Cancer cell-derived migrasomes harboring ATF6 promote breast cancer brain metastasis via endoplasmic reticulum stress-mediated disruption of the blood-brain barrier

Patients and samples

A total of 109 retrospective plasma samples were included in this study, including 32 samples from BCBM patients, 36 samples from BC patients, and 41 samples from individuals with no evidence of disease and serving as healthy controls. The sample size was determined using G*Power software (effect size = 0.4, α = 0.05, and power = 0.8) with a minimum requirement of 22 samples per group. These plasma specimens were collected before any interventional procedures were performed at Liaocheng People’s Hospital, Liaocheng Hospital Affiliated to Shandong First Medical University (Liaocheng, Shandong, China) from June 2019 to March 2023. As part of the screening process, the inclusion and exclusion criteria were as follows: (1) the BCBM group included patients with confirmed BCBM based on imaging or pathology findings and excluded patients with other malignancies (excluding breast cancer) or brain diseases (e.g., stroke and encephalitis); (2) the BC group included patients with pathologically confirmed BC and excluded patients with metastases or other malignancies (excluding breast cancer); and (3) the healthy control group included healthy individuals without any neoplasms, excluding individuals with chronic diseases (e.g., diabetes and hypertension) or recent infections and inflammatory conditions. Ethical approval was obtained from the Ethics Committee of Liaocheng Hospital Affiliated to Shandong First Medical University (Approval no. R2022035). All the participants provided written informed consent.

Cell culture

The cell source and culture conditions of the human brain-targeting breast carcinoma cell line, 231-BR, and its parental cell line, MDA-MB-231, were as described previously^35–37. The human brain microvascular endothelial cell line, hCMEC/D₃ (Mingzhoubio, Zhejiang, China), was cultured in basic medium supplemented with 5% fetal bovine serum (FBS), 1% endothelial cell culture additive for endothelial cell culture, and 1% penicillin-streptomycin in 5% CO₂ at 37°C.

Animal model

All animal experimental protocols were approved by the Laboratory Animal Welfare and Ethics Committee of Liaocheng People’s Hospital (Approval no. D2022057) and performed according to the Guide for the Care and Use of Laboratory Animals. Female BALB/c nude mice (8 weeks old weighing 18–20 g; Vital River Laboratory Animal Technology, Beijing, China) were bred and housed at a controlled temperature (22 ± 1°C) and exposed to a constant 12 h light-dark cycle in the animal facility.

Mig-BCBM and Mig-BC were injected into mice via tail veins at a concentration of 200 μg/mL twice weekly to evaluate the effects of migrasomes on BCBM. Four weeks later, the BCBM mouse models were established by injecting 231-BR or ^GFP+231-BR cells into the left ventricles, as described previously³⁵.

BBB integrity evaluation

Two percent Evans blue [EB] (2 mL/kg) was injected through tail veins 2 h prior to sacrificing the mice. Specifically, the mice were anesthetized with an intraperitoneal injection of 1% pentobarbital sodium (50 mg/kg) before removing excess EB by trascardial perfusion with 200–300 mL of heparin saline (0.9% sodium chloride and 20 U/mL heparin sodium). The mice were decapitated and the brains were collected for sagittal cutting after removing the hippocampus. One hemisphere was frozen sectioned for a BBB integrity examination. EB penetration, which indicated BBB permeability, was photographed using a fluorescence microscope (Olympus Corp., Tokyo, Japan). The other half-hemisphere was used to evaluate the EB amount. The half-hemisphere was weighed, grinded, placed in dimethylformamide (1 ml dimethylformamide per 100 mg brain tissue), incubated at 60°C for 24 h, centrifuged at 94 g for 5 min, and the absorbance at 620 nm was measured with a spectrophotometer. EB standard curve was drawn to calculate EB content.

Isolation of migrasomes

Previously established protocols³⁸ were refined and adapted to isolate migrasomes from cultured cells or human serum samples. Cells were cultured in fibronectin-coated 150 mm dishes and collected into 50 mL tubes for isolation from cultured cells. In the case of human serum samples, a 5-mL serum sample per individual was centrifugated at 845 g for 10 min. Large debris in the pellets were subsequently removed by centrifugation at 1,000 g for 10 min followed by 4,000 g for 20 min. After centrifuging at 20,000 g for 30 min, pellets were collected as crude migrasomes. To obtain precise migrasome fractionation, density gradient centrifugation was performed utilizing an OptiPrep kit (LYSISO1; Sigma-Aldrich, city, state, country). The crude migrasome pellets were resuspended with 650 μL of Optiprep dilution buffer, then mixed with 400 μL of 1 × extraction buffer, 30 μL of 2.3 M sucrose, and 505 μL of 60% OptiPrep, compounding 1,585 μL of crude migrasome suspension with 19% Optiprep. Then, a step gradient was built starting with 27% (800 μL), followed by 22.5% (1,000 μL), crude migrasome resuspension (19%, 1,000 μL), 16% (1,000 μL), 12% (900 μL), and 8% (300 μL). The gradient was centrifuged at 150,000 g for 4 h in an MLS-50 rotor. Fractions were then separately collected from each gradient with volumes ranging from 500 to 700 μL. Each fraction was then diluted with an equal volume of PBS and centrifuged at 20,000 g for 30 min to isolate the migrasomes as pellets.

