Rab1A-Mediated Exosomal Sorting of miR-200c Enhances Breast Cancer Lung Metastasis

Introduction

According to all the new cancers diagnosed recently in women, nearly one-third of the cases were breast cancer, making it the most common cancer in women.¹ Although the 5-year survival rate of breast cancer is 91%, only approximately 20% of breast cancer with metastasis survive for up to 5 years.² Lungs are one of the most common metastatic organs for breast cancer.³ Triple-negative breast cancer is more likely than other types to metastasize to the lungs with limited treatment options.⁴ Therefore, understanding the mechanisms of breast cancer metastasis is critical for cancer treatment and enhancing patient survival.

Tumor metastasis and progression depend on the microenvironment cancer cells develop to facilitate their proliferation and the ability of the tumor to evade immune surveillance. Extracellular vesicles (EVs), especially exosomes, released by cancer cells contribute to pre-metastatic niche formation, modulate the microenvironment, and suppress immune activity, eventually promoting tumor initiation, metastasis, and therapy resistance.^5,6 EVs are secreted membrane vesicles from various cell sources.⁷ Different cell types release unique patterns of EV/exosomal proteins, micro-RNAs (miRNAs), mRNAs, lipids, and metabolites. Exosomes from glioma cells with enriched miR-155 and cytokine interleukin-6 (IL-6) enhance tumor advancement by inducing M2 polarization of macrophages from M1 to M2 via activation of signal transducer and activator of transcription 3 (STAT3).⁸ Highly metastatic triple negative breast cancer cells releasing exosomes promote growth and invasion of low metastatic cells by regulating the expression of tumor suppressors.⁹ Studies over the last decade have established that the miRNA cargo in exosomes is regulated and cell-type dependent. However, the challenge is to distinguish between tumor-released EVs and EVs released from surrounding cells and also from the intracellular components of donor cells. Thus, most exosome studies are limited to cultured cell lines and liquid biopsies like peripheral blood and pleural effusion.¹⁰

Multimodal imaging reports have been used to isolate specific exosomes from solid tumors.¹¹ One study utilizing this tool reported that specific miRNAs, including tumor suppressor miR-193a, were packed inside exosomes mediated by the major vault protein (MVP), promoting colon cancer liver metastasis.¹² In recent years, differential distribution of miRNAs between cells and EVs has been reported in breast cancer. Breast cancer cells secrete miR-122-enriched exosomes suppressing glucose metabolism in the recipient cells of the pre-metastatic niche, leading to metastasis.¹³ These studies indicated that tumor EVs play a crucial role in tumor development and metastasis. However, it remains to be established how tumor cells select miRNAs and encapsulate in the exosomes in breast cancer metastasis.

Here, we generated a stable multimodal imaging report expression system of the exosomes of breast cancer cells. We isolated breast cancer metastasis exosomes from the lungs of mice using a biotin tag and examined the miRNA profile of breast cancer in many stages. Combining the protein–RNA interaction and mass spectrometric analysis, we found that Rab1A expression was upregulated in breast cancer cells, facilitating anti-growth miR-200c encapsulation into exosomes. Releasing the exosome cargo enriched in tumor suppressor miRNAs inhibited anti-tumor cytokine expression by F4/80+ macrophages, alleviating tumor suppressor anti-growth effects on the donor cells. Our study demonstrated the molecular mechanism by which breast cancer metastasizes to the lungs, providing a potential therapeutic strategy against the tumor advance by targeting the exosomal miRNA sorting proteins.

Materials and Methods

Mice

Six to eight-week-old female BALB/c mice were obtained from the Jackson Laboratory (Shanghai, China) and housed at 20–24°C with autoclaved bedding and free of specific pathogen conditions. The mice had at least 2 weeks of acclimation prior to testing conducted. The animal study protocol detailly described the proposed use of animals was approved by the Nanjing Medical University Animal Care and Use Committee prior to all animal experiments performance. All animal experiments were conducted according to the Laboratory Animal Welfare Guidelines of the Nanjing Medical University Animal Care and Use Committee.

Clinical Samples

The specimens from 53 breast cancer patients were obtained in the Department of Breast and Thyroid Surgery, The Affiliated Huai’an No.1 People’s Hospital of Nanjing Medical University, Huai’an, Jiangsu, China, with informed consent from patients. This study was reviewed and approved by the Institute Research Ethics Committee at the Health Department of Huai’an, in compliance with the ethical principles laid out in the Declaration of Helsinki.

Cell Culture

The BALB/c mice syngeneic breast cancer 4T1 cell line was obtained from the American Type Culture Collection (ATCC, Beijing, China) and grown in a humidified 37°C incubator with 5% CO₂ in Dulbecco’s Modified Eagle Medium (DMEM, Thermo Fisher Scientific).

Generation of Stable Breast Cancer 4T1 Cell Expressing pGluc and pBirA

Breast cancer 4T1 cells expressing multimodal tags on the EVs were generated as follows: lentivirus plasmid pGluc containing transmembrane peptide (TM), biotin acceptor peptide (BAP), gaussian luciferase, and green fluorescence protein (GFP), together with pBirA containing biotin ligase (BirA) and cherry fluorescence (kindly provided by Dr Xandra O. Breakefield)¹¹ were co-transduced into HEK293T cells with lentivirus packing vectors VSV-G and pCMVdelta8.2 using Lipofectamine 3000 (Thermo Fisher Scientific). Pseudovirus particles were collected from culture medium and transfected to 4T1 cells at 1×10⁸ pfu/mL. The cells were selected with puromycin, the expression of pGluc and pBirA was verified, and (GlucB) cells were sorted using the BD FACSAria™ III cell sorter (BD Biosciences). The cells with the highest expression of GlucB, verified by fluorescence microscopy (Nikon) at FITC channel, were used for further studies.

