Culturing cells

The HEK293F (Thermo Fisher Scientific) and HMEC-1 (ATCC-3243) cell lines were used in this study. The medium used to cultivate the HEK293F cells was Freestyle 293 medium (Thermo Fisher) with 100 µg/mL penicillin-streptomycin solution (Cytiva Hyclone). The HEK293F cells were cultured both adherently and in suspension. When cultured in suspension they were cultured in a shaking incubator with shaking at 130 rpm at 37 °C, 70% humidity, and 5% CO2. When cultured as adherent cultures, 10% FBS (Cytiva Hyclone) was added to the medium and the cells were cultured at 37 °C with 5% CO2 in a regular incubator. The HMEC-1 cell line was cultured adherently in MCDB131 medium (without l-glutamine, ThermoFisher Scientific) supplemented with 10 ng/mL human EGF recombinant protein (ThermoFisher Scientific), 1 µg/mL hydrocortisone-water soluble (Merck), 10 mM glutamine (Cytiva Hyclone), 100 µg/mL penicillin-streptomycin solution (Cytiva Hyclone), and 10% FBS (Cytiva Hyclone).

Developing the stable LFA-1 clone

The stable LFA-1 clone was developed via sequential transfection of wild-type (WT) HEK293F cells with human CD18 and human CD11a separately. Plasmids with the sequences for CD11a and CD18 for the transfections were purchased from GenScript (Supplementary Fig. 1A and 1B). For the transfection with CD18 the cells were grown in suspension, and for the transfection with CD11a the cells were growing adherently. The transfection with CD18 of cells growing in suspension was performed using the FectoPro system (Polyplus) according to the manufacturer’s protocol. In brief, on the day of transfection the cultures were adjusted to a concentration of 2 × 106 cells/mL in Freestyle 293 medium. For the transfection, IMDM media (Lonza) was used as the dilution medium. In a sterile tube, the IMDM (0.1 mL/mL final volume), plasmid (0.8 µg/mL final volume), and transfection reagent (0.8 µL/mL final volume) were combined. As a negative control, a mock transfection was performed using IMDM alone without plasmid. After a 20 min incubation at room temperature, the mixture was added to the cells and the cells were placed in a 37 °C incubator.

The transfection of adherently growing CD18 clones with CD11a was performed with the Lipofectamine 2000 system (Thermo Fisher) according to the manufacturer’s protocol. Briefly, two days prior to the transfection 1.0 × 106 cells were seeded on a 6-well plate (Falcon) in 2 mL growth medium (Freestyle 293 medium with 10% FBS). On the day of transfection, the wells were roughly 90–95% confluent, and the growth medium in the wells was replaced with 1.5 mL fresh growth medium. A mixture of 4 µg of plasmid with 10 µL of the Lipofectamine 2000 transfection reagent in a total volume of 500 µL IMDM medium (Cytiva Hyclone) was allowed to incubate for 20 min at room temperature. Following this incubation, the mixture was gently mixed with the growth medium in the wells. The same mixture without the plasmid was added to the cells for the mock transfection.

Generating a pure clone

For selection, the antibiotics G418 (500 µg/mL) and Hygro B (125 µg/mL) were added 48 h after transfection. The growth medium was changed every 48 h, and pools of clones were expanded or split as required. When the viability of the mock culture began to decline, a selection procedure was used to generate pure clones. Both the transfected and mock groups underwent identical selection processes involving seeding out the transfected cells in a series of 10× dilutions on Petri dishes. After around two weeks, 30–50 clones were picked from the Petri dishes using a pipette by gently scraping the surface of the petri dish and then collecting 50 µL of the media with the cells. The picked cells were added to a 96-well plate (Falcon) already containing 150 µL media for a total of 200 µL media per well. The clones were expanded as needed and analyzed via flow cytometry to identify the clones with the highest expression of the desired membrane proteins.

