The murine prostate cancer RM-1 cell line and 293T were obtained from the American Type Culture Collection (ATCC). RM-1 cells and 293T were cultured in RPMI 1640 (Gibco, C11875500BT) and DMEM (Gibco, 10313021), respectively, supplemented with 10% fetal bovine serum (Gibco, A5669701) and 1× penicillin-streptomycin (BBI, E607011-0100) at 37°C under 5% CO₂ in a humidified incubator.

Male wildtype C57BL/6J mice (WT, Stock No. 000664) and male uPA^−/− mice (Stock no. T027315) were purchased from GemPharmatech Co., Ltd. (Nanjing, Jiangsu, China). All mice (age 6–8 weeks, weight within the range of 20–25 g, 5–10 mice per group) used in this study were housed under pathogen-free conditions, with five mice per cage, and kept in the Laboratory Animal Center of Southern University of Science and Technology.

The sample size for each experiment was determined using G*Power (version 3.1.9.7) based on preliminary data and expected effect sizes. We assumed an effect size of 0.8 (max effect), a statistical power of 90%, and a significance level (α) of 0.05 for all experiments. The calculated sample size was adjusted for potential dropouts (10%) to ensure robust statistical analysis.

For all subcutaneous transplantation models used in this study, RM-1 cells (5 ×  10⁵ cells/0.1 mL of PBS) were injected subcutaneously into the back of the WT or uPA^–/– mice. To deplete CD8⁺ T cells, WT and uPA^–/– mice were randomly divided into the Anti-CD8α (WT Anti-CD8α and uPA^–/– anti-CD8α, 6 mice per group) groups and the isotype (WT Isotype and uPA^–/– Isotype) groups (6 mice per group) using a computer-based random number generator (Microsoft Excel, RAND function). The anti-CD8α group underwent CD8⁺ T cells depletion via intraperitoneal injections of anti-CD8α antibody (100 μg/mice, Selleck, A2102), while the isotype (WT Isotype and uPA^–/– Isotype) groups received isotype antibody (100 μg/mice, Selleck, A2116) every two days for a total of three times. For combined therapy, Thirty-two WT mice were randomly divided into four groups (8 mice per group) using a computer-based random number generator: Control (100 μl of DMSO), uPA inhibitor (4 mg/kg UK122; MedChemExpress, HY-111056), anti-PD-1(100 μg/mice; STARTER, S0B0594), combined UK122 (4 mg/kg), and anti-PD-1(100 μg/mice); intraperitoneal injection was performed every day for a total of seven times.

In the experiments described above, the tumor volume (mm³) was measured every 2 days from Day 2 and continued for 10–12 days. The tumor volume was calculated using the formula: volume = (length × width²)/2. When the experimental endpoint was reached, or if the tumor volume reached 1,500 mm³ or body weight loss exceeded 15% during the experiment, euthanasia was performed on the mice via cervical dislocation under anesthesia (using 1.25% avertin, 0.2 mL/10 g, intraperitoneally), with death confirmed by cessation of heartbeat. Subcutaneous tumors were then collected for further analysis, such as flow cytometry or immunohistochemistry.

All animal experimental procedures were approved by the Animal Experimentation Ethics Committee of Southern University of Science and Technology (No. SUSTech-JY202408015) and adhered to ARRIVE guidelines.

Total RNA was extracted from the tumor tissues of WT and uPA^–/– mice, and all RNA-seq samples were quality-controlled, library preparation, and analysis by Genedenovo Biotechnology Co., Ltd. (Guangzhou, China). R version 4.4.0 was used for statistical analysis. A differential gene expression (DEGs) analysis was performed using the DESeq2 R package, with genes with a P-value ≤ 0.05 and a log2 fold change of +1 or −1 considered significantly upregulated and downregulated, respectively. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of the upregulated and downregulated genes were conducted using the ClusterProfiler R package; pathways with an FDR ≤ 0.05 were considered significant.

