Exosomal B7H3 drives immunosuppression and serves as a biomarker for PD-1 blockage response in head and neck squamous cell carcinoma

Cell culture

The HNSCC cell lines {CAL27 (human tongue squamous cell carcinoma (SCC)), SCC9 (human tongue SCC), SCC25 (human tongue SCC), SCC47 (HNSCC), SCC154 (human tongue SCC), and SCC7 (mouse SCC)} were obtained from the American Type Culture Collection (ATCC) (Manassas, VA, USA). B7H3-overexpressing cell lines (SCC47-oeB7H3, CAL27-oeB7H3, and SCC154-oeB7H3) were established in our laboratory via lentiviral transduction. All cell lines were cultured in DMEM (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco) and 100 U/mL penicillin–streptomycin. Cell line identity was confirmed by DNA fingerprinting, and routine testing was performed to ensure the absence of mycoplasma contamination. Cells were used within 1 month of revival and maintained at 37°C in a humidified atmosphere containing 5% CO₂.

Generation of B7H3-overexpressing HNSCC cells

Stable B7H3 overexpression was achieved using a lentiviral expression system. Briefly, HEK-293T packaging cells were seeded and cultured for 24 h prior to transfection. The cells were co-transfected with 4 μg of the pSin-EF2-cDNA expression vector, 3 μg of psPAX2 (packaging), and 1 μg of pMD2G (envelope) using 24 μL of polyethylenimine (PEI) (2 mg/mL). Viral supernatants were harvested 48 h post-transfection and clarified through a 0.45-μm PVDF membrane filter (Millipore, Burlington, MA, USA). Target HNSCC cells were subsequently transduced with the lentiviral particles at an optimized multiplicity of infection (MOI) in the presence of 10 μg/mL of polybrene (Sigma-Aldrich, St. Louis, MO, USA). To enhance transduction efficiency, the plates were centrifuged at 2,000 ×g for 2.5 h at 37°C. Transduced cells were selected and maintained under puromycin pressure for at least 7 d to ensure stable transgene expression.

Specimen collection

Blood samples were collected from healthy donors, patients with primary SCC, and patients with metastatic SCC at The Fifth Affiliated Hospital, Sun Yat-sen University. Written informed consent was obtained from all participants prior to sample collection. All procedures involving human samples were conducted in accordance with relevant ethical guidelines and regulations. In addition, HNSCC tumor tissues and matched adjacent normal tissues were obtained from patients at the same institution. The study protocol was approved by the Ethics Committee of The Fifth Affiliated Hospital, Sun Yat-sen University and all participants provided written informed consent (chiCTR registration number: ChiCTR2400089207).

Flow cytometry

To analyze cell surface protein expression, single-cell suspensions were prepared from cultured cells. Cells were stained with primary antibodies for 30 min at 4°C. After washing with staining buffer, secondary antibodies were applied when required by the experimental design. Samples were analyzed on a CytoFLEX LX flow cytometer (Beckman Coulter, Brea, CA, USA) equipped with 405 nm, 488 nm, and 638 nm lasers. Fluorescence was detected using the following filter configurations: FITC (525/40 nm); PE (585/42 nm); and APC (660/20 nm). A minimum of 10,000–100,000 events were acquired per sample. Data analysis was performed using CytExpert 2.4 software (Beckman Coulter).

Purification of exosomes

Exosomes were purified from cell culture supernatants following established protocols. Cells were cultured in medium supplemented with 10% exosome-depleted FBS, prepared by overnight centrifugation at 100,000 × g to remove bovine-derived vesicles. Supernatants from 48–72-h cultures were collected and subjected to differential centrifugation. First, samples were centrifuged at 2,000 × g for 20 min to remove cell debris and dead cells (Sorvall ST40R; Thermo Scientific, Waltham, MA, USA). The supernatant was then centrifuged at 16,500 × g for 45 min to pellet microvesicles, which were resuspended in phosphate-buffered saline (PBS). The resulting supernatant was ultracentrifuged at 100,000 × g for 2 h at 4°C to pellet exosomes. The exosome pellet was washed and resuspended in PBS, followed by an additional ultracentrifugation step at 100,000 × g for 2 h.

Venous citrated blood from HNSCC patients or healthy donors was centrifuged at 1,550 × g for 30 min to obtain cell-free plasma (Sorvall ST40R) to purify circulating EVs from blood samples. One milliliter of plasma was centrifuged at 16,500 × g for 45 min to pellet microvesicles, which were resuspended in PBS. The supernatant was then ultracentrifuged at 100,000 × g for 2 h at 4°C to isolate exosomes. To further enhance purity, exosomes were extracted from the supernatant using a commercial exosome isolation kit (catalog # 4484450; Invitrogen, Carlsbad, CA, USA).

