Targeting MAN1B1 potently enhances bladder cancer antitumor immunity via deglycosylation of CD47

Abstract

Background

Only a few bladder cancer patients benefit from anti‐programmed cell death protein 1/programmed cell death ligand 1 immunotherapy. The cluster of differentiation 47 (CD47) plays an important role in tumor immune evasion. CD47 is a highly glycosylated protein, however, the mechanisms governing CD47 glycosylation and its potential role in immunosuppression are unclear. Therefore, this study aimed to evaluate the function of CD47 glycosylation in bladder cancer.

Methods

Western blotting, immunohistochemistry, and flow cytometry were used to measure protein expression, protein‐protein interactions, and phagocytosis in bladder cancer. A murine model was employed to investigate the impact of mannosidase alpha class 1B member 1 (MAN1B1) modification of CD47 on anti‐phagocytosis in vivo. An ex vivo model, patient‐derived tumor‐like cell clusters, was used to examine the effect of targeting MAN1B1 on phagocytosis.

Results

Our research identified that aberrant CD47 glycosylation was responsible for its immunosuppression. The glycosyltransferase MAN1B1 responsible for CD47 glycosylation was highly expressed in bladder cancer. Abnormal activation of extracellular signal‐regulated kinase (ERK) was significantly associated with MAN1B1 stability by regulating the interaction between MAN1B1 and the E3 ubiquitin ligase HMG‐CoA reductase degradation 1 (HRD1). Mechanistically, abnormally activated ERK stabilized MAN1B1, resulting in the glycosylation of CD47 and facilitating immune evasion by enhancing its interaction with signal‐regulatory protein alpha (SIRP‐α). In vitro and in vivo experiments demonstrated that MAN1B1 knockout weakened CD47‐mediated anti‐phagocytosis. MAN1B1 inhibitors promoted phagocytosis without causing anemia, offering a safe alternative to anti‐CD47 therapy.

Conclusions

This comprehensive analysis uncovered that ERK activation stabilizes MAN1B1 by regulating the interaction between MAN1B1 and HRD1, facilitates immune evasion via CD47 glycosylation, and presents new potential targets and strategies for cancer immunotherapy that do not cause anemia.

List of abbreviations

B7H3

B7 homolog 3

BLCA

Bladder cancer

BMDMs

Bone marrow‐derived macrophages

CCK8

Cell Counting Kit‐8

CD147

cluster of differentiation 147

CD47

cluster of differentiation 47

CHX

cycloheximide

CRISPR

Clustered Regularly Interspaced Short Palindromic Repeats

DHT I

dihydrotanshinone I

DMEM

dulbecco's modified eagle medium

DMSO

Dimethyl sulfoxide

ECD

extracellular domain

ER

endoplasmic reticulum

ERAD

endoplasmic reticulum‐associated degradation

ERK

extracellular signal‐regulated kinase

ERK2

extracellular signal‐regulated kinase 2

FBS

fetal bovine serum

GANAB

glucosidase II alpha subunit

GAPDH

Glyceraldehyde‐3‐phosphate dehydrogenase

GFP

green fluorescent protein

GM130

Golgi matrix protein 130

GRP94

Glucose‐Regulated Protein 94

HEK293T

Human embryonic kidney 293T cells

HRD1

HMG‐CoA reductase degradation 1

IF

Immunofluorescence staining

IHC

Immunohistochemical staining

IP

immunoprecipitation

KO

knockout

KRAS‐G12D

Kirsten Rat Sarcoma Viral Oncogene Homolog G12D

LCA

Lens culinaris agglutinin

MAN1B1

mannosidase alpha class 1B member 1

MFI

mean fluorescence intensity

MGAT2

monoacylglycerol acyltransferase 2

MOGS

mannose oligosaccharide glucosidase

N

asparagine

NSG

NoD.Cg.Prkdc^scidIl2rg^em1Smoc

NXT

N‐glycosylation

OD450

optical density at 450 nanometers

PBMCs

Peripheral blood mononuclear cells

PBS

Phosphate‐Buffered Saline

PNGase F

peptide‐N‐glycosidase F

PRAD

prostate adenocarcinoma

PTCs

patient‐derived tumor‐like cell clusters

Q

glutamine

QPCTL

glutaminyl‐peptide cyclotransferase‐like protein

qRT‐PCR

Quantitative reverse transcription PCR

RIPA

radio immunoprecipitation assay buffer

SD

standard deviation

SDS‐PAGE

Sodium Dodecyl Sulfate‐Polyacrylamide Gel Electrophoresis

SIRP‐α

surface ligand signal‐regulatory protein alpha

STAT3

Signal transducer and activator of transcription 3

SYVN1

synoviolin 1

TAM

tumor‐associated macrophages

TCGA

The Cancer Genome Atlas

WT

wild‐type

1. Background

Repairing the immune system in the tumor microenvironment is crucial for cancer treatment. Tumor cells can often achieve immune evasion through the loss of antigenic substances and the expression of anti‐immune molecules [1]. Tumors can evade the immune system by utilizing the cluster of differentiation 47 (CD47) present on their cell surface [2].

CD47, also known as integrin‐associated protein, consists of a single extracellular V‐set Immunoglobulin Superfamily domain, a presenilin domain with five membrane‐spanning sections, and a short cytoplasmic domain that is subject to alternative splicing, thereby giving rise to four isoforms [3, 4]. CD47 releases a “don't eat me” signal to macrophages by binding to the macrophage surface ligand signal‐regulatory protein alpha (SIRP‐α) [5, 6]. CD47 is significantly present in various tumor cells, including those found in acute myeloid leukemia [7], glioblastoma [8], as well as breast [9], non‐small cell lung cancer and colon cancer [10]. Blocking CD47 suppresses cancer growth in vivo [11]. CD47 monoclonal antibodies are currently being studied in clinical research on leukemia and solid tumors [12]. Despite certain benefits of CD47 monoclonal antibodies, widespread expression of CD47 in normal cells, including erythrocytes and platelets, limits its clinical application for causing adverse effects such as anemia [13, 14]. Therefore, understanding the regulatory mechanism of CD47 and exploring alternative therapeutic approaches are promising strategies to overcome these challenges and improve clinical outcomes.

CD47 expression is regulated by genomic alteration, transcriptional regulation, and post‐transcriptional and post‐translational modifications. Transcriptional regulation of CD47 has been extensively studied. Moreover, v‐myc myelocytomatosis viral oncogene homolog [15], Hypoxia‐inducible factor 1 [16], and Nuclear Factor kappa‐light‐chain‐enhancer of activated B cells increase CD47 transcription [17]. The Phosphoinositide 3‐kinase/Signal transducer and activator of transcription 3 (STAT3)/microRNA‐34a signal axis mediated regulation of CD47 expression is also observed in lung cancer [18]. Regarding post‐translational modifications of CD47, the enzyme glutaminyl‐peptide cyclotransferase‐like protein (QPCTL) is a major modifier [19]. The recent review highlights the role of CD47 in non‐small cell lung cancer [20]. However, the function and clinical significance of CD47 in bladder cancer remain unclear [21]. Consequently, we investigated whether CD47 played a key immune role in bladder cancer.

This study aimed to explore the role of CD47 in bladder cancer and to identify the glycosyltransferase responsible for the glycosylation modification of CD47. We utilized both in vivo and in vitro models to investigate the immunosuppressive impact of CD47 glycosylation in bladder cancer.

2. Materials and Methods

2.1. Cell lines and culture

Human embryonic kidney 293T cells (HEK293T), human bladder cancer cell lines J82, T24 and 5637 and murine bladder cancer cell line MB49 cells were cultured in dulbecco's modified eagle medium (DMEM) or RPMI 1640 (C11995500BT, C11875500BT, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS) (A5670701, Gibco, Burlington, MA, USA), 1% penicillin and 1% streptomycin (BL505A, Biosharp, Hefei, Anhui, China) at 37°C under 5% CO₂. All cells tested negative for mycoplasma infection. All the cell lines used in the present study were verified by short tandem repeat genotyping prior to the start of the experiments.

2.2. Vector generation

Human full‐length CD47 transcript variant 2 (RefSeq: NM_198793.3) was cloned into pLVX‐Puro (pLVX) (632164, Clontech, Mountain View, CA, USA) lentiviral expression vectors containing the long 3′ untranslated region. CD47‐FLAG mutants were generated by site‐directed mutagenesis, substituting asparagine (N) with glutamine (Q) at each of the five specific sites (N23, N34, N50, N73, and N111) or at all five sites simultaneously. These mutants included single‐site mutants N23Q, N34Q, N50Q, N73Q, and N111Q, as well as a multi‐site mutant 5NQ (N23Q, N34Q, N50Q, N73Q, N111Q). These CD47‐FLAG mutants were used to evaluate the glycosylation sites of CD47, to assess the affinity between CD47 and SIRP‐α, and to perform phagocytosis assays. All constructs were confirmed by DNA sequencing. Human full‐length mannosidase alpha class 1B member 1 (MAN1B1), mannose oligosaccharide glucosidase (MOGS), extracellular signal‐regulated kinase 2 (ERK2), and Kirsten Rat Sarcoma Viral Oncogene Homolog G12D (KRAS‐G12D) templates were obtained from Miaoling Biology (P53006, P30176, P16206, P61757, Miaoling Biology, Wuhan, Hubei, China). MAN1B1, MOGS, ERK2, and KRAS‐G12D were cloned into pLVX lentiviral expression vectors, as described above.