Immunohistochemistry (IHC) was performed using subcutaneous tumors after 12 days of inoculation. Tumor tissues were fixed in formalin and embedded in paraffin. Immunohistochemical (IHC) analyses of CD8α, GzmB, anti-TNF-α, and IFN-γ were performed on tumors from WT and uPA^−/− mice, and the expression of uPA and CD8α was also analyzed in a human prostate cancer tissue microarray. The IHC results were visualized using CaseViewer 2.3, to observe. Six to eight different areas were selected from each sample, and the proportion of positive cells was calculated using the ImageJ software. A multiplex immunohistochemistry (mIHC) for PD-1⁺CD8⁺ T cells in tissue sections was performed using a TSA ^® Plus fluorescein detection kit (Servicebio, G1256), according to the manufacturer’s instructions. The primary antibodies were listed in Supplementary Table 1 .

To produce lentiviruses, 293T cells were plated in a 10-cm dish. The next day, the lipofectamine 3000 (Thermo Fisher Scientific, L3000015) was mixed with OPTI-MEM (Gibco, 31985070). A plasmid mix included 10 μg of the vector with 5 μg of both psPAX2 and pMD2G, incubated for 15 min at room temperature before gently adding the mixture to the 293T dish and swirling. At 48–72 h post-transfection, the RM-1 cells were cultured with the viral-containing supernatant for 3 days, followed by selection with puromycin for 7 days. Finally, the expression efficiency of human CD19 (hCD19) in RM1-CD19 cells was verified using flow cytometry.

The anti-CD19-specific FMC63 chimeric antigen receptor (CAR) consists of a Flag-tagged FMC63 scFv (26), CD28, and CD3ζ intracellular domain (27). The anti-CD19 CAR sequence was cloned into the MSCV retroviral vector, followed by MSCV-CD19 CAR and MSCV (empty construct without the CD19 CAR coding sequence).

To produce the retrovirus, 20 µL of Mirius TransIT^®-LT1 (Mirus, MIR 2300) was mixed with 1980 µL of pure OPTI-MEM (without FBS, P/S, or Glutamax, Gibco, 31985070) for each 10-cm dish of 293T. The plasmid mix was prepared per dish, including 2.5 μ g of pCL-Eco and 7.5 µg of retroviral vector (MSCV-CD19 CAR). The liposome solution (20 µL) was mixed and added to the plasmid mix and incubated for 15 min at room temperature before gently adding the mixture to the 293T dish and swirling. At 48 and 72 h transfections, the supernatant was collected and filtered.

To obtain activated CD8⁺ T cells, the 24-well plate was coated with anti-CD3 (Thermo Fisher, 16-0032-81) and anti-CD28 (Thermo Fisher, 16-0281-81) at a concentration of 1 µg/mL in PBS overnight at 4°C or for 2–4 h at 37°C with RPMI 1640 (10% FBS, Glutamax (Gibco, 35050061), 1× penicillin-streptomycin, and 100 µM β-mercaptoethanol (Bestbio, BB-92003)). Next, splenic CD8⁺ T cells was isolated from WT and uPA^–/– mice, which were resuspended at 0.5 × 10⁶ cells/mL; 2 mL was added to each well at 37°C or for 48 h.

Subsequently, a 24-well plate was coated with retronectin (Takara, T100AC; 7 µL in 1 mL of PBS) and incubated overnight at 4°C. Then, the retronectin was removed, 1 mL of the virus supernatant of MSCV-CD19 CAR was added, and the mix was centrifuged in 2000 × g for 1.5 h. The 0.5 × 10⁶ cells/ml of activated CD8⁺ T cells were cultured at a 24-well plate binding virus and with a medium containing IL-7 (10 ng/mL; Novoprotein, CC73) and IL-15 (20 ng/mL; Novoprotein, GMP-C016) and used within 3–4 days. Finally, the expression efficiency of anti-CD19 CAR in CD8⁺ T cells was verified using flow cytometry.

The specific killing activity of CD8⁺CAR T cells against the target RM1-CD19 cells was measured using the CytoTox 96^® Nonradioactive Cytotoxicity Assay (Promega, G1780). This assay enables the quantification of cell lysis by measuring the release of lactate dehydrogenase (LDH) from damaged cells into the supernatant. Therefore, CD8⁺ CAR-T cells were cocultured with targeT cells (RM1-CD19) at various ratios for 24 hours. Following this incubation, the supernatants were collected, and the total cells were lysed to evaluate total LDH release. Cytotoxicity was calculated according to the manufacturer’s protocol: % Cytotoxicity = 100× (Experiment–Effector Spontaneous–Target Spontaneous)/(Target Maximum–Target Spontaneous) (28–30).