Flow cytometry

To determine the efficiency of the transfections and the expression of the desired membrane proteins, the cells were analyzed by flow cytometry on either a BD FACS Aria II cell sorter or a BD FACSVerse flow cytometer running BD FACSSuite software (BD Biosciences). One milliliter of ice-cold FACS buffer (PBS with 1% FBS) was added to 100,000 cells, which were then pelleted at 200 × g for 10 min at 4 °C. The cells were then resuspended in 50 µL of human IgG (1 mg/mL in D-PBS) and incubated for 15 min at 4 °C. After this incubation, antibodies specific for the membrane proteins of interest were then added together with a viability dye for 30 min at 4 °C. The following antibodies and viability dyes were used for the staining step: BD Pharmingen FITC mouse anti-human CD18 (Clone 6.7, BD Biosciences), BD Pharmingen PE mouse anti-human CD11a (Clone HI111, BD Biosciences), BD Pharmingen 7-AAD (BD Biosciences), and Invitrogen LIVE/DEAD Fixable Aqua Dead cell stain kit (Thermo Fisher). The cells were washed with 2 mL FACS buffer and then resuspended in a final volume of 350 µL of FACS buffer. A total of 10,000 events were collected for each sample, and the data were analyzed using FlowJo software (Tree Star Inc, Ashland, OR, USA).

Binding assay of LFA-1 cells with ICAM-1 coating

Wells in a 96-well strip plate (Thermo Fisher Scientific, Black microplate F96) were coated with ICAM-1 by introducing a 100 µL solution containing 5 µg/mL of the ligand in a coating buffer (0.1 M NaHCO3, pH 9.6) to the wells. Using the same concentration and procedure, wells were also coated with anti-CD18 (MEM-48, Thermo Fisher Scientific), bovine serum albumin, or human IgG as controls. Negative controls consisted of wells receiving only the coating buffer. The plates were subsequently sealed with parafilm and incubated overnight at 4 °C. Following incubation, the plates were inverted, and the solutions were removed from the wells, which were then washed twice with 150 µL of washing buffer (0.1% Tween in PBS). Subsequently, the wells were blocked with 150 µL of blocking buffer (1% BSA in PBS) per well at room temperature for 1 h, concurrent with cell labelling. Cell labelling was conducted using the fluorescent dye PKH67 (Sigma-Aldrich) following the manufacturer’s protocol. Briefly, the required number of cells was pelleted by centrifugation at 200 × g for 10 min, resuspended in 1 mL diluent C, and mixed with a 2× dye working solution composed of 1 mL diluent C and 4 µL PKH67. After an incubation period of 2–5 min, the reaction was halted by adding 2 mL of 1% BSA and incubating for 1 min in the dark. Excess dye was removed by washing the cells twice with PBS.

The blocking buffer was removed from the wells by inverting the plate, and the wells were washed twice with ice-cold PBS. After washing, the labelled cells were resuspended in pre-heated medium at 5 × 105 cells/mL. One hundred microliters of the labelled cells (50,000 cells) were added to each well. The plate was incubated at 37 °C for 60 min to activate the receptor so that it bound to the coated ligand. After the incubation, the plates were inverted and pressed against a paper towel, and 150 µL of ice-cold PBS was then slowly pipetted into the wells and the plate was gently shaken at a low speed for three minutes. This was repeated another two times, and 100 µL of PBS was added to the wells and the plate was analyzed with a Varioskan LUX multimode microplate reader (Thermo Fisher). The wavelengths used were 480 nm for excitation and 502 nm for emission.