Characterization of purified exosomes

The morphology of purified exosomes was verified using transmission electron microscopy (TEM). Exosomes suspended in PBS were applied to formvar carbon-coated nickel grids, stained with 2% uranyl acetate, air-dried, and visualized using a JEM-1011 transmission electron microscope (JEOL; Tokyo, Japan). The size distribution and concentration of exosomes isolated from cell culture supernatants or patient plasma were analyzed using a NanoSight NS300 system (Malvern Instruments, Malvern, Worcestershire, UK), which uses nanoparticle tracking analysis (NTA) based on fast video capture and Brownian motion measurement.

Iodixanol density gradient centrifugation was performed for further purification and validation. Exosomes harvested via differential centrifugation were layered onto a discontinuous iodixanol gradient (5%, 10%, 20%, and 40% solutions prepared by diluting 60% OptiPrep iodixanol with 0.25 M sucrose/10 mM Tris buffer) and centrifuged at 100,000 × g for 18 h at 4°C (Optima MAX-XP; Beckman Coulter). Twelve fractions of equal volume were collected from the top of the gradient. Exosome-containing fractions were identified and pooled, and exosomes were pelleted by additional ultracentrifugation at 100,000 × g for 2 h at 4°C.

ELISA

ELISA plates (96-well; Biolegend, San Diego, CA, USA) were coated with monoclonal antibody against B7H3 [0.25 μg per well (100 μL); Biolegend] overnight at 4°C for detection of B7H3 on EVs, cell supernatants, or patient plasma. Free binding sites were blocked with 200 μL of blocking buffer (Pierce, Rockford, IL, USA) for 1 h at room temperature (RT). Then, 100 μL of EV samples purified from cell culture supernatants were added to each well. The exosome samples purified from cell culture supernatants were prepared by serial dilution according to the total protein level to analyze enrichment of B7H3 on exosomes. The concentration of B7H3 on the surface of exosomes isolated from cells was calculated based on the linear range of the ELISA assay data.

Treatment of CD8⁺ T cells with exosomes

Purified exosomes (200 μg) were incubated with B7H3 blocking antibodies (10 μg/mL; Bio X Cell, Lebanon, NH, USA) or IgG isotype antibodies (10 μg/mL; Proteintech) in 100 μL of PBS to block B7H3 on the exosome surface, then washed with 30 mL of PBS and pelleted by ultracentrifugation to remove the non-bound free antibodies. Human CD8⁺ T cells purified from peripheral blood using immunodepletion on a Ficoll-Hypaque gradient (RosetteSep; StemSep Technologies, Vancouver, BC, Canada) or mouse CD8⁺ T cells purified from splenocytes using a Dynabeads Untouched Mouse CD8 Cells Kit (Invitrogen) were stimulated with anti-CD3 (2 μg/mL; Biolegend) and anti-CD28 (2 μg/mL; Biolegend) antibodies for 24 h, then incubated with human HNSCC cell-derived exosomes or mouse SCC7-oeB7H3 cell-derived exosomes with or without B7H3 blocking for 48 h in the presence of anti-CD3/CD28 antibodies. The treated cells were then collected, stained, and analysed by flow cytometry.

Exosome–T cell binding assay

To assess physical interactions between HNSCC-derived exosomes and CD8⁺ T cells, purified exosomes were labeled with carboxyfluorescein succinimidyl ester (CFSE) (5 μM; Abcam) in 100 μL of PBS for 30 min at 37°C. Unbound dye was removed by washing with 10 mL of PBS, followed by ultracentrifugation to pellet the stained exosomes. Pre-stimulated human CD8⁺ T cells (2 × 10⁵ cells per well in a 96-well plate) were incubated with CFSE-labeled exosomes (25 μg/mL) for 2 h at 37°C. Cells were then fixed and prepared for analysis by flow cytometry or confocal microscopy. Cells were immunostained with anti-CD8 antibody to confirm T cell identity for microscopy.

Immunofluorescence staining

Cells were fixed and permeabilized with 0.1% Triton X-100 for 10 min for immunofluorescence analysis, then blocked with 3% bovine serum albumin (BSA) in PBS for 1 h at RT. Primary antibodies were applied and incubated overnight at 4°C. Fluorophore-conjugated secondary antibodies were added after washing and incubated for 1 h at RT. Nuclei were counterstained with DAPI. Coverslips were mounted using antifade mounting medium and images were acquired using a Nikon confocal microscope under 100× magnification.

