A MUC1-ALIX complex regulates extracellular vesicle cargo loading with activated SRC to enhance pancreatic cancer progression

Abstract

Extracellular vesicles (EVs) are critical mediators of intercellular communication in the tumor microenvironment and play an essential role in tumor growth and metastasis. MUC1 has been identified in tumor cell-derived EVs, but its function in that context has not been well investigated. We show that MUC1 is highly enriched in EVs derived from pancreatic cancer cell lines and patient tumors but not normal pancreas. Pretreatment of mice with tumor derived MUC1 positive EVs promoted tumor cell growth and metastasis in vivo. MUC1 positive EVs enhanced tumor cell growth, motility and invasion in vitro. Proteomic profiling revealed that MUC1-positive EVs contain protein cargo distinct from MUC1-negative EVs, which implicates MUC1 in selective EV cargo regulation. Using complementary biochemical and biophysical analysis, we demonstrate that the cytoplasmic tail of MUC1 expressed in pancreatic cancer cells interacts with programmed cell death 6-interacting protein (ALIX) and influences EV biogenesis and loading of specific oncogenic cargo (Phospho_Tyr416-Src) that are known to associate with MUC1. EVs from SRC inhibitor Bosutinib treated cells showed reduced effects on cell viability, migration, and invasion of MUC1 knockout cells as compared to EVs from untreated cells. Thus, MUC1-positive EVs containing activated Src promoted tumor cell proliferation and in vivo tumor growth, migration, and metastasis of orthotopic pancreatic cancer cells. Collectively, our findings identify a previously unrecognized role for MUC1 in influencing EV composition and function: influencing loading of oncogenic protein cargoes into EVs. Our data highlight a novel mechanism controlling tumor-specific EV cargo loading and demonstrate that oncogenic cargo in MUC1-associated EVs contributes to pancreatic cancer progression.

Introduction

Extracellular vesicles (EVs) play an essential role in tumor progression and metastasis(1). Several types of EVs have been described, including ectosomes, exosomes, and oncosomes. Ranging in size from 30-150 nm, EVs circulate in the organism and contain cell-derived biomolecules, including DNA, RNA, protein, and metabolites(2). Two major categories are ectosomes, generated by direct budding of the plasma membrane, and exosomes, which are of endosomal origin. Exosomes are produced through budding of endosomal membranes of multivesicular bodies (MVBs) to create intraluminal vesicles (ILVs)(3). Tumor-derived EVs possess cargo that differs from those released from their noncancerous cellular counterparts(4); however, the mechanisms regulating differential cargo sorting are poorly understood.

EVs can serve as intercellular communication carriers that cancer cells exploit during tumor progression(5). Secreted EVs from tumor cells transport tumor-promoting signals(2, 6), influence the tumor microenvironment(7), and establish pre-metastatic niches(8–11). Recently, PDAC-derived exosomes were shown to contain macrophage migration inhibitory factors, which together with other molecules can educate the liver to form pre-metastatic niches; therefore, exosomes may be a prognostic marker for the development of PDAC liver metastasis(12). Mucins (MUC1, MUC4, MUC5AC, MUC6, and MUC16) have been identified in EVs of several types of cancer cells(13).

EVs derived from lung, colon, breast, and pancreatic cancer (PDAC), have shown selective enrichment of the type I transmembrane protein Mucin1 (MUC1)(14, 15). MUC1 is generally expressed on the surface of ductal epithelia. MUC1 has two subunits: the longer N-terminal subunit (MUC1-N) and a shorter C-terminal subunit (MUC1-C), which contains extracellular, transmembrane, and cytosolic portions. The cytoplasmic tail of MUC1 (MUC1CT) contains multiple conserved serine and tyrosine residues that are phosphorylated by growth factor receptors and intracellular kinases(15) and influence several signaling pathways(14, 15). However, the molecular functions of MUC1 in cancer progression are not fully elucidated. Cells with high levels of endogenous MUC1 expression produced significantly more ectosomes than cells with low endogenous expression, in part because the glycocalyx associated with MUC1 directly influenced the formation of membrane pearls that form vesicles(16).

Here, we examined MUC1 levels in EVs of paired tumor and adjacent tissue from pancreatic cancer patients and different cell lines and found that MUC1CT was highly associated with tumor-derived EVs. Differential proteome analysis and functional analysis of EVs revealed that MUC1-associated EVs carry unique proteins that increased properties of tumor growth and metastasis in vivo, and pancreatic cancer cell migration and invasion in vitro. We further investigated the role of MUC1 in the biogenesis of EVs and mechanisms regulating sorting of specific cargo into EVs. Our data shows that the MUC1CT influences EV cargo sorting and secretion by associating with programmed cell death 6-interacting protein (ALIX). MUC1 and associated oncogenic proteins (e.g. SRC) bind directly to ALIX to influence EV production and cargo composition.

Materials and Methods

All resources and commercial kits used in this study are detailed in the Key Resources Table, which is provided in the supplementary material.

Cell culture

The S2-013 cell line is a subclone of a human pancreatic tumor cell line (SUIT-2) derived from liver metastasis(17). The HPAF-CD11 cell line is a subclone of the human pancreatic adenocarcinoma cell line derived from ascitic fluid (HPAF) and displays a well-differentiated state, including the formation of ductal structures with polarized, long columnar-shaped cells and the presence of secretory granules in the cytoplasm(18). BxPC-3 and Capan-2 were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). The HEK293 cell line is an immortalized human embryonic kidney cell line(19). All the cell lines were grown in Dulbecco’s Modified Eagle Medium (ThermoFisher Scientific, Waltham, MA, USA) containing high glucose and L-glutamine, supplemented with 10% fetal bovine serum (Valley Biomedical Inc., Winchester, VA, USA) plus 100 μg/mL streptomycin and 100 units/mL Penicillin-Streptomycin (P/S) (Corning, NY, USA) (DMEM + 10% FBS + 1% P/S at 37°C and 5% CO2).

S2-013 MUC1F cells, HPAF-CD11 cells, BxPC-3 cells, and Capan2 cells were genetically deleted of MUC1 using CRISPR Cas9 technology. CRISPR Cas9 guide RNAs (gRNA) for human MUC1 were obtained from GenScript (Genscript Biotech Corporation, Piscataway, NJ, USA). The MUC1 CRISPR Guide RNA sequences are as follows: sgRNA1: TGAACTGTGTCTCCACGTCG, sgRNA2: TACCATCAATGTCCACGACG, and sgRNA3: GCAACAGTTGTTACGGGTTC. The knockout of the MUC1 protein was confirmed by western blot.

ALIX was knocked down in S2-013 MUC1F cells, Capan2 cells, BxPC-3 cells, and Panc1 MUC1F cells by small interfering RNAs (siRNAs) processed from short hairpin RNAs (shRNAs). MISSION shRNA plasmids were obtained from Sigma-Aldrich (SHCLND-NM_013374, Sigma-Aldrich, St. Louis, MO, USA). The target sequences are CCTGAATTACTGCAACGAAAT and CCAGAACAAATGCAGTGATAT. The knockdown of ALIX was confirmed by Western blot.

Organoid Culture and Experiment

The patient-derived xenograft (PDX) organoids PA137 and PA717 were cultured as previously described(20). For investigating the functions of MUC1 positive EVs and MUC1 negative EVs, 1 × 10³ cells were suspended in 50 μL Matrigel per well and seeded to 24-well plate and then incubated with or without 3 μg/mL MUC1 positive or MUC1 negative EVs for 6 days (fresh EVs added every other day). A Cytation 3 Cell Imaging multi-Mode Reader was used to capture the size of the organoids at Day 6. CellTiter-Glo kit was used to determine cell viability of organoids 6-day after EVs treatment.

In vivo animal studies

All animal experiments were performed in accordance with UNMC/UNO IACUC policies. Mice were housed in a pathogen-free animal facility at UNMC in Omaha. All the animals were maintained in a temperature-controlled room (22°C), with humidity at 55% and a 12-hour light-dark cycle. Mice were fed a standard laboratory chow diet and water ad libitum. Both males and females were used in this study.

6–8-week-old nude mice received 5 μg of EVs (MUC1-positive cell or MUC1-negative cell-derived EVs) or PBS in a total volume of 50 μL by tail vein injection every other day for three weeks. For the tumor progression study, pre-educated mice underwent orthotopic implantation of S2-013 MUC1KO cells or HPAF-CD11 MUC1KO cells to the pancreas. The mice were monitored every day in the first week after surgeries, then twice a week until tumor growth was observed. After that, mice were monitored 3-4 times weekly and ultrasound-imaged once a week to accurately track tumor growth and size. The ultrasound data was generated in the small animal ultrasound core of UNMC, and all the ultrasound images were analyzed by Vevo LAB 5.6.0.

Both male and female mice were included in all in vivo experiments; as no statistically significant sex-specific differences in tumor growth or treatment responses were observed, data from both sexes were pooled for analysis.

Rapid autopsy patient samples

Pancreatic tumors, metastases, and other tissue specimens were obtained with consent and Institutional Review Board (IRB) approval from surgically resected samples or decedents through the Rapid Autopsy Program at the University of Nebraska Medical Center. Organs were harvested within 3 hours postmortem, and the specimens flash-frozen in liquid nitrogen or placed in formalin for immediate fixation.

Extracellular vesicle isolation from cultured cells grown in dishes

Cells were seeded into 150 mm dishes (7 × 10⁶ cells/dish for S2-013 and 1.2 × 10⁷ cells/dish for HPAF-CD11 cells, and 9 × 10⁶ cells/dish for Capan-2 and BxPC-3 cells) and changed the medium to DMEM containing FBS depleted of bovine serum EVs by ultracentrifugation at 110,000 × g for 16 h. The cell-conditioned medium was collected after 44 hours, and the cell number was assessed using the Countess II Automated Cell Counter (Thermo Fisher, Waltham, MA, USA). The assembled media was first subjected to a centrifugation step of 300 × g for 10 min at 4°C to pellet and remove cells. All following centrifugation steps were performed at 4°C. Next, the supernatant was spun at 2,000 × g for 10 min to remove debris and apoptotic bodies. Then, to pellet and collect large EVs, the supernatant was centrifuged at 15,000 × g for 45 min. This supernatant was next subjected to ultracentrifugation at 110,000 × g for 2 h in a SW28 Swinging Bucket rotor (k factor of 204, Beckman Coulter, Fullerton, CA) to enrich the small EVs. The crude pellet was resuspended in a large volume of PBS, followed by ultracentrifugation at 110,000 × g for 2 h to wash the sample. At no time during the process were samples subjected to temperatures below 4°C.

