Abstract Recombinant oncolytic adenovirus offers a novel and promising cancer treatment approach, but its standalone efficacy remains limited. This study investigates a combination treatment strategy by co-administering recombinant oncolytic Adv-loaded silk hydrogel with a PD-L1 inhibitor for patients with bladder cancer to enhance treatment outcomes. Bladder cancer tissues from mice were collected and subjected to single-cell sequencing, identifying CRB3 as a key gene in malignant cells. Differential expression and functional enrichment analyses were performed, validating CRB3’s inhibitory role through in vitro experiments showing suppression of bladder cancer cell proliferation, migration, and invasion. Recombinant oncolytic adenoviruses encoding CRB3 and GM-CSF were constructed and encapsulated in silk hydrogel to enhance drug loading and release efficiency. In vivo experiments demonstrated that the nano-composite hydrogel significantly inhibited tumor growth and increased immune infiltration in tumor tissues. Co-administration of adenovirus silk hydrogel (Adv-CRB3@gel) with a PD-L1 inhibitor significantly enhanced T-cell infiltration and tumor killing. The combination of recombinant oncolytic Adv-loaded nano-composite hydrogel encoding CRB3 and GM-CSF with a PD-L1 inhibitor improves bladder cancer treatment outcomes by effectively recruiting T cells, providing a novel therapeutic strategy. Supplementary Information The online version contains supplementary material available at 10.1186/s12951-024-02903-9. Keywords: Recombinant oncolytic adenovirus, Silk hydrogel, Programmed death ligand-1 inhibitor, Bladder cancer, T cell infiltration, Combination therapy Introduction Bladder cancer is a common malignancy with high incidence and mortality rates worldwide [[32]1–[33]3]. Current treatment modalities for bladder cancer mainly include surgery, radiation therapy, and chemotherapy. However, these treatments still have limitations, such as the inability to effectively address metastatic lesions, limited therapeutic efficacy, and drug resistance [[34]4–[35]6]. To overcome these challenges, recent studies have explored innovative DNA nanomachine-driven chemodynamic therapy (CDT), which has demonstrated potential in penetrating the blood-brain barrier and achieving tumor-targeted treatment in glioblastoma [[36]7]. Additionally, recent research has shown that harnessing the microbiome has great potential in enhancing antitumor therapies. For example, the development of an orally administered inulin-based hydrogel that interacts with the intratumoral microbiome has shown significant efficacy in colorectal cancer treatment by modulating microbiota-related immune responses, providing new insights for bladder cancer treatment strategies [[37]8]. Therefore, it is urgently necessary to search for more effective therapeutic strategies for bladder cancer. In recent years, immune therapy has gained widespread attention and research as a novel treatment for tumors [[38]9, [39]10]. Immune checkpoint inhibitors, such as PD-L1 inhibitors, have demonstrated significant anti-tumor effects in clinical practice by inhibiting the interaction between PD-L1 and PD-1, thereby restoring the immune response of tumor-specific T cells [[40]11–[41]13]. However, there are still some issues with the use of immune checkpoint inhibitors alone in the treatment of bladder cancer, such as the need for high-dose application, the development of immune tolerance, and immune toxicity [[42]11, [43]14]. To overcome these issues and further improve the treatment outcome for bladder cancer patients, our study will explore a new combination therapy strategy by applying recombinant oncolytic adenovirus silk hydrogel (Adv-CRB3@gel) in combination with PD-L1 inhibitors on bladder cancer patients. Recent studies have shown that targeting specific genes can significantly inhibit the proliferation, migration, and invasion of cancer cells [[44]15]. Additionally, a dynamic toxicity landscape analysis of immunotherapy for solid tumors across treatment lines has provided insights into optimizing treatment strategies [[45]16]. The re-engineered oncolytic Adv-CRB3@gel is a novel therapeutic drug carrier known for its excellent biocompatibility and drug sustained-release performance [[46]17, [47]18]. The hydrogel possesses adaptive rheological properties that allow close interaction with surrounding tissues and gentle manufacturing conditions in aqueous solutions without compromising the activity of the virus vector. Silk gel, extracted from mulberry trees, serves as a biocompatible material with limited immunogenicity and cytotoxicity, facilitating sustained drug release to enhance local treatment effects. Silk gel shields AdV from host antibody neutralization and improves its sustained release, enhancing gene editing efficiency and treatment outcomes [[48]19]. Furthermore, through single-cell sequencing and machine learning methods [[49]20–[50]24], we identified the key gene CRB3 in bladder cancer and discovered its significant role in inhibiting cancer cell growth and metastasis. Building on the aforementioned bioinformatics analysis and previous research, this study explores a novel combination therapy strategy, namely, the combined application of reengineered oncolytic Adv-CRB3@gel and PD-L1 inhibitor for bladder cancer patients. Using genetic engineering techniques, we constructed a recombinant oncolytic adenovirus encoding CRB3 and GM-CSF and enhanced drug loading and release efficiency by encapsulating it in silk hydrogel. The study shows that Adv-CRB3@gel significantly suppresses bladder cancer cell proliferation, migration, and invasion. The combined application of Adv-CRB3@gel and PD-L1 inhibitor is expected to have a synergistic effect, significantly enhancing T cell infiltration and improving the ability to kill tumors. Additionally, recent advancements in visualizing cancer drug resistance have introduced nano-quenching and recovery detectors, such as Cy3-AptCD44@BPNSs, which enable dynamic monitoring of drug resistance markers like CD44. This approach allows for real-time visualization of cancer drug resistance, enhancing treatment monitoring and potentially improving therapeutic outcomes [[51]25]. The aim of this study is to improve the treatment outcome and prognosis of bladder cancer patients by combining Adv-CRB3@gel and PD-L1 inhibitors. The scientific and clinical significance of this research lies in providing a new treatment strategy for bladder cancer by enhancing T cell activity and the efficacy of immunotherapy, thus important for improving tumor treatment efficacy and reducing adverse reactions in patients. Moreover, the novelty and interdisciplinary application of this research will provide new insights and methods for the treatment of other types of tumors. Materials and methods Bladder cancer mouse model Six-week-old male BALB/C and C57B/L mice were purchased from our institution’s Animal Experiment Center. All animal studies were conducted in accordance with our institution’s “Guidelines for the Care and Use of Laboratory Animals” and were approved by the Animal Ethics Committee of Zhuhai People’s Hospital (Approval no. 2022AE00092). The mice were given tap water containing 0.05% N-butyl-N-(4-hydroxybutyl)nitrosamine (3817-11-6, Shanghai Xianding Biological Technology Co., Ltd., Shanghai, China) for 12 weeks to induce bladder cancer, followed by normal tap water. The mice were euthanized by cervical dislocation under deep anesthesia with isoflurane (R510-22-10, RWD Life Science Co., Ltd., Shenzhen, China). Bladder tissues of the mice were collected at 1, 2, 4, 12, 20, and 25 weeks after BBN induction to observe tumor formation. Histopathological staining Hematoxylin and eosin (H&E) staining: Bladder tissue samples from the mice were obtained and fixed. After sectioning, the paraffin was removed by incubating the slides in xylene. The slides were then dehydrated in 100% ethanol (64-17-5, Sigma-Aldrich, USA), followed by 95% ethanol, and 70% ethanol, before being either embedded or washed with water. The slides were immersed in a staining solution of hematoxylin (H8070, Solarbio, Beijing, China) and stained for 5–10 min at room temperature. Subsequently, the slides were rinsed with distilled water, dehydrated in 95% ethanol, stained with eosin (G1100, Solarbio, Beijing, China) for 5–10 min, and then underwent routine dehydration, clearing and mounting. Single-cell RNA sequencing (scRNA-seq) Bladder cancer tumor tissues from mice were collected and prepared into single-cell suspension using trypsin (Sigma-Aldrich, USA, 9002-07-7). The C1 Single-Cell Auto Prep System (Fluidigm, Inc., South San Francisco, California, USA) was used to capture individual cells. Upon capture, cells were lysed within the chip, releasing mRNA, and reverse transcribed into cDNA. The lysed and reverse-transcribed cDNA underwent pre-amplification in a microfluidic chip for subsequent sequencing. The amplified cDNA was used to construct libraries, and scRNA-seq was performed on the HiSeq 4000 Illumina platform (sequencing parameters: paired-end reads, read length of 2 × 75 bp, approximately 20,000 reads per cell). Data analysis was performed using the “Seurat” package in R software. Standard quality control criteria included nFeature_RNA between 200 and 5000%.mt less than 20%. From this, the top 2000 genes with the highest expression variation were selected. To reduce the dimensionality of the scRNA-seq dataset, principal component analysis (PCA) was conducted using the top 2000 genes with the highest expression variation. The first 20 principal components were chosen for downstream analysis using the Elbowplot function in the Seurat package. FindClusters function in Seurat was used to identify major cell subpopulations with a default resolution parameter (res = 1). The t-SNE algorithm was then applied for the nonlinear dimensionality reduction of the scRNA-seq data. Marker genes for different cell subpopulations were identified using the Seurat package and annotated using the “Singel R” package. Cell communication analysis was performed using the “CellChat” package, and cell trajectory analysis was conducted using the “Monocle2” package in R software. Differentially expressed genes (DEGs) in the scRNA-Seq dataset were selected using the “Limma” package in R software. The filtering criteria were set as |logFC| > 0.5 and P < 0.05. DEGs between malignant epithelial cells and normal epithelial cells were identified. The “inferCNV” package in R software was utilized to assess copy number variations (CNV) in individual cells based on tumor scRNA-seq data. This package compares the expression levels of genes between malignant and normal cells. Cell subpopulations, including granulocytes, fibroblasts, T cells, endothelial cells, and B cells, were considered while excluding sex chromosomes. The “clusterProfiler” package in R software was used to perform gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis on co-expressed genes. Data visualization analysis was conducted using the ggplot2 package. Weighted gene co-expression network analysis (WGCNA) Differential gene expression profiles were utilized to calculate the Median Absolute Deviation for each gene, and the bottom 50% of genes with the smallest MAD values were excluded. The goodSamplesGenes method from the R software package WGCNA was then employed to remove outlier genes and samples. A scale-free co-expression network was constructed using WGCNA, with the minimum module size set to 30 and sensitivity set to 3. Modules with a distance of less than 0.25 were merged, resulting in a total of 9 co-expression modules. Among these, the grey module was identified as a group of genes that could not be allocated to any specific module. Furthermore, the correlation between modules and groups was examined using the Pearson correlation test (P < 0.05). Particular attention was given to modules significantly associated with the tumor phenotype, including both positively and negatively correlated modules. Genes from these modules were selected for further analysis. Downloading TCGA database mRNA-seq expression profiles and clinical data