Transmission electron microscopy (TEM)

Migrasomes were gently resuspended in 20 μL PBS and fixed in an equal volume of 25% glutaraldehyde for 2 h for negative staining. The fixed migrasomes were absorbed onto a carbon-coated copper grid for 5 min and stained with a single drop of 2% phosphotungstic acid for 3 min. Finally, the grids were allowed to air-dry for 15 min and observed under a JEM-1200EX electron microscope (Jeol, Tokyo, Japan).

For resin-embedded ultrathin section staining, migrasomes isolated by density gradient centrifugation were fixed, pre-embedded in agarose, and post-fixed with 1% osmium tetroxide for 2 h. The samples were dehydrated using an ethanol-acetone series, infiltrated, and embedded in resin. Ultrathin sections (60–80 nm) were cut, collected on copper grids, and stained with 2% uranyl acetate (8 min) and 2.6% lead citrate [8 min (CO₂-free)]. After washing and drying, the sections were imaged by TEM for migrasome analysis.

Scanning Electron Microscopy (SEM)

Migrasomes was fixed with 25% glutaraldehyde for 2 h, followed by incubation for 12 h in a refrigerator at 4°C. The samples were then washed. After washing, the samples were dehydrated through a graded ethanol series (50%, 70%, 80%, 90%, 100%, 100%, and 100%). The dried samples were then mounted on copper plates and observed under a Hitachi SU8010 electron microscope (Hitachi, Tokyo, Japan).

Wheat-germ agglutinin (WGA) staining

Confocal Petri dishes (35 mm) were precoated with 2 μg/mL of fibronectin and incubated at 37°C for 6 h to facilitate cell attachment. The fibronectin solution was discarded after incubation. The digested cells were then seeded to attain a density of 30–40% and allowed to adhere for 20 h. The migrasomes were then stained with WGA³⁹. Finally, images were captured using a confocal microscope (Leica Microsystems, Mannheim, Germany).

Western blotting

Proteins from cells, migrasomes, or mouse tissue were extracted in RIPA reagent. Protein concentrations were determined using a BCA protein assay kit (#P0010S; Beyotime, Beijing, China). Proteins were separated by 10–15% SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% non-fat milk and incubated with primary antibodies against PIGK (#ab201693, 1:1,000; Abcam, Boston, MA, USA), CPQ (#000012845, 1:1,000; Sigma-Aldrich, St. Louis, MO, USA), ZO-1 (#13663, 1:1,000; Cell Signaling Technology, Danvers, MA, USA), VE-cadherin (#13116, 1:1,000; Cell Signaling Technology), or β-actin (#sc-47778, 1:1,000; Santa Cruz, CA, USA) overnight at 4°C. After washing with TBST, the membranes were incubated with HRP-conjugated secondary antibody [HRP-labeled goat anti-rabbit IgG (H+L), #A0208; Beyotime; HRP-labeled goat anti-mouse IgG (H+L), #A0216; Beyotime] at room temperature for 1 h. Protein-antibody complexes were visualized using an enhanced chemiluminescence reagent (#34580; Thermo Fisher Scientific, Waltham, MA, USA) and captured on a chemiluminescence imaging system (Tanon, Shanghai, China).

Real-time quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted using Trizol reagent (Thermo Fisher Scientific) to quantify gene expression in brain tissues and cells. Specific primers of targeting sequences were designed and synthesized, as listed in Table S1. The first strand of complementary DNA (cDNA) was synthesized from the isolated RNA using a reverse transcription kit (Beyotime) following the manufacturer’s protocol. The cDNA was subjected to RT-qPCR using the TB Green Premix Ex Taq II kit (Takara, Dalian, China). The amplification mix contained 0.8 μL of forward primer (10 μM), 0.8 μL of reverse primer (10 μM), 2 μL of cDNA (50 ng/μL), 10 μL of 2 × master mix, 0.4 μL of ROX II, and 6 μL of ddH₂O. PCR was performed on the 7,500 Real-time PCR System (Applied Biosystems, Foster City, CA, USA).

Digital PCR (dPCR)

Migrasomes from serum samples of BCBM patients, BC patients, and healthy individuals were isolated. Total RNA was extracted from the migrasomes using TRIzol reagent (Thermo), followed by reverse transcription according to the instructions (Beyotime). cDNA was subjected to dPCR using a HQ dPCR EvaGreen Master Mix Kit (Sniper, Suzhou, China). dPCR amplification was performed on a dPCR system (Sniper), as described previously⁴⁰. The amplification mix contained 8 μL of cDNA, 11 μL of 2 × dPCR EvaGreen Master Mix, 1.0 μL of upstream primer (10 μM), 1.0 μL of downstream primer (10 μM), and ddH₂O to adjust the total volume to 22 μL. The amplification conditions included incubation at 60°C for 5 min for droplet stabilization and activation of Taq enzyme at 95°C for 15 min. Each cycle consisted of denaturation at 95°C for 15 s and annealing/extension at 60°C for 30 s, which was repeated for a total of 40 cycles. Fluorescence signal acquisition was performed at 60°C for 60 s in a single cycle. The amplification curves were interpreted using SightPro software (Sniper).