Western blot analysis

Whole-cell lysates or exosomal proteins were separated by 10% SDS–PAGE and transferred onto 0.45 μm PVDF membranes (Millipore). Membranes were blocked with 5% non-fat dry milk for 1 h at RT, then incubated overnight at 4°C with primary antibodies at manufacturer-recommended dilutions. Membranes were probed with HRP-conjugated secondary antibodies (Cell Signaling Technology Danvers, MA, USA) for 1 h at RT after washing. Protein bands were visualized using ECL detection reagents (Pierce). Antibodies against Alix, CD9, TSG101, and CD81 were used as exosome markers; GAPDH served as the loading control.

In vivo mouse studies

All animal procedures were performed in compliance with the Principles of Vertebrate Animal Utilization and Care and the Guidelines for the Care and Use of Laboratory Animals. Female C3He and nude mice (6–8 weeks old) were purchased from Beijing Vital River Laboratory Animal Technology (Beijing, China). CAL27 cells (5 × 10⁶ cells in 100 μL medium) were injected subcutaneously into the flanks of 8-week-old female athymic nude mice to establish a human xenograft model. The tumor dimensions were measured regularly using a digital caliper and tumor volume was calculated as the (width² × length)/2. Mice were euthanized 30 d post-inoculation or when the longest diameter of the tumor reached 2.0 cm. Blood was collected via cardiac puncture immediately after euthanasia for exosome isolation. Exosomes purified from age-, gender-, and weight-matched healthy nude mice without tumors were used as controls.

SCC7 cells (5 × 10⁵ cells in 100 μL medium) were injected subcutaneously for the syngeneic HNSCC model in immunocompetent C3He mice. Mice received tail vein injections of exosomes (100 μg in 100 μL of PBS) every 2 d starting on day 7. Monoclonal antibody treatments were administered via tail vein injection on the same schedule from day 14. CD8⁺ T cells were isolated from harvested tissues and single-cell suspensions were prepared for flow cytometric analysis.

Immunohistochemistry

Tissue samples were obtained from patients at The Fifth Affiliated Hospital, Sun Yat-sen University. Written informed consent was provided from all participants and the study protocol was approved by the Medical Research Ethics Committee of The Fifth Affiliated Hospital, Sun Yat-sen University (Approval No.: 2022-L077-1) in accordance with the Declaration of Helsinki. Formalin-fixed, paraffin-embedded tissues were sectioned at a 4–5 μm thickness, mounted on adhesive slides, and dried overnight at 37°C. Sections were deparaffinized in xylene and rehydrated through a graded ethanol series [100%, 95%, and 70% (5 min each)] followed by distilled water. Heat-induced antigen retrieval was performed using citrate buffer (pH 6.0) in a decloaking chamber at 95°C for 20 min, after which slides were cooled to RT for 30 min. Endogenous peroxidase activity was blocked by incubating sections in 3% hydrogen peroxide in methanol for 10 min at RT. Non-specific binding was minimized by treatment with 5% normal serum (matched to the secondary antibody host species) in PBS for 30 min at RT. Sections were incubated with primary antibodies diluted in blocking buffer overnight at 4°C in a humidified chamber. A horseradish peroxidase (HRP)-conjugated secondary antibody was applied for 60 min at RT after washing. Antigen-antibody complexes were visualized using 3,3′-diaminobenzidine (DAB) chromogen with the development time (2–10 min) monitored under microscopy.

Antibodies

The following antibodies were used in this study: anti-B7H3 (AB134161, 1:1000; Abcam, Waltham, MA, USA); anti-ALIX (12422-1-AP, 1:1000; Proteintech, Wuhan, China); anti-CD9 (sc-13118, 1:200; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA); anti-CD81 (sc-166029, 1:200; Santa Cruz Biotechnology, Inc.); anti-GM130 (AB52649, 1:1000; Abcam): anti-GAPDH (10494–1-AP, 1:2000; Proteintech); anti-HSP70 (sc-24, 1:5000; Santa Cruz Biotechnology, Inc.); anti-fibronectin (AB2413, 1:1000; Abcam); anti-TSG101 (HPA006161, 1:1000; Sigma, St. Louis, MO, USA); anti-CD8a (AB217344, 1:1000; Abcam); and anti-PD-L1 (GTX104763, 1:1000; GeneTex, Irvine, CA, USA).

Statistical analyses

All statistical analyses were performed using GraphPad Prism software (version 8.3.0; GraphPad Software, San Diego, CA, USA). The normality of data distribution was assessed using the Shapiro–Wilk test prior to performing parametric tests and homogeneity of variances was verified using the Bartlett’s or Brown–Forsythe test, as appropriate.

