BM MSCs were isolated from young (n = 3, median age: 22 years) and aged (n = 3, median age: 69 years) healthy donors (Supplementary Table S1) after obtaining informed written consent (Ethical Approval No. EK221102004, EK47022007) as described previously (Oswald et al., 2004; Hempel et al., 2016). The cells were expanded in Dulbecco’s modified Eagle’s medium-low glucose (DMEM, Gibco, Germany) with 10% fetal bovine serum (FBS, Gibco) and characterized according to the criteria of the International Society for Cellular Therapy (Dominici et al., 2006). MSCs were used in the second or third passage for all experiments.

CD34⁺ cells were isolated from leukapheresis of patients (n = 9, 47–67 years old, median age: 52 years) suffering from non-leukemic cancers (e.g., myeloma, lymphoma and sarcoma) after obtaining informed written consent. Stem cells were collected before high-dose chemotherapy in preparation of an autologous HCT. The cells were purified by immunomagnetic sorting using CD34 MicroBead Kit UltraPure (Miltenyi Biotec, Germany). The viability was determined by live-dead-discrimination using Countess II Automated Cell Counter (Thermo Fisher Scientific, Germany) and the purity of at least 98% was confirmed by flow cytometry.

After reaching 80% confluency medium was changed to 100% DMEM. The medium was collected after 48 h and the cells were harvested and counted after staining with trypan blue for live-dead-discrimination. To deplete cell debris the medium was centrifuged at 380xg for 5 min and subsequently filtered through a 200 nm Millipore filter. Afterwards the EVs were isolated with the exoEasy Maxi Kit (Qiagen, Germany) and the EV-miRNA was purified with the exoRNeasy Maxi Kit (Qiagen, Germany) according to the manufacturer’s protocol. Briefly, the sample was mixed with the same volume of buffer XBP and given onto the column. After centrifugation at 500 g for 1 min at room temperature the column was washed with buffer XWP and centrifuged at 5.000xg for 10 min. To elute the EVs 400 μl of buffer XE was added onto the column before centrifugation at 500 g for 5 min. To increase the number of particles the column was eluted again and centrifuged at 5.000xg for 5 min.

To isolate the miRNA the EVs were bound on the columns analogously, but instead of eluting the particles were lysed in 700 μl Qiazol and the columns were spun down at 5.000xg for 5 min. The total RNA was isolated following the manufacturer’s protocol. The samples were stored at -80°C.

The data concerning size distribution and the concentration of the particles were acquired on the ZetaVIEW S/N 243 (Particle Metrix GmbH, Germany). The samples were diluted 1:500 or 1:1,000 in PBS. To process data ZetaView 8.05.05 SP2 was used.

Eluted EVs were lysed in 10X cell lysis buffer (Cell Signaling Technology, Germany) with protease inhibitor cocktail B (Santa Cruz, United States of America) and were mixed with Roti-Load 2 (Carl Roth, Germany). For reducing conditions DTT was added. 24 μl of the sample was loaded on a 10% or 12% SDS-polyacrylamide electrophoresis gel and transferred to a nitrocellulose membrane. Subsequently, it was blocked in 5% non-fat dry milk and incubated with anti-CD63 (Invitrogen, Germany) and anti-CD81 (Invitrogen, Germany) under non-reducing conditions, anti-Flotillin-1 (BD Biosciences, United States of America) and anti-Calnexin (Cell Signaling Technology, United States of America) under reducing conditions. Afterwards, the membrane was thoroughly washed and incubated with the secondary antibody coupled with a horseradish peroxidase. The chemiluminescence was detected using the ECL Plus Western Blotting Detection Reagent (Amersham, United States of America) on the Luminescent Image Analyzer LAS-3000 (FUJIFILM, Japan).

To prepare the samples for Transmission Electron Microscopy (TEM) EV solutions were applied on a 300 mesh EM grid, airdried and fixed with 1% glutaraldehyde. The samples were washed twice with water and counterstained with uranyloxalate solution. Afterwards the EVs were examined under a transmission electron microscope Jeol JEM1400 Plus (Jeol, Germany) running at 80 kV acceleration voltage.

The miRNA amount of the EV samples was calculated with the Qubit™ microRNA Assay Kit (Thermo Fisher Scientific, Germany). cDNA was synthesized using TaqMan^® MicroRNA Reverse Transcription Kit (Thermo Fisher Scientific, Germany). RT-PCR was performed using TaqMan^® universal PCR Master Mix (Thermo Fisher Scientific, Germany) on a Quantstudio three cycler (Applied Biosystems, Germany). The following miRNAs were examined: let7a, miR-10a, miR-21, miR-23, miR-155, miR-221, miR-222 and miR-486, miR29a, miR34a. All primers were purchased from Applied Biosystems (Supplementary Table S2). The samples were run in duplicates. U18 was used as endogenous control according to Vriens et al. (Vriens et al., 2012). Amplicons were normalized to U18 applying the comparative CT (∆CT) method.