Hematoxylin and eosin (H&E) staining

Mouse brain tissues were fixed with 4% paraformaldehyde for 24 h. Then, the brain tissues were embedded with paraffin and sectioned at a thickness of 5 μm. After dewaxing, the sections were stained with an H&E staining kit (Solarbio, Beijing, China). Brain metastatic nodules were counted under a microscope, as previously described³⁵.

4D label-free proteomics analysis

SDT buffer (composed of 4% SDS and 100 mM Tris-HCl at pH 7.6) facilitated sample lysis and protein extraction of migrasomes. LC-MS/MS analysis was performed on a Bruker timsTOF Pro mass spectrometer interfaced with a Nanoelute system. Peptides were loaded onto a 25 cm × 75 μm (i.d.), 1.9-μm resin C18 reversed-phase column (Thermo Fisher Scientific) in 95% buffer A (0.1% formic acid in water) and eluted with a linear gradient of buffer B (99.9% acetonitrile and 0.1% formic acid) at 300 nL/min. The mass spectrometer was operated in a positive ion mode with an electrospray voltage of 1.5 kV to analyze precursors and fragments within an m/z range of 100–1,700. The timsTOF Pro (company, city, state, country) was operated in the PASEF mode, with parameters set as follows: ion mobility coefficient (1/K0) range of 0.6–1.6 Vs cm²; 1 MS scan followed by 10 MS/MS PASEF scans; and an active exclusion period of 24 s. Raw MS data for each sample were pooled and analyzed using MaxQuant 1.6.14 software for identification and quantitation.

Immunofluorescence (IF) staining

Cells were fixed with 4% paraformaldehyde for 30 min, then permeabilized with 0.5% Triton X-100 (company, city, state, country). After blocking with 5% BSA, the cells were incubated with primary antibodies for 1 h or overnight at 4°C. After washing, the cells were incubated with the secondary antibody for 60 min. The cells were then washed with PBS and stained with DAPI for 10 min. Images were captured under a BX53 microscope (Olympus, Tokyo, Japan).

Lentivirus infection

ATF6 gene knockdown in hCMEC/D₃ or 231-BR cells was achieved by infecting the cells with an shRNA containing lentivirus Lv-shATF6 (Genechem, Shanghai, China). Lv-shNC, which had no target gene, was used as the negative control. hCMEC/D₃ cells were infected with Lv-shATF6 or Lv-shNC at a multiplicity of infection (MOI) of 100 for 12 h according to the manufacturer’s instructions. A selection process was performed with 2 μg/mL of puromycin (Beyotime) for 48 h to obtain the successfully infected cells. ATF6-knockdown migrasomes (Mig-shATF6) were subsequently isolated from the stable 231-BR cells with ATF6 knockdown for further analysis.

Endothelial cell layer permeability test

hCMEC/D₃ cells were seeded onto the polyethylene terephthalate membrane (0.4 μm pore size) in the upper chamber of a 24-well Transwell plate. The cells were allowed to grow and fuse, then treated with 8 μg/mL of Mig-BCBM for 48 h. After removing the culture medium from the upper chamber, 10 mg/mL of FITC-labeled dextran was added to the hCMEC/D₃ culture medium. The lower chamber was filled with 1 mL of hCMEC/D₃ culture medium. Fifty microliters of the culture medium in the lower chamber were withdrawn at each time point (10, 20, 30, 40, and 50 min). The optical density (OD) value was detected using a SpectraMax microplate multifunctional enzyme-linked immunosorbent assay reader (Molecular Devices, San Jose, CA, USA) with excitation and emission wavelengths set at 488 nm and 520 nm, respectively.

Tumor cell migration through the endothelial cell layer

hCMEC/D₃ cells were seeded at a density of 2 × 10⁴ on polyethylene terephthalate films (8-μm pore size) within the upper chamber of a 24-well Transwell chamber. Following growth to confluence, the cells were treated with 8 μg/mL of Mig-BCBM for 48 h. The medium in the upper chamber was aspirated and GFP-transfected 231-BR cells were added (1 × 10⁵/mL resuspended in 100 μL of serum-free high-glucose DMEM). The lower Transwell chamber was filled with 500 μL of high-glucose DMEM containing 10% FBS. The upper chamber was carefully removed and fixed in 4% paraformaldehyde following incubation for 12 h. After a gentle rinse with PBS, ^GFP+231-BR cells and hCMEC/D₃ cells remaining in the upper chamber that had not migrated to the lower surface of the filter were gently removed using cotton swabs. The nuclei were stained with Hoechst 33,342 (Solarbio, Beijing, China) at a dilution of 1:5,000 for 10 min. Finally, the lower chamber of the filter was observed under an inverted fluorescence microscope and six random fields of view were photographed per well.

Ex vivo fluorescence imaging

Whole-brain ex vivo fluorescence images were captured using an IVIS Spectrum in vivo imaging system (PerkinElmer, Waltham, MA, USA)⁴¹ with a red filter. The images were analyzed using Living Image software 4.4. The fluorescence signals were separated through the application of multispectral imaging technology.

