Hybrid bioinspired nanovescicles target tumor endothelial cells and regulate immuno-microenvironment for triple-negative breast cancer therapy
Zhengwei Gui, Lu Zhao, Shiyang Liu, Lin Zhang
TL;DR
A new nanovesicle targets tumor blood vessels and boosts immune response to fight triple-negative breast cancer.
Contribution
ox-HyVs are the first bio-nanovesicles engineered to reprogram tumor vasculature and activate immune pathways via endogenous miRNA.
Findings
ox-HyVs specifically target and internalize into tumor vascular endothelial cells.
Treatment with ox-HyVs inhibits tumor growth and extends survival in a TNBC mouse model.
ox-HyVs activate the cGAS-STING pathway and promote cytotoxic T cell infiltration into tumors.
Abstract
The treatment of triple-negative breast cancer (TNBC) remains challenging. Conventional anti-angiogenic therapies, which aim to disrupt the tumor’s blood supply, are often hampered by limited efficacy and drug resistance. Innovative strategies that specifically target the unique phenotype of tumor vascular endothelial cells (TVECs) are urgently needed. We developed peroxide-treated hybrid membrane bio-nanovesicles (ox-HyV) by fusing membranes from human induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) and breast cancer cells. The vesicles were characterized for size, morphology, and protein composition. Their targeting efficiency to TVECs was validated both in vitro and in vivo using immunofluorescence and small animal imaging. The functional effects on TVEC secretion, cGAS-STING pathway activation, and immune cell recruitment were assessed via ELISA, western blot,…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14
Figure 15
Figure 16
Figure 17
Figure 18
Figure 19
Figure 20- —http://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topicsinterferon and immune responses · Nanoplatforms for cancer theranostics · Extracellular vesicles in disease
Background
Triple-negative breast cancer (TNBC) is a biologically very aggressive subtype of breast cancer (BC) lacking estrogen and progesterone receptors and HER2 expression, rendering typical hormone and HER2-targeted therapies ineffective [1]. The heterogeneous and aggressive nature of this cancer leads to complex and challenging treatment. Although chemotherapy represents the current TNBC treatment, it has limited efficacy, especially when the disease has progressed to advanced stages and the prognosis of patients is usually poor [2–4]. Accordingly, new targeted therapeutic strategies should be developed. Notably, immunotherapy has shown potential benefits for TNBC [5]. Studies have shown that high levels of tumor-infiltrating lymphocytes exist in some TNBC subtypes, which is associated with diminished risk of disease recurrence and elevated benefit from chemotherapy [6]. Preliminary findings of immune checkpoint inhibitors in advanced TNBC are encouraging and suggest that immunotherapy may constitute a promising therapeutic option for TNBC patients. In TNBC, the immune microenvironment is crucial in determining prognosis. The tumor immune microenvironment (TIME) of TNBC is characterized by its complexity and heterogeneity, which influences both disease progression and treatment response [7, 8]. Immune cell spatial distribution and density, such as CD8 + T cells, within the tumor microenvironment (TME), have been shown to correlate with patient prognosis. A higher density of these cells is generally associated with better outcomes, whereas their absence or restricted presence within the tumor stroma is linked to poorer prognoses. This spatial heterogeneity underscores the importance of understanding the immune landscape in TNBC to develop more effective immunotherapeutic strategies [4].
While immunotherapy shows promise in many cancers, its effectiveness against TNBC remains limited. Research indicates that TNBC’s response to immunotherapy is influenced by multiple factors, including the immunosuppressive nature of the tumor microenvironment and the immune evasion mechanisms of cancer cells. First, the immune evasion mechanism of triple-negative breast cancer (TNBC) is a key factor limiting its response to immunotherapy. Research has shown that the transcription factor SOX4 plays a crucial role in immune escape by suppressing T-cell-mediated cytotoxicity in TNBC cells. Inhibiting SOX4 expression enhances the expression of genes involved in innate and adaptive immune pathways, thereby boosting tumor immunity [9]. Additionally, the integrin αvβ6 receptor on tumor cell surfaces regulates SOX4 expression through TGFβ activation. Blocking this pathway increases TNBC cells’ sensitivity to cytotoxic T cells [9]. Secondly, the immunosuppressive tumor microenvironment (TME) in triple-negative breast cancer (TNBC) is a key factor limiting the efficacy of immunotherapy. Studies indicate that tumor-associated macrophages (TAMs), the predominant infiltrating immune cells in TNBC, play a crucial role in the immunosuppressive TME. By regulating SOS1, TAM2 (pro-tumor) can be effectively polarized into TAM1 (anti-tumor), thereby reshaping the TME and enhancing immunotherapy outcomes [10]. Additionally, research has shown that PD-L1 overexpression leads to chemotherapy resistance in TNBC cells, while inhibiting PD-L1 significantly improves their sensitivity to chemotherapy [11]. Finally, while immune checkpoint inhibitors (ICIs) have shown promise in certain cancers, their efficacy in triple-negative breast cancer (TNBC) remains limited. Studies indicate that the effectiveness of ICIs in TNBC is constrained by metabolic heterogeneity among immune cell subtypes, which correlates with treatment resistance to immune checkpoint blockers (ICBs) [12]. Furthermore, research suggests that the therapeutic response to immunotherapy in TNBC may be influenced by extracellular matrix (ECM) components. Notably, specific ECM components such as versican (VCAN) may restrict tumor exposure to cytotoxic immune cells, thereby affecting the therapeutic response [13].
Recent studies have highlighted the potential of targeting specific markers on tumor endothelial cells (ECs) to inhibit angiogenesis and tumor growth. For instance, using TEM8-specific CAR T cells has displayed promise in targeting both the tumor cells and the tumor vasculature in TNBC [14]. TEM8, a marker initially identified on ECs, is upregulated in TNBC and can be targeted to disrupt neovascularization and tumor growth [15]. Moreover, the implication of centromere protein U (CENPU) in promoting angiogenesis in TNBC has been investigated [16]. CENPU inhibits COX-2 ubiquitin-proteasomal degradation, increasing COX-2-p-ERK-HIF-1α-VEGFA signaling activation, which is crucial for angiogenesis. Targeting CENPU could, therefore, be a viable strategy to reduce tumor vascularity and growth in TNBC [16].
Additionally, the concept of vascular detransformation, which involves reversing the abnormal transformation of ECs, offers another therapeutic avenue. This approach aims to normalize the tumor vasculature, thereby improving the efficacy of existing therapies and reducing tumor progression [15]. These studies underscore the importance of targeting the tumor vasculature, particularly the ECs, in developing effective treatments for TNBC. By altering the secretory phenotype of these cells, it may be possible to significantly impact tumor growth and metastasis, offering new hope for patients with this challenging form of BC.
Here, we designed hybrid membrane bio-inspired vesicles (ox-HyV) targeting tumor vascular ECs (TVECs), which could alter the secretory phenotype of TVECs and activate tumor innate immunity to kill tumors. Specifically, we first obtained peroxidized iPSC-iEC vesicles (ox-EcV) by repeated freeze-thawing and BC cell vesicles (CaM) by classical continuous extrusion and fused the two into hybrid membrane vesicles (ox-HyV). The abundance of chemokines from the surface of ECs and cancer cells made this heterodimeric membrane have good homologous targeting ability to TVECs. When these heterotrimeric membrane vesicles are targeted to the tumor vasculature, the microRNA cargoes loaded in them can activate tumor innate immunity and secrete a variety of pro-inflammatory cytokines by down-regulating Bcl2 in ECs, activating cGAS-STING signaling, and thus have a tumor-killing effect. Overall, our study suggests that targeting TVECs and altering their secretory phenotype may be a novel approach for TNBC immunotherapy.
Methods
Human induced pluripotent stem cell (iPSC) differentiation into ECs and iPSCs-ECs culture
Per protocols, human iPSC-ECs (Nuwacell Biotechnology Ltd) were cultured by first being expanded to the 3rd generation and then frozen and stored. After thawing, the cells were cultured (10,000–15,000 cells/cm^2^) onto tissue culture plates treated with 3 µg/cm^2^ fiber-conjugated proteins (Invitrogen) and passaged every 3–4 days with TrypLE (Gibco). The medium used to culture iPSC-ECs was VascuLife VEGF medium (Lifeline Cell Technologies) with 10 mL of glutamine supplement and EC growth additive per 500 mL of medium. For all experiments, cells were cultured at 37 °C and 5% CO2.
Cell lines and animals
The MDA-MB-231 and MDA-MB-468 cell lines were acquired from Shanghai Chuanqiu Biotechnology Co., Ltd, China. They were cultivated in DMEM or RPMI-1640 media (Gibco), all of which were enriched with 10% fetal bovine serum (Gibco BRL), 100 U/mL penicillin, and 100 µg/mL streptomycin. Operations on female BALB/c mice (18–22 g, six weeks, GemPharmatech) followed the regulations for the use of laboratory animal care at Tongji Hospital. Mice were euthanized by carbon dioxide asphyxiation (CO2) inhalation and cervical dislocation was performed as a secondary euthanasia procedure.
