Animals
The Animal Care and Use Committee at the University of Chicago approved all of the procedures outlined here. All animals were housed in a specific, pathogen-free, biohazard level 2 facility, maintained by The University of Chicago Animal Resources Center (Chicago, IL). Animal genotyping was performed by Transnetyx, Inc., (Cordova, TN).
NPN membranes
Chips with NPN membranes were purchased from SiMPore Inc. (West Henrietta, NY). Each chip was 300 µm thick, 5.4 mm wide and 5.4 mm long, and contained an active, 100 nm-thick NPN membrane with dimensions of 0.7 mm width and 2 mm length. Each nanomembrane was manufactured with ~5 × 107 pores with dimensions ranging from 30 to 140 nm using manufacturing conditions that allow for some pore merger events.39 Vesicles in this size range become trapped inside of the tapered pores or on top of pores, while smaller vesicles and protein aggregates will pass through.
Device fabrication
Microfluidic devices were fabricated using PDMS sheets (Trelleborg Sealing Solutions Americas, Fort Wayne, IN). Custom NuSil 4930 300 µm-thick silicone sheets were patterned using a conventional digital desktop cutter (Silhouette America, Oren, UT). The top 3 layers of the device, including the 100 µm wide channel, the sealing layer, and the PDMS slab were bonded to each other using UV/ozone treatment. The middle compartment was cut out from this assembly (6 mm × 6 mm); this middle compartment would allow for NPN chip placement after it was loaded with vesicles. The top assembly was then stacked on top of the bottom NPN housing layer using the alignment marks. Then, inlets and outlets were punched out. Finally, the PDMS assembly was bonded to a glass slide using UV/ozone treatment. For final device assembly, the NPN chip was then loaded in the EV capture section and placed within the NPN housing layer via the middle cutout compartment. The compartment was then sealed shut using PDMS as adhesive and the inlets and outlets were filled with aqueous media. The finished device was placed in an oven for 20 min at 40 °C to cure the PDMS.
Procurement of EV samples
EV samples were obtained from two sources, BAL fluids and conditioned media from cells lines in culture. BAL fluids were collected from mice using standard methods, as previously described40. After harvesting, the BAL fluids were centrifuged at 300 × g for 5 min to separate cells from the fluid. The cell pellet was frozen at 20 °C for downstream protein analysis. We then used a series of high-speed centrifugation of supernatant to concentrate EVs while also removing large debris and dead cells (2000 × g for 10 min to remove dead cells; 10,000 × g for 40 min to remove large debris; 4000 × g for 10 min in Amicon (Millipore Sigma) spin columns to concentrate vesicle samples). The final sample was between 200 and 500 µl and used for EV capture on plasma cleaned NPN membranes.
Media from cultured cell lines were obtained from HEK and J774 cells obtained from the American Type Culture Collection (ATCC) and cultured in EV-depleted DMEM supplemented with 10% FBS, 1% (vol/vol) penicillin–streptomycin. All cultures were maintained at 37 °C, 5% CO2, and 100% humidity. Cells were grown to 80% confluence in T75 flasks. At this point, the medium was removed and the cells were passaged either for future studies or trypsinized for cell protein analysis. The media was subjected to the same high-speed centrifugation cycle as the BAL fluids before concentrating to a final volume between 200 and 500 µl.
The EVs shown in the EM image in Fig. 1b were purchased from Hansa Biomed. They were plasma-derived EVs at 100 μg protein content (expected) at a concentration of approximately 10 × 1010/ml. They were reconstituted in water to reduce the total amount of salts then spun against the NPN membrane for facilitated capture at 500 × g for 30 min. The sample was then extracted from the SepCon unit, dehydrated in a sequential ethanol dilution series (50%-100% EtOH) for 10 min at each stage. The sample was then coated with ~10 nm gold before imaging in a Zeiss Auriga SEM at 10 kV.