Extracellular vesicle isolation from human tissue samples

Fresh tissues from primary tumor and metastases sites were dissected in cold PBS and incubated on ice overnight. The tissue solution was gently dissociated by pipetting (10 mL pipette). After dissociation, the sample was first applied to a 60 mm then a 40 mm mesh filter (BD Biosciences, Baltimore, MD, USA) to remove any remaining larger particles from the single-cell suspension. The supernatant containing dissociated cells and EVs was spun down at 300 × g for 10 min to pellet cells. The cell pellet was washed once in PBS by resuspension and repelling, and cellular proteins were extracted as described below. EVs were then purified from the supernatant in the same manner as described above for cell culture-derived EVs. Before ultracentrifugation, the supernatant was filtered through a 0.22 μm pore PES filter (Millipore, Burlington, MA, USA). EVs were then purified from the supernatant in the same manner as described above for cell culture-derived EVs.

Negative stain transmission electron microscopy

Samples were first fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer at RT for 30 min. Samples were then deposited on formvar/Silicone Monoxide coated grids (TED PELLA INC, Redding, CA) for 2-5 min, followed by Nanovan negative staining for 30 s at RT, blot off excess stain as before and allowed to dry 2 min. Imaging was performed on a FEI Tecnai G2 Spitit Transmission Electron Microscope. Images are acquired digitally with an AMT digital imaging system (supported by NIH 1 S10 RR024650 01A1).

Particle size and concentration measurement by Nanoparticle Tracking Analysis (NTA)

Samples in PBS were analyzed by NTA using the NanoSight NS300 system (NanoSight Ltd) configured with a high-sensitivity sCMOS camera (OrcaFlash2.8, Hamamatsu C11440, NanoSight Ltd). Previous analysis (NTA software, version 3.1, build 3.1.54). In brief, each sample was diluted in particle-free PBS and introduced manually into the machine. After optimization, settings were kept constant between measurements for that session. The ambient temperature was recorded manually and did not exceed 25°C. Five videos, with 60 s duration, were recorded for each independent replicate (n = 5).

Beads-assisted Flow Cytometry Analysis

30 μg EVs isolated by ultracentrifugation were incubated with 10 μL aldehyde/sulfate-latex beads (ø = 4 μm; 5.5 × 10⁶ particles/mL; Invitrogen, Carlsbad, CA) in 1 mL PBS for 30 minutes at room temperature. The reaction was stopped by 100 mM glycine and 2% BSA in PBS for 30 minutes, then centrifuged at 14,800 × g for 1 minute. The beads with EVs were incubated with 0.01%(v/v) Triton X 100 and 2% BSA in PBS for 10 minutes, then incubated in primary antibodies (dilution 1:500) and 2% BSA in PBS overnight at 4°C. After washing with 2% BSA in PBS, the samples were incubated with secondary antibodies and 2% BSA in PBS for one hour at 4°C. Lastly, the samples were resuspended in PBS, and data were acquired using conventional flow cytometers (LSRII, BD Biosciences) and analyzed with the FlowJo software. Bead-decorated EVs were gated for a size of 4 μm in diameter beads based on FCS/SSC parameters so that unbound EVs or possible antibody aggregates were excluded from the analysis.

Nanoscale Fluorescence Analysis

EVs were isolated and purified from the PDAC patients’ primary tumor tissues. We used PKH26, CD63-BV421, CD81-PECF594, Cy5-O14E-MUC1 and Cy7-AR20.5-MUC1 to label 100 μg EVs in bulk (antibody dilution 1:100). Murine monoclonal antibody mAb-AR20.5 recognizes the DTRPAP sequence in the VNTR region of human MUC1(21), and mAB-O14E recognizes the RYVPPSSTDRSPYEKVSAG sequence of the CT region of human MUC1. Then, the EVs were quenched by 2 mL of 10% BSA in PBS and brought the volume up to 8.5 mL in serum-free media. Added 1.5 mL 0.971 M sucrose solution by pipetting slowly and carefully into the bottom of the tube, making sure not to create turbulence. The exosomes-PKH67 solution remained on top of the sucrose cushion. Samples were centrifuged at 190,000 × g for 2 hours at 4°C. The supernatant was carefully aspirated and the crude EV pellet was resuspended in PBS by gentle pipetting. Following centrifugation at 110,000 × g for 2 hours at 4°C, the labeled EVs were present in the pellet. All nanoFACS experiments, unless otherwise stated, were performed on the Amnis ImageStream Imaging Flow Cytometer. Instrument setup and acquisition settings were performed using previously described settings(22, 23). All the data were analyzed using FlowJo.

Sample preparation for mass spectrometry

The protein concentration was estimated in each sample using a BCA Protein Assay Kit (Pierce). 100 μg of proteins from each sample was diluted to 100 μL volume with 100 mM ammonium bicarbonate (AMBIC). Proteins were reduced with 5 μL of 200 mM tris(2-carboxyethyl) phosphine (TCEP) (1 h incubation, 55°C) and alkylated with 5 μL of 375 mM iodoacetamide (IAA) (30 min incubation in the dark, room temperature). The reduced and alkylated proteins were purified with acetone precipitation at −20°C overnight. The next day, protein precipitates were collected by centrifugation at 8,000 × g for 10 minutes at 4°C, and pellets were briefly air-dried and resuspended in 100 μL of 50 mM AMBIC. Protein digestion was carried out using 2.5 μg of trypsin per sample (16 h incubation, 37°C). The next day, samples were dried using a speed vacuum and then desalted with C18 spin columns (Pierce). Clean peptides were dried out again with speed vacuum and resuspended in 0.1% formic acid, then analyzed using a high-resolution mass spectrometry nano-LC-MS/MS Tribrid system, Orbitrap Fusion^™ Lumos^™ coupled with UltiMate 3000 HPLC system (Thermo Scientific).

LC-MS/MS and bioinformatics analysis

1 μg of each sample was loaded onto trap column Acclaim PepMap 100 (75 μm×2 cm C18 LC Columns, Thermo Fisher Scientific) at a flow rate of 4 μL/min and separated with a Thermo RSLC Ultimate 3000 (Thermo Fisher Scientific) on a Thermo Easy-Spray PepMap RSLC C18 column (75 μm×50 cm C-18 2 μm, Thermo Fisher Scientific) at a flow rate 0.3 μL/min and 50°C, with a step gradient of 9%–25% solvent B (0.1% FA in 80% acetonitrile) from 10–57 min and 25%–45% solvent B from 57–62 min, with a 90 min total run time. The MS scan was done using a detector: Orbitrap resolution 120,000; scan range 350-1800 m/z; RF lens 30%; AGC target 4.0 e5; maximum injection time 100 ms. The most intense ions with charge state 2-6 isolated in 3 s cycles were selected in the MS scan for further fragmentation. MS2 scan parameters set: activation HCD with 30% normalized collision energy, detected at a mass resolution of 30,000. The AGC target for MS/MS was set at 5.0 e4, and the ion filling time was set to 60 ms.

Protein identification was performed by searching MS/MS data against (Thermo Fisher Sci, vs 2.2.) pipeline. Sequest HT was set up to search the NCBI database (selected for Homo Sapiens, 2019_06, 20431 entries), assuming the digestion enzyme trypsin. The parameters for Sequest HT were set as follows: Enzyme: trypsin, Max missed cleavage: 2, Precursor mass tolerance: 10 ppm, Peptide tolerance: ± 0.02 Da, Fixed modifications: carbamidomethyl (C); Dynamic modifications: oxidation (M).

Gene ontology (GO) was performed using DAVID Bioinformatics Resources (v2023, https://david.ncifcrf.gov/). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed with DAVID. The GO terms and KEGG pathways with p < 0.05 were considered statistically significant. The STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) algorithm was used to build protein-protein interaction networks. Only high-confidence interactions (score ≥ 0.7) were chosen.

Immunofluorescence

Cells were fixed with 4% paraformaldehyde in PBS for 30 min, followed by permeabilization in PBS containing 0.5% (v/v) Triton X-100 (Thermo Fisher Scientific, Waltham, MA, USA) for 10 min and blocking in PBS containing 5% bovine serum albumin (BSA) in TBST for 60 min. The samples were incubated overnight in TBST containing primary antibodies (dilution 1:200) and 5% BSA at 4°C and then exposed to fluorescent dye-conjugated secondary antibodies (dilution 1:400) for 60 min at RT. All confocal microscope images were generated using a Zeiss LSM 800 with Airyscan (ZEISS) with a 63 × oil objective (supported by NIH P30 GM106397). The confocal settings used for image capture were held constant when samples were being compared. Images were quantified and analyzed using the Image J software (National Institutes of Health).

Raw Zeiss image files were opened in ImageJ using the Bio-Formats plugin. Images were adjusted for brightness and contrast using the Auto function. ALIX (Pseudo-colored green) was merged with MUC1-CT, EEA1, Rab5, RAB7, Syntenin-1, or LAMP1 (Pseudo-colored red) to generate composite images. The composites were converted from 8-bit to RGB format. Colocalization analysis was performed using the Color Threshold function: the hue range was set to 24–55 to define colocalized areas and 0–255 to define the total signal area. The size of the thresholded regions were then measured for quantification.