Single-cell suspensions were isolated from the RM-1 subcutaneous tumor. The cells were washed twice with PBS and treated with LIVE/DEAD Fixable Aqua (BioLegend, 423101). After another wash with PBS, the cells were stained with Fc-blocking antibodies against CD16/32. Then the cells were stained with surface antibodies in the dark at 4°C for 30 min. For intracellular factor stimulation, cells were cultured with a cell-stimulation cocktail containing protein-transport inhibitors (eBioscience, 00-4975-03) for 4 h at 37°C. Next, the cells were harvested, fixed, and permeabilized with 100 μL of 1× fixation buffer and permeabilization buffer (eBioscience, 88-8824). Then, the cells were stained with intracellular antibodies in the dark at 4°C for 30 min. Cell suspensions were analyzed on a BD FACSAria instrument (BD, FACSAria SORP). Data were analyzed offline using FlowJo (v. 10.8.1). The antibodies used in these experiments are listed in Supplementary Table 2 .

The cytokines in the serum from WT and uPA^–/– tumor-bearing mice were measured by enzyme-linked immunosorbent assay (ELISA) kits of IFN-γ (Animaluni, LV30253M), GzmB (Animaluni, LV30229), and TNF-α (Animaluni, LV30536M), following the manufacturer’s protocol.

Tumor tissues were harvested from the subcutaneous RM-1 cell tumor model and then prepared into a single-cell suspension as previously described. The protocol used for CyTOF sample preparation and acquisition was based on a previous report (31). We utilized 16 channels for the analysis, specifically targeting various cell types and functions. The antibodies and reagents used in these experiments are listed in Supplementary Tables 3 , 4 , respectively.

The human prostate cancer tissue microarray used in this study was obtained from Servicebio Technology Co., Ltd. (Wuhan, China), comprising 32 paired peritumor-tumor tissues and 27 tumor-only specimens. Following rigorous quality control, which excluded samples with tissue fragmentation exceeding 50% or epithelial content loss greater than 30%, a total of 35 eligible samples (19 matched pairs and 16 tumor-only specimens) were retained for immunohistochemical analysis of uPA and CD8α.

The Medical Ethics Committee of Southern University of Science and Technology formally determined that this study did not require ethical review (Approval No. 20230132), as it utilized pre-existing, fully anonymized archival specimens with no access to protected health information, in compliance with the Declaration of Helsinki.

The bioinformatics analysis process performed in the study was realized through R software (4.4.0). Data are presented as the mean ± SEM and were analyzed using GraphPad Prism, version 10.1.2. A two-tailed unpaired Student’s t-test was determined to assess statistical significance; one-way ANOVA followed by the Tukey multiple-comparison test was performed to compare two groups. A two-way ANOVA with Dunnett’s multiple-comparison test evaluated the treatment effects on tumor growth. Normality was evaluated using the Shapiro-Wilk test, and homogeneity of variances was assessed using Levene’s test. If the data did not meet the assumption of normality, non-parametric tests such as the Mann-Whitney U test or Kruskal-Wallis test were applied instead. The cumulative survival time was estimated using the Kaplan-Meier method, and the log-rank test was utilized to compare the different groups. Correlation analysis was conducted using the Pearson method. Significance was set at P <0.05 (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001).

To evaluate the role of uPA in human prostate cancer, we performed IHC to assess the expression of uPA in both prostate cancer and peritumoral tissues. The results showed that uPA was significantly overexpressed in prostate cancer compared to peritumoral tissues ( Figures 1A, B , Supplementary Figure 1A ), suggesting that uPA may be an effective therapeutic target for prostate cancer. Additionally, the infiltration of CD8⁺ T cells were decreased in prostate cancer compared to peritumoral tissues ( Figures 1C, D , Supplementary Figure 1B ), and the expression of uPA was negatively correlated with CD8⁺ T cells infiltration (R = -0.38, P = 0.026; Figures 1E, F ), indicating that highly expression of uPA was associated with poorer CD8⁺ T cells infiltration in patients with prostate cancer.