EV loading onto NPN membranes
Loading of the sample fluid onto the NPN membrane was achieved by the use of SepCon™ centrifuge units (SiMPore Inc., West Henrietta, NY). The centrifuge units contained a housing area where the NPN membranes were placed. The NPN membrane was pre-treated to facilitate wetting by either exposing to UV/ozone treatment 20–30 min before loading or wetting the backside through a port provided in the SepCon unit. The range of plasma conditions tried for NPN preparation did not make an obvious difference in the wettability of the membranes or the number of vesicles captured, although this was not rigorously quantified. Once in place and pre-treated, 120 µl of the sample was deposited on top of the membrane. The sample was then spun at 3000 × g for 5 min trapping EVs within the nanopores. The NPN membrane-containing chip was then removed from the centrifuge unit and placed on the microfluidic platform. The device assembled on the stage of laser scanning confocal microscope allowed for sequential exposure of the captured EVs to buffer solutions. Inlet and outlet tubing allowed for the flow of solutions over the NPN membrane-immobilized vesicles, at a rate of 5–10 µl/min, maintained by a pressure-driven microfluidic flow system (ElveFlow®) coupled with a flow rate control system. Our measurements were taken under steady-state flow conditions. In similar devices for other studies used under constant flow rate, the pressure settles to a steady-state level within a minute of the onset of flow or solutions41. Thus, we presume that our experiments reach steady flow in under a minute which is much faster than the ~10 min measurement windows used to acquire data. Different treatment windows were tried and were not found to make any obvious difference, although we did not quantify the difference. Once the microfluidic device was placed on the microscope for imaging, NPN membrane-immobilized vesicles were equilibrated for 5–10 min before they were exposed to any stimuli, and images were taken. It is important to note, our experiments were conducted under constant pressure and, therefore, stable flow conditions. Solution changes were made such that all transient fluctuations were 10 min before the peak amplitude in response was determined and averaged over 5 time points.
EV isolation using size-exclusion column chromatography
To obtain a high purity sample of EVs, we used a size-exclusion chromatography column (Izon Science, Christchurch, New Zealand) to isolate the EV fractions of interest while separating out free-floating proteins. Eluate fractions (~500 μl per fraction) were collected individually; fraction numbers 6–10 contained most of the EV content and were pooled and concentrated to a final volume between 200 and 300 µl using Amicon 100 kDa molecular weight cut-off (MWCO) centrifugation filters (Millipore Sigma).
EV isolation using AF4
A higher resolution of vesicle size distribution as compared to that obtained with the size exclusion column strategy was obtained with the technique of AF4 (Wyatt Eclipse with Dilution Control Module, Wyatt Technology, Santa Barbara, CA). With AF4 separation we are able to concentrate, quantitate, characterize, and elute EVs, separate from contaminants, in real time with virtually no retention of contaminating soluble protein. A multitude of detectors including multi-angle light scattering, DLS, and ultraviolet gave reproducible MW measurements.
EV characterization and labeling
Size-exclusion chromatography prepared EVs were analyzed for membrane markers and size. EV content of isolated fractions was confirmed by protein analysis (Qubit, Thermo Fisher Scientific), and EV size and quantity was performed by NanoSight Nanoparticle Tracking Analysis (Malvern | PANalytical Products, Westborough, MA).
Immunoblotting analysis validated the presence of known EV membrane markers CD63 and CD9. Additionally, the lipid fluorescent label Bodipy TR (1 µM) was used to label the outer lipid structure of EVs for visualization with confocal microscopy.
Laser confocal live EV time-lapse imaging
The microfluidic platforms were sealed and loaded with the NPN chip, the inlets and outlets of the device were connected to a pressure-driven microfluidic flow control system (ElveFlow, Paris, FR). The flow rate was kept between 5 and 10 µl/min for the duration of the experiment. The microfluidic device was placed on the microscope for imaging. Control fluid flow was administered for equilibration for 5–10 min while EVs were loaded with the fluorescent dye AO before the initiation of image acquisition. Changes in intravesicular pH were monitored using AO (4 µM). EVs were pretreated with a 50 mM NH4+ solution to artificially acidify the lumen of membrane competent, sealed vesicles visualized on the chip with the fluorescent lipid dye, Bodipy TR. The fluid medium was switched after a 15 min control period to a desired ionic condition always keeping the AO concentration constant. The solution ionic conditions varied from a high Na+ solution, a Na+-free solution, or a high Na+ solution containing the NHE1 inhibitor, HOE (cariporide, 1 µM). In the pH experiments, solutions differing in pH were introduced into the device at a volumetric flow rate of 10 µl/min. Image acquisition, processing, and analysis were performed at the University of Chicago Integrated Light Microscopy Facility. Images were captured with a Leica SP5 Tandem Scanner Spectral 2-photon confocal microscope. Image processing was performed using Bitplane Imaris software v. 9.1.2 (Andor Technology PLC, Belfast, N. Ireland).
Super-resolution STED EV imaging and analysis
Super-resolution STED imaging was done using a Leica SP8 laser confocal microscope with STED functionality. STED depletion was performed with 775 nm laser set at 100% power. STED images were analyzed using the Leica LIGHTNING (Lng) image deconvolution technique and further analyzed using the ImageJ analysis software.
EV DLS size analysis
DLS EV size measurements were performed on AF4 separated vesicular fractions using the DynaPro DLS plate reader (Wyatt Technology Corp., Santa Barbara, CA) in 96-well plates. The DLS data were analyzed using Dynamics version 7.8.1.3. software (Wyatt Technology Corp).