Cloning and Constructs

Previously, our lab constructed a 42 tandem repeat MUC1 (MUC1-42TR) in pBlueScriptII KS (+/−) (pBS) plasmid vector (Agilent plasmid GenBank^®#X52327 [KS(+)] & #X5329 [KS(−)]) containing a DYKDD FLAG tag at the N terminal of MUC1(24). cDNAs for human MUC1ΔTR and human MUC1-42TR were cloned into the pAquaN1 vector (Addgene #42888, Watertown, MA, USA). For the generation of MUC1ΔTR pAquaN1, the cDNA for MUC1ΔTR was amplified with primers: 5’-TCTCGAGCTCAAGCTTCGCTCCACCTCTCAAGCAG-3’ (forward) and 5’-GTCGACTGCAGAATTCGAAGTTGGCAGAAGCGGCTG-3’ (reverse) from the MUC1ΔTR pBS. The PCR product was cloned into the pAquaN1 vector using the HindIII and EcoRI restriction sites. There is an interrupting stop code in the C terminal of MUC1-42TR pBS. To work around this stop codon, we took a small C terminal insert region from the MUC1-TR pAquaN1 excluding the stop codon and cloned this into the MscI and ApaI restriction sites of MUC1-42TR pBS. We then took the MUC1-42TR insert from this MUC1-42TR pBS without the stop codon and used SacI and EcoRI restriction sites to ligate the insert into the pAquaN1 vector. The MUC1-42TR and MUC1ΔTR cDNA were isolated from the MUC1-42TR pBS and MUC1ΔTR pBS plasmids with BamH1 and ligated into the pLVX-EF1α-IRES-ZsGreen1 vector.

The cDNA for ALIX was amplified with primers: 5’-GTCGACGGTACCGCGGGCCCTCCGGACTCAGATCTCGAGC-3’ (forward) and 5’-TAGATCCGGTGGATCCCTAGATCCGGTGGATCCTTACTGG-3’ (reverse) from mCherry-hALIX plasmid (Addgene #21504, Watertown, MA, USA). The PCR product was cloned into the pSYFP2-C1 vector (Addgene #22878, Watertown, MA, USA) using the Apal and BamHI restriction sites.

cDNA for ALIX, ALIXΔPRD, and Bro1 were cloned into the pSELECT-NHA-blasti (InvivoGen, San Diego, CA, USA). The cDNA of ALIX, ALIXΔPRD, and Bro1 was amplified with primers: 5’-AGACTATGCCGGATCCTATGGCGACATTCATCTCGGTGC-3’ (ALIX-BamH1-Forward), and 5’-GCCAGCTAGCCCATGGTTACTGGCTGTGGATAGTAAGACTG-3’ (ALIX-Nco1-Reverse), and 5’-GCCAGCTAGCCCATGGCTTAAACCATCTTCTCAAACAGATC-3’ (ALIXΔPRD-Nco1-Reverse), and 5’-GCCAGCTAGCCCATGGTTACTGGCTGTGGATAGTAAGACTG-3’ (Bro1-Nco1-Reverse) from mCherry-hALIX plasmid (Addgene #21504, Watertown, MA, USA). The PCR product was cloned into pSELECT-NHA-blasti vector using the Apal, BamH1 and Nco1 restriction sites.

FRET

HEK293 cells were seeded into 12-well TC-Plates (Sarstedt AG & Co. KG, Nümbrecht, Germany) at 1.75 × 10⁵ cells per well and 3.5x10⁵ cells per well, respectively, and allowed to adhere to the wells for at least 8 hours but no more than 24 hours. Each cell line was then transiently transfected at 70-90% confluence with Positive Control for FRET pmVenus(L68V)-mTurquoise2 (Addgene #60493, Watertown, MA, USA) and Negative Control for FRET pmTurquoise2-T2A-Venus(L68V) (Addgene #60494, Watertown, MA, USA), along with MUC1-42-TR pAquaN1 and pSYFP2-ALIX plasmids following the Lipofectamine^™ 3000 reagent protocol (Invitrogen life technologies, Carlsbad, CA, USA). No more than 20 hours following transfection, Lipomedia was replaced with DMEM + 10% FBS + 1% P/S. 24 hours post-transfection. The images were acquired 48 hours after transfection.

Cells were fixed in 4% formaldehyde in PBS (Thermo Fisher, Waltham, MA, USA) for 10 min at room temperature. Images were acquired on Zeiss LSM 710 Confocal Microscope (supported by NIH S10 RR027301) using channel-based non-spectral acceptor photobleaching Fluorescence Resonance Energy Transfer (FRET). The system was configured in channel mode with a sequential scan line, using 40 × objective and 512 × 512 at speed 6. The light path for the CFP donor was set with an excitation of 405nm with 3% laser line attenuator transmission and a master gain of 833.8. The YFP acceptor was set with excitation at 514 nm with an MBS 458/514 combo and a master gain of 1042.7. For the non-linked control samples, the individual lasers were switched on and off to verify no crosstalk between channels during collection. The imaging area and bleaching parameters were set with an area of interest within a limited number of cells/targets. Bleaching was set to collect 1 image prior to bleaching, with 30 iterations, safe GaAsP, and 100% laser at 514 nm to target the YFP. The time series was set as 4 cycles with an interval of 1 millisecond. Regions of interest were drawn for the bleaching areas with no acquisition to visualize the full area. The images were obtained at an acquisition scan speed of 6, showing CFP, YFP, and FRET channels. The confocal settings used for image capture were held constant throughout the experiment. Images were quantified and analyzed using Zeiss software analysis as well as ImageJ software (National Institutes of Health).

Co-immunoprecipitation

To generate MUC1 stable expressed cells, MUC1-42TR (MUC1-FL) or MUC1ΔTR cDNA were cloned into the pLVX-EF1α-IRES-ZsGreen1 vector. The lentivirus was produced as previously described (http://www.addgene.org/protocols/plko/#E). HEK293T cells were infected and sorted with GFP by FACS. ALIX-pSELECT-NHA, ALIXΔPRD-pSELECT-NHA, and Bro1-pSELECT-NHA were transfected into MUC1-stably-expressing HEK293T cells as described above. The cell lysate was generated by using 1 × lysis buffer (Cell Signaling, Danvers, MA, USA) with 1 × Halt Protease and Phosphatase Inhibitor Cocktail. Then, the cell lysate was rotated with anti-HA antibody conjugated magnetic bead (Cell Signaling, Danvers, MA, USA) overnight at 4°C. The supernatant was discarded, and beads were washed 5 times with 1 × lysis buffer (Cell Signaling, Danvers, MA, USA) by vortexing. Last, the elution samples were prepared for Western Blot as described below.

HPAF-CD11 EVC, HPAF-CD11 MUC1KO, HEK293T-MUC1-FL and HEK293T-MUC1ΔTR were grown to ~90% confluence in 100 mm dishes. The cell lysate was generated using Pierce^™ IP Lysis Buffer (Thermo Fisher Scientific, Waltham, MA, USA) with 1 × Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific, Waltham, MA, USA).

For MUC1CT IP, 0.5 μg of MUC1 monoclonal antibody (Thermo Fisher, Waltham, MA, USA) or 0.5 μg Armenian Hamster IgG Isotype Control (Thermo Fisher, Waltham, MA, USA) was rotated in 100 μL with 50 μL of Dynabeads Protein G (Thermo Fisher Scientific, Waltham, MA, USA) in a binding/wash buffer (PBS with 0.02% Tween 20) at room temperature (RT) for 10 min. The supernatant was magnetically separated from the beads and discarded, and the beads with antibodies were washed with binding/wash buffer 3 times. The beads with antibodies were incubated in cell lysate on the rotator for 10 min. The supernatant was again discarded, and the beads were washed 3 times with binding/wash buffer by gentle pipetting. Then, 50 μL elution buffer (0.5 M glycine at pH 2.8) was added to the beads and mixed well. 20 μL of premixed NuPAGE^™ LDS Sample buffer was then added and heated for 10 min at 70°C. The supernatant was kept for Western Blot.

Immunoblotting

Whole-cell lysates and crude EV pellet lysates were prepared by using RIPA buffer (1% sodium deoxycholate, 0.15 M NaCl, 0.1% SDS, 1% (v/v) Triton X-100, 0.05 M Tris HCl, pH 7.4) supplemented with 1 × Halt Protease and Phosphatase Inhibitor Cocktail. Protein concentrations were determined by the Pierce BCA protein assay (Thermo Fisher Scientific, Waltham, MA, USA). Equal amounts of total protein were resolved on 4-12% NuPAGE Bis-Tris Gels (Thermo Fisher Scientific, Waltham, MA, USA) and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Burlington, MA, USA). Then, the membranes were blocked with 5% bovine serum albumin and incubated with primary antibodies (dilution 1:1000) overnight at 4°C. Following incubation with secondary antibodies (dilution 1:5000), the antigen-antibody complex was developed using enhanced chemiluminescence (ECL) Prime Western Blotting detection reagent for the HRP-conjugated secondary antibodies (Thermo Fisher Scientific, Waltham, MA, USA).

Immunoblotting data was quantified with ImageJ. Protein expression in whole-cell lysates was normalized to β-actin, whereas protein levels in extracellular vesicles were quantified by band intensity and normalized to cell number (Supplementary Fig. 1).

Migration and Invasion Assay

Transwell permeable supports (8 μm) were coated with or without 50 μL Matrigel (diluted in serum-free media) and placed at 37°C for 2 hours. Then, 50,000 S2-013 MUC1KO and Capan-2 MUC1KO cells were plated in 1% EVs-free FBS media with 3 μg/mL EVs on the top chamber of the transwell supports. Media with 10% EVs-free FBS and 3 μg/mL EVs was placed in the bottom chamber as a chemoattractant. Cells were visualized and quantified using an inverted light microscope and more than five randomly chosen fields of view.

