Human PDAC cell lines BxPC-3 and previously established gemcitabine-resistant subline BxPC-3-Gem (18) were cultured in RPMI-1640 medium (Gibco, BRL Co. Ltd., USA) containing 10% fetal bovine serum (FBS) (Gibco, BRL Co. Ltd., USA)). Mia-PaCa2, PANC-1 and HEK293T cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Gibco, BRL Co. Ltd., USA) supplemented with 10% FBS. Cells were cultured in a humid atmosphere containing 5% CO₂ at 37°C. All cells were tested for mycoplasma at regular intervals.

The plasmid pLVSIN-CMV-puro carrying GFP (pLVSIN-GFP) was constructed as described previously (19). pLVSIN-CMV-puro carrying RFP (pLVSIN-RFP) was constructed using the same approach. Stable cell lines were established through lentiviral transduction. Briefly, the constructed vectors and lentiviral packaging mix (VSV-G plasmid and Gag-Pol plasmid) were co-transfected into HEK293T cells. The supernatants containing lentiviruses were collected, filtered, and added into BxPC-3 and BxPC-3-Gem cells for 2 days. The transduced cells (pLVSIN-RFP-BxPC-3-Gem, pLVSIN-GFP-BxPC-3) were selected with puromycin (Santa Cruz, Texas, USA).

Exosomes were isolated from PDAC cell lines by ultracentrifugation or PEG. After 48-hour cell culture in exosome-free medium, which was centrifuged at 100 000 g for 70 minutes to remove any exosomes from the serum in advance, the medium was centrifuged at 500 g for 5 minutes to remove any cell contamination. The resulting supernatants were centrifuged at 12 000 g for 20 minutes to remove any possible apoptotic bodies and large cell debris. The exosomes were collected as pellets after ultracentrifugation at 100 000 g for 70 minutes, then washed in PBS and pelleted again by ultracentrifugation.

For PEG approach, appropriate amount of 5X PEG8000 was thoroughly mixed with the 48-hour cell culture medium to a final 1x PEG8000 concentration. After incubation at 4°C for 12 hours, the samples were centrifuged at 12 000 g for 10 minutes. The supernatant was discarded as much as possible and exosome pellets were collected, and then washed in PBS and pelleted again by centrifugation.

Exosome preparations were verified by electron microscopy (Quanta GEG250, FEI, Hillsboro, USA) and the size and particle concentration were analyzed using the ZetaPlus nanoparticle characterization system (Brookhaven, NY, USA).

The isolated exosomes were stained with fluorescent dye PKH67 (Sigma, Saint Louis, USA) according to the manufacturer’s protocol. Briefly, exosomal protein concentration was determined by using Pierce BCA Protein Detection Kit (Thermo, Rockford, USA). Exosomes (20 μg) were suspended in 1 mL a Diluent C and incubated with equal volume of Diluent C containing 5 μL of PK67 for 5 min. Serum (2 mL) was added to terminate the staining. After washing with 1×PBS and centrifugation at 4°C, 120,000 g for 70 min, the stained exosomes were re-suspended in 1×PBS.

MMP14 overexpression vector (GFP-MMP14) was constructed by cloning the full-length cDNA of MMP14 into the Bgl II/EcoR I sites of pEGFP-C1 vector. The nucleotides targeting MMP14 and negative control were synthesized by GenePharma (Shanghai, China). The sequences of primers and siRNAs are listed in Table S1. Transient transfection was mediated by Lipofectamine 3000 (Invitrogen, Eugene, OR, USA) in pLVSIN-RFP-BxPC-3-Gem, pLVSIN-GFP-BxPC-3, PANC-1 and BxPC-3-Gem cells following the manufacturer’s protocol.

For the sphere formation assay, cells (500/well) were seeded into the ultra-low attachment 6-well plates (Corning, Inc., Corning, NY, USA) and cultured in DMEM-F12 medium (Gibco, Grand Island, NY, USA), containing 2% B27 (Gibco, MD, USA), 10 ng/mL of epidermal growth factor (EGF; Gibco, MD, USA), and 10 ng/mL of basic fibroblast growth factor (FGF; Gibco, MD, USA). After 14 days of culture, the spheres with diameter > 75 μM were counted. For colony formation assay, BxPC-3-GFP and BxPC-3-Gem-RFP cells were seeded into 6-well plates (Corning, Inc., Corning, NY, USA) individually or together, cultured for 14 days, thereafter treated with gemcitabine for 72 hours. The fluorescence of the colonies was detected by the inverted fluorescence microscope (IX71, Olympus Corporation, Japan).