Isolation and Purification of Exosomes

For isolation of exosomes from lung tissues, mice were anesthetized with isoflurane, and blood perfusion was performed by the inferior vena cava. The perfusion buffer (1x Hanks’ Balanced Salt Solution (HBSS) without Ca²⁺-Mg²⁺ containing 0.5 mM EGTA, 10 mM HEPES and 4.2 mM NaHCO₃, pH 7.2) pre-warmed at 37°C was pumped through the heart to flow freely. The lung was then perfused for 5 min using a buffer with type III collagenase (0.05%). The perfused primary xenograft tumor and lung tissues were removed and gently disaggregated in DMEM. After incubation at 37°C for 0.5–1 h, the collagenase disaggregated samples were centrifuged at 1000 g, 2000 g, 4000 g, and 10,000g for 10 min, 20 min, 30 min, and 1h, respectively. The supernatant was retained each time. To further purify the exosomes, tissue pellets were resuspended into PBS and loaded on a sucrose gradient (8%, 30%, 45%, and 60% sucrose in 20 mM zwitterionic buffer HEPES, 20 mM Tris-HCl, pH 7.2)¹² described previously.¹⁴ Sucrose gradient-purified exosomes were washed with PBS and then collected for subsequent experiments. For isolating exosomes from the tissue culture medium, EVs in the fetal bovine serum (FBS) were first removed by centrifugation at 100,000g, 4°C for 4–6 h. Ten 100-mm size plates of 4T1 cells were grown in medium DMEM supplemented with 10% EV-depleted FBS for 48 h. The exosome in the medium was collected by the differential centrifugation at 1000g to 100,000g was performed for at least 4–6 h at 4°C. The exosomes in pellet were further purified on a sucrose gradient followed by washing a couple of times with PBS. The size distribution and concentration of the 4T1 exosomes were estimated with Nanoparticle Tracking Analysis (NTA) using NanoSight NS300 (Westborough, MA).

Isolation of Biotinylated Exosomes

Exosomes isolated from 4T1 cells stably co-expressing pGluc and pBirA displaying multimodal tags including biotin could be purified using streptavidin beads. In brief, isolated exosomes (1mg) from tissues were resuspended in 500 μL of binding buffer (50 mM Tri-HCl, 0.1 M glycine-HCl) and incubated with 10 μL of streptavidin magnetic beads (Thermo Fisher Scientific) for 1 h on the rotator at 4°C. The beads were washed thrice with the wash buffer (5 mM EDTA and 0.5% BSA/PBS), biotinylated exosomes were eluted from the beads with 100 μL of 25 mM biotin at 95°C for 5 min.

Mouse Model Study

To generate animal models with metastatic breast cancer in the lung, BALB/c mice were anesthetized with isoflurane, and 1×10⁶ of breast cancer 4T1 cells were administered to the mice by the mammary fat pad injection as previously described. On day 28, mice were euthanized, and mammary xenograft primary tumor and lungs were removed after inferior vena cava perfusion. Xenograft volumes were measured in two perpendicular diameters by a caliper. The volume of individual tumor was calculated using the formula: = length × width²/2.

Isolation of RNA

The total RNA containing small RNAs in cells or tissues was isolated using miRNeasy mini kit (Qiagen) according to the manufacturer’s instructions. The tissue (50 mg) or exosome pellet was suspended into 700 μL QIAzol Lysis Reagent, 140 μL chloroform was added, and the upper aqueous phase was collected. After adding 1.5 volumes of ethanol, the mixture was loaded onto an RNeasy spin column. After centrifugation, the bound nucleic acids were washed with buffer RWT and RPE provided in the kit, and the total RNA containing small RNAs was eluted with 50 µL of RNase-free water. The quantity and quality of RNA was evaluated with NanoDrop.

Analysis of miRNA Microarray

The total RNA (100 mg) was first used as template for cDNA synthesis using miScript II RT Kit (Qiagen) and Qiagen miScript miRNA PCR Array Mouse miRBase Profiler (Cat# 331223) was used to perform the miRNA expression profiling using the Applied Biosystems ViiA 7 Real-Time PCR System. SNORD95, RNU6, SNORD72, SNORD68, and SNORD61 were used as the endogenous control genes normalized for loading RNA input. The cycle threshold (CT) values generated from the exosomes were comparatively analyzed by the Qiagen analysis tool at https://dataanalysis2.qiagen.com/pcr. Software R and GraphPad Prism were used for quantile normalization, analysis of heat maps and scatter plots, respectively. Specific miRNAs were then input into http://userver.bio.uniroma1.it/apps/mienturnet/ using miRTarBase database to determine their validated target mRNAs. Protein hits from miRTarBase with p-value <0.01 were then collectively input into the QIAGEN Ingenuity Pathway Analysis (IPA) as well as functional protein association networks STRING at http://string-db.org/ to illustrate the miRNAs and target protein interactions.

Quantitative PCR Analysis of miRNA Expression

The isolation of RNA and cDNA synthesis were described as above. The diluted cDNA (1:50) was used as template for analysis of mature miRNAs by quantitative real-time reverse transcription PCR (qPCR) using Qiagen miScript SYBR Green PCR Kit with Qiagen predesigned primers. The kits were used according to the manufacturer’s instructions. qPCRs were run in triplicate using the ABI ViiA 7 Real-Time PCR System and U6 and SNORD68 were used as endogenous control genes to be normalized for fold change (FD). The alteration in miRNA level was calculated as an FD calculated with comparative threshold cycle (Ct).¹⁵

Labeling of Exosomes with Dye PKH26

The isolated exosomes were labeled with PKH26 fluorescent dye (Sigma) according to the manufacturer’s instructions. The exosomes in PBS (10 mg/100 µL) were added 250 μL of diluent C (Sigma) with 1 μL of PKH26, mixed with 250 μL of dye solution. Then, the mixture was incubated with 1% bovine serum albumin (BSA) at the equal volume for 5 min at 22°C. Centrifugation for 30 minutes at 100,000× g, the pellet was washed three times with PBS and resuspended in 100 μL of PBS for further experiments.