Statistical analysis

All results were confirmed in at least three independent experiments. All quantitative data are presented as the mean ± standard deviation (SD). Statistical analysis was performed using GraphPad Prism 9 (San Diego, CA, USA) or SPSS26.0 software. For quantitative data, the distribution was assessed using the Shapiro-Wilk normality test. Comparisons between two groups were analyzed using the unpaired or paired t-test for normally distributed data, while comparisons among multiple groups were analyzed using one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test for pairwise comparisons. The Mann-Whitney U test (for two groups) or the Kruskal-Wallis H test (for multiple groups) was applied for non-normally distributed data, followed by Dunn’s test for pairwise comparisons. A P < 0.05 was considered statistically significant.

Mig-BCBM promotes migration of 231-BR cells across vascular endothelial layers in vitro

The unique human 231-BR/MDA-MB-231 cell model with a syngeneic nature was utilized to investigate the functions of cancer-derived migrasomes. The MDA-MB-231 parental cells exhibited the ability to metastasize to diverse sites without organ specificity, whereas the 231-BR cell variant specifically targeted the brain. 231-BR cells metastasize exclusively to the brain with 100% specificity, making 231-BR cells a valuable preclinical research model for brain metastatic breast cancer⁴². Migrasomes derived from 231-BR and MDA-MB-231 cells were isolated and purified, and were designated as Mig-BCBM and Mig-BC, respectively (Figure 1A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Mig-BCBM facilitates 231-BR cell migration through vascular endothelial layers in vitro. (A) Isolation and purification of migrasomes. 231-BR or MDA-MB-231 cell-derived migrasomes were isolated and purified using density gradient centrifugation. (B) TEM imaging of migrasomes. TEM revealed the ultrastructural features of isolated migrasomes. The green arrows indicate migrasomes. The red arrows indicate retraction fibers. In the left panel of the ultrathin section, the magnified view in the upper right corner corresponds to the area marked by the red box. Scale bars, 500 nm. (C) SEM visualization of migrasomes. The green arrows indicate migrasomes. Scale bar, 500 nm. (D) Western blotting analysis of 231-BR or MDA-MB-231 cells and cell-derived migrasome Mig-BCBM and Mig-BC using antibodies against NDST1, CPQ, PIGK, and TSG101. 231-BR and MDA-MB-231 cells, along with Mig-BCBM and Mig-BC, were normalized with total protein. The grey values of the western blot bands were quantified using ImageJ software. (E) 231-BR or MDA-MB-231 cell-derived migrasomes were stained with WGA probes. The magnified images in the right panel correspond to the areas delineated by the red boxes. Scale bars, 5 μm. (F) Schematic of endothelial cell layer permeability assay. Polyethylene terephthalate membrane, 0.4 μm pore size. (G) Effect of Mig-BCBM on hCMEC/D₃ cell layer permeability. n = 3. (H) Schematic of ^GFP+231-BR cell migration through vascular endothelial layers. Polyethylene terephthalate membrane, 8 μm pore size. (I) Effect of Mig-BCBM on the migration of 231-BR cells across vascular endothelial layers. Representative images of ^GFP+231-BR cell migration through vascular endothelial layers. Scale bars, 50 μm. (J) Quantification of ^GFP+231-BR cell migration through vascular endothelial layers. n = 5; *P < 0.05, **P < 0.01.

Migrasomes were identified using TEM, SEM, western blotting, and WGA staining (Figure 1B-1E). Under TEM, cancer-derived migrasomes were round or nearly round with membrane encapsulated and sizes ranged from 400 to 2,000 nm; some migrasomes had RFs (Figure 1B). SEM analysis further revealed the presence of RFs within some migrasomes (Figure 1C). Previous studies have demonstrated that NDST1, CPQ, and PIGK are enriched within migrasomes, whereas NDST1, CPQ, and PIGK are absent or barely detectable in exosomes, thus serving as migrasome-specific protein markers^38,43. NDST1, CPQ, and PIGK proteins were effectively detected through western blotting in 231-BR and MDA-MB-231 cells and cell-derived migrasome Mig-BCBM and Mig-BC (Figure 1D). Conversely, the exosome marker, TSG101, was absent in Mig-BCBM and Mig-BC, confirming the successful isolation of migrasomes from the cultured cells. WGA can be utilized as an effective probe for migrasomes³⁹. As shown in Figure 1E, migrasomes were conspicuously distributed around 231-BR and MDA-MB-231 cells. Taken together, these data indicated successful isolation of migrasomes from 231-BR and MDA-MB-231 cells.