The PDAC cells were seeded in the upper chamber of Transwell (Costar Corp., Cambridge, MA, USA) and allowed to translocate toward medium containing 20% FBS in the lower chamber for 48 hours. 4% formaldehyde and 0.5% crystal violet were used to fix and stain the cells that migrated to the lower surface.

Total RNA was isolated using Trizol (Invitrogen, Eugene, OR, USA) and reverse-transcribed (RT) into cDNA using ReverTra Ace (TOYOBO, Japan). RT-qPCR was performed using SYBR Premix Ex Taq (Takara, Otsu, Shiga, Japan) on ViiA7 Real-time PCR System (Applied Biosystems Inc., Foster City, CA, USA). GAPDH was used as the internal control for mRNA. Detailed information about the primers is shown in Table S1.

Cell lysis and Western blot were conducted as previously described (18). Briefly, about 20-40 μg proteins per well were resolved by SDS/PAGE and transferred on PVDF membranes (Millipore, Darmstadt, Germany). The membranes were incubated with antibodies against TSG101 (Proteintech, Wuhan, China), CD44 (Abcam, Cambridge, UK), MMP14 (Abcam, Cambridge, UK), and Activin A (ThermoFisher, Rockford, USA). β-actin (Santa Cruz, CA, USA) was used as loading control. Band intensity was quantified using ImageJ software (NIH).

The differentially secreted proteins in BxPC-3-gem vs BxPC-3 cells were expressed by fold change, and |log (FC)| > 1 was used as the cut-off value. All statistical analyses were conducted with R software (Version 4.1) and Bioconductor version 4.0. KEGG enrichment and GO function annotation analysis were performed by R package “clusterProfiler” (20). KEGG or GO terms with BH-corrected p<0.05 were considered as significance. Gene Set Enrichment Analysis (GSEA) was performed by R package “clusterProfiler” with BH-corrected p<0.05 as significance, and “enrichplot” was used to visualize the significant results.

The PDAC cells were seeded into 96-well plates and treated with various concentration of gemcitabine (LC Laboratories, Woburn, USA) for 72 hours, followed by the addition of MTT (5 mg/mL, Amresco, USA) for 4 hours. The formed MTT products was dissolved in DMSO (Sigma, Saint Louis, USA), and colorimetric analysis (wavelength, 490 nm) was performed using iMark Microplate Absorbance Reader (Bio-Rad, Hercules, CA, USA).

Cells were seeded in 24-well plates (1×10⁴/well) after pre-incubation with exosome. Wound healing assay was described as previously (Wang H., 2018)

Results for continuous variables are presented as mean ± SD unless stated otherwise. For two-group comparison, we used two-tailed Student’s t test. p< 0.05 was considered statistically significant. All analyses were performed using SPSS v.17.0 software (SPSS Inc.).

Our previous study indicates that the chemoresistant PDAC cells secreted vital factors to enhance the resistance of the sensitive PDAC cells (19). To understand the molecular mechanism, we re-visited and analyzed our comparative secretome data, and obtained the network of proteins based on cell components and molecular function (Supplementary Figures 1A, B). More proteins of BxPC-3-Gem cells were associated with cell adhesion, growth factor and receptor ligand activity, cytokine receptor binding, clathrin binding and actin binding (Supplementary Figures 1A, B). Gene Ontology (GO) analysis showed that the differential proteins of gemcitabine resistant cells were enriched in vesicle lumen, secretory granule lumen and cytoplasmic vesicle lumen relative to parental cells by cellular component (Figure 1A). Because these multivesicular components are involved in the packaging of exosomes (21), this result suggested that some differential proteins in the resistant cells might be secreted via exosomes. Among the differential proteins, MMP14 drew our attention because it was among the top 3 differentially expressed proteins in BxPC-3-Gem in relative to parental cells (Figure 1B). To compare the functions of the differential proteins, Gene Set Enrichment Analysis (GSEA) analysis was performed. Results showed that BxPC-3-Gem cells secreted more proteins which were associated with cell adhesion and cytoplasmic membrane system relative to BxPC-3 cells (Figures 1C, D; Tables S2, S3). Analysis of subfunctional pathways involved in cell adhesion system indicated that MMP14 was also related with the regulation of cell-matrix adhesion and other functional pathways (Figure 1E; Supplementary Figure 1C), indicating that MMP14 is a predominant protein secreted from chemoresistant cells in comparison with the sensitive cells. As MMP14 is a transmembrane protein, it should be engulfed into the exosome for extracellular transmission.