Transient Transfection Assay

Breast cancer 4T1 cells (3.0×10⁵) were plated in a 6-well plate at a density of cells/well in the Plasmid Transfection Medium (Santa Cruz). Two hundred nanograms of Rab1A CRISPR/Cas9 plasmid or 10 pmol of miRNA was transfected using Lipofectamin™ 3000 or Lipofectamine™ RNAiMAX with Opti-MEM^® Reduced Serum Medium (Thermo Fisher Scientific, Carlsbad, CA). Seventy-two hours after the transfection, the 4T1 cells were harvested for further analysis.

Protein Identification by MS Proteomic Analysis

To prepare protein for the analysis of mass spectrometry (MS), the 10% SDS-PAGE gel was stained with Bio-Rad Coomassie Blue R-250 and the protein bands were excised and washed with ammonium bicarbonate (AB, 50 mM). The excised protein bands were incubated with AB (50 mM)/50% acetonitrile for 20 min at room temperature. After incubation with dithiothreitol (DTT) (50 mmol/L) for 45 min, the excised gel pieces were incubated in iodoacetamide (45 mM) in AB (100 mM) for 25 min. After washing with AB (50 mM), the gel pieces were incubated in AB (50 mM)/50% acetonitrile following trypsin (10 μg/mL) in AB for 1 h. The supernatant was removed and replaced with AB. ZipTipC18 (Millipore) was used to purify peptides and directly loaded on the plate for MS analysis. Peptide mass fingerprints were identified with a MALDI-TOF mass spectrophotometer (Waters). The MS spectra will be used to search proteomics database with the Mascot program (Matrix Science) and Scaffold.

Western Blotting

The cell (10⁷), exosomes or tissue (100 mg) were incubated with 0.5 mL radioimmunoprecipitation assay (RIPA) buffer (Sigma) containing 1× fresh Roche cOmplete™ Protease Inhibitor Cocktail. Proteins in lysates (20 µg) were separated by 8% SDS-PAGE, and the separated proteins were transferred to nitrocellulose membranes (PerkinElmer). Antibodies used were as follows: Rab1A (ab302545), ZEB1 (ab276129), HMGA1 (ab129153), CD63 (ab227892) and GAPDH (ab9485) from Abcam. After washing three times with PBS/, the blot was incubated with the secondary anti-mouse antibodies (horseradish peroxidase conjugate) for 45 min, and the protein bands were detected by exposing them to enhanced chemiluminescence and X-ray film.

Flow Cytometry

Isolated cells were incubated with PBS/0.5% BSA/ for 30 min at 4°C to exclude non-specific reaction. The cells were next stained with anti-F4/80 (Abcam, ab6640) antibodies at 4°C overnight and then washed with PBS/0.5% BSA. To assess intracellular cytokines, cells were fixed and permeabilized with 2% formaldehyde for 30 min, followed by staining with anti-IFNγ (Abcam, ab6640) and anti-IL-1β antibodies for 1 h at 22°C and then washed with PBS/0.5% BSA three times. Washed cells were incubated with the appropriate fluorochrome-conjugated antibodies in PBS with PBS/0.5% BSA for 30 min at 4°C. Data were acquired using a BD FACSCalibur flow cytometer (BD Biosciences) and analyzed with FlowJo^TM version 10 software (BD Biosciences).

Histological Analysis

Tissues were fixed with 10 mL of Formaldehyde Fixative Solution (Thermo Fisher Scientific) for 12 h at 4°C, following the dehydration by passing through a gradient series of ethanol baths (70%, 80%, 90%, and 100%) for 1 h each. Tissues were embedded in paraffin and subsequently cut into 5–10 µm thick sections using a microtome. For immunohistochemistry (IHC) analysis, paraffin-embedded sections were deparaffinized by xylene (Thermo Fisher) and rehydrated by decreasing concentrations of alcohol baths (100%, 90%, 80%, 70%) and water. After incubation with blocking solution (0.5% BSA/PBS) for 30 min, the sample was incubated with primary antibody Rab1A (Abcam, ab302545) following 3 washes with PBS. The sections were subsequently incubated with biotinylated secondary antibodies for 30 min and high-sensitivity streptavidin–HRP for an additional 30 min. 3,3’Diaminobenzidine (DAB) substrate kit (Abcam, ab64238) was used to produce a brown precipitate for detection. For frozen sections, tissues were fixed with for 4 to 8 h at 4°C, following the dehydrated with sucrose (30%) in PBS at 4°C overnight. The sections were then incubated with primary anti-F4/80 antibody solution (Abcam, ab6640) for 1 h at 4°C. The signal was visualized with the secondary anti-mouse antibodies conjugated Alexa Fluor 488 dye (Thermo Fisher Scientific). The cellular nuclei were stained with fluorescent dye 4’,6-diamidino- 2-phenylindole dihydrochloride (DAPI) at a concentration of 500 mg/mL. The slides were visualized with confocal laser scanning microscopy STELLARIS (Leica).