2.3. Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/CRISPR associated protein 9‐mediated generation of knockout (KO) cells

To construct CD47‐KO cell lines, the human and mouse sgRNA sequences were synthesized: human sgCD47 #1, 5‐TTGCACTACTAAAGTCAGTG‐3, #2, 5‐CTTGTTTAGAGCTCCATCAA‐3, #3, 5‐AGCAACAGCGCCGCTACCAG‐3, #4, 5‐ATCGAGCTAAAATATCGTGT‐3, and sgControl, 5‐TTCCTGCCACGTTGCGCAG‐3; mouse sgCd47 #1, 5‐CCCTTGCATCGTCCGTAATG‐3, #2, 5‐CCACATTACGGACGATGCAA‐3, #3, 5‐ATAGAGCTGAAAAACCGCAC‐3. Then, sgRNA was constructed into LentiCRISPR V2 plasmid by molecular cloning. After plasmid extraction, cells were transfected with plasmid (strategy: #1 + #3, #2 + #3, and # 4) using lipofectamine 3000 (L3000015, Thermo Fisher Scientific). Puromycin (2 µg/mL, ST551, Beyotime, Shanghai, China) was added for 48 h for the selection of transfected cells. After selection, the monoclonal cell lines were obtained by dilution plating at a density of 50 cells per 96‐well plate, ensuring that most wells contained a single cell. The knockout cells were verified using Western blotting.

The MAN1B1, MOGS, ERK1/2 and HMG‐CoA reductase degradation 1 (HRD1) were knocked out as described above. Human sgMAN1B1 #1, 5‐ACCGAGATGAAGTCCCGATG‐3, #2, 5‐CGTCAGGAAGTCCGACTGAG‐3; Human sgMOGS #1, 5‐TTCATGCCGAAGTAGACGTG‐3; Human sgERK1 #1, 5‐ATGGACTTGGTATAGCCCTG‐3; Human sgERK2 #1, 5‐GCCTACAGACCAAATATCAA‐3 and Human sgHRD1 #1, 5‐GATGGCTCGGCGAGACATGA‐3.

2.4. Lentivirus packaging and infection

A total of 5 × 10⁶ 293T cells in good state and logarithmic growth stage were cultured in a 10 cm dish. When the cell confluence reached 60%‐80%, the lentivirus core and the packaging plasmids were co‐transfected. The medium was changed 8‐10 h after transfection. After 48 and 72 h, the virus supernatant was collected. The cells and cell fragments were removed by centrifugation at 2,000 ×g at 4°C for 10 min to concentrate the virus and the concentrate was then stored at ‐80°C for long‐term storage.

For construction of stable overexpression infection cell lines, the cells were infected with lentiviral particles expressing CD47‐WT or CD47‐5NQ for 48 h and positive stable cell lines were subsequently selected by puromycin for two days.

The shRNA sequences were inserted into the pLVX vector. Human glucosidase II alpha subunit (GANAB) shRNA: #1, 5‐GCTGTGGATAGAAGCAACTTT‐3, #2, 5‐CCCAACCTCTTTGTCTGGAAT‐3; Human monoacylglycerol acyltransferase 2 (MGAT2) shRNA: #1, 5‐GCCCAAATTGAGTCACTCTTA‐3.

2.5. Western blotting analysis and immunoprecipitation (IP)

For Western blotting, cells were collected and lysed in radio immunoprecipitation assay (RIPA) buffer (P0013C, Beyotime) supplemented with a complete protease inhibitor cocktail (11836170001, Roche, Basel, Kanton Basel‐Stadt, Switzerland). Each sample was sampled with 10‐30 µg and transferred to the nitrocellulose membrane. The final protein quantification was obtained using ImageJ software (https://imagej.net/software/fiji/downloads). For cycloheximide (CHX) experiments, cells were treated with CHX (50 µg/mL, HY‐12320, MedChemExpress, Monmouth Junction, NJ, USA) with indicated times before harvesting. For trametinib experiments, 5637 cells treated with trametinib (1 µmol/L, HY‐10999, MedChemExpress) for 24 h before harvesting.

For IP, cells were collected and lysed in IP buffer (P0013, Beyotime) supplemented with a complete protease inhibitor cocktail (Roche). Protein G beads (SM004025, Smart‐Lifesciences, Changzhou, Jiangsu, China) were incubated with the specified antibodies for 1 h at 4°C. Whole‐cell lysates were added to protein G beads. The mixture was incubated at room temperature for 1 h. For the co‐immunoprecipitation experiments assessing the affinity between CD47 and SIRP‐α, either CD47‐Fc or SIRP‐α‐Fc was used as the bait protein and incubated with Protein G beads. After the incubation, cell lysate was added and allowed to bind at room temperature for 1 h. The complexes were then eluted and detected by Western blotting. All Western blotting and IP antibodies are listed in Supplementary Table S1.

2.6. Clinical tissue specimens

The study included patients with bladder cancer, kidney cancer, and prostate cancer who were diagnosed and underwent surgical treatment at Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University between January 2022 and April 2024. Cancer tissues and adjacent non‐tumor tissues were collected from these patients. The inclusion criteria specified that patients were over 18 years of age, had indications for surgical treatment, and had no history of other malignancies. Vulnerable populations were excluded. Neoadjuvant therapy can be administered before radical cystectomy for bladder cancer according to the patient's preference. Therefore, whether chemotherapy was received before surgery was not included in the inclusion or exclusion criteria of this project. For patients with kidney cancer and prostate cancer, tissues from those who had received preoperative chemotherapy or radiotherapy were not included in this study. All patients were informed and consented, which was approved by the Ethics Committee of the Sir Run Run Shaw Hospital of Zhejiang University (No.20230312). All specimens were diagnosed by pathologists.

2.7. Immunohistochemical (IHC) staining

Fresh specimens obtained after surgery were fixed with 4% paraformaldehyde, fully dehydrated with ethanol, embedded with paraffin, and sliced into 4‐µm tumor tissue sections. The sections were placed in an incubator at 60°C for 1‐2 h until the paraffin was dissolved, dewaxed, and hydrated. After hydration, the antigen was repaired at high temperature for 15 min in pH = 9.0 ethylene diamine tetraacetic acid (EDTA) Antigen Retrieval Solution (P0085, Beyotime), and the cell membrane was broken using 0.1% Triton X‐100 (P0096, Beyotime) for 15 min. After cleaning with Phosphate‐Buffered Saline (PBS) 2‐3 times, sections were soaked in 3% hydrogen peroxide for 15 min to inactivate endogenous enzymes and enclosed with 5% FBS at room temperature for 1 h. Next, they were incubated with MAN1B1 primary antibody (1:100, #sc‐100543, Santa Cruz, Dallas, TX, USA) at 4°C overnight and cleaned with PBS 2‐3 times. After being incubated with the secondary antibody, the sections were developed with 3,3'‐diaminobenzidine (P0203, Beyotime) for 5‐10 min. The cell nuclei were then re‐stained, and the sections were stored in a sealed tablet. All of the IHC staining results were reviewed by a pathologist team blinded to the clinicopathological information. For evaluation of MAN1B1 staining, we adopted a staining score by the percentage of positive cells (0, <10%; 1, 10%‐30%; 2, 31%‐60%; 3, >60%) [22, 23, 24].

2.8. Immunofluorescence (IF)

Cells were fixed with 4% paraformaldehyde for 15‐20 min, then washed twice with PBS and permeabilized using 0.3% Triton X‐100 (Beyotime) for 15 min. After that, the cells were blocked with 5% serum (G1217, Servicebio, Wuhan, Hubei, China) for 1 h at room temperature. Primary antibodies were added and incubated at 4°C overnight. Finally, the cells were washed three times with PBS and incubated with secondary antibodies for 1 h at room temperature. Cell nuclei were stained with DAPI (5 µg/mL, D609734, Aladdin, Shanghai, China) for 15 min. After washing with PBS, the cells were mounted, and confocal images were captured using a laser scanning microscope. All IF antibodies are listed in Supplementary Table S1.

2.9. Tissue extraction

Clinical bladder cancer tissues were cut into pieces with sterilized scissors, ground with tissue grinding beads, and added to the RIPA buffer at 4°C for 60 min. DNA was separated by ultrasound. Then the sample was centrifuged at 12,000 ×g at 4°C for 20 min. The protein in supernatant was quantified using the bicinchoninic acid method. To acquire bladder cancer mRNA samples, tissue was treated in the same way as described above, and mRNA was extracted by TRIzol reagent (AG21101, Accurate Biology, Changsha, Hunan, China).

2.10. Glycosylation analysis of CD47

To confirm that CD47 was glycosylated, cell lysates were treated with peptide‐N‐glycosidase F (PNGase F) (PE011, Novoprotein, Suzhou, Jiangsu, China). Purified CD47 extracellular domain (ECD)‐biotin protein (CD7‐HM447B, KactusBio, Shanghai, China) was treated with PNGase F (50 U/10 µg) to remove N‐linked glycans, and then the protein samples were subjected to sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE). Following the completion of SDS‐PAGE, the gels were stained using the Glycoprotein Staining Kit (RTD6501, RealTimes, Beijing, China) and Coomassie Staining Kit (P0017A, Beyotime) following the manufacturer's instructions. Photos were taken for documentation and further analysis after staining.

2.11. Lens culinaris agglutinin (LCA) lectin enrichment

The cell lysate was mixed with 50 µL of biotinylated LCA lectin (B1045, vector laboratories, Newark, CA, USA) in the lysis buffer at 4°C overnight. Then, 50 µL agarose coupled streptavidin (20361, Thermo Fisher Scientific) was added, and the mixture was incubated for 1 h at room temperature. The beads were washed and then boiled with loading buffer at 100°C for 10 min. The samples were analyzed with CD47 antibodies (1:1000, #218810, Abcam, Cambridge, Cambridgeshire, England).