Protein expression

The genetic sequence of the cytoplasmic tail (CT) of MUC1 was tagged with a 10x-His-tag and an AviTag separated by thrombin and HRV3C proteases, respectively, cloned into pET28a vector, and ordered from GenScript with codon-optimization at the CQC motif (modified to SQS) for expression in E. coli. To ensure site-specific protein biotinylation, HisAvi_MUC1CT was co-expressed with a BirA expression plasmid in Tuner^™ (DE3) cells (Thermo Fisher Scientific) and grown in Luria-Bertani (LB) media (Fisher Bioreagents) under kanamycin (30 μg/mL) and chloramphenicol (34 μg/mL) selection. Biotin (10 mg/L) was added at the time of inoculation to support BirA-mediated biotinylation. Cells were incubated at 37°C with shaking at 145 rpm until an OD₆₀₀ of 0.7-0.75 was reached. Once the desired OD₆₀₀ was reached, the cell cultures were cooled for 1 hour at 4°C to improve subsequent solubility, then induced with a final concentration of 0.4 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) (Gold Biotechnology) followed by a post-induction incubation period under optimized conditions at 37°C, 145 rpm shaking for 4 hours. Initial expression trials at a lower temperature of 18°C resulted in poor induction and low yield. Post-cooling induction at 37°C provided the optimal combination of solubility and expression levels. Cells were harvested at OD₆₀₀ of 1.6-1.8, pelleted by centrifugation at 14,000 × g for 30 min, and stored at −20°C in 5 g aliquots.

The genetic sequence of the Bro1 and V domains of ALIX was fused to a 10x-His-tag with HRV3C protease, cloned into pET28a vector, and ordered from GenScript with codon-optimization at the KK motif (modified to YY(25)) for expression in E. coli. The plasmid was expressed in one-shot BL21(DE3) cells (Thermofisher) and grown in LB media (Fisher Bioreagents) under kanamycin (30 μg/mL) selection. Cells were incubated at 37°C, 145 rpm shaking until an OD₆₀₀ of 0.75–0.8 was reached. Once the desired OD₆₀₀ was reached, the cells were cooled for 1 hour at 4°C, then induced with a final concentration of 0.4 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) (Gold Biotechnology) at 18°C, 145 rpm shaking overnight (~18 hours). Cells were harvested at OD₆₀₀ of 2.4-2.6, pelleted by centrifugation at 14,000 × g for 30 min, and stored at −20°C in 5 g aliquots.

Protein Purification

Frozen cell pellets expressing HisAvi_MUC1CT were thawed on ice and resuspended at 1 g of pellet per 10 mL of resuspension buffer supplemented with 50 μL of His protease inhibitor cocktail (Sigma, Cat# P8849) per 10 mL of buffer. This buffer was optimized based on our prior solubility screening experiments to improve protein yield and minimize aggregation. The final resuspension buffer consisted of 50 mM CAPS pH 10.0, 200 mM KCl, 100 mM (NH₄)₃PO₄, 1 M Urea, 10 mM Imidazole, and 2 mM 2-Mercaptoethanol (2-Me) reducing agent. The resuspended cell pellet was then lysed by five passes through Emulsiflex C3, followed by lysate clarification via centrifugation at 40,000 × g for 30 min at 4°C. The clarified lysate containing soluble protein was then filtered through a 0.45 μM filter (Millipore) and recentrifuged at 10,000 × g for 10 min at 4°C. The filtered clarified lysate was then passed through a 5 mL HisTrap FF column (Cytiva) using AKTA FPLC (Amersham Biosciences, UK). Proteins nonspecifically bound to the column were washed using 20 column volumes (CV) of resuspension buffer. HisAvi_MUC1CT was then eluted over 20 CV by a 50% gradient method, using an elution buffer containing 20 mM CAPS pH 10, 200 mM KCl, 1 M Imidazole, and 2 mM 2-Me. Fractions yielding protein were observed by A₂₈₀ peaks on the AKTA FPLC (Amersham Biosciences, UK) chromatogram and verified using a 4-20% Bis-Tris gel (GenScript) stained with Coomassie Brilliant Blue G-250. Peaks containing HisAvi_MUC1CT were then pooled and diluted using a QFF dilution buffer containing 25 mM TAPS pH 8.5 and 2 mM 2-Me until a <20 mM KCl concentration was achieved. The diluted HisAvi_MUC1CT sample was then passed through a 1 mL HiTrap QFF column (Cytiva) using AKTA FPLC (Amersham Biosciences, UK). Proteins nonspecifically bound to the column were washed using 20 CV of QFF buffer A containing 25 mM TAPS pH 8.5, 10 mM KCl, and 2 mM 2-Me. HisAvi_MUC1CT was then eluted over 20 CV by a 100% gradient method, using an elution buffer containing 25 mM TAPS pH 8.5, 1 M KCl, and 2 mM 2-Me. Fractions yielding protein were observed by A₂₈₀ peaks on the AKTA FPLC (Amersham Biosciences, UK) chromatogram and verified using a 4-20% Bis-Tris gel (GenScript) stained with Coomassie Brilliant Blue G-250. Purified HisAvi_MUC1CT fractions were pooled, and the concentrations were determined using a Fisher NanoDrop1000 at 1 mg/mL protein absorbance. Purified HisAvi_MUC1CT was imaged on a 4-20% Bis-Tris gel (GenScript) stained with Coomassie Brilliant Blue G-250 and stored at −20°C for use in further analysis.

Each gram of His_ALIXBro1-V cell pellet was thawed and resuspended in 5 mL of a resuspension buffer containing 50 mM Tris pH 8.0, 150 mM NaCl, 10 mM Imidazole, and 2mM 2-Me and 50 μL of His protease inhibitor cocktail (Sigma, Cat# P8849). The resuspended cell pellet was then lysed by five passes through Emulsiflex C3, followed by lysate clarification via centrifugation at 40,000 × g for 30 min at 4°C. The clarified lysate containing soluble protein was then filtered through a 0.45 μM filter (Millipore) and recentrifuged at 10,000 × g for 10 min at 4°C. The filtered clarified lysate was then passed through a 5 mL HisTrap FF column (Cytiva) using AKTA FPLC (Amersham Biosciences, UK). Proteins nonspecifically bound to the column were washed using 20 CVs of resuspension buffer. His_ALIXBro1-V was then eluted over 10 CV by a 50% gradient method, using an elution buffer containing 25 mM Tris pH 8.0, 150 mM NaCl, 1 M Imidazole, and 2 mM 2-Me. Fractions yielding protein were observed by A₂₈₀ peaks on the AKTA FPLC (Amersham Biosciences, UK) chromatogram and verified using a 4-20% Bis-Tris gel (GenScript) stained with Coomassie Brilliant Blue G-250. Peaks containing His_ALIXBro1-V were then pooled and diluted using a QFF dilution buffer containing 25 mM Tris pH 8.0 and 2 mM 2-Me until <20 mM NaCl concentration was achieved. The diluted His_ALIXBro1-V sample was then passed through a 1 mL HiTrap QFF column (Cytiva) using AKTA FPLC (Amersham Biosciences, UK). Proteins nonspecifically bound to the column were washed using 20 CV of QFF buffer A containing 25 mM Tris pH 8.0, 15 mM NaCl, and 2 mM 2-Me. His_ALIXBro1-V was then eluted over 30 CV by a 100% gradient method, using an elution buffer containing 25 mM Tris pH 8.0, 1 M NaCl, and 2 mM 2-Me. Fractions yielding protein were observed by A₂₈₀ peaks on the AKTA FPLC (Amersham Biosciences, UK) chromatogram and verified using a 4-20% Bis-Tris gel (GenScript) stained with Coomassie Brilliant Blue G-250. Purified His_ALIXBro1-V fractions were pooled, and the concentrations were determined using a Fisher NanoDrop1000 at 1 mg/mL protein absorbance. Purified His_ALIXBro1-V was imaged on a 4-20% Bis-Tris gel (GenScript) stained with Coomassie Brilliant Blue G-250 and stored at 4°C for use in further analysis.

Localized surface plasmon resonance (OpenSPR) based MUC1CT binding assay

SPR experiments were performed in three replicates (n = 3) using an OpenSPR Rev4 (Nicoya) equipped with a research-grade biotin sensor chip that was bound to streptavidin using the ligand wizard Streptavidin Sensor Kit. Purified biotinylated HisAvi_MUC1CT was immobilized through the interaction between biotin on the HisAvi_MUC1CT and streptavidin on the sensor chip by injecting 35 μg/mL HisAvi_MUC1CT at a flow rate of 20 μL/min over channel 2. Channel 1 was left blank and served as a reference control. To collect kinetic binding data, His_ALIXBro1-V in 50 mM TAPS, pH 8.5, 500 mM KCl, 1% BSA was injected over the two flow cells at concentrations of 1.00 μM, 1.50 μM, 2.25 μM, 3.35 μM, and 5.00 μM at a flow rate of 30 μL/min and a temperature of 20°C. The complex was allowed to associate and dissociate for 345 seconds. The surfaces were regenerated with 1-2 injections of Glycine-HCl pH 2.5 at a flow rate of 150 μL/min. Binding kinetics data of each of the three replicates were collected on a 1:1 interaction model using the TraceDrawer software (Nicoya).

Statistical Analysis

Data was analyzed with GraphPad Prism (Version 8.00) using tests that are described in the Figure Legends. Data were generally presented as mean values ± SD of at least three independent experiments, unless otherwise indicated. Generally, and unless otherwise indicated, an unpaired Student T-test was used to compare between two groups; p-value < 0.05 was considered statistically significant.

Results

MUC1 is enriched in tumor-derived and pancreatic cancer cell line-derived extracellular vesicles.

MUC1, a type I transmembrane protein generally expressed on the surface of ductal epithelia, is overexpressed in pancreatic and other adenocarcinomas and promotes rapid growth and metastatic properties(15). Here, we demonstrate significantly higher levels of MUC1 in extracellular vesicles (EVs) isolated from primary pancreatic tumors and metastatic sites as compared to adjacent normal pancreas tissue (Fig. 1A). The N-terminal domain of MUC1 contains a variable number of mucin-type tandem repeats (VNTR) and an SEA (sea urchin sperm protein enterokinase and agrin) domain. The latter mediates an autoproteolytic cleavage event that results in two protein fragments that remain associated during further post-translational processing. The C-terminal domain of MUC1 contains a short extracellular domain (ECD), a transmembrane domain (TM), and a cytoplasmic tail (CT)(15). To further investigate MUC1 in PDAC tumor-derived EVs, we analyzed expression of the N-terminal and C-terminal domains using specific antibodies mAb-AR20.5 (MUC1 VNTR) and mAb-O14E (MUC1 CT)(21). Nanoscale Fluorescence Analysis and Cytometric Sorting (nanoFACS) analysis of tumor-derived EVs purified from different PDAC patient samples from UNMC’s Rapid Autopsy program revealed that very few exosomes expressed only the tandem repeat (4-7%); however, substantial populations (10-40%) of exosomes showed expression of tandem repeat and CT (double positive); and another population showed expression of only cytoplasmic tail (25-55%). We consistently observed that the cytoplasmic tail was expressed in greater percentages than the VNTR of MUC1 (Fig. 1B–1E and Supplementary Table 1).