Data from The Cancer Genome Atlas (TCGA) database ([52]https://cancergenome.nih.gov/) for TCGA-BLCA (Bladder Urothelial Carcinoma) were downloaded, comprising mRNA-seq expression profiles and corresponding clinical information. Patients with follow-up times of 0 days and those lacking survival status or survival time were excluded, resulting in a final cohort of 403 patients for subsequent analyses. COX regression and LASSO analysis Single-factor COX regression analysis was conducted to identify independent prognostic genes (P < 0.05) significantly associated with overall survival (OS) in patients. Subsequently, these independent prognostic genes were utilized in LASSO COX analysis for the development of a risk model. The LASSO approach incorporates an L1 norm penalty on model parameters in order to prevent overfitting. By utilizing this L1 norm penalty, the LASSO regularization can be defined as follows: graphic file with name M1.gif The complexity of the model is controlled by λ, with higher values of λ imposing greater penalties on linear models with more variables. LASSO can also be expressed as a constraint on the objective function. graphic file with name M2.gif One important feature of LASSO regularization is its ability to enforce parameter values to be equal to zero. The goal of LASSO is to create a sparse parameter space, which is an ideal characteristic for feature selection. Construction of prognostic features and development of column graph In the LASSO model, a Kaplan-Meier curve and ROC analysis were performed on the identified genes to select those with outstanding prognostic performance and diagnostic capability for constructing prognostic features. Subsequently, a risk scoring formula based on the prognostic features was utilized to assign a risk level to each patient. The equation for the risk score (RS) is defined as follows: graphic file with name M3.gif The variable “n” represents the number of genes included in the prognostic features. The LASSO coefficients, denoted as “βi,” indicate the contribution of each gene, while “Expi” represents the expression value of gene i. Moreover, based on the expression profiles of selected genes across different samples, a risk score was assigned to each patient. All patients were then divided into high-risk and low-risk groups according to the optimal cut-off point of the risk score. The prognostic significance and diagnostic ability of the risk score were evaluated using Kaplan-Meier curves and ROC analysis. The study selected a panel of 30 independent prognostic genes, and their respective risk scores were calculated using the following formula: Risk Score = -0.009 * ABHD17A + 0.003 * ACKR3–0.007 * BLNK + 0.001 * CAPG + 0.001 * CHCHD2–8.77E-4 * COX14–0.009 * CRB3–0.002 * CYB5B + 7.39E-4 * DAD1 + 0.010 * DCXR − 0.002 * EFHD2 + 0.016 * GAB2 + 8.75E-4 * GJA1–0.028 * ITPK1–4.57E-4 * KCNN4–0.057 * KLHL29 + 2.80E-4 * KRT23 + 5.20E-4 * LMNA + 0.003 * LTBP1 + 0.002 * MIEN1–2.388E-05 * MT-ATP8–0.006 * NDUFA3–0.028 * PDCL3 + 0.023 * PTPRD + 0.005 * RASD1–0.018 * SEL1L3 + 0.102 * SETBP1–0.003 * SRP9 + 0.005 * SVIL + 0.058 * USP13. Univariable and multivariable Cox regression analyses were performed to assess the prognostic independence of various clinical features in the risk score. The independent prognostic features were further incorporated into the construction of the nomogram. The performance of the nomogram in survival prediction was evaluated using the consistency index of actual survival rates and predicted survival rates. The predictive performance of the nomogram was also assessed through visual calibration curves, which compare predicted and observed results. The R package “stdca. R” was used for the prognostic decision curve analysis. Immune analysis Based on the expression profile of TCGA-BLCA, CIBERSORT analysis was performed to assess the scores of 22 immune cell types in each sample and the ESTIMATE algorithm was used to compute the immune infiltration status of each sample. Box plots were generated using R 4.3.0. To investigate the correlation between CRB3 and the 22 immune cell types in the TCGA-BLCA expression profile, the “reshape2,” “ggpubr,” and “ggExtra” packages in R were employed. The prediction of the response to immune checkpoint blockade (ICB) was conducted using the TIDE website, and the differences in TIDE scores between high and low-risk groups were compared using the Wilcoxon test. RT-qPCR Cellular or tissue lysis was performed using the Trizol reagent kit (10296010, Invitrogen, Thermo Fisher, USA), followed by total RNA extraction. The quality and concentration of RNA were assessed using UV-visible spectroscopy (ND-1000, Nanodrop, Thermo Fisher, USA). Reverse transcription was carried out using the PrimeScript™ RT-qPCR kit (RR086A, TaKaRa, Mountain View, CA, USA). Real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) was performed on the LightCycler 480 system (Roche Diagnostics, Pleasanton, CA, USA) using the SYBR Premix Ex TaqTM (DRR820A, TaKaRa). GAPDH was used as the internal reference for mRNA. The primer sequences used for amplification were designed and provided by Shanghai Universal Biotech Co., Ltd. (see Table [53]S1 for primer sequences). The fold-change in the target gene expression between the experimental and control groups was calculated using the 2^−ΔΔCt method, where ΔΔCT = ΔCt experimental group - ΔCt control group, and ΔCt = target gene Ct - internal reference gene Ct [[54]26]. Western blot Protein was extracted from cells and tissues by digestion with trypsin (T4799-5G, Sigma-Aldrich, USA). The digested cells and tissues were collected, and cell lysis was performed using an enhanced RIPA lysis buffer containing a protease inhibitor (AR0108, Wuhan Bodendi Company, Wuhan, China). The protein concentration was measured using the BCA protein quantification kit (AR1189, Wuhan Bodendi Company, Wuhan, China). Protein separation was carried out using SDS-PAGE, followed by the transfer of the separated proteins onto a PVDF membrane. The membrane was blocked with 5% BSA (9048-46-8, Sigma-Aldrich, USA) at room temperature for 1 h and then incubated overnight at 4℃ with the appropriate diluted primary antibody (for detailed information, see Table [55]S2). After washing the membrane three times with PBST (3 × 5 min), it was incubated at room temperature for 1 h with either Anti-Mouse-HRP secondary antibody (Cat # 7076, 1/5000; CST, USA) or Anti-Rabbit-HRP secondary antibody (Cat # 7074, 1/5000; CST, USA). The membrane was then washed three times with PBST (3 × 5 min). PBST was removed, and an appropriate amount of ECL working solution (Omt-01, Beijing Aomijiade Pharmaceutical Technology Co., Ltd., Beijing, China) was added. After incubating at room temperature for 1 min, excess ECL reagent was removed, and the membrane was sealed with plastic wrap, placed in a dark box, and exposed to X-ray film for 5–10 min. The film was developed and fixed. Image J analysis software was used to quantify the intensity of each band in the Western blot image, with GAPDH serving as an internal reference. Cell culture The bladder cancer cell lines used in this study were MB49-Luc (WC0016, Fenghui Biosciences, Hunan, China), MB49 (SNL-256, Shawn Biosciences, Wuhan, China), and MBT-2 (E0696, Shanghai Peak Science Co., Ltd., Shanghai, China). Mouse primary bladder epithelial cells (BEC, MIC-iCell-u007, iCell Bioscience, Shanghai, China) were also included. These cells were cultured in DMEM medium (11965092, Gibco, USA), supplemented with 1% penicillin-streptomycin (10378016, Invitrogen, USA) and 10% fetal bovine serum (FBS, 10100147, Gibco, USA), and incubated at 37℃ and 5% CO[2] in a cell culture incubator. The growth medium was replaced every 3 days. When the cultures reached approximately 80% confluency, the passage was performed using 0.25% trypsin/EDTA (T4174, Sigma-Aldrich, Shanghai, China) digestion. The cell groups were as follows: - MB49: sh-NC (cells transfected with non-targeting shRNA as control), sh-GRB3 (cells transfected with shRNA against GRB3), oe-NC (cells transfected with non-targeting overexpression plasmid as control), or-GRB3 (cells transfected with overexpression plasmid for GRB3). - MBT-2: sh-NC (cells transfected with non-targeting shRNA as control), sh-GRB3 (cells transfected with shRNA against GRB3), oe-NC (cells transfected with non-targeting overexpression plasmid as control), or-GRB3 (cells transfected with overexpression plasmid for GRB3). Bladder cancer cells were incubated with recombinant oncolytic adenovirus hydrogel for 48 h: PBS (cells incubated with PBS as control), Adv-Ctrl (recombinant adenovirus encoding EGFP and GM-CSF as control), Adv-CRB3 (recombinant adenovirus encoding EGFP, GM-CSF, and CRB3); PBS@gel (PBS encapsulated in hydrogel as control), Adv-Ctrl@gel (Adv-Ctrl encapsulated in hydrogel), Adv-CRB3@gel (Adv-CRB3 encapsulated in hydrogel). Lentivirus The GRB3 plasmid was constructed using the pCMV6-AC-GFP plasmid vector (LM-2069, LMAI Bio, Shanghai, China) by SynBio Biotech (Shanghai, China). The GRB3-shRNA derived from mice (sequence 1, 5’-3’: CCAACAAACCTCCTGTCCTTT; sequence 2, 5’-3’: CGGACCCTTTCACAAATAGCA) and sh-NC (sequence, 5’-3’: CCTAAGGTTAAGTCGCCCTCG) were purchased from Thermo Fisher (USA). Lentiviruses, including oe-GRB3 (or oe-GRB3-LTEP-s), sh-GRB3 (or sh-GRB3-LTEP-s), and control lentiviruses oe-NC (or oe-NC-LTEP-s) and sh-NC (or sh-NC-LTEP-s), were constructed based on the HEK293T cell line (CBP60661, Nanjing KeyBay BioTech, Jiangsu, China) using the gene overexpressing plasmids and lentivirus packaging services by SynBio Biotech. After transfection of the plasmids carrying the luciferase reporter gene (oe-NC-luc, oe-GRB3-luc, sh-NC-luc, and sh-GRB3-luc) along with the helper plasmids, Lipofectamine 2000 reagent (11668030, Thermo Fisher, USA) was used for co-transfection into HEK293T cells. After verification, amplification, and purification, packaged lentiviruses were obtained. For lentivirus-mediated cell transfection, 5 × 10^5 cells were seeded into 6-well plates. When cells reached 70–90% confluence, the medium containing an appropriate amount of packaged lentivirus (MOI = 10, working titer approximately 5 × 10^6 TU/mL) and 5 µg/mL polybrene (TR-1003, Merck, USA) was added for transfection. After 4 h of transfection, an equal amount of medium was added to dilute polybrene. After 24 h, the fresh medium was replaced, and after 48 h, transfection efficiency was observed through the luciferase reporter gene. Stably transfected cell lines were selected using an appropriate concentration of puromycin (A1113803, Gibco, Grand Island, NY, USA) to confirm overexpression efficacy and knockdown efficiency by RT-qPCR. CCK-8 assay Bladder cancer cells in the logarithmic growth phase were seeded in a 96-well plate at a density of 5 × 10^3 cells per well. After 24 h of seeding, 10 µL of CCK-8 reagent solution (C0038, Beyotime, Shanghai, China) was added to each well and incubated at 37 °C in a humidified incubator. The absorbance of each well at 450 nm wavelength was recorded using a Microplate Reader (abx700005, Beijing Qiwu Yicheng Technology Co., Ltd.) at 1, 24, 48, 72, and 96 h. Scratch assay Bladder cancer cells were cultured in a 24-well plate. Using a sterile 200 µL pipette tip, a scratch was made in the middle of each well, and then PBS was used to wash away the damaged cell debris. At different time intervals (0 h and 48 h), the cells migrating into the scratched area were imaged, and the percentage of the migrated area was measured using Image J software. Isolation and fluorescent labeling of T cells in vitro First, T cells were isolated from the peripheral blood of 8-week-old BALB/C mice using a T Cell Isolation Kit (11365D, Thermo Fisher, USA), following the manufacturer’s instructions. For experimental tracking, the T cells were fluorescently labelled with carboxyfluorescein succinimidyl ester (CFSE) (ab113853, Abcam, UK). T cells were activated with 1000 U/mL interferon-γ (IFN-γ) (11276905001, Sigma-Aldrich, USA) for 24 h. Transwell assay Two different devices were utilized in this study: Transwell polycarbonate membrane cell culture inserts (CLS3422, Corning, USA) and BioCoat Matrigel