Differences between two groups were assessed using two-tailed Student’s t-test. One-way ANOVA was used for comparisons among three or more groups, followed by appropriate post-hoc tests with multiple comparison correction. Specifically, Tukey’s honestly significant difference (HSD) test was applied for all pairwise comparisons using one-way ANOVA. Sidak’s multiple comparisons test was applied for two-way ANOVA analyses. All post-hoc tests were adjusted for multiple comparisons to control the family-wise error rate.

The sample sizes were determined based on preliminary experiments and relevant literature, although a formal power analysis was not performed. In addition to P-values, effect sizes [Cohen’s d for t-tests, partial eta squared (η²) for ANOVA] are reported where appropriate to provide estimates of the magnitude of observed effects.

TEM

The size, morphology, and structural features of EVs and particles (EVPs) were characterized using TEM. Briefly, 20 μL of the EVP suspension was applied to carbon-coated nickel grids (LG14-106-3S) and incubated for 10 min at RT. Grids were then carefully blotted with filter paper to remove excess liquid and placed onto droplets of 1% phosphotungstic acid solution (G1870-100 mL; Solarbio, Beijing, China) for 30 s (negative staining). Residual stain was removed by gentle blotting and the grids were air-dried. Samples were visualized using a FEI Tecnai G2 Spirit TWIN transmission electron microscope (FEI Company, Hillsboro, OR, USA) operating at 100 kV.

Generation of stable Rab27a or B7H3 knockdown HNSCC cells

Short hairpin RNAs (shRNAs) against human Rab27a (GCTGCCAATGGGACAAACATA, CAGGAGAGGTTTCGTAGCTA), human B7H3 (GGCAGCTGACAGATACCAAAC, GCTGACAGACACCAAACAGCT), or scrambled shRNA-control were packaged into lentiviral particles using 293T cells co-transfected with the viral packaging plasmids. Lentiviral supernatants were harvested 48–72 h after transfection. Cells were infected with filtered lentivirus and selected by 2 μg/mL of puromycin.

Identification of exosomal B7H3 in HNSCC

Tumor-derived exosomes function as critical mediators within the tumor immune microenvironment through the delivery of diverse bioactive cargoes, including key immune checkpoint proteins. To investigate whether B7H3 is incorporated into exosomes secreted by HNSCC cells, we isolated EVs from a series of HNSCC cell lines via standardized differential ultracentrifugation. These vesicles displayed typical exosomal morphology and size distribution, as verified by TEM and NTA (Figure 1A). Subsequent Western blot analysis detected the enrichment of canonical exosomal markers (e.g., Alix, CD9, and CD81), and importantly, verified the abundant presence of B7H3 in these exosomal fractions (Figure 1B–D). Quantitative ELISA further validated the high levels of B7H3 protein in exosomes derived from various HNSCC cell lines (Figure 1G).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Expression of B7H3 on HNSCC cell-derived exosomes. (A) Representative TEM image of exosomes purified from HNSCC cell lines and NTA for size and concentration characterization. (B–D) Western blot analysis of B7H3 expression in WCL and purified exosomes from HNSCC cell lines [CAL27 (B), SCC9 (C), and SCC25 (D)]. Equal amounts of total protein were loaded in each lane. (E, F) Western blot detection of B7H3 in WCL and exosomes from B7H3-overexpressing HNSCC cell lines [SCC47 (E) and CAL27 (F)]. (G) Quantification of exosomal B7H3 levels by ELISA in control and B7H3-overexpressing HNSCC cells. Data are represented as the mean ± SD in biologically independent experiments. (H) Iodixanol density gradient centrifugation demonstrating co-fractionation of B7H3 with exosomal markers (Alix and CD9) in exosomes derived from HNSCC cell lines. ELISA, enzyme-linked immunosorbent assay; HNSCC, head and neck squamous cell carcinoma; NTA, nanoparticle tracking analysis; SD, standard deviation; TEM, transmission electron microscope; WCL, whole cell lysate.

HNSCC cell lines with stable B7H3 overexpression were established to conclusively confirm that B7H3 is actively loaded into exosomes. Western blot analysis showed markedly increased B7H3 protein levels not only in whole-cell lysates, but also more importantly, in the corresponding exosomal fractions relative to control cells (Figure 1E, F). These findings suggest that cellular B7H3 expression levels directly determine the exosomal abundance.