Isolation and culture of human primary BC-associated ECs
Tumor tissue was obtained from surgically removed samples from BC patients. The tissue was finely chopped and treated with collagenase II in a sterile setting. Contaminated hemocytes were effectively removed by employing sucrose gradient centrifugation with histopaque1077. Finally, the resulting cell suspension was filtered. Subsequently, HBCEC were separated using a magnetic cell sorting device and identified utilizing FITC-labeled anti-CD31 antibody (WM-59, eBioscience). Moreover, we isolated CD31-positive cells and distributed them onto culture plates covered with 1.5% gelatin and propagated in EGM-2 MV BulletKit (Lonza).
The primary vascular endothelial cells utilized in this study were isolated from TNBC tissue specimens, which were procured from patients receiving treatment at the Department of Breast and Thyroid Surgery, Tongji Hospital, affiliated with Tongji Medical College, Huazhong University of Science and Technology. The Department of Pathology at Tongji Hospital provided these specimens. Informed written consent was obtained from all participants, permitting the use of their residual pathological tissue and clinical data for research purposes. All research activities involving human subjects adhered to the ethical guidelines set forth by the Tongji Hospital Ethics Committee, in accordance with the 1964 Declaration of Helsinki and its subsequent amendments or comparable ethical standards.
Preparation and characterization of ox-HyV
BC cells were cultivated in T175 culture flasks until they fully merged. Consequently, they were gathered using a spatula, centrifuged (300 rpm, 5 min), and rinsed twice with PBS. To extract membrane proteins from cancer cells, the protocol outlined by the Abcam membrane protein extraction kit was meticulously followed. Initially, an osmotic buffer containing protease inhibitors was added to lyse the cancer cells, and the mixture was incubated for 15 min to facilitate cell swelling. Subsequently, the detergent buffer provided by the kit was introduced to gently lyse the cells. The lysate was then subjected to centrifugation at 4 °C and 700 × g for 10 minutes, after which the precipitate was discarded. The resulting supernatant was transferred to a new tube and further centrifuged at 4 °C and 10,000 × g for 30 min, with the precipitate again being discarded. To enrich membrane vesicles, sucrose density gradient ultracentrifugation was employed. A sucrose gradient was prepared in an ultracentrifuge tube, consisting of layers with concentrations of 60%, 45%, 35%, and 20% from bottom to top. The supernatant was carefully layered on top of this gradient and centrifuged at 4 °C and 100,000 × g for 3 h. The milky white bands within the target density range were collected, diluted with 10 volumes of pre-chilled PBS, and subjected to a final centrifugation at 4 °C and 100,000 × g for 1 h. Remove the supernatant and resuspend the pellet in PBS to achieve a concentration of 1 × 10^10^ vesicles/mL. Subsequently, add 5 µL of CD63-PE/PD-L1-APC to each 100 µL of vesicle suspension. Incubate the mixture at 4 °C in the dark for 30 min. Following incubation, add PBS to achieve a tenfold increase in volume and centrifuge the sample at 100,000 × g at 4 °C for 1 h. Discard the resulting supernatant and resuspend the pellet in 500 µL of PBS. Finally, analyze the sample using flow cytometry to isolate the CD63⁺/PD-L1⁻ subpopulation, which constitutes the CaM utilized in this study.
ECs differentiated from iPSCs were cultured in a medium with 20 µL/mL hydrogen peroxide for 24 h. After 24 h, EC cell membranes EcV were extracted and purified by an established extrusion method. i.e., the cells were suspended in PBS (1 × 10^6^) and sequentially passed via polycarbonate membranes with 5 and 1 μm holes using a micro-extruder. The extruded specimens were subjected to centrifugation (10,000 rpm, 10 min) to eliminate any cellular waste and large vesicles. EcV was purified and concentrated utilizing a 100 kDa centrifugal filter (1000 g, 15 min), and the purified BNV was stored at − 80 °C for 15 min.
Membrane protein concentrations of CaM and ox-EcV were assayed by the BCA Protein Kit. Ox-EcV and CaM, both labeled with DiO/DiI dual dye, were combined in equal proportions based on the weight of the membrane protein at a 1:1 ratio. The mixture was then subjected to sonication at 42 kHz and 130 w for 10 min at 37 °C to enhance the fusing of the membranes. The fluorescence spectra of each sample were measured at 500–650 nm, utilizing an excitation wavelength of 490 nm (excitation/emission: DiO, 484/501; DiI, 549/565 nm). The ox-HyV size distribution was analyzed utilizing DLS employing a zeta sizer nano zs90 instrument. The morphology of the HyVs was analyzed employing TEM.
qRT-PCR
Total RNA was extracted using TRIzol reagent (Invitrogen) and reverse-transcribed into cDNA using the PrimeScript RT reagent Kit (Takara, #RR047A). Quantitative real-time PCR (qPCR) was performed in triplicate using TB Green Premix Ex Taq II (Takara, #RR820A) on a QuantStudio 5 Real-Time PCR System (Applied Biosystems). The reaction protocol consisted of an initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. The relative mRNA expression levels of target genes were normalized to the endogenous control GAPDH and calculated using the comparative 2^(–ΔΔCT) method. For the quantification of miR-429, total RNA was reverse-transcribed using the miRcute Plus miRNA First-Strand cDNA Kit (Tiangen, #KR211), followed by qPCR with the miRcute Plus miRNA qPCR Kit (SYBR Green) (Tiangen, #FP411). The snRNA U6 was used as an internal control for normalization. The specific forward primer for hsa-miR-429 was 5’-UAAUACUGUCUGGUAAAACCGU-3’, and the universal reverse primer was provided by the kit. The sequences of all primers used in this study are listed below: IFN-β: Forward: 5’-GCTTGGATTCCTACAAAGAAGCA-3’, Reverse: 5’-ATAGATGGTCAATGCGGCGTC-3’. TNF-α: Forward: 5’-CCTCTCTCTAATCAGCCCTCTG-3’, Reverse: 5’-GAGGACCTGGGAGTAGATGAG-3’. CXCL9: Forward: 5’-AGAGTTCGAGGAACCCTAGTG-3’, Reverse: 5’-GGATTGTAGTGGCCCGTGAC-3’. CXCL10:Forward: 5’-GTGGCATTCAAGGAGTACCTC-3’, Reverse: 5’-TGATGGCCTTCGATTCTGGATT-3’.
HIF-1α: Forward: 5’-GAACGTCGAAAAGAAAAGTCTCG-3’, Reverse: 5’-CCTTATCAAGATGCGAACTCACA-3’. VEGF-α: Forward: 5’-AGGGCAGAATCATCACGAAGT-3’, Reverse: 5’-AGGGTCTCGATTGGATGGCA-3’.
TGF-β: Forward: 5’-CAACAATTCCTGGCGATACCTC-3’, Reverse: 5’-GCACAACTCCGGTGACATCAA-3’. ANGPT-2: Forward: 5’-TGCAGGAACCACACTCAACC-3’, Reverse: 5’-CATGGGTCCTTGAGGCATCC-3’. GAPDH: Forward: 5’-GGAGCGAGATCCCTCCAAAAT-3’, Reverse: 5’-GGCTGTTGTCATACTTCTCATGG-3’. U6: Forward: 5’-CTCGCTTCGGCAGCACA-3’,
Reverse: 5’-AACGCTTCACGAATTTGCGT-3’.
Clone formation experiment
A colony formation assay was conducted. Breast cancer cells were harvested with 0.25% Trypsin-EDTA (Gibco, #25200056) and seeded into 6-well plates (Corning, #3516) at 2,000 cells per well. After 14 days of culture (37 °C, 5% CO₂), the cells were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. The percentage of the total area occupied by colonies in each well was automatically quantified from the scanned images using the ImageJ software (NIH), with a uniform threshold applied across all samples for consistency.
CCK-8 experiment
Breast cancer (BC) cells in the logarithmic growth phase were harvested by trypsinization, resuspended to a density of 2 × 10⁴ cells/mL in complete medium, and seeded into 96-well plates at 100 µL per well (resulting in 2,000 cells/well). To assess cell viability, one plate was assayed every 24 h for a designated period. At each time point, the culture medium was carefully removed and replaced with 100 µL of fresh serum-free medium containing 10% (v/v) CCK-8 reagent (Dojindo, Japan). Wells containing only culture medium and CCK-8 reagent served as the blank control. The plates were then incubated at 37 °C for 1 h in the dark. Following incubation, the absorbance at 450 nm was measured using a microplate reader (BioTek, Synergy H1). The blank control readings were subtracted from the sample readings for analysis.
Immunofluorescence microscopy
For immunofluorescence (IF) staining, cells or tissue sections were fixed with 4% paraformaldehyde (PFA) for 15 min and permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature. After blocking with 5% bovine serum albumin (BSA) for 1 h, the samples were incubated overnight at 4 °C with primary antibodies. The following antibodies were used: rabbit anti-phospho-STING (Ser366) (Cell Signaling Technology, CST #50907S, 1:400) and rat anti-CD31 (BD Biosciences, #553370, 1:200). The next day, after washing, the samples were incubated with a mixture of secondary antibodies for 1 h at room temperature, protected from light. The secondary antibodies used were: Alexa Fluor 594-conjugated goat anti-rabbit IgG (Invitrogen, #A-11037, 1:500) and Alexa Fluor 647-conjugated goat anti-rat IgG (Invitrogen, #A-21247, 1:500). For double-staining experiments involving cytoskeleton or apoptosis, F-actin was stained with Alexa Fluor 488-phalloidin (Invitrogen, #A12379, 1:200) for 1 h, and apoptotic cells were detected using a TUNEL assay kit (Beyotime, #C1089) according to the manufacturer’s instructions, which yields green fluorescence (FITC). Finally, all samples were counterstained with DAPI (Sigma-Aldrich, #D9542, 1 µg/mL) to visualize nuclei. Fluorescent images were captured using a fluorescence microscope (Nikon Eclipse Ti2).