invasion chambers with 8.0 μm PET membranes (354480, Corning, USA). For the cell migration assay, a suspension of bladder cancer cells (2 × 10^4 cells/well) was added to the upper chamber in serum-free medium, while the lower chamber was filled with culture medium containing 10% fetal bovine serum. After 24 h of incubation, non-migrating cells in the upper chamber were removed, and the cells that migrated through the membrane were fixed and stained with 0.1% crystal violet (C0121, Beyotime, Shanghai, China). Eight random fields were selected under an inverted microscope (XDS-900, Caikon, Shanghai, China) with a 10× objective for analysis. For co-culturing T cells, CFSE-labeled T cells (5 × 10^3 cells) were seeded along with mCherry-labeled bladder cancer cells (1 × 10^4 cells) (YT477, Beijing Bai Ao Lai Bo Technology Co., Ltd., Beijing, China) in cross-well chambers with or without a matrix gel. The lower chamber was filled with DMEM containing 10% FBS. Cells were washed with PBS and fixed in 4% paraformaldehyde 16–24 h after the seeding. Gentle wiping with a cotton swab was performed on the upper polycarbonate membrane to remove non-adhered cells. Eight random fields were selected under an inverted microscope (XDS-900, Caikon, Shanghai, China) with a 10× objective for analysis. The “Analyze Particles” function in ImageJ was used to quantify the number of migrating cells in each image. Flow cytometry assay Cell Apoptosis: The rate of cell death was detected using a flow cytometry assay. Tumor cells (1 × 10^5/well) were collected, washed with chilled PBS, and stained with the apoptosis detection kit (APOAF-20TST, Sigma-Aldrich, USA) in the dark for 15 min. The cell pellet was then resuspended in 400 µL binding buffer and stained with 5 µL Annexin-V. Flow cytometry analysis classified the cells as follows: Annexin V^+PI^+ (late-stage apoptosis), Annexin V^+PI^− (early-stage apoptosis), Annexin V^−PI^+ (necrotic cells), and Annexin V^−PI^− (live cells). Cell Cycle: Bladder cancer cells were harvested, fixed with 70% ethanol, rinsed twice with cold 1× PBS, and incubated in the dark with FxCycle™ PI/RNase staining solution ([56]F10797, Thermo Fisher, USA) for 30 min. Samples were analyzed on a flow cytometer, and the percentage of cells in each cell cycle phase was analyzed using FCS Express software (De Novo Software). Bladder cancer cells were stained with anti-PD-L1 (BD Bioscience, USA) and anti-CTLA-4 (Thermo Fisher, USA) antibodies, incubated at 4 °C for 30 min, and centrifuged. Cells were then fixed in 2% paraformaldehyde/PBS and analyzed within 24 h using the FACS Aria II flow cytometer (BD Bioscience, USA). For Tregs, cells were stained with anti-CD4-FITC, anti-CD25-PerCP-Cy5.5 (OTWO, China), and anti-Foxp3-PE-Cy7 (Abcam, UK). IFN-γ + CD8 + T cells were stained with anti-CD8-PE and anti-IFN-γ-PE-Cy7 (Abmart Medical Technology, China). Next, the harvested cells were permeabilized and subjected to flow cytometry analysis. For Tregs analysis, the cells were stained with anti-CD4-FITC, anti-CD25-PerCP-cy5.5 (HT1610179, OTWO, Shenzhen, China), and anti-Foxp3-PE-Cy7 (ab210232, Abcam, UK). For the analysis of IFN-γ^+CD8^+ T cells, the cells were stained with anti-CD8-PE and anti-IFN-γ-PE-Cy7 (FL2IFNG05, Abmart Medical Technology, Shanghai, China). After thorough mixing, the cells were incubated in the dark at 4 °C for 30 min. Subsequently, 2 mL of PBS solution (P4417, Sigma-Aldrich, USA) was added, and the mixture was centrifuged at 4 °C, 1500×g for 10 min. The supernatant was discarded, and the cells were fixed with a 2% paraformaldehyde (30525-89-4, Sigma-Aldrich, USA) / PBS solution and left in the dark at 4 °C. Within 24 h, the cells were subjected to analysis using the FACS Aria II flow cytometer (BD Bioscience, USA). Generation of multicellular spheroids (MCS) and 3D co-culture with T cells Bladder cancer cells/spheroids (1000 cells/spheroids) were seeded into 35 or 81 well-agarose tubes (A6013, Sigma-Aldrich, USA) using a 3D Petri dish (Microtissues^® Inc., RI, USA). After seeding, the tubes were incubated at 37 °C and 5% CO[2] for 1 min, followed by the addition of 1 mL (35-well) or 2 mL (81-well) of cell culture medium to form spheroids. To assess invasiveness, Type I collagen (CC050, Sigma-Aldrich, USA) was adjusted to pH 7.0–8.0 and added to the co-culture in the agarose tubes. The tubes were incubated for 4 min, inverted for 1 h, then returned to their original position and filled with RPMI medium containing 5% FBS and 1% Pen/Strep. The invasion assays were conducted for 2 days, with images captured using an inverted microscope as viable cells recovered from the collagen matrix. To evaluate MCS cytotoxicity, spheroids were co-cultured for 24 h, then washed, stained and fixed using the viability/toxicity assay kit (30002, Biotium, USA). Confocal microscopy with Z-stack scanning at 5 μm intervals was performed, followed by the generation of a maximum intensity projection and surface display (2.5D) using Zeiss software. Live/dead cell areas were quantified with Image J. Carboxyfluorescein succinimidyl ester-labeled T cells were co-cultured with the cancer spheroids at a 5:1 ratio in RPMI supplemented with 10% FBS and 100 U/mL penicillin/streptomycin (R4130, Sigma-Aldrich, USA). After 24 h at 37 °C, spheroids were washed, fixed in 4% paraformaldehyde, and imaged with a ZEN 710 confocal microscope (Zeiss, Germany). Middle position images of the spheroids were processed using Image J software. Construction of recombinant adenovirus The cDNA encoding mouse CRB3 and GM-CSF was synthesized by Genescript in Nanjing, China, ensuring appropriate restriction enzyme sites at the 5’ and 3’ ends. After purification, the cDNA was double-digested with FastDigest EcoRI (FD0274, Thermo Fisher, USA) and BamHI (FD0054, Thermo Fisher, USA) restriction enzymes. Subsequently, the coding sequences of CRB3, EGFP, and GM-CSF were inserted into the adenovirus vector pAdEasy-1 (QYV0162, Beijing Qiyan Bioscience Co., Ltd., China) using T4 DNA ligase (EL0011, Thermo Fisher, USA), resulting in the desired recombinant oncolytic adenovirus vector. The cDNA sequences and the recombinant oncolytic adenovirus vector were verified by agarose gel electrophoresis. Virus packaging and collection The recombinant vector was transfected into HEK293T cells for virus packaging. Lipofectamine 3000 transfection reagent (L3000001, Thermo Fisher, USA) was used for transfection according to the manufacturer’s protocol. After 48–72 h, the cell culture supernatant was collected and concentrated using an ultracentrifuge (Optima, Beckman Coulter, USA) to obtain viral particles. Replication of recombinant adenovirus in tumor cells The multiplicity of infection (MOI) for the infection of bladder cancer cells with recombinant adenovirus was set at 10. Samples were collected at 12, 24, 48, 72, 96, and 120 h post-infection. Two methods were used to validate the virus’s replication ability. The first was the TCID50 assay, where viral samples were serially diluted tenfold and added to cell culture plates. After incubation, the percentage of infected cells was recorded, and the TCID50 value was calculated using the Reed-Muench formula: TCID50 = (log reciprocal of the highest dilution at 50% infection) + distance ratio × log dilution factor. The distance ratio was calculated as: Distance ratio = (50% upper critical infection rate − 50%) / (50% upper critical infection rate − 50% lower critical infection rate). The second method involved extracting viral genomic DNA using a genomic DNA extraction kit (19321ES50, Yisheng Biotech, Shanghai, China) from infected samples. Virus quantification was performed using qPCR, and the copy number was normalized using the 12-hour time point as the standard. Enzyme-linked immunosorbent assay (ELISA) Mouse blood serum samples were obtained by collecting venous blood, allowing it to clot for 20–30 min at room temperature, and then centrifuging it at 2000×g for 10 min to obtain the serum samples, which were stored at -80 °C. Alternatively, cell culture supernatants were collected. Firstly, the antigen used was diluted to an appropriate concentration using a coating buffer. The wells of an ELISA plate were blocked with 5% bovine serum (F8318, MSK, Wuhan, China) at 37 °C for 40 min. The diluted samples were added to the wells, followed by the addition of CRB3 ELISA antibody (orb471576, Biorbyt, Wuhan, China), GM-CSF ELISA antibody (ab9741, Abcam, UK), TNF-γ ELISA antibody (CT68250, Genlink, Shanghai, China), and TNF-α ELISA antibody (ab183218, Abcam, UK). The plate was covered with adhesive plastic and incubated at room temperature for 2 h. Subsequently, Anti-Mouse-HRP secondary antibody (Cat# 7076; CST, USA) or Anti-Rabbit-HRP secondary antibody (Cat# 7074; CST, USA) was added. The plate was read at 450 nm using an ELISA reader (Bio-Rad, USA), and a standard curve was generated for data analysis. Optimization of Adv-CRB3@gel release efficiency Under the same power, heating and boiling for different times (0.5, 1, 2 h) resulted in degummed silk with varying molecular weights. The longer the boiling time, the smaller the molecular weight of the silk fibroin, and the slower the gelation rate, with an overall smaller hydrodynamic diameter. To reduce the degradation of silk fibroin, we ultimately selected a boiling time of 30 min. Different MOIs of the virus were used to infect MB49 and MBT-2 cells. At an MOI of 8, the highest CRB3 concentration was observed in the cell supernatant, while at an MOI of 4, the highest GM-CSF concentration was observed in the supernatant. Additionally, the 2 wt% silk hydrogel exhibited the highest transfection efficiency and extended gene expression for 21 days. Preparation of recombinant adv-loaded silk hydrogel Silkworm cocoons (5 g) were cut into small pieces and boiled in 20 mM Na[2]CO[3] solution (Sigma-Aldrich, USA) for 30 min. The cocoons were washed three times with ddH[2]O to remove the silk protein. The extracted silk fibroin protein was then air-dried at 60 °C for 12 h, then dissolved in 9.3 M LiBr salt solution (Sigma-Aldrich, USA) for 4 h. After dissolving, the solution was dialyzed for 72 h to remove the LiBr. The resulting silk fibroin solution was purified by centrifugation at 4 °C and stored for later use. To prepare a 2 wt% silk hydrogel, the solution was subjected to ultrasound for 180 s at 30% amplitude using a Scientz-iid ultrasonic probe (Ningbo ET Biotech Co., Ltd.). The mixture was then aged at 37 °C to form the hydrogel. Equal volumes of 2 wt% silk hydrogel were mixed with either a virus solution (Adv@gel) or PBS (PBS@gel). The internal structure of the hydrogel was examined using scanning electron microscopy (SEM, S-4800, Hitachi) and ZEN 710 confocal microscope. Evaluation of rheological and swelling properties To evaluate its rheological properties, Adv@gel (2 × 10^6 pfu) was subjected to rheological experiments using a dynamic shear rheometer. The storage modulus (G’) and loss modulus (G”) of Adv@gel were measured at appropriate strains and stresses. To assess its swelling behaviour, silk-water gel or Adv@gel was immersed separately in PBS, and the weight of each group was measured daily. The swelling ratio was calculated by dividing this weight by the initial weight of the gel. Shear modulus was measured using an oscillatory rheometer (Kinexus, Malvern Instruments, UK) through time sweep testing at a frequency of 1 Hz, strain of 0.5%, 37°C, and 30 minutes. Analysis of sustained release Adv@gel (2 × 10^6 pfu) was placed in a cell filter with a pore size of 8 μm. The filter was embedded in a 24-well plate containing 1 × 10^5 bladder cancer cells in culture. Every 24 h, the cells were washed twice with PBS and another 24-well plate was prepared with 1 × 10^5 bladder cancer cells that were not infected. Fluorescence imaging and flow cytometry were performed on the previous bladder cancer cells to determine the percentage of EGFP-positive cells. Cell viability assay using MTT Cells were seeded in 96-well plates at a density of 1 × 10^4 cells per well and incubated for 24 h. For different treatment groups, cells were mixed with 0.01 mL of MTT solution (CT02, Sigma Aldrich, USA) and incubated