Iodixanol density gradient ultracentrifugation was performed to further rule out potential contamination from non-exosomal vesicles or protein aggregates. The results showed that B7H3 co-localized exactly with fractions positive for classical exosomal markers, confirming a specific association with bona fide exosomes instead of other EVs or protein aggregates (Figure 1H).

Taken together, these data clearly demonstrated that B7H3 is a specific component of the exosomal cargo secreted by HNSCC cells.

Secretion of exosomal B7H3 is increased in HNSCC

A human HNSCC xenograft model was established in immunodeficient nude mice to confirm the secretion of tumor-derived exosomal B7H3 in a physiologic setting. Serum exosomes were isolated and analyzed for human-specific B7H3 by ELISA (Figure 2A). Strong signals were detected in exosomes isolated from CAL27 tumor-bearing mice using species-specific antibodies that selectively recognize human B7H3, whereas no signal was detectable in tumor-free control mice (Figure 2B). Importantly, circulating exosomal B7H3 levels showed a significant positive correlation with tumor volume (Figure 2C), indicating that exosomal B7H3 can serve as a reliable indicator of in vivo tumor burden.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Exosomal B7H3 levels distinguish patients with HNSCC from healthy donors. (A) Detection of human B7H3 by ELISA in plasma exosomes from mice bearing human HNSCC xenografts. (B) Comparison of exosomal B7H3 levels via ELISA in plasma from control nude mice and mice carrying CAL27 xenografts (n = 10). P values were calculated using an unpaired Student’s t-test. Data are represented as the mean ± SD in biologically independent experiments. Cohen’s d = 3.92. (C) Pearson correlation analysis between circulating exosomal B7H3 levels and tumor burden in xenograft-bearing nude mice (n = 10). (D) ELISA quantification of circulating exosomal B7H3 in healthy donors (n = 14), patients with primary SCC (n =36), and patients with metastatic SCC (n = 20). P values were calculated using one-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. η² = 0.52. (E) Western blot analysis of B7H3 expression in exosomes purified from healthy donors (n = 6), primary SCC patients (n = 6), and metastatic SCC patients (n = 6). ANOVA, analysis of variance; ELISA, enzyme-linked immunosorbent assay; HNSCC, head and neck squamous cell carcinoma; SCC, squamous cell carcinoma; SD, standard deviation.

Serum exosomes from healthy donors and HNSCC patients were further analyzed to assess the clinical significance of these findings. Western blot and ELISA analyses demonstrated that exosomal B7H3 levels were markedly increased in patients with primary HNSCC compared to healthy controls, which was consistent with our preclinical data. Moreover, upon stratification by clinical stage, patients with metastatic HNSCC exhibited substantially higher exosomal B7H3 levels than patients with localized disease (Figure 2D, E).

Taken together, these results validated the presence of circulating exosomal B7H3 in HNSCC patients and highlight the potential as a non-invasive biomarker for monitoring tumor progression and metastatic potential.

Exosomal B7H3 inhibits CD8⁺ T cell functions in vitro

While cell surface B7H3 has been implicated in mediating CD8⁺ T-cell immunosuppression¹⁸, the specific functional role of the exosomal counterpart remains largely unknown. In this study whether tumor-derived exosomal B7H3 directly contributes to the suppression of CD8⁺ T-cell activity was investigated.

First, a stable B7H3-EGFP overexpressing SCC154 cell line (oeB7H3-EGFP) was generated (Figure 4A, B). Quantitative analysis of EVs demonstrated that B7H3 overexpression did not significantly affect total exosome secretion levels compared to vector-control cells (Figure 3A), thereby ensuring that any observed functional differences were attributable to exosomal cargo composition rather than vesicle quantity. Internalization of exosomal B7H3 by CD8⁺ T cells was verified using confocal microscopy (Figure 3B). Exosome uptake by CD8⁺ T cells was further validated by Western blot analysis of recipient T cells (Figure 3C).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Exosomal B7H3 inhibits CD8⁺ T cell functions in vitro. (A) Quantification of exosome secretion. oeB7H3-EGFP cells did not exhibit significant changes in exosome secretion compared to vector control cells. P values were calculated using an unpaired Student’s t-test. Data are represented as the mean ± SD in biologically independent experiments. Cohen’s d = 0.410. (B) Immunofluorescence staining showing internalization of exosomal B7H3 (derived from SCC154-oeB7H3-EGFP cells) by CD8⁺ T cells. (C) Western blot analysis of B7H3 and CD8a expression in PBMCs co-cultured with exosomes from SCC154-oeB7H3 cells at varying concentrations. (D, E) Western blot analysis demonstrates that exosomes derived from oeB7H3 cells enhance phosphorylation of ERK and AKT in CD8⁺ T cells (E) and Jurakt (F) cells, indicating activation of downstream signaling pathways. This effect is abolished by anti-B7H3 treatment. (F) Representative flow cytometry contour plots and quantitative summary of granzyme B (GZMB) and Ki-67 expression in human peripheral CD8⁺ T cells. P values were calculated using one-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. GZMB η² = 0.9992, Ki-67 η² = 0.9989. (G) Representative flow cytometry contour plots and quantitative summary of GZMB and Ki-67 expression in Jurakt cells. P values were calculated using one-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. GZMB η² = 0.9982, Ki-67 η² = 0.9989. ANOVA, analysis of variance; EGFP, enhanced green fluorescent protein; GZMB, granzyme B; oe, overexpression; PBMC, peripheral blood mononuclear cells; SD, standard deviation.