Solutions
Solutions used in the functional analysis of vesicles were as follows:
High Na+ solution contained (in mM): 125 NaCl, 2.5 KCl, 10 HEPES, 1.5 MgCl2, 2.5 CaCl2, 10 glucose, 20 sucrose, pH 7.4. Na+-free solution contained (in mM): 125 N-methyl d-glucamine, 40 HCl, 10 HEPES, 1.5 MgCl2, 2.5 CaCl2, 10 glucose, 20 sucrose; pH 7.4. Acid-loading solution contained (in mM): 50 NH4Cl, 30 KCl, 10 K methane sulfonate (MeSO4), 10 HEPES, pH 7.4.
For pH measurements, we used an Intracellular pH calibration buffer Kit (Invitrogen, P35379). In all cases, the osmolarity was set to 300 ± 10 mOsm with the major salt.
Dot blot development and analysis
BAL from 3 mice was pooled together (~9 ml) and cleared by one low and two medium-speed centrifugations to remove live cells, dead cells, and cellular debris, as described above, followed by concentration down to ~500 µl in Amicon Ultra-15 Centrifugal Filters (Millipore Sigma) with a MWCO of 100 kDa.
Concentrated BAL was lysed on ice by passaging 10 times through a 30G1 needle (BD) in 10× RIPA buffer (Sigma), supplemented with cOmplete™ Protease Inhibitor Cocktail (Millipore Sigma). The lysate was further cleared from insolubilized components by centrifugation at 17,000 × g for 10 min in a tabletop centrifuge (Fisher Scientific).
BAL samples were blotted onto methanol-activated PVDF membrane (BioRad), using a Minifold-I Dot-Blot System (Schleicher and Schuell BioScience Inc., Keene N.H.). Briefly, 200 µl samples were added to each well of the manifold wells and allowed to passively filter through by gravity flow for ~2 h, followed by blocking using gravity flow with 1% bovine serum albumin (BSA; Sigma-Aldrich) solution in TBS buffer (20 mM Tris and 150 mM NaCl) for 1 h. Blocked samples were washed with 500 µl of TBS-T solution (0.1% Tween-20 in TBS) 3 times by vacuum suction, after which the PVDF membrane was removed from the manifold and washed additional three times with TBS-T for 5 min in a plastic box on a rocker (Cole Parmer). Primary antibodies, diluted 1:3000 in 0.1% BSA (Sigma-Aldrich) in TBS-T, were incubated with the membrane for 1 h at room temperature (RT), followed by three washes for 10 min with TBS-T solution. Horseradish peroxidase (HRP)-conjugated secondary antibodies, at 1:150,000 dilution, were incubated with the membrane for an hour at RT, followed by three washes with TBS-T for 10 min and an additional wash in TBS-T for 30 min. Dot Blots were developed using Super Signal West Pico PLUS HRP Substrate (Thermo Scientific) and detected by Chemidoc (Azure Biosystems, Dublin, CA.).
Western blot
Mammalian cells cultured near confluence in T75 flasks were washed twice with cold PBS and scraped into 200 µl of Pierce 1x RIPA buffer (Thermo Scientific), supplemented with protease inhibitor cocktail (Millipore Sigma), and then transferred to 1.5 ml microtube kept on ice. Cells were vortexed briefly and disrupted by trituration 10× with a 23 G ¼ inch syringe needle (Becton, Dickinson and Co. Thermo Scientific). To remove unbroken cells and cellular debris, lysates were spun at 17,000 × g for 10 min and the supernatant was collected for downstream applications, as well as for determination of protein concentration by the BCA Protein Assay Kit (Thermo Scientific).
For running Western blots, 30 µg of cell lysates in Sample Buffer was loaded per each well of 4–20% Precast Protein Gel (BioRad, Hercules, CA) and then transferred to PVDF membrane. For blocking, 3% BSA (Sigma-Aldrich) in TBS buffer was used, while antibodies were diluted in 1% BSA in TBS-T. For detection of CD63 protein, we used either anti-mouse CD63 (ab217345) or anti-human CD63 (ab134045) (Abcam, Waltham, MA); for detection of NHE1, the Anti-Na+/H+ Exchanger 1 (extracellular) was used from Alomone Labs (ANX-010, Jerusalem, Isreal); for secondary antibodies, the HRP-conjugated goat anti-rabbit IgG H&L (ab6721, Abcam) was used; primary antibodies were used at 1:2000 and secondary antibodies at 1:10,000 dilutions.
Statistics and reproducibility
To analyze the statistically significant difference between the groups, ANOVA analyses (Student’s t test and Mann–Whitney Rank-Sum Test) were used. Data are represented as mean ± SEM. In all studies, a chip represented an independent experiment. At least three mice were used in all experimental studies.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