Figure 1. MUC1 is found in EVs derived from pancreatic cancer cells and up-regulates EV secretion.

(A) Western blot of EV-related proteins and MUC1 in an untreated pancreatic cancer patient (RAP141) showing primary tumor and metastasis tumor-derived EVs; total protein mass was used as the loading control. Pan Head TT: primary tumor tissue in the head of pancreas, Pan tail TT: primary tumor tissue in the tail of pancreas, Liver Met-1: metastasis lesion in liver section A, Liver Met-2: metastasis lesion in liver section D, Liver Met-3: metastasis lesion in liver section F, Lung Met-1: metastasis lesion in left upper lung section nn, Lung Met-2: metastasis lesion in left lower lung, Lung Met-3: metastasis lesion in right lower lung section ii, Unaffected Pan: normal pancreas tissue, LN Met: lymph node metastasis, Diaphragm Met: diaphragm metastasis, Muscle: normal pectoral muscle tissue. (B - E) Nanoscale fluorescence analysis using anti-human CD63, anti-CD81, anti-O14E, and anti-AR20.5 of the EVs from pancreatic cancer patients’ primary tumor-derived EVs, RAP 137 Pan Head (B): primary tumor-derived EVs from RAP 137 pancreatic head region; RAP 140 Pan Body (C): primary tumor-derived EVs from RAP 140 pancreatic body region; RAP 141 Pan Head (D): primary tumor-derived EVs from RAP 141 pancreatic head region; RAP 142 Pan Body (E): primary tumor-derived EVs from RAP 142 pancreatic body region. Western blot of EV-related proteins and MUC1 in whole-cell lysate and EVs of S2-013 cells (F, Supplementary Fig. 1A) and HPAF-CD11 cells (G, Supplementary Fig. 1B). The quantification data are shown in Supplementary Fig. 1A and 1B. Total protein mass was used as loading control in the whole cell lysate, and EV samples were normalized by cell numbers. S2-013 NEO cells are transfected with empty vector only and S2-013 MUC1F cells are MUC1 overexpressed. HPAF-CD11 empty vector control (EVC) cell line was used as controls for MUC1 knockout cells. (H - J) Nanoparticle-tracking analysis of EVs isolated from pancreatic cancer cells; n ≥ 3, biological replicates. Representative TEM images of HPAF-CD11 EVC-derived EVs (K) and HPAF CD11 MUC1KO-1-derived EVs (L). In H and I, data is presented as mean; S2-013 MUC1F EV concentration: 7.47e+010 +/− 2.10e+009 particles/mL; S2-013 MUC1KO EV concentration: 3.66e+010 +/− 6.73e+008 particles/mL; HPAF CD11 EVC EV concentration: 2.61e+010 +/− 3.48e+008 particles/mL; HPAF CD11 MUC1KO EV concentration: 9.59e+009 +/− 1.01e+008 particles/mL. In J, the number of EV per cell was calculated with EV concentration and cell number when the EVs were harvest from the cell supernatant after 44 hours, data is presented as mean ± SD, statistical significance was determined using Student T test (***p < 0.001)

The presence of MUC1CT in EVs isolated from S2-013 MUC1F, HPAF-CD11, and Capan2 cell lines was confirmed using bead-assisted flow cytometry. Calnexin(26), an endoplasmic reticulum marker, served as a negative control. We found high expression of MUC1CT (Supplementary Fig. 2A and 2B) and no expression of Calnexin in the EVs (Supplementary Fig. 2C). Analysis of exosomes for the EV markers CD63 and ALIX and revealed that a high fraction of captured CD63- or ALIX-positive EVs co-expressed MUC1CT (Supplementary Fig. 2A and 2B).

MUC1 knockout by CRISPR-Cas9 in S2-013 (Fig. 1F) and HPAF-CD11 (Fig. 1G) cells resulted in the downregulation of key signature proteins associated with EVs, including ALIX and Tsg101 (ESCRT components) and Annexin A1, while CD9 levels remained constant. Consistent with these results, NanoSight nanoparticle tracking analysis revealed a significant decrease in EV secretion per cell in MUC1 knockout cells compared to MUC1 expressing cells. MUC1-positive cells secreted an average of 278 (S2-013 MUC1F) and 279 (HPAF-CD11 EVC) EVs, while corresponding isogenic MUC1 knockout cells secreted 206 (S2-013 MUC1KO) and 159 (HPAF-CD11 MUC1KO) (Fig. 1H–J). Nanoparticle tracking analysis validated an appropriate size range (30 - 200 nm) for EVs isolated from pancreatic cancer cell lines (Fig. 1H and 1I). Transmission electron microscopy (TEM) showed EVs from the S2-013 MUC1F and the HPAF-CD11 cell line had the predicted heterogeneous populations of small, membrane-surrounded structures (Fig. 1K and 1I, Supplementary Fig. 2D). Our results suggest that MUC1 plays a critical role in regulating EV production and content, and that MUC1CT is associated with unique forms of pancreatic tumor-derived EVs.

MUC1-positive EVs increase pancreatic cancer cell growth and metastasis in vivo and in vitro.

Given prior publications showing that exosomes affect tumor growth and metastatic properties of pancreatic cancer cells(27), we sought to evaluate the effects of MUC1-associated EVs on tumor growth properties. EVs from MUC1-expressing and MUC1 knockout cells were used to evaluate effects on tumor growth and metastasis in orthotopic xenograft models in vivo. A pre-treatment education experiment was employed in which 6–8-week-old nude mice were injected with 5 μg of EVs (MUC1-positive cell or MUC1-negative cell-derived EVs) by tail vein every other day for three weeks. Subsequently, we performed orthoptic implantation of MUC1 knockout pancreatic cancer cells (Fig. 2A). Animals pretreated with EVs from S2-013 MUC1F and HPAF-CD11 cells developed tumors earlier that grew faster to larger volumes and showed higher incidences of metastasis as compared to animals pretreated with EVs from S2-013 MUC1KO cells and HPAF-CD11 MUC1KO cells (Fig. 2B–2E, Supplementary Table 2). To validate these findings, we evaluated the effects of EVs on organoid growth. We treated the PDX-derived organoids (PA137 and PA717) with 3 μg/mL MUC1 positive EVs or MUC1 negative EVs every other day for 6 days, MUC1 positive EVs increased human PDAC organoid growth compared to MUC1 negative EVs (Fig. 2F).

Figure 2. MUC1-positive EVs increase tumor progression in vivo.

(A) The schematic workflow used for MUC1-positive EVs and MUC1-negative EVs pre-treatment in the orthotopic mouse model. Tumor volume (B) and tumor weight (C) from the mice pretreated with PBS, S2-013 MUC1F-derived EVs, and S2-013 MUC1KO-derived EVs, then the S2-013 MUC1F cells or S2-013 MUC1KO cells were orthotopically transplanted to mouse pancreases. Data representative of at least three biological replicates. Tumor volume (D) and tumor weight (E) from the mice pretreated with PBS, HPAF-CD11 EVC-derived EVs, and HPAF-CD11 MUC1KO-derived EVs, then the HPAF CD11 EVC cells or HPAF CD11 MUC1KO cells were orthotopically transplanted to mouse pancreases. (F and G) The organoid PA137 and KM5 were seeded with 1000 per well and then incubated with or without 3 μg/mL EVs isolated from S2-013 MUC1F and S2-013 MUC1KO or HPAF CD11 EVC and HPAF MUC1KO for 6 days, and additional EVs were added every other day; the cellTiter-Glo assay was applied to investigate the cell viability. Data representative of at least three biological replicates in each group. Note that the tumor volume were generated with high resolution ultrasound and the scale bars were burned into images. Data is presented as mean ± SD, significance analysis by unpaired Student’s T-test (*p < 0.05, **p < 0.005).

Additionally, we explored the effects of EVs derived from MUC1-positive cells and MUC1 knockout cells on cell proliferation, migration, and invasion in vitro. Dose-dependent titrations of 1-9 μg/mL EVs on S2-013 MUC1KO or Capan2 MUC1KO cells for 48 h revealed that treatment with MUC1-positive EVs significantly increased cell proliferation compared to MUC1-negative EVs. We found that 1 μg/mL S2-013 MUC1F-derived EVs and 3 μg/mL HPAF-CD11 derived EVs significantly increased S2-013 MUC1KO or Capan2 MUC1KO cell proliferation as compared to MUC1-negative EVs (Fig. 3A and 3B, Supplementary Fig. 3A and 3B). Treatment of S2-013 MUC1KO or Capan2 MUC1KO cells with 3 μg/mL of MUC1-positive EVs for 24 h, 48 h, and 72 h (Fig. 3C and 3D, Supplementary Fig. 3C and 3D) showed increasing and statistically significant effects on cell division (increased numbers of cells) at 48 and 72 hours, as compared to MUC1-negative controls. MUC1-positive EVs also enhanced migration and invasion of S2-013 MUC1KO (Fig. 3E–3F and Supplementary Fig. 3E–3F) and Capan2 MUC1KO (Fig. 3G–3H, Supplementary Fig. 3G–3H) cells.

Figure 3. MUC1-positive EVs increase pancreatic cancer cell growth, migration, and invasion.

