Blood extracellular vesicles from healthy individuals regulate hematopoietic stem cells as humans age

Abstract Hematopoietic stem cells (HSCs) maintain balanced blood cell production in a process called hematopoiesis. As humans age, their HSCs acquire mutations that allow some HSCs to disproportionately contribute to normal blood production. This process, known as age‐related clonal hematopoiesis, predisposes certain individuals to cancer, cardiovascular and pulmonary pathologies. There is a growing body of evidence suggesting that factors outside cells, such as extracellular vesicles (EVs), contribute to the disruption of stem cell homeostasis during aging. We have characterized blood EVs from humans and determined that they are remarkably consistent with respect to size, concentration, and total protein content, across healthy subjects aged 20–85 years. When analyzing EV protein composition from mass spectroscopy data, our machine‐learning‐based algorithms are able to distinguish EV proteins based on age and suggest that different cell types dominantly produce EVs released into the blood, which change over time. Importantly, our data show blood EVs from middle and older age groups (>40 years) significantly stimulate HSCs in contrast to untreated and EVs sourced from young subjects. Our study establishes for the first time that although EV particle size, concentration, and total protein content remain relatively consistent over an adult lifespan in humans, EV content evolves during aging and potentially influences HSC regulation.


Blood collection and plasma processing
The collection, handling and storage of blood extracellular vesicles followed guidelines outlined by the International Society for Extracellular Vesicles (ISEV) (Thery, C. (2018) J Extracell Vesicles. Nov 23;7(1):1535750). Peripheral blood was collected in the morning (9-11 am) from healthy subjects (20-85 yr) without fasting. For the isolation of plasma, blood was collected into K3E K3EDTA tubes (Greiner Bioone), inverted and processed immediately. Blood was centrifuged at 1880×g / 10min / room temperature (RT). Plasma was transferred to 50mL polypropylene centrifuge tubes and centrifuged at 2500×g / 10 min / RT to remove platelets. Plasma was frozen at -80°C. Coulter, Brea, CA, USA) for 2 h at 4°C. A visible high-density band between the 10% and 30% layers was collected. Exactly 1ml of the high-density band was loaded onto a SEC column, which was prepared using Sepharose CL-2B (Sigma, USA) to a final volume of 10 mL, equilibrated with PBS. Sample was subsequently eluted with 10mL PBS and 0.5 mL fractions were collected, with fractions 7-11 (EV-enriched fractions) pooled together and stored in cryovials at -80°C.

Determination of size and concentration of extracellular vesicles
Particle median size and concentration of the pooled fractions were measured with a ZetaView® -Nanoparticle Tracking Video Microscope PMX-120 instrument according to the manufacturer's instruction (Particle Metrix, Germany). All samples were measured in triplicate, using identical instrument settings. The chamber temperature was measured and considered when size and concentration was calculated. Camera sensitivity was set to 80 and the shutter was set to 100. Data were analysed using the ZetaView Analyze software system, with a minimum particle size of 2.5nm, a maximum size of 6000nm, a minimum brightness of 30, a maximum brightness of 255, a minimum area of 10, and a maximum area of 1000. To validate particle size, atomic force microscopy (AFM) was completed with a selection of samples. Sample concentration was calculated (incorporating dilution factors) and indicated as particles per mL of plasma.

Extracellular vesicle protein quantification
To quantitate protein, the Qubit™ Protein Assay Kit (ThermoFisher Scientific, USA) was used, in conjunction with the Qubit ® 3.0 Fluorometer (ThermoFisher, USA) as described in the manufacturer's instructions, standardized with PBS.

Preprocessing of umbilical cord blood and bone marrow
Umbilical Cord Blood (UBC) samples were obtained from healthy subjects minutes after caesarean section births at Kingston General Hospital, Canada. Umbilical cord was double clamped, and cord blood (4mL -50mL) was drawn up in a 30mL syringe attached to an 18G needle containing citrate dextrose solution (12.5mL of anticoagulant for every 50mL of cord blood). Subsequently, cord blood was diluted (1:1) with PBS/0.1% Human Serum Albumin.
Bone Marrow (BM) samples (42-80 yr) were obtained from either total knee arthroplasty or total hip arthroplasty surgeries. Fresh BM was transferred to K3E K3EDTA tubes and inverted. Samples were subsequently diluted with PBS / 0.1% Human Serum Albumin and centrifuged at 300g / 10min / RT. Post centrifugation, the upper fat layer was discarded, and the pellet was resuspended in 30mL of PBS / 0.1% Human Serum Albumin followed by two 300g / 7min /RT wash steps. After washes, the pellet was resuspended in 5-7 mL of PBS / 0.1% Human Serum Albumin depending on the original volume of bone marrow. ng/mL stem cell factor, 20 ng/mL of Interleukin-3, 20 ng/mL of Interleukin-6 and 20ng/mL granulocytecolony stimulating factor (PeproTech). EVs from the indicated age groups were added to CD34 + cells (final concentration of 2.3 X 10 4 -2.5 X 10 5 cells/mL) at a final EV concentration of 2.5-7.5 µg/mL in a round 96-well plate (total volume 200 µL) cultured for 48 hr at 37°C, 5% CO 2 .