In vivo tumor targeting of ox-HyV
Thirty-six female BALB/c nude mice (6 weeks old) bearing 4T1 tumor xenografts were randomly divided into four groups (n = 9) to evaluate the biodistribution of the following Cy5.5-labeled vesicles: (1) non-peroxidized breast cancer cell membrane vesicles (CaM) (2), non-peroxidized iPSC-endothelial cell membrane vesicles (EcV) (3), peroxidized iPSC-endothelial cell membrane vesicles (ox-EcV), and (4) peroxidized heterotrimeric hybrid membrane vesicles (ox-HyV). Following intravenous injection via the tail vein, three mice from each group were anesthetized and imaged at 4, 8, and 12-hour time points using the Bruker MI SI in vivo imaging system. For quantitative analysis of in vivo images, regions of interest (ROIs) were drawn manually to encompass the entire tumor region and major organs (e.g., heart, liver, spleen, lung, kidneys) as visualized by the background grayscale image. The total radiant efficiency within each ROI was calculated and recorded using the instrument’s software. Background fluorescence, determined from an ROI drawn over a non-signal region of the mouse, was subtracted from each measurement. At the 8-hour terminal time point, the mice were euthanized, and major organs (heart, liver, spleen, lung, kidneys) and tumors were harvested for ex vivo imaging. The fluorescence intensity of each excised organ was immediately quantified using the same imaging system. The results are expressed as the mean radiant efficiency per organ or as the tumor-to-background ratio, calculated as the signal in the tumor divided by the signal in the muscle.
In vivo antitumor properties of ox-HyV
The left mammary gland of female BALB/c mice, aged 6 weeks, was inoculated with a total of 6 × 10^6^ 4T1 cells. Afterward, the mice were randomly divided into 4 groups, with each group including six mice, once the tumors had grown to around 100 mm^3^ in size. (1) PBS. (2) Peroxide-treated tumor cell membrane ox-CaM. (3) Peroxide-treated EC cell membrane ox-EC. (4) Peroxide-treated heterogeneous membrane ox-HyV. The drugs in each of the above groups were injected into mice via the tail vein at 1.0 mg/kg body weight every three days. Meanwhile, mice were examined for body weight and tumor size every three days. On day 21, mice were euthanized, and tumors were removed and weighed.
Western blotting
Total protein was extracted from tissues using the One-Step Animal Tissue Active Protein Extraction Kit (Genefist Biotech, Shanghai, China). Protein concentration was determined with a BCA Protein Assay Kit (Beyotime Biotechnology, #P0010), using bovine serum albumin (BSA) as a standard. Equal amounts of protein (e.g., 20–30 µg per lane) were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto polyvinylidene fluoride (PVDF) membranes.
The membranes were blocked with 5% non-fat milk in TBST (Tris-Buffered Saline with 0.1% Tween-20) for 1 h at room temperature and then incubated overnight at 4 °C with the following primary antibodies: phospho-STING (Ser366) (Cell Signaling Technology, CST #50907S, 1:1000), STING (CST #13647S, 1:1000), phospho-IRF3 (Ser396) (CST #4947S, 1:1000), IRF3 (CST #4302S, 1:1000), and β-Actin (Proteintech, #66009-1-Ig, 1:5000). After washing, the membranes were incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-mouse secondary antibodies (Cell Signaling Technology, 1:2000) for 1 h at room temperature.
Protein bands were visualized using an enhanced chemiluminescence (ECL) substrate (Tanon, #180–5001) and captured with a chemiluminescence imaging system (Tanon 5200). The band intensity was quantified using ImageJ software (National Institutes of Health, USA).
miRNA sequencing
Total RNA was isolated from water and sediment samples using the RNeasy PowerWater Kit (Qiagen, #14600-50-NF) and the RNeasy PowerSoil Kit (Qiagen, #12866-50-NF) respectively, according to the manufacturers’ protocols. For the HyV and ox-HyV samples, RNA was extracted using the miRNeasy Mini Kit (Qiagen, #217004), which is specifically designed for the simultaneous purification of miRNA and total RNA. All RNA samples were treated with DNase I (RNase-free) to eliminate genomic DNA contamination. RNA concentration and purity were determined using a Qubit 2.0 Fluorometer with the Qubit RNA HS Assay Kit (Invitrogen, #Q32852) and a NanoDrop spectrophotometer, respectively. Samples with an A260/A280 ratio between 1.8 and 2.1 and an A260/A230 ratio greater than 2.0 were deemed suitable for subsequent analysis. RNA integrity was further verified by agarose gel electrophoresis to ensure the presence of distinct ribosomal RNA bands.
Sequencing services were provided by APExBIO (Shanghai, China). Sequencing libraries were constructed from total RNA using the TruSeq Small RNA Library Preparation Kit (Illumina, #RS-200-0012/RS-200-0024), which selectively ligates adapters to the 5’ and 3’ ends of small RNA molecules (14–30 nt). The protocol includes size selection steps to enrich for miRNAs. The completed libraries were quantified using the Qubit dsDNA HS Assay Kit (Invitrogen, #Q32851) and their size distribution was validated using an Agilent 2100 Bioanalyzer with the High Sensitivity DNA Kit (Agilent, #5067 − 4626). The libraries were then pooled in equimolar ratios and sequenced on an Illumina HiSeq 2500 platform to generate 50 bp single-end reads.
Raw sequencing reads were subjected to quality control using FastQC (v0.11.9). Adapters and low-quality bases (Phred score < 20) were trimmed using Cutadapt (v2.10). Clean reads were then aligned to the reference genomeGRCh38 (human) using Bowtie2 (v2.4.2). miRNA counts were quantified by mapping the reads to known miRNA precursors from miRBase (v22.0). The raw count data was normalized using the Transcripts Per Million (TPM) method to account for differences in sequencing depth and miRNA length across samples.
Differential expression analysis between HyV and ox-HyV groups was performed using the DESeq2 package (v1.30.1) in R, which employs an internal normalization based on the median of ratios method. miRNAs with a |FoldChange| > 2 and an adjusted p-value (Benjamini-Hochberg FDR) < 0.05 were considered statistically significant. The target genes of these differentially expressed miRNAs were predicted by integrating results from two algorithms: TargetScan (v7.2; context + + score percentile > 90) and miRWalk (v3.0; using the ‘miRWalk’ and ‘TargetScan’ databases with a p-value < 0.05). Only genes predicted by both tools were considered high-confidence targets. Functional enrichment analysis for Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and Gene Ontology (GO) terms (Biological Process, Molecular Function, Cellular Component) was conducted on the high-confidence target gene set using the DAVID bioinformatics database (v2021). Terms with an FDR-adjusted p-value (Benjamini-Hochberg) < 0.05 were regarded as significantly enriched.
mRNA-sequencing
Total RNA was isolated from human primary TVECs transfected with miR-NC (negative control) or miR-429 mimics using the RNeasy Mini Kit (Qiagen, #74104). The experiment included six biologically independent replicates per group. RNA integrity was confirmed using an Agilent 2100 Bioanalyzer, and all samples had an RNA Integrity Number (RIN) greater than 8.0. Sequencing libraries were constructed from 1 µg of total RNA per sample using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (New England Biolabs) and sequenced on an Illumina NovaSeq 6000 platform to generate 150 bp paired-end reads.
Raw sequencing reads were subjected to quality control using FastQC (v0.11.9), and adapter sequences and low-quality bases were trimmed using Trimmomatic (v0.39). The clean reads were aligned to the mouse reference genome (GRCm39) using STAR (v2.7.10a). Gene-level counts were generated using featureCounts (v2.0.3) based on the GENCODE M25 annotation. Differential gene expression analysis between the miR-429 and miR-NC groups was performed using the DESeq2 package (v1.34.0) in R. Genes exhibiting an absolute fold change > 2 and an adjusted p-value (Benjamini-Hochberg FDR) < 0.05 were considered statistically significant. Functional enrichment analysis of these differentially expressed genes for Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and Gene Ontology (GO) terms was conducted using the clusterProfiler package (v4.2.2), with an FDR < 0.05 set as the significance threshold. The raw sequencing data generated in this study are not publicly available due to ongoing research but are available from the corresponding author upon reasonable request.
Flow cytometry assay
Single-cell suspensions were prepared from mechanically minced mouse tumors by enzymatic digestion with Collagenase IV (2 mg/mL; Gibco, #17104019) and DNase I (50 µg/mL; Gibco, #18047019) at 37 °C for 30 min. The resulting cell suspension was passed through a 70 μm cell strainer (BD Biosciences), followed by erythrocyte lysis using ACK lysing buffer.