S2-013 MUC1KO cell numbers after 48 h incubation with or without increasing amounts of extracellular vesicles from S2-013 MUC1F and S2-013 MUC1KO (A) and HPAF-CD11 EVC and HPAF-CD11 MUC1KO (B). S2-013 MUC1KO cell numbers after incubation for 24 h, 48 h, and 72 h with or without 3 μg/mL extracellular vesicles from S2-013 MUC1F and S2-013 MUC1KO (C) and HPAF-CD11 EVC and HPAF-CD11 MUC1KO (D). Images and quantification of S2-013 MUC1KO cells incubated with or without 3 μg/mL extracellular vesicles from S2-013 MUC1F and S2-013 MUC1KO (E - F) and HPAF-CD11 EVC and HPAF-CD11 MUC1KO (G - H) for 48 h and then seeded into transwells with or without Matrigel for another 24 hours. The images were acquired randomly with 20X microscope objective (Leica), and the scale bars (233.9 μm) were burned into images at the left bottom. Data representative of at least three biological replicates. Data is presented as mean ± SD, significance analysis by unpaired Student’s T-test between control group or MUC1 negative EVs treated group (*p < 0.05, **p < 0.005, ***p < 0.001)

MUC1-positive EVs contain protein cargos that differ from MUC1-negative EVs.

Given the differences in biological effects of MUC1-positive EVs and MUC1-negative EVs and preliminary evidence of differential protein expression between these, we investigated the specific protein content of MUC1-positive and MUC1-negative EVs by Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) proteomics. In total, we identified 259 proteins in MUC1-positive cell-derived EVs and 313 proteins in MUC1-negative cell-derived EVs. There were 64 unique proteins in MUC1-positive cell-derived EVs and 118 unique proteins in MUC1-depleted cell-derived EVs (Fig. 4A and Supplementary Fig. 4A, displayed by jvenn(28)). Putative interactions and implied biological networks were determined for the 64 unique proteins in MUC1-positive cell-derived EVs and 118 unique proteins in MUC1-depleted cell-derived EVs by using the STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) algorithm to build protein and protein interaction networks (Score ≥ 0.07)(29) and the DAVID Bioinformatics resources to perform Kyoto encyclopedia of genes and genomes (KEGG) pathway analysis(30, 31). Protein-protein interaction networks are shown for MUC1-positive cell-derived EVs (Fig. 4B), MUC1-depleted cell-derived EVs (Supplementary Fig. 4B), and both EV types (Supplementary Fig. 4C). KEGG pathway analysis showed that the 64 unique proteins in MUC1-positive cell-derived EVs were associated with 18 pathways (p < 0.05) and the most enriched pathway was endocytosis (Fig. 4C and Supplementary Table 3). There were 10 pathways revealed by KEGG pathway analysis of 118 differential proteins in MUC1-depleted cell-derived EVs, with the most enriched pathway being ribosome components (Supplementary Fig. 4D and Supplementary Table 4). There were 15 pathways revealed by KEGG pathway analysis of 118 differential proteins in MUC1 depleted cell-derived EVs, with the most enriched pathway being ECM-receptor interaction (Supplementary Fig. 4E). Selected differentially expressed proteins (Src, ALDOC, Annexin A3) in MUC1-positive cell-derived EVs were validated by immunoblotting (Fig. 4D). Annexin A3 has previously been reported to be highly enriched in tumor tissue-derived EVs and particles of pancreatic cancer(5). We further evaluated Src protein in the EVs for activation (phosphorylation) status and found the presence of activated Src that was phosphorylated at Tyr 416 and dephosphorylated at Tyr 527 (Figure 4D), demonstrating that activated Src was packaged into MUC1-positive EVs. Given that numerous prior studies have documented that MUC1 binds to Src and affects its signaling properties (32) (33) (34, 35), we hypothesized that MUC1CT served as a molecular chaperone that helps load Src into EVs.

Figure 4. Phospho_Tyr416-Src packaged into MUC1 positive EV effects cancer cell proliferation, migration and invasion.

(A) Comparison of proteins in MUC1-positive cell-derived EVs and MUC1-depleted cell-derived EVs. The EVs were isolated from S2-013 MUC1F, S2-013 MUC1KO, HPAF CD11 EVC and HPAF CD11 MUC1KO, data was analyzed with two biological replicates. (B) Protein-protein interaction network for MUC1-positive cell-derived EVs. (C) KEGG pathway analysis of differentially presented proteins in MUC1-positive cell-derived EVs. (D, Supplementary Fig. 1C) Western blot of MUC1-associated EVs proteins in MUC1-positive and MUC1 depleted whole-cell lysate and EVs; whole cell lysate and EV samples were normalized by total protein concentration. The quantification data are shown in Supplementary Fig. 1C. (E) Schematic experiment plan of Src inhibitor treated-EVs. Cells were treated with 200 nM bosutinib for 24 hours, followed by incubation in EV-depleted medium containing 200 nM bosutinib for an additional 48 hours. Extracellular vesicles were then isolated from the conditioned medium. The cells were treated with 5 μg/mL EVs from HPAF-CD11 cells with or without Src inhibitor (200 nM bosutinib) treatment for 72 hours. They were supplemented with additional EVs every 24 hours. (F, Supplementary Fig. 1D) Western blot of Src-related proteins with Src inhibitor treatment in whole-cell lysate and EVs; total protein mass was used as loading control in the whole cell lysate, and EV samples were normalized by cell numbers. The quantification data are shown in Supplementary Fig. 1D. (G) The cells were treated with 5 μg/mL EVs from HPAF-CD11 cells with or without Src inhibitor (200 nM bosutinib) treatment for 72 hours. They were supplemented with additional EVs every 24 hours. CellTiter-Glo cell viability assay was applied. (H - I) The S2-013 MUC1KO cells were treated with 5 μg/mL EVs from HPAF-CD11 cells with or without Src inhibitor treatment for 48 hours. The cells were then seeded into transwells with or without Matrigel for another 24 hours and incubated with extracellular vesicles. The images were acquired randomly with 20X microscope objective (Nikon), and the scale bars (100 μm) were burned into images at the right bottom. In G and I, data points depict individual biological replicates (G) or visual fields (I). Data is presented as mean ± SD, significance analysis by unpaired Student’s T-test (***p < 0.001).

PhosphoTyr416-Src in MUC1-positive EVs plays an important role in pancreatic cancer cell proliferation, migration, and invasion in vitro.

We investigated the functional significance of activated Src in MUC1-positive EVs by inhibiting activation of Src using Bosutinib to reduce activated Src in tumor cells and isolated EVs. S2-013 MUC1F and HPAF-CD11 cells were treated with or without 200 nM Bosutinib for 24 hours; media was changed to media with EV-free FBS with or without 200 nM Bosutinib for approximately 48 hours preceding EV isolation. MUC1 negative cells were then treated for 72 hours with 5 μg/mL EVs isolated from cells with or without Src inhibitor and supplemented with additional EVs every 24 hours (Fig. 4E). Cellular levels of phosphorylated (Tyr416) Src were reduced in HPAF-CD11-EVs after Bosutinib treatment (Fig. 4F) and loading of phospho-Src (Tyr416) into EVs was significantly reduced in Bosutinib-treated HPAF-CD11-EVs (Fig. 4F). EVs from HPAF-CD11-treated with Bosutinib showed reduced effects on cells viability, migration, and invasion of MUC1 knockout cells as compared to untreated HPAF-CD11 derived EVs (Fig 4G, H, I). These data suggest that phospho_Tyr416-Src in MUC1-positive EVs is functional and may contribute to cancer cell proliferation, migration, and invasion. Src regulates the loading of β1 integrin, fibronectin, and EGFR into syntenin-positive EVs (36), cargoes that are critical for cell migration and consistent with the biological effects observed using Src inhibitor–treated EVs. These data indicate that MUC1-positive EVs contain active Src and that additional Src-regulated cargoes contribute to pancreatic cancer progression. Src inhibitors such as Bosutinib also alter EV cargo content by disrupting signaling pathways, including syndecan internalization and syntenin recruitment (36), so these effects could reflect downstream changes in addition to direct activities of Src.

MUC1 influences protein loading during extracellular vesicle biogenesis.

The function of MUC1 during membrane trafficking has not been fully clarified though it is known that MUC1 traffics through the Golgi to the cell surface(37) and can be internalized through endocytic pathways(38). To identify molecules associated with MUC1 during intracellular trafficking, we characterized MUC1-associated protein complexes by co-immunoprecipitation (co-IP) followed by shotgun mass spectrometry. MUC1-associated proteins from S2-013 MUC1F cells were subjected to one-step affinity purification using the MUC1-CT2 antibody. Immunoblot analysis of proteins pulled down by the MUC1-CT2 antibody revealed numerous discrete protein bands (Supplementary Fig. 5A), which were further analyzed by mass spectrometry (MS) (Supplementary XL File 1). We analyzed the MS dataset using DAVID, revealing Gene Ontology (GO) and KEGG pathway terms. GO term analysis revealed that a majority of the proteins were associated with cellular components of the cytosol or extracellular exosomes (Supplementary Fig. 5B). The high number of associated cytosolic proteins identified supports a current functional understanding of MUC1’s cytoplasmic domain, namely that the cytoplasmic tail (CT) interacts with and impacts numerous cell signaling functions(39). KEGG pathway analysis confirmed a relation to endocytosis and related pathways (Supplementary Fig. 5C). We mapped the MUC1-interacting proteins to known endocytosis pathways (Supplementary Fig. 5D), which revealed that MUC1 was associated with virtually all key proteins known to mediate endocytosis, the formation of early and late endosomes and multi-vesicular bodies (Supplementary Fig. 5B). These data led us to hypothesize that MUC1 could play a role in the formation, packaging and/or secretion of EVs.

MUC1 cytoplasmic tail interacts with the multivesicular bodies (MVB)-associated protein ALIX to load phosphoTyr416-Src into EVs.