EV isolation

A schematic overview of the EV isolation process is described in Fig. 1. Cells, cell debris, and large apoptotic bodies were removed from the cell culture media via centrifugation at 300 × g for 10 min and 2000 × g for 20 min. EVs were only isolated from cultures growing in suspension in FBS-free media. The supernatant was ultracentrifuged for 2.5 h at 118,000 × g at 4 °C (38,500 rpm, Type 45 Ti fixed angle rotor, 181 as k-factor, Beckman Coulter). The pellet was resuspended in 0.5–1 mL PBS, and the EVs were further purified via a bottom-loaded iodixanol density cushion. The cushion was generated by mixing the sample with 3 mL 60% iodixanol in the bottom of the tube (final concentration 45%) and then adding 4 mL of 30% iodixanol and 3 mL of 10% iodixanol to this layer. The sample was then ultracentrifuged at 100,000 × g for 2 h (28,000 rpm, SW 41 Ti swinging rotor, 265 as k-factor, Beckman Coulter). The purified EVs were collected at the 10–30% iodixanol interphase.

Fig. 1

Overview of the steps for the isolation and purification of the EVs. Cells, cell debris, and large apoptotic bodies were removed by slow centrifugation (300 × g and 2000 × g). The remaining EVs were pelleted by ultracentrifugation (118,000 × g) and were further purified via a bottom-loaded density cushion with iodixanol (45%, 30%, and 10%)

EV characterization via transmission electron microscopy (TEM)

LFA-1 EVs were negatively stained and visualized by TEM. In brief, vesicles were loaded onto glow-discharged 300-mesh copper grids (Electron Microscopy Sciences) for 30 s, then washed with water twice and further stained with 2% uranyl formate for 1 min. Negative-stained EVs were analyzed by acquisition on a TALOS L 120 C transmission electronic microscope (Thermo Fisher) at 120 kV with a BM-Ceta CMOS 4k*4k CCD camera.

Western blot

The protein concentrations of cell lysates and EVs isolated from different clones was determined with a Pierce BCA Protein assay kit (Thermo Scientific) following the recommended protocol. Samples for Western blot were made by diluting 5 µg protein from the lysates or EVs to a volume of 15 µL and then mixing it with 15 µL Sample Buffer (Bio-Rad, 2× Laemmli sample buffer), with or without dithiothreitol (DTT), to a total volume of 30 µL. DTT was used when analyzing samples for CD11a but not for CD18. Samples were added to Mini-PROTEAN TGX Stain-Free gels with seven wells (Bio-Rad) together with Precision Plus Protein Standard (Bio-Rad), using 1× Tris/Glycine/SDS Buffer (Bio-Rad). Proteins were transferred from the gel to a PVDF membrane (Bio-Rad) with Trans-Blot Turbo Mini-size Transfer Stacks (Bio-Rad) and 1× trans-blot buffer (Bio-Rad Trans-Blot Turbo 5× Transfer Buffer) using a semi-dry transfer chamber (Bio-Rad). The membrane was blocked with a blocking buffer (Bio-Rad EveryBlot Blocking Buffer) and incubated for 20 min. After this incubation, antibodies against CD11a (ab186873, Abcam, 1:1000 dilution) or CD18 (MEM-48, Thermo Fisher Scientific, 1:1000 dilution) were added and incubated overnight at 4 °C. The membrane was then washed three times with washing buffer (Bio-Rad 1× TBS with 0.1% Tween 20), incubated with the secondary antibody in blocking buffer for 60 min and then washed again three times with washing buffer. The membrane was then developed using Supersignal West Femto Maximum sensitivity substrate (Thermo Fisher Scientific) on a Chemidoc Imaging System (Bio-Rad).

Nano-FCM of EVs

Surface molecule expression was assessed by nano-FCM using a Flow NanoAnalyzer (NanoFCM Inc.) in accordance with the manufacturer’s guidelines. Prior to sample loading, the Flow NanoAnalyzer underwent a calibration process for alignment parameters as specified by the manufacturer. This alignment involved concentration/quality control (QC beads, NanoFCM Inc.), size beads (Silica nanospheres; 68–155 and 155–850 nm; NanoFCM Inc.), and a blank control. Per the manufacturer’s instructions, the measured events per minute were maintained below 12,000 events. To achieve this, samples were initially assessed, and if the count exceeded 2000 events in the first 15 s of measurement the sample was appropriately diluted to maintain counts between 11,000 and 12,000 events.