2.4 × 10⁵ CD34⁺ HSPCs were incubated with young or aged MSC-EVs collected from 1.2 × 10⁵ cells or XE buffer as control in CellGenix SCGM (CellGenix, Germany) with 10% vesicle-depleted FBS supplemented with SCF, Flt3 ligand and IL-3 (10 ng/ml each, all from Miltenyi Biotec, Germany). The EV samples were pooled beforehand. For vesicle depletion, the FBS was centrifuged with an Amicon-Ultra 100 kDa filter (Millipore, Germany) for 55 min at 3.000xg. Each sample was run in duplicates.

After the incubation, the cells were washed and counted by Countess II Automated Cell Counter (Thermo Fisher Scientific, Germany). Subsequently, samples were pooled and further analyzed using flow cytometry, clonogenic assays and RT-PCR.

HSPCs cultured with EVs were stained for surface markers using fluorescently labeled antibodies according to Supplementary Table S3. Corresponding human immunoglobulin G controls were used. Data was acquired on a BD FACS Calibur or LSRII. Data were analysed using FlowJo software (FlowJo, LLC, United States of America).

Isolated EVs were labeled with fluorescent dye 10 µM DiI (Thermo Fisher Scientific). After incubation for 20 min at 37°C the dye was removed with Exosome Spin Columns (Thermo Fisher Scientific). 10⁵ CD34⁺ cells were incubated with DiI-EVs of 3*10⁵ MSCs or buffer as control for 24 h. Afterwards, a cytospin was used to concentrate the cells onto a microscope slide. After fixation in 4% paraformaldehyde, the nuclei were counterstained with DAPI. Images were acquired on a LSM 880 (ZEISS, Germany).

Colony‐forming unit (CFU) assays were carried out using CD34⁺ HSPCs harvested after 3 days of incubation with EVs. 500 cells were plated in Stem MACS HSC-CFU complete with Epo (Miltenyi Biotec, Germany). Colonies were counted after 2 weeks and classified with the StemVision system (Stem Cell Technologies, Germany). Each sample was run in duplicates.

Total RNA were isolated from CD34⁺ HSPCs after incubation with EVs using RNeasy Micro Kit (Qiagen, Germany) and reverse transcribed into cDNA using RevertAid cDNA synthesis kit (Thermo Fisher Scientific, Germany) with oligo-dT primers. Relative target quantity was determined using the comparative CT (∆∆CT) method. RT-PCR was performed using SYBR Green/ROX PCR master mix (Thermo Fisher Scientific, Germany) and target specific primers for PTEN, CDKN2A and SIRT1 (Supplementary Table S4) on a Quantstudio three cycler (Applied Biosystems, Germany). Amplicons were normalized to endogenous GAPDH control. All samples were run in duplicates.

NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NOD/SCID/IL2rc2/2, NSG) mice were purchased from The Jackson Laboratory (Jackson Laboratory, United States of America). They were kept in the animal facility at the Medical Theoretical Center of the University of Technology Dresden in accordance with German animal welfare legislation after approval of the Landesdirektion Sachsen (TVV 25/2015). Female mice (n = 3 per group) in the age of 8–9 weeks were used for the experiments. CD34⁺ HSPCs were pre-incubated with EVs as described the in vitro studies. 1 × 10⁵ cells were injected intravenously after whole body irradiation with 1 Gy. For the first 3 weeks after transplantation, water was supplemented with neomycin (1.17 g/l, Sigma-Aldrich, United States of America). Every 4 weeks peripheral blood was taken and after 16 weeks the animals were sacrificed and bone marrow was collected. All samples were stained with monoclonal antibodies (Supplementary Table S3) after lysis of red blood cells.

Values were summarized as means, p values < 0.05 were considered significant. Statistical analyses including Student’s t-Test and One-way ANOVA were calculated with GraphPad Prism 5.00 for Windows.

After incubating a nearly confluent MSC layer in serum-free DMEM for 48 h, EVs were collected with a column-based isolation method. The MSC viability was proved to be over 95% in all samples. As a reference for EV concentration, the cell number was determined after trypsinization and was 2.4–3.3 × 10⁶ MSCs (median 2.6 × 10⁶).

NTA revealed no significant differences in concentration and size distribution between young and aged MSC-EVs. Particles larger than 400 nm could not be detected. The mean of the median diameter of the young and old MSC-EVs (n = 3 per group) was 170.7 ± 8.4 and 174.0 ± 11.0 nm, respectively (Figure 1A). The mean particle concentration was 2,83 ± 0,28 and 2,41 ± 0,49 *10³ particles per MSC.