Given prior data suggesting that MUC1CT is involved in endocytic processes(40–42) and membrane deformation(43), we posited that MUC1CT could play a role in membrane budding. It was notable that ALIX (PDCD6IP), a protein involved in the budding process through the endosomal sorting complexes required for transport (ESCRT) machinery(44) or Syndecan-syntenin-ALIX signaling processes(45), was detected in MUC1-positive EVs (Supplementary Fig. 1a and 1b) and in proteins detected by MS analysis of co-immunoprecipitations. Confocal immunofluorescence microscopy showed that MUC1CT was widely co-localized with ALIX in the S2-013 MUC1F and HPAF-CD11 cell lines (Fig 5A). We next investigated the localization of ALIX in MUC1CT-positive and MUC1-depleted cells by confocal microscopy, together with several other markers of the endocytic pathway. ALIX staining presented a punctate distribution that partially colocalized with the early endosomal markers early-endosome antigen 1 (EEA1) (Fig. 5B) and RAB5 (Supplementary Fig. 6A and 6B), and the late endosome markers RAB7 (Fig. 5C) and Syntenin-1 (Fig. 5E) in MUC1-expressing cells; these colocalization events were reduced in MUC1CT-depleted cells, where ALIX showed a distinct subcellular profile. ALIX partially colocalized with the late endosome marker LAMP1 (Fig. 5D, Supplementary Fig. 6C, Supplementary Movie 1) in MUC1-expressing cells; this colocalization was enhanced in MUC1CT-depleted cells (Supplementary Movie 2) and was prominent in enlarged perinuclear late endosomes; which was distinct from the more distributed pattern of expression see in MUC1-expressing cells. These results suggest that the subcellular distribution of ALIX was different between MUC1-expressing and MUC1 depleted cells. This redistribution suggests that MUC1 depletion shifts ALIX-expressing MVBs away from RAB7-positive late endosomes toward LAMP1-positive lysosomes, consistent with increased MVB-lysosome fusion and reduced ALIX association with RAB7. Further, this may suggest that MUC1 typically restrains lysosomal fusion of ALIX-expressing MVBs, which could, in part, explain the elevated EV release seen in MUC1-positive cells. Overall, this led us to posit that an association between MUC1CT-ALIX contributes to EV biogenesis and affects the loading of EVs with specific cargo.

Figure 5. MUC1CT regulates ALIX trafficking.

(A) Representative confocal microscopy images of S2-013 MUC1F and HPAF-CD11 cells expressing MUC1CT and ALIX. MUC1CT was detected by secondary antibodies conjugated with Alexa Fluor 488, and ALIX was detected with secondary antibodies conjugated to Alexa Fluor 647. Representative confocal micrographs showing the steady-state subcellular distribution of endogenous ALIX together with that of endogenous (detected by secondary antibodies conjugated to Alexa Flour 555) EEA1 (B), RAB7 (C), LAMP1 (D), or Synteinin-1 (E), in S2-013 MUC1F and S2-013 MUC1KO cells. See insets for high magnification. ALIX colocalization with MUC1CT, EEA1, RAB7, LAMP1, and Synteinin-1 was quantified using ImageJ (bar graphs to the right of each figure). ALIX colocalization with MUC1CT, EEA1, RAB7, LAMP1, and Synteinin-1 was assessed using ImageJ. Colocalization was analyzed in ImageJ using color thresholding, and the colocalized area was normalized to total signal area. Data represent mean ± SD from ≥3 independent experiments. In the bar graph, each dot represents individual field. Note that the scale bars were 5 μm.

We therefore sought to validate and further investigate the association between MUC1CT and ALIX. Co-immunoprecipitation (co-IP) experiments in HPAF-CD11, HEK293T-MUC1FL, and HEK293T-MUC1ΔTR cell lines were evaluated by western blotting to identify binding partners. The ESCRT-associated protein ALIX was confirmed as an interaction partner that co-immunoprecipitated with MUC1CT (Fig. 6A and 6B). Structural studies of ALIX have revealed three domains that are known to interact with different proteins: the N-terminal Bro1 domain, a central V domain, and a C-terminal proline rich domain (PRD)(46). Different truncations of ALIX (Full length, Bro1-V, and Bro1 domains) were co-expressed with MUC1 full-length or MUC1ΔTR in HEK293T cells. The C terminus of MUC1 was co-immunoprecipitated with ALIX full length, Bro1-V, and Bro1 (Fig. 6C, Fig. 6D), supporting the concept that MUC1 and ALIX interact directly.

Figure 6. MUC1 cytoplasmic tail associates with the MVB-associated protein ALIX regulating phospho_Tyr416-Src sorted into EVs.

Cell lysate extracted from HPAF-CD11 cells (A), and HEK293T-MUC1FL or HEK293TMUC1ΔTR cells (B) were co-immunoprecipitated (IP) with MUC1CT mAb or isotype control antibody and immunoblotted (IB) for ALIX, Tsg101 and MUC1CT (labeled by red star). (C) and (D) Co-IP with HA and immunoblotted for HA and MUC1CT; different ALIX truncations tagged with HA at the N-terminal (ALIX, Bro-V and Bro) were expressed in HEK293T-MUC1FL (C) or HEK293TMUC1ΔTR cells (D). (E) Coomassie blue stained 4-20% Bis-Tris gel with Precision Plus Dual Color Standard showing purified recombinant HisAvi_MUC1CT (13.38 kDa) and His_AlixBro1-V (80.65 kDa). Localized surface plasmon resonance (OpenSPR) of His_ALIXBro1-V binding to immobilized HisAvi_MUC1CT. (F) FRET microscopy images of MUC1-CFP and YFP-ALIX, MUC1-CFP (CFP tagged at C terminal of MUC1) and YFP-ALIX (YFP tagged at N terminal of ALIX) in HEK293 cells. (G, Supplementary Fig. 1E) Western blot of EVs related proteins, Src-related proteins and MUC1 in whole-cell lysate and EVs pellets of different cell lines wild type cells and ALIX knockdown cells; total protein mass was used as loading control in the whole cell lysate, and EVs samples were normalized by cell numbers. The quantification data are shown in Supplementary Fig. 1E. In E, Representative OpenSPR sensorgram showing replicate 2 of His_ALIXBro1-V binding to immobilized HisAvi_MUC1CT, with fit overlays (black lines); kinetic parameters, including association rate constant (ka), dissociation rate constant (kd), and equilibrium dissociation constant (KD), for this replicate are shown. The full set of replicate kinetic values, means, standard deviations, and %chi2/Bmax for all experiments are provided in Supplementary Table 5. Note that HEK293T-MUC1FL: MUC1 with 42 tandem repeats was expressed in HEK293T cells; HEK293T-MUC1ΔTR: MUC1 with tandem repeats deleted was expressed in HEK293T cells.

To test for a direct binding interaction between MUC1 and ALIX in the absence of other cellular components, we expressed and purified recombinant forms of both proteins (Supplementary Fig. 7A). His_ALIXBro1-V was produced using a modified established protocol(25), while purification of HisAvi_MUC1CT represents a novel method developed to overcome initial expression and solubility challenges (see Methods). We conducted localized surface plasmon resonance (LSPR) by immobilizing HisAvi_MUC1CT and titrating increasing concentrations of His_ALIXBro1-V. The results showed specific binding between the C terminal domain of MUC1 and the Bro1-V domains of ALIX with a remarkable sub-nanomolar affinity of (1.35 ± 0.13) × 10⁻¹⁰ M (Fig. 6E, Supplementary Table 5). Nonspecific binding was reduced to <5% (Supplementary Fig. 7B), supporting specificity of the interaction. We further explored this direct association between MUC1 and ALIX by performing Fluorescence Resonance Energy Transfer (FRET) using two constructs: MUC1.pAquaN1 (recombinant MUC1 with CFP added to the C-terminus - MUC1-CFP) and ALIX.pSYFP2-C1 (recombinant ALIX with YFP added to the N-terminus - ALIX-YFP). Results of acceptor photobleaching FRET (Fig. 6F) showed clear evidence of a close association between MUC1 and ALIX in vivo. Together, these studies reveal a clear and direct interaction between MUC1 and ALIX.

To explore the role of ALIX in MUC1-positive EV production and packaging we used shRNA to knock down expression of ALIX in Capan2, BxPC-3, and S2-013 MUC1F cells. EVs from those cells and their parental cells were compared to cells transfected with empty vector plasmid. We found that the knockdown of ALIX reduced MUC1CT packaging into EVs (Fig. 6G, Supplementary Fig 7C). We also found that knockdown of ALIX reduced the loading of activated phospho-Src (Tyr416) into pancreatic cancer cell-derived EVs (Fig 6C). This latter finding confirms a previous report(47) that activated Src associates with ALIX through interactions between the SH3 domain of SRC and the PRD domain of ALIX. This, alongside our findings that knocking out MUC1 reduced the packaging of SRC and ALIX into EVs (Fig. 4D) and prior findings that MUC1 binds to and is phosphorylated by Src(32) (33) (34, 35) supports the hypothesis that a complex of MUC1 and ALIX influences EV biogenesis and cargo sorting and that this MUC1/ALIX complex is responsible for binding and loading functionally active pro-oncogenic proteins such as Src into EVs.

Discussion

Here, we analyzed EVs from pancreatic cancer primary tumor, metastasis lesions, and adjacent normal tissue and found that MUC1 was expressed at high levels in tumor-derived EVs but was absent in EVs derived from adjacent normal tissue. MUC1 is a transmembrane mucin expressed on the apical surface of normal epithelial cells and aberrantly expressed on all surfaces of tumor cells(14). MUC1 is overexpressed and differentially glycosylated during pancreatic cancer progression(48) and has been reported to be packaged into tumor EVs(16, 49, 50).

Proteomic analysis of EVs derived from MUC1-positive pancreatic cancer cells was compared to EVs from MUC1 knockout cells, which revealed differential enrichment of proteins involved in oncogenic signaling, chemokine signaling, carbon metabolism, purine metabolism, phagosome, and glycolysis/gluconeogenesis signaling pathways in MUC1-associated EVs. MUC1-associated EV proteins included Annexin A3, Src, and ALDOC. Notably, Annexin A3 plays a role in regulating cell proliferation and migration through multiple signaling pathways(51), and Src is an important oncogene that regulates aspects of cell motility and invasion(47, 52). ALDOC is an aldolase family member involved in metabolism and glycolysis, and aldolase has been validated as an independent clinical prognostic marker of human cancers(53). Consistent with these contents, we demonstrated that tumor-derived EVs enriched in MUC1 promote pancreatic cancer cell growth, migration, and invasion in vitro and in vivo, as compared to EVs lacking MUC1. This leads us to posit that MUC1-associated EVs induce pancreatic tumor progression through their cargo, which contains activated oncogenes and modulators of cell proliferation and migration.