To examine the functional role of uPA in prostate cancer development, we employed both genetic and pharmacological approaches. Using a uPA^–/– mouse model or administering the uPA inhibitor UK122 to WT mice, we found uPA deficiency significantly inhibited prostate cancer progression. Both models showed reduced tumor growth (WT: 492.64 ± 143.16 mm³; uPA^–/–: 82.19 ± 59.01 mm³; UK122: 169.13 ± 104.96 mm³; Figures 2A-C ) and prolonged survival (p<0.05; Figure 2D ), with the genetic knockout exhibiting superior efficacy to pharmacological inhibition, collectively establishing that uPA deficiency inhibited prostate cancer progression.

To characterize the alterations in the tumor microenvironment induced by uPA deficiency, RNA sequencing was used to identify 3,229 differentially expressed genes (DEGs) between the groups (2,384 upregulated and 845 downregulated in uPA^–/– mice; Figure 3A ). KEGG pathway analysis revealed significant enrichment of immune-related pathways, including T cells receptor signaling, Th1/Th2/Th17 differentiation, chemokine signaling, and PD-1/PD-L1 checkpoint regulation ( Figure 3B ). GO term analysis further revealed the top-ranked biological processes, cellular components, and molecular functions associated with the DEGs ( Figure 3C ). Immunohistochemistry (IHC) confirmed elevated expression of the cytotoxic effector cytokines (GzmB, IFN-γ, TNF-α) in tumor tissues of uPA^–/– mice ( Figures 3D, E ), whereas ELISA confirmed increased serum levels of these cytokines in uPA^–/– tumor-bearing mice ( Figure 3F ). Together, these results suggested that uPA deficiency restrained prostate cancer progression via activating antitumor immunity.

Immune cell composition analysis using the CIBERSORT algorithm (34) demonstrated a marked increase in CD8⁺ T cells and resting dendritic cells, coupled with a significant reduction in myeloid populations (including monocytes and M2 macrophages), within the tumor tissues of uPA^–/– mice compared to WT mice ( Figure 4A ). CyTOF analysis of immune cell surface markers ( Supplementary Figure 1A ) further corroborated that CD8⁺ T cells were robustly elevated in tumors from uPA^–/– mice, in contrast with downregulated CD4⁺ T cells and myeloid subsets ( Figures 4B, C ). Flow cytometry and IHC further validated the accumulation of CD8⁺ T cells in tumors from uPA^–/– mice compared with WT mice ( Figure 4D ; Supplementary Figures 1B, C ). These findings collectively suggested that the antitumor effects of uPA deficiency may be mediated through enhanced CD8⁺ T cells infiltration and functional activation.

Cytotoxic effector cytokines in intratumoral CD8⁺ T cells were detected by flow cytometry. Compared to WT group, intratumoral CD8⁺ T cells isolated from uPA^–/– mice exhibited significantly elevated secretion of GzmB, IFN-γ, TNF-α ( Figures 4E, F ), indicating that an enrichment of effector CD8⁺ T cells, a subset distinguished by acquiring an optimized effector differentiation profile, conferring potent antitumor cytotoxic activity (32). However, splenic uPA^–/– CD8⁺ T cells selectively upregulated IFN-γ production but showed no concurrent increase in GzmB or TNF-α levels ( Supplementary Figure 2 ), suggesting that CD8⁺ T cells cytotoxicity in this context was tumor microenvironment-dependent. To evaluate the antitumor efficacy of uPA^–/– CD8⁺ T cells, we generated WT/uPA^–/– CD8⁺ CAR-T cells with anti-CD19 CAR, with unmodified counterparts serving as controls (WT/uPA^–/– CD8⁺ MOCK-T). Using RM1-CD19 cells ( Supplementary Figure 3 ), uPA^–/– CD8⁺ CAR-T cells executed significantly stronger cytotoxicity than WT group at effector-to-target (E: T) ratios ≥0.5 ( Figure 4G ). Most strikingly, at the E:T ratio of 8:1, uPA^–/– CD8⁺ CAR-T cells exhibited 3.2-fold greater cytotoxicity (uPA^–/– CD8⁺ CAR-T cells: 67.66 ± 0.87%; WT CD8⁺ CAR-T cells: 20.70 ± 0.73%, p<0.0001). We further assessed the migratory capacity of splenic CD8⁺ T cells toward RM-1 cell supernatants. uPA^–/– CD8⁺ T cells displayed superior migration at ratios ≥0.5:1 (CD8⁺ T:RM-1) ( Supplementary Figure 4 ). These findings demonstrated that uPA deficiency amplifies CD8⁺ T cells infiltration, cytotoxicity, and migratory capacity, positioning uPA^–/– CD8⁺ T cells as a superior candidate for adoptive cell therapy.