For intracellular cytokine analysis, cells were stimulated with Cell Stimulation Cocktail (plus protein transport inhibitors, eBioscience #00-4975-03) for 4–6 h at 37 °C. Cells were incubated with Fc receptor blocking solution (Purified anti-mouse CD16/32, Clone 93; eBioscience) for 10–15 min on ice, followed by surface staining with antibody cocktail for 20 min at 4 °C in the dark. The following antibodies were used: CD3-APC-Cy7 (Clone 17A2, #100330, 1:200), CD8-PerCP-Cy5.5 (Clone 53 − 6.7, #100734, 1:100), CD4-APC-Cy7 (Clone GK1.5, #552051, 1:400), TCR Vβ13.1-APC (Clone MR12-4, BioLegend #100208, 1:100), F4/80-APC (Clone BM8, #123116, 1:150), CD206-PE (Clone C068C2, #141706, 1:200), CD86-PE-Cy7 (Clone GL-1, #105014, 1:200), Foxp3-FITC (Clone MF-14, #320105, 1:100). For intracellular staining, cells were fixed and permeabilized using Foxp3/Transcription Factor Staining Buffer Set (eBioscience #00-5523-00) prior to staining with IFN-γ-PE (Clone XMG1.2, BioLegend #505808, 1:100) and Granzyme B-BV421 (Clone NGZB, BioLegend #515408, 1:50).
Data were acquired on an LSRFortessa flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star). The gating strategy for cytokine analysis specifically focused on the TCR Vβ13.1⁺ CD8⁺ T cell population to assess IFN-γ and Granzyme B production.
ELISA
The concentrations of specific cytokines (IFN-β, TNF-α, IL-1β, IL-4, IL-6, IL-10, TGF-β) in the conditioned media collected from the lower chamber of the breast cancer cell-macrophage co-culture system were quantified using commercial enzyme-linked immunosorbent assay (ELISA) kits, according to the manufacturers’ instructions. The following kits were used: IFN-β (R&D Systems, #DY8234-05), TNF-α (BioLegend, #430904), IL-1β (R&D Systems, #DY201-05), IL-4 (BioLegend, #431104), IL-6 (BioLegend, #431304), IL-10 (BioLegend, #431414), and TGF-β (R&D Systems, #DY240-05).
Briefly, 100 µL of standards and samples were added to the respective antibody-precoated wells and incubated for 2 h at room temperature. After washing, a biotinylated detection antibody was added to each well, followed by incubation with streptavidin-horseradish peroxidase (HRP). The colorimetric reaction was developed using a tetramethylbenzidine (TMB) substrate and stopped by the addition of stop solution. The absorbance was immediately measured at 450 nm with a wavelength correction at 570 nm using a microplate reader (e.g., Bio-Rad Model 680). The cytokine concentrations (in pg/mL) for each sample were determined by interpolating the absorbance values from the standard curve generated with the provided recombinant cytokines.
Statistical analysis
Statistical analyses were conducted with GraphPad Prism 7.0. Data are expressed as mean ± SD. Differences between two groups were assessed by an unpaired Student’s t-test. For multiple group comparisons, one-way ANOVA with a post-hoc test was used. Statistical significance was set at *P < 0.05, and denoted as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.
Results
Characterization of iPSCs-ECs and ox-HyV
To enhance vesicle targeting to tumors, we obtained and purified BC MDA-MB-231 cell-derived nanovesicles CaM using a membrane protein extraction kit. Cancer cells are lysed following the protocol provided by the Abcam Membrane Protein Extraction Kit, facilitating the separation of membrane components via differential centrifugation. Subsequently, membrane vesicles are enriched through sucrose density gradient ultracentrifugation and subjected to a washing process. The tumor cell membrane vesicles (CaM) isolated through this methodology exhibit enhanced vesicle integrity and superior membrane protein activity.
Two vesicles were fused into the heterotrimeric membrane ox-HyV with the aid of sonication (Fig. 1A). The TEM images revealed spherical particles of ox-HyV with an intact monolayer membrane structure (Fig. 1B). Herein, we conducted an SDS-PAGE analysis, aiming at examining CaM, ox-HyV, and ox-EcV protein composition. The dynamic light scattering (DLS) analysis revealed that these ox-HyVs exhibit a comparable range of sizes (PDI = 0.17) (Figure S1), with an around 100 nm peak diameter (Fig. 1C). The proteins exhibited very identical streaks, suggesting that the proteins in ox-EcV and CaM were effectively preserved in ox-HyV after the hybridized membrane formation (Fig. 1D). The successful fusion of the two biological vesicles into the hybridized membrane ox-HyV was more intuitively observed after mixing DiO/DiI dual dye-labeled ox-EcV and CaM (Fig. 1E). In addition, the zeta potential of these vesicles remained around − 20 to -30 mV, further confirming the biostability of ox-Hyv (Fig. 1F). The total amount of protein in vesicles extracted per 100,000 cells was 25.12 ± 3.26 µg/mL (ox-EcV), 23.89 ± 2.23 µg/mL (HyV), 22.74 ± 2.51 µg/mL (ox-HyV) (Figs. 1G-H).
Fig. 1. Preparation and characterization of ox-HyV. (A) Schematic illustration of the fabrication process for ox-HyV, involving membrane fusion and subsequent peroxidation. (B) Representative transmission electron microscopy (TEM) images of ox-HyV. Scale bars: 100 nm (left) and 50 nm (right). (C) Hydrodynamic diameter distribution of CaM, EcV, HyV, and ox-HyV measured by dynamic light scattering (DLS). Data are presented as mean ± SD (n = 3 independent vesicle preparations). (D) Protein profile analysis by SDS-PAGE (4–20% gradient gel). (E) Fluorescence microscopy images demonstrating the fusion of DiI-labeled CaM (red) and DiO-labeled ox-EcV (green). Yellow signal in the merged panel indicates colocalization. Scale bar: 10 μm. (F) Colloidal stability assessed by zeta potential measurements of ox-EcV, HyV, and ox-HyV in aqueous solution over 7 days. Data are presented as mean ± SD (n = 3 technical replicates). Statistical significance was determined by two-way ANOVA with Tukey’s post-hoc test. (G) Protein yield from vesicles derived from 10⁶ cells. Data are presented as mean ± SD (n = 5 independent biological replicates). No significant differences were found by one-way ANOVA. (H) Protein content per single vesicle particle. Data are presented as mean ± SD (n = 5 independent measurements). ns indicates no significant difference as determined by one-way ANOVA
ox-HyV induces BC cell apoptosis via TVECs
Direct addition of ox-EcV, HyV, and ox-HyV to BC did not show significant tumor suppression in vitro. Interestingly, significant apoptosis was observed in the lower compartment when ox-EcV and ox-HyV were added to primary TVECs cocultured with the tumor cells (Fig. 2A). Through this indirect action, the proliferative capacity of tumor cells was significantly inhibited (Fig. 2C and E), whereas it was not obvious when directly acting (Fig. 2B and D and S2). Flow cytometry demonstrated differences in tumor apoptosis (Figs. 2F–G and S2). More visually, cellular immunofluorescence staining revealed cell membrane crumpling and nuclear rupture and lysis (Fig. 2H).
Accordingly, we speculated that the promotion of tumor cell apoptosis by ox-EcV and ox-HyV through TVECs might be due to the secretion of certain inflammatory factors, so we examined the cytokines in the culture medium of the lower chamber of the coculture model, showcasing that TNF-α were elevated, whereas as TGF-β was decreased (Figure S3).
To further explore the tumor therapeutic potential of these vesicles, we added them into the body of hormonal mice by tail vein injection, manifesting that the tumors of the ox-HyV-treated group were significantly shrunk (Figs. 2I and S4). Moreover, H&E staining of the tumors of mice in the ox-HyV treatment group showed significant therapeutic effects, with a significant reduction in the Ki-67 proliferation index and an escalation in apoptosis (Figs. 2J and S5). In addition, mice in the ox-HyV treatment group had a longer survival period (Fig. 2K). Tests on major organs and blood, liver, and kidney functions of mice in the ox-HyV treatment group showed that there was no obvious damage to various organs and no abnormalities in various biochemical indexes, with no significant difference in mice body weight (Figures S6-S8). Collectively, ox-HyV treatment has good biological safety.