2.12. Macrophage generation and stimulation

Bone marrow‐derived macrophages (BMDMs) were obtained from NoD.Cg.Prkdc^scidll2rg^em1Smoc (NSG) mice (NM‐NSG‐001, Model Organisms Center, Shanghai, China). Specific pathogen‐free mice (NSG), 6‐10 weeks old, were euthanized via cervical dislocation and disinfected with 75% alcohol. The bilateral femur and tibiofibula of mice were dissected and placed into sterile PBS. The bone marrow cells were blown out with a sterile syringe, filtered with a 200‐mesh strainer, and centrifuged at 200 ×g at 4°C for 5 min. The cells obtained by centrifugation were treated with red cell lysis solution (R1010, Solarbio, Beijing, China). Cells were differentiated into macrophages by 7‐8 days of culture in DMEM [10% heated‐inactivated serum, 1% penicillin, and 10 ng/mL macrophage colony‐stimulating factor (CB34, Novoprotein)].

Peripheral blood mononuclear cells (PBMCs) from healthy donors were isolated via density‐gradient centrifugation using Ficoll‐Hypaque (DKW‐LSH‐0250, Dakewe Biotech, Shenzhen, China). Monocytes were differentiated into macrophages over 7‐10 days of culture in RPMI 1640 [10% heated‐inactivated serum, 1% penicillin, 25 ng/mL macrophage colony‐stimulating factor (C417, Novoprotein), and 2 mmol/L glutamine (C0212, Beyotime)].

2.13. Flow cytometry‐based phagocytosis assay

A green fluorescent protein (GFP)‐encoding lentivirus was prepared from the pLVX vector construct using standard techniques and transfected into J82 and T24 cells to generate GFP⁺ cells. Before the phagocytosis assay, culture medium was renewed. GFP⁺ target cells were harvested and counted. The ratio of macrophages and GFP⁺ target cells in co‐culture was 1:5. After co‐culturing for 2‐4 h at 37°C, the cells were harvested. The macrophages were stained with APC‐labeled anti‐CD11b antibody (1:200, #101212, BioLegend, San Diego, CA, USA) to identify macrophages. Assays were analyzed by flow cytometry. Phagocytosis was measured as the number of CD11b⁺GFP⁺ macrophages, quantified as a percentage of the total CD11b⁺ macrophages. All fluorophore‐conjugated antibodies are listed in Supplementary Table S1.

2.14. In vivo xenograft tumor‐growth experiments

For bladder cancer xenograft analysis, male NSG mice (6‐10 weeks old) were engrafted 5 × 10⁶ J82 or T24 bladder cancer cells in a serum‐free medium with a 1:1 mixture of Matrigel (082704, ABW, Shanghai, China). All mice were housed in specific pathogen‐free environment. Tumors were allowed to grow for 30 days. Humane endpoints specified a maximum tumor volume of 1,500 mm³ and dimensions (length and width) not exceeding 15 mm, or when weakness or body weight loss (70% of control littermates) was observed. Then the tumors were mechanically and enzymatically excised using collagenase IV (2 mg/mL, 17104019, Thermo Fisher Scientific) and DNase I (0.1 mg/mL, 18068015, Thermo Fisher Scientific). Phagocytosis was measured as the number of F4/80⁺GFP⁺ macrophages, quantified as a percentage of the total F4/80⁺ (1:200, #123149, Biolegend) macrophages.

For dihydrotanshinone I (DHT I) (5 and 25 mg/kg, B20357, Yuanye Bio‐Technology, Shanghai, China) treatment, tumors were allowed to grow for 7 ‐ 10 days. Each mouse was orally given DHT I daily. All tumor models were calculated by measuring the tumor size, V = length × width²/2.

2.15. Mass spectrometry

To detect glycosyltransferase interaction with CD47, 293T and J82 cells were transfected with CD47‐FLAG for 24‐48 h. CD47‐FLAG protein was enriched by beads, which were incubated with an anti‐FLAG antibody. The washed beads were boiled in the loading buffer (P0015A, Beyotime) for 10 min to elute bound proteins and prepare for SDS‐PAGE analysis. The whole lane was excised and digested with trypsin at 37°C overnight to attain the peptide extract. Peptides were then desalted and lyophilized. After separation on an analytical column, the peptides were analyzed through mass spectrometry. Mass spectrometry analysis was assisted by Beijing Qinglian Biotech Co., Ltd. CD47‐interacting proteins from 293T and J82 cells were intersected with glycosyltransferases (Supplementary Table S2) using Sanger Box (http://www.sangerbox.com/).

2.16. Culture and treatment of patient‐derived tumor‐like cells (PTCs) from bladder cancer

Samples collected from bladder cancer patients were conditioned in ice‐cold PBS with 4‐(2‐Hydroxyethyl)‐1‐piperazineethanesulfonic acid (10 mmol/L, H43340, Aladdin) and penicillin‐streptomycin (100 U/mL, 15140122, Thermo Fisher Scientific). Tissues were washed with PBS at least three times. Necrotic areas and adipose tissue were removed. Tissues were minced into small pieces and digested in 10 mL PBS containing 1 mmol/L EDTA (E301914, Aladdin), 200 U/mL collagenase I, II, and IV (17018029, 17101015, 17104019, Thermo Fisher Scientific), at 37°C for 1 h. Dissociated cells were collected using filters with a pore size of 40 µm. Cells were centrifuged at 200 ×g at 4°C for 5 min. Cells were resuspended in a PTC growth medium and seeded in the low‐attachment‐surface dish. Cells were cultured in an incubator at 37°C, 5% CO₂. PTC growth medium was refreshed every 2‐3 days, as necessary. After five days of culture, the antibody and MAN1B1 inhibitor (DHT I) were added to the specified concentration for 24 h. Phagocytosis was detected following digestion.

2.17. Subcellular fractionation assays

Cells harvested with a cell scraper were collected into a centrifuge tube and washed twice with PBS. The Golgi apparatus and endoplasmic reticulum (ER) extraction kit (BB‐3604, BB‐314541, BestBio, Shanghai, China) was used according to the manufacturer's instructions. Briefly, the cells were fully cracked by a homogenizer, and impurities were removed by several centrifugations. Finally purified Golgi and endoplasmic reticulum solution was obtained.

2.18. Molecular docking

The X‐ray crystal structures of CD47 (7MYZ), QPCTL (3PB6) and MAN1B1 (1×9D) were retrieved from the Protein Data Bank (https://www.wwpdb.org/). To ensure the accuracy of the docking results, the protein was prepared by the AutoDockTools‐1.5.7 (https://autodocksuite.scripps.edu/adt/). Water molecules were removed from the proteins, and polar hydrogen atoms were incorporated. Docking Web Server (GRAMM) (https://gramm.compbio.ku.edu/) was used for protein‐protein docking. The resulting protein‐protein complex was also manually optimized by removing water and adding polar hydrogen by the AutoDockTools‐1.5.7. Ultimately, the protein‐protein interaction was predicted, and a visual representation of the interaction was generated using PyMOL (https://www.pymol.org/).

2.19. Cell viability assay

The cells were cultured in 96‐well plates at a density of 5000 cells per well for 24, 48, and 72 h. Following this incubation period, the culture medium was discarded, and a medium containing Cell Counting Kit‐8 (CCK8) (HY‐K0301, MedChemExpress) was introduced. The cells were then incubated for an additional hour before measuring optical density at 450 nm (OD450). Proliferation rate (%) = [OD450 (experiment) ‐ OD450 (blank)] / [OD450 (control) ‐ OD450 (blank)] × 100%.

2.20. Quantitative reverse transcription PCR (qRT‐PCR) analysis

Total RNA was extracted using the TRIzol reagent. The reverse transcription process was conducted using the Evo M‐MLV kit (AG11728, Accurate Biology). The following primers were used for qRT‐PCR. For human MAN1B1: 5‐TCACAGGGGACCGCAAATAC‐3 (forward), 5‐ TGAGCAGGTTTGGGTCATCG ‐3 (reverse); for human MGAT2: 5‐GTGCATAACCGGCCCGAATA‐3 (forward), 5‐ACGAGGACGTTGTCAATTCCC‐3 (reverse); for human 18S: 5‐GGATGCGTGCATTTATCAGA‐3 (forward), 5‐GTTGATAGGGCAGACGTTCG‐3 (reverse); for human ACTIN: 5‐ACAGAGCCTCGCCTTTGC ‐3 (forward), 5‐GATATCATCATCCATGGTGAGCTGG‐3 (reverse). PCR conditions: 30 s at 95°C for the preliminary denaturation, and 40 cycles of 5 s at 95°C, 30 s at 60°C for the amplification.

2.21. The Cancer Genome Atlas (TCGA) analyses

The expression levels of CD47 and MAN1B1 in different tumors were evaluated utilizing the tumor immune estimation resource (TIMER) platform base on TCGA dataset (http://timer.cistrome.org/). Kaplan‐Meier survival curves were generated to assess the overall survival after surgery of bladder cancer and kidney cancer patients from TCGA dataset (https://www.cancer.gov/ccg/research/genome‐sequencing/tcga), categorized by high and low expression levels of MAN1B1 and CD47, using the median as cut‐off.