Chromatin Immunoprecipitation (ChIP) Assay

The ChIP assay was carried out to identify the direct interaction of Rab1A and miR-200c, 1 mg of recombinant mouse Rab1A protein (Abcam, ab90795) was incubated with tumor RNA for 3 h at 4°C followed by 1% formaldehyde/PBS for one more hour for cross-linking. Rabbit anti-Rab1A antibody and Dynabeads protein G (50 µL, Thermo Fisher Scientific) were added for 1 h at room temperature. Normal lgG incubated with the beads was utilized as the negative control for the specific interaction. After rinsing the beads with PBS/0.05% Tween 20, the RNA bound to Rab1A/Ab/protein G complex was isolated with a miRNeasy mini kit. Two hundred micrograms of RNA was used to estimate the level of miR-200c with RT-qPCR.

Co-Immunoprecipitation (Co-IP) Assay

To further confirm the direct interaction, DIG-conjugated miR-200c or let-7a (Takara Biomed) was incubated with whole cell extracts of 4T1 lung metastasis, then the miRNA complex was pulled down using anti-Dig Alpha Donor beads (PerkinElmer) in binding buffer (0.1% NP-40, 5 mM EDTA, 0.5 M NaCl, 0.01% SDS). After washing with PBS/0.01%SDS, the miR-200c or let-7a complex was eluted with the protein loading buffer (BioRad), and then analyzed by electrophoresis with 10% SDS-PAGE following an staining with Bio-Rad Coomassie Blue R-250 and standard Western blotting or MS analysis.

Surface Plasmon Resonance (SPR) Analysis

SPR analysis was conducted to identify the direct interaction between Rab1A and miR-200c on an OpenSPR^TM (Nicoya, Lifesciences, CA). The streptavidin sensor chip was first immobilized on the surface of equipment. HBS running buffer (20 mM HEPES, 150 mM NaCl, pH 7.4) was run with the flow rate of 25 µL per minute. After rinsing with 40 mM octyl-β-D-glucopyranoside and 20 mM CHAPS, 300 mg of synthetic miR-200c labeled with biotin at 3’ end was applied on the sensor chip at 1 µg/mL for 20 min until a stable resonance was observed. After the biotin-miR-200c was immobilized on the sensor chip, 3% BSA in the running buffer (0.5% SDS, 50 mM NaOH, 100 mM HCl) was used to block non-specific interactions on the surface. Once the signal was stabilized, recombinant Rab1A protein was run over the sensor surface at 0.1 mM/mL for 20 min until a stable resonance was emerged. A negative control test was also performed by applying Rab1A onto a biotin-conjugated scramble miRNA run-over sensor chip to exclude the non-specific signals. The sensorgrams were versatilely evaluated by kinetic analysis using TraceDrawer^TM software.

Enzyme-Linked Immunosorbent Assay (ELISA) Assay

The cytokine IFNγ and IL-1β in mice lungs were estimated using eBioscience^TM ELISA kits (Thermo Fisher) according to the manufacturer’s protocols. The anti-mouse IFNγ or IL-1β antibody was coated in a 96-well microtiter plate at 1 µg/mL for 12 h at 4°C. The antibody coated plate was rinsed with PBST (PBS/0.05% Tween 20) and then blocked with 150 μL of 1x eBioscience^TM ELISA/ELISPoT Diluent and incubate at 22°C for 1 h. The plate was next washed with PBST and incubated with the cytokine secondary antibody in eBioscience^TM ELISA/ELISPoT blocking solution for 1 h. The plates were rinsed with PBST following the incubation with the avidin conjugated with HRP for 1 h. The substrate was added and incubated for 30 min. The absorbance of plates at 405 nm was estimated using a microtiter plate spectrophotometer (BioTek Synergy HT).

Detection of Exosome Marker CD63 by Immunogold TEM

4T1/GlucB-derived exosomes were isolated from the tumor and resuspended in PBS. Formvar/carbon-supported copper grids (Polysciences, Inc) were floated on droplets of the diluted exosomes (10⁷/mL) for 30 min and blocked with 10 mg/mL BSA in PBS. The adsorbed exosomes on the grids were then floated on 10 µL droplet of anti-CD63 Ab (1 mg/mL, Abcam) in blocking solution at room temperature for 30 min. The grids were rinsed with PBS/BSA (10 mg/mL) following the incubation with protein A-assembled gold nanoparticles (10 nm, OD=10, Cytodiagnostics) (1:45 in PBS/1% BSA) for 45 min. After washing with blocking buffer, the grids were fixed with glutaraldehyde (1%), followed by staining with ammonium molybdate (AM, 1%) on the surface of the grid and washed with 10 mg/mL BSA in PBS. The images of exosomes were visualized using transmission electron microscope (TEM, Thermo Fisher, Spectra 300) and images were recorded by the AMT CCD camera.

Statistical Analysis

All statistical analyses in this study were assessed with GraphPad Prism. Data are presented as mean ± standard deviation (SD). The Student’s t-test was utilized to compare the mean values between two groups and one- or two-way ANOVA test was used to compare the means among three or more groups. The chi-square test was used to estimate the differences between percentages of signal-positive cells (Rab1A, F4/80, eg) in IHC, flow cytometry and confocal microscopy. The p-value <0.05 (*) or 0.01 (**) was considered significantly different.