A crucial event in brain metastasis involves the traversal of cancer cells across the BBB. Cerebral vascular endothelial cells have a pivotal role in determining BBB permeability⁴⁴. First, an endothelial cell layer permeability test was performed to assess the effect of 231-BR cells-derived migrasomes on vascular endothelial barrier permeability. The hCMEC/D₃ cell layers were exposed to Mig-BCBM or Mig-BC for 48 h, followed by the addition of FITC-labled dextran (Figure 1F). Upon addition of dextran for 10, 20, 30, 40, and 50 min, the fluorescence intensity in the Mig-BCBM treatment group was significantly higher than the Mig-BC group (Figure 1G). The effect of Mig-BCBM on the metastatic capacity of 231-BR cells when traversing vascular endothelial barriers in vitro was then determined. The hCMEC/D₃ cell layers were treated with Mig-BCBM or Mig-BC for 48 h, then ^GFP+231-BR cells were introduced and allowed to migrate for 12 h (Figure 1H). The number of ^GFP+231-BR cells that traversed the hCMEC/D₃ cell layer in the Mig-BCBM group was significantly higher than the Mig-BC group (Figure 1I and 1J). Taken together, these findings suggest that Mig-BCBM effectively enhances the permeability of vascular endothelial layers and facilitates the metastasis of 231-BR cells across endothelial barriers in vitro.

Mig-BCBM enhances BBB permeability and promotes BCBM in vivo

Mig-BCBM was pre-injected into the tail veins of mice to evaluate the effect of 231-BR cells-derived migrasomes on BCBM, then the BCBM models were established by intracardiac injection of 231-BR cells (Figure 2A). Four weeks later, the brain tissues were collected and the brain metastases nodules were counted using H&E staining. As shown in Figure 2B and 2C, Mig-BCBM significantly enhanced brain metastases compared to the Mig-BC and control groups, with mean brain metastatic nodule counts of 23.6, 10.0, and 3.8, respectively. No detectable metastatic foci were observed in the liver and lungs (Figure S1). These data indicated that Mig-BCBM promotes BCBM in vivo.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Mig-BCBM enhances BBB permeability and promotes BCBM in vivo. (A) Schematic illustration of a model designed to detect the effect of Mig-BCBM or Mig-BC on BCBM. (B and C) Effect of Mig-BCBM on BCBM in vivo. H&E staining of brain tissues (B) and quantification of brain metastases (C). Top right: H&E-stained whole brain section (scale bars, 1 mm). Brain metastasis nodule images (scale bars, 50 μm) show magnified view of black-framed area. n = 5. (D) Schematic illustration of a model designed to detect the effect of Mig-BCBM on disrupting BBB. (E and F) Effect of Mig-BCBM on BBB permeability. Evans blue (EB) fluorescence images (E) and concentration (F) in the brains. Scale bars, 10 μm. n = 5; *P < 0.05, **P < 0.01, ***P < 0.001.

Given that BBB maintains brain microenvironment homeostasis and restricts the entry of exogenous macromolecules and cells into the brain parenchyma, whether Mig-BCBM promotes BCBM via disrupting the BBB was determined. Mig-BCBM was pre-injected into the tail veins of mice. EB, a tracer for monitoring BBB disruption, was injected into the mice after 4 weeks (Figure 2D). Compared to the Mig-BC and control groups, EB-stained brain sections exhibited increased EB uptake in the Mig-BCBM-injected models (Figure 2E). Quantitative analysis of EB extravasation demonstrated significantly elevated cerebrovascular permeability in the Mig-BCBM group (8.729 μg/mL) compared to the control (6.772 μg/mL) and Mig-BC groups (7.857 μg/mL; Figure 2F). These results indicated that Mig-BCBM enhances BBB permeability in vivo.

Mig-BCBM internalizes into hCMEC/D₃ cells and downregulates t ZO-1 and VE-cadherin expression

Having determined that Mig-BCBM enhances the permeability of BBB in vivo and vascular endothelial layers in vitro, the impact on vascular endothelial cells was determined. First, it was imperative to ascertain whether Mig-BCBM could be transferred to hCMEC/D₃ cells. Recipient cells were incubated with WGA-labeled Mig-BCBM for 24 h. The results showed a large amount of Mig-BCBM internalized into hCMEC/D₃ cells (Figure 3A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Mig-BCBM modulates BBB permeability by altering ZO-1 and VE-cadherin expression in hCMEC/D₃ cells. (A) Internalization of Mig-BCBM into hCMEC/D₃ cells. hCMEC/D₃ cells were incubated with WGA-labeled Mig-BCBM for 24 h. The right panel shows the line profile of representative cell images in the dashed box using Image-Pro Plus software. The right panel shows the line profile of the image of a representative cell within the dashed box, generated using the Image-Pro Plus software. The fluorescence images revealed efficient endocytosis of Mig-BCBM into hCMEC/D₃ cells. Scale bar, 10 μm. (B) Downregulation of ZO-1 and VE-cadherin mRNA expression by Mig-BCBM. n = 3 (C and D) Decrease in ZO-1 (C) and VE-cadherin (D) protein expression following Mig-BCBM treatment. n = 3. (E and F) Effects of Mig-BCBM on the expression, but not subcellular localization, of ZO-1 and VE-cadherin. Immunofluorescence staining revealed that Mig-BCBM treatment reduced the levels of ZO-1 and VE-cadherin expression in hCMEC/D₃ cells without affecting the subcellular localization. Scale bar, 20 μm. (G and H) Effects of Mig-BCBM on ZO-1 and VE-cadherin expression in mouse brains. Mig-BCBM (200 μg/mL) was injected into the mice brain. ZO-1 and VE-cadherin mRNA (G) and protein (H) expression was downregulated. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001.