2.22. UbiBrowser 2.0 and N‐glycosylation (NXT) motif analysis

UbiBrowser 2.0 database (http://ubibrowser.bio‐it.cn/ubibrowser_v3/) was used to predict the E3 ubiquitin ligase of MAN1B1. NXT motif of CD47 was analyzed with Uniprot (https://www.uniprot.org/).

2.23. Statistical analysis

Quantitative data were presented as mean ± standard deviation (SD). Mean normalization was applied to all quantitative data for statistical analysis. Each quantitative experiment was conducted with a minimum of three replicates to ensure reliability. Statistical analyses were performed using GraphPad Prism 8 (https://www.graphpad.com). The paired or unpaired two‐tailed Student's t‐test was utilized to assess differences between two groups. A P value < 0.05 was considered statistically significant.

3. Results

3.1. CD47 was heavily glycosylated in bladder cancer

CD47 is a major immune checkpoint site for cancer immunotherapy [25]. Due to limited reports about its function in bladder cancer, we knocked out Cd47 in the mouse bladder cancer cell line MB49 (Figure 1A, Supplementary Figure S1A). After Cd47 was knocked out, there was a significant reduction in tumor burden in vivo (Figure 1B‐D, Supplementary Figure S1B‐D), indicating the key role of CD47 in bladder cancer. To investigate the expression pattern of CD47 in bladder cancer, we collected tissue samples from bladder cancer patients. Two bands of CD47 were observed in bladder cancer and adjacent normal tissues in Western blotting (Figure 1E).

FIGURE 1.

The glycosylation of CD47 was critical for its interaction with SIRP‐α. (A) Cd47 was knocked out in the bladder cancer cell line MB49. (B‐D) Tumor growth of MB49 in mice (n = 4 mice per group). (B) Images of endpoint subcutaneous tumors. (C) Tumor volumes of MB49 in vivo. (D) Tumor weight of MB49 at endpoint. (E) Western blotting analysis of CD47 in normal and tumor tissues of bladder cancer. (F) Western blotting analysis of cell lysates from tumor tissues and 293T treated with PNGase F for 1 h at 37°C in vitro. (G) Glycoprotein staining and Coomassie blue staining of PNGase F treated purified CD47 (ECD)‐biotin. Horseradish peroxidase and soybean trypsin inhibitor served as positive and negative controls, respectively. (H) Schematic diagram of CD47 amino acid sequence alignment among different species. The five putative NXT motifs are highlighted in red. (I) Western blotting of the protein expression pattern of CD47 WT and mutants overexpressed in 293T cells. (J) Western blotting of the protein expression pattern of CD47‐WT and 5NQ treated with PNGase F for 1 h at 37°C in vitro. (K‐L) Analysis of CD47‐SIRP‐α binding by IP. (K) The lysates of 293T cells overexpressing FLAG‐tagged CD47 WT or 5NQ mutant were incubated with beads which had conjugated to SIRP‐α‐Fc and then observed by Western blotting with anti‐FLAG antibody. (L) The lysates of 293T cells overexpressing FLAG‐tagged SIRP‐α were incubated with beads which had conjugated to CD47‐WT or 5NQ‐Fc and then observed by Western blotting with anti‐FLAG antibody. (M) The lysates of 293T cells overexpressing CD47 WT or the indicated NQ mutants were incubated with beads which had conjugated to SIRP‐α‐Fc and then observed by Western blotting with anti‐FLAG antibody. (N‐O) Analysis of CD47 knockout in J82 and T24 cells by Western blotting. (P) CD47‐WT or 5NQ was overexpressed in CD47‐knocked out J82, T24 or 293T cells to detect the expression of CD47 and its affinity with SIRP‐α by flow cytometry. (Q‐R) Representative flow cytometry images (Q) and quantification (R) depicted the In vitro phagocytosis of J82 cells, which were knocked out CD47 and then stably rescued with either CD47‐WT or CD47‐5NQ. Boxes were representative of GFP⁺ CD11b⁺ macrophages. (S‐W) Tumor growth of J82 in NSG mice (n = 4 mice per group). (S) Images of endpoint subcutaneous tumors. (T) Tumor volumes of J82 in vivo. (U) Tumor weight of J82 at endpoint. (V‐W) Representative flow cytometry (V) and quantification (W) demonstrated TAM phagocytosis in GFP⁺ CD47‐WT group and GFP⁺ CD47‐5NQ group (n = 4). Numbers indicated frequency of phagocytosis events out of all TAMs after engraftment. Boxes were representative of GFP⁺ CD11b⁺ macrophages. P values in (C‐D, P‐R, T‐W) were determined by a two‐tailed unpaired Student's t‐test, error bars represent mean ± SD. NS, not significant; * P < 0.05, ** P <0.01, *** P <0.001.

Abbreviations: CD47, cluster of differentiation 47; SIRP‐α, signal‐regulatory protein alpha; TAM, tumor‐associated macrophages; PNGase F, peptide‐N‐glycosidase F; GFP, green fluorescent protein; ECD, Extracellular domain; NXT motif, N‐glycosylation motif; WT, Wild type; 5NQ, 5 substitution of asparagine with glutamine mutant; NSG, NoD.Cg.Prkdcscidll2rgem1Smoc mice; gCD47, glycosylated CD47; ngCD47, nonglycosylated CD47; IP, immunoprecipitation; SD, standard deviation.

Considering post‐translational modifications, particularly glycosylation, could result in proteins with different molecular weights in Western blotting analysis [26]. Thus, we hypothesized that the two bands observed in Western blotting represented the glycosylated and non‐glycosylated forms of CD47. To further validate this hypothesis, tumor tissues and 293T cells lysates were treated with PNGase F, which reduced the molecular weight of CD47 as observed in Western blotting (Figure 1F). Purified CD47 protein treated with PNGase F was then subjected to glycoprotein staining. Adding PNGase F reduced the molecular weight of CD47 on SDS‐PAGE, and the band disappeared in the glycoprotein stain (Figure 1G), confirming that the CD47 of higher molecular weight was the glycosylated form.

Mawby et al. [27] predicted possible glycosylation sites in CD47. To determine the glycosylation sites, we searched for evolutionarily conserved NXT motifs in the CD47 amino acid sequences from different species. We identified five conserved NXT motifs (N23, N34, N50, N73 and N111) (Figure 1H). Substitution of asparagine (N) with glutamine (Q) at each or all of the five sites (N23Q, N34Q, N50Q, N73Q, N111Q, or 5NQ) resulted in reduced molecular weight of the CD47 mutants compared to their wild‐type (WT) counterpart (Figure 1I). Treatment of CD47‐WT and CD47‐5NQ with PNGase F revealed that the molecular weight of the CD47‐WT reduced to that of the 5NQ mutant, while the band location of the 5NQ mutant remained unchanged (Figure 1J). This suggested that glycosylation modification occurred at these five sites, and CD47‐5NQ represented the non‐glycosylated form of CD47.

Immune checkpoint programmed cell death protein 1 (PD‐1) [28], programmed cell death ligand 1 (PD‐L1) [29], and B7 homolog 3 (B7H3) [22] can be glycosylated and are closely related to immune evasion in cancers. We wondered whether the integrity of CD47 glycans is vital for immune evasion. Therefore, SIRP‐α‐Fc fusion proteins were incubated with the lysates from 293T cells expressing FLAG‐tagged CD47‐WT or 5NQ. CD47‐SIRP‐α interaction was analyzed by IP and Western blotting. The binding between CD47‐5NQ and SIRP‐α decreased significantly compared with CD47‐WT (Figure 1K). Consistent results were found when SIRP‐α was overexpressed in 293T cells using CD47‐WT and CD47‐5NQ as bait (Figure 1L). To investigate the impact of specific glycosylation sites, we overexpressed single‐point mutant CD47 (N23Q, N34Q, N50Q, N73Q and N111Q) and CD47‐5NQ. It was found that each glycosylation site enhanced its affinity for SIRP‐α, with particularly strong interactions observed at the N34 site (Figure 1M).

In the analysis of CD47 expression in bladder cancer cell lines, the J82 cell line showed the lowest level of CD47, whereas the T24 cell line had the highest level of CD47 expression (Supplementary Figure S1E). The J82 cell line was derived from a male bladder cancer patient, while T24 cell line was obtained from a female patient. Furthermore, CD47 was knocked out in J82 and T24 (Figure 1N‐O). When CD47‐WT or CD47‐5NQ was overexpressed in CD47‐KO J82, T24, and 293T cells, CD47‐5NQ hardly bound with SIRP‐α compared to WT (Figure 1P). The anti‐phagocytosis ability of J82 cells overexpressing CD47‐5NQ was notably diminished compared to CD47‐WT (Figure 1Q‐R). CD47‐KO J82 cells, overexpressed CD47‐WT or CD47‐5NQ, were transplanted into NSG mice. CD47‐5NQ overexpression significantly decelerated tumor growth (Figure 1S‐T) and reduced tumor burden in vivo compared with CD47‐WT (Figure 1U). Simultaneously, CD47‐5NQ increased the phagocytosis to tumor cells compared to CD47‐WT (Figure 1V‐W). These results strongly indicated that CD47 glycosylation was critical for its interaction with SIRP‐α and played a vital role in phagocytosis.

3.2. MAN1B1 involved in the glycosylation of CD47 was highly expressed in bladder cancer

Since glycosylation of CD47 is critical for its immunosuppressive function, we studied the mechanisms of abnormal CD47 NXT regulation. We performed the IP assay using the lysates from J82 or 293T cells transfected with CD47‐FLAG to obtain complexes containing CD47 and its interacting proteins. The immuno‐complex was analyzed using mass spectrometry to identify potential CD47‐interacting proteins in cells.