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Intracellular fate of exosomes in T cells. (A) Immunofluorescence validation of stable cell lines. SCC154 cells were stably transduced with a control EGFP or a B7H3-overexpressing construct (B7H3-EGFP). EGFP fluorescence (green) confirms successful expression of the transgenes. (B) Western blot analysis of exosomes isolated from SCC154 cells transfected with a B7H3-EGFP overexpression (oeB7H3-EGFP) or control (EGFP) plasmid. Blots show expression of exogenous and endogenous B7H3, and the presence of B7H3-EGFP in exosomes further verifies successful construction of the B7H3-overexpressing cell line and efficient loading of B7H3-EGFP into exosomes. (C) Representative confocal immunofluorescence images showing internalization of control EGFP exosomes (green) by recipient cells. Co-staining with LAMP1 (lysosomal marker, magenta) and PSMB5 (proteasome marker, red) indicates the subcellular distribution of the internalized exosomes. Merged images and corresponding line-scan profiles (right panels) demonstrate no significant co-localization of control exosomes with lysosomes (LAMP1) or proteasomes (PSMB5). (D) Representative confocal immunofluorescence images showing internalization of B7H3-EGFP exosomes (green) by T cells. Co-staining with LAMP1 (lysosomal marker, magenta) and PSMB5 (proteasome marker, red) indicates the subcellular distribution of the internalized exosomes. Merged images and corresponding line-scan profiles (right panels) demonstrate significant co-localization of exosomal B7H3 with lysosomes (LAMP1), while lower overlap is observed with proteasomes (PSMB5). EGFP, enhanced green fluorescent protein; LAMP1, lysosome-associated membrane protein 1; PSMB5, proteasome subunit beta 5.

Importantly, treatment with B7H3-enriched exosomes significantly impaired T-cell immune function, as reflected by inactivation of key T-cell activation signaling pathways (Figure 3D, E). This immunosuppressive effect was further confirmed by flow cytometry, which showed that exposure of T cells to B7H3-enriched exosomes markedly reduced the expression of the cytotoxic effector molecule granzyme B (GZMB) and the proliferation marker, Ki-67 (Figure 3F, G). Notably, this inhibitory effect was significantly reversed by treatment with B7H3-specific neutralizing antibodies, formally demonstrating that the observed immunosuppression was directly and specifically mediated by exosomal B7H3.

Co-immunofluorescence staining was performed using markers for major protein degradation pathways to further characterize the intracellular trafficking and fate of internalized exosomes. While control EGFP exosomes exhibited a random, dispersed pattern throughout the cytoplasm after internalization (Figure 4C), confocal imaging revealed substantial co-localization between exosomal B7H3 and the lysosomal marker, LAMP1. In contrast, no obvious co-localization was detected between internalized exosomes and the proteasomal subunit, PSMB5 (Figure 4D). Collectively, these findings indicated that following internalization by CD8⁺ T cells, exosomal B7H3 is selectively sorted into lysosomes for degradation rather than undergoing proteasomal processing.

