumor cell senescence, we analyzed publicly available transcriptomic and clinical data from The Cancer Genome Atlas (TCGA). A validated senescence gene signature (seno-sig), encompassing key markers of cellular senescence (e.g., CDKN1A, CDKN2A, LMNB1, SASP components), was utilized. This signature score was calculated for each tumor sample using the single-sample Gene Set Enrichment Analysis (ssGSEA) algorithm as implemented in the R package GSVA. Patients within each cancer type were then stratified into ‘High seno-sig’ and ‘Low seno-sig’ cohorts based on the median value of the senescence signature score. Overall survival (OS) analysis between the high and low seno-sig groups was performed using the Kaplan-Meier method. The statistical significance of differences in survival curves was assessed using the two-sided log rank test. To identify biological pathways differentially activated in tumors with a high senescence burden, we performed Gene Set Enrichment Analysis (GSEA) using the GSEA software (version 4.x.x) from the Broad Institute. Pre-ranked gene lists were generated based on the signal-to-noise ratio comparing the high seno-sig versus low seno-sig cohorts. The analysis was run against the KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway database. Differential expression of specific antigen presentation machinery (APM) genes (e.g., HLA-A, HLA-B, HLA-C, CD80, CD86) between the high and low seno-sig cohorts was analyzed using Wilcox.test. All bioinformatics analyses were conducted using R software (version 4.x.x) and relevant Bioconductor packages.

Induction and identification of senescent tumor cells

Tumor cells were pre-seeded in 6-well plate (2.0 × 10⁶/well) and radiated by 8-Gy X-ray, followed by another 48 h culture with or without IFN-γ treatments. The senescent phenotype of sMB49 or sKPC was confirmed by the high expression of p21 by WB and the β-Gal by β-Gal stain assay.

Fabrication of the senescent or non-senescent tumor cell-derived nano-APM

The cell membrane from senescent or non-senescent tumor cell was collected according to the protocol of Membrane Extraction Kit. Briefly, cells were resuspended in lysis buffer A and subjected to two complete freeze-thaw cycles (liquid nitrogen for 2 min followed by room temperature for 5 min) to achieve cellular lysis. Lysis efficiency was verified by microscopic examination, with additional cycles performed if < 70% cell disruption was observed. The cellular membrane fraction was isolated through differential centrifugation. First, the lysate was centrifuged at 700 × g for 10 min at 4°C to pellet nuclei and residual intact cells (P1). The supernatant was then subjected to high-speed centrifugation at 14,000 × g for 20 min at 4°C to obtain a crude membrane pellet (P2). For Nano-APM preparation, the membrane pellet was resuspended in nuclease-free ddH₂O at a 1:10 (w/v) membrane-to-water ratio, followed by bath sonication (2 min, 40 kHz, 100 W) using a Fisher Scientific FS30D sonicator (thermostated at 4°C).

RNA sequencing analysis

Senescent or non-senescent tumor cells (MB49 cellS, sMB49 cells, KPC cells, and sKPC cells) were treated with 8 Gy-RT and cultured for 48 h. Then the cells were collected for total RNA extraction using TRIzol Reagent. Genomic DNA was removed by DNase I, and the 1% agarose gels were used to monitor the RNA degradation and contamination. Next, RNA quality was identified by 2100 Bioanalyser (Agilent Technologies) and quantified by ND-2000 (NanoDrop Technologies). RNA sample with high-quality was utilized to establish sequencing library (OD260/280 = 1.8–2.2, OD260/230 ≥ 2.0, RIN ≥8.0, 28 S:18S ≥ 1.0 and ≥1 μg).