We identified that glycosyltransferases MAN1B1, MOGS, MGAT2 and GANAB might be involved in CD47 glycosylation (Figure 2A and Supplementary Figure S2A). The CD47‐MAN1B1 interaction score was ‐726 according to molecular docking (Figure 2B), representing a stable structural relationship between CD47 and MAN1B1. Meanwhile, the CD47‐QPCTL interaction score was ‐573 (Supplementary Figure S2B). The interaction between MAN1B1 and CD47 was better than QPCTL in silico docking. IP assays demonstrated that both MAN1B1 and MOGS interacted with CD47, especially the non‐glycosylated form CD47‐5NQ, while MGAT2 and GANAB did not interact with CD47 (Figure 2C and Supplementary Figure S2C‐E). This indicated that the non‐glycosylated form of CD47 may require more glycosyltransferases for proper glycosylation.

FIGURE 2.

MAN1B1 was involved in the glycosylation process of CD47 and exhibited high expression levels in bladder cancer. (A) MS analysis result of CD47. Venn diagram was performed according to glycosyltransferases identified in CD47 MS analysis from 293T and J82 cells and well‐known glycosyltransferase. (B) Diagram of docking between CD47 and MAN1B1. The binding score was ‐726. (C) Analysis of interaction between CD47 and MAN1B1 by IP. (D) Expression levels of CD47 and MAN1B1 in 12 human BLCA fresh samples by Western blotting. N, normal matched tissue; T, tumor tissue. (E‐F) The band intensity was quantified and normalized to compare CD47 and MAN1B1 levels (n = 12). (G‐H) Images and statistical results of IHC staining of MAN1B1 in a tissue microarray (n = 47). (I) Analysis of MAN1B1 knockout in J82 and T24 cells by Western blotting. (J) Detected changes in LCA affinity of whole cell lysates under MAN1B1 knockout or control conditions in J82 and T24. (K) Survival analysis of MAN1B1 and CD47 in bladder cancer. P values in (E‐H) were determined by a two‐tailed unpaired Student's t‐test, error bars represent mean ± SD. NS, not significant; * P < 0.05, ** P <0.01, *** P <0.001.

Abbreviations: CD47, cluster of differentiation 47; SIRP‐α, signal‐regulatory protein alpha; MAN1B1, mannosidase alpha class 1B member 1; IHC, Immunohistochemical staining; LCA, Lens culinaris agglutinin; MS, Mass spectrometry; IP, immunoprecipitation; BLCA, bladder urothelial carcinoma; B7H3, B7 homolog 3; SD, standard deviation.

Glycosylation of proteins often occur in the ER and Golgi apparatus [30], and we wondered whether CD47 and MAN1B1 co‐localized in the ER and Golgi apparatus. IF experiments and subcellular fractionation assays revealed the co‐localization of MAN1B1 and CD47 in the ER and Golgi apparatus (Supplementary Figure S3A‐F), indicating that CD47 and MAN1B1 interacted in the ER and Golgi apparatus. IF revealed that the localization of CD47‐5NQ in the ER and Golgi apparatus increased compared with CD47‐WT (Supplementary Figure S3G‐H). This indicated that MAN1B1 was involved in modifying CD47 and completing glycosylation in the ER and Golgi apparatus.

Western blotting further verified the expression of MAN1B1 and CD47 in bladder cancer tissues. Surprisingly, MAN1B1 was significantly increased in bladder cancer, while CD47 did not change significantly (Figure 2D‐F). MAN1B1 in bladder cancer was analyzed by immunohistochemistry, and consistent results were found (Figure 2G‐H, Supplementary Figure S4 and Supplementary Table S3). MAN1B1 is a glycosidase that cuts mannose during glycosylation, allowing for the N‐acetylglucosamine to mannose residues [31]. Lectin from LCA specifically recognizes mannose [32]. Therefore, this reagent was used to confirm whether CD47 was directly glycosylated by MAN1B1. Interestingly, MAN1B1 knockout did not alter the expression level of CD47 (Figure 2I). Meanwhile, in MAN1B1‐KO cells, the affinity between CD47 and lectin was increased significantly, while B7H3, an immune checkpoint protein with glycosylation modification, did not exhibit a similar change (Figure 2J).

Survival analysis showed that bladder cancer patients with high expression of MAN1B1, but not CD47, had significantly shorter survival compared to those with low expression of MAN1B1 (Figure 2K). The survival analysis of MAN1B1 partially demonstrated that high CD47 glycosylation is associated with short overall survival in bladder cancer patients.

Logtenberg et al. [19] reported QPCTL as a major post‐translational modification enzyme of the CD47‐SIRP‐α axis which also did not influence the expression level of CD47. QPCTL showed no significant increase in bladder cancer (Supplementary Figure S5A‐B) and was not significantly associated with the prognosis of bladder cancer (Supplementary Figure S5C).

Additionally, analysis of the TCGA database revealed that the expression levels of CD47 and MAN1B1 in the majority of tumors were distinct (Supplementary Figure S6A‐B). Notably, the expression of MAN1B1 were found to be significantly higher in a variety of cancers, such as invasive breast carcinoma, cholangiocarcinoma, and adenocarcinomas of the colon, lung, and stomach, compared to their normal tissue counterparts. Moreover, kidney renal clear cell carcinoma (KIRC), rectum adenocarcinoma and uterine corpus endometrial carcinoma, exhibited a similar pattern to bladder cancer. The expression levels of MAN1B1 and QPCTL in KIRC were associated with patient survival, whereas the CD47 expression was not significantly associated with patient survival (Supplementary Figure S6C). In KIRC, MAN1B1 did not show a significant increase compared to the matched normal tissues; however, QPCTL expressed significantly higher in tumor tissues than in matched normal tissues in KIRC (Supplementary Figure S6D‐F). In prostate cancer, the expression of MAN1B1 in cancer tissues was higher than that in matched normal tissues (Supplementary Figure S6G‐H). Thus, we speculated that both QPCTL and MAN1B1 were important modifying enzymes of CD47 in cancers, and their expression levels and dominance varied in bladder cancer and kidney cancer. These findings mechanistically demonstrated that the abnormally elevated glycosyltransferase MAN1B1 acted as a glycosyl‐modifying enzyme for CD47 in bladder cancer.

3.3. Identification of MAN1B1 as a modulator of the CD47‐SIRP‐α axis

Since glycosylation was essential for the immune evasion of CD47, and MAN1B1 was the glycosyltransferase of CD47, we investigated the impact of MAN1B1 on immune evasion. IP revealed that the decrease of MAN1B1 resulted in a significant decrease in the affinity between CD47 and SIRP‐α (Figure 3A‐B). MAN1B1 knockout did not affect the expression of CD47 on the cell membrane, and simultaneously, the affinity between CD47 and SIRP‐α was significantly weakened (Figure 3C‐F). MOGS was knocked out, while GANAB and MGAT2 were knocked down in bladder cancer cell lines (Supplementary Figure S7A‐C). Despite this, these glycosyltransferases had no effect on the affinity between CD47 and SIRP‐α (Supplementary Figure S7D‐G). IF revealed that the binding capacity of SIRP‐α to CD47 was reduced following MAN1B1 knockout (Figure 3G). IP demonstrated that elevated levels of MAN1B1 corresponded with a marked enhancement in the binding affinity between CD47 and SIRP‐α in bladder cancer tissues (Figure 3H‐I).

FIGURE 3.

Identification of MAN1B1 as a modulator of CD47‐SIRP‐α axis. (A‐B) Analysis of CD47‐SIRP‐α binding was measured in MAN1B1 knockout and control cells by IP. (A) The lysates of sgControl or sgMAN1B1 293T cells overexpressing FLAG‐tagged CD47 were incubated with beads which had conjugated to SIRP‐α‐Fc and then observed by Western blotting with anti‐FLAG antibody. (B) The whole cell lysates of sgControl or sgMAN1B1 T24 cells were incubated with beads which had conjugated to SIRP‐α‐Fc and then observed by Western blotting with anti‐CD47 antibody (left). The band intensity was quantified and normalized to exhibit changed affinity after MAN1B1 knockout (right). (C‐F) Flow cytometry plots of surface binding of anti‐human CD47 antibody clone B6H12 (anti‐hCD47) and SIRP‐α to T24 (C‐D) or J82 (E‐F) sgControl and sgMAN1B1 cells. Values indicated MFI relative to sgControl cells stained with the same reagent (D, F). (G) IF staining showed bound SIRP‐α protein (red) and DAPI‐stained nuclei (blue) in T24 cells. (H) Analysis of CD47‐SIRP‐α binding between normal and tumor tissues by IP. (I) The affinity intensity was quantified and normalized to show the association between the expression levels of MAN1B1 and the CD47‐SIRP‐α binding levels. (J‐M) Representative flow cytometry images (J, L) and quantification (K, M) demonstrated the phagocytic activity of BMDMs to both control and MAN1B1 knockout J82 cells (J‐K, n = 9) or T24 cells (L‐M, n = 4). Boxes were representative of GFP⁺ CD11b⁺ macrophages. (N‐Q) Representative flow cytometry images (N, P) and quantification (O, Q) demonstrated the phagocytosis of PBMCs to both control and MAN1B1 knockout J82 cells (N‐O, n = 4), as well as in control and MAN1B1 and CD47 knockout T24 cells (P‐Q, n = 3). Boxes were representative of GFP⁺ CD11b⁺ macrophages. (R) Human bladder cancer cells J82 were transplanted subcutaneously into NSG mice. (S‐Y) Tumor growth of J82 sgControl and sgMAN1B1 in mice (n = 4 mice per group). (S) Images of endpoint subcutaneous tumors. (T) Tumor grow curves of J82 cells in vivo. (U) Tumor weight of J82 cells at endpoint. (V) Representative flow cytometry plots demonstrated TAM phagocytosis in GFP⁺ control group (left) versus GFP⁺ MAN1B1 knockout group (right); numbers indicated the frequency of phagocytosis events out of all TAMs after engraftment. Boxes were representative of GFP⁺ CD11b⁺ macrophages. (W) Phagocytosis ratio was analyzed by statistics (n = 4). (X‐Y) Frequency of TAMs positive for CD80 (M1‐like) as per gating in vivo (n = 4). P values (A‐B, D, F, I‐Q and T‐Y) were determined by a two‐tailed unpaired Student's t‐test, error bars represent mean ± SD. NS, not significant; * P < 0.05, **P < 0.01, *** P < 0.001.