Fig. 2. The indirect mechanism by which ox-HyV induces apoptosis in BC cells through TVECs. (A) This panel presents a schematic diagram depicting the process of BC cell apoptosis mediated by ox-HyV via TVECs. (B-C) The panels demonstrate the formation of BC cell clones when treated with ox-HyV, either directly (B) or indirectly through TVECs (C). (D-E) CCK-8 assays were conducted to assess BC cell viability following direct (D) or indirect (via TVECs) (E) treatment with ox-HyV. Data are presented as mean ± SD (n = 3 independent experiments). Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test for all pairwise comparisons. **P < 0.01; ***P < 0.001. (F-G) Flow cytometry analysis was performed to detect apoptosis in BC cells treated with ox-HyV directly (F) or indirectly via TVECs (G). (H) Immunofluorescence imaging reveals morphological changes in BC cells following treatment with PBS, ox-EcV, HyV, and ox-HyV indirectly via TVECs. (Green: phalloidin; Blue: DAPI). Scale bar = 10 μm. (I) Tumor size measurements were taken after 20 days of treatment in mice with PBS, ox-EcV, HyV, and ox-HyV. (J) Histological analyses, including H&E staining, Ki-67, and TUNEL assays, were conducted on tumor samples from mice treated with PBS, ox-EcV, HyV, and ox-HyV. Scale bar = 100 μm. (K) Survival curves are presented for mice subjected to treatment with PBS, ox-EcV, HyV, and ox-HyV
ox-HyV demonstrates effective targeting of TVECs
Both ox-EcV and ox-HyV exhibited significant indirect anti-tumor effects in vitro. However, in a murine model, ox-EcV failed to achieve effective tumor suppression compared to ox-HyV. We hypothesized that this discrepancy stemmed from inferior in vivo targeting and tumor accumulation of ox-EcV. In vitro experiments confirmed that Cy5.5-labeled vesicles were effectively internalized by tumor vascular endothelial cells (TVECs) (Figs. 3A, B). To assess their in vivo distribution, we performed live imaging and quantitative analysis. Following intravenous injection, both CaM and ox-HyV showed significantly greater accumulation in tumors compared to EcV and ox-EcV across multiple time points, with peak enrichment observed at 8 h post-injection (Fig. 3C, D; n = 3). At this 8-hour peak, the tumor fluorescence intensity of ox-HyV was approximately 3.2-fold and 3.4-fold higher than that of EcV and ox-EcV, respectively (Fig. 3D). Ex vivo fluorescence quantification of major organs and tumors at the 8-hour peak further validated the specificity of ox-HyV for tumor tissue (Figure S9)(n = 5). Critically, the fluorescence signal in tumors was 12.7-fold higher than in the heart and 3.6-fold higher than in the liver, demonstrating superior tumor-specific accumulation over both a circulatory background and the primary metabolic organ(Fig. 3E). Furthermore, immunohistochemical analysis of tumor sections co-stained with the vascular marker CD31 demonstrated that CaM and ox-HyV were preferentially localized to tumor blood vessels (Figs. 3F, S10), indicating that the heterodimeric membrane’s targeting ability is primarily inherited from the tumor cell membrane component. In conclusion, ox-HyV inherits potent tumor-targeting capability from its incorporated tumor cell membrane, enabling its specific and effective enrichment at tumor endothelial sites in vivo. This targeted delivery underlies its superior therapeutic performance and is a crucial foundation for its application in tumor therapy.
Fig. 3ox-HyV targeting to TVECs in breast cancer. (A-B) Immunofluorescence analysis (A) demonstrates the uptake of Cy5.5-labeled vesicles (red) by human primary TVECs, with phalloidin staining in green and DAPI in blue. Scale bar = 20 μm. The accompanying quantitative analysis (B) presents data as mean ± SD, with n = 3, indicating no significant difference (ns). (C-D) Small animal imaging (C) and subsequent quantitative analysis (D) were conducted for CaM, EcV, ox-EcV, and ox-HyV at 4, 8, and 12 h post-injection via the tail vein in mice. (E) The fluorescence intensity of major organs and tumors was measured 8 h following the injection of ox-HyV into the mice via the tail vein. (F) Co-localization studies of each vesicle with vascular endothelial cells were performed in immunofluorescent mouse tumors, with CD31 in green, Cy5.5 in red, and DAPI in blue. Scale bar = 100 μm
The secretory phenotype of ox-HyV-transformed TVECs and the underlying mechanisms
To investigate the mechanism by which ox-HyV indirectly kills tumors by acting on TVECs, we conducted a comparative analysis of HyV and ox-HyV miRNA profiles utilizing NGS technology. The NGS data showcased a set of either upregulated or downregulated miRNAs (Figs. 4A–D). Afterward, we performed GO and KEGG enrichment analyses to comprehend the roles of these miRNAs better, focusing on miRNAs that were upregulated in ox-HyV relative to HyV, which may account for ox-HyV’s indirect killing of BC through TVECs. Biological process (BP) outcomes manifested that these miRNAs-linked genes primarily contributed to positive regulation of cell adhesion, mononuclear cell differentiation, Wnt pathway, lymphocyte differentiation, cell-substrate adhesion, mitotic nuclear division, T cell differentiation, and stem cell differentiation (Fig. 4E). The KEGG enrichment data identifies several pathways that are potentially significant in ox-HyV potentiation. These pathways include the PI3K-Akt, MAPK, Transcriptional misregulation in cancer, Cellular senescence, Cell cycle Hippo, and p53, as well as signaling pathways governing the pluripotency of stem cells (Fig. 4F).
Fig. 4ox-HyV reprograms the secretory phenotype of TVECs through miRNA-mediated mechanisms. (A) Venn diagram illustrating the distribution of miRNAs identified in HyV versus ox-HyV. (B) Volcano plot of differentially expressed miRNAs between HyV and ox-HyV. Red and blue dots represent significantly up- and down-regulated miRNAs, respectively (fold change > 2.0, adjusted P value < 0.05 by Benjamini-Hochberg correction). (C, D) Heatmap visualization of the most significantly dysregulated miRNAs in HyV versus ox-HyV. Color scale indicates Z-score normalized expression levels. (E) Gene Ontology (GO) enrichment analysis of predicted target genes of differentially expressed miRNAs. The top significantly enriched biological processes are shown (adjusted P value < 0.05 by Fisher’s exact test with FDR correction). (F) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of predicted target genes. The most significantly enriched pathways are displayed (adjusted P value < 0.05 by Fisher’s exact test with FDR correction)
miR-429 in ox-HyV activates the cGAS-STING pathway in TVECs and enhances the TIME
The outcomes of miRNA sequencing in HyV and ox-HyV showcased the most obvious difference in miR-429 expression; therefore, we demonstrated the enrichment of miR-429 in ox-HyV by qRT-PCR.
To demonstrate that the indirect therapeutic effect of ox-HyV on tumors is mainly dependent on miR-429 in it, we used the coculture model (Fig. 5A), where ox-HyV, miR-429mimic acted on human primary TVECs, showcasing that they impeded the proliferative potential of tumors and prompted apoptosis, and that miR-429 inhibitors could reverse this effect (Figs. 5B-C, S11). This suggests that the main active ingredient in ox-HyV is miR-429. Since miR-429 plays very different roles in different types of tumors, we further explored its specific mechanisms in BC treatment. We found that the miRNA sequencing results in ox-HyV included biological processes such as Mononuclear cell differentiation, lymphocyte differentiation, regulation of response to DNA damage stimulus and T cell differentiation. Therefore, we speculated that ox-HyV might activate cGAS-STING signaling in ECs, thus affecting tumor cell growth and apoptosis. Immediately, we examined the expression of relevant genes in tumor cells. Among them, IFN-β1 and TNF-α genes and CXCL9/10 genes were significantly upregulated, whereas genes involved in the generation of vascular destabilization and affecting the normalization of blood vessels, such as HIF-1α, VEGF-α, TGF-β, and ANGPT2 were significantly down-regulated. similarly, miR-429inhibitor could reverse the ox-HyV-related effects (Figs. 5D-K). At the protein level, phosphorylated IRF3 was significantly increased in the ox-HyV and miR-429mimic treatment groups, again demonstrating potent activation of the cGAS-STING pathway (Fig. 5L). ELISA assay of cytokine levels in the lower chamber of the coculture model showed that the pro-inflammatory cytokines IFN- β, TNF-α, IL-1β, and CXCL9 were elevated, while TGF-β and IL-10 were decreased (Fig. 5M). When the coculture lower chamber was switched to undifferentiated M0, macrophage conversion to M1 morphology was clearly observed in the ox-HyV and miR-429mimic treatment groups, while the miR-429inhibitor prevented this trend (Fig. 5N). In the tumors of the ox-HyV and miR-429mimic treatment groups, p-STING was significantly elevated (Figs. 5O, S12), and the changes of related gene expression in the tumors aligned with the in vitro experiments (Figure S13). In order to investigate whether miR-429 in ox-HyV played the above-mentioned effects on the TIME of BCs, we examined the tumors after various treatments by flow cytometry, revealing a significant increase in M1 macrophages and a decrease in M2 macrophages in tumors of the ox-HyV and miR-429mimic treatment groups, and the M1/M2 ratio was decreased (Figs. 5P-Q). The high M2 macrophage density in tumor tissues has been reported to be often associated with poorer cancer-specific survival rates, whereas the high M1 macrophage density does not necessarily correlate directly with better survival rates. However, a higher M1:M2 density ratio in the tumor stroma is linked to better cancer-specific survival rates, indicating that the polarization state of macrophages, rather than their overall density, is related to cancer-specific survival rates [17]. In addition, the polarization state of macrophages not only affects tumor progression but is also closely linked with the effectiveness of immunotherapy. The M1-type macrophages can enhance immune responses by promoting the infiltration and survival of tissue-resident memory T cells, thereby improving patient survival rates [18]. In certain cancer types, M1-type macrophages are related to better outcomes, while M2-type macrophages are linked to poorer outcomes [19]. In T cell-mediated antitumor immunity, CTLs are the main antitumor effector cells, whereas Tregs are involved in antitumor immunosuppression. miR-429 in ox-HyV significantly increased the infiltration of CTLs (Figs. 5R-S) and decreased the proportion of Tregs (Figs. 5T and S14). DCs are pivotal in the immune system, particularly in cancer immunotherapy. Their maturation is crucial for the effective tumor antigen presentation, which is a key step in initiating an immune response against cancer cells. To investigate the DC maturation rate after various treatments, we collected lymph nodes from the ipsilateral groin of each 4T1 hormonal mouse and analyzed them by flow cytometry. The results showed that ox-HyV and miR-429mimic could promote DC maturation in mice (Figs. 5U-V).