Results

Generation of Stable Multimodal Imaging Report Expression of Exosomes in Breast Cancer Cell

To isolate the exosomes from cancer solid tissue excluding the contamination of intracellular components, we generated the stable multimodal imaging report expression of exosome in breast cancer cells. The lentivirus vectors pGluc and pBirA were stably co-transduced into breast cancer 4T1 cells, providing the biotin acceptor peptide (BAP) and biotin ligase (BirA) for exosome isolation using streptavidin beads and luciferase Gaussia (Figure 1A). The green fluorescent protein (GFP) and cherry fluorescent tags conjugated on pGluc and pBirA were visualized by confocal microscopy (Figure 1B). The 4T1 cells stably expressing pGluc and pBirA (GlucB) were inoculated in BALB/c mice (1×10⁶ 4T1 cells/each) via mammary fat pad injection. The Gaussia activity was visualized by the substrate coelenterazine (CTZ) administered by intravenous injection. The luciferase assay indicated successful generation of 4T1 cells with multimodal imaging report expression, and the released exosomes/GlucB could be visualized and tracked in vivo (Figure 1C). After differential centrifugation, tumor exosomes were isolated using streptavidin beads and subsequently purified by sucrose gradient centrifugation (Figure 1D). The size distribution (Figure 1E and F) and yield (Figure 1G) of exosomes were analyzed by NanoSight LM10 (NanoSight Ltd.). The purified exosomes were stained with streptavidin-conjugated gold nanoparticles and fixed with glutaraldehyde following the analysis of morphology with transmission electron microscopy (TEM) (Figure 1H). Immunoblot analysis in TEM demonstrated the existence of the exosomal marker CD63 and the lack of the endoplasmic reticulum (ER) marker calnexin from exosomes of naïve 4T1 cells or 4T1/GlucB cells (Figure 1I), excluding the contamination from intracellular proteins.

Figure 1 Identification of exosomes with multimodal imaging report in breast cancer 4T1 cells. (A) Schematic diagram for generation of stable expression of multimodal imaging report in breast cancer 4T1 cells and isolation of extracellular vesicles (EVs) with biotin tag and Gaussia luciferase. 4T1 cells stably transduced with a lentiviral vector expressing membrane-bound Gaussia luciferase (Gluc), biotin acceptor peptide (BAP), and biotin ligase (BirA). (B) Visualization of live 4T1 cells stably expressing Gluc-GFP and BirA-cherry by confocal microscopy. Scale bars, 50 μm. (C) In vivo imaging of EV-Gluc/BirA (GlucB) expression in 4T1 cells. BALB/c mice were administered substrate coelenterazine (CTZ) by intravenous injection 2 weeks after mammary fat pad injection of 1×106 4T1 cells with or without GlucB expression. (D) Sucrose-banded particles from tumor after differential centrifugation. Exosomes were purified using a sucrose gradient (8, 30, 45, and 60% sucrose in 20 mM Tri-Cl, pH 7.2) and collected from the band between 8% and 30% sucrose. (E) Size distribution of exosome populations using NanoSight LM10 (NanoSight Ltd.). (F) Quantification of the exosome size distribution (Mean ± SD). (G) Quantification of exosome yield from tissue (n = 5) by weight. (H) A representative TEM image of released exosomes from naïve 4T1 and 4T1/GlucB cells. Arrows showing anti-biotin staining with streptavidin conjugated with gold nanoparticles. Scale bars, 100 nm. (I) Expression of CD63 and endoplasmic reticulum protein calnexin in exosomes and tissues from breast cancer lung metastasis assessed by Western blotting. Data are representative of three independent experiments (error bars, (SEM.).

Comparative Analysis of Exosomal miRNAs

Given the critical role of miRNA on gene regulation and its enrichment in extracellular vehicles (EVs), we next performed a comparative analysis of miRNomes in the exosomes isolated from normal breast cells, primary breast cancer, and lung metastasis in 4T1/GlucB-inoculated mice using miScript miRNA PCR-Array (Qiagen). The criteria for determining the significance of miRNAs distribution between normal breast tissue versus primary 4T1 breast tumor, and primary breast tumor versus metastatic tumor were deemed on the fold-changes of >3.0 or <-3.0. A total of 58 miRNAs meeting these requirements were accepted for further mechanism analysis (Figure 2A). A heat map of the 26 high- and 32 low-expression miRNAs in metastasis demonstrated gene clustering and sample clustering based on miRNA levels in the exosomes from normal breast tissue, primary breast tumor, and metastatic breast cancer in the lung (Figure 2A). The results of the high-throughput quantitative real-time RT-PCR (qPCR) array were confirmed by qPCR of selected miRNAs, including miR-200c, let-7a, miR-221, miR-18, miR146b, and miR-196a (Figure 2B). The individual qPCR results were consistent with the alterations observed in the qPCR array.

Figure 2 Converse miRNA levels in exosomes and breast cancer tissues. (A) Comparative analysis of exosomal miRNAs from normal breast cells, primary breast cancer, and lung metastasis quantitatively analyzed by microarray. (B) Verification of microarray results with selective miRNAs using individual qPCR. (C) Network analysis of exosomal miRNAs with high and low expression in lung metastasis analyzed by miRecords and STRING. (D) Western blot analysis of miR-200c target gene ZEB1 and Let-7a target gene HMGA1 (left panel). Quantification of band intensity normalized by the intensity of GAPDH (right panel). (E) Analysis of miR-200c and let-7a in tissues using qPCR. *P < 0.05 and **P < 0.01. Data are representative of three independent experiments (error bars, SD).

The biological impact of the differentially expressed miRNAs in tumor exosomes was analyzed using Ingenuity Pathways Analysis (IPA) (Figure 2C). Significantly, miR-200c, let-7a, and miR-221 levels, identified as tumor suppressors, were reduced in tumor tissues but were high in tumor-derived exosomes and highest in the exosomes from lung metastasis. In contrast, as oncogenes, miR-18, miR146b, and miR196a exhibited high expression in tumor tissues and were excluded from tumor-derived exosomes. We then analyzed miRNA target genes in the normal breast tissue, primary tumor, and lung metastasis tissue. Western blot analysis demonstrated that the miR-200c target ZEB1 and let-7a target HMGA1 were induced in metastasis tissues (Figure 2D). The results showed that miR-200c and let-7a levels were decreased in tumor tissues and metastasis but increased in the respective exosomes (Figure 2E). These conflict findings prompted us to hypothesize that breast cancer cells shunt anti-tumorigenesis factors like tumor suppressor miRNAs into exosomes to escape their growth-suppressing effect.