Next, the extracted EVs of young and old MSCs were characterized by Western Blot. The exosomal proteins CD63, CD81 and Flotillin-1 could be detected in all samples, whereas Calnexin, a typical cellular marker, was absent in all samples but present in the cell lysate (Figure 1B). The Transelectron Microscopy showed round vesicular structures with a hypodense center in isolates from both young and aged MSCs. The typical bilayer membrane could be observed. There were no obvious differences in the ultrastructure and the morphology between EVs from young and aged MSCs (Figure 1C).

MSCEVs can transfer the cargoes to the recipient cells and thereby alter their activities.

Since small RNAs are highly abundant in EVs, the miRNA cargo was isolated and the concentration determined by Qubit analysis. Interestingly, the total amount of miRNA was significantly decreased in EVs derived from aged compared to young MSCs (Figure 2A). In a next step a panel of miRNAs relevant for hematopoiesis were quantified by PCR. U18 was used as endogenous control. The Ct value in the group of young MSC-EVs was 29.16 ± 0.96 and in the group of aged MSC-EVs 30.43 ± 0.86 confirming stable U18 expression in both sample cohorts. Whereas no relevant differences could be detected for let-7a, miR-10a, miR-21, miR-23, miR-155, miR-221, miR-222 and miR-486, the levels of miR-29a and miR-34a were significantly higher in EVs from aged MSCs (Figures 2B,C). This suggests an accumulation of specific miRNAs in aged EVs whereas the overall amount was decreased.

To investigate whether MSC-derived EVs can be incorporated by HSPCs, confocal microscopy was applied. After incubation of HSPCs with Dil labeled EVs for 24 h, an uptake could be detected independent on the age of the parental MSCs. Unstained EVs and buffer, respectively, were used as control (Figures 3A,B).

To analyze the impact of incorporated EVs on HSPC characteristics in more detail, the incubation time was increased for up to 72 h. Aged HSPCs (n = 4) were treated with EVs of young or aged MSCs, respectively, or with XE buffer only as control. EVs derived from young MSCs caused a significant increase in the cell number compared to cells treated with buffer (1.2-fold) or with old EVs (1.8-fold) (Figure 3C). Furthermore, the viability of HSPCs was significantly improved after incubation with young EVs (Figure 3D).

Incorporation of EVs and miRNA release can modulate gene expression of HSPCs. Therefore, the expression of the tumor suppressor gene PTEN, a known target of mir-29a, was decreased in HSPCs treated with aged EVs (Figure 4A). Interestingly, the expression of CDKN2A, encoding for the cell cycle inhibitors p16^INK4A and p19, was significantly lower in HSPCs incubated with aged MSC-EVs. In contrast, the gene was upregulated in cells treated with young MSC-EVs (Figure 4B). SIRT1, encoding for the histone deacetylase Sirtuin-1 that regulates a balanced hematopoietic differentiation, is upregulated in both groups receiving EVs (Figure 4C).

Moreover, expression of p53, FOXO3, AKT1, Beclin-1 and ATG12 were investigated. However, no significant differences were detected in the expression levels of these genes.

Surface marker expression is an important indicator of differentiation and stemness of HSPCs and contributes to the definition of the functional properties, e.g., after transplantation. Therefore, the expression of the proteins CD38 and CD90 on CD34⁺ population was investigated (Figure 5A). We could not detect significant changes in the absolute expression of these surface markers owing to relevant inter-individual differences. The mean fluorescence intensity (MFI) of the stemness marker CD90 normalized to the respective donor was increased after treatment with MSC-EVs with no respect to the donor’s age (Figure 5B). However, the normalized MFI of CD38, a protein expressed on committed progenitors, was also significantly higher on HSPCs treated with young EVs too (Figure 5C).

To evaluate the clonogenic potential of HSPCs in vitro after EV treatment, colony-forming unit (CFU) assays were performed. The HSPCs of each group were able to differentiate into each type of colony. HSPCs treated with aged MSC-EVs showed a significant higher number of BFU-E/CFU-E colonies. Furthermore, the number of CFU-GM colonies was reduced in the group of HSPCs treated with young MSC-EVs compared to the control group (Figure 5D).

To prove the beneficial effect of MSC-EVs on the engraftment potential of HSPCs in vivo, xenotransplantation assays in immunodeficient NSG mice were conducted. 1 × 10⁵ pretreated HSPCs were transplanted. Long-term engraftment could be observed in all three groups assessed by the de novo generation of myeloid (CD33⁺) and lymphoid (CD3⁺ and CD19⁺) cells (Figure 6A). There were no significant differences concerning the phenotype and the compartmental attribution of the blood cells neither in the spleen and the peripheral blood nor in the BM (Figure 6B).

The HSPCs treated with aged MSC-EVs showed a significantly higher week 12 peripheral blood chimerism (Figure 6C). The chimerism in bone marrow and spleen was also evaluated after 16 weeks. No significant differences could be observed, although the chimerism was always higher after EV treatment with no respect to the donor’s age (Figures 6D,E).