Overall, our results suggest that miR-429 in ox-HyV can play a tumor-killing role by activating cGAS-STING signaling in BC tumors, which leads to the secretion of various cytokines and the improvement of the TIME.
Fig. 5. The role of miR-429 in ox-HyV in activating the cGAS-STING pathway within TVECs. (A) Quantitative real-time PCR (qRT-PCR) was conducted to assess the differential expression of miR-429 between HyV and ox-HyV. Data are shown as the mean ± SD, n = 3. ns, no significant difference. ***P < 0.001 by one-way ANOVA with Tukey’s test. (B-C) In ox-HyV, miR-429 suppresses BC proliferation (B) and enhances apoptosis (C) through its interaction with TVECs. (D-K) qRT-PCR analysis was employed to evaluate the expression levels of inflammatory factors (D-E), angiogenic factors (F-G), chemokines (H-I), and vascular destabilizing genes (J-K) following treatment with PBS, ox-HyV, miR-429 mimic, and ox-HyV combined with miR-429 inhibitor in BCs via TVECs. Statistical analysis was performed using one-way ANOVA with Tukey’s post-hoc test for all pairwise comparisons. Data are presented as mean ± standard deviation (SD), with n = 3. Statistical significance is indicated as follows: ns denotes no significant difference, *P < 0.05, **P < 0.01, and ***P < 0.001. (L) The activation of the cGAS-STING pathway in TVECs treated with PBS (a), ox-HyV (b), miR-429 mimic (c), and ox-HyV combined with miR-429 inhibitor (d). (M) A heatmap depicting variations in cytokine secretion in TVECs treated with PBS, ox-HyV, miR-429 mimic, and ox-HyV combined with miR-429 inhibitor. (N) Immunofluorescence analysis illustrating the morphology of M0 macrophage polarization induced by TVECs following various treatments, with green indicating phalloidin and blue indicating DAPI. Scale bar = 10 μm. (O) Immunofluorescence showing the expression of p-STING in the tumors of homozygous mice treated with PBS, ox-HyV, miR-429mimic and ox-HyV + miR-429 inhibitor. (red: p-STING, blue: DAPI) Scale bar = 100 μm. (P-Q) Flow cytometry was performed to detect M1-like macrophages and M2-Like macrophages (P) and M1/M2 (Q) in the tumors of mice in each treatment group. Statistical analysis was performed using one-way ANOVA with Tukey’s post-hoc test for all pairwise comparisons. Data are shown as the mean ± SD, n = 3. ns, no significant difference. *P<0.05.(R-S) Flow cytometry detection of CD8 + T cells (R) and quantitative analysis of CD8 + T cells (S) in the tumors of mice in each treatment group. Statistical analysis was performed using one-way ANOVA with Tukey’s post-hoc test for all pairwise comparisons. Data are shown as the mean ± SD, n = 3. ns, no significant difference. ***P < 0.001. (T) Quantitative analysis of Treg cells in the tumors of mice in each treatment group. Statistical analysis was performed using one-way ANOVA with Tukey’s post-hoc test for all pairwise comparisons. Data are shown as the mean ± SD, n = 3. ns, no significant difference. ***P < 0.001.(U-V) Flow cytometry for detection of mature DCs (U) in lymph nodes of mice in each treatment group and quantitative analysis (V). Statistical analysis was performed using one-way ANOVA with Tukey’s post-hoc test for all pairwise comparisons. Data are shown as the mean ± SD, n = 3. ns, no significant difference. ***P < 0.001
miR-429 targets Bcl2 in TNBC
The results of mRNA sequencing on miR-NC and miR-429-treated human primary TVECs, as shown in the volcano diagram (Fig. 6A), demonstrated that among the 81 differentially expressed genes (DEGs), 30 were upregulated and 51 were down-regulated after miR-429 treatment. The DEGs were analyzed for GO/KEGG pathway enrichment, which included biological processes such as cytokine receptor binding, cytokine activity, T cell differentiation, cytokine-cytokine receptor interaction, and TNF signaling pathway, among other signaling pathways (Figs. 6B-C). This is consistent with the previous cGAS-STING signaling pathway activation, cytokine secretion, and tumor apoptosis results. To further predict miR-429 target genes, we took the intersection of BC-related genes, DEGs, and TargetScan database predicted target genes and targeted two candidate genes, SOX5 and Bcl2 (Fig. 6D). Notably, SOX5 did not show the same trend as the ox-HyV-treated group in the subsequent validation (Fig. 6E). Therefore, we assumed that Bcl2 was the target gene of miR-429. At the protein level, ox-HyV and miR-429mimic similarly significantly downregulated Bcl2 expression (Fig. 6F). To determine whether Bcl2 is a direct target of miR-429, we predicted the conserved miR-429 binding site within the 3′-UTR region of the Bcl2 gene for a luciferase reporter assay (Fig. 6G). As a control, the same 3′-UTR region of Bcl2 was mutated to eliminate the seed sequence. Cotransfection of miR-429 with wild-type (WT) Bcl2 3′-UTR luciferase reporter vector in HEK293 cells resulted in a significant reduction in luciferase activity (Fig. 6H).
In contrast, cotransfection with a mutant (MUT) Bcl2 3′-UTR luciferase reporter vector did not affect luciferase activity. Subsequently, for miR-429 and Bcl2 overexpression, plasmids were added to MDA-MB-231 cells, respectively, and at the protein level, miR-429 reversed the overexpression of Bcl2 (Fig. 6I).
Fig. 6miR-429 targets Bcl2. (A) The volcano plot displays the differentially expressed genes (DEGs) following treatment of TVECs with miR-NC and miR-429. (B-C) Gene Ontology (GO) analysis (B) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis (C) were conducted on the DEGs. (D) Venn diagrams depict the overlap among DEGs, breast cancer-associated genes, and targets predicted by TargetScan. (E) Quantitative real-time PCR (qRT-PCR) was performed to assess the expression levels of SOX5 and BCL2 in TVECs treated with PBS, HyV, ox-HyV, and miR-429. Statistical analysis was performed using one-way ANOVA with Bonferroni correction. Data are presented as mean ± standard deviation (SD), with n = 3. No significant difference was observed (ns), while ***P < 0.001 indicates high significance. (F) The protein levels of Bcl2 in TVECs were evaluated following various treatments. (G) Fragments of the 3’ untranslated region (UTR) of Bcl2, containing either wild-type (WT) or mutant (MUT) seed sequences, were cloned into pmirGLO vectors for luciferase reporter assays. (H) The luciferase reporter assay demonstrated that miR-429 targets the 3′UTR region of Bcl2 in HEK293 cells. Statistical significance between multiple groups was determined by one-way ANOVA followed by post-hoc tests for all pairwise comparisons. Data are shown as the mean ± SD, n = 5. (I) Western blot analysis (WB) indicated that miR-429 downregulates Bcl2 expression in TVECs
miR-429 modulates TIME by downregulating Bcl2 in TVECs and cGAS-STING signaling activation
To elucidate that miR-429 activates cGAS-STING signaling via Bcl2 in TVECs, we designed five subgroup controls of miR-NC, miR-429mimic, miR-429mimic + OE-Bcl2, miR-429 inhibitor and miR-429 inhibitor + siBcl2 five subgroup controls. The clone formation results established that Bcl2 overexpression reversed miR429 inhibition on the tumor proliferative ability. Similarly, tumor proliferation was still inhibited after the knockdown of Bcl2 even in the presence of miR-429 inhibitor (Figs. 7A-B), suggesting that Bcl2 may be a downstream molecule for the action of miR-429. IFN-β1, TNF- α, and CXCL10 expression was elevated in miR429-mimic (compared to the miR429-mimic + OE-Bcl2 group) and miR-429inhibitor + siBcl2 (compared to the miR-429 inhibitor group), and intra- and extracellular cGAMP was also significantly increased, and cytokine-wise, IFN-β, and TNF-α secretion were similarly elevated, indicating that cGAS-STING signaling was activated in both miR-429mimic and miR-429 inhibitor + siBcl2 groups (Figs. 7C-D, S15). At the protein level, the p-IRF3/IRP3 and p-STING/STING ratios exhibited the highest ratios in the above two groups (Figs. 7E, S16). Thirty 6-week-old hormonal BALB/c mice were randomly allocated into five groups (n = 6). miR429mimic and miR-429 inhibitor + siBcl2 groups showed increased apoptosis of tumor cells, decreased proliferation index, and significantly overexpressed p-STING, unlike the other three groups (Figs. 7F-G, S17). The expressions of other related genes in the tumor were also similar to the aforementioned trend (Figure S18). The above results indicated that miR-429 activated cGAS-STING signaling in BC by down-regulating Bcl2. Detecting the content of immune cells in the tumors of each group by flow cytometry, M1 macrophages, and CTLs were significantly higher in miR-429mimic and miR-429 inhibitor + siBcl2 groups were significantly increased, whereas M2-like macrophages and Treg were decreased. miR-429mimic and miR-429 inhibitor + siBcl2 groups of mice had a higher proportion of mature DCs in lymph nodes (Figs. 7H-L, S19).