Abbreviations: CD47, cluster of differentiation 47; SIRP‐α, signal‐regulatory protein alpha; MAN1B1, mannosidase alpha class 1B member 1; BMDMs, bone marrow‐derived macrophages; PBMCs, peripheral blood mononuclear cells; TAM, tumor‐associated macrophages; NSG, NoD.Cg.Prkdcscidll2rgem1Smoc mice; GFP, green fluorescent protein; MFI, mean fluorescence intensity; IF, immunofluorescence; IP, immunoprecipitation; BLCA, bladder urothelial carcinoma; B7H3, B7 homolog 3; SD, standard deviation.

To study the effect of MAN1B1 on CD47‐mediated anti‐phagocytosis, BMDMs were differentiated in vitro and subsequently co‐cultured with cancer cells. It was found that MAN1B1 knockout inhibited the CD47‐mediated anti‐phagocytosis (Figure 3J‐M and Supplementary Figure S8A). The in vitro differentiation of PBMCs into macrophages was followed by co‐culture with J82 and T24 cells, yielding consistent outcomes (Figure 3N‐Q). The enhancement in phagocytosis induced by CD47 knockout exhibited a slight distinction from the augmentation observed subsequent to MAN1B1 knockout.

MAN1B1 knockout displayed no effect on the proliferation rate in vitro (Supplementary Figure S8B‐C). J82 and T24 cells were transplanted into NSG mice (Figure 3R). MAN1B1 knockout significantly decelerated tumor growth (Figure 3S‐T and Supplementary Figure S8D‐E) and diminished tumor burden in vivo when compared with control group (Figure 3U and Supplementary Figure S8F). Simultaneously, we observed that the knockout of MAN1B1 resulted in a three‐fold increase in the phagocytosis to tumor cells compared to the control group (Figure 3V‐W, Supplementary Figure S8G‐I and Supplementary Figure S9A‐B). There was no notable alteration in the infiltration of M1‐like macrophages in vivo (Figure 3X‐Y and Supplementary Figure S9C). In summary, the findings indicated that MAN1B1 played a crucial role in mediating immunosuppression via CD47 in bladder cancer.

3.4. Targeted therapy against MAN1B1 inhibited the binding of SIRP‐α and promoted macrophage phagocytosis

Alvarez et al. [33] predicted that DHT I could potentially inhibit MAN1B1. We verified the possibility of DHT I as an inhibitor of MAN1B1 via molecular docking. The binding energy of the protein‐ligand complex was ‐8.4 kcal/mol (Figure 4A‐B), demonstrating an excellent performance. Cell proliferation experiments were performed to assess the toxicity of DHT I. The results showed that 0.625 µmol/L of DHT I did not inhibit cell proliferation (Supplementary Figure S10A). After exposure to non‐cytotoxic concentrations of DHT I, an increased affinity between CD47 and lectin was observed. This enhancement occurred without any change in the expression level of MAN1B1 (Figure 4C). This confirmed that DHT I was a potent inhibitor of MAN1B1.

FIGURE 4.

The MAN1B1 inhibitor DHT I reduced the affinity between CD47 and SIRP‐α, promoted phagocytosis of bladder cancer cells via macrophage in vitro and in vivo. (A) Chemical structure of dihydrotanshinone I (DHT I). (B) Predicted three‐dimensional model of interaction binding between DHT I (ZINC ID: 8681480) and MAN1B1 (PDB: 1×9D) as shown by computational docking. Binding energy: ‐8.4 kcal/mol. (C) Detected changes in LCA affinity of whole cell lysates from J82, T24 or 293T cells treated with DHT I (0.625 µmol/L). (D) Flow cytometry plot of surface binding of anti‐human CD47 antibody clone B6H12 (anti‐hCD47) and SIRP‐α to J82 cells which were treated with DHT I (0.625 µmol/L). Values indicated MFI relative to control cells stained with the same reagent (bottom). (E) 293T cells overexpressing FLAG‐tagged CD47 were treated with DHT I for 24h, then the whole cell lysates were incubated with beads which had conjugated to SIRP‐α‐Fc and observed by Western blotting with anti‐FLAG antibody. (F) Representative flow cytometry image depicted the process of BMDMs mediated phagocytosis of J82 sgControl or sgMAN1B1 cells treated with or without DHT I for 24h in vitro. Boxes were representative of GFP⁺ CD11b⁺ macrophages. (G) Phagocytosis ratio was analyzed by statistics. (H) A schematic treatment plan for NSG mice bearing subcutaneous T24 tumors. Mice were subcutaneously implanted with T24 sgControl or sgMAN1B1 cells and treated with either control vehicle or DHT I, respectively. (I‐M) Tumor growth in DHT I‐treated mice (n = 4 mice per group). (I) Images of endpoint subcutaneous tumors. (J) Tumor weight of T24 tumors at endpoint. (K) Tumor volumes of T24 in vivo. (L‐M) Representative flow cytometry plots demonstrated TAM phagocytosis in T24 tumor, with numbers indicated the frequency of phagocytosis events out of all TAMs after treatment with DHT I. Boxes were representative of GFP⁺ CD11b⁺ macrophages. (N) A representative case of bladder cancer with PTCs culture was presented in the bright field view. (O) Representative flow cytometry image depicted phagocytosis in PTCs model treated with CD47 antibody or DHT I. Boxes were representative of KRT18⁺ CD11b⁺ macrophages. (P) Analysis of phagocytosis ratio after CD47 antibody or DHT I treatment in the PTCs model. P values in (D, G and J‐M) were determined by a two‐tailed unpaired Student's t‐test, error bars represent mean ± SD. NS, not significant; * P < 0.05, **P < 0.01, ***P < 0.001.

Abbreviations: CD47, cluster of differentiation 47; SIRP‐α, signal‐regulatory protein alpha; MAN1B1, mannosidase alpha class 1B member 1; DHT I, dihydrotanshinone I; TAM, tumor‐associated macrophages; PTCs, patient‐derived tumor‐like cell clusters; DMSO, Dimethyl sulfoxide; GAPDH, Glyceraldehyde‐3‐phosphate dehydrogenase; LCA, Lens culinaris agglutinin; KRT18, Keratin 18; GFP, green fluorescent protein; MFI, mean fluorescence intensity; BMDMs, bone marrow‐derived macrophages; NSG, NoD.Cg.PrkdcscidIl2rgem1Smoc mice; SD, standard deviation.

DHT I (0.625 µmol/L) did not reduce the expression of CD47 on cell membrane; meanwhile, the affinity between CD47 and SIRP‐α decreased, similar to the MAN1B1 knockout group (Figure 4D). The affinity between CD47 and SIRP‐α decreased with increasing concentration of DHT I in the 293T cells, while the weak interaction observed between CD47 and SIRP‐α did not vary with the increasing concentration of DHT I in MAN1B1‐KO 293T cells (Figure 4E). Additionally, DHT I weakened the anti‐phagocytosis effect in J82 cells and the anti‐phagocytosis effect with the function of DHT I remained unchanged in MAN1B1‐knockout cells (Figure 4F‐G).

DHT I can inhibit the proliferation of hepatoma cells by inhibiting STAT3 phosphorylation through proto‐oncogene tyrosine‐protein kinase Src [34]. However, in T24 cells, when treated with the reported effective concentration, no significant effect on p‐STAT3 expression was observed, even at the concentration of 2.5 µmol/L (Supplementary Figure S10B). Therefore, this effect was found to be imperceptible in T24. T24 cells were transplanted in the NSG mice. At 7 days after transplantation, DHT I was orally administered to the mice (Figure 4H). The tumor burden decreased significantly with the use of DHT I in the control group. Meanwhile, the MAN1B1‐KO group showed no significant difference in response to the use of DHT I (Figure 4I‐K). The trends of phagocytosis in vivo and tumor burden was consistent (Figure 4L‐M).

Animal experiments demonstrated the ability of DHT I to inhibit MAN1B1 in vivo. In addition, validation of the effectiveness of inhibitors in real‐world tumors in patients is important. However, DHT I had still not been approved for clinical use. Recently, several patient‐derived tumor models have been developed to screen drug candidates for cancer patients. An ex vivo model called PTCs has been applied. PTCs result from the self‐assembly and proliferation of primary epithelial, fibroblast, and immune cells, which structurally and functionally recapitulate original tumors [35]. This model was used to culture patient‐derived samples of bladder cancer (Figure 4N). After the culture was completed, DHT I, CD47 antibody and SIRP‐α were added to observe the effect on phagocytosis in our model respectively. DHT I promoted the phagocytosis of macrophages in the PTC model, whereas the CD47 antibody and its ligand showed a less pronounced effect (Figure 4O‐P, Supplementary Figure S10C‐D and Supplementary Table S4). The antibody's poor effect might be attributed to the difficulty of crossing the biological barrier, highlighting one of the reasons why MAN1B1 inhibitors were considered to have more promising prospects for application. This indicated that targeting MAN1B1 using DHT I could significantly inhibit CD47‐mediated immune evasion in bladder cancer.