Ras-Related Protein (Rab1A) Facilitates Sorting of miR-200c from Breast Cancer Cells to Exosomes

To test our hypothesis and identify how miRNA repertoires of exosomes differ from those of their donor cells, digoxin (DIG)-labeled miR-200c and let-7a were incubated with exosomal lysates. miRNA complexes were obtained using anti-Dig Alpha Donor beads (PerkinElmer) and separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (Figure 3A). The staining bands were collected, and proteins were identified by in-gel digestion followed by MALDI-TOF mass spectrometry (MS) analysis. By searching MS database (Mascot) and analyzing peptide spectra from MS using Scaffold 4, Rab-1A Rab1A, protein transport protein Sec61 subunit beta (Sec61β), and α-catenin were identified as potential miR-200c interacting molecules. Further, we detected desmoglein-1-alpha (Dsg1α), guanine nucleotide-binding protein 4 (GNB4), and high mobility group protein B1 (HMGPB1) as potential let-7a-binding proteins (Figure 3B).

Figure 3 Sorting of miR-200c from cells to exosomes via Rab1A in breast cancer. (A) Dig-conjugated miR-200c and let-7a complex pulled down from whole cell extracts of 4T1 lung metastasis using anti-digoxin beads and analyzed by electrophoresis followed by Coomassie blue staining. (B) MALDI-TOF-MS analysis of tryptic peptides from the band indicated by red arrows in (A). (C) Western blot analysis of Rab1A proteins from before (top panel) and after anti-Dig Alpha Donor beads pull-down (bottom panel) in 4T1 lung metastasis lysate. (D) ChIP analysis of the interaction between Rab1A and miR-200c using anti-Rab1A antibody pull-down exosomal RNAs and access the miR-200c with miRNA qPCR. (E) SPR analysis of the interaction between Rab1A recombinant protein and biotinylated (Bio-) miR-200c or miRNA scramble control covalently immobilized onto the streptavidin-coated sensor chip. (F) Rab1A knockout (KO) by transfection of Rab1A CRISPR/Cas9 plasmid with and without miR-200c inhibitor in 4T1 cells. qPCR analysis of miR-200c in exosome and donor cell 3 days after transfection. (G) Representative Western blot analysis. (H) Quantification of cell number after 48 h culture in 24-well plate. **P < 0.01. Data are representative of three independent experiments (error bars, SD). Abbreviation: NS, not significant.

Given the critical role of miR-200c in breast cancer progression, we focused on miR200c/Rab1A complex to understand the molecular mechanism of miR-200c sorting to exosomes and its effect on tumor metastasis. The interaction between miR-200c and Rab1A was confirmed by co-immunoprecipitation (Co-IP) and chromatin immunoprecipitation (ChIP) in the DIG-miR-200c complex recognized by the Rab1A antibody (Figure 3C) and by miR-200c qPCR (Figure 3D), respectively. Next, we conducted an optical technique, surface plasmon resonance (SPR) analysis to identify the interaction of miRNA and protein, to provide direct binding evidence and verify the dynamic interaction of Rab1A and miR-200c. For this, biotin-miR-200c was immobilized on the surface of a streptavidin sensor chip in an SPR instrument, and recombinant Rab1A was then flowed on the chip cell to allow coupling. The data of the SPR assay suggested that biotin-miR-200c, not the control scramble miRNA, was directly interacted by Rab1A (Figure 3E).

The role of Rab1A on the miR-200c distribution in breast cancer cells was further analyzed by knocking out (KO) the Rab1A gene by transfection of Rab1A CRISPR/Cas9 KO plasmid (Santa Cruz, sc-422550) followed by treatment with or without miR-200c inhibitor in 4T1 cells. The results showed that Rab1A KO repressed the miR-200c level in exosomes and induced intracellular miR-200c (Figure 3F). Western blotting verified that Rab1A deficiency in breast cancer cell by Rab1A CRISPR/Case9 and miR-200c target ZEB1 alleviated (Figure 3G). And 2×10⁴ of 4T1 cells with and without Rab1A KO were plated into 24-well plate for 48 h culture, and cell counting suggested that Rab1A KO inhibits cell growth due to the intracellular accumulation of miR-200c (Figure 3H).

Impact of miR-200c Sorting to Exosomes by Rab1A on Breast Cancer Lung Metastasis

We demonstrated that breast cancer metastasis cells release EVs highly enriched in tumor suppressor miR-200c mediated by Rab1A trafficking. Next, to identify the effect of miR-200c sorting from cells to exosomes on breast cancer metastasis by immunohistochemistry (IHC) analysis, whereas its expression in adjacent areas of lung metastasis was relatively low (Figure 4A). Using Rab1A expression and percentage progression, we sought to investigate the expression of miR-200c trafficking protein Rab1A in lung sections of breast cancer patients (n=53). High levels of Rab1A were detected in lung of Rab1A⁺ cells in the metastasis from triple-positive (Tri⁺) breast cancer patients (n=28). However, their levels were significantly lower compared to the specimens from triple-negative (Tri⁻) patients (n=25) (Figure 4A and B). Similar results were observed by Western blot analysis of Rab1A expression (Figure 4C). In contrast, miR-200c levels in the metastases from Tri⁻ patients were lower than in Tri⁺ patients (Figure 4D). The Rab1A lentiviral activation particles (Santa Cruz, sc-422550-LAC) were transduced into 4T1/GlucB to overexpress Rab1A before inoculation into mice by mammary fat pad injection. Remarkably, overexpression of Rab1A promoted tumor growth and lung metastasis (Figure 4E and F). The induction of Rab1A by lentiviral activation particles was confirmed by Western blotting (Figure 4G, left panel). Tumor exosomes were isolated from tumor tissues with streptavidin beads, and the qPCR analysis of exosomal miR-200c showed the induction of miR200c by Rab1A overexpression (Figure 4G, right panel). Survival analysis indicated that overexpression of Rab1A resulted in increased death risk in breast tumor xenografted mice (Figure 4H). We investigated the effect of Rab1A attenuation on tumors by administering anti-Rab1A antibody (Ab) into mice with 4T1 tumor xenografts every other day for 2 weeks. We found that the exosomal miR-200c was repressed by anti-Rab1A (Figure 5A). As expected, the anti-Rab1A Ab significantly alleviated tumor growth and lung metastasis (Figure 5B). Further analysis indicated that anti-Rab1A Ab inhibited the trafficking of miR-200c from cells to exosomes (Figure 5C) and prolonged the survival of 4T1 tumor-bearing mice (Figure 5D). These data indicated the therapeutic potential of the novel immunotherapy on breast cancer by targeting Rab1A.