RNA purification, reverse transcription, library establishment and sequencing were completed at Shanghai Major Bio-pharm Biotechnology Co., Ltd. (Shanghai, China), according to the manufacturer’s instructions (Illumina, San Diego, CA). The RNA-seq transcriptome library was prepared following Illumina Stranded mRNA Prep, Ligation from Illumina (San Diego, CA) using 1μg of total RNA. Shortly, messenger RNA was isolated according to polyA selection method by oligo(dT) beads and then fragmented by fragmentation buffer firstly. Secondly double-stranded cDNA was synthesized using a SuperScript double-stranded cDNA synthesis kit (Invitrogen, CA) with random hexamer primers (Illumina). Then the synthesized cDNA was subjected to end-repair, phosphorylation and ‘A’ base addition according to Illumina’s library construction protocol. Libraries were size selected for cDNA target fragments of 300 bp on 2% Low Range Ultra Agarose followed by PCR amplified using Phusion DNA polymerase (NEB) for 15 PCR cycles. After quantified by Qubit 4.0, paired-end RNA-seq sequencing library was sequenced with the NovaSeq Xplus sequencer (Illumina, San Diego, CA, USA). To identify DEGs (differential expression genes) between two different samples, the expression level of each transcript was calculated according to the transcripts per million reads (TPM) method. RSEM was used to quantify gene abundances. Essentially, differential expression analysis was performed using the DESeq2or DEGseq. DEGs with |log2FC| ≥ 1 and FDR≤ 0.05 (DESeq2) or FDR ≤0.001 (DEGseq) were significantly differentially expressed genes. In addition, functional-enrichment analyses including GO and KEGG were performed to identify which DEGs were significantly enriched in GO terms and metabolic pathways at Bonferroni-corrected p-value ≤0.05 compared with the whole-transcriptome background. GO functional enrichment and KEGG pathway analysis were carried out by Goatools and KOBAS, respectively.

Extraction, purification, and culture of primary splenic CD8⁺ T lymphocytes

The tumor-bearing mice were killed by CO₂ asphyxiation and the spleen were removed. The spleen was dissociated by gentle MACS Tissue Dissociator and run through a 70-μm cell strainer to obtain a single-cell suspension. CD8⁺ T cells were isolated by magnetic-activated cell sorting (MACS) using a commercial CD8⁺ T cell Isolation Kit (EasySep Mouse CD8a Positive Slctn Kit II, Stem Cell), according to the manufacturer’s instructions. Splenic CD8⁺ T (1.0 x10⁶/well) were seeded into 6-well plate and stimulated with immobilized antiCD3 mAb (1 μg/mL), anti-CD28 mAb (2 μg/mL), co-cultured with senescent or none-senescent tumor cells, or these cell-derived nano-APMs for 48 h–72h.

Evaluation of activation of CD8⁺ T lymphocytes in vitro

Splenic CD8⁺ T cells (1.0 x10⁶/well) or OT-1 T cells (3.0 x10⁵/well) were seeded into 6-well plate and stimulated by nano-APM (20 μg/ml) for 48h, or cocultured with senescent or none-senescent tumor cells. Then, these cells were harvested for FCM analysis of the expression of CD69-APC, TNF-PE, Granzyme B-FITC, or IFNγ-APC.

Establishment of PD-L1 knockdown KPC cell line

For PD-L1 knockdown, three independent shRNA sequences targeting mouse Cd274 (shPD-L1 #1-#3) and a scrambled control (shCtrl) were cloned into the pLKO.1 vector (Addgene #10878). Lentiviral particles were produced in HEK293T cells using psPAX2 and pMD2.G packaging plasmids (Addgene #12260 and #12259). KPC cells were infected at an MOI of 5 with 8 μg/mL polybrene, followed by puromycin selection (4 μg/mL, 7 days).

Synthesis of DBCO-Conjugated Cyanine dye (DBCO-Cy7)

To synthesize DBCO-conjugated Cy7 (DBCO-Cy7), compound Cy7 (100 mg, 0.2 mmol) was first dissolved in 5–10 mL of anhydrous N, N-dimethylformamide (DMF), followed by the addition of piperazine (86 mg, 1 mmol, 5 equivalents). The mixture was stirred at room temperature overnight. After completion, DMF was removed under reduced pressure, and the residue was washed with anhydrous diethyl ether and filtered to afford the intermediate NH-Cy7, which was used directly in the next step without further purification.