The hCMEC/D₃ cell tight junction protein, ZO-1, and adherens junction protein, VE-cadherin, have major roles in BBB integrity²⁵. The effect of Mig-BCBM on the expression and subcellular localization of ZO-1 and VE-cadherin was further investigated. As shown in Figure 3B–3F, Mig-BCBM significantly decreased ZO-1 and VE-cadherin mRNA and protein expression in hCMEC/D₃ cells compared to the Mig-BC control group. However, Mig-BCBM did not alter the subcellular localization of ZO-1 and VE-cadherin within the recipient cells (Figure 3E and 3F).

To further validate these findings in vivo, ZO-1 and VE-cadherin expression in brain tissues of mice following the injection of Mig-BCBM was determined. ZO-1 and VE-cadherin mRNA and protein expression in brain tissues was downregulated after Mig-BCBM injection compared to Mig-BC and the blank control groups (Figure 3G and 3H), which was consistent with the in vitro data. Additionally, immunofluorescence staining confirmed that Mig-BCBM did not alter the membrane localization of ZO-1 and VE-cadherin in brain tissues (Figure S2A and S2B). Taken together, these data suggested that BCBM-derived migrasome Mig-BCBM increased BBB permeability by downregulation of vascular endothelial cell ZO-1 and VE-cadherin.

Mig-BCBM induces ER stress via ATF6 in hCMEC/D₃ cells

Having determined that Mig-BCBM promotes BBB permeability by downregulating ZO-1 and VE-cadherin, the underlying mechanism was clarified. 4D label-free proteomics analysis was performed on Mig-BCBM and Mig-BC. Quantitative mass spectrometry revealed differential protein enrichment between the two groups with ATF6 notably more abundant in Mig-BCBM compared to Mig-BC (Figure 4A). Western blotting analysis was performed to verify enrichment of ATF6 in Mig-BCBM. Consistent with the proteomics analysis results, ATF6 was significantly more abundant in Mig-BCBM compared to Mig-BC (Figure 4B). ATF6 expression in both cell lines was determined to ascertain whether enrichment of ATF6 in Mig-BCBM was due to differential expression between 231-BR and MDA-MB-231 cells. ATF6 was significantly upregulated in 231-BR cells compared to MDA-MB-231 cells (Figure 4C and 4D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Mig-BCBM induces ER stress in hCMEC/D₃ cells. (A) Volcano plot depicting differential abundance proteins in Mig-BCBM compared to Mig-BC. Mig-BCBM and Mig-BC underwent 4D label-free quantitative mass spectrometry analysis. The red dots signify a Mig-BCBM/Mig-BC abundance ratio ≥ 2 (P < 0.01), while the green dots denote a Mig-BCBM/Mig-BC abundance ratio ≤ 0.5 (P < 0.01). n = 3 biologically independent experiments. (B) ATF6 expression in Mig-BCBM compared to Mig-BC. Mig-BCBM and Mig-BC were normalized with total protein and subjected to western blotting analysis. The gray values of the western blotting bands were quantified using ImageJ software. (C and D) Enhanced ATF6 expression in 231-BR cells. RT-qPCR (C) and western blotting (D) analyses revealed a significant upregulation of ATF6 in 231-BR cells compared to MDA-MB-231 cells. n = 3. (E) Elevated expression of ER stress markers in hCMEC/D₃ cells treated with Mig-BCBM. RT-qPCR analyses revealed significant upregulation of six ER stress markers, including ATF6, GRP78, CHOP, ATF4, EIF2AK3, and PPPIR15A, in hCMEC/D₃ cells that had been exposed to Mig-BCBM. n = 3. (F) Enhanced protein expression of ER stress markers in Mig-BCBM-treated hCMEC/D₃ cells. Western blotting analyses confirmed the elevated protein levels of three ER stress markers (ATF6, GRP78, and CHOP) in hCMEC/D₃ cells that were subjected to treatment with Mig-BCBM. n = 3. (G) Protein expression of additional ER pathway markers (PERK and IRE1 pathways). Western blot analyses revealed no significant changes in p-PERK, p-eIF2α, or IRE1α in Mig-BCBM-treated hCMEC/D₃ cells compared to Mig-BC. n = 3; *P < 0.05, **P < 0.01; ns, no significance.

Significant enrichment of pathways related to the endoplasmic reticulum and Golgi vesicle-mediated transport and the UPR within Mig-BCBM was confirmed based on Gene Ontology (GO) analysis (Figure S3). Notably, ATF6 has emerged as a pivotal regulator of the UPR and has a crucial role in maintaining ER function and protein homeostasis⁴⁵. Therefore, we postulated that entry of Mig-BCBM into hCMEC/D₃ cells may elicit ER stress in recipient cells. The mRNA levels of various ER stress markers, including ATF6, GRP78, CHOP, ATF4, EIF2AK3, and PPPIR15A, were all increased in Mig-BCBM-treated hCMEC/D₃ cells compared to the Mig-BC control group (Figure 4E). Upregulation of ATF6, GRP78, and CHOP protein expression further confirmed the ability of Mig-BCBM’ to induce ER stress in hCMEC/D₃ cells (Figure 4F). Additional ER stress pathways were further examined by assessing p-PERK, p-eIF2α (PERK pathway), and IRE1α (IRE1 pathway). Mig-BCBM did not significantly alter activation of p-PERK, p-eIF2α, and IRE1α compared to Mig-BC (Figure 4G). These results suggested Mig-BCBM may induce ER stress of hCMEC/D₃ cells via the ATF6 pathway.