Figure 4 Increased miR-200c trafficking from cell to exosomes promotes breast cancer lung metastasis. (A) Representative IHC analysis of Rab1A level in Tri+ (n=28) or Tri− (n=25) breast cancer tumor (T) specimens and adjacent normal tissues (A) (n=20). Scale bars, 50 μm. (B) Quantification of PD-L1+ cell ratio in the specimens. (C) Representative immunoblot analysis of Rab1A in the tumor tissues. (D) Analysis of miR-200c in lung metastasis using qPCR. (E) Representative 4T1 primary breast tumor (top left) and lung (bottom left) with metastatic nodules from tumor-bearing mice (n=5) at 28 days injected with 4T1 cells with Rab1A activation particles via intravenous injection (50 mg/kg, body weight, every other day). (F) Quantification of primary tumor volumes (left panel) and metastasis nodule number (>1 μm) (right panel). The tumor size was assessed by the following formula: volume = (width)2 × length/2. (G) Representative immunoblots of Rab1A in tumors (left panel) and qPCR analysis of miR-200c in tumor exosomes (right panel). (H) Survival rate of mice in (E). *P < 0.05 and **P < 0.01. Data are representative of three independent experiments (error bars, SD).

Figure 5 Inhibition of miR-200c sorting protein interferes with breast cancer lung metastasis. (A) Survival rate of 4T1 tumor-bearing mice (n=8) treated with different doses of anti-Rab1A antibody or normal lgG control via intravenous injection. (B) Representative 4T1 breast primary tumor (top panel) and lung (middle panel) with metastatic nodules from tumor-bearing mice at 28 days after intravenous injection of 4T1 cells with anti-Rab1A antibody (10 mg/kg, body weight, every other day). Quantification of primary tumor volume (bottom left) and metastasis nodule number (>1 μm) (bottom right). (C) qPCR analysis of miR-200c in tumor tissues and exosomes. (D) Survival rate of mice. *P < 0.05 and **P < 0.01. Data are representative of three independent experiments (error bars, SD).

4T1 Exosomes Inhibit Immune Cell Response in the Lung

We explored the impact of tumor exosomes on their microenvironment that influence tumor growth, invasion, and metastasis, by labeling 4T1-derived exosomes with fluorescence dye PKH26 (Exo/PKH26) and administering them to mice by intravenous injection. Immunofluorescence analysis indicated the presence of Exo/PKH26 in the lung, and the uptake by macrophages was manifested by co-staining exosomes (Exo/PKH26) with F4/80, a marker for macrophages (Figure 6A). Flow cytometry demonstrated that 4T1 exosomes reduced the cytokines interferon-gamma (IFNγ) and interleukin-1 beta (IL-1β) expression in F4/80⁺ cells (Figure 6B). Moreover, anti-Rab1A Ab attenuated the uptake of Exo/PKH26 by F4/80⁺ cells (Figure 6C) and reversed the repression of IFNγ (Figure 6D) and IL-1β (Figure 6E) by tumor exosomes. Collectively, our study demonstrated the molecular mechanism of breast cancer lung metastasis via altered tumor suppressor miRNA200c sorting from cells to exosomes and suggested that disruption of Rab1A-mediated miR-200c sorting could be a viable strategy for the treatment of breast cancer lung metastasis in clinical settings.

Figure 6 4T1 exosomes inhibit the immune response of immune cells in the lung. (A) 4T1-derived exosomes labeled with fluorescence dye PKH16 (Exo/PKH26) and administered to mice by intravenous injection. Immunofluorescence analysis of Exo/PKH26 and F4/80. Cell nuclei stained with DAPI. Scale bars, 50 μm. (B) 4T1 tumor-bearing mice (n=8) treated with 4T1-derived exosomes at 50 mg/kg for 3 days via intravenous injection. Concentration of IFNγ and IL-1β in monocytes of breast cancer lung metastasis assessed by flow cytometry. Numbers above the box indicate percent cells. (C) 4T1 tumor-bearing mice (n=8) treated with different doses of anti-Rab1A antibody via intravenous injection. Flow cytometry analysis of exosome uptake efficacy by monocytes from lungs. (D) Analysis of IFNγ in the lung by ELISA. (E) Analysis of IL-1β in the lung by ELISA. *P < 0.05 and **P < 0.01. Data are representative of three independent experiments (error bars, SD).

Discussion

About 60% of the patients diagnosed with metastatic breast cancer have spread in either the lungs or bones.⁴ Triple-negative breast cancer (ER⁻/HER2⁻/PR⁻) is more likely to metastasize to the lungs than any other type of breast cancer.¹⁶ Recently, there has been growing interest in small RNAs, especially in miRNAs, as mediators of cross-talk between tumor cells and immunocytes via EVs. Given this critical function, identifying the EV communications in normal physiology and disease mechanisms holds great therapeutic promise.