An in vitro co-culture system comprising HNSCC cells and immune cells was established to better model the functional dynamics of exosomes within the in vivo TME (Figure 5E). In this system, HNSCC cells were genetically modified to knock down two critical molecules: B7H3; and Rab27a, a key GTPase that regulates exosome secretion (Figure 5A, B). The efficiency of these genetic manipulations was first validated by measuring exosome release. As expected, Rab27a knockdown markedly reduced the number of exosomes secreted by tumor cells, consistent with the well-characterized function in exosome biogenesis and trafficking (Figure 5C, D).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Exosomal B7H3 mediated immunomodulation in HNSCC cells. (A, B) Western blot analysis of protein levels in whole cell lysates (WCLs) and isolated exosomes from SCC9 (A) and CAL27 (B) cell lines to verify the effect of knock down. Cells were transduced with scrambled shRNA (shNC) or shRNAs targeting B7H3 (shB7H3#1, #2) or RAB27A (shRAB27A#1, #2). (C, D) Relative exosome secretion levels from SCC9 (C) and CAL27 (D) cells were determined by BCA. Knockdown of RAB27A, but not B7H3, significantly reduced exosome secretion. P values were calculated using one-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. η² = 0.9926 (C), η² = 0.9960 (D). (E) Schematic of the exosome-mediated immunomodulation assay. SCC9 or CAL27 cells are cultured in chambers; only exosomes can interact with the underlying human peripheral blood mononuclear cells (PBMCs) or Jurkat T cells to evaluate the effects on immune cell gene expression. (F, G), PBMCs and Jurkat cells were co-cultured with control (shNC), B7H3-knockdown (shB7H3), or RAB27A-knockdown (shRAB27A) SCC9 (F) or CAL27 (G) cells with or without B7H3/PD-1 antibody. The expression of immune-related genes (IFN-γ, GZMB) was analyzed by ELISA. Data are presented as the mean ± SD. ANOVA, analysis of variance; BCA, bicinchoninic acid; ns, no significance; PBMC, peripheral blood mononuclear cells; SD, standard deviation; sh, short hairpin; WCL, whole cell lysate.

Intriguingly, B7H3 knockdown resulted in significant restoration of immune cell function in the co-culture system. More notably, even when Rab27a was knocked down, leaving only a minimal residual pool of exosomes, these remaining vesicles still exerted potent immunosuppressive effects. These findings suggested that B7H3-carrying exosomes exhibit high biological potency and retain immunosuppressive activity, even at low concentrations (Figure 5F, G).

Neutralizing antibodies against B7H3 and PD-1 were used to further dissect the underlying mechanisms and compare the relative contributions. Strikingly, the observed immunosuppression was effectively reversed by B7H3 blockade, whereas PD-1 inhibition did not produce a comparable restorative effect (Figure 5F, G). These data indicated that exosomal B7H3 may have a more dominant role in HNSCC immune evasion than the classical PD-1/PD-L1 axis, identifying B7H3 as a more promising therapeutic target in this setting.

Taken together, these results demonstrated that exosomal B7H3, like the membrane-bound counterpart, directly interacts with immune cells to induce robust immunosuppression, thereby sustaining an immune-suppressive TME in HNSCC.

Exosomal B7H3 inhibits CD8⁺ T cells and facilitates HNSCC progression in vivo

A syngeneic HNSCC mouse model was utilized via subcutaneous implantation of SCC7 cells into immunocompetent C3H/He mice to formally establish the role of exosomal B7H3 in shaping the tumor immune microenvironment in vivo (Figure 6A). Systemic delivery of exosomes isolated from B7H3-overexpressing SCC7 cells (SCC7-oeB7H3) significantly promoted tumor growth compared to control groups (Figure 6B, C). The specificity of this pro-tumorigenic effect was further validated by pretreatment of SCC7-oeB7H3 exosomes with a B7H3-neutralizing antibody, which effectively reversed the tumor-promoting activity, whereas isotype-control IgG showed no significant effect.

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Exosomal B7H3 inhibits CD8⁺ T cell function in vivo. (A) Schematic representation of the experimental design for in vivo animal studies. (B) Tumor growth curves of SCC7 implants under different treatment conditions (n = 5 mice per group). The main effect [F(5, 48) = 2,248, P < 0.0001, η² = 0.89] is statistically significant. (C) Final tumor weights of SCC7 implants across treatment groups (n = 5 mice per group). P values were calculated by two-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. The main effect [F(3, 12) = 74.82, P < 0.0001, η² = 0.91] is statistically significant. (D) Representative flow cytometry contour plots showing IFN-γ expression in peripheral CD8⁺ T cells. (E) Quantitative analysis of IFN-γ⁺ CD8⁺ T cells among tumor-infiltrating lymphocytes (TILs), splenic, or lymph node populations following indicated treatments (n = 5). P values were calculated by two-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. The main effect [F(3, 12) = 99.15, P < 0.001, η² = 0.95] is statistically significant. (F) Representative flow cytometry contour plots showing granzyme B (GZMB) expression in peripheral CD8⁺ T cells. (G) Proportions of GzmB⁺ CD8⁺ T cells among TILs, splenic, or lymph node populations after indicated treatments (n = 5). P values were calculated by two-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. The main effect [F(3, 12) = 35.70, P < 0.001, η² = 0.087] is statistically significant. (H) Immunofluorescence staining of CD8⁺ T cells co-cultured with exosomes derived from SCC7-oeB7H3 cells in the presence of control IgG or anti-B7H3 antibody. (I) Statistical summary of CD8⁺ T cell counts per 10,000 cells under each culture condition. P values were calculated using one-way ANOVA. Data are represented as the mean ± SD in biologically independent experiments. η² = 0.9788. ANOVA, analysis of variance; GZMB, granzyme B; IFN-γ, interferon gamma; ns, no significance; SD, standard deviation; TILs, tumor-infiltrating lymphocytes.