3.5. Erythrocyte CD47 lacked MAN1B1 modification

We also assessed the in vivo toxicity of DHT I on the levels of red blood cell, hemoglobin and platelet content in mice. The results indicated that DHT I did not elicit anemia, which was commonly associated with targeting CD47 (Figure 5A). Furthermore, DHT I showed no adverse effects on the hepatorenal function of mice (Figure 5B).

FIGURE 5.

The targeted therapy directed at MAN1B1 exhibited negligible adverse effects. (A) Analysis of erythrocytes, hemoglobin, and platelets of mice treated with DHT I. (B) Analysis of hepatorenal function of mice treated with DHT I. (C) Analysis of MAN1B1 expression levels in erythrocytes from healthy donors compared to bladder cancer tissues by Western blotting. (D) Analysis of expression of MAN1B1 between paired erythrocytes and tumor tissues in bladder cancer patients by Western blotting. (E‐F) Representative flow cytometry plots of surface binding of anti‐human CD47 antibody clone B6H12 (anti‐hCD47) and SIRP‐α to red blood cells of healthy donors which were treated with DHT I (0.625‐10 µmol/L). (F) Values indicated MFI relative to control cells stained with the same reagent (n = 3 donors). P values in (F) were determined by a two‐tailed unpaired Student's t‐test, error bars represent mean ± SD. NS, not significant.

Abbreviations: CD47, cluster of differentiation 47; SIRP‐α, signal‐regulatory protein alpha; MAN1B1, mannosidase alpha class 1B member 1; DHT I, dihydrotanshinone I; GAPDH, Glyceraldehyde‐3‐phosphate dehydrogenase; RBC, red blood cell; MFI, mean fluorescence intensity; BLCA, bladder urothelial carcinoma; SD, standard deviation.

To investigate why DHT I did not affect red blood cell level in mice, we conducted experiments using blood cells from healthy donors and tissue samples from bladder cancer patients. Western blotting demonstrated minimal expression of MAN1B1 in red blood cells, while bladder cancer tissue exhibited a high level of MAN1B1 expression (Figure 5C). Despite the high expression of MAN1B1 in cancer tissues, red blood cells from the same patients still lacked MAN1B1 modification (Figure 5D). In vitro, red blood cells were exposed to DHT I for 24 h at concentrations ranging from 0.625 to 10 µmol/L. The affinity between CD47 and SIRP‐α on the erythrocyte membrane remained unchanged (Figure 5E‐F). This indicated that targeting MAN1B1 represented a promising approach in immunotherapy, effectively addressed the challenge of adverse effects such as anemia associated with existing monoclonal antibody treatments.

3.6. Abnormal ERK phosphorylation stabilized MAN1B1 in bladder cancer

Considering the significant upregulation of MAN1B1 in bladder cancer, our study aimed to uncover the molecular mechanisms behind its dysregulated expression. We utilized qRT‐PCR to quantify the mRNA levels of MAN1B1, revealing no significant variation between tumor and adjacent normal tissues (Figure 6A and Supplementary Table S5). This finding indicated that the increased presence of MAN1B1 in bladder cancer did not stem from mRNA‐level regulation. Instead, it implicated post‐translational modifications as the probable cause of the aberrant high expression of MAN1B1 in this disease.

FIGURE 6.

ERK stabled MAN1B1 by regulating its interaction with E3 ligase HRD1 in bladder cancer. (A) Analysis of mRNA levels of MAN1B1 in bladder cancer by qRT‐PCR (n = 8). (B) Western blotting analysis of WCL derived from 5637 cells treated with trametinib (1 µmol/L) for 24 h before harvesting. (C) Western blotting analysis of WCL derived from J82 cells treated with trametinib (1 µmol/L) before harvesting. (D‐G) Western blotting analysis of WCL derived from bladder cancer cells transfected with KRAS‐G12D or ERK2 as indicated. (H) Analysis of mRNA level of MAN1B1 in 5637 cells which were stably transfected KRAS‐G12D by qRT‐PCR. (I) Analysis of interaction between ERK2 and MAN1B1 by IP. (J) Western blotting analysis of WCL derived from HEK293T cells co‐transfected MAN1B1‐FLAG with KRAS‐G12D or ERK2 as indicated. (K) Western blotting analysis of stability of MAN1B1 treated with 50 µg CHX when co‐transfected with or without ERK2. (L) Western blotting analysis of WCL and anti‐FLAG from lysates of sgControl or sgERK1/2 treated 293T using indicated antibodies. Cells were pretreated with 5 µmol/L MG132 for 12 h. (M‐N) Representative flow cytography (M) and quantification (N) depicted the process of macrophages mediated phagocytosis of J82 cells treated with trametinib In vitro compared to the control group. Boxes were representative of GFP⁺ CD11b⁺ macrophages. (O) Network view of predicted E3 ubiquitin ligase‐MAN1B1 interactions by UbiBrowser 2.0. (P) Analysis of interaction between HRD1 and MAN1B1 by IP. (Q) Western blotting analysis of HRD1 knockout in J82 cells. (R) Western blotting analysis of stability of MAN1B1 in J82 sgControl or sgMAN1B1 cells. (S‐T) Correlation between MAN1B1 and HRD1 in ten human fresh bladder cancer samples using Western blotting. N, matched normal tissue; T, tumor tissue. (U‐V) Analysis of MAN1B1‐HRD1 binding by IP when ERK was phosphorylated in J82 and 293T cells. (W) Analysis of MAN1B1‐HRD1 binding by IP when ERK was dephosphorylated in UMUC3 cells. P values (A, H and N) were determined by a two‐tailed unpaired Student's t‐test, error bars represent mean ± SD. P value (T) was calculated by the Pearson correlation test. NS, not significant; **P < 0.01.

Abbreviations: CD47, cluster of differentiation 47; SIRP‐α, signal‐regulatory protein alpha; MAN1B1, mannosidase alpha class 1B member 1; WCL, whole cell lysates; IP, immunoprecipitation; HRD1, HMG‐CoA Reductase Degradation; ERK2, extracellular signal‐regulated kinase 2; KRAS‐G12D, Kirsten Rat Sarcoma Viral Oncogene Homolog G12D; GAPDH, Glyceraldehyde‐3‐phosphate dehydrogenase; EGF, epidermal growth factor; MG132, Carbobenzoxy‐L‐leucyl‐L‐leucyl‐L‐leucinal; GFP, green fluorescent protein; SYVN1, synoviolin 1; CD147, cluster of differentiation; DMSO, dimethyl sulfoxide; Vec, vector; CHX, cycloheximide; SD, standard deviation.

Phosphorylation and ubiquitination are two crucial post‐translational modifications that can influence the function and expression of target proteins [36, 37]. Rampias et al. [38] revealed that mutations in the NOTCH receptor could result in abnormal phosphorylation of ERK, a hallmark of ERK signaling pathway activation, and were found in 43% of bladder cancer.We investigated whether the ERK phosphorylation increased MAN1B1 expression. Upon treating 5637 and J82 cells with trametinib, a mitogen‐activated protein kinase kinase inhibitor, we observed a decrease in MAN1B1 expression which was associated with reduced levels of ERK phosphorylation (Figure 6B‐C). Consistently, we performed transient and stable transfections of bladder cancer cells with KRAS‐G12D and ERK2, which resulted in an increased expression of MAN1B1 (Figure 6D‐G). qRT‐PCR confirmed that the mRNA level of MAN1B1 was not significantly elevated when cells were stably transfected with KRAS‐G12D (Figure 6H). This suggested that ERK activation affected the stabilization of MAN1B1 via post‐translational modification.

MAN1B1 and ERK2 were found to interact by IP (Figure 6I). The expression of exogenous MAN1B1 was closely linked to the ERK activated level, while the co‐expressed GFP protein exhibited no significant changes (Figure 6J). The CHX experiments revealed that the stability of MAN1B1 was positively associated with the ERK phosphorylation (Figure 6K). ERK1/2 knockout markedly increased the ubiquitination of MAN1B1 (Figure 6L). Inhibiting ERK activity in bladder cancer cells with trametinib markedly elevated the phagocytosis by macrophages (Figure 6M‐N).

ERK phosphorylation typically stabilized proteins by modulating their affinity to E3 ubiquitin ligases [39]. Building on this understanding, we explored the E3 ligase of MAN1B1. As per the database UbiBrowser [40], synoviolin 1 (SYVN1), also known as HRD1, was identified as a potential E3 ligase for MAN1B1 (Figure 6O). IP revealed that MAN1B1 interacted with HRD1 (Figure 6P). MAN1B1 increased significantly following HRD1 knockout in J82 (Figure 6Q). The CHX experiments demonstrated that the stability of MAN1B1 and CD147 increased in HRD1‐KO J82 (Figure 6R). CD147 has been reported as a substrate of HRD1 [41]. Significantly reduced expression of HRD1 was negatively correlated with the high expression of MAN1B1 in bladder cancer tissue (Figure 6S‐T and Supplementary Table S6).