For immunofluorescent staining, 1 µL of the prediluted antibodies at a 6-fold dilution was added to 3 µL of the diluted EV sample, followed by a 40-minute incubation at room temperature in the dark. Subsequently, samples were diluted 50 fold with PBS and promptly loaded onto the nano-FCM for data acquisition. The antibodies used for the staining step were BD Pharmingen FITC mouse anti-human CD18 (Clone 6.7, BD Biosciences) and BD Pharmingen PE mouse anti-human CD11a (Clone HI111, BD Biosciences). The instrument parameters for the Flow NanoAnalyzer were configured as follows: laser power of 10 mW at 488 nm; laser power of 20 mW at 638 nm; SS decay at 10%; sampling pressure at 1.5 kPa; time to record at 1 min; and a 525/40 filter for FITC and a 488/30 filter for PE. The nano-FCM software (NF profession V1.0) was used to calculate the percentage of positive signal, particle concentration, and size distribution.

Proteomics—sample preparation

The amount of each sample for the proteomic analysis was normalized based on the protein content. Each sample contained 40 µg protein (9 samples), and a reference pool was also constructed with contributions from all samples containing 40 µg in total (4.44 µg/sample). Sodium dodecyl sulfate (SDS) was added to all samples to a final concentration of 2%.

The samples and reference pool were processed using a modified filter-aided sample preparation method [18]. In short, samples were reduced with 100 mM DTT at 60 °C for 30 min, transferred to Microcon-30 kDa Centrifugal Filter Units (Merck), and washed several times with 8 M urea and once with digestion buffer (DB; 50 mM TEAB and 0.5% sodium deoxycholate (SDC)) prior to alkylation with 10 mM methyl methanethiosulfonate in DB for 30 min at room temperature. Samples were digested with trypsin (Pierce MS-grade trypsin, Thermo Fisher Scientific, 1:100 ratio) at 37 °C overnight, and an additional portion of trypsin was added and incubated for another 2 h. Peptides were collected by centrifugation and labelled using tandem mass tag (TMT) 11-plex isobaric mass tagging reagents (Thermo Fisher Scientific) according to the manufacturer’s instructions. The samples were combined into one TMT-set and SDC was removed by acidification with 10% TFA. The TMT-set was further purified using a High Protein and Peptide Recovery Detergent Removal Spin Column and Pierce peptide desalting spin columns (both from Thermo Fischer Scientific) according to the manufacturer’s instructions prior to basic reversed-phase chromatography fractionation. Peptide separation was performed using a Dionex Ultimate 3000 UPLC system (Thermo Fischer Scientific) and a reversed-phase XBridge BEH C18 column (3.5 μm, 3.0 × 150 mm, Waters Corporation) with a gradient from 3 to 100% acetonitrile in 10 mM ammonium formate at pH 10.00 over 23 min at a flow of 400 µL/min. The 40 fractions were concatenated into 18 fractions, dried, and reconstituted in 3% acetonitrile and 0.1% trifluoroacetic acid.

Proteomics—nanoLC-MS/MS analysis and data base search

Each fraction was analyzed on an Orbitrap Lumos Tribrid mass spectrometer equipped with the FAIMS Pro ion mobility system interfaced with an nLC 1200 liquid chromatography system (all from Thermo Fisher Scientific). Peptides were trapped on an Acclaim Pepmap 100 C18 trap column (100 μm × 2 cm, particle size 5 μm, Thermo Fischer Scientific) and separated on an in-house-constructed analytical column (350 × 0.075 mm I.D.) packed with 3 μm Reprosil-Pur C18-AQ particles (Dr. Maisch, Germany) using a gradient from 3 to 80% acetonitrile in 0.2% formic acid over 85 min at a flow of 300 nL/min. Precursor ion mass spectra were acquired at 120,000 resolution, scan range 375–1375, and maximum injection time 50 ms. MS2 analysis was performed in a data-dependent mode, where the most intense doubly or multiply charged precursors were isolated in the quadrupole with a 0.7 m/z isolation window and dynamic exclusion within 10 ppm for 45 s. The isolated precursors were fragmented by collision-induced dissociation at 35% collision energy with a maximum injection time of 50 ms for 3 s (‘top speed’ setting) and detected in the ion trap, followed by multinotch (simultaneous) isolation of the top 10 MS2 fragment ions within the m/z range 400–1400, and further fragmentation (MS3) by higher-energy collision dissociation (HCD) at 65% collision energy and detection in the Orbitrap at 50,000 resolution, m/z range 100–500, and a maximum injection time of 105 ms.