Expression of ER stress-related genes in brain tissues of mice following injection of Mig-BCBM was determined to further validate the capacity of Mig-BCBM to induce ER stress in vivo. The levels of ATF6, GRP78, and CHOP expression were significantly upregulated in the Mig-BCBM group compared to Mig-BC or PBS group mRNA and protein expression (Figure 5A and 5B). These findings confirmed that Mig-BCBM induces ER stress in vivo.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Mig-BCBM induces ER stress in mouse brains in vivo. (A) Elevated transcript levels of GRP78, CHOP, and ATF6 in Mig-BCBM-exposed mouse brains. (B) Enhanced expression of GRP78, CHOP, and ATF6 protein in Mig-BCBM-treated mouse brains. *P < 0.05, **P < 0.01.

Mig-BCBM reverses the upregulation of ZO-1 and VE-cadherin caused by ATF6 knockdown in hCMEC/D₃ cells

ATF6 is transferred to the Golgi apparatus during ER stress, where ATF6 is cleaved by S1P and S2P proteases, and subsequently transferred to the nucleus to activate ER stress-related genes⁴⁵. We hypothesized that the ATF6 pathway has an important role in downregulation of ZO-1 and VE-cadherin induced by Mig-BCBM; inhibiting this pathway might reverse the downregulation effect. Lentivirus was used to knockdown ATF6 expression in hCMEC/D₃ cells. As shown in Figure 6A–C, sh-ATF6 lentiviruses were successfully transduced into hCMEC/D₃ cells, resulting in a significant downregulation of ATF6 in mRNA and protein expression. Mig-BCBM was added after knocking down ATF6 in hCMEC/D3 cells. The levels of ZO-1 and VE-cadherin expression were upregulated after ATF6 knockdown in hCMEC/D₃ cells but this phenomenon was reversed after addition of Mig-BCBM (Figure 6D and 6E). Moreover, knockdown of ATF6 in hCMEC/D₃ cells led to suppression of the metastatic potential of 231-BR cells traversing vascular endothelial barriers. However, the addition of Mig-BCBM into hCMEC/D₃ cells reversed this inhibitory effect (Figure 6F). These findings indicated that ATF6 knockdown elevates the expression of ZO-1 and VE-cadherin in hCMEC/D₃ cells, whereas Mig-BCBM reverses this upregulation.

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Mig-BCBM reverses ZO-1 and VE-cadherin upregulation induced by ATF6 knockdown in hCMEC/D₃ cells. (A–C) Efficient knockdown of ATF6 in hCMEC/D₃ cells. The knockdown efficiency of ATF6 in hCMEC/D₃ cells was examined using immunofluorescent staining (A), RT-qPCR (B), and western blotting (C). Scale bars, 50 μm; n = 3. (D–E) ATF6 knockdown upregulates ZO-1 and VE-cadherin expression, which is reversed by Mig-BCBM. The mRNA (D) and protein (E) levels were upregulated after ATF6 knockdown in hCMEC/D₃ cells, but this phenomenon was reversed after the addition of Mig-BCBM. n = 3. (F) ATF6 knockdown inhibits the metastatic potential of 231-BR cells traversing vascular endothelial barriers, but this phenomenon was reversed after the addition of Mig-BCBM. Scale bars, 50 μm; n = 5. (G) Efficient knockdown of ATF6 in Mig-BCBM (Mig-shATF6). n = 3. (H) Effect of Mig-shATF6 on hCMEC/D₃ cell layer permeability. n = 3. (I) Scale bars, 50 μm. Effect of Mig-shATF6 on the migration of 231-BR cells across vascular endothelial layers. Quantification of ^GFP+231-BR cell migration through vascular endothelial layers. n = 5; *P < 0.05, **P < 0.01, ***P < 0.001.

ATF6 was knocked down in Mig-BCBM (Mig-shATF6) to further investigate the functional role of ATF6 in migrasome-mediated BBB disruption and 231-BR cell transendothelial migration (Figure 6G). Quantitative analysis revealed that ATF6 depletion significantly reduces BBB permeability compared to the Mig-shNC control, as demonstrated by decreased dextran fluorescence intensity across all time points (10–50 min; Figure 6H). Furthermore, transendothelial migration assays using GFP-labeled 231-BR cells showed a marked reduction in cell penetration through hCMEC/D₃ monolayers in the Mig-shATF6 group vs. Mig-shNC (Figure 6I). These findings suggested that ATF6 is a critical mediator in migrasomes by regulating BBB integrity and metastatic cell transmigration, and thereby influencing final metastatic outcomes.