Fig. 7miR-429 modulates the immune microenvironment of breast cancer tumors via Bcl2-mediated activation of the cGAS-STING pathway in tumor vascular endothelial cells (TVECs). (A-B) The effects on clone formation and quantitative analysis were assessed for miR-NC, miR-429 mimic, miR-429 mimic combined with OE-Bcl2, miR-429 inhibitor, and miR-429 inhibitor combined with siBcl2 in breast cancer cells treated with TVECs. Statistical significance between multiple groups was determined by one-way ANOVA followed by post-hoc tests for all pairwise comparisons. Data are presented as the mean ± standard deviation (SD), with a sample size of n = 3. Statistical significance is indicated by *P < 0.05. (C) Quantitative reverse transcription PCR (qRT-PCR) was employed to measure the expression levels of IFN-β1, TNF-α, and CXCL10 in TVECs following various treatments. Statistical significance between multiple groups was determined by one-way ANOVA followed by post-hoc tests for all pairwise comparisons. Data are presented as the mean ± SD, n = 3, with significance levels denoted as *P < 0.05, **P < 0.01, and ***P < 0.001. (D) The concentration of cGAMP inside and outside of breast cancer cells was measured after different treatments. Statistical significance between multiple groups was determined by one-way ANOVA followed by post-hoc tests for all pairwise comparisons. Data are shown as the mean ± SD, n = 3, with ***P < 0.001 indicating statistical significance. (E) Examination of various treatment groups involving cGAS-STING activation in tumor vascular endothelial cells (TVECs) was conducted. The groups were as follows: A: miR-NC, B: miR-429 mimic, C: miR-429 mimic combined with OE-Bcl2, D: miR-429 inhibitor, and E: miR-429 inhibitor combined with siBcl2. (F) Tumor growth curves were analyzed for mice in each treatment group. (G) Histological and immunohistochemical analyses, including Hematoxylin and Eosin (H&E) staining, Ki-67, TUNEL staining, and p-STING immunofluorescence staining, were performed on tumor samples from mice in each treatment group. (H-I) Flow cytometry was utilized to detect M1-like macrophages (H) and M2-like macrophages (I) within the tumors of mice across the treatment groups. (J-K) Flow cytometry analysis was also conducted to assess the presence of CD8 + T cells (J) and regulatory T cells (Tregs) (K) in the tumors of mice in each treatment group. (L) Additionally, flow cytometry was employed to detect mature dendritic cells (DCs) in the lymph nodes of mice from each treatment group
ox-HyV triggers anti-tumor immunity via the miR-429/Bcl2/cGAS-STING axis
To definitively establish that the anti-tumor efficacy of ox-HyV is contingent upon the activation of the cGAS-STING pathway, we performed a comprehensive set of loss-of-function studies in vivo. We employed a selective STING inhibitor in a murine TNBC model to determine whether pharmacological blockade of STING signaling could abrogate the therapeutic effects of ox-HyV.
Consistent with our previous findings, ox-HyV monotherapy potently inhibited tumor growth compared to the vehicle control. However, this therapeutic benefit was completely abolished in mice that were co-treated with the STING inhibitor (Fig. 8A). This critical result indicates that a functional STING pathway is indispensable for the anti-tumor effect of ox-HyV.
At the molecular level, Western blot analysis of tumor lysates confirmed that ox-HyV robustly activated the STING pathway, as evidenced by increased phosphorylation of STING. As expected, this activation was effectively suppressed in the combination treatment group. Importantly, the downregulation of Bcl2 by ox-HyV was unaffected by STING inhibition, positioning Bcl2 upstream of STING activation within the signaling cascade (Fig. 8B).
We further investigated the cellular consequences of STING pathway blockade. Immunofluorescence analysis revealed that ox-HyV treatment enhanced the phosphorylation of STING in the tumor microenvironment, an effect that was attenuated upon STING inhibition. Concordantly, the potent anti-proliferative and pro-apoptotic effects of ox-HyV, demonstrated by a significant reduction in Ki-67-positive cells and an increase in TUNEL-positive cells, were substantially reversed in the presence of the STING inhibitor (Figs. 8C-F).
Finally, we assessed the impact on tumor immunity by flow cytometry. ox-HyV treatment alone promoted a favorable immune microenvironment, characterized by an increased ratio of M1 to M2 macrophages and a decrease in regulatory T cells. Moreover, it enhanced the infiltration of CD8⁺ T cells and, crucially, the frequency of tumor-infiltrating CD8⁺ T cells producing IFN-γ and Granzyme B. Strikingly, all these immunomodulatory effects were significantly mitigated when STING signaling was inhibited (Figs. 8G-L and S20).
Collectively, these data from genetic and pharmacological loss-of-function experiments demonstrate that the cGAS-STING pathway is the critical downstream mechanism through which ox-HyV reprograms the tumor immune microenvironment and elicits its potent anti-tumor immunity.
Fig. 8. Genetic and pharmacological inhibition of STING abrogates the anti-tumor efficacy of ox-HyV. (A) Tumor growth curves of mice treated with vehicle, ox-HyV, STING inhibitor (STINGi), or their combination (ox-HyV + STINGi). Data are presented as mean ± SD (n = 6 mice per group). ***P < 0.001 for ox-HyV + STINGi vs. ox-HyV, by two-way ANOVA with Tukey’s post-hoc test. (B) Western blot analysis of tumor lysates showing protein levels of Bcl2, p-STING (Ser366) and total STING, and. β-Actin serves as a loading control. (C) Representative histological and immunofluorescence images of tumor sections. From top to bottom: H&E staining; immunohistochemical (IHC) staining for Ki-67 (proliferation); and immunofluorescence (IF) staining for TUNEL (green) and p-STING (red) with DAPI (blue) nuclear counterstain. Scale bars = 100 μm. (D-F) Quantitative analysis of Ki-67-positive cells (D), and TUNEL-positive cells (E), p-STING mean fluorescence intensity (MFI) (F) from images in (C). Data are presented as mean ± SD (n = 5 random fields per mouse from 3 mice). **P < 0.01, ***P < 0.001 (one-way ANOVA with Tukey’s post-hoc test). (G, H) Flow cytometric analysis of tumor-associated macrophages. Gating strategy based on F4/80 versus SSC-A is shown in representative plots (G). (I) The percentage of regulatory T cells (Tregs, CD4⁺Foxp3⁺) within CD4⁺ T cells is shown. (J) Representative flow cytometry plots showing the gating strategy for CD4⁺ and CD8⁺ T cells from live CD3⁺ T lymphocytes. (K, L) Functional analysis of the specific TCR Vβ13.1⁺ CD8⁺ T cell subset. The percentages of IFN-γ⁺ (K) and Granzyme B⁺ (L) cells within the TCR Vβ13.1⁺ CD8⁺ T cell population are shown
Discussion
The treatment of triple-negative breast cancer (TNBC) remains a significant challenge due to its high heterogeneity, aggressive phenotype, and unique immunosuppressive tumor microenvironment [1, 4]. Traditional therapeutic approaches often struggle to overcome its complex biological barriers, necessitating the development of novel treatment strategies. This study innovatively constructed peroxidized hybrid membrane nanovesicles (ox-HyV), which achieve multi-targeting functionality towards tumor vascular endothelial cells through precise fusion of membranes from human induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) and breast cancer cells. This sophisticated design not only fully utilizes the inherent vascular homing properties of iPSC-ECs and the homologous targeting capability of breast cancer cell membranes [20, 21] but, more importantly, endows the nanovesicles with unique functional characteristics through engineered peroxidation treatment.
In recent years, nanocarriers such as lipid nanoparticles (LNPs) and exosomes have demonstrated tremendous potential as gene and drug delivery platforms for regulating endothelial cell function and cancer immunotherapy. Studies have shown that engineered cationic LNPs can efficiently deliver mRNA encoding Tie2 agonists to pulmonary endothelial cells, effectively alleviating inflammation-induced vascular leakage and pulmonary edema, providing proof-of-concept for precise vascular regulation via nanotechnology [22]. Concurrently, exosomes, as endogenous natural nanocarriers, have garnered significant attention for their targeting capabilities. For instance, engineered endothelial cell-derived exosomes can deliver miRNA-125b-5p to lung tissue, exerting therapeutic effects through multiple mechanisms including protecting endothelial barrier integrity, inhibiting apoptosis, and promoting angiogenesis [23]. Furthermore, exosomes have been confirmed to accelerate the diabetic wound healing process by promoting endothelial cell proliferation and migration, highlighting their value in regenerative medicine [24]. In the field of oncology, the application strategies for exosomes are even more diverse. For example, their combination with nanozymes can be used to neutralize pro-inflammatory factors and regulate immune cell balance for treating inflammatory diseases like rheumatoid arthritis [25]; alternatively, loading them with specific siRNA can inhibit tumor-derived exosome secretion, thereby enhancing the efficacy of immune checkpoint inhibitors [26]. On another front, nanodelivery systems for STING agonists are rapidly developing. Systemic administration of these systems enables selective targeting of the tumor immune microenvironment, simultaneously inhibiting tumor angiogenesis and inducing potent anti-tumor T cell immunity [27]. More importantly, STING signaling activation has been proven to promote tumor vascular normalization and tertiary lymphoid structure formation, effectively controlling tumor growth by increasing the infiltration of immune cells such as CD8⁺ T cells and CD11c⁺ DCs, thereby laying a solid theoretical foundation for vascular-immune combination therapy [28]. In summary, these cutting-edge studies collectively outline a clear path towards achieving synergistic “vascular reprogramming - immune activation” through nanotechnology. Our ox-HyV platform emerges precisely within this context. It innovatively integrates the innate targeting capability of iPSC-ECs, the homologous targeting propensity of tumor cell membranes, and the endogenous immunostimulatory capacity triggered by peroxidation treatment into a single hybrid nanovesicle. This achieves a more streamlined and integrated therapeutic modality that does not rely on exogenous genes or synthetic agonists.