for 4 hours in a CO[2] incubator. Then, 0.1 mL of isopropanol containing 0.04 N HCl was added to each well. The mixture was thoroughly mixed by pipetting with a multichannel pipette. The HCl converted the phenol red in the culture medium to a yellow colour that does not interfere with the measurement of MTT formazan. Cell viability was determined by measuring the absorbance at 570 nm at 0, 24, 48, and 72 h [[57]27]. Biochemical analysis Peripheral blood from mice was collected and analyzed using a Vitros 5.1FS automatic biochemical analyzer (Ortho Clinical Diagnostics, USA) to measure serum albumin (ALB), alkaline phosphatase (ALP), aspartate transaminase (AST), alanine transaminase (ALT), total cholesterol (CHOL), red blood cell count (RBC), white blood cell count (WBC), and haemoglobin (HGB). Immunofluorescence Cells or tissues were washed with cold PBS and fixed with 4% paraformaldehyde (P885233, Macklin, USA) for 15–30 min. Subsequently, they were treated with 0.1% Triton (L885651, Macklin, USA) for 15 min. After two washes with PBS, they were incubated with PBS containing 15% FBS overnight at 5 °C. The cells or tissues were then stained with rabbit anti-CD25 antibody (ab238272, Alexa Fluor^® 568, Abcam, UK; 1:100) or rabbit anti-CD69 antibody (PA5-102562, Alexa Fluor™ 594, Thermo Fisher, USA; 1:100) and incubated overnight at 4 °C. After washing three times with TBST (TBS containing 1% Tween-20), the cells were incubated with goat anti-rabbit Alexa Fluor^® 568 secondary antibody (A-11011, Thermo Fisher, USA) or goat anti-rabbit Alexa Fluor™ 594 secondary antibody (A-11012, Thermo Fisher, USA) for 2 h at room temperature. DAPI (D1306, Thermo Fisher, USA) was used for counterstaining. Fluorescence intensity was measured using a fluorescence microscope (Zeiss Observer Z1, Germany), and target areas were selected in the images for quantification using ImageJ software. The number of positive cells was counted. In vivo fluorescence imaging To establish a subcutaneous bladder cancer tumor model, 200 µL of MB49-Luc (2 × 10^6) cell line was injected into the lower abdomen of 6-week-old BALB/C mice. The tumor formation, approximately 2 mm in size, was observed macroscopically after approximately 4 weeks. A total of 40 mice were used. In vivo, fluorescence imaging of the mice was performed using the IVIS Lumina Series III imaging system (PerkinElmer, USA). Prior to imaging, the mice were briefly anesthetized to ensure their immobility throughout the procedure. The mice were grouped as follows: PBS (mice injected with PBS as a control), Adv-Ctrl (recombinant adenovirus encoding EGFP and GM-CSF as a control), Adv-CRB3 (recombinant adenovirus encoding EGFP, GM-CSF, and CRB3); PBS@gel (PBS encapsulated in a hydrogel as a control), Adv-Ctrl@gel (Adv-Ctrl encapsulated in a hydrogel), Adv-CRB3@gel (Adv-CRB3 encapsulated in a hydrogel). The dosage was 1 × 10^7 pfu with a volume of 50 µL, injected directly into the tumor. The injection was performed in the 7th week after subcutaneous tumor formation as a single injection. The internal organs and tumors of mice were collected at 1, 3, and 5 days after injection with Adv-CRB3 and Adv-CRB3@gel. The ex vivo images were obtained using a near-infrared fluorescence imaging system (UNITED WELL) to observe the fluorescence signals at different time points. Combined treatments were administered as follows: PBS + IgG (mice injected with PBS and IgG as a control; IgG: BE0083, bioxcell, Shenzhen, China); PBS + Anti-PD-L1 (mice injected with PBS and Anti-PD-L1; Anti-PD-L1: BP0101, bioxcell, Shenzhen, China); Adv-CRB3@gel + IgG (mice injected with Adv-CRB3@gel and IgG); Adv-CRB3@gel + Anti-PD-L1 (mice injected with Adv-CRB3@gel and Anti-PD-L1). IgG or Anti-PD-L1 was co-injected with the recombinant adenovirus hydrogel. The dosage was 20 mg/kg. Immunohistochemistry Tissues or cells to be examined were extracted, fixed, and embedded. The embedded tissues were sectioned and subjected to dewaxing treatment, which removes the wax from the sections and renders them hydrophilic for subsequent immunostaining procedures. Dewaxed tissue sections were then treated with specific Ki67 antibody (SAB5700770, 1:200; Sigma Aldrich, USA) and PCNA antibody (13110, 1:200; CST, USA), followed by treatment with the Anti-Rabbit-HRP secondary antibody (12–348, 1:1000; Sigma Aldrich, USA). DAB staining reagent (ab64238, Abcam, USA) was used to visualize the sites of antibody binding. The stained tissue sections were dewaxed and coverslipped for observation. The sections were observed under a microscope, and the expression patterns were recorded. To assess the staining results, five random areas of the lesion were selected under the microscope, and the number of positively stained cells was counted. Statistical analysis The data in this study were obtained from at least three independent experiments and presented as mean ± standard deviation (mean ± SD). For comparisons between the two groups, we employed the two-sample independent t-test. For comparisons among three or more groups, a one-way analysis of variance (ANOVA) was used. If the ANOVA results revealed significant differences, Tukey’s HSD post-hoc test was conducted to compare differences between groups. For data with non-normal distribution or unequal variances, the Mann-Whitney U test or Kruskal-Wallis H test was employed. All statistical analyses were performed using GraphPad Prism 9 software (GraphPad Software, Inc.) and R language. A significance level of 0.05 was used for all tests, and a two-sided p-value less than 0.05 was considered statistically significant. Results Successful construction and validation of a BBN-induced mouse model of bladder cancer Bladder cancer is one of the most common malignancies of the urinary tract and a leading global cancer [[58]28]. The N-butyl-N-(4-hydroxybutyl) nitrosamine (BBN)-induced mouse model has been widely adopted as an experimental model for studying bladder cancer [[59]29]. To induce bladder cancer, we administered tap water containing 0.05% BBN to 6-week-old C57B/L mice for 12 weeks, followed by normal tap water (Figure [60]S1A). As the duration of BBN treatment increased, the proportion of mice with urinary epithelial lesions continued to rise (Figure [61]S1B), with increasing severity of variation. At 1 week of treatment, moderate reactive variation was observed; at 2 weeks, extensive reactive atypia was observed; and at 4 weeks, concentrated reactive atypia and dysplasia were observed. By the 12th week of treatment (after switching to normal water), most mice exhibited focal urothelial dysplasia and focal CIS (carcinoma in situ); at 20 weeks of treatment, the majority of mice had tumors in the early stage of invasion into the subepithelial connective tissue; and by the 25th week, all mice had developed invasive tumors with glandular and squamous differentiation (Figure [62]S1C). Additionally, through anatomical examination, we visually observed the morphology of bladder tumors in the mice (Figure [63]S1D). Using the BBN-induced mouse model, we successfully established a stable and reliable mouse model of bladder cancer. scRNA-seq reveals cellular heterogeneity and T-cell exhaustion in bladder cancer We obtained tumor tissues from a mouse model of bladder cancer after 25 weeks of treatment and performed scRNA-seq analysis (Fig. [64]1A). The data were integrated using the Seurat package to investigate gene counts, mRNA molecule numbers, and the percentage of mitochondrial genes. The majority of cells had nFeature_RNA < 5000, nCount_RNA < 20,000, and percent.mt < 20% (Figure [65]S2A). Post-filtering retained 15,119 genes and 34,011 cells, with a correlation coefficient of r = -0.15 for nCount_RNA with percent.mt and r = 0.92 for nCount_RNA with nFeature_RNA (Figure [66]S2B). The top 2000 highly variable genes, determined by variance, were selected for downstream analysis (Figure [67]S2C). Linear dimensionality reduction through PCA was carried out, illustrating the main gene expression heatmaps of PC_1-PC_6 and the cell distributions on PC_1 and PC_2 (Figure [68]S2D-E), revealing batch effects. The harmony package was used for batch effect correction (Figure [69]S2F). Fig. 1. [70]Fig. 1 [71]Open in a new tab Revealing cellular heterogeneity in the bladder cancer microenvironment. Note (A) Schematic diagram of single-cell sequencing process; (B) Visualization of cell annotation results based on single-UMAP clustering; (C) Heatmap showing the correlation of marker genes for six different cell types; (D) Identification of 12 clusters after extracting epithelial cells; (E) Calculation of copy number amplification and deletion features using InferCNV, with myeloid cells, fibroblasts, T cells, endothelial cells, and B cells as controls; (F) UMAP visualization of the re-annotated seven cell types after annotation of malignant epithelial cells Nonlinear dimensionality reduction using the UMAP algorithm was performed on the top 20 principal components, clustering all cells into 23 cell clusters (Figure [72]S3A-B). These were automatically annotated into six cell types-T cells, myeloid cells, epithelial cells, fibroblasts, endothelial cells, and B cells using the SingleR package (Fig. [73]1B). UMAP and correlation plots of marker genes were presented (Figure [74]S3C, Fig. [75]1C), showing a predominance of epithelial cells (Figure [76]S3D). Large-scale chromosomal copy number variations (CNVs) were inferred in epithelial cells based on the results from the InferCNV tool (Fig. [77]1D), defining Clusters 1, 3, and 9 as malignant epithelial cells and the rest as normal epithelial cells (Fig. 1F). The proportions of re-annotated cells in each sample were statistically analyzed (Figure [78]S3E). The CellChat package was used to analyze intercellular pathway activity, revealing interactions among seven cell types (Fig. [79]2A). Signal role analysis within the cellular aggregation communication network indicated that malignant epithelial cells were the primary signaling entities (Fig. [80]2B). Their interactions with other cells suggested T cell exhaustion (Fig. [81]2C). Temporal analysis using the Monocle2 package showed cellular transition trajectories indicating state changes (Fig. [82]2D), dividing into five States (Fig. [83]2E). By cellular type analyses, malignant and normal epithelial cells diverged at “2,” with T cells predominantly in the early stage (Fig. [84]2F). Fig. 2. [85]Fig. 2 [86]Open in a new tab Cellular communication and cell trajectory analysis. Note (A) Circular plot of cell communication in the sample, with line thickness representing the number of pathways and interaction strength; (B) Signal role analysis on the aggregated cell communication network; (C) Separate display of communication between malignant epithelial cells and normal epithelial cells with other cells; (D) Trajectory skeleton plot with pseudo time colouring, with each point representing a cell; (E) Trajectory skeleton plot with state colouring, where different states are represented by different colours; (F) Trajectory skeleton plot with cell type colouring In conclusion, seven cell types were identified, with T cells exhibiting exhaustion during bladder cancer progression. scRNA-seq combined with TCGA-BLCA dataset identifies CRB3 as a key gene in bladder cancer and its association with prognosis Through scRNA-seq analysis, crucial prognostic genes were identified in bladder cancer patients. Differential expression analysis revealed 896 genes significantly different between malignant epithelial cells and normal epithelial cells, with 524 genes upregulated and 372 genes downregulated in malignant cells (Fig. [87]3A). GO and KEGG enrichment analyses indicated significant enrichment of these genes in multiple cancer-related signaling pathways, such as bladder cancer, the P53 signaling pathway, and cell cycle G1/S phase transition (Fig. [88]3B). Further, WGCNA categorized these differential genes into 9 modules, with modules significantly related to bladder cancer, including brown, pink, blue, and black modules (Figure [89]S4A-D). These modules collectively contained 380 genes significantly associated with bladder cancer. In the TCGA-BLCA