In the second step, dibenzocyclooctyne carboxylic acid (DBCO-COOH, 34 mg, 0.112 mmol, 1.2 equivalents) and 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU, 42 mg, 0.112 mmol, 1.2 equivalents) were dissolved in DMF, and N,N-diisopropylethylamine (DIEA, 49 μL, 0.279 mmol, 3 equivalents) was added at 0°C for pre-activation over 5 min. NH-Cy7 (50 mg, 0.093 mmol) was then added, and the reaction mixture was stirred at room temperature for 24 h. The reaction progress was monitored by thin-layer chromatography. After completion, the solvent was evaporated under reduced pressure, and the crude product was extracted using saturated ammonium chloride solution and ethyl acetate. The organic layer was dried over anhydrous sodium sulfate and purified by silica gel column chromatography using dichloromethane/methanol (10:1, v/v) to yield the final DBCO-Cy7 product with an isolated yield of 85%. ¹H NMR (400 MHz, Chloroform-d) δ 7.70–7.67 (m, 2H), 7.40–7.28 (m, 13H), 7.16 (d, J = 7.5 Hz, 1H), 7.01 (s, 1H), 5.82 (d, J = 10.7 Hz, 2H), 5.18 (t, J = 13.8 Hz, 2H), 4.23–4.13 (m, 1H), 4.08 (t, J = 6.7 Hz, 1H), 3.72–3.66 (m, 4H), 3.64–3.57 (m, 4H), 3.50 (s, 3H), 2.88–2.79 (m, 2H), 2.53–2.37 (m, 6H), 1.82 (p, J = 6.4 Hz, 3H), 1.70–1.64 (m, 4H), 1.39–1.35 (m, 2H), 1.27 (d, J = 12.0 Hz, 6H), 0.93 (q, J = 7.4, 6.7 Hz, 3H). HRMS (ESI): m/z [M]⁺ calcd for C55H58N5O2, 820.4585; found, 820.4584.

Establishment of Cy7-linked KPC or senescent KPC cells

Tumor cells in vitro were cultured in six-well plates at a density of 150,000 cells per well incubating with 25 μM Ac₄ManAz in cell culture medium for 3 days. After washing with PBS, the cells were collected via centrifuge and incubated with 10 μM DBCO-Cy7 for 1.0 h at 4°C. Then these tumor cells were subjected to senescence induction or nano-APM fabrication. The successful link of Cy7 was confirmed by flow cytometry and CLSM analysis.

Biodistribution of Cy7-linked nano-APM in vivo

The subcutaneous mouse KPC cell model of PDAC was established by subcutaneously injection of 5.0 × 10⁶ KPC tumor cells on the right flank of mice. Tumor-bearing mice were administered Cy7-linked nano-APM via slow intravenous injection at an identical dosage of 5 mg/kg. Biodistribution of Cy7-linked nano-APM was determined by IVIS Lumina II In Vivo Imaging System (Caliper Life Sciences) (Ex/Em = 670 nm/720 nm). Major organs (the heart, liver, spleen, lungs, and kidneys) and tumors were collected and likewise imaged ex vivo at 4 h post-administration. Accumulation of Cy7-linked nano-APM within major organs and tumors was assessed by the relative Avg Radiant Efficiency.

Study of tumor senescence in patient-derived tumor tissues and murine tumor tissues

Formalin-fixed, paraffin-embedded (FFPE) tumor sections derived from tumor patients or murine tumor model were deparaffinized and rehydrated. Antigen retrieval was performed using citrate buffer (pH 6.0) in a microwave. Endogenous peroxidase activity was blocked with 3% H₂O₂. Sections were then incubated with a primary antibody against p21 overnight at 4°C, followed by incubation with an HRP-conjugated secondary antibody. Signal was developed using a DAB substrate kit, and sections were counterstained with hematoxylin. p21-positive cells were quantified by counting positive nuclei in 3 random high-power fields per sample. To precisely identify STCs within the tissue architecture, multiplex immunofluorescence (mIF) was performed on FFPE sections according to a commercial kit (AFIHC034). This highly sensitive iterative staining protocol allowed for sequential labeling with primary antibodies against key markers, including the tumor cell marker Cytokeratin 19 (CK19), senescence markers (p21, p16), the heterochromatin marker H3K9me3, and MHC-I.

Briefly, for each staining cycle, sections were incubated with a primary antibody, followed by an HRP-conjugated secondary antibody. The signal was then amplified using a fluorophore-conjugated tyramide reagent. After each round, the HRP-conjugated secondary antibody was inactivated, and the primary-secondary antibody complex was gently stripped to prepare the tissue for the next cycle of staining. Nuclei were counterstained with DAPI in the final step. High-resolution images were acquired using a fluorescence microscope or a confocal laser scanning microscope.