We administered anti-CD8α or isotype antibodies to WT and uPA^–/– tumor-bearing mice ( Figure 5A ). Compared with isotype control group, depletion of CD8⁺ T cells with anti-CD8α antibodies accelerated tumor progression in both WT and uPA^–/– mice (WT- Isotype: 388.70 ± 163.53 mm³; WT- Anti-CD8α: 659.83 ± 329.32 mm³; uPA^–/– Isotype: 65.39 ± 34.54 mm³; uPA^–/– Anti-CD8α: 256.24 ± 109.25 mm³; Figures 5B, C ), which underscored the centrality of CD8⁺ T cells in antitumor immunity. Although CD8⁺ T cells depletion partially reversed tumor suppression in uPA^–/– mice, the resultant tumor volumes remained 21% smaller than WT isotype controls ( Figure 5B ). This result demonstrated that uPA deficiency mediated tumor suppression primarily through CD8⁺ T cells-dependent mechanisms, while potentially involving CD8⁺ T cells-independent pathways. Flow cytometry confirmed near-complete elimination of CD8⁺ T cells in the peripheral blood, spleen, and tumors across both groups ( Figures 5D, E ), which validated the specificity and efficacy of the depletion regimen.

The PD-1/PD-L1 axis is a central regulator of T and B cell activation, maintaining immune tolerance through inhibitory signaling (33). In solid tumors, the upregulation of PD-L1 on malignant cells engages PD-1 receptors on T cells, thereby dampening cytotoxic activity and facilitating immune evasion (37, 38). RNA sequencing revealed profound PD-1 upregulation in tumors from uPA^–/– mice compared to WT mice ( Figures 6A–C ). This finding was corroborated by CyTOF and flow cytometry, which demonstrated elevated intratumoral PD-1⁺ CD8⁺ T cells infiltration in uPA^–/– mice ( Figures 6D-F ). mIHC further demonstrated that more PD-1⁺CD8⁺ T cells in tumor tissues from uPA^–/– mice ( Figure 6G ). Collectively, uPA deficiency drove a higher frequency of exhaustion-prone CD8⁺ T cells, characterized by elevated PD-1 expression.

To evaluate the synergistic potential of uPA inhibitor and immune checkpoint blockade, we treated WT mice with (1) DMSO, (2) UK122 monotherapy (4 mg/kg), (3) anti-PD-1 monotherapy (100 μg), or (4) the combination of UK122 and anti-PD-1 ( Figure 7A ). Both monotherapies suppressed tumor growth, yet the combination regimen achieved superior efficacy, suggesting additive antitumor effects (Control: 617.93 ± 118.01 mm³; UK122: 257.78 ± 71.54 mm³; anti-PD-1: 353.77 ± 167.47 mm³; combination: 100.16 ± 49.79 mm³, Figures 7B, C ). Flow cytometry and immunostaining revealed more CD8⁺ T cells infiltration in tumors treated with UK122 or anti-PD-1 monotherapy, and combination therapy. Notably, although combination therapy enhanced CD8⁺ T cells infiltration compared to anti-PD-1 monotherapy, it did not exceed UK122 monotherapy levels, suggesting that the superior efficacy likely stemmed from the concerted actions of UK122 promoting T cells recruitment and anti-PD-1 restoring cellular functionality ( Figures 7D-H ).