The data files for each set were merged for identification and relative quantification using Proteome Discoverer version 2.4 (Thermo Fisher Scientific). The search was against Homo sapiens (Swissprot Database, April 2023, 20,422 entries) using Sequest as a search engine with precursor mass tolerance of 5 ppm and fragment mass tolerance of 0.6 Da and Sequest HTXCorr set to 2. Tryptic peptides were accepted with one missed cleavage, variable modifications of methionine oxidation, and fixed cysteine alkylation, and TMT-label modifications of the N-terminus and lysine were selected. Percolator was used for PSM validation with the strict FDR threshold of 1%. TMT reporter ions were identified with 3 mmu mass tolerance in the MS3 HCD spectra, and the TMT reporter abundance values for each sample were normalized to the total peptide amount. Only the quantitative results for the unique peptide sequences with a minimum SPS match of 65% and an average S/N above 10 were considered for the protein quantification. The reference samples were used as the denominator and for the calculation of the ratios. The quantified proteins were filtered at 1% FDR and grouped by sharing the same sequences to minimize redundancy.

Binding assay of LFA-1-expressing EVs to ICAM-1 cells

A binding assay was utilized to determine whether the membrane proteins CD11a and CD18 present on the EVs form a functional receptor. Wells in a 96-well plate (Thermo Scientific, Black microplate F96) were coated with EVs from the LFA-1 clone (LFA-1 EVs) by adding 10 µg/mL of the EVs in 100 µL PBS. Alternatively, wells were coated with EVs isolated from the CD18 clone (CD18 EVs) or EVs isolated from WT HEK293F cells (WT EVs), and LFA-1 EVs that had been pre-incubated with 100 µg/mL anti LFA-1 neutralizing antibody (InVivoMAb, Clone TS-1/22.1.1.13). PBS alone was added to the wells as negative controls. In these binding assays, labelled ICAM-1-expressing HMEC-1 cells were added to the wells. The HMEC-1 cells were treated with 15 ng/mL human tumor necrosis factor alpha (TNF-α; 210-TA-005, R&D systems) overnight to increase ICAM-1 expression on the cells. The washing and blocking of the plates and labelling of the cells were the same as the previous binding assay. After the labelled cells were added to the wells, the plate was incubated for 30 min at 37 °C and then washed and analyzed as for the previous binding assay.

Uptake experiment

100,000 HMEC-1 cells were seeded in a 24-well plate (Falcon) (0.5 mL). After the cells had been allowed to adhere (3–4 h), they were treated with 15 ng/mL TNF-α (210-TA-005, R&D systems) and incubated overnight. EVs were stained with the fluorescent dye DiO by incubation for 30 min at 37 °C. The next day the cells were washed with PBS and then incubated with an anti ICAM-1 neutralizing antibody (R&D Systems, clone BBIG-I1; using concentrations 0.3–30 µg/mL) or media for 90 min at 37 °C. Following this incubation the cells were again washed and incubated with either 5 × 109/mL DiO-stained EVs or 5 × 109/mL revesiculated EVs loaded with the fluorescent peptide and incubated for 30 min in the incubator. The wells were then washed twice with PBS and treated with 0.2 mL trypsin (Cytiva Hyclone) for 5 min at 37 °C. The trypsinated cells were then transferred to 15 ml Falcon tubes, diluted in 5 mL culture medium, and pelleted via centrifugation at 300 × g for 7 min. Following the centrifugation, the supernatant was removed, and the cells were resuspended in 500 µL fixation buffer (BD CellFIX) and incubated for 15 min at 4 °C. The fixed cells were washed using 1.5 mL FACS buffer (PBS with 1% FBS) followed by centrifugation at 200 × g for 10 min. After the centrifugation the supernatant was removed, and the cells were resuspended in 500 µL FACS buffer and analyzed via flow cytometry in which 10,000 data points were collected. The data were analyzed with FlowJo Software (Tree Star Inc.).