Extracellular Vesicle Sonication
EVs were sonicated at high power for 20 second intervals on ice, with 1-minute rest, repeated 20 times and added to CD34 + cells immediately.

Colony forming cell (CFC) assay
After culturing cells under various conditions as indicated, cells were washed with PBS / 0.1% Human Serum Albumin and centrifuged at 300g / 10min / RT. After counting viable cells using trypan blue dye exclusion, approximately 1000 CD34 + cells were seeded into 1.2 mL of Methocult™ Medium H4435 Enriched (StemCell Technologies) and plated in 35mm plates using 16 gauge blunt-end needles followed by an incubation period of 10-14 days at 37°C and 5% CO 2 . Colonies were subsequently both quantified and qualified according to their morphology as lineage-committed progenitor colonies (CFU-GEMM: Colony To assess protein abundance between young, middle and older groups, the proteins were ranked based on the median level of expression with each group. To identify proteins that discriminate between three age groups, we used an established feature selection algorithm (Ren, R. et. al (2017). Oncotarget 8, 70982-71001) with leave-one-out cross-validation. From the top 5% of ranked proteins from each age group against all other samples 31 proteins were selected as the best discriminators and were visualized by hierarchical clustering using complete linkage and Euclidean distance for rows (proteins) and Spearman correlation for columns. The selected proteins were also used to show the age group clustering using t-Distributed Stochastic Neighbor Embedding (t-SNE)and Principal component analysis (PCA) plots. Kruskal Wallis -test was used to assess gender differences for the selected proteins. The top 5% of ranked proteins were used to identify the tissue of origin by linking with the human protein atlas (Uhlen, M. et al. (2015) Science 80-347, 1260419-1260419) for each age group.

Electron microscopy
For the negative-staining preparation, approximately 5µL of concentrated EV sample were loaded onto mesh carbon film grids (with glow discharge). The sample was left to adsorb onto the grid for 5 minutes, rinsed with distilled water, and stained with 2% uranyl acetate/H2O for 1min. After the sample preparations were air dried, they were examined using a JEM-2100 Electron Microscope with the follow settings: 80KV, Magnification is 50K to 80K.

Bead-based flow cytometry analysis of extracellular vesicle markers
To select for tetraspanin positive populations, anti-CD9, anti-CD63, or anti-CD81-coated superparamagnetic polystyrene beads (4.5µm) (Dynabeads®, Thermo Fisher Scientific) were used as per manufactures instructions (Thermo Fisher Scientific MAN0007670). For counterstaining of the captured extracellular vesicle populations, phycoerythrin (PE)-conjugated anti-CD9 antibody (mouse anti-human clone H19a) was used and compared with isotype-matched negative control (PE-mouse anti-human clone MOPC-21). Cells were analyzed using a CytoFLEX-S flow cytometer. Analysis of flow cytometry data was performed using FlowJo Software.

Western blot
To demonstrate the presence of protein markers associated with EVs, equal EV protein from young, middle-aged and old subjects were loaded and ran on 10% Bis-Tris polyacrylamide gels. For detection of CD63, non-reducing conditions were used. Membranes were blocked for 1 hour using 5% non-fat dry milk in 1X TBS-T (Tris Buffered Saline with Tween 20). Membranes were probed with anti-CD63 antibody (Purified mouse anti-human CD63; BD PharmingenTM) overnight at 4°C. Blots were developed using ECL luminol-based enhanced chemiluminescent substrate (UltraScence Pico Ultra Western Substrate; FroggaBio) and visualized using Azure c600 imaging system.

Apoptosis assay
After incubation with or without extracellular vesicles, CD34 + cells were stained with annexin-V (Allophycocyanin APC (Biolegend)) and DAPI (Sigma-Aldrich) in the presence of Hank's balanced salt solution (HBSS) (Millipore-Sigma) for 20 min at room temperature. Cells were immediately analyzed using a CytoFLEX-S flow cytometer. Analysis of flow cytometry data was performed using FlowJo Software Results are expressed as a mean ± SD.

Statistical analysis
Data analysis and graphical presentations were performed using GraphPad Prism version 8.0 (GraphPad Software, USA). Statistical analysis for the characterization of blood extracellular vesicles in terms of median size, protein concentration, and particle concentration between sex (male n=18, female n=17) was conducted using an unpaired t-test, and between age groups (young n=12, middle-aged n=11, older n=12) using ordinary One-Way ANOVA. CFC statistical analysis was conducted using mixed-effects analysis, with the Geisser-Greenhouse correction, along with Tukey's multiple comparisons test with individual variances computed for each comparison. Statistical analysis of blood extracellular vesicles in terms of median size, protein concentration, and particle concentration plotted against donor age included linear regression analysis (n=35). Data is displayed as mean ± SEM unless stated otherwise.