Upon the activation of ERK through epidermal growth factor, we found that the affinity between MAN1B1 and HRD1 decreased with increased ERK phosphorylation (Figure 6U‐V). Similarly, inhibiting ERK phosphorylation by trametinib, the affinity between MAN1B1 and HRD1 was significantly increased (Figure 6W). Collectively, these findings indicated that elevated level of ERK activation in bladder cancer contributed to the increased stability of MAN1B1 by regulating the interaction between MAN1B1 and HRD1, resulting in high expression levels of MAN1B1 and promoted bladder cancer immune evasion.

4. Discussion

This study investigated the role and regulation of CD47 to thoroughly comprehend immunotherapy for bladder cancer. Our study found that the glycosylation of CD47 played a vital role in bladder cancer, and the glycosyltransferase MAN1B1 mainly mediated this effect. Additionally, we identified that targeting MAN1B1 can achieve the therapeutic strategy of anti‐CD47 immunotherapy for bladder cancer without causing anemia. This was critical in CD47‐associated anti‐tumor immunotherapy. This study demonstrated that ERK could stabilize MAN1B1 by regulating its interaction with HRD1. These results underscored the promising therapeutic potential of targeting MAN1B1 in bladder cancer.

Our study supports the notion that CD47 releases a “don't eat me” signal, enabling cancer cells to evade the immune system. CD47 was found to play a pivotal role in cancers [42]. Our findings confirmed that extensive glycosylation of CD47 was essential in the immune evasion, modulating the affinity between CD47 and its ligand. This offered a fresh perspective on bladder cancer immunotherapy.

Protein glycosylation is primarily regulated by glycosyltransferases. Multiple glycosyltransferases, including MAN1B1, MOGS, MGAT2, and GANAB, were identified using mass spectrometry from J82 and 293T cells. However, subsequent knockout or knockdown experiments with MOGS, MGAT2, and GANAB demonstrated that their presence did not alter the affinity between CD47 and its ligand. IP and functional validation identified that MAN1B1 was the specific glycosyltransferase for CD47 in bladder cancer. Our study confirmed that MAN1B1 deficiency weakened CD47‐mediated anti‐phagocytosis through in vivo and in vitro experiments. MAN1B1 was reported to be associated with a poor prognosis of bladder cancer, consistent with our findings [43, 44, 45]. MAN1B1 modified CD47 and led to immune evasion in bladder cancer, indicated a potential mechanism by which elevated expression of MAN1B1 influenced the prognosis of bladder cancer. Reportedly, MAN1B1 mediated solute carrier family 5 member 5 in breast cancer [46] and is highly expressed in prostate cancer histochemical staining, indicating its diverse roles in cancer progression [47].

Although CD47 is an important immune checkpoint, and immunotherapies targeting CD47 are undergoing clinical trials in the 2a/2b phases. However, its broad expression in normal cells, especially erythrocytes and platelets, has exhibited adverse effects such as anemia in animal experiments and clinical trials. This limits the application of CD47 monoclonal antibody treatment [48]. Meanwhile, recent studies on CD47 expression in bladder cancer illustrated an indefinite relationship between CD47 expression and prognosis or pathological type [21]. These findings corresponded to the results of this study, indicating that CD47‐mediated tumor immunosuppression might be mediated via regulating post‐translational modifications. QPCTL is a published enzyme responsible for the interaction of CD47 with SIRP‐α. ISM004‐1057D is a small molecule inhibitor that targets QPCTL to achieve anti‐tumor effects by blocking the signal. In vivo preclinical studies demonstrated anti‐tumor efficacy in hematologic tumors and solid tumors. This suggests the possibility of immunotherapy by targeting CD47‐modifying enzymes.

Our study provided evidence that glycosyltransferase MAN1B1 might inhibit the innate tumor immune response through glycosylation of CD47. It provided an important indicator for the selection of targeted CD47 therapeutic strategies. Interestingly, we found low level of MAN1B1 in erythrocytes. Our data implied that inhibitors targeting MAN1B1 perform immunotherapeutic functions without adverse effects such as anemia in vivo. Thus, MAN1B1 as a target could treat CD47‐mediated immune evasion in bladder cancer with highly expressed MAN1B1. Monoclonal antibodies are widely used in tumor immunotherapy. However, the large molecular weight of monoclonal antibodies limits them from crossing the tumors' biological barrier [49, 50, 51]. Small‐molecule drugs can more easily penetrate biological barriers than monoclonal antibodies [52]. In contrast to the clinically required concentration of monoclonal antibodies, only a small concentration of MAN1B1 inhibitors was required to achieve effective tumor immunotherapy in the PTCs model in this study. Inhibitor therapy targeting MAN1B1 could be an innovative method in immunotherapy. In addition, proteolysis targeting chimera and ubiquitin‐proteasome systems are alternative strategies for tumor immunotherapy in glioma and breast cancer [53, 54]. This novel drug strategy provides a variety of options for bladder cancer immunotherapy.

Our results indicated that ERK activation could stabilize MAN1B1 by regulating its interaction with HRD1. Our study found that increasing the affinity of HRD1 and MAN1B1 through ERK inhibitors could effectively reduce MAN1B1 expression and promote tumor immune efficacy. This suggested that, in addition to MAN1B1 inhibitors, the combined use of ERK inhibitors could also be considered in immunotherapy strategies in bladder cancer. Applying ERK inhibitors in T cells could affect the interaction between PD‐1 and ubiquitin specific peptidase 5, degrading PD‐1 and promoting tumor immunotherapy [39]. Thus, targeting ERK with trametinib in tumors may enable tumor immunotherapy through multiple mechanisms. Previous studies reported that HRD1‐mediated proteostasis network promotes tumor resistance [55]. HRD1 is an E3‐ubiquitin ligase associated with ER degradation pathways. Studies reported that PD‐L1 can be ubiquitinated by Ovarian tumor domain‐containing ubiquitin aldehyde‐binding protein 1/2 and then degraded by endoplasmic reticulum‐associated degradation (ERAD) to enhance tumor immune evasion [56, 57]. This suggested that targeting ERAD could be a possible anti‐tumor therapy. Our findings offer many potential immunotherapy approaches against bladder cancer.

Although this study provided sufficient evidence from biochemical analyses and several mouse tumor models to highlight the important role of MAN1B1 in regulating CD47‐mediated immune evasion in bladder cancer, as well as the safety of targeting MAN1B1. Our study still had limitations. First, the sample size of the overall experiment was small. Many of the experiments relied on cell lines such as J82 and T24, which represented bladder cancer cell lines with the highest and lowest CD47 expression levels and male and female origin, respectively. However, there could be discrepancies among various cell lines of the same tumor. Knock‐in and knock‐out mice can be used to investigate the function of MAN1B1 in other cell types. Due to the rigorous sample collection requirements for PTCs, only data from three patients were included in this study. We hope for more data from the PTCs model in the future. Further basic and clinical data are still needed to confirm the function of targeting MAN1B1 in bladder cancer immune evasion, as well as to assess its safety.

5. Conclusions

This study found that glycosylation of CD47 in bladder cancer was critical for releasing immunosuppressive signals by influencing the interaction of CD47 with SIRP‐α. The expression of glycosyltransferase MAN1B1 was significantly elevated in bladder cancer. The absence of MAN1B1 considerably impaired the anti‐phagocytosis mediated by CD47 in bladder cancer. Furthermore, activation of ERK was found to be an essential factor in regulating the stability of MAN1B1 by modulating the interaction between MAN1B1 and the E3 ligase HRD1 in bladder cancer. Targeting MAN1B1 demonstrated significant antitumor activity and exhibited no adverse effects such as anemia, offering a promising strategy for bladder cancer immunotherapy. This study systematically demonstrated the functional importance of glycosylated CD47 and proposed a potentially safe treatment strategy for targeting CD47‐specific glycosyltransferase, MAN1B1, in bladder cancer.

AUTHOR CONTRIBUTIONS

Jie Zhang, Chen Zhang, Guoqing Ding, Yicheng Chen and Yanlan Yu conceived the study and designed the experiments; Jie Zhang, Chen Zhang and Ruichen Zang performed most of the experiments assisted by Weiwu Chen, Haofei Jiang, Yining Guo, Jing Le, Kunyu Wang, Haobo Fan, Xudong Wang, Sisi Mo, Peng Gao, Wenhao Guo, Xinrong Jiang, Fengbin Gao, Yuxing Chen, Junming Jiang and Juyan Zheng. Jie Zhang, Yicheng Chen, Guoqing Ding, Yili Fu, Yanlan Yu and Fengbin Gao provided funding as well as technical and material support. Jie Zhang and Chen Zhang analyzed the data and wrote the manuscript. Yuxing Chen, Junming Jiang and Juyan Zheng generated SIRP‐α‐Fc protein; All authors reviewed the manuscript and provided final approval for submission.

CONFLICT OF INTEREST STATEMENT

The authors declare no potential conflict of interest.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

This study was approved by the Ethics Committee of the Sir Run Run Shaw Hospital of Zhejiang University (No.20230312), all patients were informed and consented, and it was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. All animal experiments were approved by the Animal Care and Use Committee of the Sir Run Run Shaw Hospital of Zhejiang University (SRRSH202202086), and performed in accordance with this committee's guidelines.

Supporting information

Supporting Information

Zhang J, Zhang C, Zang R, Chen W, Guo Y, Jiang H, et al. Targeting MAN1B1 potently enhances bladder cancer antitumor immunity via deglycosylation of CD47. Cancer Commun. 2025;45:1090–1112. 10.1002/cac2.70040

DATA AVAILABILITY STATEMENT

All data in this article are legally available after publication