dataset, 48 genes significantly associated with overall survival were determined through single-factor COX regression analysis. LASSO COX regression analysis ultimately selected 30 independent prognostic-related genes (Table [90]1; Fig. [91]3C-D). Patients were stratified into high-risk and low-risk groups based on the calculated risk scores from the expression levels of these genes. The high expression group of GJA1, LTBP1, PTPRD, SETBP1, SVIL, and USP13 showed poorer survival outcomes, while the high expression group of BLNK, COX14, CRB3, EFHD2, KCNN4, MIEN1, NDUFA3, PDCL3, and SEL1L3 exhibited longer survival times. The expression of the remaining 15 genes did not show significant survival differences (Figure [92]S5). Kaplan-Meier curves demonstrated significantly lower survival rates in the high-risk group compared to the low-risk group (Fig. [93]3E, Figure [94]S6A). The accuracy of survival prediction for 1, 3, and 5 years was assessed through ROC analysis, with AUC values of 0.79, 0.80, and 0.81, respectively (Fig. 3F). A Nomogram chart was constructed based on risk score and clinicopathological parameters to predict survival probabilities for 1, 3, and 5 years (Figure [95]S6B-C), along with decision curve analysis (DCA) (Figure [96]S6D), demonstrating high prediction accuracy and clinical utility. Fig. 3. [97]Fig. 3 [98]Open in a new tab Analysis of key genes associated with bladder cancer prognosis using integrated scRNA-seq and TCGA-BLCA datasets. Note (A) Volcano plot displaying DEGs between normal and malignant epithelial cells in the scRNA-seq dataset; red dots represent significantly upregulated genes, blue dots represent significantly downregulated genes, and gray dots indicate genes with no significant difference; (B) GO and KEGG functional enrichment analysis of significantly DEGs between normal and malignant epithelial cells; (C-D) Identification of 30 candidate genes through LASSO analysis based on the TCGA-BLCA dataset, with Lambda set at 0.03; (E) Visualization of trends in risk levels and survival status in BLCA; (F) Risk scores based on LASSO analysis in the TCGA-BLCA dataset for ROC analysis at 1, 3, and 5-year time points; (G) Correlation plot showing the relationship between 10 genes significantly associated with TCGA-BLCA prognosis and 7 cell types in the scRNA-seq dataset; (H) Biased jitter plot illustrating the distribution of CRB3 expression levels in the scRNA-seq dataset, with cell types on the x-axis and gene expression levels on the y-axis; (I-K) Expression levels of CRB3, MIEN1, and LTBP1 in mouse bladder tissues detected via RT-qPCR (C) and Western blot (D-E). Each group includes 6 mice, with values presented as mean ± standard deviation. P-values < 0.05, < 0.01, and < 0.001 are denoted as, , and **, respectively Table 1. Single factor COX regression analysis and LASSO COX analysis Gene Univariate COX Regression Analysis LASSO COX HR HR.95 L HR.95 H P value Coefficients ABHD17A 0.932487 0.872775 0.996283 0.038428 -0.00858 ACKR3 1.003688 1.0014 1.005981 0.00157 0.003045 BLNK 0.952831 0.911912 0.995587 0.03097 -0.00749 CAPG 1.002104 1.000592 1.003617 0.006357 0.001296 CHCHD2 1.001527 1.000218 1.002837 0.022212 0.001057 COX14 0.988859 0.979754 0.998049 0.017609 -0.00088 CRB3 0.971306 0.952873 0.990095 0.002898 -0.00906 CYB5B 0.98089 0.962342 0.999794 0.04758 -0.00235 DAD1 1.003243 1.000805 1.005687 0.009113 0.000739 DCXR 1.007852 1.001716 1.014026 0.012065 0.009956 EFHD2 0.994196 0.990677 0.997728 0.001293 -0.0017 GAB2 1.054115 1.008564 1.101724 0.019371 0.015717 GJA1 1.002873 1.000789 1.004962 0.006861 0.000875 ITPK1 0.968267 0.939607 0.997801 0.035417 -0.02804 KCNN4 0.993011 0.986144 0.999925 0.047574 -0.00046 KLHL29 0.840391 0.718937 0.982364 0.029008 -0.05679 KRT23 1.001826 1.000777 1.002877 0.000645 0.00028 LMNA 1.003489 1.00027 1.006718 0.033638 0.00052 LTBP1 1.012952 1.005825 1.02013 0.000354 0.003332 MIEN1 1.001712 1.000036 1.003391 0.04522 0.001681 MT-ATP8 0.999912 0.99984 0.999985 0.017873 -2.39E-05 NDUFA3 0.981927 0.968038 0.996016 0.012101 -0.00619 PDCL3 0.953009 0.927344 0.979384 0.000549 -0.0284 PTPRD 1.124828 1.023496 1.236192 0.014601 0.022775 RASD1 1.007293 1.002035 1.012578 0.006501 0.004882 SEL1L3 0.968845 0.944954 0.993339 0.012971 -0.01827 SETBP1 1.267323 1.127874 1.424012 6.80E-05 0.101847 SRP9 0.993596 0.987492 0.999738 0.041019 -0.00251 SVIL 1.01839 1.006468 1.030453 0.002422 0.004949 USP13 1.156983 1.051946 1.272508 0.002675 0.057599 [99]Open in a new tab Note HR Hazard Ratio In scRNA-seq datasets, further analysis revealed that GJA1, LTBP1, EFHD2, KCNN4, PDCL3, PTPRD, and SVIL exhibit high expression levels in malignant cells, whereas BLNK, COX14, CRB3, MIEN1, NDUFA3, SETBP1, SEL1L3, and USP13 show significantly lower expression levels (Figure [100]S7). UMAP plot analysis demonstrated that GJA1 and LTBP1 are predominantly expressed in malignant epithelial cells, while MIEN1 and CRB3 are mainly expressed in normal epithelial cells (Figure [101]S8). Correlation analysis further revealed a significant negative correlation of CRB3, MIEN1, and NDUFA3 with malignant cells, while a significant positive correlation was found for LTBP1 and PTPRD with malignant cells (Fig. [102]3G). Expression levels of CRB3, MIEN1, and LTBP1 in normal and bladder cancer mouse bladder tissues were validated using RT-qPCR and Western blot techniques. The results indicated a significant decrease in the expression levels of CRB3 and MIEN1 in bladder cancer mice, while LTBP1 showed a significant increase, with CRB3 exhibiting the most notable change in expression level (Fig. [103]3I-K). A biased jitter plot demonstrating the distribution of CRB3 expression levels showed an initial increase followed by a decrease over time (Fig. [104]3H). In conclusion, CRB3 plays a crucial role in bladder urothelial cancer, with its low expression correlating with increased malignancy and poor prognosis. As a key gene, CRB3 may suppress bladder cancer progression by regulating biological processes such as epithelial cell polarity and cell adhesion. High-risk BLCA patients have lower immune infiltration In the scRNA-seq dataset, a lower proportion of immune cells, such as T cells, was observed. Studies have shown that high immune infiltration is associated with better prognosis in bladder cancer patients [[105]30]. To assess the immune cell infiltration in bladder urothelial carcinoma (BLCA) in detail, the TCGA-BLCA dataset was analyzed using the “CIBERSORT” algorithm. The results show differences in the infiltration of various immune cells, including B cells, CD4 + T cells, CD8 + T cells, macrophages, and dendritic cells (DC cells), between the high-risk and low-risk groups (Fig. [106]4A). Using the Estimate package to calculate immune infiltration scores, stromal scores, and overall estimation scores, all scores in the high-risk group were significantly higher than those in the low-risk group (Fig. [107]4B). TIDE scoring analysis revealed that the Dysfunction and Exclusion scores in the high-risk group were higher than those in the low-risk group, indicating a poorer response to immune therapy in the high-risk group (Fig. [108]4C-D). Analysis of MSI scores showed that the low-risk group was more sensitive to immune therapy (Fig. [109]4E). Based on CIBERSORT analysis, the correlation between CRB3 and 22 immune cells was further investigated. The results indicate a significant correlation between CRB3 and 9 immune cells, including B cells, dendritic cells, monocytes, CD4 + T cells, and Tregs (Fig. [110]4F). In-depth data analysis revealed that high-risk BLCA patients exhibit lower immune infiltration, poorer response to immune therapy, and correlation between CRB3 and multiple immune cells. This finding provides new insights for immune therapy in bladder cancer. Fig. 4. [111]Fig. 4 [112]Open in a new tab Exploration of immune cell infiltration and immune therapy response characteristics in a high-risk group of BLCA. Note (A) The infiltration of 22 immune cells in the high-risk group and low-risk group was evaluated using the CIBERSORT method; (B) The infiltration of stromal and immune cells in the high and low-risk group tumors was estimated using the ESTIMATE algorithm; (C) The Dysfunction score in the high and low-risk groups, indicating P < 0.05; (D) The Exclusion score in the high and low-risk groups, * indicates P < 0.001; (E) The MSI score in the high and low-risk groups, indicating P < 0.05; (F) Correlation scatter plot of CRB3 with 9 immune cells The correlation between CRB3 expression and proliferation, migration, and invasion abilities in bladder cancer cells Bioinformatics analysis identified CRB3 as a key gene significantly associated with bladder cancer. In bladder cancer, CRB3 expression is significantly reduced and correlated with various immune cells. Initially, we conducted a quantitative analysis using RT-qPCR on two mouse bladder cancer cell lines, MB49 and MBT-2. The results indicated a significant decrease in CRB3 gene expression in both MB49 and MBT-2 cell lines compared to normal cells (mouse primary bladder epithelial cells, BECs), with MB49 showing a more pronounced decrease (Figure [113]S9A). To further investigate the function of CRB3, we performed CRB3 gene knockout experiments (transfection with sh-CRB3) on the MB49 cell line. The results showed that sh-CRB3-1 significantly reduced CRB3 expression, leading us to select sh-CRB3 for subsequent experiments (referred to as sh-CRB3) (Figure [114]S9B). Similarly, in the MBT-2 cell line, we carried out CRB3 overexpression experiments using the lentiviral system, resulting in a significant enhancement of CRB3 expression (Figure [115]S9C). To understand the role of CRB3 in regulating bladder cancer cell proliferation, we conducted CCK-8 experiments for validation. The experimental data revealed a significant increase in cell proliferation capacity in CRB3-interfered MB49 cells, while the proliferation capacity of MBT-2 cells overexpressing CRB3 was significantly inhibited (Fig. [116]5A). Furthermore, through flow cytometry, we detailedly analyzed the cell cycle and apoptosis status. The results showed that the cell cycle was accelerated and apoptosis was reduced in the MB49-sh-CRB3 group, whereas the MBT-2-oe-CRB3 group exhibited the opposite trend (Fig. [117]5B-E). Fig. 5. [118]Fig. 5 [119]Open in a new tab Investigating the Regulatory Role of CRB3 in Bladder Cancer Cell Function. Note (A) CCK-8 assay to assess tumor cell proliferation at different time points; (B-C) Flow cytometry analysis of tumor cell apoptosis, with apoptotic cells outlined in red boxes in panel C for statistical representation; (D-E) Cell cycle analysis of tumor cells using flow cytometry, with statistical representation in panel E; (F-G) Scratching of a single layer of bladder cancer cells with a sterile micropipette tip for migration observation under a microscope, scale bar = 100 μm, and statistical analysis of cell migration in panel G; (H-I) Transwell assay evaluating the migratory capacity of bladder cancer cells, scale bar = 50 μm, and statistical representation of migrated cells in panel I; (J-K) Collection of cell culture images using inverted microscopy in bright-field mode, outlining the invasive cell area with a white dashed line, scale bar = 50 μm, and statistical analysis of invasive areas in panel K; (L-M) Confocal microscopy images from live/dead viability assay in MCS, where live cells are stained with Calcein AM (green) and dead cells with EthD-1 (red), scale bar = 50 μm, and statistical analysis of live/dead cells in panel M. All cell experiments were performed thrice, and values are presented as mean ± standard deviation. Significance levels are denoted as * for P < 0.05, ** for P < 0.01, and *** for P < 0.001 Scratch test results demonstrated that interference with CRB3 resulted in significantly enhanced migration ability in MB49 cells, while the migration ability of MBT-2 cells overexpressing CRB3 was significantly inhibited (Fig. [120]5F-G). Similarly, in the Transwell experiment, we observed a similar migration pattern, with CRB3 interference enhancing the migration ability of MB49 cells while CRB3 overexpression inhibited the migration ability