Generating and loading EVs via the open-and-close procedure

To improve the anti-inflammatory potential of the LFA-1 EVs, they were loaded with the anti-Myd88 anti-inflammatory peptide via an open-and-close procedure to generate revesiculated EVs as previously described [15]. Anti-inflammatory peptides targeting Myd88 were synthetized by JPT Peptide Technologies. Attached to the Cys side chain was a fluorescent tag (Fluorescein), and the peptide sequence was Arg-Asp-Val-Leu-Pro-Gly-Tr-Cys-Val-Asn-Ser-cholesterol. This cholesterol tag allows the peptide to associate with the EVs strongly.

The LFA-1 EVs (1011 EVs) were treated with a high pH solution and incubated at room temperature for 30 min, after which the membrane sheets were washed with PBS. After the washing, the membranes were resuspended in 400 µL PBS with 50 µg of the peptide. The membranes were incubated with the peptides at 37 °C for 30 min after which they were sonicated for 30 min at room temperature. Peptide-loaded revesiculated EVs were separated from non-loaded peptides through iodixanol-based ultracentrifugation, using a gradient consisting of 3 mL of 60% iodixanol mixed with the sample, 4 mL of 30% iodixanol, and 3 mL of 10% iodixanol, at 100,000 × g for 2 h (28,000 rpm, SW 41 Ti swing rotor, 265 as k-factor, Beckman Coulter). The vesicles located in the interphase between the 10% and 30% iodixanol layers were collected. A multimode microplate reader (Varioskan LUX, Thermo Fisher Scientific) was used to determine the peptide loading efficiency with excitation and emission wavelengths set at 495 nm and 520 nm, respectively. The quantification of peptides was determined based on a calibration curve, enabling accurate calculation of the peptide loading efficiency in the collected vesicles.

Evaluating the anti-inflammatory potential of LFA-1 EVMyd88

An assay was developed to determine how efficient LFA-1 EVMyd88 was at reducing inflammation in ICAM-1-expressing cells. HMEC-1 cells expressing ICAM-1 were seeded out onto a 24-well plate (Falcon; 100,000 cells per well in 500 µL medium) and incubated overnight. The culture medium was replaced with 500 µL fresh medium along with TNF-α (3 ng/mL; 210-TA-005, R&D systems) to induce high ICAM-1 expression and inflammatory responses in the HMEC-1 cells. The pre-treatment with TNF-α was 3 h in the incubator after which the cells were washed with PBS once and then treated with either WT EVMyd88 or LFA-1 EVMyd88 (2 × 109/mL; 10,000 EVs per cell), with anti-Myd88 peptide as the positive control (2 µg; corresponding amount of peptide which was loaded in EVs) or PBS and not loaded LFA-1 EVs (2 × 109/mL; 10,000 EVs per cell) as negative controls in 500 µL fresh medium for 30 min. Following this treatment, the cells were again washed with PBS and 500 µL fresh medium was added to the wells along with bacterial outer membrane vesicles (OMVs; 3 ng/mL) as a post-treatment to induce strong inflammatory responses. The OMVs were isolated as previously described using Escherichia coli strain DH5α [15]. The post-treatment lasted 6 h, and then the supernatant was collected and the concentration of the cytokine IL-8 was measured using a DuoSet ELISA Development kit (R&D Systems, Minneapolis, MN).