Compared to conventional hybrid vesicles primarily used for drug delivery, ox-HyV is endowed with intrinsic immunogenicity through precisely controlled peroxidation treatment, enabling it to directly reprogram the tumor vascular system and efficiently activate the cGAS-STING signaling pathway via its endogenous miRNA cargo [29–31]. This characteristic transforms it from a traditional delivery vehicle into a biological agent with active therapeutic functions, representing a significant innovation in the field of tumor immunotherapy.
At the molecular mechanism level, our systematic transcriptomic analysis revealed the complete signaling network of ox-HyV action. The specific enrichment patterns of differentially expressed genes (DEGs) provide substantial evidence supporting the activation of the cGAS-STING signaling pathway. In-depth analysis indicates that this complex regulatory network achieves its function through multi-level cascading reactions of downstream immune events: the significant enrichment of “cytokine activity” and “cytokine-cytokine receptor interaction” pathways directly reflects the typical transcriptional characteristics of cGAS-STING signaling, demonstrating coordinated expression of type I interferons, chemokines, and other pro-inflammatory cytokines [32]. The synergistic enrichment of the “TNF signaling pathway” reveals a crucial signal amplification mechanism, where cGAS-STING and TNF-NF-κB pathways form a positive regulatory loop, collectively maintaining a sustained immune activation state [33]. Particularly noteworthy is the significant enrichment of the “T cell differentiation” process, which builds a critical bridge connecting innate immunity to adaptive immunity, proving that initial immune signals can be effectively transformed into a functional T-cell-inflamed microenvironment [34]. These findings collectively depict a complete signal regulation network centered on the cGAS-STING pathway.
In comparative analysis of tumor treatment strategies, ox-HyV demonstrates unique comprehensive advantages. Although researchers have developed various hybrid membrane systems in recent years, including lemon-derived vesicles (LEVBD) for enhanced targeting [29], M1 macrophage-tumor cell hybrid nanovesicles (hNVs) for PD-1/PD-L1 blockade [30], and doxorubicin liposome-extracellular vesicle composite systems for improved cytotoxicity [31], ox-HyV adopts a fundamentally innovative strategy. Compared to other STING-activating nanocarriers such as phototherapeutic micelles (TPC@M), cocktail nanoparticles, and self-assembled prodrug nanoparticles (G-M NPs) [35–37], ox-HyV achieves specific activation of the cGAS-STING pathway through the endogenous miR-429/Bcl2 signaling axis while possessing the unique ability to remodel the tumor vascular system, representing a qualitative leap from single-function delivery systems to multi-functional therapeutic platforms.
In terms of clinical application prospects, ox-HyV demonstrates significant potential for synergistic effects with existing standard treatment regimens. Specifically, ox-HyV can significantly enhance TNBC sensitivity to DNA-damaging chemotherapeutic drugs such as anthracyclines and platinum-based agents by improving tumor cell recognition of iatrogenic DNA, providing new insights for overcoming chemotherapy resistance [38, 39]. In the field of radiotherapy, ox-HyV serves as an effective radiosensitizer, capable of establishing an “innate immune alert” state to optimize immune recognition of radiation-induced DNA damage, thereby transforming local radiotherapy into an in situ vaccine with systemic anti-tumor effects [40, 41]. Particularly important is the combined application of ox-HyV with anti-PD-1/PD-L1 immune checkpoint inhibitors using an innovative sequential treatment strategy. This approach addresses two core challenges in TNBC immunotherapy - insufficient T-cell infiltration and functional exhaustion - through a coordinated model of first inducing T-cell infiltration and then maintaining T-cell function [42, 43].
From a translational medicine perspective, we have systematically planned the implementation pathway for clinical translation. Regarding GMP production feasibility, the human iPSC source provides a reliable foundation for establishing standardized cell banks, though process optimization is still required for large-scale expansion and directed differentiation. Safety control of peroxidation modification requires precise definition of the therapeutic window, determining optimal modification conditions through systematic dose-gradient studies. The quality control system needs to establish full-process monitoring indicators from iPSC differentiation to final formulation to ensure batch-to-batch consistency. These considerations provide the technical foundation for the industrial development of ox-HyV.
We fully acknowledge the limitations of the current study, particularly that all in vivo experimental data are derived from mouse syngeneic tumor models. Although these studies provide an important platform for mechanistic exploration, their differences from the human tumor microenvironment require careful evaluation. We have established clear follow-up research plans and will conduct in-depth validation using more clinically relevant systems such as humanized mouse models and patient-derived xenografts (PDX). These studies will become key components in advancing preclinical development.
In summary, ox-HyV represents an innovative vascular-targeted immunotherapy platform that reprograms tumor vascular endothelial cells and remodels the tumor microenvironment through the well-defined miR-429/Bcl2/cGAS-STING signaling axis. Future research will focus on addressing scaling challenges in production processes, validating therapeutic efficacy in more advanced preclinical models, and deeply exploring synergistic effects with existing standard treatment regimens to promote the translation of this innovative strategy into clinical applications.
Conclusions
In this study, we developed a hybrid membrane, designated as ox-HyV, by integrating peroxidation-treated iPSCs-ECs with BC tumor cell membranes. This hybrid membrane specifically targets TVECs and delivers a unique miRNA payload, notably miR-429. The miR-429 effectively down-regulates Bcl2 expression in tumors, alters the secretory phenotype of ECs, and activates cGAS-STING signaling. Consequently, this activation leads to pro-inflammatory cytokine secretion, thereby significantly enhancing the TME and achieving therapeutic efficacy in TNBC treatment.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1
Supplementary Material 2
Supplementary Material 3
Supplementary Material 4
Supplementary Material 5
Supplementary Material 6
Supplementary Material 7
Supplementary Material 8
Supplementary Material 9
Supplementary Material 10
Supplementary Material 11
Supplementary Material 12
Supplementary Material 13
Supplementary Material 14
Supplementary Material 15
Supplementary Material 16
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Wang Y, Lyu Y, Tu K, Xu Q, Yang Y, Salman S, et al. Histone citrullination by PADI 4 is required for HIF-dependent transcriptional responses to hypoxia and tumor vascularization. Sci Adv. 2021;7(35). 10.1126/sciadv.abe 3771 PMC 839727234452909 · doi ↗ · pubmed ↗
- 2Garrido-Martin EM, Mellows TWP, Clarke J, Ganesan AP, Wood O, Cazaly A, et al. M 1(hot) tumor-associated macrophages boost tissue-resident memory T cells infiltration and survival in human lung cancer. J Immunother Cancer. 2020;8(2). 10.1136/jitc-2020-000778 PMC 737546532699181 · doi ↗ · pubmed ↗
- 3Chelvanambi M, Fecek RJ, Taylor JL, Storkus WJ. STING agonist-based treatment promotes vascular normalization and tertiary lymphoid structure formation in the therapeutic melanoma microenvironment. J Immunother Cancer. 2021;9(2).10.1136/jitc-2020-001906 PMC 785294833526609 · doi ↗ · pubmed ↗
- 4Paschoal Barbosa LM, Gomes ER, Barros ALB, Cassali GD, Carvalho AT, Silva JO, et al. Hybrid nanosystem formed by DOX-loaded liposomes and extracellular vesicles from MDA-MB-231 is effective against breast cancer cells with different molecular profiles. Pharmaceutics. 2024;16(6). 10.3390/pharmaceutics 16060739 PMC 1120686638931861 · doi ↗ · pubmed ↗
- 5Hato L, Vizcay A, Eguren I, Perez-Gracia JL, Rodriguez J, Gallego Perez-Larraya J, et al. Dendritic cells cancer immunol immunotherapy cancers (Basel). 2024;16(5). 10.3390/cancers 16050981 PMC 1093049438473341 · doi ↗ · pubmed ↗
- 6Anders CK, Woodcock MG, Van Swearingen AED, Moore DT, Sambade MJ, Laurie S, et al. Evaluating the efficacy of a priming dose of cyclophosphamide prior to pembrolizumab to treat metastatic triple negative breast cancer. J Immunother Cancer. 2022;10(2). 10.1136/jitc-2021-003427 PMC 881978735121644 · doi ↗ · pubmed ↗