of MBT-2 cells (Fig. [121]5H-I). Closer to in vivo conditions, we first constructed 3D cancer spheroids of MB49 and MBT-2 cell lines and observed the invasion of spheroids 48 h after transfection. Results showed a sharp increase in the invasive spread of MB49-sh-CRB3 spheroids, while MBT-2-oe-CRB3 significantly reduced the invasive spread of spheroids (Fig. [122]5J-K). We used Calcein AM to stain live cells and EthD-1 to stain dead cells. Microscopic observation revealed that CRB3 overexpression caused most tumor cells to die, while most tumor cells in the CRB3 knockout group survived (Fig. [123]5L-M). These experimental results demonstrate the significant role of CRB3 in inhibiting bladder cancer cell proliferation, migration, and invasion. Construction and expression of recombinant adenoviruses carrying CRB3 and GM-CSF Oncolytic adenoviruses (OVs) are viruses that can selectively replicate and destroy tumor cells [[124]31]. Increasing evidence suggests that OVs can serve as exogenous gene delivery vehicles to enhance anti-tumor immune responses in the tumor microenvironment [[125]32]. Granulocyte-macrophage colony-stimulating factor (GM-CSF) is a potent stimulator of antigen-presenting cells (APCs) [[126]33] and can enhance the activation and proliferation of T cells [[127]34]. Although we have validated the relationship between CRB3 and immune cells through bioinformatics, there have been no direct reports of the recruiting role of CRB3 on immune cells. Therefore, we constructed a recombinant oncolytic adenovirus carrying CRB3 and GM-CSF to further enhance the immunotherapy effect on bladder cancer and increase treatment efficacy. First, the coding sequences of mouse CRB3 and GM-CSF were successfully amplified using PCR techniques, ensuring that they carried appropriate 5’ and 3’ end restriction enzyme sites. The size of the amplified products was confirmed to be as expected through 1% agarose gel electrophoresis (Fig. [128]6A). Following double digestion and T4 DNA ligase treatment, the coding sequences of CRB3 and GM-CSF were successfully inserted into the adenovirus vector (pAdEasy-1) (Fig. [129]6B). The correctness of the recombinant oncolytic adenovirus vector was confirmed through gel electrophoresis and sequencing (Fig. [130]6C). Fig. 6. [131]Fig. 6 [132]Open in a new tab Construction of Recombinant Oncolytic Adenovirus. Notes (A) Agarose gel electrophoresis showing bands of GM-CSF and GRB3; (B) Strategy for constructing recombinant adenovirus; (C) Agarose gel electrophoresis demonstrating recombinant adenovirus; (D) Fluorescent images of HEK293T cells infected with Adv-Ctrl or Adv-CRB3 after 24 h, Bar = 25 μm; (E) Viral replication ability assessed by TCID50 titration method; (F) Viral DNA extraction and quantification of viral copy numbers using quantitative PCR; (G) ELISA detection of CRB3 and GM-CSF concentrations in cell supernatant; (H-J) Expression levels of CRB3 in MB49 and MBT-2 cells after incubation with recombinant adenovirus using RT-qPCR (H) and Western blot (I-J). All cell experiments were conducted in triplicate, and values are expressed as mean ± standard deviation, where “ns” indicates no significant difference, and “***” indicates P < 0.001 Next, the recombinant vector was transfected into 293T cells for viral packaging (Fig. [133]6D). After 48–72 h, the cell culture supernatant was collected and subjected to centrifugation to obtain high-titer viral particles. Virus titer measurements indicated that the obtained virus activity was within the range of 1 × 10^9 PFU/ml (Fig. [134]6E). To determine whether the insertion of CRB3 affected the replication ability of the oncolytic adenovirus, MB49 and MBT-2 cells were infected with viruses at different MOIs. Compared to Adv-Ctrl, Adv-CRB3 showed the same replication ability in MB49 cells (Fig. [135]6E-F). Furthermore, ELISA confirmed that CRB3 and GM-CSF were expressed and secreted into the cell culture supernatant by Adv-CRB3, while Adv-Ctrl also secreted GM-CSF (Fig. [136]6G). Meanwhile, after infecting MB49 and MBT-2 cells with adenoviruses for 48 h, the expression level of CRB3 was detected by RT-qPCR and Western blot. The results showed that compared to the Adv-Ctrl group, Adv-CRB3 significantly increased the expression of CRB3 (Fig. [137]6H-J), with consistent effects observed in MB49 and MBT-2 cells. Therefore, MB49 cells were chosen for subsequent experiments. In summary, we successfully constructed a recombinant adenovirus encoding and expressing CRB3 and GM-CSF without altering the replication ability of CRB3. Silk hydrogel as a biocompatible material for tumor-targeted virus delivery system Silk hydrogel derived from mulberry silkworm silk is a biocompatible material capable of sustained drug release with limited immunogenicity and cytotoxicity, thereby enhancing therapeutic efficacy through local injection [[138]35, [139]36]. In this study, we introduced silk hydrogel produced from silk cocoons as a delivery system for tumor-targeted virus delivery (Fig. [140]7A). Subsequently, we encapsulated Adv-CRB3 in the silk hydrogel (Adv-CRB3@gel) and verified its biological characteristics. Adv-CRB3@gel could be injected and form a gel state (Fig. [141]7B). Fig. 7. [142]Fig. 7 [143]Open in a new tab Evaluation of cell toxicity of Adv-CRB3@gel. Note (A) Schematic representation of the gelation process of silk gel solution; (B) Injectable characteristics and gelation of encapsulated silk gel; (C) Effect of water gel on cell viability; (D) SEM image of water gel, Bar = 300µm; (E-F) Confocal microscopy observation of water gel (E) and fluorescence spectrum of HSV-1Dylight 550 excited at 520nm (F); (G) Storage modulus (G’) and loss modulus (G”) of Virus@gel under appropriate strain; (H) Swelling ratio of silk gel and Virus@gel; (I) Representative fluorescence images of EGFP-positive MB49 cells in release experiment; (J) Cell viability of bladder cancer cells after co-culture with recombinant adenovirus water gel determined by MTT assay. All experiments were repeated three times, and the values are expressed as mean ± standard deviation. * indicates P < 0.01, ** indicates P < 0.001 Our study confirmed the lower cytotoxicity and good biocompatibility of silk hydrogel through the MTT assay conducted on primary mouse urothelial cells (Fig. [144]7C). Scanning electron microscopy and confocal microscopy were employed to confirm the porous network structure of Adv-CRB3@gel and the distribution of virus particles within the hydrogel, respectively (Fig. [145]7D-F). Further evaluation of rheological and swelling properties demonstrated that Adv-CRB3@gel possessed favourable mechanical properties and material stability, maintaining a gel state with a low swelling ratio (Fig. [146]7G-H). Fluorescence observation indicated that Adv-CRB3@gel sustained virus release within 7 days, exhibiting superior sustained release capability compared to Adv-CRB3 (Fig. [147]7I). Cell viability assay revealed that after co-culturing with bladder cancer cells for 3 days, both Adv-CRB3 and Adv-CRB3@gel induced tumor cell death, with the oncolytic effect of Adv-CRB3@gel gradually increasing, suggesting its excellent oncolytic efficacy (Fig. [148]7J). In conclusion, these findings demonstrate that Adv-CRB3@gel exhibits low systemic toxicity and good biocompatibility. Recombinant oncolytic adenovirus hydrogels recruiting T cells and enhancing immunotherapy efficacy To validate the recruitment of T cells and the enhanced immunotherapy function of the recombinant oncolytic adenovirus hydrogel we constructed, we first successfully isolated T cells from mouse peripheral blood using a T cell isolation kit. Next, we labelled MB49 cells with mCherry (red) and T cells with carboxyfluorescein succinimidyl ester (CFSE, green). Bladder cancer cells treated with the hydrogel loaded with recombinant oncolytic adenovirus were co-cultured with labelled T cells (Fig. [149]8A). Immunofluorescence staining data revealed that co-culture of bladder cancer cells treated with Adv-CRB3@gel exhibited a significant increase in the expression of activation markers CD25 and CD69 in T cells, surpassing levels seen in both Adv-Ctrl@gel and Adv-CRB3 groups. Conversely, Adv-Ctrl and Adv-Ctrl@gel were found to activate T cells to some extent by releasing GM-CSF (Fig. [150]8B-C). Fig. 8. [151]Fig. 8 [152]Open in a new tab Interactions between recombinant oncolytic Adv-treated bladder cancer cells and T cells. Note (A) Transwell system used to study virus release from silk gel for in vitro infection of cancer cells; (B-C) Immunofluorescence detection of CD25 and CD69 levels in MB49 cells, with images shown at Bar = 25 μm and statistical graph for positive cells in Image C; (D-E) Detection after co-culture of T (CFSE) cells and MB49 (mCherry) cells for 12 h, with images shown at Bar = 25 μm and statistical graph for migrated cells in Image E, binding rate of T (CFSE) cells to MB49 (mCherry) cells also measured (F); (G-H) Penetration of CFSE-labeled CD8^+ T cells in 3D cancer spheroids MCS observed by confocal microscopy, with images shown at Bar = 50 μm and statistical graph for average CFSE fluorescence intensity (FL) in 3D cancer spheroids MCS (H); (I) Concentration of GM-CSF, IFN-γ, and TNF-α detected by ELISA; (J-K) Levels of PD-L1 and CTLA-4 measured by flow cytometry, with statistical graph for positive cells in Image K. All cell experiments were repeated three times, and the values are expressed as mean ± standard deviation. P < 0.05, P < 0.01, and ** indicate P < 0.001 In the Transwell experiment, the number of T cells in the lower chamber was observed using fluorescence microscopy, and confocal imaging was used to explore the interaction between bladder cancer cells and T cells. The results revealed that the Adv-CRB3@gel treatment group had the highest number of migrating T cells compared to Adv-Ctrl@gel and Adv-CRB3 groups, suggesting that bladder cancer cells treated with the hydrogel loaded with recombinant oncolytic adenovirus released specific chemokines or signals to recruit T cells towards them (Fig. [153]8D-F). To further investigate the recruitment capability of bladder cancer cells on T cells, we co-cultured CFSE-labeled T cells with MB49 cells to form 3D cancer spheroids. The results revealed that the infiltration rate of T cells into the 3D cancer spheroids was highest in the Adv-CRB3@gel group, while Adv-Ctrl@gel also significantly increased the infiltration of T cells into the 3D cancer spheroids (Fig. [154]8G-H). The expression of multiple cytokines in the co-culture system was measured using ELISA assay kits. The results showed that the concentrations of GM-CSF, IFN-γ, and TNF-α were highest in the Adv-Ctrl@gel group, suggesting an increase in the secretion of pro-inflammatory cytokines by T cells after co-cultivation with bladder cancer cells treated with the hydrogel loaded with recombinant oncolytic adenovirus. This may be related to the activation of T cells and their interaction with bladder cancer cells (Fig. [155]8I). Furthermore, flow cytometric analysis of isolated bladder cancer cells using fluorescently labelled anti-PD-L1 and anti-CTLA-4 antibodies was performed. The results showed that compared to the Adv-Ctrl@gel and Adv-CRB3 groups, the expression levels of PD-L1 and CTLA-4 were significantly reduced in the Adv-CRB3@gel group, suggesting suppression of the tumor cells’ ability to evade the immune system (Fig. [156]8J-K). The above experimental results indicate that CRB3 can synergistically increase the recruitment ability of T cells with GM-CSF. Additionally, the hydrogel loaded with the recombinant oncolytic adenovirus maximizes T cell activity and significantly inhibits the immune evasion ability of tumor cells. Anti-tumor efficacy and immune activation effects of Adv-CRB3@gel in a subcutaneous bladder cancer mouse model Based on the aforementioned findings, we further investigated the anti-tumor potential of Adv-CRB3@gel in a mouse model of subcutaneous bladder cancer. We constructed the mouse model using bioluminescently labelled MB49 bladder cancer cells, successfully forming tumors after 7 weeks. Subsequently, we treated these tumors with different hydrogel formulations (Fig. [157]9A). The results demonstrated that Adv@gel, following subcutaneous injection of Adv or Adv@gel, effectively prevented the viral spread and infection of healthy tissues. This was corroborated by the levels of Adv genomic DNA in the nervous system, blood, and peripheral organs (Figure [158]S10A). Biochemical and routine blood tests in mice showed low toxicity of Adv@gel (Figure [159]S10B). H&E images of major organs indicated minimal damage (Figure [160]S10C). Thus, we confirmed that hydrogels exhibit lower toxicity. Fig. 9. [161]Fig. 9 [162]Open in a new tab Study on the anti-tumor potential of Adv-CRB3@gel in a subcutaneous bladder cancer mouse model. Note (A) Schematic representation of silk hydrogel injection of recombinant oncolytic Adv-loaded tumor-bearing mice; (B-C) Representative bioluminescence imaging (BLI) images, with image C showing semi-quantitative statistical graph of signals; (D) Relative tumor volume changes in each group of mice; (E) Schematic representation of subcutaneous tumors in each group of mice; (F) H&E staining showing tumor tissue histopathology, Bar = 50 μm; (G-H) Immunohistochemical detection of Ki67 and PCNA expression in tumor tissue, with arrows indicating tumor cell nuclei, Bar = 50 μm, and image H showing statistical graph of positive cells; (I-J) Immunofluorescent detection of CD25 and CD69 expression in tumor tissue, Bar = 100 μm, with image J showing statistical graph of positive cells; (K-L) Flow cytometry detection of Treg + CD4 + T cells and TNF-γ^+CD8^+ T cell numbers in tumor tissue, with L being the statistical graph; (M-N) Western blot detection of PD-L1 and CTLA-4 expression in tumor tissue, with N being the statistical graph. Each group consisted of 6 mice, and values are presented as mean ± standard deviation. ns indicates non-significant, * indicates P < 0.05, ** indicates P < 0.01, *** indicates P < 0.001 In addition, we investigated the biodistribution of the virus after intratumoral injection. At different time points post-injection, we isolated the tumors and major visceral organs for ex vivo fluorescence imaging. One day after injection, the free Adv-CRB3 was evenly distributed throughout the tumor area, indicating rapid diffusion from the injection site. Besides the tumor site, a significant amount of Adv-CRB3 fluorescence signal was also detected in the liver. However, the distribution of Adv-CRB3@gel was markedly different; it was present only in part of the tumor, with a strong fluorescence signal around the injection site due to the slow release of Adv-CRB3 from the silk gel. Three days later, the signal in the tumors treated with free Adv-CRB3 sharply declined and was significantly lower compared to the Adv-CRB3@gel group. By extending the observation time to 5 days, no viral fluorescence signal was detected in the tumor tissues or major organs of the mice treated with free Adv-CRB3, suggesting clearance of these free viruses from the body. In contrast, the fluorescence signal in the tumor site of the Adv-CRB3@gel group remained strong even after 5 days of injection. These results validate our hypothesis that the silk gel can achieve sustained viral release in vivo, thereby reducing its leakage into other tissues (Figure [163]S10D). By measuring the bioluminescent signals of the tumors, we observed the growth of fluorescent signals at 0, 3, and 5 weeks after hydrogel treatment and found that Adv-CRB3@gel significantly slowed down the signal growth compared to the control group (Fig. [164]9B-C). Actual tumor volume measurements also confirmed the pronounced anti-tumor effect of Adv-CRB3@gel (Fig. [165]9D). After 5 weeks of treatment, we excised the tumors from the mice and observed that the hydrogel-treated group had the smallest tumors (Fig. [166]9E). Furthermore, we performed pathological analysis on the tumor tissues to validate the therapeutic effects after different treatments. H&E staining in histopathology revealed that mice tumors in the control and adenovirus groups displayed typical malignant tumor features, including high cell density, disorganized cell arrangement, noticeable nuclear pleomorphism, and pathological nuclear division. However, after treatment with Adv-CRB3@gel, these malignant characteristics significantly improved, such as reduced tumor cell count and more regular cell arrangement (Fig. [167]9F). We investigated the proliferation markers Ki67 and PCNA in tumor tissues through immunohistochemical analysis. The results demonstrate a significant decrease in the positive expression of Ki67 and PCNA in tumor tissues of mice treated with Adv-CRB3@gel compared to the control group (Fig. [168]9G-H). To further explore the immune infiltration status in the tumor microenvironment, we quantified the expression of CD25 and CD69 in tumor tissues using immunofluorescence technology. The results reveal a significant enhancement in the expression of these two markers in mice treated with Adv-CRB3@gel, compared to other groups, suggesting an improved immune activation status (Fig. [169]9I-J). In the tumor microenvironment, Treg cells can be exploited by tumors to evade the immune system’s attack, thereby promoting tumor growth and dissemination [[170]37]. On the other hand, TNF-γ can enhance the cytotoxic function of CD8^+ T cells, facilitating their recognition and elimination of tumor cells [[171]38]. We further assessed the differences in Treg cells and TNF-γ^+CD8^+ T cells among the groups. Remarkably, Adv-Ctrl@gel treatment significantly reduced Treg infiltration while increasing the number of TNF-γ^+CD8^+ T cells (Fig. [172]9K-L). This strongly suggests that Adv-Ctrl@gel can enhance the adaptive immune response against tumors. To further explore the underlying mechanisms, we utilized Western blotting to measure the expression levels of the immune checkpoint proteins PD-L1 and CTLA-4 in tumor tissues. Encouragingly, compared to the Adv-Ctrl@gel group, the expression of PD-L1 and CTLA-4 was significantly downregulated in the Adv-CRB3@gel group, indicating effective disruption of the tumor’s immune escape mechanisms (Fig. [173]9M-N). The above experiments demonstrate the potent tumor growth inhibition capabilities of the recombinant oncolytic adenovirus silk gel Adv-CRB3@gel in a mouse model of subcutaneous bladder cancer. Compared to the use of adenovirus alone, it exhibits better therapeutic effects and lower toxicity. Furthermore, it helps GM-CSF to significantly enhance the immune response in the tumor microenvironment, thereby effectively inhibiting tumor immune escape. Adv-CRB3@gel combined with PD-L1 inhibitor enhances bladder cancer treatment: T cell infiltration and tumor cell necrosis Immune checkpoint blockade (ICB) improves the prognosis of patients with advanced bladder cancer who are resistant to treatment, particularly with programmed cell death-1 (PD-1)/programmed cell death ligand-1 (PD-L1) targeted antibodies that reactivate the anti-tumor immune response. This approach is considered innovative in the treatment of refractory advanced MIBC patients [[174]39]. However, the use of PD-L1 as a single biomarker is controversial due to its dynamically regulated expression [[175]40]. The recombinant oncolytic adenovirus hydrogel has demonstrated effective oncolytic effects and moderate recruitment of T cells. Therefore, we hypothesized that combining Adv-CRB3@gel with PD-L1 inhibitors could enhance T cell aggregation and increase the anti-tumor effect. We conducted the following injections in tumor-bearing mice: PBS + IgG, Adv-CRB3@gel + Anti-PD-L1, PBS + Anti-PD-L1, Adv-CRB3@gel + IgG. After four weeks of observation, we found that the tumor volume in the Adv-CRB3@gel + Anti-PD-L1 treatment group was significantly lower compared to the PBS + Anti-PD-L1 and Adv-CRB3@gel + IgG groups, and the tumor volume had decreased compared to pre-treatment (Fig. [176]10A-E). Fig. 10. [177]Fig. 10 [178]Open in a new tab Evaluation of the immune-enhancing effect of Adv-CRB3@gel combined with PD-L1 inhibitor in bladder cancer. Note (A) Schematic representation of the combined treatment of bladder cancer in mice; (B-C) Representative bioluminescence imaging (BLI) images, with image C showing a semi-quantitative statistical graph of signals; (D) Relative changes in tumor volume in each group of mice; (E) Schematic representation of subcutaneous tumors in each group of mice; (F) H&E staining showing histopathology of tumor tissue, Bar = 50 μm; (G-H) Immunofluorescent detection of CD25 and CD69 expression in tumor tissue, upper image Bar = 50 μm, the lower image is an enlarged view of the white box in the upper image, and image H shows a statistical graph of positive cells; (I-J) Flow cytometry detection of Treg cells and TNF-γ^+CD8^+ T cell numbers in tumor tissue, with J being the statistical graph. Each group consisted of 6 mice, and values are presented as mean ± standard deviation. * indicates P < 0.05, ** indicates P < 0.01, *** indicates P < 0.001 Using H&E staining to examine tumor tissues, we found a substantial area of cell necrosis in the tumors of the Adv-CRB3@gel + Anti-PD-L1 treatment group, indicating its significant therapeutic effect (Fig. [179]10F). By utilizing immunofluorescence to measure the expression levels of CD25 and CD69 in tumor tissues, we observed significantly higher expression levels of CD25 and CD69 in the combination treatment group compared to the single treatment groups (Fig. 10G-H). Flow cytometry also confirmed a decrease in Treg cells and a significant increase in TNF-γ^+CD8^+ T cells in the combination treatment group (Fig. [180]10I-J). This suggests that the combination of PD-L1 inhibitors and recombinant oncolytic Adv-CRB3@gel synergistically enhances T cell infiltration, increasing their killing capacity against bladder cancer. Discussion Bladder cancer is a common malignancy that presents numerous challenges in treatment, including surgical limitations, limited effectiveness of chemotherapy, and the development of drug resistance [[181]28, [182]41–[183]43]. Therefore, the search for new treatment strategies has become particularly important. Immunotherapy, as an emerging approach, aims to inhibit tumor growth and metastasis by activating the patient’s own immune system [[184]44, [185]45]. In recent years, the successful application of immune checkpoint inhibitors has brought new hope to bladder cancer treatment [[186]46–[187]48]. A genetically engineered adenovirus was developed as a vector to deliver CRISPR/Cas9 (sgCas9-AdV) for permanent PD-L1 gene editing, successfully inhibiting the growth of Hepa 1–6 liver cancer cells, PD-L1-resistant CT26 colon tumors, and highly invasive orthotopic 4T1 mouse breast tumors [[188]19]. Building on this study, our research employed a novel combination therapy strategy by co-administering recombinant oncolytic Adv-loaded silk hydrogel with a PD-L1 inhibitor. This combination therapy offers two main advantages. First, the recombinant oncolytic adenovirus exhibits strong tumor-lytic capabilities, directly killing cancer cells. Second, the PD-L1 inhibitor lifts the suppression of immune attacks on tumor cells, enhancing the immune response. Compared to monotherapy, combination therapy more effectively activates the immune system, enhancing tumor antigen-specific immune responses [[189]49–[190]51]. Moreover, recent studies have shown that novel strategies combining tumor sensitization and suppression of MDSC recruitment can significantly improve the effectiveness of radiotherapy, providing new avenues for enhancing anti-tumor immune responses (10.1002/adfm.202000189). Previous studies have indicated that CRB3 serves as an independent prognostic factor for clear cell renal cell carcinoma patients and may represent a potential broad-spectrum tumor suppressor gene [[191]52]. Crb3 significantly accelerates cell migration in colon cancer cells by interacting specifically with FGFR1, thereby promoting invasion and metastasis of human colon cancer [[192]53]. It has been observed that CRB3 inhibits the stemness of tamoxifen-resistant cells by suppressing the β-catenin signaling pathway, leading to a decrease in the number of breast cancer cells [[193]54]. This study, for the