Mig-BCBM promotes BCBM by triggering ER stress in hCMEC/D₃ cells

Cells were pretreated with ER stress inhibitors (4-PBA), followed by Mig-BCBM exposure for 48 h, then ER stress markers expression was determined to further clarify the Mig-BCBM’ induction effect on ER stress in hCMEC/D₃ cells. As shown in Figure 7A, upon pretreatment with 4-PBA, the levels of CHOP, GRP78, and ATF6 expression in hCMEC/D₃ cells were suppressed, but Mig-BCBM significantly attenuated downregulation of these molecules caused by 4-PBA. Moreover, 4-PBA led to upregulation of ZO-1 and VE-cadherin in hCMEC/D₃ cells but this effect was reversed upon addition of Mig-BCBM (Figure 7B).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

Mig-BCBM enhances BCBM via the induction of ER stress. (A) Mig-BCBM rescued the downregulation of CHOP, GRP78, and ATF6 caused by 4-PBA in hCMEC/D₃ cells. n = 3. (B) Mig-BCBM rescued the upregulation of ZO-1 and VE-cadherin caused by 4-PBA in hCMEC/D₃ cells. n = 3. (C) 4-PBA reversed the Mig-BCBM-induced downregulation of ZO-1 and VE-cadherin in vivo. n = 3. (D) Schematic illustration of an in vivo rescue experiment to confirm the role of Mig-BCBM in enhancing BCBM through ER stress. The mice were intraperitoneally injected with 4-PBA, followed by the administration of Mig-BCBM through the tail vein 2 h later. The mice underwent a twice-weekly injection schedule for 4 weeks. The BCBM models were established by injecting ^GFP+231-BR cells into the left ventricle. (E–F) 4-PBA reduced the brain metastases caused by Mig-BCBM in vivo. Fluorescence images of ^GFP+231-BR cells (E). Whole-brain ex vivo fluorescence images and ROI value of fluorescence intensity of mouse metastases (F). n = 3; *P < 0.05, **P < 0.01.

Rescue experiments were performed to further validate the effect of Mig-BCBM on ZO-1 and VE-cadherin expression through induction of ER stress in vivo. The mice were initially injected with Mig-BCBM via the tail veins and after a 2-h interval 4-PBA was administered intraperitoneally. 4-PBA reversed Mig-BCBM-induced downregulation of ZO-1 and VE-cadherin (Figure 7C).

In vivo rescue experiments were performed to further confirm the mechanism underlying Mig-BCBM in enhancing brain metastases via induction of ER stress. The mice were injected intraperitoneally with 4-PBA and 2 h later Mig-BCBM was administered via the tail veins. This injection schedule was repeated twice weekly for a total of 4 weeks. The BCBM models were established by injecting ^GFP+231-BR cells into the left ventricles (Figure 7D). Whole-brain ex vivo fluorescence imaging revealed that Mig-BCBM facilitates the development of brain metastases, whereas 4-PBA effectively mitigates the brain metastases triggered by Mig-BCBM (Figure 7E and 7F).

ATF6 expression in human serum migrasomes as a potential molecular marker for diagnosing BCBM

Having confirmed that Mig-BCBM downregulates ZO-1 and VE-cadherin expression in hCMEC/D₃ cells via the release of ATF6 resulting in BBB impairment and ultimately facilitating the process of BCBM, the association between the levels of ATF6 in human serum migrasomes and BCBM were determined. First, serum samples were collected from patients with BCBM and primary BC, and healthy individuals. The serum migrasomes were purified using a density gradient centrifugation technique (Figure 8A). The SEM and TEM images clearly demonstrated that the majority of serum migrasomes exhibited a round or nearly round morphology enveloped by membranes (Figure 8B and 8C). Furthermore, western blotting analysis revealed the presence of NDST1, CPQ, and PIGK in the serum migrasomes (Figure 8D), confirming successful isolation from serum migrasome samples.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8

Expression of ATF6 in serum migrasomes of BCBM patients. (A) Flowchart of isolation and purification of serum migrasomes. (B–D) Identification and characterization of serum migrasomes. The human serum migrasomes were identified using TEM (B), SEM (C), and western blotting (D). The green arrows indicate migrasomes in (B). Scale bars, 0.3 μm in (B) and 1 μm in (C). (E) Elevated protein levels of ATF6 in serum migrasome samples of BCBM patients compared to BC patients and healthy controls. n = 3. (F and G) Evaluation of ATF6 mRNA expression in serum migrasomes utilizing digital PCR (dPCR). Representative imaging obtained from the dPCR analysis of ATF6 mRNA expression within serum migrasomes (F) and quantitative analysis (G). BCBM, n = 32; BC, n = 36; control, n = 41; **P < 0.01, ***P < 0.001.

Western blotting analysis was performed to determine the association between the levels of ATF6 protein in human serum migrasomes and BCBM. The findings revealed significantly elevated levels of ATF6 in the serum migrasome samples of BCBM patients in contrast to the levels of ATF6 in BC patients and healthy controls (Figure 8E). Given the limitations of western blotting in clinical molecular diagnosis, dPCR experiments were performed to further assess the diagnostic potential of ATF6 levels in serum migrasomes. Interestingly, ATF6 was detected in the serum migrasomes of BCBM and BC patients, and healthy controls. Importantly, a significant increase in ATF6 expression was noted in the serum migrasome samples of patients with BCBM compared to BC and healthy controls (Figure 8F and 8G), suggesting that detection of ATF6 expression in human serum migrasomes may potentially serve as a molecular marker for diagnosing BCBM.