MUC1 has been shown to regulate the differentiation and proliferation of cancer cells through ligand-receptor interactions and signal transduction(14) events. Phosphorylation of MUC1CT by specific receptor tyrosine kinases (RTKs), including hepatocyte growth factor receptor (Met), epidermal growth factor receptor (EGFR), and HER2, promotes their activation and downstream signaling(39, 54, 55). Those RTKs regulate cancer cell proliferation, differentiation, and survival. Importantly and germane to this paper, MUC1 has been shown to be phosphorylated by and to bind to SRC (32) (33) (34, 35). This association likely explains, at least in part, our observation of the presence of activated and biologically active Src in MUC1-associated EVs. MUC1CT is also known to exert other pro-oncogenic activities. For example, MUC1 interactions with β-catenin induce activation of the β-catenin/TCF4 target gene MYC(56, 57). The major downstream effectors of MYC include a broad range of biological functions, such as cell proliferation, differentiation, and survival(58, 59). MUC1CT has been reported to directly activate the TGF-β pathway(60), WNT/β-catenin pathway(57), and STAT3(61). Thus, the current results extend previous findings that MUC1 contributes to in vivo invasive and metastatic potential of pancreatic cancer cells(58), by showing that molecular functions and effectors of MUC1 biological activity can be packaged into EVs and transferred to other cells.

Previous studies have highlighted a role of EVs in traveling to distant metastatic sites where they contribute to forming a premetastatic niche of PDACs(62, 63). Hoshino et al. analyzed all proteins in pancreatic tumor tissue (TT)-derived EVPs (PaCa EVPs) and adjacent normal tissue-derived EVPs. They found that proteins associated with epithelial-mesenchymal transformation, coagulation, and actin signaling were upregulated in PaCa EVPs(64). Tumor-derived exosomes polarize macrophages to an immunosuppressive phenotype by activating the NF-kB pathway to induce PD-L1 expression, which is dependent on glycolytic-dominant metabolic reprogramming(65, 66). Those studies demonstrated that EVs promote metastasis by acting directly on cancer cells and educating the microenvironment, including immune cells, to form a pre-metastatic niche. The results presented here support those findings and extend them by showing that pretreatment of animals with MUC1⁺ EVs significantly enhances tumor growth and metastasis in orthotopic models, further demonstrating that the differential packaging of molecular components into EVs is biologically significant.

Mechanistically, we present evidence that MUC1 influences the formation and packaging of EVs by associating with ALIX, a protein that has been shown to modulate exosome formation by recruiting ESCRT proteins to endosomes, which in turn drives membrane deformation and fission to produce intraluminal vesicles that may become exosomes(45, 46, 67, 68). ALIX is ordinarily present in a closed conformation due to an intramolecular interaction that renders ALIX inactive for binding. Structural studies of ALIX show that ALIX has an N-terminal banana-shaped Bro1 domain, a middle V-shaped domain, and an intrinsically disordered C-terminal proline-rich domain (PRD)(46). We demonstrated that the localization of ALIX is influenced by MUC1 and that MUC1 increases the amount of ALIX in EVs. We show that MUC1CT associated with ALIX in vivo by FRET analysis and we observed a remarkable affinity (KD of 10⁻¹⁰ M) of MUC1CT for ALIX Bro1-V domains in vitro. We posit that MUC1CT interacts with the open conformation of ALIX in the Bro1-V domains and that this complex is targeted to the membrane of MVBs. As well, MUC1 bound to other oncogenic molecular effectors likely contribute to ALIX-dependent EV formation and cargo sorting. Of particular interest is Src, which has previously been shown to associate with ALIX through interactions between the Src SH3 domain and the ALIX PRD domain. Given that MUC1 interacts strongly with the ALIX BRO1 domain and Src, we propose that a complex of MUC1, ALIX and Src is packaged into EVs through a unique association and molecular pathway that creates tumor-associated cargos. The fact that MUC1-expressing cells produced higher numbers of EVs is consistent with prior reports that Src increases exosome production(69) and supports the hypothesis that these complexes impact ESCRT-mediated EV secretion. Thus, the association between MUC1, ALIX and Src reported here is highly significant and may explain, in part, the process of differential packaging of cargo into unique MUC1-positive tumor EVs that we observed in both clinical samples and pancreatic tumor cell lines.

In summary, the results presented here show that EVs from pancreatic tumors contain MUC1 and a number of differentially packaged proteins that are biologically significant and that enhance tumor growth, motility, invasion, and metastasis. EVs from normal pancreatic cells or tumor cell lines with MUC1 knocked out do not show these activities. We present evidence that MUC1 influences the production of tumor-associated EVs and their contents by associating with ALIX and other oncogenic molecular effectors such as Src to form and load oncogenic contents into the EVs, which can be transported to other cells. These processes are predicted to create protumorigenic and prometastatic niches at both local and distant sites. While our study revealed the influence of MUC1 in cancer EVs using PDAC as a model, the implications of our results can be expanded to include numerous other cancers that have also shown overexpression of MUC1.

Key Resources Table

Reagent	Source	Identifier
Antibodies
ALIX (Host: mouse)	Santa Cruz	Cat#: sc-53540
ALIX (Host: rabbit)	Cell Signaling Technology	Cat#: 92880
Annexin A1 (Host: rabbit)	Abcam	Cat#: ab214486
Annexin A3 (Host: mouse)	ThermoFisher	Cat#: MA5-36106
Aldolase C (Host: mouse)	ThermoFisher	Cat#: MA5-37623
Calnexin (Host: mouse)	Novus	Clone: 1E2.1C12; Cat# NBP2-36571APC
Calnexin (Host: rabbit)	Cell Signaling Technology	Cat#: 2679
CD71 (D7G9X)	Cell Signaling Technology	Cat#: 13113S
EEA1	Cell Signaling Technology	Cat#: 3288
Flotillin-2 Antibody (B-6)	Santa Cruz	Cat#: 28320
GM130 (Host: rabbit)	Cell Signaling Technology	Cat#: 12480
Goat Anti-Armenian Hamster IgG H&L (Alexa Fluor 488)	Abcam	Cat#: ab173003
Goat Anti-Mouse IgG H&L (Alexa Fluor 647)	Abcam	Cat: ab150115
Goat anti-Rabbit IgG (H+L) Secondary antibody, HRP	ThermoFisher	Cat#: 31460
Mouse Anti-rabbit IgG (Conformation Specific) (L27A9) mAb (HRP Conjugate)	Cell Signaling Technology	Cat#: 5127
Goat Anti-Rabbit IgG H&L (Alexa Fluor 555)	Abcam	Cat#: ab150078
HA tag antibody (Host: rabbit)	Abcam	Cat#: ab236632
HA-Tag (C29F4) Rabbit mAb (Magnetic Bead Conjugate)	Cell Signaling Technology	Cat#: 11846
Human CD63 (Host: mouse)	Biolegend	Cat#: 353010
Human CD9 (Host: rabbit)	Abcam	Cat#: ab92726
Human CD9 (Host: rabbit)	Abcam	Cat#: ab236630
IRDye 680RD Donkey anti-Rabbit IgG	Li-COR	Cat#: 926–68073
IRDye 800CW Donkey anti-Mouse IgG	Li-COR	Cat#: 926-32212
LAMP1 (Host: rabbit)	Cell Signaling Technology	Cat#: 9091
MUC1 (Host: Armenian hamster)	ThermoFisher	Clone: MH1(CT2); Cat# MA5-11202
MUC1 (Host: rabbit)	Abcam	Cat#: ab109185
Phospho-Src (Tyr527)	Cell Signaling Technology	Cat#: 2105
Phospho-Src Family (Tyr416)	Cell Signaling Technology	Cat#: 6943
Rab7 (Host: rabbit)	Cell Signaling Technology	Cat#: 9367
Src (Host: rabbit)	Cell Signaling Technology	Cat#: 2109
Syntenin-1 (Host: rabbit)	Cell Signaling Technology	Cat#: 27964
Tsg101 (Host: mouse)	Santa Cruz	Cat#: sc-7964
Tsg101 (Host: mouse)	BD Transduction Laboratories	Clone: 51/TSG101; Cat# 612696
β-Actin Antibody (C4)	Santa Cruz	Cat#: sc-47778
Critical Commercial Assay
Lipofectamine 3000	ThermoFisher	Cat#: L3000015
APC/Cy7 Conjugation Kit	Abcam	Cat#: ab102859
Cell Counting Kit 8	Abcam	Cat#: ab228554
CellTiter-Glo® Luminescent Cell Viability Assay	Promega	Cat#: G7572
Cy5™ Mono-Reactive Dye	Amersham	Cat#: PA25001
Cy5 Mono-Reactive Dye Pack	GE Healthcare	Cat#: PA25001
Hi-T4 DNA ligase	New England Biolabs	Cat#: M2622S
NEB® Turbo Competent E. coli (High Efficiency)	New England Biolabs	Cat: C2984I
TC Plate Insert, PC Membrane (24-well, 8 μm)	VMR	Cat#: 10769-234
Matrigel	Corning	Cat#: 354234
Cell Lysis Buffer (10X)	Cell Signaling Technology	Cat#: 9803
pLVX-EF1α-IRES-ZsGreen1	TaKaRa	Cat#: 631982
pSELECT-NHA-blasti	InvivoGen	Cat#: psetb-nha
Stellar™ Competent Cells	TaKaRa	Cat#: 636763

Significance Statement.

Our study demonstrates that MUC1-ALIX-reglated cargo sorting plays a pivotal role in determining the molecular composition of MUC1-positive EVs. These EVs are selectively enriched with Phospho_Tyr416-Src, a key effector of oncogenic signaling. The presence of Phospho_Tyr416-Src within MUC1-positive EVs enhances intercellular communication and promotes pancreatic cancer progression. Collectively, these findings define a mechanistic link between MUC1-mediated EV biogenesis, ALIX-dependent cargo selection, and the propagation of tumor-promoting signals in the pancreatic tumor microenvironment.