first time, demonstrates that CRB3 can inhibit the proliferation and migration of bladder cancer cells. The critical aspect of tumor immunotherapy lies in the recruitment and activation of T cells [[194]55–[195]57]. In this research, the nanocomposite hydrogel loaded with recombinant oncolytic adenovirus encoding CRB3 and GM-CSF not only significantly inhibits bladder cancer but also recruits T cells, effectively promoting their activation. The role of PD-L1 inhibitors in cancer treatment has garnered widespread attention [[196]58–[197]60]. When used in combination with the hydrogel and recombinant oncolytic adenovirus, the activation of T cells and their cytotoxicity against tumor cells are markedly enhanced. This discovery opens up a new direction for immunotherapeutic strategies in cancer treatment (Fig. [198]11). Fig. 11. [199]Fig. 11 [200]Open in a new tab Molecular mechanism of enhanced anti-bladder cancer effect by injection of recombinant oncolytic adenovirus encoding CRB3 and GM-CSF in silk hydrogel combined with PD-L1 inhibitor to recruit T cells The findings of this study demonstrate that the silk fibroin hydrogel containing the PD-L1 inhibitor and recombinant oncolytic adenovirus can synergistically enhance T cell infiltration and increase the cytotoxicity against bladder cancer. Previous research has not investigated the combined application of Adv-CRB3@gel and PD-L1 inhibitors in bladder cancer patients. However, there have been studies on the use of gel carriers in combination with PD-L1 for cancer treatment, such as the gel system DOX/ICG/CpG-P-ss-M/CD based on α-cyclodextrin (CD), encapsulating doxorubicin (DOX) and the photothermal agent indocyanine green (ICG). Following irradiation, the gel system exhibits a thermal responsive release of DOX and vaccine-like nano-particles CpG-P-ss-M, as well as the in-situ storage of tumor-specific antigens generated by chemotherapy and phototherapy. The released CpG-P-ss-M acts as a carrier to adsorb and deliver antigens to lymph nodes, promoting antigen uptake by dendritic cells and dendritic cell maturation. Combined with PD-L1 blockade, this therapy effectively inhibits primary tumor growth, induces tumor-specific immune responses, and prevents tumor recurrence and metastasis [[201]61]. The bee venom-derived-(RADA) (MR) hydrogel scaffold, loaded with the specific Ca/calmodulin-dependent protein kinase II (CAMKII) inhibitor KN93, possesses characteristics of direct anti-tumor effects, controlled drug release, and enhancement of intracellular cargo uptake. The combination of anti-PD-L1 therapy achieved a cure rate of approximately 30% in malignant ascites in a mouse model, demonstrating promising potential as an anti-tumor application platform [[202]62]. This study holds significant scientific value. Through the adoption of single-cell sequencing and machine learning methods, this study identified the key gene CRB3 and confirmed its inhibitory effect on bladder cancer cell proliferation, migration, and invasion. This research further emphasizes the importance of CRB3 in bladder cancer and provides a theoretical foundation for treatment strategies. In clinical practice, the combination application of Adv-CRB3@gel and PD-L1 inhibitors significantly enhances T cell recruitment and tumor inhibition. This offers a new option for improving the treatment effectiveness of bladder cancer and holds important clinical implications. However, this study does have limitations. First, the findings are based solely on mouse models and in vitro cell experiments, without validation in clinical patients. Therefore, the feasibility of clinical translation requires further investigation. Second, the drug loading and release efficiency of Adv-CRB3@gel need further validation and optimization, and its production process requires exploration for improvement, especially the large-scale production of silk hydrogel and the refinement of viral encapsulation methods [[203]63, [204]64], to enhance therapeutic efficacy and ensure treatment consistency. Additionally, although this study demonstrated the efficacy of the combination therapy, its specific therapeutic effects and side effects still need to be evaluated through more clinical trials and long-term follow-up. Future research could further explore the regulatory mechanisms of CRB3 and investigate its interactions with other signaling pathways, offering deeper insights into the role of CRB3 in the development and progression of bladder cancer and providing more potential targets for precision therapy. Given the importance of PD-L1 inhibitors in immunotherapy, further studies could explore the optimal combination of different PD-L1 inhibitors and dosages with Adv-CRB3@gel to further enhance treatment efficacy. The combination effects of Adv-CRB3@gel with other immune checkpoint inhibitors, such as CTLA-4 or TIGIT inhibitors, could also be explored. These immune checkpoints are expressed in tumor cells, stromal cells, and immune cells, promoting the generation of Tregs and myeloid-derived suppressor cells and enhancing their suppressive functions, thereby reinforcing the immunosuppressive environment within tumors. This includes CD73 [[205]65], which is expressed in human tumor cell lines such as carcinoma, melanoma, and sarcoma, and regulates T cell activation and plays a crucial role in the suppressive function of regulatory T (Treg) cells; CTLA-4 [[206]66], which regulates T cell activation and is vital for Treg cell suppressive function; and TIGIT [[207]67], an inhibitory receptor that activates T cells, natural killer (NK) cells, and Tregs. Moreover, preclinical studies, including further validation experiments in animal models and drug safety assessments, should be conducted. These will help to verify the therapeutic effects and potential side effects of Adv-CRB3@gel in the treatment of human bladder cancer. In conclusion, this study provides a novel strategy for the combination therapy of bladder cancer, but further experimental and clinical research is needed to verify its application prospects in humans. Through continuous improvement of study design and more in-depth mechanism research, this combination therapy strategy has the potential to become one of the more effective methods for bladder cancer treatment. Electronic supplementary material Below is the link to the electronic supplementary material. [208]12951_2024_2903_MOESM1_ESM.docx^ (13.5KB, docx) Supplementary Material 1: Histological changes observed at different stages of BBN treatment, with 10 mice in each group; (C) H&E staining results of mouse bladder tissue at different time points, upper image Bar=100μm, lower image Bar=25μm; (D) Image of bladder cancer specimen. [209]12951_2024_2903_MOESM2_ESM.jpg^ (231.3KB, jpg) Supplementary Material 2: Figure S2. Quality control, filtering, and principal component analysis of scRNA-seq data. Note: (A) Violin plots showing the gene count (nFeature_RNA), mRNA molecule count (nCount_RNA), and mitochondrial gene percentage (percent.mt) for each cell in scRNA-seq data; (B) Scatter plots showing the correlation between filtered data nCount_RNA and percent.mt, and nCount_RNA and nFeature_RNA; (C) Variances analysis used to select highly variable genes, where red represents the top 2000 highly variable genes and black represents low-variable genes, with the top 10 gene names annotated; (D) Heatmap of the top 20 genes with the highest correlation with PCs 1-6 in PCA, where yellow indicates upregulation and purple indicates downregulation; (E) Distribution of cells before batch correction in PC_1 and PC_2, where each point represents a cell; (F) Distribution of cells after Harmony batch correction in PC_1 and PC_2, where each point represents a cell. [210]12951_2024_2903_MOESM3_ESM.jpg^ (295.7KB, jpg) Supplementary Material 3: Figure S3. Cell annotation. Note: (A) Visualization of UMAP clustering results showing cell clustering and distribution, with each colour representing a cluster; (B) UMAP plot based on six cell marker genes; (C) Visualization of UMAP clustering results showing cell clustering and distribution among the five samples, with different colours representing different samples; (D) Proportion of different cell subtypes in each sample, shown with different colours representing different cell subtypes; (E) Proportion of different cell subtypes in each sample after malignant cell annotation, shown with different colours representing different cell subtypes. [211]12951_2024_2903_MOESM4_ESM.jpg^ (292.8KB, jpg) Supplementary Material 4: Figure S4. WGCNA Analysis of Differential Genes in the scRNA-seq Dataset. Note: (A) Scale independence, average connectivity, and scale-free topological structure graph were used to determine the weighted value β=3 that satisfies the scale-free network law; (B) Dendrogram of co-expression network modules; (C) Correlation analysis between the modules and normal/tumor groups; (D) Scatter plot analysis of the Brown, Pink, Blue, Black modules, and Tumor. [212]12951_2024_2903_MOESM5_ESM.jpg^ (173.8KB, jpg) Supplementary Material 5: Figure S5. Kaplan-Meier curves of 30 genes significantly correlated with survival in BLCA patients. Note: Genes with higher expression associated with poorer survival are marked with a red circle, while genes with lower expression associated with poorer survival are marked with a green circle. [213]12951_2024_2903_MOESM6_ESM.jpg^ (715.1KB, jpg) Supplementary Material 6: Figure S6. Revealing the key genes associated with bladder cancer and their prognosis evaluation. Note: (A) Using risk scores generated from LASSO analysis, Kaplan-Meier curves were created to compare the prognostic differences between high-risk and low-risk groups; (B-C) Based on the risk scores (RS) and different clinical-pathological parameters, a Nomogram chart was constructed (B) and the predicted survival probabilities at 1, 3, and 5 years were compared with the actual survival probabilities (C); (D) Decision curve analysis (DCA curve) was performed to plot the 1, 3, and 5-year overall survival rates, where the horizontal axis represents the threshold probability - the proportion of samples above a certain value after grouping samples based on it compared to the total number of samples, and the vertical axis is the net benefit calculated by subtracting true positive results from false positive results. None and ALL are two reference lines; the closer the model curve is to these lines, the lower the application value. A higher vertical axis value for the same horizontal coordinate indicates a better model. Within a larger range of horizontal threshold values, if the vertical axis value is consistently higher compared to other models, it indicates that the model is superior. [214]12951_2024_2903_MOESM7_ESM.jpg^ (182.5KB, jpg) Supplementary Material 7: Figure S7. Violin plots of the expression of 15 genes significantly associated with BLCA survival in scRNA-seq data. Note: Violin plots of the expression of 15 genes significantly associated with BLCA survival in normal epithelial cells and malignant epithelial cells in the scRNA-seq dataset. Genes with expression trends opposite to survival analysis are indicated by blue boxes. [215]12951_2024_2903_MOESM8_ESM.jpg^ (2.6MB, jpg) Supplementary Material 8: Figure S8. UMAP plot showing 10 genes significantly associated with TCGA-BLCA prognosis in the scRNA-seq dataset, with shades of blue indicating the level of average expression. [216]12951_2024_2903_MOESM9_ESM.jpg^ (281.1KB, jpg) Supplementary Material 9: Figure S9. RT-qPCR Analysis of CRB3 Expression. Note: (A) RT-qPCR results showing the expression levels of CRB3 in BECs, MB49, and MBT-2 cell lines; (B) Results indicating the knockdown efficiency of CRB3 in MB49 cells; (C) Results demonstrating the overexpression efficiency of CRB3 in MBT-2 cells. Cell experiments were repeated three times, with results presented as mean ± standard deviation. Significance levels are denoted as * for P < 0.05, ** for P < 0.01, and *** for P < 0.001. [217]12951_2024_2903_MOESM10_ESM.jpg^ (458.6KB, jpg) Supplementary Material 10: Figure S10. Safety Assessment [218]Supplementary Material 11^ (12.2MB, jpg) Acknowledgements