We isolated all microparticles (microvesicles and exosomes) secreted from BxPC-3-Gem and its parental BxPC-3 cells using differential centrifugation or with polyethylene glycol (PEG) approach. NanoSight particle tracking analysis showed that the predominant microparticles in both BxPC-3-Gem and BxPC-3 cells were of exosomal size (30~100 nm) (6) (Figure 2A). The typical size of the exosomes extracted from BxPC-3-Gem ranged 40~110 nm while exosomes isolated from BxPC-3 cells ranged 20~50 nm (Figure 2A). In addition, scanning electron microscopy revealed that the isolated microparticles were of exosomal morphology (Figure 2B). To see whether MMP14 was engulfed in the exosome, the isolated exosomes were subjected to Western blot. Exosomes were confirmed by the expression of exosome marker, the tumor susceptibility gene (TSG101), while MMP14 was detected in the portion of exosome pellets (Figure 2C), but not in the supernatant (Figure 2D). Furthermore, more exosomes were secreted from chemoresistant BxPC-3-Gem than parental cells through a quantified analysis of exosomes. Equal amounts of cells were indicated by equal loading control in the two cell lines (Figures 2C, E). In addition, higher level of MMP14 was detected in both the exosomes (Figure 2C) and whole cell lysates (Figure 2E) of the resistant cells in comparison with parental cells.

Neighboring cells can uptake exosomes (22). To evaluate whether the exosomes secreted from the resistant PDAC cells can be internalized by the neighboring sensitive cells, BxPC-3 and Mia-PaCa-2cells were incubated with PKH67-dyed exosomes isolated from the gemcitabine-resistant BxPC-3-Gem cells. Using confocal microscopy, we observed PKH67-dyed exosomes internalization into BxPC-3 and Mia-PaCa-2 cells within 48-hour co-incubation (Figure 3A). To determine whether exosome-transferred MMP14 could confer the resistant phenotype to recipient sensitive PDAC cells, the exosome donor cells BxPC-3-Gem were overexpressed with MMP14 (Figure 3B) or knocked down of MMP14 (Figure 3C). The isolated exosomes carried more MMP14 proteins when it was overexpressed, while less exosome-MMP14 proteins were detected when MMP14 was knocked down in BxPC-3-Gem cells compared with the control (Figures 3D, E). MTT assay showed that the collected exosomes had no effect on the proliferation of recipient Bx-PC-3 or Mia-PaCa-2 cells upon gemcitabine treatment no matter whether MMP14 was overexpressed or knocked down in the donor resistant cells (Supplementary Figures 2A, B). To see whether a long period of incubation with exosomes could educate the recipient sensitive cells in obtaining resistance to gemcitabine, co-culture colony formation assay was performed using GFP-labeled sensitive BxPC-3 cells and RFP-labeled resistant BxPC-3-Gem cells. The BxPC-3 cells became resistant to gemcitabine when co-cultured with gemcitabine-resistant cells for two weeks and MMP14 overexpression in the resistant cells further increased this effect (Figures 3F, G). To further address whether exosome-transferred MMP14 was a key intercellular messenger for mediating chemoresistance, MMP14 was knocked down in RFP-labeled BxPC-3-Gem cells. The colony formation of BxPC-3 cells post gemcitabine treatment was increased when co-cultured with RFP-labeled resistant BxPC-3-Gem cells, but was significantly abrogated when MMP14 was knocked down in the resistant cells (Figures 3H, I). We noticed that the abrogation effect was comparable to MMP14 knockdown efficiency, indicating that MMP14 is a key player in exosome-mediating transmission of chemoresistance.