Study of splenic immunity activation in vivo

To evaluate the splenic immunity activation in vivo, C57 mice were intravenously injected with sKPC^KD APM (500 μg/kg) at predetermined time points (3, 5, 7 days). Then, spleens were harvested for H&E, mIF, or FCM analysis. FCM analysis was performed to evaluate the DC maturation (CD11c⁺CD80⁺CD86⁺), CD8⁺ T cell proliferation and activation (CD69⁺), and the frequency of memory T cells. mIF was performed to detect the proliferation of B cells (CD20), T cells (CD3), and endothelia cells (CD31).

The establishment of the subcutaneous tumor model and antitumor study in vivo

The subcutaneous BLCA and PDAC tumor models were established by subcutaneously injection of 5.0 × 10ˆ⁵ MB49 cells and 5.0 × 10ˆ⁶ KPC tumor cells on the right flank of mice, respectively. In the antitumor study of nano-APMs in BLCA model, when the tumor volume reached about 100 mm³, the mice were randomly divided into the following groups (n = 6): Control (G1), RT(G2), sMB49 APM (G3), MB49 APM+RT (G4), and sMB49 APM+RT (G5). KPC tumor-bearing mice were randomly divided into the following groups (n = 6): Control (#1), sKPC^KD APM (#2), RT (#3), RT+ABT-263 (#4), sKPC^KD APM+RT (#5). These mice were intravenously injected with nano-APM at an equal dose of nano-APM 500 μg/kg and received 8 Gy RT 7-day post nano-APM injection. 2 mg/kg free ABT-263 was intravenously injected into tumor-bearing mice at 5-day post RT for 3 times. The tumor size and body weight were recorded every other day during the observation period. The tumor volume was determined following the formula: Tumor volume = length × width × width/2 (length, the longest dimension; width, the shortest dimension).

The establishment of the orthotopic tumor model of PDAC in vivo

For the establishment of orthotopic PDAC tumor model, KPC cells were dissociated from culture using 0.25% trypsin (Gibco), washed and resuspended in fresh culture medium. A total of 1 × 10⁵ KPC cells were orthotopically implanted into the pancreas. Mice were anesthetized under continuous isoflurane. The left abdominal side was shaved to expose the incision site, which was sterilized using povidone–iodine followed by an alcohol wipe. A small incision in the abdominal skin was made using scissors to expose the underlying muscle. A secondary small incision was made in the muscle layer using sterile scissors, and the pancreas was gently exposed from the abdominal cavity using sterile ring forceps. The KPC cells were injected using a 0.3 cc insulin syringe with a 31-gauge needle into the tail of the pancreas. The pancreas was returned into the abdominal cavity and the incision site was closed using polysorb sutures for the inner muscle layer and the outer skin. All surgical tools were sterilized between animals using 70% isopropyl alcohol and a heated bead sterilizer. For growth monitoring of PDAC, mice were subjected to abdominal ultrasound at day 9∼11 after implantation using a Vevo 3100 ultrasound machine (Visualsonics).

Evaluation of antitumor immunity in vivo

To investigate the induction of the protective antitumor immunity, the mice were randomly divided into the groups same as those in antitumor performance studies (n = 5). Tumors, TdLNs, and spleens were harvested at 7-day post treatment. Briefly, these tissues were cut into small pieces after weighing, then suspended in RPMI1640 with indicated digestive enzymes and dissociated by the gentle MASC dissociator (Miltenyi, German). The single cell suspension was obtained by filtering through 70 μm filters. For detection of lymphocyte subsets, e.g., intratumoral IFN-γ⁺ CD8⁺ T cells, the single cell suspension was pretreated with permeabilization, and stained with anti-CD45-PE, anti-CD3-Percp-Cy5.5, anti-CD8-PE/Dazzle-594, anti-CD4-APC-Cy7, anti-Granzyme B-FITC, and IFN-γ-APC antibodies for FCM analysis. All samples were incubated with specific antibodies for 30 min at room temperature in the dark, then measured by FACS Fortessa flow cytometry (Bioscience) and analyzed with FlowJoV10 software.

Quantification and statistical analysis

Quantitative data were presented as mean ± SD, and analyzed using either a two-tailed Student’s t test or one-way ANOVA followed by Tukey’s post hoc test, as appropriate. p < 0.05 was accepted as a statistically significant difference. The survival study was analyzed through log rank (Mantel-Cox) test. No statistical method was used to predetermine sample size. No data were excluded from the analyses. The investigators were not blinded to allocation during experiments and outcome assessment.

Published: March 30, 2026