Further profiling of the TME uncovered the immunologic mechanisms underlying accelerated tumor progression. Exposure of immune cells to oeB7H3 exosomes resulted in significant downregulation of key cytotoxic molecules, including IFN-γ and granzyme B (GZMB), to assess the functional impact of exosomal B7H3 on immune effector functions (Figure 6D–G). In vivo analyses revealed a marked reduction in CD8⁺ tumor-infiltrating lymphocyte (TIL) infiltration following treatment with B7H3-enriched exosomes. Notably, this inhibitory effect on T-cell recruitment and activation was largely restored by co-administration of a B7H3-blocking antibody (Figure 6H, I).

Taken together, these consistent and complementary findings demonstrated that exosomal B7H3 has a critical role in systemically suppressing antitumor immunity and driving HNSCC progression in vivo.

High expression of B7H3 indicates poor immunotherapy efficacy

Although ICIs targeting the PD-1/PD-L1 axis have revolutionized the management of advanced HNSCC, a substantial proportion of patients exhibit intrinsic resistance, resulting in suboptimal objective response rates. The heterogeneous expression patterns of alternative immune checkpoints are increasingly recognized as a key determinant of these variable clinical outcomes⁸. We hypothesized that HNSCC cells might preferentially exploit B7H3 as an alternative immune evasion mechanism, particularly in tumors with low PD-L1 expression or activity.

IHC staining was performed and tumor tissues from 32 HNSCC patients were analyzed to investigate the potential reciprocal relationship between B7H3 and PD-L1. The results revealed a significant inverse correlation between the expression levels of these two checkpoints: tumors with high B7H3 expression consistently had low PD-L1 levels; and high PD-L1 expression was associated with reduced B7H3 expression (Figure 7A, B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

Level of B7H3 expression is negatively correlated with PD-L1. (A) Representative immunohistochemical staining of B7H3 (left) and PD-L1 (right) in HNSCC tissue sections. (B) Pearson correlation analysis between B7H3 (H-score) and PD-L1 expression (H-score) in HNSCC patients (n = 32). (C) Pearson correlation analysis between circulating exosomal B7H3 and PD-L1 levels in plasma from HNSCC patients (n = 61). (D) Western blot analysis of B7H3 and PD-L1 expression in exosomes purified from HNSCC patient samples. (E) ELISA quantification of exosomal B7H3 levels in plasma from HNSCC patients with or without response to immunotherapy. P values were calculated by unpaired Student’s t-test. Data are represented as the mean ± SD in biologically independent experiments. Cohen’s d = 2.69. (F) ELISA quantification of exosomal PD-L1 levels in plasma from HNSCC patients with or without response to immunotherapy. P values were calculated by unpaired Student’s t-test. Data are represented as the mean ± SD in biologically independent experiments. Cohen’s d = 1.00. ELISA, enzyme-linked immunosorbent assay; HNSCC, head and neck squamous cell carcinoma; P1–P12, patient 1 to patient 12; SD, standard deviation.

Whether this reciprocal relationship was reflected in the circulating exosomal compartment was explored next. ELISA-based quantification of exosomes purified from the sera of 61 HNSCC patients validated a significant negative correlation between exosomal B7H3 and PD-L1 levels (Figure 7C, D). Most importantly, analysis of serial serum samples from patients receiving anti-PD-1 therapy unveiled a striking divergence in protein profiles. Specifically, patients who achieved a clinical response exhibited low baseline levels of circulating exosomal B7H3 in conjunction with elevated exosomal PD-L1, while non-responders displayed the opposite profile, characterized by high exosomal B7H3 and low PD-L1 levels (Figure 7E, F).

These findings suggested that HNSCC cells may selectively prioritize the upregulation of B7H3 or PD-L1 as a dominant immune escape mechanism. Furthermore, high exosomal B7H3 expression emerges as a major contributor to primary immunotherapy resistance, indicating that targeting the B7H3 axis could represent a promising salvage strategy to restore therapeutic efficacy in PD-L1-low HNSCC patients.