Cancer cells that acquire cancer stem cell-like properties confer to chemoresistance (23). To determine whether exosome-transferred MMP14 could educate the recipient cells to obtain more aggressive properties, sphere formation assay was performed in the recipient cells post exosome uptake from the resistant cells. The sphere-forming abilities were increased in BxPC-3 and Mia-PaCa2 cells through uptaking exosomes from resistant PANC-1 cells, which were further enhanced when MMP14 was overexpressed in those cells (Figures 4A, B). Knockdown approach confirmed that exosome-transferred MMP14 participated in the regulation of cancer stemness of the recipient cells. The sphere-forming abilities were greatly increased in BxPC-3 and Mia-PaCa2 cells by uptaking exosomes from BxPC-3-Gem cells, which were dramatically decreased when MMP14 was knocked down in those cells (Figures 4C, D). GSEA enrichment analysis of comparative secretome also showed that MMP14 was involved in the regulation of cell growth and differentiation of stem cells (Supplementary Figures 3A-C; Tables S4, S5).

Transwell assays showed that the exosomes shed by MMP14-overexpressing PANC-1 cells could improve the migration abilities of the sensitive BxPC-3 and Mia-PaCa-2 cells (Figures 5A, B). Knockdown approach confirmed that the exosomes from MMP14 knockdown BxPC-3-Gem cells led to decreased migration abilities of BxPC-3 and Mia-PaCa-2 cells in comparison with that from control BxPC-3-Gemcells (Figures 5C, D). Scratch assay confirmed that the sensitive cells could obtain increased motility upon receiving exosome-transferred MMP14 (Supplementary Figures 4A, B).

Epithelial to mesenchymal transition (EMT) is regarded as one of the sources of cancer stem cells (24). Our GSEA analysis and point plots on EMT suggested that both MMP14 and CD44 were involved this process (Supplementary Figures 4C, D; Table S6). Given that MMP14 can form a complex with CD44 on the cell membrane (25), we speculated that exosome-transferred MMP14 might affect recipient cells’ stem-like properties via modulating the activity of CD44. Western blot showed that the exosomes from either MMP14 overexpressing PANC-1 cells or GFP control cells could increase CD44 protein levels in the recipient BxPC-3 cells (Figure 6A). Furthermore, exosomes from MMP14-overexpressing PANC-1 cells led to greater CD44 level in comparison with that from the GFP control cells (Figure 6A). We presumed that there are two possibilities associated with CD44 protein increase in the recipient cells. One possibility is the exosomes from MMP14-overexpressing PANC-1 cells might carry more CD44 protein than the GFP control cells. The other possibility is that the exosomes from MMP14-overexpressing PANC-1 cells might lead to more CD44 expression or protein accumulation in the recipient cells than that from GFP control cells. To rule out the possibility that the increased CD44 might have resulted from exosomal transfer, CD44 levels in the exosomes were determined. Results showed that CD44 did exist in the exosomes, however the exosomes from MMP14-overexpressing PANC-1 cells did not carry more CD44 protein in comparison with that from the GFP control cells (Figure 6B). mRNA analysis indicated that CD44 mRNA level was not affected by the exosomes (Figures 6C, D). In addition, the exosomes from either MMP14-overexpressing PANC-1 or MMP14 knockdown BxPC-3-Gem cells did not change CD44 expression relative to the exosomes from the control cells (Figures 6C, D). Then, we used cycloheximide (CHX) to block protein synthesis to see CD44 protein accumulation. The exosomes from MMP14-overexpressing PANC-1 cells led to increase of CD44 protein in the recipient BxPC-3 cells in comparison with the exosomes from GFP control cells (Figure 6E). Moreover, the exosomes from MMP14 knockdown BxPC-3-Gem cells led to CD44 protein levels dramatically decreasing in BxPC-3 cells in comparison with the control cells (Figure 6F). These results indicate that exosome-transferred MMP14 promotes CD44 stability in recipient cells.