Accumulating evidence suggests that miRNA dysregulation is linked to all stages of the metastatic cascade in breast, colorectal, and ovarian cancers.^17–19 Circulating miRNAs and tumor tissue miRNAs have shown promise as biomarkers for tumor recurrence, metastasis, and response to clinical therapy. Here, we identified a novel mechanism of metastatic progression of breast cancer to the lung through exosomal sorting of tumor suppressor miR-200c mediated by Rab1A. Modulating miRNA-200c distribution alleviated its anti-tumorigenesis impact on breast cancer cells and promoted primary tumor proliferation and lung metastasis. We also showed that miR-200c target ZEB1 was elevated when miR-200c was sorted into exosomes. Our approach to modeling breast cancer lung metastasis by inoculating breast cancer 4T1 cells with multimodal imaging report expression using a lentiviral vector might be complementary to current in vivo studies of exosomes in breast metastasis mouse models. In this study, using a biotin-streptavidin-based approach, we demonstrated that tumor cells selectively eliminate tumor suppressor miRNA like miR-200c, let-7a and miR-221 from the donor cell mediated by exosomes. An improved understanding of the sorting tumor suppressor miR-200c cross-talk with an oncogene Rab1A provides a novel mechanism for the tumor advance and provide new therapeutic potential for the treatment of cancer.

The miR-200 family members, including miR-200c, downregulate the expression of transcription factors ZEB1 and ZEB2,²⁰ which are transcriptional inhibitors of E-cadherin and contribute to epithelial to mesenchymal transition (EMT).²¹ ZEB1 and ZEB2 are widely expressed by tumor cells and myeloid immune cells, including macrophages, monocytes, and dendritic cells (DCs). ZEB1 promotes tumorigenesis by activating the oncogene p53 expression, abandoning addiction to Ras, and against antiproliferative therapies. ZEB1 contributes to maintaining a bivalent chromatin configuration mediated by TGF-β in breast cancer cells.²²

Recent studies suggest that the patients with metastatic breast cancer present the higher levels of miR-200c and miR-141 in peripheral blood compared to those with localized breast cancer without metastasis or healthy controls. The miR-200c level in breast cancer patients was modulated by the FOXP3-KAT2B-miR-200c/141 axis.²³ The deficiency of miR-200 in lung cancer cells enhanced lung cancer metastasis mediated by Notch signaling Jagged1 and Jagged2 activation.²⁴ The miR-200c level in the lung tumor tissues was lower than in normal tissues.²⁵ Thus, the majority of the findings in the studies focused on the prognostic and tumorigenic role of miR-200c in tumor tissues including breast and lung tissues as well as peripheral blood are conflicted. Our data clarified the contradictory phenomenon that cancer cells sort miR-200c into exosomes and secret it into biofluids, decreasing intracellular miR-200c but elevating its level in biofluids like plasma and pleural effusion in cancer patients. Another breast cancer metastasis small RNA let-7a was also enriched in breast cancer-released exosomes but lower in donor cells, and the let-7a target HMGA1 was induced in tumor cells. In this study, we selected miR-200c for further exploration.

We explored the mechanism underlying the uneven distribution of intracellular and exosomal miRNAs by integrating Co-IP and MS strategies²⁶ and identified potential miR-200c-binding molecules, including Rab1A, Sec61β, and α-catenin, and three proteins Dsg1α, GNB4, and HMGPB1 as potential let-7a-binding molecules. As the small GTPase Ras-related protein, RAB1A, regulates the intracellular membrane trafficking and formation of fusion membranes in transport vesicles. Upregulation of RAB1A contributes to breast cancer, especially triple-negative type.²⁷ Sec61β is the component of SEC61 and mediates the transport of signal peptides across the endoplasmic reticulum (ER). SEC61G, a member of SEC61 family, promotes breast cancer development and metastasis. Although there is no evidence showing that Sec61β is involved in breast cancer, the accumulation of nuclear EGFR, which is associated with poor clinical prognosis for breast cancer, is Sec61β-dependent.²⁸ α-Catenin promotes hepatocellular carcinoma metastasis mediated by ubiquitin ligaseRNF219 and acts as a tumor suppressor via E-cadherin in breast cancer.²⁹

We also confirmed that breast tumor exosomes were taken up by F4/80⁺ macrophages in the lung and reduced the expression of anti-tumor cytokines. As the exosome recipient cells, macrophages represent more than 50% of the tumor-infiltrating immune cells.³⁰ In vitro and in vivo studies have revealed that macrophages play diverse roles in inflammation^7,31 and cancer development, ranging from the anti-tumor activity in early progression to the tumor-promoting role in established cancers by activating anti-apoptotic programs.³² Despite extensive studies, the mechanisms underlying the modulation of macrophages by exosomal miRNAs involved in tumor metastasis are unclear and remain to be explored.

Currently, antibodies against PD-1/PD-L1, atezolizumab and pembrolizumab, have been approved for various anti-tumor therapies.^33–35 Similarly, given the observation that Rab1A-mediated elimination of miR-200c by exosomes promotes breast cancer development and progression, interfering with miR-200c sorting to exosomes by Rab1A antibody represents a potential therapeutic strategy for breast cancer. Consistent with this notion, anti-Rab1A Ab inhibited miR-200c trafficking from cells to exosomes, decreasing breast cancer metastasis to the lung and prolonging survival of 4T1 tumor-bearing mice.

Our data indicate that the generation of an exosome tracing system with multimodal imaging report elucidates mechanisms of breast cancer metastasis. The distribution of miR-200c and sorting protein Rab1A are potential biomarkers for early detection of breast cancer metastasis. Inhibition of the sorting protein with the antibody provides a potential novel immunotherapy for breast cancer.